Advances in Agronomy, Volume 21

ADVANCES IN AGRONOMY VOLUME 21 CONTRIBUTORS TO THIS VOLUME F. J. CARLISLE D. J. GREENLAND R. B. GROSSMAN S. B. HEA...

Author: N. C. Brady

21 downloads 1093 Views 18MB Size Report

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

Report copyright / DMCA form

DOWNLOAD PDF

ADVANCES IN

AGRONOMY VOLUME 21

CONTRIBUTORS TO THIS VOLUME

F. J. CARLISLE D. J. GREENLAND

R. B. GROSSMAN S. B. HEATH CHARLES E. KELLOGG

OLIVER E. NELSON J. M. OADES ARNOLD C . ORVEDAL W. F. RAYMOND

G. D. SWINCER R. W. WILLEY

ADVANCES IN

AGRONOMY Prepared under the Auspices of the

AMERICAN SOCIETY

OF

AGRONOMY

VOLUME 21

Edifed by N. C. BRADY Roberts Hall, Cornell University, ithaca, New York

ADVISORY BOARD

R. R. DAVIS F. A. HASKINS W. D. KEMPER

.I.P. MARTIN

J . W. PENDLETON W. A. RANEY 1969

@

ACADEMIC PRESS 0 N e w York and London

COPYRIGHT ‘C 1969,

BY

ACADEMIC PRESS, INC.

ALL RIGHTS RESERVED. N O PART OF THIS BOOK MAY BE REPRODUCED IN ANY FORM, BY PHOTOSTAT, MICROFILM, RETRIEVAL SYSTEM, OR ANY OTHER MEANS, W I T H O U T WRITTEN PERMISSION FROM T H E PUBLISHERS.

ACADEMIC PRESS, INC. 1 1 1 Fifth Avenue, New York, New York 10003

United Kingdom Edition published by ACADEMIC PRESS, INC. (LONDON) LTD. Berkeley Square House, London W 1X 6BA

LIBRARY OF

CONGRESS CATALOG C A R D

NUMBER50-5598

PRI N TED IN T H E UNITED STATES O F AMERICA

CONTRIBUTORS TO VOLUME 21 Numbers in parentheses indicate the pages on which the authors’ contributions begin.

F. J . CARLISLE ( 2 3 7 ) , Soil Conservation Service, United States Department of Agriculture, Hyattsville, Maryland D. J . GREENLAND (195), Depurtment of Agricultural Biochemistry and Soil Science, Waite Agricultural Research Institute, University of Adelaide, South Australia R. B. GROSSMAN ( 2 3 7 ) , Soil Conservation Service, United States Department of Agriculture, Lincoln, Nebraska S . B. HEATH(28 11, Department of Agriculture, University of Reading, Reading Berkshire, England CHARLES E. KELLOGG( I09), Soil Survey, Soil Conservation Service, United States Department of Agriculture, Washington, D.C. OLIVER E . NELSON* (17 l ) , Purdue University, Lafayetre, Indiana J . M . OADES(195), Department of Agricultural Biochemistry and Soil Science, Waite Agricultural Research Institute, University of Adelaide, South Australia ARNOLDC. ORVEDAL ( 109), Soil Survey, Soil Conservation Service, United States Department of Agriculture, Washington, D.C. W. F . RAYMOND( I ) , The Grassland Research Institute, Hurley, England G . D. SWINCER (195), Department of Agricultural Biochemistry and Soil Science, Waite Agricultural Research Institute, University of Adelaide, South Australia R. W. WILLEY t (28 I ) , Department ofAgriculture, University of Reading, Reading Berkshire, England

*Present address: University of Wisconsin, Madison, Wisconsin. t Presenr address: Makerere University College, Kampala, Uganda. V

This Page Intentionally Left Blank

PREFACE

This volume marks a significant milestone in the history of Advances in Agronomy. The first 20 volumes were compiled under the very capable editorship of Dr. A. G. Norman, now Vice-president for Research at the University of Michigan. Mounting pressures of other responsibilities have prompted Dr. Norman to ask to be relieved as editor of this serial publication; this is the first volume that does not carry his name. It is fitting that we reflect briefly on the contributions he has made, not only to this work but to his profession as well. As editor of Advances in Agronomy, Dr. Norman has given 20 years of faithful service and leadership to agronomists and soil and crop scientists throughout the world. His extraordinarily good judgment in the selection of authors and of subject matter has been largely responsible for the success of this publication. His guidance to authors has helped both them and the quality of their papers. He has seen Advunces in Agronomy grow from a struggling review journal of concern only to American scientists to a prominent review series with contributors and subscribers in many nations. Dr. Norman has found other ways to benefit his profession. He has contributed directly as an active researcher in soil microbiology and in soil and plant biochemistry. He has served as director of a large, interdisciplinary research unit and has enriched the education and training of many soil and crop scientists as well as biologists. We are also indebted to Dr. Norman for his service in scientific societies. He served as vice-president and later president of the American Society of Agronomy during a very critical period in the Society’s history. In addition, for a period of three years he served as chairman of the Division of Biology and Agriculture of the National Research Council. Even though Dr. Norman has resigned his editorial responsibilities, Advances in Agronomy fortunately will reflect his influence for some time to come. The challenge of maintaining Dr. Norman’s high standards and the broad subject matter coverage he provided is materially aided by the efforts of such men as the eleven who have contributed to this Volume 2 I.

N. C. BRADY Ithaca, New York August, I969

vii

This Page Intentionally Left Blank

CONTENTS CONTRIBUTORS TO VOLUME 21

.

.

.

.

.

.

.

.

.

.

.

PREFACE .

.

.

.

.

.

.

.

.

.

.

.

vii

2 3

.

.

.

.

V

THE NUTRITIVE VALUE OF FORAGE CROPS W . F . RAYMOND

. . . . . . . . . . . I . Introduction . . . . . . . I 1 . The Components of Nutritive Value . . . . . . 111. The Digestibility of Forage Crops . . . . . IV . The Digestibility of Different Forage Species . . . . . . V . The Voluntary Intake of Forages . . . VI . The Efficiency of Utilization of Digested Nutrients . VII . The Relationship between Forage Quality and Forage Yield . . . . VIII . Forage Breeding for Improved Nutritive Value . IX . The Effects of Processing on the Components of Forage . . . . . . . . . . Nutritive Value . X . The Nutritive Value of Grazed Forage . . . . . . References . . . . . . . . . . . .

.

.

.

.

. . .

. . .

.

.

. .

. .

65

4 16

27 38 68

.

.

.

.

69 80

.

.

97

. . . . .

. . . . .

POTENTIALLY ARABLE SOILS OF THE WORLD A N D CRITICAL MEASURES FOR THEIR USE CHARLES E. KELLOGGA N D ARNOL.D C . ORVEDAL

I . Introduction . . . . . . . . I 1 . The Principle of Interactions in Soil Use . 111 . Higher Production from Existing Arable Soils IV . New Potentially Arable Soils . . . . References . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

109

I12 122 140 169

GENETIC MODIFICATION OF PROTEIN QUALITY I N PLANTS OLIVERE . NELSON

I . lntroduction . . . . . . . . . . . . I I . The Genetic Control of Protein Structure . 111 . The Relative Constancy of Leaf Protein Composition

ix

. .

. .

. .

.

.

.

.

.

.

171 173 177

X

CONTENTS

IV. The Storage Proteins of Seeds . . . . V. Theopaque-2 andJloury-2 Mutations in Maize VI. The Prospects of Improvements in Other Plants VI1. Summary . . . . . . . . References . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

178 I80 I87 I90 191

THE EXTRACTION, CHARACTERIZATION, A N D SIGNIFICANCE OF SOIL POLYSACCHARIDES G. D. SWINCER, J. M. OADES,A N D D. J. GREENLAND I. Soil Carbohydrates . . . . . . . . . . 11. The Significance of Soil Polysaccharides . . . . . I l l . Studies on Soil Polysaccharides . . . . . . . IV. Methods for the Analysis of Complex Polysaccharide Materials V. Summary and Conclusions. . . . . . . . . References . . . . . . . . . . . .

.

.

. .

. .

.

.

. .

. .

195 196 I99 222 229 230

FRAGIPAN SOILS OF THE EASTERN UNITED STATES R. B. GROSSMAN A N D F. J. CARLISLE 1. 11. 111.

IV. V. VI. VII. VIII. IX.

x.

Introduction . . . . . . Horizons of Fragipan Soils . . Occurrence of Fragipan Soils . . Properties of Fragipans . . . Fragipans and the Soil Water Regime . . . Genesis of Fragipans . Fragipans and Soil Use . . . Classification of Fragipan Soils. . Unresolved Problems. . . . Summary . . . . . . References . . . . . . Appendix . . . . . .

. . .

. . .

. . .

. . .

. . .

. . .

. . .

. . .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

. .

. .

. .

. .

. .

. .

. .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . . . .

231 240 244 246 254 256 263 265 269 27 I 272 276

THE QUANTITATIVE RELATIONSHIPS BETWEEN PLANT POPULATION A N D CROP YIELD R. W. WILLEYA N D S. B. HEATH

I. 11.

Introduction . . . . . . . . . . Relationships between Plant Density and Crop Yield.

.

.

.

.

.

.

.

.

28 I 283

xi

CONTENTS

I l l . The Relationship between Plant Rectangularity IV. The Variation in Yield of the Individual Plant . . . . . . . V. Conclusions . . . . . . . . References .

AUTHORI N D E X . SUBJECT INDEX .

. .

. .

. .

. .

. .

. .

. .

. .

. . .

. . .

. . .

. . .

. . . .

.

. . .

3 14 317 319 3 20

. .

. .

. .

. .

. .

. .

323

and Crop Yield

338

This Page Intentionally Left Blank

THE NUTRITIVE VALUE OF FORAGE CROPS W. F. Raymond The Grassland Research Institute, Hurley, England

I. 11. 111.

IV.

V.

VI.

VII. v111. IX.

X.

Page Introduction. ...... ........................................................... 2 The Components of Nutritive Value .................................. .... 3 The Digestibility of Forage Crops ........... .................. 4 A. The Measurement of Digestibility in .................. 4 B. The Prediction of Forage Digestibility from Chemical Composition ....... 6 .... 7 C. Improved Chemical Techniques ..... . . . . . ................... ..... D. Estimation of Forage Digestibility by in Vitro Techniques ................. 10 E. The Relative Utility of Chemical and in Virro Estimations of Forage Digestibility _ . _ _____ . _,_. .._ _ _ _ ___._.__..__, _ _ _ _ . .._....._.___. ._.......... 16 The Digestibility of Different Forage Species ..._...._.... ... . ..... 16 16 A. Basic Patterns of Digestibility ....................... B. The Digestibility of Different Plant Fractions .. _........_____..._. 20 C. The Effect of Environmental and Other Factors on ...... ......... Forage Digestibility , . , . , . . . . . . . _ ._ . _ _ _ _ _._. ._. . _. . ,_..., __............. 23 .............................................. 27 The Voluntary Intake of Forages 27 A. The Factors Controlling Feed Intake ........................... 28 B. Intrinsic Factors Determining Forage Intake ..._.. .........._ 35 C. The Nutritive Value Index ........................................................... D. The Crude Protein Content of Forage and Voluntary Intake .............. 36 E. The Effect of Supplementary Feeds on Forage Intake ..... 37 The Efficiency of Utilization of Digested Nutrients. .................. 38 A. Methods of Expressing Energy Values ....... . , , .. .. .. ................ . .. ........ 38 B. The Role of Volatile Fatty Acids in Ruminant 40 C . The Utilization of the Crude Protein in Forages .................. 48 D. The Use of Nonprotein Nitrogen in Ruminant 51 E. The Mineral Nutrients in Forages ................................................. 54 F. Pharmacologically Active Components in Forage .................. 61 The Relationship between Forage Quality and Forage Yield ...... .............. 65 Forage Breeding for Improved Nutritive Value ....... . . ........... ..... 68 The Effects of Processing on the Components of Forage Nutritive Value , . _.. .._,, . . . _ .. . . . . . , , _ _ _...... . . ..., . . . . . ............. .............. 69 .......................................... 69 B. The Grinding and Pelleting of Dehydrated Forages ........ 71 C . The Ensiling of Forage Crops .......... ................................... 74 The Nutritive Value of Grazed Forage .... ................................... 80 A. Measurement of the Nutrient Intake by Grazing Animals ..... 80

2

W. F. RAYMOND

B. The Botanical Composition of Grazed Forage ................................. C. The Nutrient Intake of Grazing Livestock ...................................... D. The Effect of Management on the Productivity of Grazing Animals ........................................................................ References .......................................................................................

90 91 94 97

1. Introduction

Forages are grown for ruminant feeding, most ruminant animals eat forages. Thus a review of the nutritive value of forages is essentially a review of ruminant nutrition, yet with the difference that the nutritionist can treat the animal and the forage it eats in isolation, whereas the agronomist must also consider the problems that arise when the animal and its feed are brought together in practical systems of forage production and utilization. Ruminant feeding to date has been a nonintensive system of land use, in comparison with crop farming or the feeding of nonruminant livestock. This has been justified on the basis that forages are cheap to grow, and that harvesting by grazing is a cheap method of utilization. However, as Melville ( 1960) emphasized, present extensive pastoral systems produce very low outputs of human food per acre; as world demand for food increases, pastures will either have to become markedly more productive, or be replaced by crops that can be used by nonruminants, or directly by humans. In the latter cases forage-ruminant systems would be confined to noncultivable agricultural areas, and so would contribute only marginally to the nutrition of the world’s population. At that time of apparent food surpluses the replacement of ruminant products seemed remote; today analogue substitutes are already serious competitors to meat and milk in North America, the most sophisticated consumer market in the world. To date this competition has been in terms of cost and convenience; in future it will increasingly be in terms of competition for land, as foreseen by Melville. This means that the efficiency of soil-forage-ruminant systems must be greatly increased if they are to continue as a significant sector of agriculture. Raymond (1968) has considered this problem in terms of (a) the efficiency of use of incident light energy by the growing plant, (b) the proportion of the energy in the plant which is actually eaten by the ruminant animal, and (c) the efficiency with which different animal populations convert the energy they eat into products which can be used by humans. In many cases it appears that stages (b) and (c) are the main factors limiting the output of ruminant products per acre; until we can

THE NUTRITIVE VALUE OF FORAGE CROPS

3

ensure that a high proportion of the forage grown is eaten by efficient animals, there may be little advantage in concentrating effort on growing more forage. Efficiency of feed conversion (c) depends on many factors, including the structure of the animal population (the proportion of adult breeding animals to productive offspring; Spedding, 1965) and the genetic potential of the animals. But a dominant factor is the level of nutrient intake of the animals being fed: the higher the level of nutrient intake, the higher the level of productivity of the animals, and the lower the nutrient requirement for each unit of animal output. Thus, as the daily nutrient intake of the 300-kg. steer increases from 15.3 to 20.3 Mcal. of metabolizable energy, its daily rate of liveweight gain increases from 0.5 to 1.25 kg. per day: the corresponding requirement of metabolizable energy per kilogram gain decreases markedly from 30.6 to 16.2 Mcal. (Raymond, 1968). II. The Components of Nutritive Value

Thus the nutritive value of a forage should be considered not as a single parameter, but as composed of a complex of parameters that determine the nutrient intake of ruminant animals fed on that forage. In this it differs from the classical concept of nutritive value as a feed concentration (starch equivalent, total digestible nutrients, or net energy) by including feed intake as an integral component of nutritive value. With the major economic feeds of earlier feeding systems (cereals and pulses, oilseed residues, and industrial by-products) this was not necessary, as the quantity eaten was controlled by rationing; with forages, on the other hand, there is seldom any formal control of the amount eaten, which therefore depends on factors in the forage and in its method of presentation. This review therefore considers the nutritive value of forages in terms of the factors that determine the level of nutrient intake by ruminant livestock. It has proved useful to treat nutrient intake as the product of three parameters (Raymond, 1969b): Nutrient intake

= intake X

of feed x digestibility of feed efficiency of utilization of digested feed

(1)

each of which can be investigated separately, before their interactions in practical systems of ruminant feeding are considered. The importance of this approach is indicated by the conclusion of lngalls er al. ( 1 965) that 70 percent of the variation in production potential between forages can be accounted for in terms of differences in voluntary intake, compared with 30 percent by differences in digestibility, the nutrient concentration

4

W. F. RAYMOND

measure of earlier systems. Equation (1) is also possibly more informative than the analogous nutritive value index (Crampton et al., 1960), which combines intake and digestibility in one parameter, so that the relative importance of these two factors is not immediately evident. However, nutritive value and nutrient intake can have no real meaning except in relation to the needs of ruminant animals. Different animals have different nutritional requirements, depending on their species, sex, physiological status, and level of production, and this means that the nutritional adequacy of a forage diet can be assessed only in terms of the nutritional needs of the particular animals to be fed. The requirements by different classes of stock for energy, protein, minerals, and vitamins have been tabulated (Morrison, 1957; Agricultural Research Council, 1965; National Academy of Sciences, 1966). The objective must then be to establish relevant parameters to describe the nutritive value of forages which can be equated with these nutrient needs. The components in Eq. (1) provide a framework within which to assess our current knowledge of these nutritional parameters. Of these components, the digestibility of forage is considered first, because of the important influence which digestibility exerts on the other two components, intake and efficiency of utilization. These components are discussed in relation to fresh forages, but particular emphasis is then given to the effects of processing methods, feed interactions, and methods of feeding, all of which can markedly alter the basic nutritional features of forages. The practical aim must be to exploit this new information so as to improve the nutritional potential of forage feeding systems, and the effectiveness with which soil-forage-ruminant systems can compete for the world’s increasingly scarce land resources. Ill. The Digestibility of Forage Crops

A. THEMEASUREMENT OF DIGESTIBILITY in

ViVO

The digestibility of a feed is defined: Digestibility = ‘Ieed

- cfses

X

100

cfeed

Where Cfed is the amount of feed or feed component eaten (organic matter, cellulose, protein), and Cfecesis the corresponding amount of fecal excretion. The measurement of digestibility requires a preliminary feeding period during which the experimental animals adapt to the feed under test, followed by a test period, during which feed eaten and fecal output are measured. For precise measurement preliminary and test

THE NUTRITIVE VALUE OF FORAGE CROPS

5

periods of at least 10 days are recommended (Raymond et al., 1953); this presents the particular difficulty with fresh forages that the forage must be cut daily, and so may change in digestibility and chemical composition during this experimental period. In many studies this day-to-day variation in feed characteristics has been overcome by cutting at one time sufficient fresh forage for the complete digestibility experiment, and preserving this forage so that it can be fed over an extended period. Storage as hay (J. R. Jones and Hogue, 1963) or after artificial drying (Kivimae, 1959) has been used but, because of the changes in digestibility possible with these methods, cold storage of forages has been adopted by some workers (Raymond et al., 1953; Pigden et al., 196 1 ; Minson, 1966). The technique of storage at 5°F. has been described in detail (Commonwealth Agricultural Bureaux, 1961, pp. 88 and 150); it has been shown to have a negligible effect on the digestibility of the dry matter or organic matter in forage (Raymond et al., 1953) or of the rate of digestion within the rumen (Pigden et al., 1961), but slightly reduces the digestibility of the crude protein fraction (Raymond et al., 1953; Minson, 1966). An alternative technique, the continuous digestion trial with fresh forage, is now being increasingly widely used (Greenhalgh er al., 1960; Commonwealth Agricultural Bureaux, 196 1 ; Ademosum et al., 1968). Herbage is cut and fed daily over an extended period, and the amounts of forage eaten and feces voided are measured daily throughout the experiment. The amounts of forage eaten and feces are summed over 5day subperiods, allowing a 2-day lag for passage of the feces, and digestibility coefficients are calculated on these subperiods, each of which serves as the preliminary (adaption) treatment for the succeeding subperiod. This technique has proved of particular use in association with grazing experiments (see Section X,A,2), but it is less accurate than the cold-storage technique because of the shorter balance periods used. The measurement of the digestibility of forages conserved by natural or artificial dehydration presents no such problem, and most of the reported data on forage digestibility relate to such feeds. Silage is generally removed daily from silos for feeding, but cold storage of silage (Harris and Raymond, 1963) requires less labor, and eliminates any risk of day-to-day variation in silage quality. The many thousands of recorded determinations of forage digestibility have been collated at intervals and provide the broad background to our present understanding of forage nutritive value (Schneider, 1952; Leitch, 1969; tropical forages, Butterworth, 1967). However, such compila-

6

W. F. RAYMOND

tions may be of limited value in indicating the digestibility of an “unknown” forage, because of the difficulty of identifying it with a particular feed class. This problem, long recognized, led to the development of techniques such as the Weende feed analysis for estimating nutritive values; a major advance in the period under review has been in the development of improved laboratory techniques for predicting the nutritive value of forages, to replace wherever possible the laborious and expensive in vivo determination.

B. THE PREDICTION OF FORAGE DIGESTIBILITY FROM CHEMICAL COMPOSITION As Van Soest (1968) has noted, animal nutrition has had a history of inertia and complacency, each further experiment carried out with old techniques and old terminologies making it yet more difficult to adopt new ones. But it is still difficult to create a logical pattern from the torrent of new analytical techniques and new parameters of nutritive value that have recently been put forward to replace these older concepts. The requirement is to establish a relationship between a nutritional parameter (e.g., digestibility) of forages, measured in controlled in vivo experiments, and the chemical composition of the same forages, from which the nutritive value of other forages can be predicted. Digestion of forage by the ruminant is a most complex process; yet for nearly a century the attempt was made to predict the extent of forage digestion in terms of its proximate analysis based on Weende crude fiber, crude protein and nitrogen-free extractives. Sullivan (1 962) and Dijkstra ( 1 966) have both shown that when these analyses are applied to a limited range of forages close relationships between digestibility and chemical composition can be established, but that these relationships become less precise as the range of forages included is increased. As a forage crop matures its fiber content increases and it becomes less digestible; a close negative relationship between fiber content and digestibility is found. But this relationship is likely to differ from that with a different forage species (in particular, tropical forages; Butterworth, 1963) or from that with the same forage species at a different time of year; in each case the forage becomes less digestible as it becomes more fibrous, but at a given fiber content different forages can have markedly different levels of digestibility. To some extent this can be overcome by using tabulations of relationships, each based on a limited feed class (Dijkstra, 1966). But again these pose the problem of allocation to a particular feed class; more seriously, they add little to our basic understanding of the factors that determine forage digestibility.

T H E NUTRITIVE VALUE OF FORAGE CROPS

7

The inadequacy of crude fiber as a determinant of nutritive value was clearly established by Norman ( 1 935). Tentative alternatives to crude fiber were proposed: cellulose (Crampton and Maynard, 1938), holocellulose (Ely and Moore, 1955), modified acid-detergent fiber (Clancy and Wilson, 1966). Each of these aimed to analyze a more precise chemical grouping than crude fiber, but each perhaps reflected the same basic thinking, that the complex process of forage digestion can be quantified by a single chemical analysis. The relationships between these “fiber” components and forage digestibility (reviewed by Miller, 196 1 ; Sullivan, 1962), are often more precise than those based on crude fiber; they are still inadequate for predictive purposes. This conclusion, which had become evident by 1960, stimulated the two main developments discussed below: the study of chemical techniques more relevant to the digestion process, and of biological techniques that attempt to simulate the process of rumen digestion by a laboratory technique. C. IMPROVEDCHEMICAL TECHNIQUES Forage digestibility, Eq. (2), is the summation C% content X % digestibility of all the different chemical components in the forage. Some of these components, such as soluble carbohydrates and organic acids, are completely digested as the forage passes through the ruminant tract; others, in particular the polysaccharides and lignin, are much less completely digested and comprise most of the feed residue excreted as feces. The “classical” approach, discussed above, assumes that the extent to which the fiber fraction is digested is directly related to the proportion of that fraction in the forage. Detailed studies of the digestibility of different fiber fractions, based on in vivo experiments, have clearly shown that this is not so. Thus Jarrige and Minson (1964) found that there was no decrease in the digestibility of the cellulose in S.24 ryegrass as the cellulose content increased from 14.1 to 19.0 percent of the dry matter in early spring, while Gaillard (1962) and others showed that the cellulose in alfalfa is much less digestible than that in grasses with the same content of cellulose. This led to the development of techniques of graded extraction with reagents of increasing concentration (Gaillard, 1958; Jarrige, 196 1 ; Burdick and Sullivan, 1963) and of cellulose solubility in cupriethylenediamine (Dehority and Johnson, 1963) which take some account of the chain length and resistance to digestion of the different polysaccharide fractions. However, no single procedure is likely to give results relevant to the wide range of polysaccharides and lignin that comprise the fiber

8

W. F. RAYMOND

fraction in forages, and Gaillard (1966) has developed a more comprehensive relationship between forage digestibility and composition: Digestibility of organic matter % = 0.37(C-19.19) - 5.51(L-5.58) - 0.51(H-18.10) (3) + 4. I I(U-3.80) 65. I

+

which includes the percent contents of cellulose (C), lignin (L), hemicellulose (H), and anhydrouronic acid (U). More recently Gaillard and Nijkamp (1968) have proposed a less complex analytical system, which replaces the separate determinations of cellulose and hemicellulose with neutral-detergent fiber (N DF, v.i.): Digestibility of organic matter % = 66.7 - 4.64(L-5.19)- 0.14(NDF-48.05) + 2.95(U-3.47)

(4)

An alternative approach, developed by Van Soest (1 967) and Terry and Tilley ( 1964a), emphasizes the contribution to total forage digestibility of the highly digestible cell-contents fraction in forages. These workers have considered forage to contain two main fractions, the cell contents which are almost completely digested, and the cell-wall constituents, which are only partly digested, and they have proposed analytical systems that (a) separate these two fractions and (b) indicate the extent to which the cell-wall fraction would be digested in the ruminant tract. In a series of papers (summarized by Van Soest, 1967) this author has described methods for separating a forage sample into a cell-contents fraction soluble in neutral detergent (S), and an insoluble cell-wall fraction (neutral-detergent fiber, NDF), as well as a fiber fraction insoluble in acid detergent (acid-detergent fiber, ADF) and lignin (L). In a key paper (Van Soest and Moore, 1966), the digestibility of the N D F fraction was shown to be negatively correlated with log X (r = -0.98**) where X , the concentration of lignin in the A D F fraction, effectively measures the extent of lignification of the cellulose in the forage (in that paper X was denoted as L, which was confused with percent lignin). The mechanism by which lignin reduces fiber digestibility probably includes the effects of physical incrustation, of lignin-carbohydrate complexes, and of molecular bonds. Van Soest (1967) also showed that the cell-content fraction in forages is almost completely digested (98 percent) by the ruminant. However, a significant amount of material soluble in neutral detergent occurs in ruminant feces. This is not undigested plant cell contents, but consists of endogenous materials (mucus, salts, bile residues, and undigested bacteria) resulting from the digestion process; digestibility as measured by Eq. (2) is not the “true” digestibility of the forage material, but the “apparent” digestibility, (feed - feces) measuring the amount of feed digested, less this inevitable

THE NUTRITIVE VALUE O F FORAGE CROPS

9

endogenous loss associated with the passage of the feed through the tract. Based on in vivo results with a limited range of feeds, Van Soest (1967) calculated this fecal loss to be 12.9 percent of the dry weight of forage eaten. Van Soest (1967) was then able to compute the apparent digestibility of forage: Apparent digestibility of dry matter % = 0.98s

+ W ( 1 . 4 7 3 - 0.789 log X ) - 12.9

(5a)

comprising the almost completely digested cell-contents (S),plus the cellwall constituents (W=NDF) digested to an extent depending on the degree of lignification of the A D F fraction ( X ) , and less the endogenous excretion. It has not yet been possible to test this relationship on a wider range of forages than those studied by Van Soest. But by taking account of the differing contents and digestibilities of the two main fractions in herbage, the cell contents and the cell-wall material, Eq. (5a) clearly represents an important advance over the more empirical methods summarized by Miller ( 1 96 1) and Sullivan (1 962). In the course of the development of the detergent-fiber methods, Van Soest ( 1 965b) examined the effect of the method of drying herbage samples before analysis on the measured levels of acid-detergent fiber and lignin. Drying temperatures above 50”C., particularly over an extended period, significantly increased the levels of both these fractions; this artifact “fiber” was shown to result from a nonenzymatic browning reaction, in which protein polymerizes with products of carbohydrate breakdown, so that the “lignin” fraction in particular contains an abnormally high percentage of nitrogen. In earlier work this had been corrected by subtracting %N X 6.25 from the apparent lignin analysis. However, Van Soest recognized that natural plant lignins may contain some nitrogen, and derived a relationship that would correct only for the nitrogenous matter which might be attributed to heat damage: % corrected lignin (L,)

=

1.208 X % measured lignin (LA)- 10.75 x %N in A D F + 0.42

(6)

The acid-detergent fiber (ADF) content is then corrected: % A D F corrected = % A D F observed - (LA- L,)

(7)

From Eqs. (6) and (7) the factor log X in Eq. (5a), based on corrected values for A D F and lignin, can be calculated. The need for this correction must reduce the utility (and precision) of Eq. (5a) and emphasizes the importance of adequate drying methods for preparing herbage samples for analysis. The method of choice must surely

10

W. F. RAYMOND

be freeze-drying (lyophilization). But the great majority of freeze-driers in current laboratory use are of small capacity (< 1 kg. water/24 hours), and this can introduce a source of error which is seldom recognized- that the sample of forage which is dried by this ideal method may be so small as to be quite unrepresentative of the material sampled. This possible contradiction between the precision of the drying method and the accuracy of sampling has been discussed (Commonwealth Agricultural Bureaux, 1961, p. 135); until much larger freeze-driers become available, the solution in many cases may be to dry forage samples of adequate size as rapidly as possible at 100"C.,so as to minimize the time during which nonenzymatic browning (which occurs only in the presence of water) can take place. The individual investigator can then test the success of his own drying method by the application of Eq. (6) to analyses on representative samples. Recently Van Soest and Jones (1969) suggested a further refinement to the concept summarized in Eq. (5a), by indicating that the silica present in plant material may exert much the same effect as lignin in reducing the digestibility of the neutral-detergent cell-wall fraction. L. H. P. Jones and Handreck (1967) discussed the forms and reactions of silica in the food chain from soil to plant to animal. They pointed out that silica absorbed by plant roots is carried in solution to the actively metabolizing tissues. As the transporting water is transpired, solid silica is deposited on to the cell walls so that as these develop the polysaccharides are intimately associated with encrusting silica as well as lignin. From examination of the digestibility in vitro of forage samples of silica content ranging from 0.5 percent to 5.4 percent, Van Soest and Jones (1969) proposed a modified form of Eq. (5a): Apparent digestibility of dry matter % = 0.98s + W(1.473 - 0.789 log X) - 3.O(SiO2)- 12.9

(5b)

As yet the evidence is restricted to relatively few forages, but further study may clearly indicate the need for refinement of the biological concepts implicit in Eqs. (5a) and (5b).

D. ESTIMATION OF FORAGE DIGESTIBILITY BY in Vitro TECHNIQUES The inclusion of silica as a further component which may influence forage digestibility illustrates the trend toward multicomponent chemical techniques for predicting digestibility. In effect, this accepts that no single component can quantify the complex process of ruminant digestion, and that this must be treated as a series of stages, each described by a logical chemical evaluation, as in the decreasing digestibility of the N D F fraction as the fiber becomes more lignified.

T H E NUTRITIVE VALUE OF FORAGE CROPS

11

The inclusion of silica also illustrates a basic problem with chemical methods of evaluation, that a relationship such as Eq. (5a), which is found to be adequate with one population of forages, may give inaccurate prediction of the digestibility of other forages-in this case, of forages of unusually high silica content. This could arise from two causes: (a) the original relationship did not include all the components that exert a significant effect on forage digestibility and (b) chemical methods measure the content of different components in forage samples; they do not measure the physical distribution and organization of these different components within the plant, which must to some extent determine how far the plant fibers are digested by the microorganisms within the rumen. The chemical approach tends to treat a forage as a homogeneous material, an increase in lignin content, for instance, being visualized as an increase in lignification throughout the whole plant. In practice the forage plant is more realistically considered as made up of morphologically “distinct” fractions, each of which can be changing in chemical composition and digestibility in a way not necessarily related to the other fractions, so that chemical analysis (an average of the whole plant material) may well not describe the summation of the individual plant fractions that make up the digestibility of the whole plant. Thus, parallel to the development of chemical methods of forage evaluation, described in the previous section, has been the development of biological methods of evaluation, the artificial rumen or in vitro digestion methods. Essentially these have attempted to simulate the process of ruminant digestion by methods that can take account both of the overall chemical composition of the forage plant and of the distribution and physical interrelations of the chemical components within the different morphological parts of the plant. With the recognition that the digestibility of the “fiber” fraction of forages would be most affected by these physical characteristics the initial investigations of biological methods were concerned with fiber digestibility, and in particular with the digestibility of the cellulose fraction in forages. Although details of technique differed, all these methods were based on the incubation, under controlled conditions, of a sample of the test forage with a mixed culture of the microflora taken from the rumen of a forage-fed animal; the aim was to standardize the conditions of incubation so that the fiber in the forage sample was digested to the same extent as in the same forage when fed in an in vivo experiment (Quicke et al., 1959; Lefevre and Kamstra, 1960; Karn et al., 1967). These techniques were also used to measure the extent to which the dry matter (Clark and Mott, 1960), organic matter (R. L. Reid el al., 19601, or energy content (R. L. Reid et al., 1960; Baumgardt el al., 1962; Naga and El-Shazly,

12

W. F. RAYMOND

1963) in forage was digested in vitro. In most cases the extent of digestion in vitro was found to be less than that in vivo, and regression equations were developed to allow prediction of in vivo values. In an alternative system, a sample of dried forage is enclosed in a nylon or dacron mesh bag suspended within the rumen in vivo, and digestibility and rate of digestion are measured by the loss of dry matter or of cellulose from the sample (Lusk et al., 1962; Hopson et al., 1963). This technique could have the advantage that a normal microfloral population will be maintained, although this will tend to be that characteristic of the feed eaten by the host animal, rather than of the sample under test. However the technique does permit the comparison of large numbers of feed samples under standard conditions, and it could be of use in ranking forage samples in a breeding selection program. This approach was analogous to that with the earlier chemical methods, in attempting to predict the complex process of forage digestion by a single procedure. As with the chemical methods, the accuracy of prediction was found to decrease as the range of forages examined was widened; in particular marked divergences were found between results for grasses and legumes (Shelton and Reid, 1960). Tilley and Terry (1963) suggested that these discrepancies might be the result of correlating data from a single digestion with rumen organisms with those from digestion within the animal, which involves a mainly bacterial digestion within the rumen followed by a mainly enzymatic digestion in the remainder of the digestive tract. Within the rumen, the “digestible” polysaccharides, carbohydrates, and protein in the feed are broken down by the action of the microorganisms there; some of the products of digestion are absorbed directly through the lumen wall, but a considerable part serves as the substrate for microbial growth, and is resynthesized into protein, polysaccharides, and lipids within the proliferating bacterial and protozoal population. These microorganisms, entrained in the residues of undigested fiber and other feed components, then pass from the rumen to the abomasum and duodenum. In these organs the digesta are acidified and further digested by secreted enzymes that hydrolyze much of the bacterial and residual plant proteins to amino acids. These are then absorbed as the main source of amino acids for the metabolism of the host animal. The undigested residue from the in vitro digestion of forage material with rumen microorganisms is thus seen to contain, in addition to undigested feed, the rumen organisms which, in vivo, would be enzymatically digested in the ruminant hind tract. Tilley and Terry ( 1 963) proposed that this second stage should be simulated by subjecting the residue from the in vitro bacterial digestion to a second enzymatic digestion. They ex-

THE NUTRITIVE VALUE OF FORAGE CROPS

13

amined several enzymes and concluded that a two-stage procedure comprising digestion by rumen microorganisms followed by digestion by acid-pepsin gave the closest agreement with in vivo digestibility values for the dry matter and organic contents in forages. This method showed a correlation of 0.97 between in vitro and in vivo values when tested on a wide range of forages, including grasses fertilized with different levels of nitrogen, and legumes: Digestibility in vivo = 0.99 X digestibility in vitro - 1.O I (S.E. = f 2.3 I )

(8)

Similar high degrees of correlation have been found by O’Shea and Wilson ( I 965; r = 0.94), Wedin el al. ( 1966; r = 0.996) and Ademosum et af. ( 1 968; r = 0.96); Dent ( 1 963) found close agreement between two-stage in vitro and in vivo digestibility results with brassicas and forage maize. In a number of studies the precision of prediction of digestibijity in vivo by this in vitro technique has been compared with the chemical methods already discussed. Armstrong et al. (1964a) found that the metabolizable energy and net energy contents of a series of dried grass feeds were more accurately correlated with organic matter digestibility in vitro than with cellulose or lignin contents; Bosman (1967) and Ademosum ef al. (1968) have reported the in vitro method to correlate more closely with in vivo digestibility than the chemical methods tested; Engels and Van der Merwe (1967) have found the same result with lowdigestibility hays in South Africa. However, to date no direct comparison has been reported with the improved chemical technique proposed by Van Soest ( 1 967, Eq. 5 ) , and it is possible that these two techniques differ little in the precision with which they allow prediction of forage digestibility in the laboratory. In fact the stage may be approaching at which little further improvement in precision can be expected. In interpreting the error terms of these relationships, it is important to recognize that this error does not arise solely from deficiencies in the laboratory technique used (chemical or in vitro), but that errors are also associated with the actual measurement of the in vivo digestibility of the forages, and with the fact that digestibility in vivo is not a constant parameter of a particular forage. Thus digestibility determined in an animal experiment may depend on the amount of forage fed, digestibility decreasing as the level of feeding increases (Moe et al., 1963, and can be significantly reduced if the animal is parasitized with stomach worms-probably the rule rather than the exception with sheep (Spedding, 1954; Shumard et al., 1957). The standard deviation of an estimate of digestibility is between 1.0 and 1.3% (Raymond et al.,

14

W. F. RAYMOND

1953); as most digestibility determinations are made with only 2 or 3 sheep it is evident, even where the factors noted above are standardized, that much of the error in the relationships noted must result from errors in the in vivo determination, rather than in the concept or precision of the laboratory determination. Probably the main area for improvement lies then in the better standardization of the in vivo digestibility experiments, of the preparation of the forage samples for analysis, and of the conduct of the laboratory procedures. R. L. Reid et al. (1964) and Noller et al. ( 1966) have both reported significantly higher levels of dry-matter digestibility in vitro in forages prepared by freeze-drying than by oven-drying, presumably because of the production of indigestible artifacts during oven-drying, as suggested by Van Soest ( I965b). Tilley and Terry (personal communication), however, found no advantage of freeze drying compared with rapid ovendrying at 1Oo”C., although the same authors (1963) had reported considerable depression of digestibility in vitro in samples dried at temperatures above 1 10°C. The concept that fiber digestibility is limited by physical incrustation with lignin indicated that digestibility should be increased by fine subdivision, and both Dehority et al. (1962) and Tilley and Terry (1 963) reported increases of up to 50 percent in digestibility as a result of grinding forage samples in a ball mill before in vitro digestion. However, within the range of fineness of grinding found in forage samples ground by hammermill, Tilley and Terry (1 963) found no significant effect of particle size. The most serious problem arises, however, in the lack of standardization of in vitro procedures between different laboratories. Barnes (1967) reported the results of a collaborative study in which the in vitro digestibilities of the dry matter and cellulose in three forages were measured at 17 laboratories. The mean values for cellulose digestibility after 24 hours ranged from 40.0 to 63.9 percent, reflecting the use of different techniques in terms of sample size, preparation of the rumen inoculum, pH control, etc. In contrast, Raymond and Terry (1966) have reported close agreement between in vitro results from two laboratories using identical procedures, and have stressed the importance of different laboratories using the same “standard” forage samples as an additional check on the reproducibility of the method. Tilley and Terry ( 1 963) found that rumen liquors taken from donor animals fed on several contrasting forages were of similar digestive efficiency in the two-stage in vitro system, and Troelsen and Hanel(l966)

THE NUTRITIVE VALUE OF FORAGE CROPS

15

have reported that the potency of liquors taken from different sheep differs less than the in vivo digestive efficiency between sheep. I n general the most important consideration seems to be that the diet of the donor animals should contain at least 10 percent of crude protein (see below), that it should give a sample of rumen contents from which a strained liquor can readily be separated (i.e., the animals should be fed on a coarsely chopped hay rather than on a pelleted ground feed), and that the sample should be kept in an anaerobic condition and be prepared for inoculation of the digestion tubes as rapidly as possible. It must be accepted, however, that this in vitro procedure, developed with temperate grass and legume species, may not be directly applicable with other temperate species, or with tropical forage species, and Drew (1 966) has stressed that the system should wherever possible be checked with relevant samples of known in vivo digestibility. Thus Raymond and Terry (1 966) reported low in vitro digestibility levels when both the test forage (0.7 percent N) and the feed of the donor animal were of low nitrogen content, as can often occur with tropical forage species. The higher level of digestibility in vivo resulted from the animal’s ability to recycle urea via salivary and ruminal secretions, whereas digestibility in vitro was limited by a deficiency of nitrogen in the combined sample and inoculum. Addition of 6 mg. of N , as urea, to the in vitro system increased sample digestibility to the in vivo level. Engels and Van der Merwe (1967) found that the difference between in vitro and in vivo digestibility values of veldt grasses became greater as the nitrogen content of the test forages decreased. Addition of 20 mg. of urea N to each digestion tube gave in vitro values in close agreement with those in vivo. In a modification of the method of Tilley and Terry ( 1963), Alexander and McGowan ( 1966) have included ammonium sulfate in the buffer added to each tube. However in the author’s opinion this is advisable only where depressed levels of in vitro digestibility result from low nitrogen contents. Engels and Van der Merwe (1967) showed a marked depression in digestibility when 60 mg. of urea N was included in the digestion system, and a similar depression might occur when urea is added to an in vitro digestion of a forage sample already of high nitrogen content. Although some modification may be necessary in particular situations, the study reported by Barnes (1967) does emphasize the importance of the general adoption of standardized in vitro digestibility procedures, without which results reported by different laboratories cannot be directly comparable. The two-stage procedure described by Tilley and Terry (1963) is now used by many laboratories, and there seems a strong case

16

W. F. RAYMOND

that this should be adopted as a standard procedure in agronomic studies. But such a technique, based on digestion in vitro with a mixed culture of rumen organisms, must always be sensitive to uncontrolled biological variation, and there is a clear need for the future for a technique that can be more rigidly defined. A promising development, reported by Dehority et al. (1968), is the use of a pure culture of cellulolytic bacteria to replace the mixed inoculum taken from the rumen of a donor animal. Two of the strains of bacteria tested gave cellulose digestibility values in vitro in close agreement with the measured in vivo values; further development of this work could well lead to significant improvement in the standardization of laboratory in vitro techniques.

E. THERELATIVEUTILITY OF CHEMICAL AND in Vitro ESTIMATIONS OF FORAGE DIGESTIBILITY As has been indicated, the two-stage in vitro technique appears to give a better prediction of in vivo forage digestibility than any of the chemical methods yet investigated, and the use of this technique, reported below, has contributed greatly to our knowledge of forage digestibility. However this is an integrative technique, the measured digestibility being the sum of the digestibilities of the many different chemical fractions within the forage; without additional chemical information it can only describe, rather than explain, the differences in digestibility observed among different forage samples. This suggests that in vitro and chemical techniques should be considered as complementary, rather than competitive, methods of forage evaluation, the in vitro techniques being used to establish that forages differ in digestibility, and the chemical techniques to study the probable reasons for these differences. Such an understanding is essential if further improvement of forage digestibility is to be based on nonempirical concepts. IV. The Digestibility of Different Forage Species

A. BASIC PATTERNSOF DIGESTIBILITY The use of these more precise laboratory techniques for estimating the digestibility of forages has led to considerable progress in extending in vivo studies that were started in the 1950’s. It had long been recognized that as a forage becomes more mature it also becomes less digestible; the possibility that this imprecise statement could be quantified was proposed by Homb (1953), who showed a close relationship between the age (maturity) of a timothy-clover mixture and its digestibility, and by J. T. Reid et al. (1959), who suggested that the digestibility of a range of

THE NUTRITIVE VALUE OF FORAGE CROPS

17

forage species harvested during first growth during the spring could be estimated: % digestibility of dry matter = 85.0 - 0.48X (9) where X = number of days to harvest from April 30. These reports stimulated further investigations that have greatly increased our understanding of forage digestibility. Clearly the “date of harvest” in Eq. (9) can only be relevant in the area close to Cornell (Ithaca, New York), where the forages used in these in vivo digestibility determinations were harvested, but the basic principle, that digestibility is closely related to forage maturity, with “date of cutting” being used to describe stage of maturity in a particular location, was soon confirmed by other workers (Kane and Moore, 1959, and others; reviewed by Blaser, 1964). However, several divergences from the original concept, that all forages are of similar digestibility at a given date, became evident. Thus the digestibility of timothy (Phfeumpratense) (Mellin et a f . , 1962; Minson et a f . , 1964), of lucerne (Medicago sativa) (Demarquilly, 1966a), and in particular of white clover (Trifofium repens) (Harkess, 1963) were shown to decline less rapidly with advancing maturity than the 0.5 unit per day indicated in Eq. (9). Small differences in the digestibility of a given forage variety cut on the same date in different years may be due to the delayed onset of active spring growth in a “late” season (Minson et af., 1960), or to differences in leaf percentage in the forage grown in different years (R. H. Brown et a f . , 1968). But the most important observation has been that there can be large and consistent differences in digestibility between forage species and forage varieties. In a detailed series of in vivo experiments, Minson et a f . ( I 960, 1964) found that certain species (Lofium spp. and Festuca pratense) were considerably more digestible than others such as cocksfoot (Dactylis gfomerata) and tall fescue ( F . arundinacea); within a species late-maturing varieties (e.g., S.23 ryegrass) maintained a high level of digestibility to a later date than early-maturing varieties (e.g., S.24). These differences, illustrated in Fig. 1 , were considered large enough to be of agronomic and nutritional significance. Figure 1 also shows that the digestibility of each grass studied did not fall at a uniform rate as it matured. There was an initial period of almost constant digestibility before digestibility began to decrease. In the case of Lofium and Dactylis this change to a more rapid fall in digestibility, at a rate very similar to the 0.5 unit per day recorded by Reid et a f . ( 1 959), was closely associated with the first emergence of flowering heads; the digestibility of timothy (S.48), however, decreased well before this stage, and this species also showed the much slower rate of fall in digestibility noted above.

18

W. F. RAYMOND

In establishing these novel digestibility patterns, Minson et al. (1964) had two main advantages compared with their colleagues in North America-in the more general use of the genus Lofium in the United

I

I

10 20 April

30

Ib

20

May

3b

Ib

$0

$0

June

date of first cuttinq

FIG. I . The percent digestibility of the organic matter in grass varieties during first growth in the spring. S.23 and S.24, ryegrass; S.37 and Germinal, cocksfoot; S.215, meadow fescue; S.48, timothy; S. 170, tall fescue; O.M., organic matter. 0 indicates date of first ear emergence. (Data from Minson et al., 1964.)

Kingdom than in North America; and in the availability of widely differing maturity types, within this genus, developed at the Welsh Plant Breeding Station. These results stimulated other in vivo studies, which have in general confirmed the important conclusions implicit in Fig. 1. Thus Castle et a f . (1962) and Harkess (1963) have found consistently higher digestibility of ryegrass than of cocksfoot, and Lowe et af. (1962) reported that late-maturing varieties of grasses were more digestible than early-maturing at a given cutting date in the spring. These in vivo studies, initially on first spring growth, have also been extended to regrowths during the rest of the growing season, and consistent patterns have again emerged. The digestibility of regrowth cocksfoot and tall fescue is always lower than that of the corresponding regrowth of ryegrass, and the rate of fall of digestibility with time is much less in these later, largely vegetative, regrowths than in the first, reproductive, growth (Minson et af., 1960, 1964); a similar slower rate of fall in digestibility has been shown with regrowths of lucerne (Demarquilly, 1966a).

THE NUTRITIVE VALUE OF FORAGE CROPS

19

However, the digestibility of regrowth forage is likely to be less predictable than that of first growth. The latter comprises largely reproductive tillers which develop from the start of active growth in the spring; regrowths can contain both reproductive and vegetative tillers, the relative proportions of each depending on the management of earlier harvests. Thus the regrowth from a tiller whose reproductive growing point (ear) is harvested will comprise mainly leaf, which will decrease only slowly in digestibility at ca. 0.1 percent per day. In contrast, a tiller which is harvested before the ear has reached the height of cutting or grazing will continue to develop, and its digestibility will decrease at the 0.5 percent per day characteristic of first growth. Harvesting up to the time of first ear emergence will remove only some of the developing ears, and the digestibility of the regrowth will decrease at an intermediate rate, e.g., 0.3 percent per day. A later harvest, which removes most of the potential ears, will give a leafy regrowth which decreases more slowly in digestibility. Under a cutting regime regrowths are mainly leafy, and of relatively predictable digestibility (Minson et al., 1960, 1964). Under a grazing situation the prediction of the digestibility of regrowths is less precise. The grazing animal generally leaves some of the herbage on offer, and this continues to decrease in digestibility, so that the combined regrowth and remaining herbage is of lower digestibility than the corresponding regrowth from a cut sward (Tayler and Deriaz, 1963). As noted in Section IV, C , 2, the digestibility of a regrowth from grazing will also depend considerably on the moisture and nitrogen status of the sward, which determines the proportion of the next harvest that is composed of new growth of high digestibility. Limited information indicates that the digestibility of forage mixtures can be calculated from the proportions and digestibilities of the constituent species at the time of harvest (Harkess, 1963). An exception may be the case where the digestibility of one species is low because it is deficient in protein content (as has been shown with some tropical forages; Smith, 1962). Such a species, grown with another species of higher protein content, could have an enhanced digestibility, so that the digestibility of the “mixture” would be somewhat higher than predicted. The patterns of forage digestibility discussed above were all based on in vivo experiments; they have been considerably extended by the use of laboratory in v i m techniques. The validity of the two-stage in vitro technique was indicated by Terry and Tilley (1964a), who showed that the basic patterns of digestibility shown in Fig. 1, and in particular the nonlinear fall in digestibility with time and the consistently higher diges-

20

W. F. RAYMOND

tibility of ryegrass than of cocksfoot, could be closely reproduced by in vitro measurements. With this method many more forage samples can be examined than would ever be possible by in vivo experiments, but the need to confirm, in vivo, the more important conclusions indicated from in vitro studies must be emphasized. Thus Dent and Aldrich (1968) have measured the digestibility, in both first growth and regrowths, of numbers of varieties within several forage species. They have shown consistent differences in digestibility between species, and between maturity types within species, at different centers and in different years. Their work indicated that different varieties of the same maturity type within a species might differ in digestibility. Thus a small number of varieties of cocksfoot (e.g., Roskilde I1 and Scotia) appeared to be more digestible than the varieties Germinal and S.37, shown in Fig. 1, and Reveille tetraploid ryegrass was more digestible than S.24 ryegrass. These in vitro results, subsequently confirmed in in vivo experiments (Osbourn, unpublished) have particular relevance to the possibility of breeding more digestible forage varieties, discussed in Section VIII.

B. THEDIGESTIBILITY OF DIFFERENT PLANTFRACTIONS The use of in vitro techniques to measure the digestibility of different plant fractions has also provided a more logical understanding of these patterns of forage digestibility. These are in many cases contrary to what would have been predicted in terms of the earlier concepts of forage nutritive value, that leaf was more digestible than stem, and that forage fractions high in protein (N) content would be more digestible than those of lower N content. The results of Minson et al. ( 1 960) showed that these concepts might be incorrect. Thus in first growth, S.24 ryegrass forage cut in 1959 on May 1 , with a nitrogen content of 2.62 percent and 53 percent of leaf lamina, was of exactly the same digestibility as that cut on April 20, which had contained 3.66 percent N and 77 percent leaf, S.37 cocksfoot cut on May 5 , with 2.78 percent N and 58 percent leaf, was 4.5 units less digestible than the ryegrass cut on May 1 . In later experiments Tayler and Rudman (1966) harvested a S.24 ryegrass sward in two horizons, a top fraction cut above 13.5 cm., and a bottom fraction from 6 to 13.5 cm. The digestibilities of the organic matter in the two fractions, in vivo, were 84.0 and 82.6 percent, respectively. This indirect in vivo evidence has been followed up in detail by in vitro digestibility determinations on forage samples separated into fractions of leaf, leaf sheath, stem, inflorescences, and dead material. Terry and Tilley ( 1 964a) analyzed these fractions from the forages fed in vivo by Minson et al. ( 1 960, 1964) (Fig. 1). They showed that in all species the

THE NUTRITIVE VALUE OF FORAGE CROPS

21

digestibility of the leaf fraction fell only slowly with advancing maturity (0.13 percent per day) whereas that of the leaf sheath (0.4 percent) and stem fractions (0.7 percent) fell much more rapidly. In the immature forages the stem was always more digestible than the other components. On any given date each fraction in S.24 ryegrass was more digestible than the corresponding fraction in S.37 cocksfoot, of equivalent maturity type. A typical set of results is shown in Fig. 2. The digestibility of the

whole plant leaf sheath

----

leaf blade stem

--

-----.

FIG.2. The digestibility in vifro of the dry matter in the whole plant, and in the leaf blade, leaf sheath, and stem fractions of S.37 cocksfoot during first growth in the spring. Figures in parentheses are the percentage of stem in the whole plant. (Data from Terry and Tilley, I 964a. )

whole forage material, calculated from the proportions and digestibilities of the constituent fractions, changed with maturity just as in the in vivo experiments-that is, with a slow fall in digestibility up to the time of ear emergence, followed by a more rapid fall as the stem and leaf sheath fractions, by now less digestible than the leaf, comprised an increasing proportion of the total forage. The slower and more steady fall in the digestibility of S.48 timothy could be explained by the much higher proportion of leaf sheath in this species than in ryegrass or cocksfoot.

22

W. F. RAYMOND

These conclusions have been confirmed in considerable detail by Pritchard et al. (1963), Wedin et al. (1966), Walters et al. (1967), and Dent and Aldrich (1968); Mowat et al. (1965) found similar results with timothy and bromegrass, but were unable to show higher digestibility of cocksfoot stems than of leaves even in immature forage. Similar logical patterns of digestibility have been shown with legume forages (Terry and Tilley, 1964a; Mowat et al., 1965). With lucerne ( M . sativa), red clover ( T . pratense), and sainfoin (,Onobiychis viciifoliu) the older leaves tend to senesce and fall, so that the digestibility of the leaf fraction decreases very little as the plant matures. By separating the stem fraction into 6 inch subfractions, measured from the top of the plant, it was shown that digestibility decreased down the stem, but that the digestibility of any given fraction changed relatively little with maturity. A similar result has been found with the stems of brassicas and forage maize (Zea mays), the stem tip of marrow-stem kale being highly digestible (81.8 percent dry matter) and the stem base of lower digestibility (62.3 percent) (Dent, 1963). Within a particular forage species, Walters et al. (1967) and Dent and Aldrich (1968) have shown, at a similar stage of morphological development, that “late” varieties tend to be less digestible than “early” varieties because both the stem and leaf fractions are somewhat less digestible, and because they contain a higher proportion of senescent and dead material. Small differences in digestibility of a given variety at ear emergence in different years can be attributed to differences in leafistem ratios, stem at this stage being less digestible than leaf (R. H. Brown et al., 1968). Of particular importance is the observation, already noted, of differences in digestibility between varieties of similar maturity type within a species (Dent and Aldrich, 1968). The higher digestibility of REVEILLE ryegrass than S.24 was not accounted for in terms of leafstem ratio, but because both the leaf and stem fractions in REVEILLE were more digestible than the same fractions in S.24. Differences in digestibility have also been shown between the individual plants (genotypes) within a variety (Cooper et al., 1962; Walters et al., 1967; Mowat, 1969), resulting from differences in digestibility of the plant fractions rather than from different leaf stem ratios. However, while the in virro techniques used in these studies can describe the changes in digestibility within and between different forage. species, they cannot explain them; for this the newer chemical techniques (Section 111, C ) are needed, to analyze digestibility measured in vitro into its component parts. In detailed studies by Tilley, Terry, and Outen (unpublished) the forage sample is separated into a cell con-

THE NUTRITIVE VALUE O F FORAGE CROPS

23

tents fraction, soluble in acid pepsin, and a cell wall fraction (analogous to the fraction soluble in neutral detergent and the cell wall residue of Van Soest, 1967). The digestibility of the cell wall fraction is also measured by in vitro (rumen organism) digestion. Consistent differences have been shown between S.24 ryegrass and S.37 cocksfoot, harvested at the same stage of maturity: (a) the content of pepsin-soluble material is higher in the ryegrass than in the cocksfoot; (b) as a result there is a higher content of cell wall fraction in the cocksfoot; and (c) this fraction in the cocksfoot is less digestible than in the ryegrass, so that the content of “digestible cell wall material” in the two species is very similar. As a result the digestibility of the ryegrass (cell contents X 0.98 digestible cell wall fraction) is higher than that of the cocksfoot. Within different plant fractions, the small decrease in digestibility of the leaf fraction as the plant matures (Fig. 2) is accounted for by the consistently high level of cell contents and high digestibility of the cell wall fraction in the leaf. Young stem material contains an even higher proportion of cell contents than the leaf (hence its higher digestibility); but as the stem matures the cell content fraction decreases rapidly and is replaced by cell wall material which becomes less digestible with increasing lignification, so that the digestibility of the stem decreases rapidly with advancing maturity. Extension of these more detailed studies of the components of digestibility to the genotypes within a species which have been found to be of higher digestibility (Section VIII) may provide a more objective basis for selection for improved digestibility than the in vitro techniques that have so far been used.

+

c. THEEFFECTOF ON

ENVIRONMENTAL A N D OTHER FACTORS FORAGEDIGESTIBILITY

1 . Environmental Effects The difference in digestibility between forages cut on the same date at Cornell (J. T. Reid et al., 1959) and in Maryland (Kane and Moore, 1959) appeared to be due to the forages in these two locations being at different stages of physiological maturity. It is of course possible that, even at the same stage of maturity, the digestibility of a forage may differ between locations; thus Aldrich and Dent (1 967) have found an indication of higher digestibility at a northern than at a southern latitude in the United Kingdom in cocksfoot cut 10 days after 50 percent ear emergence. Deinum et al. (1 968) measured the in vivo digestibility of perennial ryegrass grown under high and low light intensities, and with low and high levels of nitrogen manuring. Considerable differences in chemical com-

24

W. F. RAYMOND

position of the grass were found between treatments, but these had no significant effect on dry matter digestibility at any one sampling. These experiments showed lower levels of digestibility on all treatments during the summer when temperatures were higher, confirming earlier results of Deinum, based on chemical analysis of ryegrass grown in controlled environment cabinets. Deinum et al. ( 1 968) postulated that this effect of high temperature might partly account for the generally lower level of digestibility of tropical than of temperate forages. Hiridoglou et al. ( 1966) have also shown that high summer temperatures were associated with low in vitro digestibility levels. Forages growing in summer also tend to contain lower moisture contents than late season forages; however, the only report found on the effect of water intake on forage digestibility (Thornton and Yates, 1968) has indicated a small increase in the digestibility by cattle of the dry matter and fiber in chaffed oat straw-lucerne hay when water intake was restricted. 2 . Fertilizers and Forage Digestibility The effects of fertilizer nitrogen on forage digestibility have been studied in numerous experiments; most of these have reported an insignificant effect from the use of widely differing levels of application (summarized by Blaser, 1964): thus Minson et al. ( 1 960) found no effect on the digestibility of ryegrass or cocksfoot from levels of nitrogen application varying from 0 to 175 pounds/acre. However, Raymond and Spedding (1 965) have indicated several situations in which fertilizer nitrogen is likely to affect forage digestibility: (a) in a mixed grass-clover sward the use of this fertilizer may reduce the contribution, in the forage harvested, of the more digestible clover complement: (b) uneaten herbage left on a sward after stock have grazed will continue to decrease in digestibility, and by diluting the highly digestible new growth will depress the digestibility of herbage available at the next grazing; fertilizer nitrogen will increase the proportion of new growth in this harvest, and so may increase its digestibility; (c) unfertilized herbage may contain an inadequate level of nitrogen for the growth of rumen microorganisms: thus Smith ( 1 962) found an increased level of digestibility after application of nitrogen to such forage; 40 pounds of fertilizer nitrogen per acre increased the crude protein content of late-cut veldt hay from 3.6 to 6.8 percent and the digestibility of the forage dry matter from 5 1.7 to 59.5 percent. These effects of fertilizer nitrogen are consistent with the concepts of forage digestibility already discussed. McIlroy ( 1 967) has summarized results showing that fertilizer nitrogen increases the crude protein content and decreases the soluble carbohydrate content of herbage. But

THE NUTRITIVE VALUE OF FORAGE CROPS

25

there is little net change in the content of crude protein plus soluble carbohydrate, which largely comprises the cell contents fraction, or in the composition of the cell wall fraction. Thus little change in the digestibility of the forage would be expected except when the digestibility of the cell-wall fraction is limited by the low content of protein in the forage. There is also the possibility that certain forages contain specific components which reduce bacterial activity within the rumen. Hawkins ( 1 959) suggested that the low digestibility of the protein fraction in Sericea lespedeza and vetches might be due to the formation of insoluble protein complexes with the tannin in these forages, and Smart et al. ( 1 96 I ) found a depression of cellulose activity in vitro by an extract from Lespedeza cuneata. Schillinger and Elliott ( 1 966) observed differences of u p to 15 percent between the digestibilities in vitro of different lucerne plants. Low levels of digestibility could be raised by addition of amino acids (glycine, aspartic acid, glutamine) to the in vitro system, and these amino acids also increased the growth rate of voles fed on the lucerne forage. These authors attributed these differences to the presence of water-soluble antimetabolites in the low digestibility plants. Such cases are, however, likely to be exceptional, and the digestibility of most forages appears to be in line with the systems of evaluation proposed by Van Soest ( 1967) and Terry and Tilley ( 1964a).

3. The Effect of Feed Supplements on Forage Digestibility These systems do not, however, predict adequately the digestibility of forages fed in mixed rations. Thus when carbohydrate (starch) supplements are fed with forages, there can be a significant decrease in the digestibility of the fiber (cell wall) fraction of the forage, unaccounted for by any change in the composition of the cell wall (Eq. 5 ) . This has been attributed to a preferential digestion of the starch by the rumen microorganisms, so that the extent of digestion of the plant fibers is reduced, or alternatively that the amylolytic bacteria compete preferentially for ammonia against the cellulolytic bacteria, so reducing cellulose digestion (El-Shazly et al., 1961). More recent work has indicated an alternative explanation. First, it is known that when a starch supplement is fed with a forage there is a reduction in the pH of the rumen contents compared with that when the forage is fed alone (Topps et al., 1965). Second, Tilley et al. ( 1964) have shown a marked reduction in the rate and extent of digestion of dry matter and cellulose in vitro when the pH of the in vitro system is reduced (Table I); a decrease, with decrease in rumen pH, has also recently been

26

W. F. RAYMOND

indicated in the rate of “digestion” of cotton threads suspended within the rumen in vivo (Wilkins, unpublished). Tilley et al. (1964) have therefore suggested that the lower rumen pH when a starch supplement is fed TABLE I The Effect of the pH of the in Vitro Digestion on the Extent of Dry Matter and Cellulose Digestibility of Samples of Cocksfoot, S.37, by the 2-Stage in Vitro Method“

Sample No.

In vivo

Percent of forage dry matter digested In vitro (48 hours) pH 6.8 pH 6.0 (normal)

76 73 72 72 64 63 59 59 5 49 49 “From Tilley et al. ( 1964). *Figures in parentheses are percentages matter.

(15)

(17) (16) (12)

67 (12) 63 ( I 1) 50 (10) 49 (10) 41 (7)

pH 5.5 54 53 37 36 29

(4) (5) (3) (2) (1)

of digestible cellulose in the digestible dry

may provide a less favorable environment for the cellulolytic and other bacteria that are able to digest plant fiber (see also Head, 1961). In anthropomorphic terms, the natural rumen microflora has become adapted to the pH 6.6 to 6.8 characteristic of the rumen contents of the grazing herbivore; any marked divergence from this pH finds a microfloral population increasingly unable to digest fiber. This hypothesis, if substantiated, could lead to the development of feeding regimes, aimed at optimizing rumen pH (and redox potential) to ensure maximum digestion of the cell wall fraction of forages. As already noted, forages may be of low digestibility because they are of very low protein content (< 4% crude protein). The digestion of these forages can be increased by feeding protein supplements, and there has been much interest in the use of urea for this purpose. Thus Campling ef al. (1962) measured an increase in organic matter digestibility from 41 to 50 percent when a urea supplement was fed with oat straw of 3.0 percent crude protein, and other examples have been reported (see M. H. Briggs, 1967). Within the rumen the urea is rapidly deaminated, the ammonia produced then being used by the celluloytic and other bacteria, whose digestive activity would have been limited by deficiency of protein in the unsupplemented forage. The feeding of urea to increase the digestibility of forage is not often used in practice, because a crude protein level in forage of less than 4

THE NUTRITIVE VALUE OF FORAGE CROPS

27

percent is uncommon. The main use of urea is likely to be as a substitute for protein in productive rations (Section VI, D). V. The Voluntary Intake of Forages

A. THEFACTORS CONTROLLING FEEDINTAKE

It was noted in the Introduction that the quantity of forage that ruminant animals eat is in practice seldom controlled in the same way that most other farm feeds are rationed. Yet the amount of forage that animals eat is in many cases the major factor determining their level of nutrient intake and their output of useful products. Increasing emphasis has been given in the last decade to the study of voluntary intake. Earlier studies of forage intake were undoubtedly hindered by the confusion of “intake” with “palatability” (Blaxter et al., I96 1 ; Campling, 1964), but it is now accepted that forage intake is mainly controlled by largely involuntary physiological reflexes within the animal, rather than by its subjective liking for different feeds. The development of some of the current concepts on voluntary intake have been reviewed by Balch and Campling (1962), Conrad ( 1 966), and L. D. Brown (1966). From these a broad distinction appears between the factors determining intake by ruminants and by nonruminants. Intake by nonruminants is controlled mainly by levels of blood metabolites, the animal ceasing to eat when these reach a threshold level. Intake by ruminants depends much more on the capacity of the digestive tract, particularly the rumen, eating ceasing when a certain degree of “fill” has been reached, and starting again when “fill” has been reduced by digestion and movement of food residues through the digestive tract: only on feeds of high energy concentration does blood metabolite level, rather than gastrointestinal fill, begin to control the amount of food that ruminants will eat (Conrad, 1966). However, as with other aspects of forage nutritive value it is essential to recognize that the amount of forage that animals will eat is likely to be determined by a complex of factors. Some earlier investigations may have oversimplified the problem; the observation by Blaxter et al. ( I96 1) that the newer concepts related to digestibility and rate of passage of foods are “attributes which are hardly consonant with their acceptability to the palate or taste” perhaps dismissed too lightly the possible significance of these other attributes. It is evident also that the amount of forage that animals eat may depend as much on the amount of forage available as on any inherent characteristics of the forage itself. Raymond (1966a) has proposed that the factors determining forage intake can be usefully divided into intrinsic factors (i.e., features inherent in the forage)

28

W . F. RAYMOND

and extrinsic factors, which depend on the method of presentation of the forage, on the effect of processes such as ensilage and dehydration, and on environmental factors. The intrinsic factors determine how much of a forage a ruminant animal could eat under ad libitum conditions; the extrinsic factors determine how much the animal is able to eat under the particular feeding conditions imposed upon it. B. INTRINSIC FACTORS DETERMINING FORAGE INTAKE It is first necessary to emphasize the great importance of rigid standardization of the conditions under which intake measurements are made, if comparisons are to be made between results from different centers (Chalupa and McCullough, 1967). Voluntary intake is generally defined as the amount animals will eat when an excess of 15 percent is offered (Blaxter et al., 1961). But fresh feed should be presented at least twice daily and uneaten residues removed to avoid soiling this fresh feed, so that limitation of intake by extrinsic factors is avoided. Ruminants are also found to vary much more in their capacities for feed intake than, say, in their digestive capacities, and adequate numbers of animals must be used to obtain reliable intake data. Heaney et al. (1968) summarized CV’s* of individual animal intakes ranging from 10.5 percent (Minson et al., 1964) to 16.4 percent in their own experiments. Complicating factors are also the species and size of the experimental animals, and most intake results are now reported in terms of the metabolic weight of the animal, expressed as liveweight to the power 0.75 (although agronomists must have detected that some of their colleagues in ruminant nutrition still prefer 0.73). In this way the voluntary intakes of both sheep and cattle could be expressed on a comparable basis (Blaxter and Wilson, 1962). But it must be accepted that intake cannot be a precise parameter of a forage, because the amount that an animal eats depends on the individuality of the animal (its species, sex, physiological status, health, etc.) as well as on the intrinsic features of the forage. 1 . The Relation between Forage Digestibility and Intake It is now accepted that the major factor limiting the amount of forage eaten by ruminants is the capacity of the rumen and digestive tract. Ruminants are able to eat much more of highly digestible forages than of less digestible forages, because the latter occupy more volume and are within the rumen for a longer time and because from them more indigestible residue has to be passed down the hind tract (Balch and Campling, 1962). A decrease in voluntary intake as forage becomes more mature, *Coefficient of variation.

THE NUTRITIVE VALUE OF FORAGE CROPS

29

and so less digestible, has been shown in many experiments both with temperate forages (Crampton et a / . , 1960; Minson et al., 1964; Demarquilly, 1966b; Osbourn et al., 1966; Heaney et al., 1966) and with tropical forages (Grieve and Osbourn, 1965; Milford, 1967; da Silva and Gomide, 1967). These results appeared to confirm the concept that rumen fill controls voluntary intake. Unfortunately this led to the more generalized concept that the voluntary intake of a forage could thus be predicted from its digestibility, and that a single forage parameter, related to digestibility, would be adequate to estimate both digestibility and intake, the two main components of nutrient intake. There is now increasing evidence that this is a too simplified concept. While with most forage species intake decreases as the forage becomes less digestible, the relationships between intake and digestibility can differ markedly between different forages. Therefore, different forage species at the same level of digestibility may be eaten in quite different amounts, a fact that is in conflict with the earlier concept. This first became apparent with the observation by several workers of higher intakes from legumes than from grasses of the same digestibility (e.g., Van Soest, 196%; Osbourn et a / . , 1966; Milford, 1967; Weston and Hogan, 1967) and subsequently of different levels of intake between grass species. Thus, low intakes have been reported with timothy (Phleum prarense) (Minson e f al., 1964; R. L. Reid and Jung, 1966; Miles and Walters, 1966), with tall fescue ( F . arundinacea) (Van Soest, 19641, and with Phalaris arundinacea (0. N. Andrews and Hoveland, 1965; O’Donovan et al., 1967) compared with other grass species. With timothy the rate of fall in intake with decreasing digestibility was found to be less than with ryegrass (Minson et al., 1964), and Heaney et a / . ( 1966) found only a negligible change in intake with decreasing digestibility of Phleum nodosum. Within a species, differences in intake characteristics have also been shown. Thus Osbourn et al. ( I 966) found the intake of a diploid variety of Italian ryegrass to be 16 percent greater than that of a tetraploid variety during first growth in the spring. Large differences in intake between different lines of Phalaris arundinacea have been shown by Roe and Mottershead ( 1962) and O’Donovan et al. ( 1967). 2 . Differences in Intake between Forage Species Observations of different intake-digestibility relationships for different forages are of considerable importance. As long as voluntary intake was accepted as being determined mainly by level of digestibility, there appeared to be little prospect of improving the nutritive potential (intake X digestibility) of forages except by an improvement in digesti-

30

W. F. RAYMOND

bility. The evidence that factors in forage other than digestibility can also affect intake now offers a much wider scope for the improvement of forage nutritive value. Two possible lines of development are indicated in the work of Osbourn et al. (1966) and O’Donovan et al. (1967), noted above. Thus Osbourn et al. (1966) and Osbourn (1967) showed marked differences in voluntary intake, in the order lucerne > ryegrass > timothy, at a given level of digestibility. Chemical analysis of these forages showed that the “digestible” fraction in lucerne contained a higher proportion of pepsin-soluble material, and a lower proportion of digestible fiber than the “digestible” fraction in timothy, with the levels in ryegrass intermediate between those for lucerne and timothy (Fig. 3).

PEPSIN SOLUBLE MATERIAL

DIGESTIBLE FIBER

INDIGESTIBLE MATERIAL

LUCERNE Voluntary intake qDM/kqO’~dldoy

82

S 24

5 48

73

bl

FIG. 3. The composition of the digested fraction, and the voluntary intakes of lucerne, S.24 ryegrass, and S.48 timothy forages of the same dry matter (D.M.)digestibility. (From Osbourn, 1967.)

Van Soest ( 1 965c) reported a similar conclusion, that lucerne contains a higher proportion of cell contents (material soluble in neutral detergent) and a. lower proportion of cell wall constituents, than grass of the same level of digestibility. While the grass and lucerne forages are of the same digestibility, it is likely that the lucerne will reach the “digested” stage more rapidly than the ryegrass, and the ryegrass than the timothy. The digestible fraction of the lucerne could thus occupy less volume X time within the rumen; as a result the animal could eat more of it than of the grasses. Osbourn ( 1 967) also reported that, in the experiment described by Osbourn et al. (1966), the diploid ryegrass had a higher content of pepsin-soluble material than the tetraploid ryegrass, which could have been associated with the higher intake of the diploid variety.

THE NUTRITIVE VALUE OF FORAGE CROPS

31

The observations illustrated in Fig. 3 offer one possible explanation as to why different forages of the same level of digestibility may be digested at different rates, and so have different intake levels. But there is also the possibility that rate of digestion, and in turn the rate of intake, may be affected by conditions within the rumen. As Tilley et al. ( 1 964) have shown (Table 1) the rate and extent of cellulose digestion decreases as the pH of an in vitro system (and by analogy of the rumen in vivo) diverges from the physiologically normal level of about pH 6.8. Some forages, particularly highly buffered, low-sugar, forages such as the legumes, are found to give a characteristically higher rumen pH (6.6 to 6.8) than perennial ryegrass (<6.5), and Raymond (1966a) has suggested that these differences in rumen pH could lead to a higher rate of digestion of the cell wall fraction of lucerne than of ryegrass, and that this could partly account for the higher level of intake of lucerne and of other legumes. This introduces the possibility of developing feeding regimes which, by maintaining rumen pH close to the optimum level of 6.8, could ensure the most rapid rate of digestion of the cell wall content in the feed, with a resulting high level of voluntary intake. The results of Roe and Mottershead (1962) and O’Donovan et al. (1967) have indicated that other factors may also modify the simple relationship between forage digestibility and forage intake. These workers showed that, within the species Pharlaris arundinacea, sheep showed a marked preference for certain strains compared with others. An extract from one low-preference strain, sprayed on a higher-preference strain, made it unacceptable to the sheep (Roe and Mottershead, 1962). O’Donovan el al. (1967) also established differences in intake between lines of this grass, and suggested that it might be possible to breed lines of improved intake characteristics. No intake problem appears to have arisen with P . tuberosa (0.N. Andrews and Hoveland, 196% but it is of interest that the Aberystwyth variety S.230 ( P . tuberosa X P. arundinacea) was found to be very unpalatable to sheep (U. I. Jones, 1959). Considerable differences have been found between the intake-digestibility relationships for different forage species but different workers are not in agreement about the form of these relationships. Blaxter and Wilson (1962) found a curvilinear relationship, intake increasing less rapidly with digestibility at higher levels of forage digestibility. J. B. Hutton ( 1963) and Conrad et al. ( 1 964) found no increase in intake above a forage dry matter digestibility of about 7076, and Baumgardt ( 1 967) found that the intake of a roughage-concentrate mixture did not increase when the energy digestibility was above 67 percent (ca. percent dry

32

W. F. RAYMOND

matter digestibility). In contrast Osbourn et al. (1966, and unpublished) have found no divergence from linearity with a wide range of grasses and legumes up to 80 percent dry matter digestibility. The results of J. B. Hutton (1963) and Conrad et al. ( 1 964) suggested that at high levels of forage digestibility intake begins to be limited by metabolic factors (blood levels of organic acids, glucose, etc.) rather than by gastrointestinal fill. That is, the ruminant begins to behave, in intake terms, like a nonruminant (Conrad, 1966). Freer and Campling (1963) also showed that the intake of (highly digestible) concentrated feeds was not limited by bulk. Van Soest ( 1965c) has suggested that this is likely to occur when forage dry-matter contains less than 5 5 to 60 percent of cell wall constituents. However, the level at which forage digestibility no longer limits intake must to some extent depend on the physiological condition of the animal being fed, i.e., on the critical level of blood metabolites at which it ceases to eat. The experiments of Osbourn et al. ( 1 966) were all carried out with “thin” sheep, with a high growth potential, which may therefore have had a higher threshold intake level than some other experimental animals, such as the mature sheep used by Blaxter et al. (1 961). Critical studies to compare the voluntary intake levels of the same forage by ruminants in different physiological states are needed: particular examples would be animals in early or late lactation, or subjected to climatic stress.

3. The Effect of Fertilizers on Forage Intake In many experiments initial observations of differences between forages have been made in “preference-ranking” tests in which sheep were offered a free choice from a number of different forages. In only few experiments have differences in voluntary intake characteristics then been established by feeding high and low preference strains separately to sheep under ad libitum conditions. This illustrates a most important feature in the testing of forage intake, that differences in preference ranking must be confirmed in intake experiments before superiority of the higher-ranking forages can be accepted. Thus Bland and Dent (1 962) reported a positive correlation between the preference ranking and the content of water-soluble carbohydrates in a number of varieties of cocksfoot, and Plice ( 1952) suggested that the low preference ranking of grass fertilized with nitrogen was due to the lower content of sugar in such grass. But experiments under ad libitum feeding have shown no effect on voluntary intake from widely different levels of N fertilization of grass. Thus Marten and Donker (1964) found no differences in voluntary intakes of Bromus inermis fertilized with no nitrogen (24 percent

THE NUTRITIVE VALUE OF FORAGE CROPS

33

crude protein; 7.2 percent ethanol-soluble sugar) or with 300 pounds of N per acre (28 percent and 5.4 percent). Likewise, R. L. Reid and Jung (1965) found similar levels of intake from Festuca arundinacea grown with or without nitrogen fertilizer, although sheep showed a marked preference for the low-nitrogen grass. Holmes and Lang (1963) and Hight er al. (1968) have also reported that nitrogen fertilization had no effect on the voluntary intake of grass fed to either dairy cows or sheep. Cameron ( 1967) found similar intakes of hay fertilized with 0 or 1 12 kg. of N per hectare. We cannot conclude from these results that nitrogen fertilization never decreases the voluntary intake of forage - the reports of feed refusal and poor animal performance sometimes experienced in practice cannot be ignored- but experimental confirmation is still lacking. And in other cases, described in Section V, D, there is clear evidence that the intake of forages of very low nitrogen content can be increased by nitrogen fertilization. Greenhalgh and Reid (1967) have suggested that the relative importance of “rumen fill” and “palatability” in determining the voluntary intake of forages may be studied by introducing feed directly into the rumen via a fistula. An equivalent decrease in the intake per 0s would demonstrate that intake is being controlled by rumen fill. If intake per 0s remains at the original level some other factor, such as palatability (or more strictly, unpalatability) would be indicated. When ad libitum chopped straw was fed to sheep per os, and an equal quantity of grass was introduced through the fistula, total daily intake was 23.5 g. dry matter per kilogram liveweight to the power 0.73, but when the feeding regime was reversed, total intake increased to 48.8 g., indicating that the amount of straw the sheep ate was not limited by rumen fill only. However, further examination of this technique, and in particular of the effect of the absence of mastication and salivary secretion on the feed introduced via a fistula is still needed. Sonneveld (1965) has suggested that forage intake by cattle may be related to the moisture content of the forage. Herbage was cut daily and fed to cattle throughout a complete year, and the voluntary intake on each day was related to a number of forage parameters by multiple regression. Intake on any day was slightly related to the level of forage digestibility, but more closely to day-to-day changes in moisture content. lntake was found to increase with a decrease in forage moisture content, relative to the previous day’s feed, when forages were at a low fiber stage, but the reverse was found when the forage was of high fiber content. Conversely Holmes and Lang (1963) found that intake did not change when forage moisture content increased to 84 percent, compared with 80 percent in

34

W. F. RAYMOND

the control forage, following nitrogen fertilization. The difference between these observations indicates a need for further study. While the higher intake of legumes than of grasses appears, at least in part, to be due to the higher proportion of cell contents in the legumes, this explanation does not appear to account for differences in intake which have been found between different legumes. Thus Van Soest ( 1 965a) reported a higher intake of birdsfoot trefoil (Lotus corniculatus) than of lucerne of the same level of digestibility, and Osbourn et al. (1 966) found a similarly higher intake of sainfoin than of red clover or lucerne. In the latter experiment the content of pepsin-soluble material in the red clover (36.4%)was higher than in the sainfoin (33.5%). Troelsen and Bigsby (1964) had demonstrated an inverse relationship between the resistance to maceration and the voluntary intake of a series of hays; part of this relationship is likely to be due to a direct association of physical strength of plant materials with increasing fibrousness (and so to decreasing digestibility), but Osbourn et al. (1 966) suggested that there might remain differences in intake, as between sainfoin and red clover. A similar concept is seen in the fibrousness index of Chenost ( 1 966), which is measured as the electrical energy needed to pulverize a 5-g. sample of the dried forage, and which was shown to be closely related to dry matter intake. Evans ( 1 964) has shown differences in breaking strength of grass leaves, and Theron and Booysen (1966) have related breaking strength positively to the acid-insoluble lignin content of 8 grass species. Johnston (1967) has shown that the force needed to penetrate the stems of kale was closely related to in vitro digestibility values. L. H. P. Jones and Handreck (1967) raise the interesting possibility that variations in the degree of silicification in grasses may contribute to variations in culm strength. In addition grass may become less palatable to the grazing animal as a result of heavily silicified hairs on the leaf surface. These concepts of the intrinsic factors that may affect forage intake represent an important advance from the earlier, oversimplified, concept that forage intake is mainly determined by the level of forage digestibility. The demonstration that forages of the same digestibility can differ markedly in level of voluntary intake, and the elucidation of some of the possible causes for this, now indicates the prospect of breeding for improved forage intake. Any argument that the extreme cases noted are exceptional is surely irrelevant. Thus the observation, in Phalaris arundinacea, of strains of high and low preference rating suggests that there are also strains of intermediate preference. More important, knowledge of the factors causing these differences in preference should

THE NUTRITIVE V A LU E OF FORAGE CROPS

35

allow the breeding of new strains of even higher preference rating. “Palatability” may be an imprecise concept, but the observation of extreme unpalatability in some species indicates the likelihood of intermediate unpalatability in other species, which it would be unwise to dismiss. For a difference of only 10 percent in voluntary intake between two forages could have as great an effect on their relative values for animal production as the difference of 6 units in digestibility between S.24 ryegrass and S.37 cocksfoot, about which much has been written. C. THENUTRITIVEVALUEINDEX With the recognition that both voluntary intake and the level of digestibility of a forage play major roles in determining its nutritive value, Crampton et al. (1960) suggested that intake and digestibility should be combined in a single parameter, the nutritive value index. The concept was possibly complicated by relating the actual intake of a given forage, expressed per kilogram liveweight to the power 0.75 (kg. LW0.7s), to a theoretical standard forage of intake 80 g. dry matterlkg. LW0.75,the relative intake being calculated: Relative intake =

actual intakelkg. LW0.75 80

and the nutritive value index: NVI

= relative

intake X

% digestibility of forage energy content I00

( 1 1)

Crampton er al. (1960) proposed, from initial studies with 9 forages, that the rate of cellulose disappearance in a 12-hour standard in vitro digestion ( x ) could be closely related to the NVI: (12) Donefer et al. ( 1 966) subsequently reported an improved correlation between NVI and the percentage of the forage dry matter soluble in acid pepsin (P): NVI

=

1.3 1 4 ~ 7.8

NVI = 1.60P - 0.75 ( r = 0.95**)

(13)

In reporting these results, Donefer er al. (1966) suggested that both of the measures used, rate of cellulose digestion and pepsin solubility, are likely to be more closely related to the intake characteristics of forages than to digestibility. The high correlations with N V I result from the feature, already noted, that intake changes considerably more with changes in forage maturity than does digestibility (Ingalls et al., 1965). The relationship of pepsin solubility to voluntary intake is in line with the

36

W. F. RAYMOND

conclusions of Osbourn et al. (1966), illustrated in Fig. 3. However the relatively poor correlation of pepsin solubility with dry matter digestibility (Terry and Tilley, 1964b; Donefer et al., 1966) indicates that a more precise estimate of NVI might be obtained by a combination of analyses for pepsin solubility (relative intake) and in vitro digestibility. In that case, however, the separate estimates of intake and digestibility could be of more use than the single parameter of N V I , by indicating the relative importance of the differences in intake and in digestibility among the forages being studied. D. THE CRUDEPROTEINCONTENTOF FORAGE AND VOLUNTARY INTAKE

While nitrogen fertilization generally seems to have little effect on the intake characteristics of forages, an exception occurs in the case of forages of very low crude protein content, for which rate of digestion within the rumen, and so level of intake, seems to be limited by the lack of nitrogenous substrates for the rumen microorganisms. The critical level of feed protein depends on the type of forage, but it is commonly in the range of 4-6 percent crude protein. Thus Minson (1 967) showed a 54 percent higher intake of pangola grass (Digitaria decumbens) fertilized with 250 kg. of urea per acre (7.2 percent crude protein in the forage dry matter) than at 100 kg./acre (3.7 percent crude protein). Minson and Milford (1967) showed that the intake of the same species (at 3.6 percent crude protein) was increased when a supplement of lucerne was fed, to a maximum intake at a total diet content of 6 percent crude protein, after which the intake of the grass decreased as it was progressively replaced by lucerne. In most experiments the available feed nitrogen has been increased indirectly by supplementary feeding, particularly with' sources of nonprotein nitrogen, such as urea. This is discussed more fully in Section V , D; but in relation to forage intake there is evidence that the intake of low-protein forages can be considerably increased by feeding urea. Thus Campling et al. (1962) found a 40 percent increase in the intake of oat straw (3.0 percent crude protein), and Graham (1 967) found a 50 percent increase with Digitaria decumbens (4.0percent crude protein), resulting from urea supplementation. However, at a higher level of feed protein (5 to 7 percent), Kay et al. (1968) found no significant improvement in the intake of barley straw when supplementary nitrogen was fed. Weston (1967) showed that at very low levels of crude protein the main factor limiting the intake of wheaten hay was nitrogen deficiency; when this was remedied, intake was then limited by the rate of removal of indigesti-

THE NUTRITIVE VALUE OF FORAGE CROPS

37

ble residues from the rumen. However, Egan and Moir (1965) have queried whether the intake of low-nitrogen hay is limited solely by deficiency of feed nitrogen within the rumen, as infusion of casein directly into the abomasum or duodenum was found to increase the intake of such hay. This source of nitrogen could have benefited bacterial digestion within the rumen only indirectly, and these authors suggested that the amount of protein reaching the duodenum may itself affect voluntary intake. E. THEEFFECTOF

SUPPLEMENTARY

FEEDSON FORAGEINTAKE

The use of nonprotein.nitrogen compounds is one aspect of the larger subject of feed interactions, including the effect of other components of the total ration on the amounts of different forages that the ruminant will eat. Most studies to date have been essentially practical and have recorded mainly the changes in forage intake when supplements are fed. As with the effects of supplementary feeding on forage digestibility, noted in Section IV, C, 3 , an understanding of the biological basis of these interactions is essential if supplementary feeding is to be other than an empirical procedure. As might be expected, the most general observation is that, as increasing amounts of supplementary concentrates are fed, the ruminant animal eats less forage. This decrease in forage intake appears to be more marked with forages of high digestibility (i.e., of initially high intake) than with forages of lower digestibility and intake (Blaxter and Wilson, 1963; Campling and Murdoch, 1966). The considerable reduction in the voluntary intake of forages of high digestibility may well be associated with the observations, noted in Section V, B, 1, that such forages tend to behave like concentrated feeds, so that their voluntary intake may be determined by levels of blood metabolites, rather than by rumen fill; in that case level of forage intake might be expected to decrease as a result of the enhanced levels of blood metabolites following concentrate feeding. The reduction in intake of forages of lower digestibility may in part be related to the effects of supplementary feeding on rumen pH, noted in Section IV, C, 3. As the level of concentrate feeding increases, rumen pH tends to decrease (P.K. Briggs et al., 1957; Topps et al., 1965) and it is possible that this causes a reduced rate of digestion of the cell wall constituents in the forage (Tilley et al., 1964), and so a reduced level of voluntary intake. The depression in rumen pH could perhaps be greater with concentrates of low protein content (containing mainly cereals) than with those of higher protein content, and this may partly account for

38

W. F. RAYMOND

the greater reduction in forage intake when low-protein concentrates are fed. However, the most important feature is evident, that in many cases supplementary feeds partly replace rather than supplement the forage with which they are fed; the significance of this in determining the economic response to supplementary feeding emphasizes the urgent need for more detailed studies of these food interactions. VI. The Efficiency of Utilization of Digested Nutrients

The previous sections have examined the factors that determine the quantity of digested nutrients (intake x digestibility) made available to the ruminant; the following sections consider the efficiency with which it utilizes the various components within these digested nutrients, in particular the energy, protein, and minerals, as well as other components in forages which may modify their potential as practical ruminant feeds. A. METHODSOF EXPRESSING ENERGYVALUES 1 . Digestible Energy

Most results of digestibility experiments with forages have been reported in terms of the digestibility of the dry matter or of the organic matter in the forage (Eq. 1) or of its content of total digestible nutrients (TDN), which effectively measures the percentage of digestible organic matter in the forage dry matter (Minson et a f . , 1960). Digestibility is primarily an index of the energy value of a feed. Crampton et al. ( 1 960) have proposed that all results should be reported in terms of digestible energy by measuring C (Eq. 1) in kilocalories; TDN data already published could be converted on the basis of 4.4 kcal. of digestible energy per gram of TDN. The latter factor can only be approximate, as it varies with the crude protein and ether extract contents of the forage (Heaney and Pigden, 1963), and direct calorific determinations on feed and feces samples provide the only reliable method. 2 . Metabolizable Energy All the expressions of digestibility based on Eq. (1) measure the “apparent” digestibility, rather than the “true” digestibility of the feed energy, for the feces contain metabolic residues not derived from the feed (Section 111, C ) . But these are in practice more useful measures than “true” digestibility, because these metabolic residues represent an inevitable energy loss associated with the processes of digestion and

THE NUTRITIVE VALUE OF FORAGE CROPS

39

assimilation of the feed. Energy is also lost in the methane produced from microbial digestion of carbohydrates within the rumen, and in the waste products excreted in the urine. The metabolizable energy (ME) in the forage: ME = digestible energy - methane energy - urine energy

(14)

thus more usefully expresses the proportion of the energy in the feed which is available for the metabolic and productive activities of the animal. To measure metabolizable energy precisely requires, in addition to total collection of feces and urine, the measurement of methane production in a respiration chamber. For many purposes, however, an estimate of methane production: Methane (g.) = 2.4 I X

+ 9.80 (sheep; Swift ef al., 1948)

(15)

where X = grams of carbohydrate digested + 100, based on data from respiration chamber experiments, is likely to be sufficiently accurate; but wherever possible this should be checked, as Eq. (15) may not be valid in novel feed situations -for instance, when milled dehydrated forages are fed to ruminants (Blaxter and Graham, 1956). Furthermore, Flatt ( 1 966) has shown that methane and urine production is not constant, but decreases as the level of feeding of a particular feed is increased. As noted earlier, digestibility decreases as level of feeding increases (e.g., Moe et al., 1965). These changes in digestibility and in methane and urine production tend to compensate. As a result Flatt ( 1 966) found that the metabolizable energy content of a feed changes relatively little with change in the level at which it is fed, but this conclusion is not in complete agreement with that of Blaxter (1962), who concluded that metabolizable energy decreases with an increase in level of feed intake. Such divergences are in fact to be expected, because metabolizable energy is not thefixed characteristic of a feed that is sometimes implied. Urine energy, in particular, reflects the physiological status of the animal, the animal with a high positive nitrogen balance (e.g., the dairy cow, used by Flatt, 1966) giving a lower urine energy loss than one with lower nitrogen balance (e.g., the mature sheep, used by Blaxter, 1962). Because of this the dairy cow may show less decrease in the metabolizable energy of a feed than the mature sheep at feeding levels above maintenance. In general terms, Armstrong et af. ( 1964b) and Graham ( 1964) found that the metabolizable energy of grass was about 81 percent of the digestible energy content, and could therefore be calculated from its energy digestibility measured in vitro.

40

W. F. RAYMOND

3 . N e t Energy The classical studies of Kellner and Armsby established that as feeds become less digestible their content of metabolizable energy is less efficiently used by the ruminant animal. The methods of expression of useful energy used by the European workers (starch equivalent) and the North American workers (net energy) were different, as were some of the underlying concepts, but for some purposes the two systems can be equated, 1 pound of starch equivalent being taken as equal to 1,071 kcal. of net energy. These systems of feed evaluation accept that when the animal digests and metabolizes food there is a production of heat within the body which represents a net loss of energy that is an inevitable part of the utilization of the food (an exception is when the animal is in a very cold environment, and can use this heat production to maintain body temperature instead of needing to oxidize useful food for this purpose). In precise terms, net energy (NE) is an incremental measure: N E = (ME,

-

ME,)

-(HB

- HA)

(16)

where ME and H are the metabolizable energy and heat production from the same feed fed at levels A and B. But for practical feed evaluation it is treated: Net energy = (metabolizable energy) - (heat production) (17) As feeds become less digestible an increasing amount of heat is produced for each unit of metabolizable energy derived from the feed, so that the net energy falls more rapidly than metabolizable energy or total digestible nutrients content (L. A. Moore et al., 1953): N E = 1.393 x T D N

-

34.63 ( r = 0.977)

(18)

A similar result is achieved in the starch equivalent system by making a correction, based on the crude fiber content of the feed, which is thus inversely related to the level of digestibility. This wastage of heat was considered to be due to the energy of digestion, and the increase in heat production with less digestible feeds was attributed to the increased energy needed to eat and digest such feeds. This somewhat imprecise concept, challenged by recent developments in knowledge of ruminant digestion and metabolism, does still appear to account for much of the difference in energy value between feeds. B. THEROLEOF VOLATILE FATTY ACIDSIN RUMINANT METABOLISM 1 . The U s e of Volatile Fatty Acids as Energy Metabolites It has long been known that among the main products of microbial

THE NUTRITIVE VALUE O F FORAGE CROPS

41

fermentation within the rumen are steam-volatile fatty acids, in particular acetic, propionic, and butyric acids, with smaller amounts of isobutyric and longer-chain acids. Early in the 1950’s it was shown that these acids are absorbed from the ruminant digestive tract and provide a major source of energy metabolites for the animal (reviewed by Annison and Lewis, 1959). Recent work has suggested that some 70 to 80 percent of the energy supply of the ruminant on normal feeds is provided by volatile fatty acids (A. C. I. Warner, 1964). These acids are produced by fermentation of a very wide range of feed substrates-sugars, cellulose, hemicellulose, proteins, etc. As I. W. McDonald ( I 968) has emphasized, digestion within the rumen serves the remarkable function of converting a most diverse range of chemical constituents in the feedstuffs eaten by the ruminant into the relatively simple range of compounds that it absorbs and uses as metabolites. The acid produced by fermentation of feed within the rumen comprises mainly acetic, propionic, and butyric acids, but it was found that the proportions in which these are produced differs among classes of feeds: in general, as feeds become less digestible the combined proportions of propionic and butyric acids decreases, and that of acetic acid increases (Rook and Balch, 196 1). This observation assumed particular significance with the demonstration by Blaxter and his associates (Armstrong and Blaxter, 1957; Blaxter, 1962) that these three acids appear to be used by the ruminant with different efficiencies for different metabolic purposes. In a detailed series of calorimetric experiments these workers showed that, when these acids were infused into the rumen in vivo, they were used with equal efficiency for maintenance purposes (respiration, circulation, etc.), but that acetic acid was much less efficiently used for lipogenesis (body fat synthesis) than propionic and butyric acids. These results indicated that the decrease in efficiency ofuse of metabolizable energy as feeds become less digestible might at least in part be the result of theassociated shift in rumen fermentation products from propionate plus butyrate on highly digestible feeds to predominantly acetate on feeds of low digestibility. The observation that the three acids were equally efficiently used for animal maintenance purposes led Blaxter ( 1 962) to propose a system of feed evaluation in which the efficiency of utilization of metabolizable energy for maintenance (K,J is assumed to be constant, and the same for all feeds, but the efficiency of utilization for lipogenesis (fattening KJ)decreases as feed becomes less digestible, that is: K , = constant

K j = 0.8lQ + 3.0

(19)

42

W. F. RAYMOND

where Q is the ME, expressed as a percentage of the gross energy in the feed. The concept that the efficiency with which the ruminant uses its digestible and metabolizable energy intake is largely determined by the pattern of volatile fatty acid production within the rumen has had a most profound effect on ruminant research in the last decade. In many experiments in which feeds or rations have been compared, higher levels of animal production have been recorded on the feed or ration giving the higher proportion of propionate plus butyrate in the total rumen acidsand the conclusion has been drawn that the higher level of production was the result of the higher proportion of these acids. But this elegant explanation of differences between ruminant feeds contains anomalies, and its generalized validity must now be questioned. First, the original infusion experiments examined the efficiency of utilization of the different acids for lipogenesis in the mature sheep; they cannot equally well explain the decrease in efficiency of bovine milk production with decreasing level of feed digestibility, for as Rook and Balch (196 1) indicated, acetic acid, the predominant acid produced in the rumen from feeds of low digestibility, is more efficiently used for milk production than is propionic acid. Neither is there any evidence that acetic acid is less efficiently used than propionic or butyric acids for body protein synthesis, which comprised only a small part of the body-energy gain of the sheep used in the experiments reported by Blaxter; in fact Rook and Balch ( 1 96 1) indicated that efficiencies may be in the reverse order. Second, the conclusion that the ratio of propionic plus butyric acids to acetic acid is closely related to feed digestibility does not hold for all feeds: Terry and Tilley (1962) found very little relationship between rumen acid pattern and forage digestibility, and showed that the molar proportion of propionic acid in the rumen acids ( P ) was more closely related to the soluble carbohydrate content (C) in a range of forages: P

= 0.48C

+ 16.5

( r = 0.75)

(2 1)

As forages of the same digestibility can differ markedly in their content of soluble carbohydrate, a system of feed evaluation, based on rumen acid proportions, would indicate a more efficient utilization for lipogenesis of the metabolizable energy in a high-sugar than in a low-sugar forage of the same digestibility. Yet Thomson (1964) found identical energy gains when ryegrass samples of high or low sugar contents were fed to lambs. There was also an indication that a higher proportion of the energy gain by the lambs on the low-sugar grass (7 percent soluble carbohydrate, 29 percent crude protein) was in the form of body protein

THE NUTRITIVE VALUE OF FORAGE CROPS

43

than on the high-sugar grass (19 percent and 16 percent), a result which could be of economic significance. For, paradoxically, the one form of animal production, lipogenesis, for which propionic and butyric acids have been indicated as superior to acetic acid, is the one least required by the meat industry and the consumer. Third, while the measured differences in the proportions of volatile fatty acids between rations have generally been in the predicted direction, they have also nearly always been much too small to account for the observed differences in animal production. Blaxter and Wainman ( 1 964, Fig. 4) showed a decrease of 1.091 in K f for each 1 percent increase in the molar proportion of acetic acid in the rumen acids. In a comparison of immature and mature ryegrass feeds, Armstrong ( 1 960) found the Kf value of the immature feed to be 26 percent greater than of the mature feed; the corresponding molar proportions of acetic acid in the rumen acids, 61.0 and 65.4, would indicate a difference of only 5 percent in KJ values. On the basis of these and other experiments, in which no overall relationship could be established between the molar proportions of rumen acids in animals fed on different grass feeds and the efficiency of lipogenesis, Armstrong (1964) concluded that mechanisms, in addition to that based on rumen volatile fatty acids, must be sought for the observed differences in energy value between feeds. Finally the concept of energy metabolism based on the rumen acids indicated that the metabolizable energy of all feeds should be equally efficiently used for maintenance, and that differences would occur only in their use for productive purposes. However, Brouwer et al. (1961) reported that the requirement of metabolizable energy for maintenance of the 500-kg. cow increased from 11.3 Mcal. with an early-cut hay to 12.7 Mcal. for a mature hay. From a recalculation of earlier calorimetric experiments, Armstrong ( 1964) concluded that the efficiency of utilization of metabolizable energy for maintenance (K,,J decreased with a decrease in feed digestibility (metabolizable energy concentration in the feed, Q ) :

K,,, = 50.9 + 0.3759

(22)

From data on a wider range of feeds, Blaxter (1964) derived the relationship: K,,, = 54.8

+ 0.30Q

(23)

and Graham (1964) also concluded that metabolizable energy is less efficiently used for maintenance as feeds become more fibrous, i.e., as Q decreases.

44

W. F. RAYMOND

“It seems clear then that the earlier assumption that the metabolizable energy of all feeds is used with equal efficiency for maintenance purposes involved too wide an extrapolation of the results of experiments with rumen infusions of pure volatile fatty acids. In effect it implied that a constant proportion of the M.E. of feeds of widely differing digestibilities is absorbed from the rumen as volatile fatty acids. Armstrong ( 1 964) has suggested that, with diminishing digestibility in a series of feeds, an increasing fraction of the M.E. is lost as heat of fermentation, and a lower proportion absorbed as useful energy” (Raymond, 1966b). It is also most likely that maintenance requirement itself will effectively increase as a result of the greater muscular energy involved in chewing and movement of the larger quantity of more fibrous feed required to provide a given amount of metabolizable energy from a low than from a high digestibility feed. These mechanisms, viz. the decreased proportion of metabolizable energy absorbed as volatile fatty acids and the increased energy of digestion as feeds become less digestible, could both also be expected to occur in the utilization of metabolizable energy above maintenance. This increased wastage of energy might thus be expected to contribute to the coefficient 0.8 1 in Eq. (20); the remainder of this coefficient may then be accounted for by less efficient lipogenesis as a result of the increasing proportion of acetic acid in the rumen acids when feeds of low digestibility are fed. This analysis of the original proposals of Blaxter (1962) offers some explanation of the discrepancies noted. Thus both Kf and K , can be expected to decrease with decreasing feed digestibility; the quite small differences in rumen acid proportions often found could be adequate to explain the residual differences in Kf between feeds of different digestibilities: the decreasing efficiency of the lactating animal with decreasing feed digestibility would be due to the dominance of losses of energy in digestion, despite the higher efficiency of use of acetic acid for milk production (Rook and Balch, 1961). On this basis, the decreasing efficiency of utilization of metabolizable energy as feeds become less digestible would result mainly from the decrease in digestibility per se (as in the original Kellner system) with a further decrease in efficiency for lipogenesis due to the less favorable rumen acid pattern, which is not implicit in the Kellner system. This concept would place emphasis on level of feed digestibility, rather than on rumen acid pattern, as the main determinant of the eficiency with which the digestible and metabolizable energy in feeds are utilized by the ruminant. Yet since 1964 the number of publications in which differences in

THE NUTRITIVE VALUE OF FORAGE CROPS

45

nutritive value between feeds have been attributed mainly to differences in rumen acid patterns has increased, rather than decreased. It would be invidious-and unhelpful - to summarize them here; but it does appear that emphasis on this mechanism has diverted attention from differences in feed digestibility and feed intake, which have often been neither controlled, nor measured, in these experiments. This comment implies, not that the rumen acids are not important, but that interpretations of animal production results based on rumen acid patterns must be considered more critically. First, it is important, where rumen acid proportions are measured, that this is done accurately. Many experiments have relied on single samples of rumen fluid taken at a fixed time after feeding; but the concentration and proportions of the volatile fatty acids in rumen fluid can vary markedly throughout the day (Terry and Tilley, 1964b), reflecting the pattern of feeding of a given forage (I. H. Bath and Rook, 1963; Faichney, 1 9 6 8 ~or ) of a forage with concentrate ration (McCullough and Smart, 1968). The more irregular the feeding, the wider the fluctuations in rumen acids and the greater the need for an intensive sampling schedule. 2 . Rumen Condition and Rumen Acid Patterns There is also need for further study of the causes of the particular rumen acid patterns produced when different feeds are digested. Some evidence suggests that these are the result of the conditions established within the rumen by the particular feed and feeding regime adopted, rather than of the digestibility of the feed per se. Thus there was a shift from a higher to a lower acetate:propionate fermentation when the proportion of cereal in a mixed cereal-hay diet was increased, but I. H. Bath and Rook ( I 965) showed that this shift could not be accounted for solely by the increasing digestibility of the diet. I t is known that increasing the proportion of cereal in the diet leads to a decrease in the pH of the rumen content (P. K. Briggs et al., 1957; Topps et al., 1965). This led R. L. Reid et al. (1 957), C . L. Davis et al. ( 1 964), and Raymond ( 1 966a) to suggest that the rumen conditions associated with cereal feeding may favor the microbial production of propionic and butyric acids and discourage the production of acetic acid. This could be the result of a shift either in the bacterial population or in the nature of the end products of fermentation. Much reported evidence is consistent with this hypothesis. Thus C. L. Davis et al. (1964) found that sodium bicarbonate, fed with concentrates plus limited alfalfa hay, increased pH and led to an increased acetate and decreased propionate proportion in the rumen acids. Terry and Tilley (1963) reported a reduction in rumen pH and a de-

46

W. F. RAYMOND

crease in the acetate:propionate ratio when the level of feeding of S.24 ryegrass was increased. McCullough and Smart ( 1968) found that a corn silage-flaked corn mixture, fed in four equal feeds, gave a fairly level rumen pH throughout the day. When the grain and silage were fed separately, the rumen pH and rumen acetate level were both markedly depressed after the grain feeding, and did not increase until the silage was fed. Faichney (1968b) found a wide fluctuation in rumen pH when alfalfa pellets were fed in a single feed, with pH below 6.0 and an acetate: propionate ratio below 3.5 during the period of maximum acid concentration in the rumen. When the same quantity of pellets was fed in eight feeds, rumen pH remained above 6.0 and the acetate:propionate ratio was in the range 5.3 to 5.7 throughout the 24 hours. Von Kaufmann and Rohr (1967) have summarized the changes in rumen acid patterns with changes in rumen pH, measured in their own experiments (Fig. 4) which are in close accord with the foregoing observations.

RUMEN pH FIG.4. Schema of the changes in the molar contents of volatile fatty acids and lactic acid within the rumen at different levels of rumen pH: at low pH levels, either lactic or propionic acid may occur under different conditions. (From von Kaufmann and Rohr, 1967.)

At present these results indicate an association between rumen pH and rumen acid patterns, rather than a direct causation, but they help to explain the conclusion that rumen acid patterns can differ between feeds of the same digestibility: for rumen pH is clearly not solely determined by level of feed digestibility, but depends also on feed composition and pattern of feeding. This has led C. L. Davis et al. (1964) and McCullough and Smart ( 1 968) to suggest that it may be possible to predetermine the pattern of feeding of a given ration so as to shift the

THE NUTRITIVE VALUE OF FORAGE CROPS

47

rumen acid pattern in the required direction. This could be toward a propionate fermentation if body fat synthesis is required. Of much greater significance could be the possible control of milk composition; thus low levels of butterfat are often found in milk produced from rations containing low levels of fiber, and this has been shown to result from the low production of acetate (a precursor in milk-fat synthesis) in the rumen of cows fed on these rations (Balch et al., 1952). The above authors have suggested that by feeding concentrates and limited roughage together (McCullough and Smart, 19681, or by supplementing with sodium bicarbonate (C. L. Davis er d.,1964), mean rumen pH would be increased, and that this might lead to an increased acetate production and increased butterfat levels in milk. Conversely, too low a level of rumen propionate appears to reduce the level of solids-not-fat in milk. Baumgardt (1967) has suggested that the maximum overall efficiency of use of metabolites for milk production occurs with an acetate:propionate molar ratio of 2.75:1, and this could indicate the target toward which feeding regimes should aim; this ratio is very close to that indicated by von Kaufmann and Rohr ( 1967) in Fig. 4. In this perspective, the patterns of rumen volatile fatty acids produced from different forages are clearly of interest and importance. Although, as noted above, these patterns may not be mediated directly through rumen pH, the association seems close enough to indicate that, as the pH at which forage is digested in the rumen decreases, there is likely to be a shift in the products of rumen fermentation from acetic acid to propionic plus butyric acids. Thus forages of high soluble carbohydrate content tend to give the lowest rumen pH and the highest propionate values (Eq. 21). Application of nitrogen fertilizer can decrease the soluble carbohydrate content of grass (McIlroy, 1967). When this grass is fed it gives a higher rumen pH level and a higher proportion of acetate in the rumen acids than similar but unfertilized grass (Thomson, 1964; A. M. Bryant and Ulyatt, 1965; Grimes, 1967). R. L. Reid and Jung (1965) found no effect on rumen acid patterns resulting from nitrogen fertilization of grass; in that experiment there was also no effect on the soluble carbohydrate content of the grass. In these and other experiments no clear relationship has been shown between rumen acid pattern and level of forage digestibility. Thus Italian ryegrass of high digestibility and high sugar content has given 62 percent of acetate in the rumen acids; white clover of the same digestibility has given 67 percent (Terry and Thomson, unpublished). These results with white clover at the Grassland Research Institute differ, however, from those in New Zealand, where higher levels of sugar content and

48

W. F. RAYMOND

lower proportions of lumen acetate have been found with white clover (R. W. Bailey, 1964). The recent emphasis on lumen acids has also led to investigations of in vitro techniques of estimating rumen acid production, despite the indication by A. C. I. Warner (1964) of the assumptions involved in extrapolating from in vitro to in vivo results. The more recent work reported here reinforces this qualification. Thus Raymond and Terry ( 1 966) and Griffiths (1967) have reported that the proportions of acids produced in an in vitro system can be changed by altering the pH of the system: a reduction in pH produces, as might be predicted, a shift from acetate to propionate or butyrate production. This means that the acid production measured in an in vitro system may reflect the pH at which the system is buffered, rather than the nature of the feed being tested. As rumen pH in vivo can vary between 5.0 and 7.0, in vitro systems are most unlikely to give estimates of rumen acid patterns free from operator bias. It might be suggested that the in vitro system should be buffered at the pH measured in vivo when the test feed is fed-but the rumen acids could then be measured in vivo. However, the different activity in vitro than in vivo (A. C . I. Warner, 1964) must indicate the need for further research before in vitro systems can be considered for the assessment of rumen acid patterns. Increasing attention is also being given to techniques for measuring the rates of production of the volatile fatty acids during microbial digestion of feeds within the rumen in vivo. One technique is based on measurement of the rate of isotope dilution of the different acids labeled with I4C (Bergman et al., 1965; C . L. Davis, 1967; Weller et al., 1967), and this has also been used to study the absorption of the acids in different parts of the digestive tract (Weston and Hogan, 1968). In an alternative technique, samples of lumen contents are removed and incubated in vitro, and the rate of acid production is measured over periods of less than 40 minutes to maintain physiological relevance. The acid production is either extrapolated back to zero time (A. C . I. Warner, 1964) or used as a direct estimate of rate of acid production in vivo (Faichney, 1968a). These techniques, applied to forages, must increase our basic understanding of the nutritive value of these feeds; Leng et al. ( 1 968) have recently reported studies with grazing sheep, for which they used an isotope dilution technique.

c. THEUTILIZATION

FORAGES I . The Digestion of Protein within the Rumen OF THE C R U D E PROTEIN IN

Many experiments have shown that the digestibility (Eq. 2) of the

THE NUTRITIVE VALUE OF FORAGE CROPS

49

crude protein (percent N X 6.25) in forages is very closely related to the crude protein content (e.g., Minson and Kemp, 1961; O’Shea and Maguire, 1967). As with other feed components, Eq. (2) measures the “apparent” digestibility of the feed nitrogen: the “true” digestibility is very close to 100% (Van Soest, 1967), and the decrease in “apparent” digestibility results from the relatively constant excretion of endogenous fecal nitrogen per unit of feed dry matter eaten. The utilization of this “digestible protein” fraction has been the subject of much study; this has established that the rumen microbial population digests and modifies almost all the nitrogenous constituents eaten in the ruminant diet, so that the nature of these nitrogenous constituents may bear little relation to the form in which the nitrogen is absorbed from the digestive tract. The first stage in this reorganization of the feed protein is deamination, the ammonia produced forming one of the main nitrogen substrates for the rumen microbes which proliferate within the rumen medium, to bring about the digestion of the feed polysaccharides, characteristic of ruminant digestion. Entrained organisms then pass with the undigested feed residues from the rumen into the hind tract, where they are subjected to intense proteolytic action in the acid conditions in the abomasum. Here a high proportion of the microbial protein is broken down into amino acids, which are then absorbed and used as the main source of amino acids by the ruminant animal (Hungate, 1966). It is for this reason that the biological value of feed protein, in the terms appropriate to the nonruminant, has little relevance for the ruminant. As I. W. McDonald ( I 968) has emphasized, the biological value will be determined by the amino acid composition of the microbial protein digested in the hind tract, which need bear little relation to the amino acid composition of the feed. The main consideration then is with the factors that determine how efficiently the feed protein (nitrogen) is converted into microbial protein. Immediately after ingestion of food by the ruminant, there is a rise in the ammonia level within the rumen as a result of microbial degradation of nitrogenous compounds. However, the rate at which this ammonia is released may be greater than the rate at which the total rumen population can utilize it, and excess ammonia may then be absorbed through the rumen epithelium into the blood, there to be detoxified and excreted in the urine (I. W. McDonald, 1968). The efficiency of feed protein utilization thus depends on the relative rates of the two processes of ammonia release within the rumen and ammonia assimilation into bacterial protein, and optimum feeding should ensure that the former is not greatly in excess of the latter, so that ammonia absorption from the rumen is mini-

50

W. F. RAYMOND

mized. This is not to suggest that such absorption is necessarily wasteful; forage protein levels are frequently well in excess of animal requirements, and excretion of nitrogen via the rumen wall and urine may then be merely an alternative to excretion after the catabolism of absorbed amino acids. But as level of protein intake decreases toward theoretical requirements, loss of ammonia through the rumen wall must begin to deprive the animal of a supply of amino acids (from microbial protein) which it could effectively use. Factors which determine the level of free ammonia within the rumen include the solubility of the feed protein, rate of deamination decreasing with decreasing solubility of the protein (Annison et al., 1954), and the pattern of feeding, frequent small feeds tending to give lower rumen ammonia levels than few large feeds (Satter and Baumgardt, 1962). The rate at which the ammonia produced is assimilated by the rumen microflora depends mainly on the proportion of readily available energy in the ration, starch and soluble carbohydrates encouraging rapid microbial proliferation and so reducing rumen ammonia levels. Active carbohydrate fermentation also leads to a lower rumen pH, at which absorption of ammonia through the rumen wall tends to be reduced (I. W. McDonald, 1968). These considerations are particularly significant with forages. The nitrogen constituents of fresh forages are mainly in solution or suspended in the cell sap. These are released during mastication of the feed, and are rapidly deaminated in the rumen (Hogan, 1965); yet forages are often low in starch or soluble carbohydrate content, so that rate of microbial growth may not keep pace with rate of ammonia production, and high levels of rumen ammonia may result, with consequent high blood urea levels (Abou Akkada and El Sayed Osman, 1967). This would indicate advantage in more frequent feeding of forage as total nitrogen intake decreases toward animal requirements, but this does not appear to have been tested. At lower levels of nitrogen intake, rate of ammonia production in the rumen begins to limit rate of microbial growth, and can lead to the reductions in herbage digestibility and voluntary intake noted in previous sections. This is one of the situations in which ration supplementation with nonprotein nitrogen may be advantageous (Section VI, D, 2).

2 . .The m e c t of Processing of Forage on Nitrogen Utilization (See Also Section ZX) Many experiments have shown that heat-dehydration of forage can reduce the apparent digestibility of the crude protein in the forage

THE NUTRITIVE VALUE O F FORAGE CROPS

51

(Watson and Nash, 1960, p. 1101, and it has become accepted that this is undesirable. However, Annison et al. ( 1954) had shown that increased nitrogen retention by sheep could result when the solubility of the protein in casein or fish meal, fed in a mixed ration, was reduced by heating; this was despite a decrease in the digestibility of the protein and was mediated through a decreased rate of release of ammonia within the rumen. This led Synge (unpublished) to suggest that a similar advantage might result from a controlled heat-denaturing of the protein in grass. The recent results of Ekern et al. ( 1 965) showing similar levels of nitrogen retention by sheep fed on fresh or dehydrated forage, despite a lower digestibility of the protein in the dried forage, are consistent with this suggestion. This result, if confirmed, must indicate the need to establish optimum conditions for high-temperature dehydration, by which protein solubility is depressed without an excessive depression in protein digestibility. The reverse situation may exist with silage, in which the nitrogenous constituents may be in an even simpler and more soluble form than in the original herbage (Synge, 1952), so that a rapid increase in rumen ammonia level may follow the ingestion of silage. Little work has been reported on this, although McCarrick (1966) has reported that a higher proportion of the body energy gains of silage-fed animals than of hay-fed animals was in the form of body fat, which might indicate a deficiency of available amino acids from the silage diet. D. THEUSE

OF

NONPROTEIN NITROGENIN RUMINANT DIETS

1 . Sources of Nitrogen for Rumen Digestion Under extensive or range-grazing conditions, much of the herbage on offer contains levels of crude protein (N x 6.25) that are inadequate to support economic levels of animal production, and that in some cases may be inadequate for the needs of the rumen microbial population and its animal host even at the maintenance level of feeding. In some areas it is possible to increase the nitrogen level of the diet by sowing legumes within the grazing area or by fertilizing with nitrogen; an alternative is to supplement the animal’s diet either with a protein concentrate o r with nonprotein nitrogen. The possibility of using nonprotein nitrogen for this purpose, rather than preformed protein, arises from the mechanism of protein metabolism within the rumen, in which feed proteins are deaminated and the ammonia produced is assimilated into microbial protein. Simple chemical supplements, such as urea, which are degraded to ammonia within the rumen, might thus be expected to be used almost as efficiently by the

52

W. F. RAYMOND

ruminant as the normal feed proteins. Some success has been achieved in this direction, but despite a considerable research effort during the last thirty years some aspects of the use of nonprotein nitrogen supplements remain obscure (see reviews by M. H. Briggs, 1967: Chalupa, 1968). A more objective approach to the subject becomes increasingly possible, and recent information on nitrogen metabolism within the rumen indicates two distinct areas of interest, with forages, respectively, of very low and of medium protein contents.

2 . Low Protein Forages As discussed in Sections IV, C, 3 and V, E, the digestibility and the level of voluntary intake of forage both may be very low because the forage does not contain enough nitrogen compounds to sustain active bacterial digestion in the rumen. With forages of less than 4 percent crude protein, supplementation with urea (Campling et al., 1962) or other nonprotein nitrogen sources (e.g., biuret; Schaadt et al., 1966) has increased both level of digestibility and voluntary intake. At higher crude protein levels little increase in digestibility has been found: thus no response was found when urea was added to barley straw of 5 to 7 percent crude protein (Kay et af., 1968) or to oat straw of 4.4percent crude protein (Faichney, 1968b). The practical result is that urea supplementation will improve lowprotein pasture from a nutritive value at which animals lose weight to one where they maintain weight; but it will seldom lead to body weight gain, for even with urea, the level of digestibility and intake of such feeds is still too low to sustain production. This may be accentuated by deficiencies of other components in the diet, including minerals; thus it is possible that sulfur deficiency in the diet may limit microbial action even when urea is added, although I. W. McDonald (1 968) has questioned whether inorganic sulfur can be used by rumen organisms. 3 . Medium-Protein Forages

With many types of low-protein forage the response of animals to urea supplementation is limited by the low inherent digestibility and intake of these forages. But many other forages, despite a low protein content, may be of high digestibility and intake characteristics (e.g., forage maize). In other cases the overall digestibility and intake of a forage-based ration may be increased by supplementation with low-protein cereal grains or starch. Available energy need then no longer be limiting in the diet, and with informed use urea can contribute significantly as a protein supplement to such diets.

T H E NUTRITIVE V A L U E OF FORAGE CROPS

53

The main limitation to the use of urea then appears to lie in the very rapid rate at which it is deaminated in the rumen; unless urea is fed in phase with an appropriate energy source rumen ammonia and blood ammonia levels may rise so rapidly that the animal suffers, and may die, from ammonia toxicity. This problem can be most acute with feeds of low digestibility; with these rumen p H tends to be high, and ammonia is more rapidly absorbed into the blood than at the lower p H associated with the rapid fermentation of most feeds of high digestibility; but urea toxicity can still be a hazard with high-digestibility feeds, and feeding regimes must aim to keep rumen ammonia levels at a low level. Recent research indicates a number of advances in nonprotein nitrogen feeding. Thus J. R. Campbell et al. ( 1963) found lower rumen ammonia levels and more efficient nitrogen utilization when a grain ration with 3.3 percent urea was fed six times rather than twice daily. It has been suggested that urea toxicity may be avoided by reducing urease activity within the rumen by feeding barbituric acid (Clifford and Tillman, 1966) or by the use of antibiotics (Chalupa, 1968). Other workers have suggested the replacement of urea by biuret, which is less rapidly deaminated within the rumen, but results with this supplement have been conflicting: thus Schaadt et al. ( 1 966) found similar forage cellulose digestibility with urea or biuret, but Oltjen ef al. ( 1 968) found lower fiber digestibility with biuret than with urea even after an adaptation period of 32 days. Perez et al. ( 1967) have reported that a urea-phosphate combination is less toxic than urea alone, and can supply up to 40 percent of the nitrogen in the total ration. Of particular interest is the demonstration by Virtanen ( 1966) and Oltjen and Bond (1967) that the ruminant can adapt to gradually increasing levels of urea feeding, which makes possible a synthetic diet for dairy cows in which all the nitrogen is derived from urea. It has also recently been shown that urea is particularly efficiently used in ruminant rations when it is incorporated in a pellet with dehydrated lucerne (Karr et al., 1965). This has led to the development of feed supplements (e.g., Conrad and Hibbs, 1966), in which lucerne and urea are combined to give high crude protein CN X 6.25) levels; these have been used to replace soybean meal completely in dairy rations, without the toxic symptoms that might have been expected at the high levels of urea fed. The exact mechanism is uncertain; it has been attributed to “unidentified growth factors,” which could include minerals (Karr et al., 1965), or branch-chain fatty acids, which have been found to increase efficiency of utilization of urea nitrogen (Cline et al., 1966). And in relation to the possible effect of heat dehydration on the rate and site of digestion of the protein fraction in forages (Section IX, B), it is interesting to speculate that urea and dehydrated lucerne may act as

54

W. F. RAYMOND

complementary sources of nitrogen, the urea being deaminated rapidly and the lucerne slowly, so giving a steady provision of ammonia for microbial growth.

E. THEMINERALNUTRIENTS IN FORAGES The requirements of different classes of ruminant livestock for mineral nutrients, and the ability of forages to provide these requirements, cannot be fully covered in the present review; but a number of recent and significant developments are noted. Requirements are best stated as the total intake of a particular mineral needed by each class of livestock at different levels of production (Agricultural Research Council, 1965). Both the quantity of forage eaten (Section V) and the concentrations of minerals in this forage are important, but the relation of these to total requirements depends also on the availability to the animal’s metabolism of these minerals, and of interactions between them (Underwood, 1962, 1966). The first concern of the agronomist must be to establish the factors that determine the total intake of different mineral elements (forage dry matter X percent mineral content in the dry matter). Only when total intake is above the theoretical requirement do availability and interactions become of major concern. 1 . The Concentration of Mineral Elements in Forages Very many analyses of the mineral composition of forages, reported in the literature, have been summarized by Whitehead ( 1966). Such a summary is of value in indicating broad differences in composition between forage species: but it also indicates a wide range of contents of particular minerals within each forage species, which cannot always be related logically to stage of physiological maturity, in the same way as, for instance digestibility, protein, and fiber contents. Thus the copper content of perennial ryegrass can vary from 2 to 12 p.p.m., depending on soil copper status, soil pH, and fertilizer application. This means that an average figure for copper content may be of limited usefulness in indicating the adequacy of a particular crop of ryegrass as a source of copper, and emphasizes the importance of understanding the effects of these other factors. The mineral contents of temperate forage species grown under the same conditions have been reported by many workers, including Davies et al. ( 1 966) and Fleming and Murphy (1 968), of farm grassland heavily fertilized with nitrogen by de Groot and Keuning ( 1 967), and of tropical forages by Dougall and Bogdan ( 1966). From these, and from the summary already noted, a number of trends emerge. As a given forage ma-

T H E NUTRITIVE VALUE OF FORAGE CROPS

55

tures, the percentage content of some elements, in particular, P, S, and Cu, tends to decrease. Among species the legumes have higher contents of some minerals than the grasses, particularly of Ca; within the grasses there are some notable differences, as in the consistently lower content of sodium in timothy (Phleum pratense) and meadow fescue (Festuca pratense) than in most other grasses (ap Griffith and Walters, 1966; de Loose and Baert, 1966). Mineral content varies markedly between different parts of the same plant, leaves being higher in Ca, Mg, S, Fe, and Si than stems and older tissues (Davey and Mitchell, 1968). When forages are being grazed this can be of importance because of the selective nature of grazing (Section X). Within these overall patterns, certain features of particular elements have emerged. 2 . Phosphorus 1. W. McDonald (1968) has pointed out that while phosphorus deficiency has been clearly established with cattle feeding on forages of low P content, this has not been shown with sheep. He suggests that this may be because, although the bodies of cattle and sheep contain very similar concentrations of phosphorus, the feed intake of sheep is 1.5 to 2 times that of cattle, per unit of body weight, so that sheep may require only about half the concentration of P in their feed compared with cattle. Under grazing conditions sheep may also be able to select plant fractions of higher P content-a difference of general significance, because of the higher content of most elements in plant leaves, which sheep are better able to select than are cattle.

3 . Sodium, Calcium, and Potassium The particular importance of potassium is that it can be absorbed from the soil in luxury amounts, and this can lead to marked reductions in the contents of the elements Na and Mg in forage plants. Ap Griffith and Walters ( 1966) and de Loose and Baert ( 1 966) have shown large differences in Na content between different species: thus ryegrass commonly contains several times as much Na as timothy, and Poa pratensis and Festuca pratense also have contents well below the concentration of 0. I 3 percent Na suggested as necessary for the high-yielding dairy cow (Kemp, 1964). Yet I. W. McDonald (1968) considers that there is no unequivocal evidence for sodium deficiency in ruminants that is not preceded by other deficiencies. Considerable differences in sodium content have been found between different timothy genotypes (de Loose and Baert, 1966). But there can be no case for breeding timothy varieties of higher

56

W. F. RAYMOND

sodium content; if sodium deficiency does occur it should be remedied by making a salt supplement available, rather than by introducing a marginal objective into a breeding program. With regard to calcium, Langlands et al. (1967) and I. W. McDonald ( 1 968) have suggested that deficiency in intake is seldom likely to occur on forage feeding; where it does occur other deficiencies, e.g., in crude protein or phosphorus intake, are likely to be primary. 4 . Magnesium

This element remains of particular interest because of the widespread occurrence of hypomagnesemic tetany in forage-fed animals in many parts of the world (reviewed by Allcroft and Burns, 1969; Grunes and Stout, 1969). Two forms of the disease have been distinguished. A chronic form is associated with a slight but prolonged deficiency of Mg in the diet, leading to low plasma-Mg levels; however, symptoms seldom occur until the animal is subjected to stress, and magnesium therapy may not then be effective. The acute form is most common when stock have access to rapidly grown fresh forage. This may stimulate higher production, e.g., of milk, but may not provide the Mg needed for this higher production: although body reserves of Mg may be high, this is mainly in bone, and being only slowly available cannot satisfy the suddenly increased demand, so that plasma-Mg falls to a critical level. The resulting tetany thus cannot be obviated by building up body reserves by feeding Mg supplements before the forage is fed, and supplementary feeding must continue throughout the period when animals are a t risk. The currently useful supply of magnesium depends on two main factors, viz. the daily intake of magnesium, and its availability to the animal. Many workers have reported Mg levels in forages (e.g., Davies et al., 1966; Hedin and Duval, 1966; de Groot and Keuning, 1967; Fleming and Murphy, 1968). Within the grasses no consistent differences appear: there is a trend to lower Mg levels in the spring, and a rise later in the season. Legumes may contain more Mg than grasses, and inclusion of clover in a ryegrass sward may improve the magnesium status of grazing stock, particularly in the spring. Nitrogen fertilizer has little effect on Mg levels, but they can be markedly reduced by potassium fertilization. Tetany is most commonly found when the Mg content of forages is below 0.18 percent of the dry matter, but Kemp ( 1 960) has suggested that it is the ratio [K] to ([Ca] [Mg]), expressed in milliequivalents per kilogram, that is important, the risk of tetany increasing when this ratio approaches 2.2. However, the low correlation in practice between such measures of total Mg intake and the incidence of hypomagnesemic tetany

+

THE NUTRITIVE VALUE OF FORAGE CROPS

57

indicates that availability, rather than intake per se, must often be the determining factor. In general the digestibility of magnesium in forages is about 20 percent; any reduction in this already-low digestibility must have a major effect on the quantity available to the animal’s metabolism. Thus Wind et al. ( 1966) have found high levels of higher fatty acids and lipids in young grass, and have suggested that these may form insoluble soaps with both magnesium and calcium. O’Sullivan ( 1968) has suggested that high levels of histamine may precipitate tetany; the histamine content of forage rises during dry cold weather, when tetany is most prevalent. Recently Stout et al. (1967) have shown high levels of trunsaconitic acid in some species growing on rangeland areas where tetany in cattle is prevalent. Grasses contained higher levels than other species, with a maximum content of 6.2 percent being recorded. A possible mode of action indicated is by inhibition of the enzyme aconitase in the tricarboxylic acid cycle, but the relation with serum-Mg level is not clear. However, Kennedy (1968) has reported that a supplement of trunsaconitate, fed to sheep at 7 percent of the feed dry matter intake, had little effect on blood levels of citrate or aconitate, on plasma levels of C a or Mg, or on urinary excretion of Mg. H e suggested that this was to be expected, because the rapid rate of metabolism of this anion within the rumen would preclude significant absorption into the blood. In the author’s view, hypomagnesemic tetany should now be considered a problem mainly of extensive grazing. Under intensive grazing or indoor feeding, the magnesium content of forage can be increased by fertilization with magnesium salts, or supplementary magnesium can be fed. With neither of these is there any evidence of damage resulting from an excessive intake of magnesium, and there can now surely be no excuse for tetany among intensively managed livestock. And with animals under extensive management a practical solution may emerge with the development of a slow-release magnesium bullet (University of Glasgow, 1965). 5 . Cobalt

The occurrence of areas of cobalt-deficient soils, the induced vitamin BI2 deficiency in ruminants eating forage grown on these areas, and methods of treatment by the use of cobalt fertilizers or by dosing livestock with cobalt bullets, have been fully documented (Underwood, 1966). E. D. Andrews (1966) has shown that the cobalt content of legumes is higher than of grasses, but on soils where lambs respond to cobalt dosing the cobalt content of all the forage species present is generally below 0. I p.p.m.

58

W. F. RAYMOND

6 . Copper and Molybdenum These two elements interact in ruminant feeding, a high level of molybdenum leading to effective copper deficiency even on pasture of high copper content, although this can be prevented by feed supplements or injections of copper (Dick, 1954; Underwood, 1966; I. W. McDonald, 1968). However, the more common problem appears to arise either from deficiency or low availability of copper in forage. Minimum copper levels for sheep have been reported to vary between 1 and 8 p.p.m. (Dick, 1954), the highest level being associated with high forage molybdenum content. However, Macpherson and Hemingway ( 1968) found swayback in lambs grazed on limed pasture with a copper content of 8.9 p.p.m., a level only slightly lower than on the control pasture, 9.6 p.p.m., on which no disease occurred. This was not due to high forage molybdenum content, and indicated a reduced absorption or retention of copper, reflected by a marked reduction in blood and liver copper levels. A daily supplement of 10 mg. of copper administered in several forms gave an adequate blood copper level of 1 p.p.m., without risk of toxicity. Alternatively herbage copper levels can be increased by fertilization, a single dressing of 10-20 pounds copper sulfate per acre giving a response for up to eight years (Reith, 1968). Excessive copper intake is also possible, and this could become more common if indiscriminate mineral supplementation follows awareness of a possible deficiency. The immediate response of the animal is to store excess copper in the liver, but under stress conditions this can become metabolized and cause toxicity, as hemolytic jaundice (I. W. McDonald, 1968); certain forages also contain hepatoxins which damage liver cells, and can lead to accumulation of copper (Bull and Dick, 1959). As with other mineral elements, analysis of the forage available may give an erroneous measure of the copper content of the forage eaten under grazing conditions, because of the considerable difference in content between forage species and between different parts of the same plant (Gladstones, 1962). Thus Van der Merwe and Perold (1967) reported that no symptoms of copper deficiency were found with cattle grazing grass of very low copper content; this was because the cattle were browsing the leaves of a shrub which contained 4 to 15 times as much copper as the grass. These authors pointed out the dangers that could arise from subdivision of extensive grazing areas if this reduces the opportunity for feed selection. 7 . Lead Mitchell and Reith (1 966) have discussed the possible risks to stock health of the considerable rise in lead content (up to 40 p.p.m.) which

THE NUTRITIVE VALUE OF FORAGE CROPS

59

can occur in forage in autumn and winter, compared with the normal level of about 1 p.p.m. in actively growing grasses and clovers.

8. Iodine Alderman and Jones ( 1 967) have reported on the changes in iodine content of six grass species at different stages of maturity. The average level was lower in ryegrass and fescues (0.22 p.p.m.) than in cocksfoot and timothy (0.32 p.p.m.); iodine levels fell with increasing maturity of the crop, and returned to early spring levels only in the autumn; content in all species was reduced by nitrogen fertilization. In very few instances did the iodine content reach the level recommended for pregnant or lactating animals. Butler and Glenday ( 1 962) showed that the iodine content of a forage species is strongly inherited, and this could make selection for higher iodine content possible. The practical solution, however, seems to make available an iodized salt lick, with which the danger of toxicity is remote. 9. Sulfur

The animal’s requirement for sulfur is mainly in the form of the sulfur amino acids cystine and methionine; thus the primary impact of any deficiency of S will be on the rumen microorganisms, which make these amino acids available to the host animal by synthesis from a wide range of nitrogenous compounds in the feed. Moir and Bray ( 1 967) showed that rumen bacteria have a relatively uniform content of sulfur amino acids, with a S:N ratio of 1:5, which is similar to that in the true protein in leaves. Barrow and Lambourne ( 1 962) found a ratio close to 1 :13 (based on total forage N content) in forages ranging from 0.05 to 0.58 percent sulfur in the dry matter. They emphasized, however, that there might be an effective deficiency of sulfur in feeds of low sulfur content, because the ruminant is able to recycle nitrogen as urea via the saliva, but this barely occurs with sulfur. This could account for the poor use of urea when fed with forages of low sulfur content, as noted by I. W. McDonald (1968). Sulfur deficiency in ruminants appears rare, but must be kept in mind with the increasing trend toward the use of chemical fertilizers that contain no sulfur. 10. Selenium

Selenium in forage has assumed importance because of the association in several countries of low forage selenium levels with white-muscle disease in young stock (a detailed review by Rosenfeld and Beath, 1964). Hamilton and Beath (1964) suggest that a level of 0.1 to 0.3 p.p.m. Se should give protection. Only small differences have been found in the Se

60

W. F. RAYMOND

contents of different forages growing on deficient soils (Hamilton and Beath, 1964), but on soils of high Se content, or with Se fertilization, 10-fold differences between species have been found (Ehlig et al., 1968), lucerne having a relatively high content compared with the grasses. Selenium content tends to be lower in rapidly growing grass, and to be reduced by nitrogen fertilization or irrigation, under which a marginal situation may become critically deficient (Hamilton and Beath, 1964). Correction of selenium deficiency can be by pasture topdressing (Allaway and Hodgson, 1964) or by drenching the stock, but Paulson et al. ( 1 968) suggested that use of a selenium-fortified salt lick may be the safest method. However, care is needed in all cases to avoid building up toxic levels of selenium in the animal carcass, and the possibility must be considered of a buildup in pasture selenium content as a result of absorption from the excreta of grazing stock (Peterson and Spedding, 1963). I. W. McDonald ( 1 968) has pointed out that white-muscle disease may not be directly attributable to selenium deficiency, as no specific role has been defined for this element in ruminant metabolism; thus it might operate by counteracting some other toxic component in the feed. There is clearly an association with the vitamin E status of the young animal, as dosing with this vitamin can prevent the disease; but administration of the vitamin to the ewe will not prevent the disease in the lamb, indicating that selenium, but not vitamin E, is transferred from the ewe to the lamb (Paulson et al., 1968). However, whatever the mechanism of the disease it is clearly associated with selenium. Feeds on which young stock have developed white-muscle disease have always contained less than 0.1 p.p.m. of selenium, although not all feeds with this low level cause the disease (Allaway and Hodgson, 1964). 11. Silica

Sheep and cattle in some areas in North America and Australia may suffer severely from the deposition of siliceous uroliths in the urinary tract. The etiology of urolithiasis in sheep and cattle has been recently reviewed by L. H. P. Jones and Handreck (1967). The calculi in both sheep and sattle have been shown to consist largely of amorphous silica. Despite a better understanding of the formation of these calculi, effective prophylactic measures under rangeland conditions are still needed. Baker et al. ( 196 1) have suggested that the solid silica of plants may be an agent of wear in sheeps’ teeth, because it is harder than the dental tissues. In addition soil quartz may be ingested in such quantities as to be equally important as a cause of wear in teeth (Healy and Ludwig, 1965).

T H E NUTRITIVE VALUE OF FORAGE CROPS

61

F. PHARMACOLOGICALLY ACTIVECOMPONENTS I N FORAGE The practical use of a forage species, otherwise nutritionally adequate in terms of digestibility, protein, and minerals may be restricted because it contains also a pharmacologically active component. The occurrence of such toxic forage plants is widespread (reviewed by E. G. C. Clarke and Clarke, 1967), and the surprising feature is perhaps that toxic symptoms in ruminant livestock are not more common. I. W. McDonald ( 1968) has suggested three possible reasons for this; many toxic plants are distasteful to livestock; animals reared in a given environment seem to develop a capacity to avoid toxic plants; and where toxic plants comprise only part of a mixed community, the quantity eaten may represent only a minor fraction of the total forage intake. He noted, however, that main emphasis has usually been placed on obvious symptoms of toxicity, and less on subclinical effects that may still reduce productivity or reproductivity efficiency. Most serious is the limitation placed on the use that can be made of agronomically and nutritionally desirable forage species because they may at times contain toxic components. I . Estrogenic Compounds Many forage species, in particular the legumes, contain compounds that exhibit estrogenic activity in ruminants. This was first observed in Australia, where the breeding efficiency of ewes grazing subterranean clover ( T . subterraneum) was found to be impaired. This had most serious implications because of the great potential this species offered for improving forage production in the subcontinent, and a major research effort has been directed to the problem (reviewed by Moule er al., 1963; Bickoff, 1968). Biochemical studies isolated a number of isoflavone compounds, including formononetin, genistein, and biochanin A, which have estrogenic activity when injected into mice, of which the change in uterine weight is used as a bioassay. However, potency, assessed by this method, sometimes showed little correlation with observed responses in sheep, and Moule et al. ( 1963) concluded that it was advisable to use relevant animals. In many cases bioassay has been replaced by chemical assay, but these methods also do not always appear to correlate closely with animal response. Although Beck ( I 964) considered that chemical assay could give a reasonable estimate of estrogenic activity, Lindner ( 1 967) found that it could not be closely correlated with reproductive disturbance in ruminants. This is partly because of the different activity of the different isoflavones, and as Bickoff (1968) has emphasized, interactions are also possible, and the presence of estrogen inhibitors must also be con-

62

W. F. RAYMOND

sidered. However, with these limitations, chemical methods have been of value, particularly in screening large numbers of forage samples, which would not be practicable in vivo. Thus Beck (1964) and Francis and Millington (1 965) showed that the leaves of subterranean clover contain much more isotlavones than the petioles -a feature of practical importance under conditions where sheep may graze a higher proportion of leaves than petioles. Rossiter and Beck (1966) showed isoflavone content to increase with increasing environmental temperature, and under conditions of soil phosphate deficiency. Morley and Francis (1968) reported the isoflavone contents of 151 lines of subterranean clover. Differences between the same line grown at different centers were small. Differences between varieties within a subspecies were large enough to indicate the possibility of breeding for low levels of particular isoflavones, although evidence of a negative correlation between the contents of genistein and biochanin A might present problems in selection. Reduced breeding efficiency in sheep has also been found with red clover ( T . prutense; Barrett et al., 1965) and with lucerne (Coop and Clark, 1960). Assays for isoflavone content were made on both legumes, but it was subsequently found that lucerne also contains a compound, coumestrol, which is 15 to 100 times more potent than the most active isoflavone, formononetin (see Bickoff, 1968). Hanson ( 1965) showed considerable differences in coumestrol content among lucerne varieties, and Loper (1968) reported that fungal and aphid attack could considerably increase the coumestrol content of lucerne; aphid-resistant varieties, such as MOAPA, thus had a lower content than a susceptible variety, VERNAL. Newton and Betts ( 1 968) have recently shown considerable estrogenic activity in legumes grown in the United Kingdom, using as bioassay techniques both the effect on reproductive efficiency in ewes and the increase in teat length of wether sheep. High levels of activity were found in S.123 red clover, lower levels in lucerne and white clover, and no activity in sainfoin. While most experimental evidence has been that the effects from pasture estrogens are undesirable, at one stage it was considered that the stimulation of milk production in ruminants grazing spring pasture might be due to estrogens (Pope, 1954). However no correlation or causation could in fact be established (Pope et al., 1959). More recently it has been suggested that pasture estrogens may stimulate growth and feed conversion efficiency in growing and fattening stock, by mechanisms similar to those from dosing or implantation with stilbestrol or hexestrol. Thus Oldfield et al. ( 1966) considered that the coumestrol content of lucerne may benefit the growth and carcass quality of lambs. However, in gen-

THE NUTRITIVE VALUE OF FORAGE CROPS

63

era1 it is likely that estrogens in pasture are undesirable, and Bickoff ( 1 968) has concluded that even low levels may be unacceptable because their effects, although unspectacular, may impair overall economic efficiency. 2 . Phalaris Alkaloids A chronic condition, phalaris staggers, may affect sheep grazing P . tuberosa. This can sometimes be prevented by regular cobalt dosing, but R. M. Moore and Hutchings ( 1 967) showed that this was not associated with vitamin BI2 status, as is the case with straight cobalt deficiency. An acute syndrome (Gallagher et al., 19661, however, is not cured by cobalt therapy. Biochemical fractionation of extracts from toxic forage have isolated a number of alkaloids, derivatives of dimethyl tryptamine, which have produced symptoms of acute phalaris staggers when administered to sheep (Gallagher et al., 1966). Studies using this chemical assay have shown that factors leading to nitrate accumulation in Phalaris also produce high alkaloid levels (R. M. Moore and Hutchings, 1967): of particular importance, 20-fold differences in alkaloid content have been found between different lines of Phalaris. Commercial Australian varieties were consistently high, whereas lines from Algeria were all of low alkaloid content, and this indicated the possibility of breeding nontoxic varieties (Oram and Williams, 1967).

3 . Alkaloids in Tall Fescue Cattle grazing tall fescue suffer from a reduction in skin temperature and from gangrene which is sometimes fatal. Yates and Tookey (1965) found, in extracts from tall fescue, an alkaloid, festucine, which is possibly a toxic principle. 4 . Bloat-Producing Compounds

A most serious economic disease of grazing is bloat, in which gas produced by microbial fermentation of feed is trapped within the rumen, and the animal may die from the resulting pressure on the heart. Earlier theories were that this was due to paralysis of the muscles involved in eructation, but it is now accepted that the main cause is the formation of a stable foam within the rumen which traps the fermentation gases ( C . S. W. Reid, 1960). The main foaming agent appears to be plant cytoplasmic protein (Laby and Weenink, 1966), and a specific protein fraction, 18 S protein, has been isolated (McArthur and Miltimore, 1966). Bloat-producing forages, in particular lucerne, were found to contain over 4.5 percent of this protein compared with contents of less than 1

64

W. F. RAYMOND

percent in nontoxic forages. It has become clear that a number of factors can reduce the stability of the foam produced within the rumen, so that not all forages with high levels of 18 S protein will produce bloat. Thus the foam is most stable at a pH of 5.5 to 6.0 (characteristic of the diseased rumen condition in vivo); restriction of the rate at which forage is eaten so as to reduce the extent to which rumen pH falls below the resting level, ca. 6.8, is indicated. E. M. Hutton and Coote (1966) and Kendall (1 966) have suggested that a high tannin content may prevent bloating with potentially toxic forages by forming a nonfoaming complex with the active protein. Thus crownvetch (Cornilla varia L.), containing 10 percent tannin, and birdsfoot trefoil (Lotus corniculatus L.1 do not cause bloat (Kendall, 1966); of the legumes introduced into Australia Desmodium may contain over 9 percent of tannin (E. M. Hutton and Coote, 1966). The most effective treatment is to dose the animal or to spray the pasture with antifoaming agents, such as peanut oil (Johns, 1959). R. T. J . Clarke ( 1 965) has suggested that rumen protozoa may produce foam-stabilizing secretions and that antiprotozoal drugs may reduce bloat. McArthur and Miltimore ( 1 966) consider that the best solution may be to prevent the risk of bloat by agronomic measures; by growing potentially toxic legumes in combination with nontoxic varieties; and by studying the effect of environment, so as to manipulate environmental factors to minimize the 18 S content of legumes.

5 . The Nitrate Content of Forages M. J. Wright and Davison (1964) have reviewed in detail the factors determining toxicity arising from high nitrate content in forages: this anion is reduced to nitrite within the rumen, and when absorbed this can lead to the toxic condition of methemoglobinemia in which ferrous iron is oxidized to ferric, so that the blood cannot carry oxygen to the tissues. However, no specific level of nitrate intake can be stated as toxic; a drench of nitrate, or very rapid forage intake, can lead to toxicity at a level of nitrate intake that would be safe under normal feeding and nitrate is less toxic if the ration also contains readily digested feeds, such as carbohydrates, which can reduce the nitrite before it is absorbed. This suggests that nitrate poisoning may be particularly associated with a high content of nitrate in a forage of low digestibility. If symptoms do appear, and they cannot be reproduced by feeding nitrate, then some other causal agent must be sought; but the effect of sublethal doses of nitrate is not well understood (M. J . Wright and Davison, 1964). However, high nitrate contents should clearly be avoided; even if they do not

THE NUTRITIVE VALUE OF FORAGE CROPS

65

always indicate a forage toxic to animals, they do indicate a forage that has made inefficient use of this nutrient, which it has absorbed from the soil but has not metabolized. 6 . Other Problems

Sweet clover (Melilotus alba) contains a coumarin which can cause hemorrhages in sheep. Linton et al. ( 1964) showed considerable protection against this disease when a low-coumarin variety, CUMINO, was compared with a toxic variety, ARCTIC. A serious photosensitization disease, facial eczema, has been suffered by sheep grazing autumn pasture in New Zealand. Detailed research showed that this is due to a toxin produced by a fungus, Pithomyces chartarum, which proliferates on the senescent and dead litter within these swards (reviewed by Brook, 1963). This disease has not been reported elsewhere; however, the possibility of similar, but subclinical, infections in other environments cannot be ignored. A less serious photosensitization, due possibly to an alkaloid, perloline, has also been reported with perennial ryegrass (Bennett, 1963). Recent research has begun to clarify a most unusual toxicity, enteque seco, found in restricted areas in Argentina; cattle suffer calcification of the blood vessels due to high blood levels of calcium and phosphate (Worker and Carrillo, 1967). Observation of the areas of incidence implicated an erect plant, Solanum mufacoxylon.Although this appears to be unpalatable to stock, as the plants senesce the leaves fall on the pasture plants below and are ingested as the animals graze these plants. A different problem is also indicated in the possibility that chemicals in certain forage plants may be absorbed and secreted in milk, to which they give an off-flavor. These seem to be associated with legumeswhite and red clovers (Honkanen and Moisio, 19631, ladino clover (Woods and Aurand, 1963), and lucerne (Dunham et al., 1968)-but no clear relationships have been shown, although dimethyl sulfide from lucerne appears in milk some 3 hours after feeding. Many livestock men are also convinced that pasture composition can affect the flavor of meat. VII. The Relationship between Forage Quality and Forage Yield

Both experimental and practical evidence shows that as a forage crop grows its yield increases but its nutritive value tends to decrease. Thus the practical exploitation of any crop involves a compromise between yield and nutritive value. This requires, in addition to the information on nutritive value reported in the previous sections, data on yields at dif-

66

W. F. RAYMOND

ferent stages of maturity, as given by Minson et al. (1960, 1964) in association with the digestibility data shown in Fig. 1 . The use of in vitro digestibility techniques has made possible the compilation of yielddigestibility relationships for many different forage materials; these are most usefully reported in the form used by Green ( I 966) and Aldrich and Dent (1967) for grasses and legumes, and by Rogers (1967) for arable forage crops (Fig. 5 ) . From these graphs different forage species and

2

diqestible orqanic matter content

\

75

655500

I

I

I

Ib./ocre YIELD

I

! I

10,000-

8,000-

I

1 0 20

Aprd

i

3 0 10 2 0 3 0 10 2 0 May

June

FIG. 5. The changes in yield and digestibility of the first growth of S.24 ryegrass during first growth in the spring. (From Green, 1966.)

varieties can be compared in terms of their yields at a given level of digestibility. Figure 6 shows the yields per acre of different grass varieties during first spring growth when they all contain 65 percent of digestible organic matter (Green, 1966). This figure illustrates clearly the higher yields of ryegrass varieties than of other grass species harvested at the

THE NUTRITIVE VALUE O F FORAGE CROPS

67

same level of digestibility, and the interval of 23 days between the dates at which the digestibility of the earliest (S.24) and the latest ryegrass variety (VERTAS HARDGRAZING) falls to 65 percent digestible organic matter content. The interpretation of such results is complicated, however, by the fact that the harvest of a perennial forage, unlike that of an arable crop, is also a treatment for the following regrowth. A forage harvested a t an Ib./acre YIELD OF DRY MATTER

I0,OOO

1

eS 321 Q RvP

s 22

6,000

1

Q .

QTetronc

OS 152

20 MAY

oPetro

0523

524

0 5.215

Ib

ORcveille

5.53 5.48

30

A

TOII fescue

A

Cockrfmt

4

I?

JUNE

FIG.6. The date at which the content of digestible organic matter in the dry matter falls to 65 percent during first growth of different grass varieties. (From Green, 1966.)

immature, high digestibility stage will regrow more rapidly than if harvest is delayed until a more mature stage; as a result, the total yield, of the first growth plus regrowth, from early or late harvested forage may differ much less than would be indicated from Fig. 5. In terms of digestible organic matter yield the difference may be even less. Very few data are available on the comparative annual yields from different timesequences of harvesting a forage through a complete growing season, yet such data are essential if harvesting or grazing programs are to be planned on an objective basis.

68

W. F. RAYMOND

VIII. Forage Breeding for Improved Nutritive Value

The first objectives in a forage breeding program must concern agronomic characters -yield, persistence, disease resistance, growth habit in relation to planned method of use-and only when these are achieved does it seem justified to select specifically for improved nutritional features. But such further selection can already be considered within the numerous forage materials of high agronomic quality now available. Until recently the main criteria used in selection for nutritive value have been those accepted for many years, namely crude protein content and leaf stem ratio. Crude protein content assumed less importance with the recognition that it was only casually related to the digestibility and energy content of forages, but the usefulness of leafstem ratio was accepted until the more recent studies (Section IV, B) which demonstrated that, in immature plants, stem may in fact be more digestible than leaf. The importance of digestibility as a measure of energy value, and also as an index of voluntary intake characteristics, and the availability of in vitro techniques for estimating forage digestibility, have emphasized digestibility as a promising selection parameter (Rogers and Whitmore, 1966). Cooper et al. ( 1962) found considerable differences in digestibility between individual plants of S.37 cocksfoot, all the plants being at a young leafy stage. Furthermore, the progeny of crosses between the genotypes of highest digestibility showed a mean value higher than the main population, with a heritability of 0.5. This early study, showing the possibility of objective selection for high digestibility in herbage plants, has been confirmed by other workers. Julen and Lager ( 1966) found that the more digestible plants within a cocksfoot population maintained this higher digestibility at different growth stages, and that both the leaves and the stems of some plants were more digestible than in others. The same authors later reported similar differences in digestibility between plants of other grass species, as well as with several legumes and fodder rape, which might form the basis for a breeding program, and Mowat (1 969) reported similar results from Canada. Knight and Yates (1 968), however, found no consistent digestibility ranking for cocksfoot plants grown in a Mediterranean environment in Australia. They concluded that under these conditions responses to temperature, resistance to rust diseases, and the observed 4-fold differences in yield between genotypes were much more important than possible marginal improvements in digestibility. This concept is most important. For if one genotype is of higher yield than another when both are at the same level of digestibility,

THE NUTRITIVE V A L U E O F FORAGE CROPS

69

then the first genotype is likely to be of higher digestibility when both are harvested at the same yield-and improved yield is in other respects a rewarding objective. Significant genetic variation has been found in other nutritionally important characters of forages, in particular of minerals, including sodium (de Loose and Baert, 1966). This illustrates an important aspect of the responsibility of the nutritionist in advising the plant breeder on nutritional objectives, for selection for improved sodium content may well be irrelevant in comparison with direct sodium supplementation of the animals that are to be fed. This means that the nutritionist must consider most carefully before advising the plant breeder to start a selection program. In particular he must recognize that a breeding program takes many years, so that the nutritional objectives he defines must be relevant to the likely feeding systems of the future, rather than to those presently operating (Raymond, 1968). Within the context of the nutritional developments outlined in other sections, including forage processing, feed combinations, and mineral and urea supplementation, the impact of selection for improved digestibility or for higher contents of soluble carbohydrates or minerals could well be marginal. In particular a more effective use of nonprotein nitrogen supplements would relieve the breeder of the necessity to ensure “feeding standard” levels of crude protein in forages. The most important developments may lie in improving the intake characteristics of forages, and in eliminating toxic components. Selection for improved palatability within Phalaris tuberosa (Section V, B, 2 ) and for reduced levels of estrogens witin subterranean clover and of coumarins in sweet clover (Section V, F) are examples of breeding objectives that are unlikely to be made obsolete by other advances in nutritional science. IX. The Effects of Processing on the Components of Forage Nutritive Value

The previous sections have considered some of the basic nutritional features of forage crops. However, many forage crops are cut and processed before they are fed; it is clear that such processing may at times markedly modify the nutritive features found in the fresh crop, and an understanding of these changes is essential to the development of improved processing techniques. A. DEHYDRATION The effects of dehydration on the digestibility and voluntary intake of

70

W. F. RAYMOND

forages has been considered in Sections IV, C and V, B. Both the digestibility and intake of dehydrated forage have been found to be similar to those of the fresh forage, but the apparent digestibility of the nitrogen fraction is reduced if the temperature of the forage is maintained above 105°C. for a significant period during drying (see, e.g., Ekern et al., 1965). Despite this reduction in nitrogen digestibility, body nitrogen retention from dried forages is not reduced (Ekern et al., 1965) unless heat damage is excessive: possible reasons for this were considered in Section VI, C. In practice the main method of forage dehydration is haymaking. Considerable changes in nutritive value can occur in this process, due not to dehydration per se, but to deficiencies in the techniques used, leading to dry-matter losses in the field which may rise to 25 percent or more under unfavorable conditions. Because of differential loss of leaf the loss of crude protein from the crop may be as high as 40 percent and dry-matter digestibility may be reduced. Shepperson (1960) compared the digestibility of hay made by three methods (slow field making, rapid field making, and barn drying) with the digestibility of the herbage cut. The digestibility of all the hay samples was lower than that of the original crop, the depression in digestibility being directly related to the time the cut crop remained on the field before it was lifted, the amount of mechanical tedding needed before lifting, and the extent of leaching by rain. The hay under the least favorable conditions was 7 units less digestible than the herbage cut. Many other similar experiments (Watson and Nash, 1960, p. 53) have confirmed the importance of rapid field curing if depression of digestibility is to be avoided; but it is certain that most of the hay made today still suffers high losses of both dry matter and nutritive value. As Shepperson ( 1960) indicated, nutrient losses during haymaking can be so great that crop digestibility, based on species and stage of maturity, may give only a very approximate guide to the digestibility of the resulting hay. Further deterioration can occur in storage. Even under optimal conditions there is a slow loss of dry matter; thus Melvin (1963) found a 10 percent loss of weight from ryegrass hay stored for 40 weeks, and Greenhill et al. (1961) showed that storage losses increased with increased moisture content in the hay and at higher temperatures. These losses are mainly in the sucrose, glucose, and fructose fractions in the hay and may lead to some decrease in digestibility. More serious is the depression in digestibility which occurs when hay overheats in storage (Watson and Nash, 1960), and molding can also lead to reduced intake of hay (Demarquilly, 1966a) and may present a toxic hazard to both animals and man (Lacey, 1968).

THE NUTRITIVE VALUE O F FORAGE CROPS

71

B. THEGRINDING A N D PELLETING OF DEHYDRATED FORAGES One of the most important advances in the period under review has been the discovery that grinding and pelleting a dried forage may lead to a considerable increase in its nutritive potential. Earlier studies were mainly practical, but recent research has greatly clarified the mechanisms involved (see reviews by Minson, 1963; Beardsley, 1964; L. A. Moore, 1964). Most workers agree (a) that the voluntary intake of pelleted ground forages is considerably higher than of the corresponding long forage, (b) that under ad libitum conditions the long forage is more digestible, but (c) that there is usually an increase in the net intake of digestible energy on the pelleted forage. The magnitude of these effects differs, however, among forage species and with the stage of maturity of a given forage (Heaney et al., 1963). These results can primarily be attributed to the markedly higher rate of passage through the digestive tract of the small particles in ground forages compared with the larger particles in long or chopped forages. The latter must remain within the rumen until they are broken down by mastication and microbial digestion to a size small enough to pass through the reticulo-omasal orifice (Campling et a/., 1963). As voluntary intake of forage depends largely on rate of passage (Section V), the animal is able to eat more of ground than of chopped forage. However, because the ground forage passes more rapidly through the reticulorumen there is much less time for it to be subjected to microbial digestion, and digestibility, particularly of the fiber fraction, is reduced (Campling et al., 1963). Dehority et al. ( I 962) and Tilley and Terry ( 1963) had shown that the digestibility of forage in vitru is increased by fine grinding (presumably by making the fiber structure more accessible to microbial attack), but in vivo any such effect seems to be outweighed by the reduced time available for digestion. Hinders and Owen ( I 968) have also shown a change in the site of digestion of fiber, only 60 percent of the fiber which is digested from pelleted lucerne being digested within the rumen (the remainder being digested in the hind tract, including the cecum), compared with over 90 percent with lucerne fed long. As a given crop becomes more mature, the increase in intake following grinding is greater, however, than the corresponding decrease in digestibility, so that the effect of grinding on digestible intake increases. The digestibility of legumes is depressed by grinding less than that of grasses; thus Demarquilly and Journet ( 1 967) found an average decrease of 9 units in digestibility for grass compared with 4 units for lucerne, and

72

W. F. RAYMOND

Buchman and Hemken ( 1964) reported little reduction in digestibility when lucerne was ground and pelleted. The difference in response between grasses and legumes is in line with the data in Fig. 3. At a given level of digestibility the “digestible” fraction of lucerne contains much less digestible fiber than does grass, and so might be expected to be less sensitive to reduction in fiber digestibility following grinding. Demarquilly and Journet (1967) also suggested that grinding lucerne led to an increased rate of fermentation within the rumen, in contrast to evidence of a marked depression in rate of cellulose digestion with grass hay (Campling et al., 1963). The stage has now been reached when the terms drying, grinding, and pelleting may be inadequate descriptions of these processes, and more detailed specifications, e.g., of temperature and duration of drying, and of particle size and pellet dimensions, will be needed. Possible effects of different drying conditions have been noted (Sections IV, C, VI, C). Demarquilly and Journet ( I 967) found that very fine grinding through a 1.5-mm. screen depressed grass digestibility more than through a 5-mm. screen, but that there was little increase in intake consequent on the finer grinding, which may thus be disadvantageous. Kamstra and Jahn (1966) found that very high pellet pressure may increase cellulose digestibility. The pelleting process itself may also be important, by making ground herbage (a dusty material) more acceptable to stock: conversely practical experience confirms a marked reduction in intake when pellet density is too hard. Ronning and Dobie ( 1962) have reported detailed studies on the effects of pellet size on voluntary intake, which may be reduced below that of the long forage when unsuitable pellets are made. Pelleting may also have less effect when it is applied to forages of very low protein content. Thus Minson ( 1 967) found an increase in intake of only I4 percent when unfertilized Pangolagrass (3.7 percent crude protein) was pelleted, compared with a 30 percent increase with fertilized grass of 7.2 percent crude protein; intake of the low protein grass may have been limited by the low nitrogen status of the animal (Egan and Moir, 1965) rather than by rate of passage, and so have been relatively insensitive to grinding. It also appears that the metabolizable energy value of ground forages cannot be calculated from Eqs. ( 1 9) and (20), because of the reduction in methane production consequent on the reduced fiber digestion within the digestive tract, compared with long forage: net energy values also cannot be calculated because of the lower heat increment with milled than with long forages. As a result Blaxter and Graham (1956) found very similar net energy values for chopped dried grass and finely ground

THE NUTRITIVE VALUE OF FORAGE CROPS

73

grass, even though the energy digestibility of the latter was 1 1 . 1 units lower. Recently Graham ( 1967) has reported that the net energy value of pelleted rations can be further improved if they are fed at frequent intervals rather than once or twice per day. Using a comparative slaughter technique, Paladines et al. ( 1964) found very similar levels of digestibility when equal intakes of chopped o r pelleted hay were fed to lambs, but lower body energy gains by the lambs on the chopped hay. This they accounted for in terms of higher heat increment (wastage) on the chopped hay. Under ad libitum feeding conditions they attributed 78 percent of the higher energy gains on the pelleted hay as resulting from increased intake and 2 2 percent resulting from increased efficiency of utilization of metabolizable energy; the possible role of decreased methane production (found by Blaxter and Graham, 19S6) could not be assessed with the experimental technique used. Numerous experiments (reviewed as above) have shown the effects of processing dried forages on animal production, and many of these have related the results to measured intake and digestibility data. These results are often difficult to interpret because of the different significance of digestibility for long and ground forages. They are also complicated when concentrates are fed with the forage, and particularly when concentrates and forage are pelleted together. Thus the intake of pellets containing ground forage and concentrates may be less than that of the long forage and concentrates fed separately (McCroskey et al., 196 1 ) . There is often a marked effect of grinding and pelleting on rumen volatile fatty acids, and many experiments have shown higher propionate: acetate proportions in the rumen liquor of animals fed on pelleted compared with long feeds. Earlier experiments had indicated that this would only occur with mixed forage-concentrate feeds, but P. L. Wright et al. ( 1963) showed a marked shift from acetate to propionate production when hay was ground, and Demarquilly and Journet (1967) have found a similar but smaller effect with dried lucerne. As with other ruminant feeds, the higher nutritive value of pelleted feeds has sometimes been attributed to this higher proportion of propionic acid in the rumen acids, but it appears likely that the lower methane production and the lower heat increment with ground feeds may be more important factors than the differences in rumen acids. An important exception occurs with milk production, where there is clear evidence of a depression in the butter fat content of milk associated with the decreased acetate production when ground forage is fed. G. D. Thomas et al. ( 1 968) found a depression in milk fat from 4.6 to 3.9 percent when coarsely ground hay in a mixed ration was replaced by finely ground hay; at the same time there

74

W. F. RAYMOND

was a decrease from 68.3 to 59.2 percent in the molar proportion of acetate in the rumen acids. Another quite different effect from diets of ground forages is the disease of rumen parakeratosis, which is characterized by hardening of the small papillae on the rumen wall, and which leads to reduced nutrient absorption and decreased animal production. The problem is most acute with finely ground feeds, but is reduced when small amounts of long roughages are fed (Garrett et al., 1961). Processing forages by grinding and pelleting clearly offers exciting possibilities for improving levels of ruminant production. Much more detailed work is needed, however, if such processing is to become an integral part of a planned program of crop production and animal feeding, rather than a somewhat empirical operation, as at present. Its implications may then extend beyond purely nutritional advantages, with drying and pelleting providing efficient conservation of forage crops in a form suited for mechanized handling and feeding (Raymond, 1968).

c. THE ENSILINGOF

FORAGE CROPS

Ensilage can be simply described as the storage of wet crops under anaerobic conditions by preservatives that inhibit microbial or enzymatic changes in the crop. The most common preservatives are acids-either mineral or organic acids added to the crop, or organic acids produced by bacterial fermentation of carbohydrates present in, or added to, the crop. In general, the higher the moisture content and the buffering capacity from organic salts in the forage, the lower is the pH needed to ensure stability (Playne and McDonald, 1966). Crops high in crude protein and moisture content, such as legumes and nitrogen-fertilized grass, are the most difficult to ensile; with these, wilting to reduce moisture content, or the use of additives, is advised, in particular when the ratio of crude protein:soluble carbohydrate in the forage exceeds 2 (Gordon et al., 1964). 1 . The Voluntary lntake of Silage The very considerable literature summarized by Watson and Nash (1960) reveals a remarkable imbalance of work to that date toward studies of the chemistry and microbiology of the ensiling process compared with studies of the effect of ensilage on the nutritive value of the product. Deficiencies in silage as a ruminant feed were often noted in practice, but the first formal definition of the problem appears to have been by Presthegge ( 1 959) and L. A. Moore et al. ( 1 960), who showed that the voluntary intake of silage was less than that of hay. This has

T H E NUTRITIVE V A L U E OF FORAGE CROPS

75

been confirmed by many other workers, and Harris and Raymond ( 1963) have also shown that the intake of silage is generally lower than that of the fresh crop. In all these experiments products from the same crop were compared: most previous comparisons of hay and silage had been confounded by silage being made at the “silage stage of growth” and hay at the “hay stage,” that is from a more mature and less digestible crop. Presthegge (1959) and L. A. Moore et al. (1960) also showed that the intake of silage was increased if the crop was wilted before ensiling. These results appeared just at the time that the significance of voluntary intake was becoming widely recognized, and as a result particular emphasis came to be placed on the importance of wilting crops before ensiling so as to ensure high intake levels. As noted above, wilting may be advantageous in ensuring an efficient ensilage process, and is essential if silage is to be made in tower silos. But many crops, such as forage maize and ryegrass, contain sufficient available carbohydrate to give stable, low pH, silage with minimal wilting. Chemical additives can also be used with unwilted crops. In these cases wilting, which makes the silage system more weather-sensitive, appears unnecessary as an essential part of the ensilage process, but might appear advantageous because it leads to an increase in silage intake. Because of the practical problems in wilting, recent research on the causes of low intake of silage has thus aimed to develop alternative methods of increasing intake which will obviate the need for wilting. As with forage intake, it is clear that there is no single factor causing the intake of unwilted silage to be lower than that of the equivalent fresh or dried forage. It is not due to the high moisture content per se in the silage, for addition of water to hay or wilted forage does not decrease voluntary intake (J. W. Thomas et al., 1961). Campling (1964) has suggested that silage within the rumen forms a fibrous dough from which “digested” feed particles can pass only with difficulty to the hind tract, so that rate of passage, and as a result level of feed intake, is restricted. Most workers, however, have considered that during the ensilage process chemical compounds are produced which limit intake. In the case of poorly fermented silages with high contents of butyric acid and amines, this could be a direct effect of unpalatability, but Neumark ( I 962) suggested that protein breakdown products of specific pharmacological activity might also be present in such silages. Neumark’s original suggestion that the active compound might be histamine was not confirmed by P. McDonald et af. (1963), who found no decrease in voluntary intake when histamine was added to silage. More recently, however, Neumark and Tadmor ( 1968) have indicated that histamine is active only when

76

W. F. RAYMOND

acetic or formic acid is also present, and that these compounds may act within the abomasum rather than within the rumen itself. More serious is the problem of low intake with the low pH silage which the efficient preservation process aims to make, and in which there is minimal degradation of the protein fraction. There is now some evidence that this is at least partly due to the high content of free organic acids in such silage (Harris et al., 1966; McCullough, 1966). It is recognized that some chemical additives, such as mineral acids or ammonium bisulfate may reduce silage intake (Watson and Nash, 1960, p. 641; McCarrick et al., 1965), but this has been attributed to the ruminant animal's inability to metabolize inorganic anions, in contrast to the organic acid anions, which comprise its main energy metabolite. However, it has been found that a number of silage constituents, including lactic, acetic, and propionic acids and longer-chain fatty acids, may lead to reduced feed intakes when they are added to silage before feeding, or infused directly into the rumen (J. W. Thomas et al., 1961; Rook et al., 1963; Ulyatt, 1965). Thus McLeod, Wilkins, and Wilson ( 1968, unpublished) have shown a linear depression of silage intake by lambs as the pH of silage was progressively lowered by the addition of lactic acid (Table 11). Lambs were used because they appear to be most sensitive to the factors in silage that depress intake. In the reverse direction,' McCarrick et al. (1965) and McLeod, Wilkins, and Wilson (1968, unpublished) have found significant increases in voluntary intake by the addition of sodium bicarbonate to silage. In the latter experiments the intake of unwilted silage by lambs and calves was raised by I5 percent when the silage was partially neutralized from pH 4. I to pH 5.3 before feeding. TABLE I I The Voluntary Intake, by Lambs, of Silage Acidified to Different Levels of pH by Addition of Lactic Acid".b Silage Oeginal Acidified

PH

I 2 3

5.4 4.8 4.4 3.8

Intake (DM % LWY 2.4 2.3 2.2 I .9 S.E. of treatment mean, 0.12 ~~~

"FromMcLeod, Wilkins, and Wilson (unpublished). "The original silage, pH 5.4, was prepared by partially neutralizing an acid silage, pH 4.0, with sodium bicarbonate. 'Dry matter as percentage of liveweight.

THE NUTRITIVE VALUE OF FORAGE CROPS

77

The exact mechanism whereby the organic acids in silage may limit the voluntary intake of silage is still uncertain. I t appears unlikely to operate directly within the rumen; while R. G. Warner et al. ( 1966) found a 40 percent reduction in hay intake when rumen pH was reduced to 6.0 by infusions of lactic or citric acids, this depression of hay intake could have been the result of decreased cellulose digestibility at this low pH (Section IV, C, 3) which would be quite uncharacteristic of that of the animal fed on silage, which is generally in the pH range 6.6 to 6.8. It appears that silage is largely neutralized by saliva before it is swallowed. Thus Lambourne ( 1 965, unpublished) used esophageal-fistulated sheep to collect the feed bolus swallowed by the animal, and showed that silage (pH 4.0) was partially neutralized (ca. pH 5.8) before it was swallowed. Unwilted silage (pH 4.0, 20 percent dry matter) can contain 20 times as much H+ ion per unit of feed dry matter as unwilted silage (pH 5.0, 40 percent dry matter); thus it may be postulated that the intake of unwilted silage is lower than that of wilted because of the much greater quantity of acid to be neutralized by saliva during the process of ingestion. Orth and von Kaufmann ( 1966) have also suggested that saliva secretion may be depressed by the high acid content and lack of physical structure in high-moisture silage. There is as yet no firm evidence that silage intake is limited via restriction of salivation. But if the apparent effect of high silage acidity in depressing voluntary intake is further confirmed it may indicate a possible contradiction between high intake and the requirement for high acidity in the ensilage process. The comparison of the voluntary intakes of hay, wilted silage, and unwilted silage is an important and legitimate research study. But the results of these studies, which have compared hay or silage when each was fed as a sole feed, have perhaps been extrapolated too widely in advice on practical feeding systems, in which hay or silage are most unlikely to comprise the sole feed of productive livestock. The relevant question then is how much hay or silage is eaten when both are fed in combination with the other ration components. Recent information suggests that the superiority, in intake terms, of hay o r wilted silage compared with unwilted silage, when each is fed alone, may largely disappear when they are fed in mixed rations. Thus Osbourn (1967) found that the intake of hay by lambs was 35 percent higher than that of unwilted silage (pH 4.4, 2 3 percent dry matter) made from the same crop. As increasing amounts of rolled barley were fed, the hay intake decreased by 0.68 g. per gram of barley fed, but the silage intake decreased by only 0.28 g.; when barley comprised 40 percent of the total dry matter intake, there was no difference between the hay and silage intakes (Fig. 7). Campling and Murdoch

78

W. F. RAYMOND

(1966) also found that the intake of silage by dairy cows was depressed less than the intake of hay when concentrates were fed, and S. M. Brown ( 1 960) found that the intake of unwilted silage by cows was depressed FORAGE INTAKE q. D. M{ kq.Wo'7J/24hr.

"1

---

HAY y = - O . b 4 ~ + 5 2 . 9 0 SILAGE y = - 0 . 2 6 ~ + 3 8 . 6 0

20

1 0 20 INTAKE OF BARLEY q. D.M./ kq.Wo.7y24 hr.

30

40

FIG.7. The voluntary intake, by lambs, of hay and silage made from the same crop of ryegrass and fed with increasing levels of rolled barley. D.M. = dry matter. (Based on Osbourn, 1967.)

less than that of wilted when both were supplemented by concentrates. Of particular interest are recent experiments in which supplements of pelleted dehydrated forages have been fed with silage (Wilkins and Osbourn, unpublished). S.24 ryegrass was conserved either as silage (pH 4.3, 20 percent dry matter) or as pellets after dehydration and grinding. The dry matter intakes of the silage and pellets when fed as sole feed to sheep were 15.3 g./kg. LW and 26 g./kg. LW, respectively. Increasing amounts of pellets fed with the silage resulted in no decrease in the amount of silage dry matter eaten until pellets comprised 40 percent of the total ration: at this point the intake of silage plus pellets was similar to that of the pellets fed alone, and the intake of digestible dry matter was in fact higher because of the lower digestibility of the pellets than of the silage. The nutritional and practical implications of this observation could

THE NUTRITIVE VALUE O F FORAGE CROPS

79

be of considerable importance, for instance, in the use of dehydrated lucerne (high crude protein and minerals) as a supplement to low protein silage made from ryegrass or wholecrop cereals (Raymond, 1968). A further interesting development has been the addition of urea to the crop before ensiling, with the aim of increasing the crude protein content of silages made from low protein crops, in particular corn silage (reviewed by M. H. Briggs, 1967; Essig, 1968). An addition of 0.5 to 1.0 percent urea appears to increase dry matter loss slightly, and markedly to increase the free ammonia content in the silage; a perhaps unexpected observation is that simultaneous addition of 0.5 percent ground limestone improved silage fermentation. Essig ( 1968) reported a 5 percent increase in gains when treated corn silage was fed to cattle. In general it seems likely that, in mixed rations, the intakes of wilted and well-preserved unwilted silage may not differ greatly (S. M. Brown, 1960). This could mean that a decision to wilt a crop before it is ensiled should be made on the basis that wilting is needed to ensure successful ensilage, rather than with the aim of increasing voluntary intake. 2 . The Digestibility of Silage While most attention has been paid to the intake characteristics of silage, the digestibility of silage is also an important determinant of its nutritive value. Earlier studies (Watson and Nash, 1960, p. 397) had indicated a rather lower digestibility of silage than of the crop ensiled. In some cases, however, this may have been because the intake of silage was based on dry matter determinations made at 100°C. As Harris and Raymond ( 1963) have shown, samples of silage from digestibility experiments may suffer considerable losses of volatile constituents during oven-drying. This leads to an underestimate in the amount eaten and in the measured digestibility (Eq. 2). These authors measured the true dry matter content of silage by toluene distillation and showed almost identical digestibilities for silage and the original crop. Harris ( 1 963) also found that the digestibility of corn silage was very similar to that of the fresh crop, and Johnson and McClure ( 1968) found the digestibility of corn silage cut at increasing stages of maturity to be 65.3 percent, 71.9 percent, and 69.8 percent, in line with the expected changes in crop digestibility. In this experiment, voluntary intake was closely related to silage digestibility (42.9, 58.9, and 54.0 g./kg. LW0.75, respectively), but this has not been found with unwilted grass silages (Harris and Raymond, 1963). Ensilage under poor conditions can lead to depressed digestibility for instance, when there are high losses of soluble constituents in effluent,

80

W. F. RAYMOND

or when overheating and caramelization occur (Watson and Nash, 1960; Wieringa et al., 1961). A small decrease in dry matter digestibility may also occur when forage is wilted before ensiling. Thus Harris et al. (1966) found a mean dry matter digestibility of 63.9 percent for a number of wilted silages (26 to 54 percent dry matter) compared with 66.3 percent for unwilted silages from the same crops (16 to 23 percent dry matter); Schulz and Oslage (1967) found a similar fall in digestibility with some silages, but concluded that this did not necessarily result from wilting. The reason for this lower digestibility is not known, but if it occurs generally it must somewhat reduce the advantage attributed to higher intake with wilted silages. X. The Nutritive Value of Grazed Forage

Grazing, the most common method of using forages, imposes a number of modifying factors on the basic components of nutritive value already discussed. Thus “forage grazed” may differ markedly in composition from “forage on offer,” because animals are able to graze selectively; the amount of forage eaten may depend as much on extrinsic factors such as feed availability and environment as on the intrinsic factors discussed in Section V ; and the nutrient requirements of grazing livestock may differ from those of housed livestock, with which all basic data on requirements have been determined. A. MEASUREMENT OF THE NUTRIENT INTAKE BY GRAZING ANIMALS

Central to any consideration of these factors is the need for information on the nutrient intake by grazing ruminants. Yet despite much research effort in the last twenty years the techniques now available still appear inadequate, and they limit the precision with which grazing studies can be interpreted. Two distinct types of technique have been used, involving, respectively, measurements on the pasture and measurements on the animal. The former, which include “IN - OUT” pasture sampling and “browse unit” estimates, have been fully reviewed by D. Brown ( 1 954) and J. T. Reid (1962). They are not considered further here, except to note that the technique of estimating forage intake as the difference between yield samples before and after grazing appears to be reasonably precise only when applied to “strip” grazing, where the difference between the yield estimates is large and pasture growth during the period of grazing is small.

81

T H E N U T R I T I V E V A L U E O F FORAGE CROPS

The most commonly used techniques are based on a rearrangement of Eq. (2): Intake of forage (1)

= weight

of feces voided (F) X

I00 (100- 7% digestibility of forage eaten (D)

1

(24)

and requires estimates to be made of (F) and (D); from (D) the metabolizable and net energy values of the forage eaten can also be calculated to give estimates of nutrient intake (Eqs. 19 and 20). (See reviews by Commonwealth Agricultural Bureaux, 196 I , p. 92; 1. W. McDonald, 1962; J. T. Reid, 1962.)

1. The Estimation of Fecal Output by Grazing Animals The most precise method of measuring fecal output is undoubtedly by total collection, using harness and fecal bags. However this method is laborious, is difficult under extensive grazing conditions, and may interfere with animal behavior and production, so that most workers have preferred the indirect estimation of fecal production by the use of indigestible tracers. A known weight of the tracer is fed to each animal daily; from the concentration of tracer in a sample of feces the weight of feces which would contain the total weight of tracer is calculated as the total output of feces. This requires (a) that the tracer is completely recovered in the feces and (b) that the sample of feces analyzed is truly representative. a . The Recovery of Tracers. Many different substances have been examined as fecal tracers, but the most commonly used is still chromic oxide, Cr203,which is cheap and nontoxic. Administration is easy with the use of factory-produced capsules (Commonwealth Agricultural Bureaux, I96 I , p. 93), as is also the analysis of the oxide in feces (Commonwealth Agricultural Bureaux, 196 I , p. 152). Unfortunately the fecal recovery of this tracer has so often been found to diverge from the theoretical 100 percent that the assumption of complete recovery, without confirmation under the conditions of the particular experiment, must be questioned. Thus Kane et al. ( 1950) and Raymond and Minson ( 1955) found mean recoveries close to 100 percent, whereas several other workers, including Christian et al. (1965; 92.8 percent) and McGuire et al. (1966; 94.2 percent) have found low recoveries. The reasons for such low recoveries, particularly over the 66-day feeding period studied by Christian et al. (1965). which must have eliminated “end errors,” are still uncertain; they may include regurgitation and rejection of cap-

82

W. F. RAYMOND

sules, an impurity of soluble chromate in the chromic oxide, and differential loss of fine particles of the oxide during milling of the feces sample prior to analysis (see, e.g., P. H . Bailey et al., 1957). But they do indicate that error, in an overestimation of fecal output, can arise unless an independent check on recovery is made within the experiment. Attention has thus been given to other tracers, including techniques based on the use of the rare earth elements, which have been shown to be almost completely indigestible (Bell, 1964). With the use of radioactive materials, only minute amounts of tracer are needed, and physical, rather than chemical methods of fecal analysis can be used; furthermore the rare earths appear to become firmly bound to plant materials and to pass through the digestive tract with the undigested fiber residues (Bell, 1964) in contrast to chromic oxide, which appears to move with the liquid digesta phase (Raymond and Minson, 1955). As a result the diurnal variation of excretion of tracer in the feces seems to be reduced (Section X, A, 1 , b). Radiocerium oxide has been used by Huston and Ellis (1968), but dysprosium appears to have particular advantage (Ellis, 1968) as its concentration in feces can be measured by activation analysis, so avoiding the need to feed radioactive material, with the consequent problem of waste disposal. The fecal recovery of dysprosium was 99.8 percent, and this tracer showed only a small diurnal variation in excretion (Ellis, 1968). Radiocerium also gave high recovery and low variation, and variation among animals was found to be less than with chromic oxide or polyethylene glycol (Huston and Ellis, 1968). Further development of the use of radioactive tracers is awaited with much interest. b. The Pattern of Excretion of lndigestible Tracers. Where the tracer can be thoroughly mixed with all the feed eaten it is fairly uniformly excreted in the feces; but this is not possible with grazing livestock, and here the pattern of dosing of the tracer, which is generally given once or twice daily, is clearly reflected in a markedly nonuniform excretion in the feces. Kane et al. ( I 952) studied this under indoor feeding conditions, and proposed that a representative sample of feces could be obtained by “grab” sampling directly from the animal’s rectum at fixed times each day. Raymond and Minson ( 1 955) showed, however, that even under indoor conditions marked variations in excretion pattern could occur, depending among other factors on the type of feed and the level of feeding. Both of these can affect the rate of passage of digesta through the tract, and so the time between chromic oxide dosing and the peak of excretion in the feces. These authors concluded that “grab” sampling would thus be invalid under grazing conditions, where the excretion pattern would

THE NUTRITIVE VALUE O F FORAGE CROPS

83

depend on the kind and amount of forage grazed, the very parameters the technique is intended to measure. They proposed that feces should be sampled directly from the sward being grazed, and details of sampling methods have been described (Commonwealth Agricultural Bureaux, I96 1 , p. 94). To permit the feces from different animals to be sampled separately, each grazing animal is dosed daily with a capsule containing polystyrene granules of a particular color, which can readily be identified in the feces lying on the field (Commonwealth Agricultural Bureaux, I96 I , p. 95). This technique is suitable only for cattle, as sheep are able to chew the granules. Another approach has been to minimize the extent of the diurnal variation in the content of tracer in the feces by feeding it in a form which passes through the tract with the undigested residues. Corbett et a f . ( I 960) developed a kraft paper impregnated with chromic oxide, with which a high proportion of the oxide remains associated with the feed residues, rather than moving with the liquid digesta phase. They showed a marked reduction in diurnal variation of oxide excretion in the feces and, as expected, a change in the times of maximum and minimum excretion, compared with the oxide dosed in capsules. Cowlishaw and Alder (1963) showed a similar improvement in uniformity of excretion with the chromic oxide paper, but concluded that the greater ease of administration of commercially produced capsules justified the effort involved in sampling feces from the sward. The use of the rare earth elements, noted in the previous section, may well be a promising solution to this problem. 2 . The Estimation of the Digestibility of Grazed Forage The earliest applications of Eq. (24) made use of a determination of the digestibility of forage cut from the sward being grazed. Observations showed that this could vary markedly from the forage actually being eaten by the grazing animals. These earlier studies are likely to have underestimated the amount and digestibility of the forage grazed, because this forage tends to be the more digestible part of the forage on offer - as indicated by the generally higher fiber content and lower protein content in forage sampled after grazing than before grazing. However, estimates of the composition (digestibility) of forage grazed, based on samples taken before and after grazing, are subject to considerable error (D. Brown, 1954). Thus attention has been given to methods of estimating the digestibility of grazed forage based on chemical analysis of the feces produced. In the ratio techniques, digestibility is calculated from the relative contents of a naturally occurring “indigesti-

84

W. F. RAYMOND

ble” tracer in samples of “forage grazed” and in feces (summarized by J. T. Reid, 1962). But because of the difficulty of obtaining relevant forage samples, these methods have been superseded by the fecal index techniques (see Raymond et al., 1954). These were developed from observations by Lancaster ( 1949) from which it appeared that the nitrogen content of feces was correlated with the digestibility of the forage eaten; thus the nitrogen content of the feces of grazing livestock should allow an estimate of the digestibility of the forage selectively grazed from a sward. Fecal contents of chromogen (J. T. Reid et al., I952), nitrogen and fiber (Raymond et al., 1954), lignin and methoxyl (Richards et al., 1958), phosphorus (Barrow and Lambourne, 1962), and copper, magnesium, and silica (McManus et al., 1967) have all been shown to relate to forage digestibility and these fecal index markers have been widely used in grazing experiments. In a related technique Owen and lngleton (1963) found that the intake of forage dry matter could be related directly to a dissolved fecal fraction (feces extracted for 18 hours with 0.2 N hydrochloric acid at room temperature) thus:

+

Intake of forage (g.) = 250 17.8 X weight of dissolved fecal fraction excreted (g.) (25) Langlands and Corbett (1964) reported a less precise estimate by this method than by the use of a fecal nitrogen relationship. In all these studies the precision of each technique was assessed in terms of the statistical correlation between forage digestibility and fecal composition, based on indoor feeding experiments with cut forages. The aim was to develop a relationship based on a wide range of forages, so as to be generally applicable, yet with low errors of prediction (Raymond et al., 1954). Arnold and Dudzinski ( 1963) tested several increasingly complex statistical relationships with only a marginal reduction in prediction errors. McManus et al. ( 1 967) found that a multiple fecal index relationship based on fecal contents of nitrogen, copper, and silica gave slightly lower errors than one based only on fecal nitrogen. Hennaux et al. ( 1967) found similar precision of estimation of forage digestibility using eithep chromogen ratio or fecal chromogen techniques. L. E. Davis et al. (1 967), however, found better prediction using fecal nitrogen because of the low recovery of forage chromogen in the feces, particularly with lucerne: fecal nitrogen reflected differences in digestibility between Bromus and lucerne which were not picked up by fecal chromogen, indicating that the latter might not be valid in comparisons between forage species. It is now evident, however, that other, possibly more serious, errors

THE NUTRITIVE VALUE O F FORAGE CROPS

85

may be involved in the practical application of fecal index methods. First, there is the possibility that “application” errors may arise when an indoor-based regression is applied in the field, because of the different levels of forage intake likely in the two situations (Raymond et al., 1956). Indoor experiments with the same forage fed at different levels of intake indicated that fecal nitrogen and chromogen contents may change in a self-compensating way, but that regressions based on fecal contents of fiber, lignin, and methoxyl may introduce error if forage intake in the field differs from that indoors. J. B. Hutton and Jury ( I 964) have also found that the statistical errors of fecal index relationships based on forages fed under ad libitum conditions are greater than under the restricted intake conditions of most of the earlier digestibility experiments ( e g . Raymond et al., 1954). Second, it was found that the assumption in the regression analyses that the errors were randomly distributed was incorrect, and that different populations of feeds showed a biased distribution about the “general” regression line. Thus Greenhalgh and Corbett ( 1960), Minson and Kemp ( 196 1 ), and Lambourne and Reardon ( 1963) showed considerable differences in the relationships between forage digestibility and fecal nitrogen content with the same forage species at different seasons of the year, or at different levels of nitrogen fertilization, and similar differences occur in relation to fecal chromogen content. As a result the feces from animals grazing pasture heavily fertilized with nitrogen contain more nitrogen than feces from less heavily fertilized pasture; this would indicate a higher level of digestibility for the fertilized forage, but this is not found in practice (Minson and Kemp, 196 I). This indicated the need for “restricted” regressions, each based on forage cut from the sward under study, in order to reduce error due to bias. The continuous digestibility experiment, using fresh forage cut daily from the sward, was considered appropriate for this purpose (Greenhalgh and Corbett, 1960; Commonwealth Agricultural Bureaux, 1961, p. 97), but the need for such associated indoor feeding must limit the situations in which this technique can be applied. More serious, however, is the recent evidence which questions the basic premise of the fecal index technique, that a relationship based on the feeding of cut forage indoors can be used to predict the digestibility of that part of the forage that the grazing animal selects in the jield. This was indicated by the results of Minson and Kemp ( 196 I ) that forages of the same digestibility, but containing different levels of nitrogen, give feces of different nitrogen contents. Lambourne and Reardon ( 1 962) showed that the “stem” and “leaf’ fractions of a sward may produce

86

W. F. RAYMOND

quite different digestibi1ity:fecal nitrogen relationships (Fig. 8), and Pearce et al. (1962) found different relationships for the “top” and “bottom” fractions of a sward, harvested and fed separately in digesti-

& I

2

b I

3 % N in fecal O.M.

4

FIG. 8. Relationship between percent digestibility of forage organic matter and percent nitrogen in fecal organic matter (O.M.), for whole-cut forages and for leaf-rich and stemrich fractions separated from these forages. As fecal nitrogen decreases from 3.6 percent to 2.6 percent, the whole-plant regression indicates a fall in digestibility of herbage grazed from 8 1 percent to 73 percent ( x to y). The actual fall in digestibility, as the type of feed selected changes from leafy to stemmy, would be much less, i.e., from x’ toy’, as indicated by the dashed line. (From Lambourne and Reardon, 1962; Raymond, 1966b.)

bility experiments. The effect of this is illustrated in Fig. 8; use of a regression equation based on whole forage plants will indicate a much greater effect of selective grazing on the digestibility of forage eaten than is in fact the case. Confirmation of this conclusion has been provided by studies with fistulated animals (below). These show that the earlier concept that the digestibility of grazed forage decreases markedly on the successive days of grazing on a paddock is often erroneous. This must mean for instance that the fecal index technique cannot validly be used to study the effects of grazing pressure on forage selection or intake. As Langlands (1 967b) concludes, the appropriate fecal index relationship must be derived from herbage similar to that being selected by each particular group of animals-in which case a less indirect method of estimating forage digestibility must surely be possible. “This must mean that, even in experiments where ‘local’ regressions have been used, incorrect estimates may have been made of the effects of selective graz-

THE NUTRITIVE VALUE OF FORAGE CROPS

87

ing, whilst there must be even greater doubt of the validity of the results of the large number of (grazing) experiments which have used ‘general’ regressions” (Raymond, 1966b).

3 . The U s e of Fistulate Techniques This conclusion has stimulated investigation into other methods of studying the forage intake of grazing animals, and in particular the use of fistulated animals for obtaining samples of the herbage actually being selected, on which botanical and chemical determinations can be made. The establishment of esophageal fistulas in ruminants was first described by Torell ( 1954), and their earlier experimental use was reviewed by Van Dyne and Torell (1 964). The more general application of the technique was advanded by the development of the split-plug closure technique, which reduces the hazards of the initial operation and simplifies the handling of the animals in the field (McManus et al., 1962). The technique was initially used for species-selection studies (see below), but many workers have also used it to estimate the nutritive value of the herbage grazed (e.g., Bedell, 1967; Bredon and Torell, 1967). I t has been important, then, to examine how far the extrusa sample collected through the fistula is fully representative of the feed eaten, and whether the chemical composition of the sample differs from that of the feed. D. L. Bath et al. ( 1956) and Grimes et al. ( 196.5) have shown reasonable agreement between the botanical composition of herbage eaten and that measured on the extrusa sample; but greater differences can occur in chemical composition, due both to the contamination of the sample with saliva before it can be collected and to the preparation of the sample for analysis. Thus Lombard and van Schalkwyk ( 1963) and C. M. Campbell et af. (1968) found higher ash contents in extrusa samples than in the herbage eaten, and Hoehne et al. ( 1967) found higher ash, phosphate, and chloride levels, lower nitrogen and water-soluble carbohydrate contents, but similar contents of crude fiber and lignin. Gonzalez Gonzalez and Lambourne ( 1 967) showed that when a hay-saliva mix was dried, the measured in vitro digestibility of the hay increased by up to 4 percent units, indicating the importance of removing saliva contamination before drying extrusa samples. Hoehne et al. ( 1 967) suggested that this should be by squeezing the sample to remove excess saliva before it is dried; Robards and Wilson (1967) have shown that washing the sample to remove saliva reduced in vitro digestibility by as much as 7 units. Thus the method generally adopted is gentle squeezing of the sample in a muslin bag, but the validity of any procedure used must be checked against forage samples wherever possible.

88

W. F. RAYMOND

A similar problem may arise when forage samples are collected manually via a rumen fistula: this technique has been used by, among others, Ridley et al. ( 1 963) and Tayler and Deriaz (1963), but it is applicable only to cattle because of limitation in the diameter of the cannula with sheep. The observer has to walk alongside the animal to make the collection, whereas with an esophageal fistula the extrusa samples are collected in a bag tied under the animal’s neck. An advantage of sampling from the rumen is that the sample can be related directly to the particular sub-area of pasture being grazed; measurement of the in vitro digestibility of such samples has provided valuable information on the effects of management and grazing pressure on selective grazing, which differs significantly from earlier evidence based on the fecal index technique (Tayler and Deriaz, 1963). The digestibility of the forage eaten by animals grazing immature forage changes relatively little as the sward is grazed down, despite a marked decrease in the nitrogen content in the feces produced. Where uneaten residues remain from a previous grazing, these form a low-digestibility component which dilutes the high digestibility of the new growth. If this residue is in a lower horizon than the new growth, cattle can select forage of high digestibility, but little selection occurs where the old and new growth are intimately mixed within the forage on offer, or where high grazing pressure is imposed so that selection is difficult. There are several limitations in the use of fistulated animals for estimating the digestibility of “forage grazed” and in the prediction of forage intake from Eq. (24). First, there is some uncertainty as to the validity of the digestibility estimate, because of the possible effect of saliva contamination, noted above. Second, in vitro digestibility is a parameter of a forage measured under standard conditions. It is not necessarily the same as the in vivo digestibility of the forage by the particular animals being studied, yet it is this measure of digestibility which is required in Eq. (24). Finally, extrusa samples are collected only at intervals from a proportion of the animals within a flock or herd, and intensive sampling may be necessary to obtain a representative sample. Langlands (1967b) has thus suggested that a “local” fecal index regression should be calculated between the in vitro digestibility of the extrusa samples and the chemical composition (nitrogen and cellulose) of the feces voided by the fistulated animals. He showed different relationships for whole forage fed indoors and for forage selected in the field as would be expected from Fig. 8. The latter relationship should allow corrections for selection on the basis of contents of nitrogen and cellulose in feces from the nonfistulated animals within the herd. A multiple-regression approach to the same problem has been described by Arnold and Dudzinski ( 1967a).

T H E NUTRITIVE VALUE OF FORAGE CROPS

89

An alternative method of using extrusa samples has been described by Cook et al. (1963), the contents of lignin or silica in the extrusa samples, and in the corresponding feces, being measured, and the digestibility of the forage eaten estimated by the ratio method. This method may well have advantage over the direct measurement of in vitro digestibility on the extrusa sample, although it relies on the assumption of 100% indigestibility of the lignin or silica, or the use of an indigestibility coefficient measured in vivo on herbage cut from the sward under test. For precise results it is also necessary for fistulated animals to be adapted to the pasture being examined, before they can be considered to give a representative sample of the forage grazed. Thus Langlands ( 1967a) found that fistulated sheep newly moved onto a ryegrass-clover pasture gave extrusa samples containing 3.8% nitrogen, compared with 3 . 3 3 % nitrogen in extrusa samples from sheep already adapted to the pasture. This indicates that fistulated sheep should not be moved, as tester animals, between different experimental groups. A novel approach, outlined by Wilkins ( 1966), is now being examined. This is based on the premise that there is a maximum extent to which the cellulose in a given forage can be digested (potential digestibility) which is controlled by the cellulose structure and degree of lignification. Cellulose digestibility in vivo will be less than this either because the feed passes through the digestive tract too rapidly for all the cellulose to be digested, or because the rumen medium is deficient in nutrients needed for efficient microbial digestion, or because rumen pH is not optimal for cellulose digestion. As a result the feces still contain cellulose which can be digested by an in vitro incubation. The level of cellulose digestibility in vivo (CD) is related to the potential cellulose digestibility (PCD, determined in a 6-day in vitro incubation) as under: CD =

100 (PCD - FCD) 100 - FCD

where FCD is the potential digestibility of the cellulose in the feces. The intake of cellulose can then be calculated from the quantity of cellulose excreted in the feces: Intake of cellulose =

O0 x fecal production 100-CD

(26)

x % cellulose in feces Analysis of the potential cellulose digestibility in a sample of forage eaten, collected via an esophageal fistula, and of the potential cellulose digestibility in the feces, permits calculation of C D from Eq. ( 2 5 ) . Substitution

90

W. F. RAYMOND

of this value for C D in Eq. (26) gives an estimate of the intake of cellulose. From the percentage content of cellulose in the forage sample the intake of forage dry matter or organic matter can be calculated: Intake offorage= intake of cellulose X

100 100 - content of cellulose in forage

7%

(27)

This may appear to be a complex procedure. But it is based solely on the determination in feed and feces samples of the content and potential digestibility of a chemical entity which can be analyzed simply and with considerable precision (cellulose by the method of Crampton and Maynard, 1938). More important, it appears to compensate for the animal factors (level of intake, effect of feed supplements, parasite damage) which, by affecting the in vivo digestibility of a feed, may invalidate both fecal index techniques and the use of in vitro digestibility determinations on fistula extrusa samples. It does, however, make the assumption that the process of digestion in vivo does not render potentially digestible in the feces cellulose which was not measured as potentially digestible in the feed. Despite this, it appears to offer the most logically valid approach to the estimation of grazing intake that has yet been proposed. B. THE BOTANICAL COMPOSITION

OF

GRAZEDFORAGE

The original use of the esophageal fistula by D. L. Bath et al. (1956) was to measure the botanical composition of grazed forage, and the technique has been widely used for that purpose to replace the earlier swardsampling techniques. This has been more successful with cattle than with sheep, presumably because sheep masticate feed more completely before it is swallowed. Ridley et al. ( 1 963) were able to distinguish smoothstalked meadowgrass from cocksfoot and tall fescue in herbage sampled from the rumen of cattle grazing a mixed sward, whereas Cook et al. (1958) were not able to distinguish between species grazed by sheep from cultivated pasture. However with rangeland forages Cook et al. (1967) were able to study species selection by both cattle and sheep and to show differences in preference between the two classes of stock. The technique has also been used successfully by Van Dyne and Heady (1965) and Leigh and Mulham (1966) on rangeland vegetation and by Hodge and Doyle ( 1967) on cultivated pasture. The latter workers were primarily concerned with selection of either ryegrass or clover from a mixed sward, and this degree of separation was readily achieved in extrusa samples taken from lambs and yearling sheep. In nutritional

THE NUTRITIVE VALUE OF FORAGE CROPS

91

terms selection of different plant fractions may often be as important as interspecies selection, and Tayler and Deriaz ( 1 963) were able to draw important conclusions from samples of leaf, stem, and dead and senescent forage separated from samples collected via a rumen fistula. This separation was essentially quantitative, whereas species separations must normally be accepted as qualitative, because of the considerable part of an extrusa sample which cannot be identified by species-but which could be allocated to a leaf or stem separate. The same restriction to a qualitative botanical separation also applies to the microscopic examination of feces samples for residues of epidermis and cuticle, characteristic of the different species in the plant community being grazed (e.g., Hercus, 1963; Stewart, 1967). C. THENUTRIENTINTAKEOF GRAZING LIVESTOCK It is evident from the consideration of the possible errors involved, discussed in the previous sections, that some reservation must be applied to many published estimates of the nutrient intake by grazing livestock. However, where the experimenter has used a “local,” rather than a “general,” fecal index relationship, the results are likely to be of reasonable validity.

I . The Energy Requirements of Grazing Animals

1. W. McDonald (1968) has emphasized the considerable assumptions implicit in the extrapolation, to field conditions, of energy standards based on controlled calorimetric studies. Most of the experiments in which fecal techniques have been used to estimate grazing intake indicate that the maintenance energy requirement of grazing livestock is considerably higher than of stall-fed animals (Coop and Hill, 1962; Langlands et al., 1963), particularly when the amount of pasture available is restricted (Lambourne and Reardon, 1963). Clapperton (1 964) showed that maintenance requirement is increased by walking activity; Blaxter ( 1 964) calculated that this was unlikely to increase the requirement by more than 18 percent, and that environmental effects, such as wind, rain, and air temperature would also be important. Graham ( I 966) has developed equations for calculating the maintenance requirement of grazing livestock in terms of locomotion, grazing energy, and environmental stress, and Lambourne and Reardon ( 1963) suggested that heat production may be increased by endocrine stimulation initiated by stress when the supply of forage is inadequate.

92

W. F. RAYMOND

2 . Factors Affecting Intake by Grazing Ruminants: “Extrinsic” Factors Under grazing conditions it is important to know to what extent forage intake is controlled by the “intrinsic” factors, discussed in Section V , and how far it depends on “extrinsic” factors associated with management and environment. Under effectively ad libitum grazing conditions, Hodgson ( I 968) found a linear decrease in intake by calves as the dry matter digestibility of the forage on offer fell from 82 percent to 68 percent, but under most practical conditions forage availability appears to be the more important factor (see review by Fontenot and Blaser, 1965). Thus Arnold and Dudzinski (1967b) concluded that the amount of dry matter available per acre accounted for 40 percent of the differences in intake between grazing treatments, and also that leaf length was important. But Greenhalgh et al. (1967) found that the organic matter intake by grazing cows decreased only from 26.4 to 23.9 pounds/day when the quantity of dry matter available was reduced from 45 to 25 pounds a remarkable result indicating virtually 100 percent utilization of the forage with only a marginal reduction in forage intake. In a comparison between ryegrass and cocksfoot, Greenhalgh and Reid ( 1968) found the ryegrass to be consistently more digestible than the cocksfoot (average 4.6 units), but only in grazing periods later in the season was intake on the ryegrass higher. Evidence on other extrinsic factors which may affect forage intake by grazing animals, such as sward structure, weather conditions, or soiling by excreta and fungal infections, is quite inadequate, and these urgently need detailed study. An example is recent work by Marten and Donker ( 1 966) in which forage, heavily fertilized with dung, was not unpalatable when fed to housed stock. This indicates that rejection of such forage by grazing animals may be due to the physical presence of residual dung in proximity to the forage, which MacLusky (1960) found might be affecting S O percent of the pasture at any one time. 3 . The Level of Production by Grazing Animals In most cases the effects of pasture species, management, and environment on the amount and quality of forage grazed can only be inferred indirectly from measurements of the resulting levels of animal production. Thus Cameron (1965) and Large and Spedding (1967) found very little effect on liveweight gains by either lambs or calves from very high levels of nitrogen fertilizer, and they concluded that intakes at low and high levels of fertilizer were very similar. In a series of papers (Rae et al., 1963, to Cramer et al., 1967), New Zealand workers reported markedly

THE NUTRITIVE VALUE OF FORAGE CROPS

93

higher rates of liveweight gain by lambs grazing short-rotation ryegrass than perennial ryegrass, and also higher gains when white clover was grown in mixture with either type of ryegrass. This was attributed to the production of a higher proportion of propionate to acetate during rumen fermentation of the more productive forages, but this conclusion perhaps underestimated the likely significance of differences in voluntary intake and digestibility between the forages. If priority must be placed on the order of importance of the components in Eq. ( 1 ) in determining the productivity of grazing ruminants it must surely be that intake > digestibility > utilization of digested nutrients. It is now also recognized that, under grazing conditions, nutritional differences between forages or between management treatments can be confounded by animal health factors, in particular by the effects of gastrointestinal parasites. In classical terms, parasites were considered to be of significance oniy when grazing livestock showed clinical symptoms of parasitism; however Spedding and his co-workers (Spedding, 1965) have shown that low levels of parasitism (subclinical infestation) can markedly reduce the productivity of grazing ruminants by reducing the amount of forage the animals will eat and the extent to which they are able to digest this forage, and possibly by stress on the animal. This effect was shown by Lindahl et al. ( 1 963), who found that lambs fed on cut forage gained more rapidly than those grazing on initially “clean” pasture, and both groups more rapidly than lambs grazing on infected pasture, even though parasitism in the latter two groups was a t a subclinical level. It is evident that grazing experiments, particularly with sensitive animals such as lambs, may often be measuring primarily the effects of management and pasture structure on the level of parasiticworm infestation on the forage, rather than any specific nutritional effects of the treatment. A response to supplementary feed could thus be because this feed is “worm free” rather than because it supplies additional nutrients. Moreover, the usual techniques of fecal egg counts are quite inadequate measures of level of animal infestation, from which to assess the relative contributions of nutrition and infection to the observed animal production response (Spedding, 1952). This led Spedding to develop the concept of the “worm-free” pasture as a tool in nutritional studies under grazing conditions, and of rearing “worm-free’’ animals for use in critical grazing experiments (Spedding et al., 1965). This has allowed nutritional effects to be studied at stocking rates in excess of 40 lambs per acre, which could have been catastrophic if carried out with normally reared lambs. In this way lamb gains on ryegrass pasture fertilized with 200, 600, or 1,000 pounds of

94

W. F. RAYMOND

nitrogen per acre could be compared, without the differential effects of parasitism obscuring the nutritional conclusion, that there was no difference in rates of gain on the three treatments (Large and Spedding, 1967). These techniques, originally conceived as experimental tools, are also likely to be of increasing practical and economic significance. Despite recent developments in anthelmintic treatment, the risk of parasitic infestation is a major factor limiting pasture intensification -the problem of stocking pasture with the number of productive animals needed to graze the high yields of forage than can now be grown. The concepts of the “worm-free” pasture, of the early separation of the young from its dam (which is generally the primary source of infection), and of the management of the breeding and productive units within a livestock population in separate nutritional environments, could be key features in future systems of exploiting the nutritional potential of forages (Spedding, 1965).

D. THE EFFECTOF MANAGEMENT ON THE PRODUCTIVITY OF GRAZINGANIMALS With adult animals the effects of parasitism are likely to be less serious, and their level of productivity is likely to reflect more closely the nutritional parameters in Eq. (1). A more critical approach to grazing experiments has followed the paper by McMeekan (1956) at the 7th International Grassland Congress. This led to the recognition of the importance of stocking rate in determining level of animal productivity, as against earlier emphasis on grazing method. Harlan (j958) showed that, as stocking rate was increased beyond a certain point, rate of production per animal began to decrease. Rate of production per acre however continued to increase beyond this stocking rate, producing the familiar double-exponential equations, expressed graphically by Mott (1960) and by Riewe ef al. (1 963) (Fig. 9) and recorded subsequently in numerous grazing experiments (e.g., McMeekan and Walshe, 1963; Conway, 1968a). This is a most important concept in forage research, for it poses the problem of how a high proportion of the forage grown can be grazed without reducing the level of productivity of the individual grazing animal. While the practical implications may differ between milking animals (where output per acre is immediately salable) and meat animals, and in different economic conditions, biological efficiency of land use demands a solution to the problem of achieving a high intensity of forage utilization (point S ) by high-producing animals ( P ) (Raymond, 1968,1969a).

95

THE NUTRITIVE VALUE OF FORAGE CROPS

The relative importance of intake, digestibility, and utilization will differ in different situations. At very low stocking rates, daily output per animal may in fact decrease as a result of the very low digestibility of

output Per animal

lutput Per acre

A

B

Crazinq intensity+ FIG.9. The responses in animal output, per animal and per acre, with increasing grazing intensity in pasture systems. (From Raymond, 1968.)

undergrazed forage. But at higher stocking rates, intake is likely to be the critical factor, with a secondary effect from the reduced possibility for selection of forage of high digestibility from that on offer. Raymond ( I 968) has used Fig. 9 to indicate three possible methods of combining high production per animal with high intensity of forage use: a. The division of the animal population into animals with currently high or low nutrient requirements; the former group graze first at low intensity, and are followed in the rotation by the lower-producing animals, which can still satisfy their nutrient requirements at a high intensity of grazing. Examples are “forward-creep’’ grazing, in which the lamb grazes ahead of the ewe (Dickson, 1959), and the split-herd technique with high- and low-yielding dairy cows. b. The feeding of supplements to animals grazing at high intensity. H. T. Bryant et al. ( 1 96 1) found a greater response in milk yield to supplementary feeding of dairy cows grazing the “bottom” horizon of a sward than by “top” grazers, and Conway ( 1 968b) has shown a similar effect with beef cattle, which gave a response to carbohydrate supplements at a stocking rate of 2.5 animals per acre, but no response at 1.75 animals per acre. Both these experiments illustrate a possible deficiency in the current description of grazing in terms of stocking rate, for the relevant parameters are grazing pressure - the number of animals per

96

W. F. RAYMOND

unit of available forage-and the amount of forage per unit area. A fixed stocking rate within a treatment can conceal wide variations in grazing pressure at different seasons of the year, as well as on the successive days of grazing on a given paddock; yet it is only at those times when grazing pressure is high that a significant response to supplementary feeding can be expected. It is not surprising then that some experiments in which supplements have been fed throughout have shown only marginal responses (e.g., Castle et al., 1968); supplementary feeding must surely be a tactical management tool, to be used only when grazing pressure is so high that forage intake-and so level of animal productionis likely to be significantly reduced. c. Forages can be harvested and fed, either directly or after some form of conservation. In theory this should allow a high efficiency of utilization of forage ( B , Fig. 9) by high-producing animals ( P ) ; yet many conservation systems involve high losses ( A ) and the conserved products are only of low nutrient potential. Raymond ( 1 968) has suggested that this reflects preconceived attitudes to forage conservation, rather than inherent deficiencies in conserved forages per se. The information outlined in earlier sections indicates current developments in forage conservation that may allow a large proportion of the forage grown to be fed to productive animals. The solution of this problem, by creating a storage buffer between the growing and the utilization of forages, will present new perspectives for the agronomist. For grazing requires a succession of growths of forage in phase with current animal requirements, and the agronomist must also aim to make this forage as nearly as possible a complete food for the animals to be fed. By contrast, forages for conservation can be grown at times when they make best use of the environment. The range of species grown is not restricted by the requirement that they must be suitable for grazing; and no single conserved forage need be a complete feed, so that several forages and other feeds, each outstanding in a particular nutritional feature, can be fed in appropriate combination so as to provide optimal nutrition for different classes of ruminant livestock. Thus systems of forage Conservation and nutritional interactions between feeds must be major subjects for future investigations. The agronomist who must produce forage as a complete feed for grazing is involved in compromise - between yield and digestibility, between yield and protein content, between annual and seasonal production. In contrast, forages for conservation, each presenting a much less demanding specification, may offer a more rewarding field for agronomic advance. Figure 9 thus provides a basic, though simple, biological framework

T H E NUTRITIVE VALUE OF FORAGE CROPS

97

within which the problem of land use, presented in the Introduction, can be considered. Most of the world’s ruminant livestock are now at a low level of productivity because they are underfed; the need for the future will be to integrate these new nutritional concepts into practical systems of exploiting, with high-producing animals, the increasing levels of forage production made possible by agronomic research. REFERENCES Abou Akkada, A. R., and El Sayed Osman, H. 1967.J. Agr. Sci. 69,25-3 1. Ademosum, A . A., Baurngardt, B. R., and Scholl, J . M. 1968. J. Animal Sci. 27,818-823. Agricultural Research Council. 1965. “The Nutrient Requirements of Farm Livestock, No. 2 Ruminants.” H.M. Stationery Office, London. Alderman, G., and Jones. D. I. H. 1967.J. Sci. FoodAgr. 18,197- 199. Aldrich, D. T. A., and Dent, J. W. 1967.J. Nail. 1nsr.Agr. Botany 11,104-1 13. Alexander, R. H., and McGowan, M. 1966. J . Brit. Grassland Soc. 21, 140-147. Allaway, W. H., and Hodgson, J. F. 1964. J . Animal Sci. 23,27 1-277. Allcroft, R., and Burns, K. N. 1969. New Zealand Vet. J . In press. Andrews, E. D. 1966. New ZealandJ. Agr. Res. 9,829-838. Andrews, 0. N., and Hoveland, C. S. 1965.Agron. J . 57,3 15-3 16. Annison, E. F., and Lewis, D. 1959. “Metabolism in the Rumen,” Chapter 3. Methuen, London. Annison, E. F., Chalmers, M. 1.. Marshall, S. B. M., and Synge, R. L. M. 1954. J . Agr. Sci. 44,270-273. apGriffith. G.. and Walters, R. J . K. 1966. J. Agr. Sci. 67,8149. Armstrong, D. G. 1960. Proc. 8th Intern. Grassland Congr., Reading, Engl., 1960, pp. 485-489. Armstrong, D. G. 1964.J. Agr. Sci. 62,399-4 16. Armstrong, D. G., and Blaxter, K. L. 1957. Brif. J . Nutr. 11,247-272.392-408, and 4 I3425. Armstrong:, D. G., Alexander, R. H., and McGowan, M. 1964a. Proc. Nurr. Soc. f E n g l . Scot.) 23, xxvi. Armstrong, D. G., Blaxter, K. L., and Waite, R. 1964b.J.Agr. Sci. 62,417-424. Arnold, G. W., and Dudzinski, M. L. 1963.J.Agr. Sci. 61,33-43. Arnold, G . W., and Dudzinski, M. L. 1967a.J. Agr. Sci. 68,2 13-2 19. Arnold, C. W., and Dudzinski, M. L. 1967b.Australian J. Agr. Res. 18,657-666. Bailey, P. H., Hughes. M., and McDonald, A. N . C . 1957. J. Brif. Grassland Soc. 12,157165. Bailey, R. W. 1964. New ZealandJ. Agr. Res. 7,496-507. Baker, G., Jones, L. H. P., and Wardrop, 1. D. 1961. Nature 189,682-683. Balch, C . C., and Campling, R. C. 1962. Nu!r. Abstr. Rev. 32,669-686. Balch, C. C.. Balch, D. A., Bartlett, S., Cox, C. P., and Rowland, S. J. 1952.J. Dairy Res. 19,39-50. Barnes, R. F. 1967.J. Animal Sci. 26, I 120- 1 130. Barrett, J. F., George, J. M., and Lamond, D. R. 1965.AustralianJ.Agr. Res. 16,189-200. Barrow, N . J., and Lambourne, L. J. 1962.Ausrralian J. Agr. Res. 13,46 1-47 I . Bath. D. L., Weir, W. C., and Torell, D. T. 1956.J. Animal Sci. 15,i 166- I 17 1.

98

W. F. RAYMOND

Bath, 1. H., and Rook, J. A. F. 1963. J . Agr. Sci. 61,341-348. Bath, I . H., and Rook, J. A. F. 1965. J . Agr. Sci. 64,67-75. Baumgardt, 9. R. 1967. J . Animal Sci. 26,1186- I 194. Baumgardt, 9. R., Cason, J. L., and Taylor, M. W. 1962. J . Dairy Sci. 45, 59-61 and 62-68. Beardsley, D. W. 1964.J. AnimalSci. 23,239-245. Beck, A. B. 1964. Australian J . Agr. Res. 15,223-230. Bedell, E. 1967. J . AnimalSci. 26,933-934. Bell, M. C. 1964. Proc. 6th Intern. Congr. Nutr., Edinburgh, 1963, p. 185. Livingstone, Edinburgh and London. Bennett, W. D. 1963. New Zealand J . Agr. Res. 6,3 10-3 13. Bergman, E. N., Reid, R. S., Murray, M. G., Brockway, J. M., and Whitelaw, F. G. 1965. Biochem. J. 97,53-58. Bickoff, E. M. 1968. Commonwealth Bur. Past. Crops, Rev. Ser. 111968. Bland, B. F., and Dent, J. W. 1962.J. Brit. Grassland Soc. 17,157- 158. Blaser, R. E. 1964. J . Animal Sci. 23,246-253. Blaxter, K. L. 1962. “The Energy Metabolism of Ruminants.” Hutchinson, London (2nd Impression, 1967). Blaxter, K. L. 1964. Proc. Nutr. SOC. (Engl.Scot.) 23,62-7 1. Blaxter, K. L., and Graham, N. McC. 1956. J . Agr. Sci. 47,207-21 7. Blaxter, K. L., and Wainman, F. W. 1964. J . Agr. Sci. 63,113-128. Blaxter, K . L., and Wilson, R. S. 1962.Animal Prod. 4,35 1-358. Blaxter, K. L., and Wilson, R. S. 1963.Animal Prod. 5,27-42. Blaxter, K. L., Wainman, F. W., and Wilson, R. S. 196 1. Animal Prod. 3,5 1-62. Bosman, M. S. M. 1967. Jaarb. Inst. Biol. Scheik. Onderz. LandbGewass. pp. 97-100. Bredon, R. M., and Torell, D. T. 1967. J . Range Management 20,3 17-320. Briggs, M. H. 1967. “Urea as a Protein Supplement.” Pergamon Press, Oxford. Briggs, P. K., Hogan, J. P., and Reid, R. L. 1957.Australian 1.Agr. Res. 8,674-690. Brook, P . J. 1963. New ZealandJ. Agr. Res. 6,147-228. Brouwer, E.. van Es, A. J. H., and Nijkamp, H. J. 1961. Proc. 2nd Intern. Symp. Energy Merab., Wageningen, 1961, pp. 153-163. Brown, D. 1954. Commonwealth Bur. Past. Field Crops, Bull. 42. Brown, L. D. 1966. J . Dairy Sci. 49,223-230. Brown, R. H., Blaser, R. E., and Fontenot, J. P. 1968. J . Animal Sci. 27,562-567. Brown, S. M. 1960. Res. Exptl. Record Min. Agr. Northern Ireland 10,9-19. Bryant, A. M., and Ulyatt, M. J. 1965. New ZealandJ. Agr. Res. 8,109-1 17. Bryant, H. T., Blaser, R. E., Hammes, R. C., and Hardison, W. A. 1961. J . Dairy Sci. 44, 1733- I74 I. Buchman, D. T., and Hemken, R. W. 1964. J . Dairy Sci. 47,861-864. Bull, L. B., and Dick, A. T. 1959. J . Pnthol. Bacteriol. 78,483-502. Burdick, D., and Sullivan, J. T. 1963.J . Animal Sci. 22,444-447. Butler, G . W., and Glenday, A. C. 1962. Australian J . Biol. Sci. 15,183-187. Butterworth, M . H. 1963. J.Agr. Sci. 60,341-346. Butterworth, M. H. 1967. Nutr. A bstr. Rev. 37,349-368. Cameron; G . D. T. 1965. Can. J . AnimalSci. 45,79-83. Cameron, G. D. T. 1967. Can. J . Animal Sci. 47,123- 125. Campbell, C. M., Eng, K. S., Nelson, A. B., and Pope, L. S. 1968. J . Animal Sci. 27,23 I233. Campbell, J. R., Howe, W. M., Martz, F. A., and Merilan, C. P. 1963. J . Dairy Sci. 46,I3 1134.

THE NUTRITIVE VALUE OF FORAGE CROPS

99

Campling, R. C. 1964. Proc. Nutr. Soc. (Engl. Scot.) 2 3 , 8 0 4 8 . Campling, R. C., and Murdoch, J . C. 1966. J. Dairy Res. 33,l - I I . Campling, R. C., Freer, M., and Balch, C. C. 1962. Brir. J. Nurr. 16, 115-124. Campling, R. C., Freer, M., and Balch, C. C. 1963. Brif.J. Nutr. 17,263-272. Castle, M. E., Drysdale, A. D., and Watson, J. N . 1962.5. Dairy Res. 29,199-206. Castle, M. E., Drysdale, A. D.,andWatson, J. N. 1968.1.Brit.GrassZandSoc. 23,137-143. Chalupa, W. 1968. J. Animal Sci. 27,207-2 19. Chalupa, W., and McCullough, M. E. 1967. J. Animal Sci. 26,l 135- 1 143. Chenost, M. 1966. Proc. Intern. Grassland Congr., Helsinki, 1966, pp. 406-41 I . Christian, K. R., Williams, V. J., and Carter, A. H. 1965. N e w Zealand J . Agr. Res. 8, 530-541. Clancy, M. J., and Wilson, R. K. 1966. Proc. 10th Intern. Grassland Congr., Helsinki, 1966, pp. 445-453. Clapperton, J. L. 1964. Brit. J. Nutr. 18,39-46 and 47-54. Clark, K. W.,and Mott,G. 0. 1960. Can.J. PlantSci. 40,123-129. Clarke, E. G . C., and Clarke, M. L. 1967. i n “Garner’s Veterinary Toxicology,” 3rd ed. Tindall & Cassall, London. Clarke, R. T. J. 1965. Proc. 25th Con$ N e w ZealandSoc. AnimalProd. pp. 96-105. Clifford, A. J . . and Tillman, A. D. 1966.5. Animal Sci. 25,899. C1ine.T. R.,Garrigus, U. S.,and Hatfield, E. E. 1966.J.AnimalSci. 25,734-739. Commonwealth Agricultural Bureaux. 196 I . Commonwealth Bur. Past. Crops, Bull. 45. Conrad, H. R. 1966. J. AnimalSci. 25,227-235. Conrad, H. R., and Hibbs, J. W. 1966. Dairy Sci. Deprl. Ser. 10, No. 4, Ohio Agr. Res. Devel. Ctr., Wooster, Ohio. Conrad, H. R., Pratt, A. D., and Hibbs, J. W. 1964. J . Dairy Sci. 47,54-62. Conway, A. 1968a. IrishJ. Agr. Res. 7,93-104. Conway, A. I968b. Irish J . Agr. Res. 7,105- 120. Cook, C. W., Thorne, J. L., Blake, J. T., and Edlefsen, J. 1958. J . Animal Sci. 17, 189-192. Cook, C. W., Blake, J. T., and Call, J. W. 1963. J. Animal Sci. 22,579-581. Cook, C. W., Harris, L. E., and Young, M. C. 1967.5. AnimalSci. 26, I 169-1 174. Coop, 1. E.,andClark,V. R. 1960. New ZealandJ. Agr. Res. 3,922-933. Coop, 1. E., and Hill, M. K. 1962.J. Agr. Sci. 58, 187-199. Cooper, J. P., Tilley, J. M. A., Raymond, W. F., and Terry, R. A. 1962. Nature 195, 1276- 1277. Corbett, J. L., Greenhalgh, J. F. D., McDonald, I., and Florence, E. 1960. Brit. J . Nutr. 14, 289-299. Cowlishaw, S . J., and Alder, F. E. 1963.5. Brit. GrasslandSoc. 18,328-333. Cramer, D. A,, Barton, R. A., Shorland, F. B., and Czochanska, Z. 1967. J . Agr. Sci. 69, 367-373. Crampton, E. W., and Maynard, L. A. 1938.1. Nutr. 15,383-395. Crampton, E. W.. Donefer, E., and Lloyd, L. E. 1960. J. Animal Sci. 19, 538-544 and 545-552. da Silva, J. F. C., and Gomide, J. A. 1967. Rev. Ceres 13,255-275. Davey, B. G.,and Mitchell,R. L. 1968.5. Sci. FoodAgr. 19,425-431. Davies, W. E., ap Griffith, G., and Ellington, A. 1966.5. Agr. Sci. 66,35 1-357. Davis, C. L. 1967. J . Dairy Sci. 50, I62 1- 1625. Davis, C. L., Brown, R. E., and Beitz, D. C. 1964. J . Dairy Sci. 47,12 17-1 2 19. Davis, L. E., Marten, G. C., and Jordan, R. M. 1967. Agron. J . 59,544-546. de Groot, T., and Keuning, J. A. 1967. Stikstof5,289-293 and 32 1-327. Dehority, B. A., and Johnson, R. R. 1963.5. AnimalSci. 22,222-225 and 1 134.

100

W. F. RAYMOND

Dehority, B. A,, Johnson, R. R.,and Conrad, H. R. 1962. J . Dairy Sci. 45,508-5 12. Dehority, B. A,, Scott, H. W., and Johnson, R. R. 1968.J . Dairy Sci. 51,567-572. Deinum, B., van Es, A. J. H., and Van Soest, P. J. 1968. Netherlands J . Agr. Sci. 16, 217-223. de Loose, R.,and Baert, L. 1966. Plant Soil 24,343-350. Demarquilly, C. 1966a. Fourrages 26,12-33. Demarquilly, C. 1966b.Ann. Zootech. 15, 147-169. Demarquilly, C., and Journet, M. 1967. Ann. Zootech. 16,123-150. Dent, J. W. 1963. J . Brit. GrasslandSoc. 18,181-188. Dent, J. W., and Aldrich, D. T. A. 1968. J . Brit. Grassland Soc. 23,13- 19. Dick, A. T. 1954. Australian J . Agr. Res. 5 , 5 I 1-544. Dickson, G . R. 1959. J . Brit. Grassland Soc. 14, 172- 176. Dijkstra, N. D. 1966. Proc. 10th Intern. Grassland Congr., Helsinki, 1966, pp. 393-397. Donefer, E.. Crampton, E. W., and Lloyd, L. E. 1966. Proc. 10th Intern. Grassland Congr., Helsinki, 1966, pp. 442-445. Dougall, H. W.. and Bogdan, A, V. 1966. East African Agr. Forestry J . 32,45-49. Drew, K. R. 1966. Proc. N e w Zealand Soc. Animal Prod. 26,52. Dunham, J . R., Ward, G . . Bassette, R., and Reddy, M. C. 1968. J . Dairy Sci. 51,44-46. Egan, A. R., and Moir, R. J. 1965.Australian J . Agr. Res. 16,437-449 and 45 1-472. Ehlig, C. F., Allaway, W. H., Cary, E. E., and Kubota, J. 1968. Agron. J . 60,43-47. Ekern, A., Blaxter, K. L.,and Sawers, D. 1965. Brit. J . Nutr. 19,417-434. Ellis, W. C. 1968. J . Agr. FoodChem. 16,220-224. El-Shazly, K., Dehority, B. A., and Johnson, R. R. 1961.J . Animal Sci. 20,268-273. Ely, R. A., and Moore, L. A. 1955. J . Animal Sci. 14,7 18-724. Engels, E. A. N., and Van der Merwe, F. J. 1967. South African J . Agr. Sci. 10,983-995. Essig, H. W. 1968. J . AnimalSci. 27,730-738. Evans, P. S. 1964. N e w Zealand 1. Agr. Res. 7, 508-5 13. Faichney, G. J . 1968a. Australian J . Agr. Res. 19,791-802. Faichney, G . J. 1968b.Australian J . Agr. Res. 19,803-81 I . Faichney, G. J. 1 9 6 8 ~Australian . J. Agr. Res. 19,8 13-8 19. Flatt, W. P. 1966.J . Dairy Sci. 49,230-237. Fleming, G. A., and Murphy, W. E. 1968. J. Brit. GrasslandSoc. 23, 173- 185. Fontenot, J. P., and Blaser, R. E. 1965. J . Animal Sci. 24,1202- 1208. Francis, C. M., and Millington, A. J. 1965. Australian J . Agr. Res. 16,557-573. Freer, M., and Campling, R. C. 1963. Brit. J . Nutr. 17,7948. Gaillard. 8. D. E. 1958. J.Sci. FoodAgr. 9, 170-177and 346-353. Gaillard, 9. D. E. 1962. J . Agr. Sci. 59,369-373. Gaillard, 9. D. E. 1966. Netherlands J . Agr. Sci. 14,2 15-223. Gaillard, B. D. E., and Nijkamp, H . J. 1968. Netherlands J . Agr. Sci. 16,2 1-24. Gallagher, C. H., Koch, J. H., and Hoffman, H. 1966.Australian Vet. J . 42,279. Garrett, W. N., Meyer, J. H . , Lofgreen, G. P., and Dobie, J. 8.1961. J . Animal Sci. 20, 833-838. Gladstones, J. S. 1962. Australian J . Exptl. Agr. Animal Husbandry 2,2 13-220. Gonzalez Gonzalez, V., and Lambourne, L. J. 1967. Rev. Nutr. Animal, Madrid 4,34-40. Gordon, C. H., Derbyshire, J. C., Wiseman, H. G . , and Jacobson, W. C. 1964. J . Dairy S C ~47,987-992. . Graham, N. McC. 1964.AustralianJ.Agr. Res. 15,100-1 12. Graham, N . McC. 1966. Proc. Australian Soc. Animal Prod. 6,364.

THE NUTRITIVE VALUE OF FORAGE CROPS

101

Graham, N . McC. 1967.Australian J . Agr. Res. 18,467-485. Green, J. 0. 1966. Ann. Rept. Grassland Res. Inst., Hurley, 1965, pp. 15-1 8. G reenhalgh, J. F. D., and Corbett, J . L. 1960. J . Agr. Sci. 55,37 1-376. Greenhalgh, J . F. D., and Reid, G . W. 1967. Nature 214,744. Greenhalgh, J . F. D., and Reid, G . W. 1968. Animal Prod. 10,23 I . Greenhalgh, J. F. D., Corbett, J . L., and McDonald, 1. 1960.5.Agr. Sci. 55,377-386. Greenhatgh, J. F. D., Reid, G . W., and Aitken, J . N. 1967. J . Agr. Sci. 69,2 17-223. Greenhill. W. L., Couchman, J. F., and de Freitas. J . I96 I . J . Sci. Food Agr. 12,293-297. Grieve, C. M., and Osbourn, D. F. 1965.J.Agr.Sci. 65,411-417. Griffiths, T. W. 1967.5.Agr. Sci. 69,355-366. Grimes, R. C. 1967. J . Agr. Sci. 69,33-4 I . Grimes, R. C., Watkin. B. R., and May, P. F. 1965. J . Brit. Grussland Soc. 20, 168-173. Grunes, D. L., and Stout, P. R. To be published. Hamilton, J . W., and Beath. 0. A . 1964. Agr. Food Chem. 12,37 1-374. Hanson, C. H. 1965. U S .Dept. Agr., Tech. Bull. 1333. Harkess, R. D. 1963. J . Brit. Grassland Soc. 18,62-68. Harlan, J . R. 1958.5. RangeManagement 11,140-147. Harris, C. E. 1963. J . Brit. Grassland Soc. 18, 189. Harris. C. E.. and Raymond. W. F. 1963. J . Brit. Grussland Soc. 18,204-2 12. Harris. C . E., Raymond. W. F., and Wilson, R. F. 1966. Proc. loth Intern. Grassland Congr., Helsinki, 1966, pp. 564-568. Hawkins. G. E. 1959. J . Animal Sci. 18,763-769. Head, M. J . I96 I . In “Digestive Physiology and Nutritionbf the Ruminant” (D. Lewis, ed.), Chapter 23. Butterworth, London and Washington, D.C. Healy, W. B., and Ludwig, T. C. 1965. N e w ZealandJ. Agr. Res. 8,737-752. Heaney, D. P., and Pigden, W. J. 1963. J . Animal Sci. 22,956-960. Heaney, D. P., Pigden. W. J., Minson, D. J., and Pritchard, G . I. 1963. J . Animal Sci. 22, 752-757. Heaney, D. P., Pigden, W. J., and Pritchard, G . 1. 1966.J.AnimalSci. 25,142-149. Heaney, D. P., Pritchard, 0.I., and Pigden, W. J. 1968.5.AnimalSci. 27,159-164. Hedin, L., and Duval, E. 1966. Fourrages 28,3- 15. Hennaux. L., Compere, R.,Frankinet, M., and Fabry. J. 1967. Bull. Rech. Agron. G e m bloux 2,473-506. Hercus, J. M. 1963. N e w ZealandJ. Agr. Res. 6 , 8 3 4 9 . Hight. C . K., Sinelair, D. P., and Lancaster. R. J. 1968. N e w ZealandJ. Agr. R e s . 11,286302. Hinders, R. G.,and Owen, F. G . 1968.5. Dairy Sci. 51,1253-1257. Hiridoglou, M., Dermine, P., Hamilton, H. A , , and Troelsen, J. E. 1966. Can. J . Plant Sci. 46, I0 I - 106. Hodge, R. W.. and Doyle, J. J . 1967. Australian J . Exptl. Agr. Animal Husbandry 7, 141143. Hodgson, J. 1968. J . Agr. Sci. 70,47-5 I . Hoehne, 0. E., Clanton, D. C., and Streeter, C. L. 1967.5.Animal Sci. 26,628-63 I . Hogan, J . P. 1965. AustralianJ. Agr. Res. 16,855-862. Holmes, J . C., and Lang, R. W. 1963. Animal Prod. 5.17-26. Hornb,T. 1953.ActaAgr. Scand. 3,1-32. Honkanen, E., and Moisio, T . 1963.Acta Chem. Scand. 17,588. Hopson, J. D.. Johnson, R. R., and Dehority, B. A . 1963.J . Animal Sci. 22,448-453.

102

W. F. RAYMOND

Hungate, R. E. 1966. “The Rumen and its Microbes.” Academic Press, New York. Huston, J. E., and Ellis, W. C. 1968. J . Agr. Food Chem. 16,225-230. Hutton, E. M., and Coote, J. N. 1966. J . Australian Ins?.Agr. Sci. 32,139- 140. Hutton, J . B. 1963.Proc. NewZealandSoc. Animalprod. 23,39-51. Hutton, J. B., and Jury, K. E. 1964. New Zealand J . Agr. Res. 7,583-595. Ingalls, J. R., Thomas, J. W., Benne, E. J., and Tesar, M. 1965. J. Animal Sci. 24, 11591164. Jarrige, R. 1961. Ann. Biol. Animale, Biochim., Biophys. 1,163-205. Jarrige, R., and Minson, D. J. 1964. Ann. Zootech. 13,118- 150. Johns, A. T. 1959. New Zealand J . Agr. 99,2-9. Johnson, R. R., and McClure, K. E. 1968. J . Animal Sci. 27,535-540. Johnston, T. D. 1967. Euphytica 16,365-369. Jones, J. R., and Hogue, D. E. 1963. J . AnimalSci. 22,881-885. Jones, L. H. P., and Handreck, K. A. 1967.Advan.Agron. 19,107-149. Jones, U . I. 1959. Rep?. Welsh Plant Breeding Sta., 1956-1 958, p. 52. Julen, G., and Lager, A. 1966. Proc. 10th Infern. Grassland Congr., Helsinki, 1966, pp. 652-657. Kamstra, L. D., and Jahn, J. R. 1966. J . Range Management 19,196-200. Kane, E. A., and Moore, L. A. 1959. J . Dairy Sci. 42,936. Kane, E. A., Jacobson, W. C., and Moore, L. A. 1950.J . Nutr. 41,583-596. Kane, E. A., Jacobson, W. C., and Moore, L. A. 1952.J . Nutr. 47,263-273. Karn, J. F., Johnson, R. R., and Dehority, B. A. 1967.J . Animal Sci. 26,38 1-384. Karr, M. R., Garrigus, U. S., Hatfield, E. E., and Norton, H. W. 1965. J . Animal Sci. 24, 459-468. Kay, M., Andrews, R. P., MacLeod, N. A., and Walker,T. 1968.AnimalProd. 10,171-175. Kemp, A. 1960. Netherlands J . Agr. Sci. 8,28 1-304. Kemp, A. 1964. Netherlands J . Agr. Sci. 12,263-280. Kendall, W. A. 1966. Crop Sci. 6,487-489. Kennedy, G . S. 1968. Australian J . Biol. Sci. 21,529-538. Kivimae, A. 1959.Acta Agr. Scand. Suppl. 5. Knight, R., and Yates, N. G. 1968. Australian J . Agr. Res. 19,373-380. Laby, R. H., and Weenink, R. 0. 1966. New ZealandJ. Agr. Res. 9,839-850. Lacey, J. 1968. J . Gen. Microbiol. 51,173-177. Lambourne, L. J., and Reardon, T. F. 1962. Nature 196,96 I , Lambourne, L. J., and Reardon, T. F. 1963. Australian J . Agr. Res. 14,239-256,257-27 1 , and 272-293. Lancaster, R. J. 1949. New ZealandJ. Sci. Technol. A31,3 1-38. Langlands,J. P. 1967a.AnimalProd. 9,167-175. Langlands, J. P. l967b. Animal Prod. 9,325-33 I . Langlands, J. P., and Corbett, J. L. 1964. J . Agr. Sci. 63,305-3 10. Langlands, J. P., Corbett, J. L., McDonald, I., and Reid, G. W. 1963. Animal Prod. 5, 11-16. Langlands, J. P., George, J. M., and Lynch, J. J. 1967. Australian J . Expfl. Agr. Animal Husbandry 7,325-328. Large, R. V., and Spedding, C. R. W. 1967. J . Agr. Sci. 67,41-52. Lefevre, C. F., and Kamstra, L. D. 1960.J . Animal Sci. 19,867-872. Leigh, J. H., and Mulham, W. E. 1966. Australian J . Exptl. Agr. Animal Husbandry 6, 460-467. Leitch, Isabella. 1969. In preparation.

T H E NUTRITIVE V A L U E OF FORAGE CROPS

103

Leng, R. A., Corbett. J. L., and Brett, D. J. 1968. Brit. J . Nutr. 22,57-68. Lindahl, I . L., Kates. K. C., Turner, J. H., Enzie, F. D., and Whitmore, G . E. 1963. J. Parasitol. 49,209-2 17. Lindner, H. R. 1967.AustralianJ. Agr. Res. 18,305-333. Linton, J. H., Goplen, B. P., Bell, J. M., and Jaques, L. B. 1964. Can. J . Animal Sci. 44, 76-86. Lombard, P. E., and van Schalkwyk, A. 1963. South African J . Agr. Sci. 6, 205-21 I. Loper, G . M. 1968. Crop Sci. 8, 104-106. Lowe, C . C., Reid, J. T., and Meredith, W. R. 1962. Agron. Abstr. p. 94. Lusk, J. W., Browning, C. B., and Miles, J . T. 1962. J. Dairy Sci. 45,69-73. McArthur, J. M.. and Miltimore, J. E. 1966. Proc. 10th Intern. Grassland Congr., Helsinki, 1966, pp. 5 18-524. McCarrick, R. B. 1966. Proc. 10th Intern. Grassland Congr., Helsinki, 1966, pp. 575-580. McCarrick, R. B., Keane, E., and Tobin, J. 1965. Irish J. Agr. Res. 4 , l 15-1 23. McCroskey, J. E., Pope, L. S., Stephens, D . F., and Waller, G . 1961.J.AnimalSci. 20,4245. McCullough, M. E. 1966. Proc. 10th Intern. Grassland Congr., Helsinki, 1966, pp. 58 1584. McCullough, M. E., and Smart, W. W. G . 1968. J. Dairy Sci. 51,385-391. McDonald, 1. W. 1962. Commonwealth Bur. Past. Crops, Bull. 46,43. McDonald, I . W. 1968. Nutr. Abstr. Rev. 38,38 1-400. McDonald, P.. MacPherson, H. T., and Watt, J. A. 1963. J. Brit. Grassland Soc. 18,230231. McGuire, R. L., Bradley, N. W., and Little, C. 0. 1966. J . Animal Sci. 25,185- I9 I . Mcllroy, R. J . 1967. HerbageAbstr. 37,79-87. MacLusky, D. S. 1960.J. Brit. Grassland Soc. 15, I 8 I - 188. McManus, W. R., Arnold, G . W., and Hamilton, F. J . 1962.Australian Vet. J . 38,275-28 I . McManus, W. R., Dudzinski, M. L., and Arnold, G . W. 1967. J. Agr. Sci. 69, 263-270. McMeekan. C . P. 1956. Proc. 7th Intern. Grassland Congr., Palmerston, N e w Zealand, 1956, pp. 146-155. McMeekan, C. P., and Walshe, M. J. 1963. J. Agr. Sci. 61,147- 163. Macpherson, A.,and Hemingway, R. G . 1968.5. Sci. FoodAgr. 19,53-56. Marten, G . C., and Donker, J. D. 1964.J. Dairy Sci. 47,87 1-874. Marten, G. C., and Donker, J . D. 1966. Proc. 10th Intern. Grassland Congr., Helsinki, 1966, pp. 359-363. Mellin, T. N., Poulton, B. R., and Anderson, M. J. 1962. J. Animal Sci. 21,123- 126. Melville, J . 1960. Proc. 8th Intern. Grassland Congr., Reading, Engl., 1960, pp. 35-40. Melvin,J. F. 1963.J.Sci. FoodAgr. 14,281-283. Miles, D. G., and Walters, R. J. K. 1966.J . Agr. Soc. Univ. Coll. Wales 47,4-9. Milford, R. 1967.Australian J. Exptl. Agr. Animal Husbandry I, 540-545. Mi1ler.T. 6. 1961. HerbageAbstr. 31,81-85and 163-167. Minson, D. J. 1963.5. Brit. GrasslandSoc. 18,39-44. Minson, D. J . 1966.5. Brit. GrasslandSoc. 21,123-126. Minson, D. J . 1967. Brit. J . Nutr. 21,587-597. Minson, D. J . , and Kemp, C. D. I96 I . J . Brit. Grassland Soc. 16,76-79. Minson, D. J., and Milford, R. 1967. Australian J . Exptl. Agr. Animal Husbandry 7,5465 5 I. Minson, D. J., Raymond, W. F., and Harris. C . E. 1960. J. Brit. Grassland SOC. 15, 174180.

104

W. F. RAYMOND

Minson, D. J., Harris, C. E., Raymond, W. F., and Milford, R. 1964. J . Brit. Grassland SOC.19,298-305. Mitchell, R. L., and Reith, J. W. S. 1966. J. Sci. FoodAgr. 17,437-440. Moe, P. W., Reid, J. T., and Tyrrell, H. F. 1965. J. Dairy Sci. 48,1053- I06 1. Moir, R. J., and Bray, A. C. 1967. Sulphur Inst. J . 4,M- 18. Moore, L. A. 1964. J . Animal Sci. 23,230-238. Moore, L. A., Irvin, H. M., and Shaw, J. C. 1953. J . Dairy Sci. 36,93-97. Moore, L. A., Thomas, J. W., and Sykes, J. F. 1960. Proc. 8th Intern. Grassland Congr., Reading, Engl., 1960, pp. 701-704. Moore, R. M., and Hutchings, R. J. 1967. Australian J . Exptl. Agr. Animal Husbandry 7, 17-21. Morley, F. W. H., and Francis, C. M. 1968. Australian J . Agr. Res. 19,I5-26. Morrison, F. B. 1957. “Feeds and Feeding,” 22nd ed. Morrison, Ithaca, New York. Mott, G. 0. 1960. Proc. 8th Intern. Grassland Congr., Reading, Engl., 1960, pp. 606-61 1. Moule, G . R., Braden, A. W. H., and Lamond, D. R. 1963. Animal Breeding Absrr. 31, 139-157. Mowat, D. N. 1969. “Symposium: Forage Quality Evaluation.” Am. SOC. Agron. Spec. Publ. 13,85-95. Mowat, D. N., Fulkerson, R. S., Tossell, W. E., and Winch, J. E. 1965. Can. J . Plant Sci. 4 5 3 2 1-33 I . Naga, M. M. A., and El-Shady, K. 1963. J . Agr. Sci. 61,73-79. National Academy of Sciences. 1966. Natl. Acad. Sci. -Natl. Res. Council, Publ. 1349. Neumark, H. 1962. Nature 195,626. Neurnark, H., and Tadmor, A. 1968. J . Agr. Sci. 71,267-270. Newton, J . E., and Betts, J. E. 1968. J . Agr. Sci. 70,77-82. Noller, C . H., Prestes, P. J., Rhykerd, C. L., Rumsey, T. S., and Burns, J. C. 1966. Proc. 10th Intern. Grassland Congr., Helsinki, 1966, pp. 429-434. Norman, A. G. 1935. J . Agr. Sci. 25,529-540. O’Donovan, P. B., Barnes, R. F., Plumlee, M. P., Mott, G. 0.. and Packett, L. V. 1967. J . Animal Sci. 26,1144- I 152. Oldfield, J. E., Fox, C. W., Bahn, A. V., Bickoff, E. M., and Kohler, G. 0. 1966. J. Animal Sci. 25,167-174. Oltjen, R.R., and Bond, J. 1967. J.Animal Sci. 26,225. Oltjen, R. R., Slyter, L. L., Kozak, A. S., and Williams, E. E. 1968. J . Nutr. 94, 193-202. Oram, L. R. N., and Williams, J. D. 1967. Nature 213,946-947. Orth, A., and von Kaufmann, W. 1966. Z. Tierphysiol., Tierernaehr. Futtermittelk. 21, 327. Osbourn, D. F. 1967. Brit. GrasslandSoc. Occ. Symp. 3,20-28. Osbourn, D. F., Thomson, D. J., and Terry, R. A. 1966. Proc. 10th Intern. Grassland Congr., Helsinki, 1966, pp. 363-366. O’Shea, J., and Maguire, M. F. 1967. Irish J . Agr. Res. 6,127-129. O’Shea, J . , and Wilson, R. K. 1965. Irish J . Agr. Res. 4,235-237. O’Sullivan, M. 1968. J . Sci. FoodAgr. 19,12-15. Owen, J . B., and Ingleton, J. W. 1963. J.Agr. Sci. 61,267-274. Paladines, 0. L., Reid, J. T., van Niekerk, B. D. H., and Bensadoun, A. 1964. J . Nutr. 83, 49-59. Paulson, G. D., Broderick, G. A , , Baumann, C. A., and Pope, A. L. 1968. J . Animal Sci. 27,195-202. Pearce, G . R., Vercoe, J . E., and Freer, M. 1962.J. Agr. Sci. 59,397-402.

THE NUTRITIVE VALUE O F FORAGE CROPS

105

Perez, C. B., Warner, R. G., and Loosli, J. K. 1967.J . Animal Sci. 26,8 10-8 19. Peterson, P. J., and Spedding, D. J. 1963. N e w ZealandJ. Agr. Res. 6, 13-23, Pigden, W. J., Pritchard, G. I., Winter, K. A., and Logan, V. S. 1961. J . Animal Sci. 20, 796-80 I . Playne, M. J., and McDonald, P. 1966.J . Sci.FoodAgr. 17,264-268. Plice, M. J. 1952.J . Range Management 5,69-75. Pope, G. S. 1954. DairySci. Abstr. 16,333-356. Pope, G. S., McNaughton, M. J., and Jones, H. E. H. 1959. J . Dairy Res. 26,196-202. Presthegge, K. 1959. 93 Berern. Insr. Husdryrernaering og FBringslaere Agr. Coll. Norway, Vollebekk. Pritchard, G. I., Folkins, L. P., and Pigden, W. J. 1963. Can.J . Plant Sci. 43,79-87. Quicke, G. V., Bentley, 0. G., Scott, H. W., and Moxon, A. L. 1959. J . Animal Sci. 18, 275-287. Rae, A. L., Brougham, R. W., Glenday, A. C., and Butler, G. W. 1963. J . Agr. Sci. 61, 187- 190. Raymond, W. F. 1966a. I n “The Growth of Cereals and Grasses’’ (F. L. Milthorpe, ed.), pp. 259-27 I . Butterworth, London and Washington, D.C. Raymond, W. F. 1966b. In “Recent Advances in Animal Nutrition” (J. T. Abrams. ed.), pp. 81-1 18. Churchill, London. Raymond, W. F. 1968. J . Roy. Agr. Soc. Engl. 129,85-105. Raymond, W. F. I969a. “Symposium: Forage Economics.” A m . Soc. Agron. Spec. Publ. 13,25-35. Raymond, W. F. 1969b. “Symposium: Forage Quality Evaluation.” A m . SOC.Agron. Spec. Pub/. 13,47-63. Raymond, W. F., and Minson, D. J. 1955. J . Brit. Grassland Soc. 10, 282-296. Raymond, W. F., and Spedding, C. R. W. 1965. Proc. Fertiliser Soc. 88, 5-34. Raymond, W. F., and Terry, R. A. 1966. Outlook Agr. 5,60-68. Raymond, W. F., Harris, C. E., and Harker, V. G . 1953. J . Brit. Grassland Soc. 4, 3 I5320. Raymond, W. F., Kemp, C. D., Kemp, A. W., and Harris, C. E. 1954. J . Brir. Grassland SOC.9,69-82. Raymond, W. F., Minson, D. J., and Harris, C. E. 1956. Proc. 7th Intern. Grassland Congr., Palmerston, New Zealand, 1956, pp. 123- 133. Reid, C. S. W. 1960. Proc. 8th Intern. Grassland Congr., Reading, Engl., 1960, pp. 668671. Reid, J. T. 1962. “Pasture and Range Research Techniques,” p. 43. Am. Soc. Agron., Comstock, Ithaca, New York. Reid, J. T.. Woolfolk, P. G., Hardison, W. A., Martin, C. M., Brundage, A. L., and Kaufmann, R. W. 1952.J . Nutr. 46,255-269. Reid, J. T., Kennedy, W. K., Turk, K. L., Slack, S. T., Trimberger, G . W.. and Murphy, R. P. 1959.J . Dairy Sci. 42,567-57 1. Reid, R. L., and Jung, G . A. 1965.J . Animal Sci. 24,6 15-625. Reid. R. l.., and Jung, G . A. 1966. Proc. 9th Intern. Grassland Congr., Suo Paulo, 1966, p. 863-869. Reid, R. L., H0gan.J. P.,and Briggs, P. K. 1957.AustralianJ.Agr. Res. 8,691-710. Reid, R. L., Clark, B., Welch, J. A., Jung. G. A., and Shelton, D. C. 1960. J . Animal Sci. 19,13 12. Reid, R. L., Jung, G . A., and Murray, S. 1964.J . Animal Sci. 23,700-7 10. Reith, J. W. S. 1968.J . Agr. Sci. 70,39-45.

106

W. F. RAYMOND

Richards, C. R., Weaver, H. S., and Connolly, J. 0. 1958. J. Dairy Sci. 41,956-962. Ridley, J. R., Lesperance, A. L., Jensen, E. H., and Bohman, V. R. 1963. J. Dairy Sci. 46, 128-130. Riewe, M . E., Smith, J. C., Jones, J. H., and Holt, E. C. 1963. Agron. J. 55,367-369. Robards, G. E., and Wilson, A. D. 1967. Australian J. Exptl. Agr. Animal Husbandry 7, 22-24. Roe, R., and Mottershead, B. E. 1962. Nature 193,255. Rogers, H. H. 1967. Brit. GrasslandSoc. Occ. Symp. 3,66-75. Rogers, H. H., and Whitmore, E. T. 1966. J. Brit. GrasslandSoc. 21,150-152. Ronning, M., and Dobie, J. B. 1962. J. Dairy Sci. 45,969-97 1. Rook,J. A. F., and Balch, C. C. 1961. Brit. J. Nutr. 15,361-369. Rook, J. A. F., Balch, C. C., Campling, R. C., and Fisher, L. J. 1963. Brit. J. Nutr. 17, 399-406. Rosenfeld, I., and Beath, 0. A. 1964. “Selenium. Geobotany, Biochemistry, Toxicity and Nutrition.” Academic Press, New York. Rossiter, R. C., and Beck, A. B. 1966. Australian J. Agr. Res. 17,29-37. Satter, L. D., and Baumgardt, B. R. 1962. J.Animal Sci. 21,897-900. Schaadt, H . , Johnson, R. R., and McClure, K. E. 1966.J.Animal Sci. 25,73-77. Schillinger, J. A., and Elliott, F. C. 1966. Quart. Bull. Michigan Agr. Expt. Sta. 48, 570579. Schneider, B. H. 1952. “Feeds of the World.” West Virginia Univ., Morgantown, West Virginia. Schulz, E., and Oslage, H. J. 1967. Wirtschaftseigene Futter Part 4, pp. 298-306. Shelton, D. C., and Reid, R. L. 1960. Proc. 8th Intern. Grassland Congr., Reading, Engl., 1960, pp. 524-528. Shepperson, G . 1960. Proc. 8th Intern. Grassland Congr., Reading, Engl., 1960, pp. 704708. Shumard, R. F., Bolin, D. W., and Eveleth, D. F. 1957. A m . J. Vet. Res. 18,330-337. Smart, W. W. G., Bell, T. A., and Stanley, N . W. 196 1. J. Dairy Sci. 44,1945- 1946. Smith,C. A. 1962. J. Agr. Sci. 58,173-178. Sonneveld, A. 1965. Veerelt Zuivelber. 8,409-420. Spedding, C. R. W. 1952. Vet. Record 64,813-815. Spedding, C. R. W. 1954. J. Comp. Pathol. Therap. 64,5-14. Spedding, C. R. W. 1965. “Sheep Production and Grazing Management.” Bailliere, London. Spedding, C. R. W., Large, R. V., Brown, T. H., and Wilson, 1. A. N. 1965. J. Agr. Sci. 64,283-290. Stewart, D. R. M. 1967.J. Appl. Ecol. 4,83-1 I I . Stout, P. R., Brownell, J., and Burau, R. G . 1967. Agron. J. 59,2 1-24. Sullivan, J. T. 1962. Agron. J. 54,5 1 1-5 15. Swift, R. W., Bratzler, J. W., James, W. H., Tillman, A. D., and Meek, D. C. 1948. J. Animal Sci. 7,475-485. Synge, R. L. M. 1952. Brit. J. Nutr. 6,100-104. Tayler, J. C., and Deriaz, R. E. 1963. J. Brit. GrasslandSoc. 18,29-38. Tayler, J. C., and Rudman, J. E. 1966. Proc. 9th Intern. Grassland Congr., Sclo Paulo, 1966, pp. 1639- 1644. Terry, R. A., and Tilley, J. M. A. 1962. Exptl. Grassland Res. Inst., Hurley 14,58. Terry, R. A., and Tilley, J. M. A. 1963. Exptl. Grassland Res. Inst., Hurley 15,47. Terry, R. A., and Tilley, J. M. A. 1964a.J . Brit. GrasslandSoc. 19,363-372. Terry, R. A., and Tilley, J. M. A. 1964b. Exptl. GrasslandRes. Inst., Hurley 16,67.

THE NUTRITIVE VALUE O F FORAGE CROPS

107

Theron, E. P., and Booysen, P. de V. 1966. Proc. Grassland SOC.South Africa 1, I 1 1- 120. Thomas, G . D., Bartley, E. E., Pfost, H. B., and Meyer, R. M. 1968.J. Dairy Sci. 51,869875. Thomas, J. W., Moore, L. A., Okamoto, M., and Sykes, J . F. 1961. J . Dairy Sci. 44, 147 I 1483. Thomson, D . J . 1964. Exptl. Grassland Res. Inst., Hurley 16,67-68. Thornton. R. F., and Yates, N . G. 1968. AustralianJ. Agr. Res. 19,665-672. Tilley, J. M. A., and Terry, R. A. 1963.J . Brit. Grassland Soc. 18, 104- I 1 1. Tilley, J. M. A., Terry, R. A., Deriaz, R. E., and Outen, G . E. 1964. Exprl. Grassland Res. Inst., Hurley 16,64-67. Topps, J. H., Reed, W. D. C., and Elliott, R. C. 1965. J . Agr. Sci. 64,387-402. Torell, D. T. 1954. J . Animal Sci. 13,878-884. Troelsen, J. E., and Bigsby, F. W. 1964.J . Animal Sci. 23, I 139- 1 142. Troelsen, J. E., and Hanel, D. J. 1966. Can. J . Animal Sci. 46,149- 156. Ulyatt, M. J. 1965. N e w Zeal7nd.f.Agr. Res. 8,397-408. Underwood, E. J. 1962. “Trace Elements in Human and Animal Nutrition,” 2nd ed. Academic Press, New York. Underwood, E. J. 1966. “Mineral Nutrition of Livestock.” F.A.O. and Commonwealth Agr. Bureaux. Farnham Royal, Bucks, England. University ofGlasgow. 1965. British Patent 1,030,101. Van der Menve, F. J., and Perold, I . S. 1967. J . South African Vet. Med. Assoc: 38,355363. Van Dyne, G. M., and Heady, H. F. 1965. Hilgardia 36,465-492. Van Dyne, G . M., and Torell, D. T. 1964. J . Range Management 17,7-19. Van Soest, P. J . 1964.J . Animal Sci. 23,838-845. Van Soest, P. J. I965a. J . Dairy Sci. 48,8 IS. Van Soest, P. J. 1965b. J . Assoc. Ofic. Agr. Chemists 48,785-790. Van Soest, P. J. I965c. J . Animal Sci. 24,834-843. Van Soest, P. J. 1967.J . Animal Sci. 26,119- 128. Van Soest. P. J. 1968. I n “International Encyclopaedia of Food and Nutrition,” Chapter 2. Pergamon Press, Oxford. Van Soest. P. J., and Jones, L. H. P. 1968.J . Dairy Sci. 51, 1644- 1649. Van Soest, P. J . , and Moore, L. A. 1966. Proc. 9th Intern. Grassland Congr., Siio Paulo, 1966, pp. 783-789. Virtanen, A. I. 1966. Science 153, 1603-1614. von Kaufmann, W., and Rohr, K. 1967. Z . Tierphysiol. Tierernaehr. Futtermirtelk. 22,l-8. Walters, R. J. K., ap Griffith, G., Hughes, R., and Jones, D. 1. H. 1967. J . Brit. Grassland SOC.22,112-1 16. Warner, A. C. I . 1964. Nutr. Abstr. Rev. 34,339-352. Warner, R. G., Battacharya, A. N.,and Davison, K. L. 1966. Cornell Nutr. Con$, pp. 24-29. Watson. S. J., and Nash, M. J . 1960. “The Conservation of Grass and Forage Crops.” Oliver & Boyd, Edinburgh and London. Wedin, W. F., Carlson. 1. T., and Vetter, R. L. 1966. Proc. 10th Intern. Grassland Congr., Helsinki, 1966, pp. 424-428. Weller, R. A., Gray, F. V., Pilgrim, A. F., and Jones, G. B. 1967. Australian J . Agr. Res. 18,107-1 18. Weston, R. H. 1967.AustralianJ. Agr. Res. 18,983- 1002. Weston, R. H., and Hogan, J . P. 1967.AustralianJ. Agr. Res. 18,789-801. Weston, R. H.. and Hogan, J. P. 1968.AusfralianJ.Agr.Res. 19,419-432.

108

W. F. RAYMOND

Whitehead, D. C. 1966. Commonwealth Bur. Past. Crops, Mimeo Publ. I / 1966, pp. 1-83. Wieringa, G. W., Schukking, S., Kapelle, D., and D e Haan, Sj. 1961. Netherlands J . Agr. Sci. 9,2 10-2 16. Wilkins, R. J . 1966. Thesis, University of New England, Australia. Wind, J., Deijs, W. B., and Kemp, A . 1966. Jaarb. Inst. Biol. Scheik. Onderz. LandbGewass., pp. 9 1- 100. Woods, A . E., and Aurand, L. W. 1963. J . Dairy Sci. 46,656-659. Worker, N . A., and Carrillo, B. J . 1967. Nature 215,72-74. Wright, M. J., and Davison, K. L. 1964. Advan. Agron. 16,197-241. Wright, P. L., Pope, A . L., and Phillips, P. H. 1963. J . Animal Sci. 22,586-591. Yates, N. G . , and Tookey, H. L. 1965. Australian J . Chem. 18,53-60.

POTENTIALLY ARABLE SOILS OF THE WORLD A N D CRITICAL MEASURES FOR THEIR USE Charles E. Kellogg and Arnold C. Orvedal Soil Survey, Soil Conservation Service United Stater Deportment of Agriculture, Washington, D.C.

Introduction ........................................................................ 11. The Principle of Interactions in Soil Use ............................................... A. In the Field ..................................................................... i d Farm .................. ......................... C . In the Community ............................................... D. Country Planning .......................................................................... 111. Higher Production from Existing Arable Soils ....................... A. Soil Surveys ................................................... B. Transfer of Science and Technology ............................................... C. Incentives .................................................................................... D. Water Control ...... ........................

I.

F. Soil Erosion Control ...................................................... G . Soil Blowing IV. New Potentially ........................... ................................................................. A. Soils of the References .........................................................................................

I.

Puge 109 I12 I12 I16 116

117

122 123 124 129 130 135 I38 I40 I40 141 I53 I69

Introduction

Soils are the basic resource for farm production. Thus it seems reasonable to begin with the soils of the world, and how they respond to management, to assess what farm production is biologically and physically possible with at least moderately efficient use. In any such assessment we must recognize the vital need for those combinations of private and public institutions that meet the requirements for efficient sustained production and that fit the genius of the many peoples in the 95 less developed countries.* But the advanced countries can help. If all farmers in the world were as *For “less developed countries” there are many near synonyms, including “undeveloped countries” and “underdeveloped countries.” Those moving forward are called “newly developing countries” and “newly emerging nations.” Then, too, many “advanced countries” have “less developed areas” within them.

I09

110

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

efficient as those, say, in Holland, it would be easy. But most are not. Building the necessary institutions for effective technical assistance requires time and patience. As pointed out by Broekmeijer (1966), we are thinking in terms of centuries, not simply decades. Much has been written recently about the world food problem, sometimes garnished with such catch phrases as “the race between food and population” or “the oncoming famine.” Yet it is highly doubtful that food situations in the 95 less developed countries are as bad now as they were earlier. Most people in these countries formerly took famine, flood, and pestilence for granted as man’s inevitable tragedies. Since 1940, world communications have improved. Many people from advanced countries have traveled in the less developed countries. In turn, people from the less developed countries have visited the advanced countries. They now know that floods, famine, and pestilence can be avoided. Certainly most have food problems. With the wide introduction of medicine and sanitation, many of the less developed countries have experienced great decreases in death rates with corresponding increases in population. Bogue ( 1 967) points out that current evidence suggests that the birth rate may be on the way down, especially in those countries where individuals have some social security besides many sons. Poor people in these countries hope for better standards of living, including many of the conveniences developed by the advanced countries only within this century. Yet they need to be reminded that each of the advanced countries without exception got its start toward economic development from a reasonably efficient agriculture. Although people in the less developed countries want some of the new things, the great majority of them do not want to become Englishmen, Frenchmen, or Americans. Instead they want to develop their own culture in their own way and fit the new ways into their culture. As pointed out by Martin and Knapp (19671, economic planning must be done differently in different countries. Based on our experience, it seems clear that one cannot discuss even briefly suggestions for increased food production in one or more countries without also relating the suggestions to general economic development, better use of all resources, equitable distribution of the national product, improved living standards, and local cultures. Since much that has been learned about science and its application to farming and to the other sectors of agriculture has come from western Europe and the United States, difficult problems of communications arise in giving people the benefit of this knowledge in terms they can use with understanding. Change needs to start where the people are, with their

POTENTIALLY ARABLE SOILS OF THE WORLD

11 I

skills, their culture, and their hopes and aspirations. Not much in the way of advanced techniques can be transferred, even if it were desirable to do so, from farmers in the advanced countries to farmers in the less developed countries. Nor are any two of these countries facing exactly the same problems. Nor do they have the same potentials. Nor are they at similar stages of cultural and economic development. In some of the less developed countries there are religious and social ideas that conflict with science or with the technology based on science. With great patience and tact ideas can be changed slowly without damage to the cultural systems, but certainly few other changes should be even attempted at the start. The modern systems of agriculture contrast with most of those in the less developed countries. In the advanced countries the proportion of agricultural workers on farms has been declining at an accelerating rate for many years. In the United States, for example, perhaps only about one-fifth of the labor force in agriculture works on farms. The majority of people working in agriculture live in towns and cities. They make machines, chemicals, and other supplies for farmers; process, transport, and market farm products; and perform research, educational, financial, and other services for both the farming and industrial sectors of agriculture. This highly complex system we call “modern agriculture” requires a host of highly skilled workers, technicians, managers, and scientists to keep it going. Such systems evolved in educated societies along with scientific research and its application, educational institutes, and dynamic governmental systems. Then too, the agricultural system evolved in close association with many other production and service activities. It should be evident that technical assistance to the less developed countries requires highly trained people with deep knowledge in those scientific disciplines essential to agriculture and enough knowledge of the broad field to understand how the several parts fit together. Those giving technical assistance must be able to diagnose problems and potentials in natural and cultural environments unlike their own. Most improvements need to be invented on the spot to fit both the local kinds of soil and the social environment. Scientists must be able to communicate with the governments and people of the countries they assist in terms of the local cultural systems. By three broad ways can the total amount of food, fiber, and other agricultural products be increased: 1. By improving the management of the many kinds of soil already being used for crops;

112

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

2. By developing and managing well potentially arable soils not now used for crops; and 3. By reducing preventable losses of farm products between the fields of the cultivators and the kitchens of the consumers. Such losses are less in modern agricultural systems. They result from diseases, insects, rodents, and poor harvesting methods on farms, and from inadequate processing, storage, and transport. Improvements along any of these lines or some combination of them depend on parallel development of the industrial and service sectors of agriculture, including adequate transport, advisory services, and the like. II.

The Principle of Interactions in Soil Use

The most important single principle for guiding improved farming and agricultural development is the principle of interactions: Each practice or each program within a system affects all other components of the system, so that a proper combination gives a far greater result than the sum of the several components considered singly (Kellogg, 1962a). The principle of interactions has been cryptically defined as a peculiar kind of mathematics from which the whole is much greater than the sum of its parts. In technical assistance for agricultural development, this principle guides the process at several levels: (1) reliable interpretations of the responses of a soil to alternative management systems from field study of its characteristics; (2) working out the most effective combination of practices for a system of managing each specific local kind of soil by people of known skills and with the facilities to give optimum production on a sustained basis; (3) fitting together a management scheme for a whole farm or group of holdings having unlike kinds of soil and different enterprises; (4)an effective balance of services needed for an existing or potential farming community to be successful; ( 5 ) organizing common services for two or more resource uses in the economic development of a large area or region; and ( 6 ) a balance of emphasis between the many aspects of a national plan or system for economic and agricultural progress, considering the skills of citizens, patterns of resources, sources of revenue, transport facilities, and the like. A. IN THE FIELD Each farmer makes his arable soil from either a natural soil or an old arable soil. He may change it only a little or he may change it drastically by reshaping the surface for water control, by adding fertilizers to correct plant nutrient deficiencies, by adding other materials to correct acidity or alkalinity or to improve the structure of the soil, or by tilling in depth.

POTENTIALLY ARABLE SOILS OF THE WORLD

I I3

Within his skills and with the facilities available to him, he tries to make the soil as nearly ideal as practicable for the crops most likely to grow well. Only a tiny fraction of the soils of the world produce well under simple management that includes only clearing, plowing, seeding, and harvesting. For most crops the farmer needs a stable soil that water, air, and roots can penetrate deeply in order to have a large volume of soil for storing water and plant nutrients and for releasing them to the roots of the plants. For each kind of soil, and the associated climate and length of day, standards can be designed for the combinations of characteristics that an arable soil should have for the most rewarding results. No soilmanagement practice or system is universally beneficial. A well-proved system for one kind of soil may be harmful or perhaps wasteful on another kind, even in an adjoining field. Because the effect of any single soil characteristic depends on the host of other characteristics in the combination that makes a soil of a certain kind, we cannot make useful generalizations in terms of a single soil characteristic or even of only three or four. Few useful statements about soil management can be made about salty soils, hilly soils, clayey soils, red soils, or the like. Because of differences in clay minerals, some kinds of gently sloping clayey soils are slowly permeable to water and have a high erosion hazard under cultivation, whereas other kinds of clayey soils are so permeable to water that crops can be grown without significant erosion even on soils with 40 percent slope. (Fig. I). Every productive hectare of arable soil in the world has at least four basic conditions, appropriately related to one another and to the local kind of soil. A balanced combination may have been arrived at over years of trial-and-error or through planning with the aid of science. If adequate practices for any one of the conditions are neglected, little can be expected from the other practices. Some unique kinds of soil respond well to only a very few combinations of practices; other kinds of soil have a wider range of alternatives. The four basic conditions are: 1. The arable soil needs a balanced supply of plant nutrients for acceptable yields. For the best results, at least some chemical fertilizers are nearly always necessary to supplement the supply of one or more of the many essential plant nutrients within the soils plus that contributed by compost, manures, crop residues, and green manures. Thus the kinds and amounts of fertilizers to use depend on the soil, other practices, and expected costs and returns. Fertilizers are especially critical in areas of high rainfall and leaching, all the way from the cool temperate regions to the humid Tropics (lgnatieff and Page, 1958).

114

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

Recently some have overemphasized the critical role of fertilizers in tropical countries in relation to the other inputs. That is, high yields on a sustained basis do require considerable fertilizer, but fertilizer alone is

FIG. 1 . A strongly sloping, yet only slightly erodible excellent clayey soil used for potatoes near Ootacamund, India. The trees are Eucalyptus. Pruned branches are used for manufacturing a special oil. (These are examples of some of the best upland soils of the Tropics -typical Reddish-Brown Latosols, likely Ustropepts in the new system of classification.)

not enough. Fertilizers are economically effective only i f ( a ) the correct kinds and amounts are used for the local kind of soil and the crops to be grown and ( b ) the practices needed to meet the other requirements for good harvests are adopted at the same time. 2. A productive arable soil has adequate moisture in the rooting zone when the plants need it, without the waterlogging that deprives the roots of air. A good arable soil holds abundant water and lets the excess drain on through. Few kinds of soil have a combination of characteristics and associated climate for ideal water control. Some need shaping, diversions, or terraces (bunds) for runoff control; some need irrigation; others need drainage; and many need a combination of two or more of these practices. Recently much emphasis, perhaps overemphasis, has been given to irrigation. We must recall that by far the greatest amount of water used by the crop plants of the world is that which falls as rain on the surface of

POTENTIALLY ARABLE SOILS OF THE WORLD

115

fields. Practices to conserve this water are normally cheaper than those for irrigation. Nor will irrigation be successful unless a good job is done of shaping the surface and of avoiding waterlogging and accumulation of salts. This is not to say that irrigation cannot be highly rewarding if a good enough job is done to eliminate moisture deficiency as a limiting factor, and if the other essential practices are also carried out. In a great many regions with erratic rainfall, more can be had for the labor and capital with improved runoff control. 3 . Success requires a variety of crop (or varieties of several crops in rotation or in a mixed culture) adapted to the environment and with the genetic potential to respond to the most favorable arable soil that it is practicable to develop. Occasionally but not commonly, an improved variety by itself can lead to significant increases in yield, at least for a while. That is, a new crop variety placed in a primitive system may result in a 2-fold increase; yet if combined with adequate fertilizers and water control, it might lead to a 10-fold increase. Neither crop variety nor fertilizer is any kind of panacea by itself. 4 . Arrangements are needed to protect plants (and animals for which the crops may be grown) against diseases, insects, weeds, predatory animals, and other hazards. The problem ofweeds, for example, commonly becomes very difficult after clearing for continuous culture as compared to primitive shifting cultivation, especially in the Tropics. With 3 to 4 years in crops, let us say, and 12 years of rejuvenating forest trees, the biological cycles of insects, diseases, and weeds are interrupted, but if the soils are used in open fields under permanent cultivation, much more attention must be given to pest control. 5 . Supplementary practices. In addition to these basic and universal sets of practices, some potentially arable soils can be used with heavy investments for protection against the sea, mountain torrents, and floods. From smaller investments some can be used with protection against high winds. We emphasize again: Each vital practice in the combination depends on and supports the others. We can never be sure how much of a crop yield can be attributed to fertilizer, how much to a new variety, or how much to water control. Many interactions in farming are a bit obscure, and unless we find them, we can be seriously misled. For example, people have interviewed cultivators using fertilizer. Those cultivators may get higher yields than others. But it is wrong to attribute all the differences in yield to fertilizer or to any other single item under study. Cultivators who take the trouble to use fertilizers usually carry out other

116

CHARLES

E.

KELLOGG A N D ARNOLD C. ORVEDAL

improved practices that help to make them effective. The same is true of those who use a new variety or some other improved practice. Besides the “pessimistic” or “scare” statements about the hopelessness of the world food problem - “no new land,” “erosion,” “leached tropical soils,” and so on, we have had overemphasis on fertilizer alone, on irrigation alone, and recently on the new “miracle” varieties. Clearly fertilizers and high-yielding varieties are essential parts of combinations of practices suited to the local kinds of soil. So are water control and plant protection. A failure of any one can make the others ineffective. Perhaps it is only human to look for a single answer or slogan to promote increased farm production. The hopeless search results partly because people unfamiliar with farming may think that agriculture is very simple. Nothing could be further from the truth (United Nations, 1963).

B. ON A FARMOR POTENTIAL FARM The principle of interactions also guides the working out of a good system for an operating unit made up of fields of different sizes and kinds of soil. The management system best for any one field of a farm depends partly on the alternatives for use of other fields, especially in mixed farming either with crops and livestock or with industrial crops and food crops. Soils that do not response economically to systems with cultivated crops may support permanent or semipermanent forage, trees, palms, or shrubs. Industrial or forage crops can be grown on some of these soils, and food crops on those easily made arable. Commonly, sloping soils unstable in cultivation can be changed gradually from natural forest cover to rubber trees, coffee, cacao, or oil palms without exposing more than narrow contour strips to the sun and rain, and these only for a short time. Small areas too wet for normal crops can gradually be built up by bedding with soil from the outside or from ditches between beds for crops that can be sold or othenvide used within the farming enterprise (Fig. 2). Thus, the measure of success is not simply what can be done on the most productive soils but the production that can be had from the whole land tract that makes a farm. C. IN THE COMMUNITY The same principle of interactions needs to be observed in planning essential services for a village or rural community so that cultivators can have available the materials and services for effective combinations of soil management practices. If cultivators farm better, they must have

POTENTIALLY ARABLE SOILS OF THE WORLD

117

incentives; this means local markets with fair grading and prices. They need appropriate tenure arrangements and opportunities for basic education, for learning technical skills, for credit, for industrial supplies, and

FIG.2. A cultivator and his family in Karala State, India, are bringing in soil material from the outside to build a well-drained soil for coconut palms that give higher returns than rain-fed rice. As the palms grow, the intervening areas are filled.

for technical assistance on special practices. As pointed out earlier, many of these services are more economically supplied if costs are shared with other economic enterprises. Each of these services influences the effectiveness of the others and which of the alternative farming systems is best for each cultivator.

D. COUNTRY PLANNING Effective programs of technical assistance require some kind of country planning. To an observer from abroad it may seem that many things could and should be done in most newly developing countries. But the local people cannot do everything themselves and also get the economy going. Usually the total number of both professional people and skilled technicians is small. If too many kinds of projects are attempted with subsidies or grants from various sources, efforts are likely to get spread

118

CHARLES E. KELLOCG AND ARNOLD C. ORVEDAL

too thin for the small corps of trained people to be effective. Since every country gets its start toward economic growth out of agriculture, emphasis needs to be placeti on farming and on the industries that serve farming (Bauer, 1965).

1 . Combined Resource Use Planning for economic growth should begin by taking advantage of the combinations of resources and skills already available. The plan should include early efforts for agricultural research, including soil surveys. Appraisals will be needed of the resources of water, forests, and minerals and of the skills of the population. In most of the less developed countries, a large part of the population has skill in farming. Even though the farming may not be near the potential, we should not assume that cultivators lack all basic skills for farming or that they are unable to take on new ones fairly rapidly with good demonstrations and the incentives of good prices and of local markets with fair grading. The goal in planning for economic growth needs to be income to the people-money in their pockets - rather than “monuments” -capital goods per se (Johnson, 1967). Then too, local communities need to be able to arrange necessary training, credit, fertilizers and other chemicals, and adapted machines. 2. Education Most writers have given great emphasis to education, perhaps with some overemphasis on primary and university education and not nearly enough emphasis on intermediate technical training in the local language. Agricultural development and associated economic development require large numbers of skilled people- machinists, electricians, truck drivers, and the whole list. Then too, many have encouraged bright young men and women in the less developed countries to go to either Europe or North America for their university education. This has not seemed to work well, partly because of an unhappy social gap between educated people able to attend college and the mass of cultivators and laboring people. Also a young man or woman taken out of his own society and placed for 4 to 7 years in Europe or North America becomes adjusted to a quite different culture. When he finishes, he really belongs to neither the new culture nor the old. Back in his home country with the prestige of a Western degree, the social gap between the graduate and his fellow citizens commonly is further widened. The mass of simple people tend not to accept him, nor can he easily communicate with them. Comparable expenditures to help the less developed countries develop

POTENTIALLY ARABLE SOILS OF THE WORLD

I19

their own colleges and training institutes for their young people are much more productive to the economy. After graduates have completed their training and demonstrated their ability to analyze problems and work well in their trades or professions with the mass of people in their country, tours and additional training in other countries are useful. As soon as possible, such local training through the Ph.D. degree should be arranged and should come before much study abroad. The work of the Rockefeller Foundation and the Government of India has demonstrated what can be done. In some countries scholarship standards and salary opportunities in agriculture tend to be lower than in law, medicine, and literature. For economic development additional emphasis needs to be given to training scientists and engineers in agriculture, including both advisors in farming and professional workers in related public and private services, because agriculture is so vital for getting a start toward economic development. 3 . Research

The total effort in tropical agriculture must be greatly stepped up in both basic and adaptive research (Moseman, 1964). Many of the existing research institutes are working along narrow lines, even on a single function or crop. The great need is for general research stations with good staff in all lines, including the social sciences, soil surveys with accompanying laboratory work, and related field research. For results to be effective in developing farming systems and broader agricultural systems, people of many skills must work together and with cultivators to take full advantage of the principle of interactions. In research programming, special emphasis is needed on field and on combined field and laboratory studies. Our knowledge of how tropical plants and soils get their nitrogen is inadequate, for example. Many observers have reported good supplies that cannot be accounted for by currently known mechanisms. If they were known, no doubt much greater advantage could be taken of them. Similarly, more study of improving supplies of calcium in many tropical soils is needed. This is not to suggest that nothing can be done now. A great deal can be done through soil surveys, exploratory studies of existing farming, and field testing of promising new combinations of practices, including fertilizers, based on principles already known from research and experience elsewhere in the world on similar kinds of soil.

4 . Industrial Sectors Generally speaking, agriculture is most efficient in countries with a

120

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

growing industry. Yet the fact that capital accumulation to get industry started comes mainly out of agriculture has been overlooked in some of the newly developing countries. The lack of adequate rural statistics has been another reason for neglecting the importance of agriculture as a source of saving. Steel mills, great hydroelectric plants, huge dams for irrigation, and the like are easy to count. Yet if most cultivators in a newly developing country developed one-fifth of an acre or more of greatly improved arable soils with runoff controlled, bought two or three bags of the correct fertilizer for their soils, got improved seed, and bought a duster for insecticides, the total investment would be large. If the advisory services were reasonably sound, this investment would lead at once to increased farm production, income, and opportunities for savings. And with more money, the many cultivators buy more, which further boosts the economy. Thus, we should argue that high emphasis be placed on investments for transport, marketing, and the other industries that serve agriculture or are developed with it. If these service industries can be properly organized, they also give income to people and return savings that can be used later for still more basic industries. Then as industry develops and furnishes more materials for farming, countries experience the usual substitution of off-farm labor for direct farm labor.

5 . Public Versus Private Sector Every country has some balance of economic activity in both the public and the private sectors. Too far one way or the other may lead to inefficiency. Certainly the balance varies from country to country. Private firms competing under a fair set of rules, honestly enforced by government, can lead to efficiency. Public agencies with leadership devoted to the public interest can most effectively handle many kinds of essential services. Yet either an entrenched private or public monopoly, without fair competition and with indifferent management resorting to nepotism, can lead to inefficiency and high cost for goods and services essential to agricultural and economic development. Decisions on these matters can be very important to farmers. If fertilizers and machines, for example, are too costly or of low quality, farmers are severely handicapped in developing balanced systems of soil management for high yields and efficiency. Of course, what they can pay depends partly on prices they receive. In calculating benefits from fertilizer one should think not only of cost per kilogram of fertilizer in currency but also of cost in terms of kilograms of the farmers’ production, be it rice or oil palm nuts. The incentive to use fertilizer on rice, for ex-

POTENTIALLY ARABLE SOILS OF THE WORLD

121

ample, is quite different if one cultivator can get 1 kg. of fertilizer for 1 kg. of rice whereas another cultivator must give 6 kg. of rice for the same fertilizer. If the methods for processing and marketing farm products are unreasonably costly, farm prices are likely to be too low to give farmers the incentives for full production. The same principles apply to costs for credit and other essential services. Thus, in country planning, decisions about how essential services are organized and operated have a great deal to do with expansion of farming and its efficiency and with progress toward a modern agricultural system for the country. 6. Trade Relations

Trade relations have a strong influence on farm development. Especially in the early stages, many newly developing countries must depend a great deal on farm products, such as tropical woods, tropical fruits, palm nuts, rubber, fiber crops, and other industrial crops for a part of their foreign exchange. Thus, we cannot think of effective agricultural development in terms of food alone. In fact, some have gone to such extremes in this direction as to gauge agricultural development only by international trade in grain -as if this were equivalent to food. Actually people vary widely in the amount of grain they eat, from none to a high percentage of their diet. Further, much trade in grain and other foods takes place in rural villages where none of it is counted. Much better statistics on all important kinds of food and industrial crops, at least in sample villages, are urgently needed for reasonable estimates of actual food production and consumption, and of that available for trade in the general market. Trade relations between the newly developing countries and the advanaced countries are highly important to both. The developing countries need foreign exchange in order to buy equipment, instruments, and other supplies necessary to get started toward improved agricultural systems. In fact, many of the advanced countries need to search more carefully for ways to expand their imports from the newly developing countries if they are to move ahead in an orderly manner. Ultimately efforts toward increasing trade are bound to benefit both groups of countries. These trade relationships need continual study because of changing requirements and technological substitutions (Johnson, 1967).

7 . Finance and Taxation Several of the less developed countries have traditional finance and taxing arrangements that seriously handicap farming and discourage

122

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

investments in agriculture. Some have special export taxes on farm products, so that the cultivators get considerably less than the prices on the world markets. In some countries schemes may allow sales of certain products only to a government monopoly at low prices; then the products are sold at higher prices on the world market. This amounts to a tax on farming. In a few less developed countries cultivators are handicapped by being charged higher prices for local fertilizers or other materials produced within the countries than the sale prices for export. This also amounts to a special tax on farming. Such unfair taxes or low local prices for farm products and high local prices for materials and services are used in some countries to favor the city population. In the end, this sort of practice hurts everyone by holding back agricultural progress, which is the normal forerunner of general economic development. It is highly important to correct these handicaps- to recognize that cultivators need fair prices and to replace unfair taxes on farming with other sources of revenue. Where income taxes are difficult to enforce, substantial sales taxes on imported luxury items would be paid by those most able to pay, whether cultivators or urban dwellers. Similarly cultivators should not be required to pay unreasonably high rates for credit needed to make investments in machines, chemicals, improved seeds, and the other inputs essential to modern management systems. Ill. Higher Production from Existing Arable Soils

Nearly one-half of the soils of the world that have good physical and biological potential for farming are already in use. On these arable soils yields could be considerably higher and the efficiency of producing them could be greater. This is true for large areas in advanced countries, whereas in many other parts of the world yields are only one-tenth to onehalf of those had by skilled farmers elsewhere on similar kinds of soil. It commonly costs less to improve farming on arable soils in use than to develop farming on new, unused potentially arable soils. First of all, the cultivators are there and already know something about the farming of their soils. Second, at least part of the essential infrastructure-roads, markets, and the like-is in place. Yet cultivators obtaining low yields on soils responsive to management may be handicapped in several ways, e.g., by inadequacy of education, transport, markets, and access to products of the industrial sector of agriculture or by prices and taxes that discriminate against farm people. Most cultivators are willing and able to work hard to support their

POTENTIALLY ARABLE SOlLS O F THE WORLD

123

families, but returns in the market place may be so low that they have little incentive to use added inputs for production beyond family needs. As pointed out in the previous section, several practices must be combined, in proper relationship to the local kind of soil and to one another, for rewarding harvests in relation to inputs of labor, skill, and materials. Because of the great contrasts among the many thousands of kinds of soil and their associated crops, the “packages” of practices vary widely from place to place, even between one kind of soil and another within a small holding. Especially in tropical areas, more local research is needed to learn how best to combine machines, chemicals, and other products of the industrial sector of agriculture to have better farming systems for the many kinds of tropical soils. A. SOILSURVEYS Soil surveys are essential for the full use of the results of research and of experience related to soil management. For orderly use in technical assistance, these results need to be reported and summarized by specific kinds of soil. No one has yet found a reasonable alternative for assernbling knowledge about soils for application to the specific land tracts where it applies. 1 . Operational Planning

For operational planning and technical assistance the maps in such soil surveys must be in enough detail to show clearly the kinds of soil in each important field and holding. The degree of detail depends on the relative intensity of potential use, the relative value per hectare of labor and other inputs, the complexity of the pattern of unlike soils in the landscape, and their degrees of contrast. Then too, where expensive structural works, such as those for water control - runoff control, drainage, irrigation, or land shaping-are in prospect, detailed soil mapping is required for satisfactory results without waste. Soil surveys are essential for carrying out programs to consolidate fragmented holdings: programs that are essential to effective water control and field design in millions of farm villages. Thus the scale for adequate soil mapping ranges from about 1 :50,000in areas with simple soil patterns or only extensive uses to about 1 :8000 for highly intensive farming. (At and around the site of expensive structures, along planned highways, and for housing clusters, field mapping may need to be done at a large scale, say 1:1000,where the pattern of soils is very complex.) The soil map must also have plotted on it accurately the roads, streams, houses, and other features that a user can see and that help him to locate

124

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

himself precisely on the map. These can be plotted during the soil survey or taken partly from adjusted aerial photographs or from detailed planimetric surveys. Usually soil maps are published at a somewhat smaller scale than the field sheets. For example, field sheets at a scale of about 1:15,000 can be published at 1:20,000 or 1:24,000 if the symbols and mapping are properly planned in advance. Any soil map to be used in operational planning or technical assistance must be accompanied by a text that describes the soils in terms of the lay reader and in the technical language of scientists and engineers. If new practices are developed after the soil survey is published, the scientists and engineers need the technical descriptions to provide new or revised interpretations. Besides the descriptions, the principal alternative uses, management systems, and expected yields of adapted crops must be given in specific terms for each kind of soil. If relevant in the area, similar interpretive ratings are needed for native and sown grasses, trees, and any limitations of the soils for suggesting houses and other structures.

2 . General Planning For general planning of the approximate location of projected broad changes in land use, highways, and other community facilities, a smallscale soil map, generalized from the detailed one or made by reconnaissance, is highly useful. It too must include roads, villages, and other important features to help the reader locate himself. The map units on this kind of map are associations of kinds of soil that naturally occur together in the landscape, whether they are similar or highly contrasting. Each soil association is named from the two or three principal kinds of soil that occur together. Its description includes the proportion of each kind of soil that is an important part of the mapping unit. Depending on the local area, descriptions and interpretations of potential uses can be given specifically for each kind of soil, but only in broad terms for each soil association. OF SCIENCE A N D TECHNOLOGY B. TRANSFER

Compared to temperate regions, less research for soil management systems has been done in the Tropics and less experience with modern systems is available. Yet statements that little is known about tropical soils are quite untrue. A great deal is known and published. But the number of people well informed about this knowledge is small in relation to the number of cultivators in both the less developed countries and the ad-

POTENTIALLY ARABLE SOILS OF T H E WORLD

125

vanced countries. Then too, much useful material is available in many scientific institutes that has not been summarized by kinds of soil and published. Considerable research has been done on several kinds of soil in the Tropics, as in northern Queensland, Hawaii, Puerto Rico, and especially in the Congo between 1935 and 1958. During the past few years much important research has been done in other centers in both East and West Africa, in south Asia, and in parts of South and Central America, including some that is not yet published. Each year more soil surveys are being made in tropical areas. Detailed soil maps are available at a good many locations and small-scale soil maps are available for all the Tropics and subtropics at scales ranging from about 1:250,000 to 1 : 1,000,000. Unfortunately not many of these vitally needed soil maps have been published. Several schematic soil maps at scales of 1:5,000,000 are either available or becoming available, but these soil maps have far too few base data - villages, cities, roads, streams, and other features. These features are needed to enable the user to locate himself well enough to note the kinds of soil on the tracts of land in which he is interested. They are also needed to enable him to read from the map the kinds of soil on tracts of land supporting experimental stations or successful farm communities in other parts of the world. It is from these places that are on similar kinds of soil that he can obtain valuable data for planning. Even at a scale of 1:I ,000,000, the smallest practicable individual mapping unit is about 4000 hectares. At a scale of 1:50,000 one can show areas of about 10 hectares. Obviously, this is about the smallest practicable scale of soil maps for operational planning. Where the individual areas of soil are small and the kinds of soil are highly contrasting, the scale should be larger for operational planning of enterprises requiring moderate to large inputs of labor and capital. Much more research and study has been done with the reasonably wellwatered tropical soils than with the dry ones, with those of moderate to high fertility than with those of low fertility, and with those near existing transport than with those far from existing roads or navigable streams. Yet much useful material is becoming available within the Tropics and subtropics for wide use to transfer experience and knowledge between countries and continents. “The World Food Problem” (President’s Science Advisory Committee, 1967) emphasizes the urgent need for their synthesis and publication. For operational planning where the individual areas are small and the soils are highly contrasting, the scale should be larger for operational planning of enterprises requiring considerable inputs of labor and capital.

126

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

The results of research and experience gained on tropical soils are far more useful for estimating the potentials and best farming systems on similar tropical soils than results brought from temperate regions. Despite the urgent need, improved technology is not easily transferred from temperate regions to tropical ones. But the principles developed from scient$c research in the sciences basic to soil use and the scholarly methods for both basic and adaptive research can be transferred. In technical assistance, either to improve the use of existing arable soils or to suggest systems for new arable soils, one should take orderly account of the available experience in the world on comparable soils and of the skills of the cultivators or of the proposed settlers. The basic task in technical assistance is for the scientist or a small group of scientists to connect the basic principles of the sciences relevant to farming and the other sectors of agriculture to the specific natural, social, economic, and political environment where the work is to be done -in other words to invent the most efficient technologies and systems of soil use for the local kinds of soil that can be carried out by the local people. Obviously, success requires excellent scholars with deep knowledge of the basic principles and with the skills to work in contrasting social and natural environments and to communicate with the people. When we read of “practice” application or that “cultivators are slow to accept a new practice,” we can be fairly certain that the advisor is too narrow in his viewpoint. The cultivator is always dealing with a combination of practices. Here and there good starts have been made and others could be, especially with local adaptive research for guidance. Yet more basic research needs to be carried out in the newly developing countries to reach the best results. New principles from that research are bound to lead to more adaptive research for the most practicable local systems. This need is great not only in farming, but also in the service and industrial sectors of agriculture.

1. Adaptive Research Even with the best summaries of what has been learned from research and experience elsewhere, adaptive research in local areas is usually essential for modern systems. First of all, one can learn much of importance by studying existing practices, even primitive ones that cultivators have developed over the years and passed on to their children. Mere description, however, is nearly useless unless the soils are named in a standard system or are

POTENTIALLY ARABLE SOILS OF THE WORLD

127

described well enough so they can be named later. Otherwise there is no way to apply the results elsewhere. Second, one can lay out simple field trials as small rectangular plots on cultivators’ fields located on the different kinds of soil, which are named or described. These trials can have different combinations of fertilizers, water control practices, new varieties, and so on. Of course, knowledge gained from elsewhere on how similar soils respond under good management form the basis for planning such trials. In areas with generally low yields, precise yield measurement is not necessary. A plot can be treated, for example, with a moderate application of nitrogen alone, and other plots with nitrogen in various combinations with phosphorus, potassium, magnesium, calcium, or other plant nutrients expected to be in short supply from the appearance of the plants. The yields and condition of the crops should be recorded during the growing season. In some areas the appearance of the plants suggests deficiencies of one or more of the secondary nutrients, such as boron, iron, zinc, or manganese. If such plant symptoms are evident early in the trial, small strips across the plots can be sprayed with a dilute solution of the individual trace nutrients to learn whether or not plants respond to them. Some soil scientists tend to make this kind of adaptive research much more sophisticated than is necessary. Eventually it is desirable to have long-term experiments accompanied by soil tests for available nutrients as well as many detailed soil descriptions, but from many of these test plots a good deal can be learned that is highly useful in making a start quickly. A farmer in an advanced country, if convinced of a new practice in his system, will normally adopt it if he expects a 10 or 15 percent increase in yield. We must recall that this 10 or 15 percent increase is from a high base yield. The yields of many cultivators are already so low that a 10 or 15 percent increase could hardly be seen, kt alone be enough to compensate him for probable inputs. In many areas of the world, unless the yield can be at least doubled, a cultivator will not be persuaded to adopt another combination of practices. Commonly the opportunities are far greater than for mere doubling of the yield. On some old tropical soils, for example, a crop like grain sorghum will little more than germinate in the absence of phosphatic fertilizer (Fig. 3). Yet with it the crop responds to nitrogen and other inputs to give yield increases of severalfold. 2. Continued Research Continued research and examination of agricultural systems by scientists are essential, not only for further progress but also to maintain

128

CHARLES E. KELLOGG A N D ARNOLD C . ORVEDAL

the progress that has been gained. This important point is often overlooked. Let us say that a new package of soil management practices, including new varieties, fertilizers, and water control systems, is estab-

FIG.3. An experiment showing the enormous effect of phosphatic fertilizer on a tropical soil with a sandy surface and clayey B horizon at the Nungua Station near Accra in Ghana. With nitrogen but no phosphatic fertilizer, the crop in the foreground is worthless. With fairly high rates of both, good yields of grain sorghum are seen in the background. (The soil is probably a sandy Paleustult.)

lished and proves to be highly successful. Unless competent agricultural scientists are available, some new disease or insect could wipe out the crop. Modern plant breeding, for example, produces plants of high yield having rather narrow genetic bases. Where cultivators in a broad area have grown the same new variety and a new disease or insect went unnoted for a short time, the result has been catastrophic. This hazard is great with many newly introduced varieties. Similarly unnoticed changes in the soil as a result of a new management system can lead to serious difficulties unless they are noted promptly and the necessary corrective measures are taken. In other words, the more efficient and productive the soil management system, the more important it is that scientists be available to detect a serious problem before it becomes catastrophic. Most experienced scientists from the advanced countries are familiar

POTENTIALLY ARABLE SOILS OF THE WORLD

129

with the kinds of dramatic failures that can happen after the adoption of a new combination of high-yielding varieties and other practices - failures resulting from a new disease or insect, waterlogging and salinity, new gullies from poorly located terrace outlets, and so on. Most are familiar with weed problems, but some may not visualize the sudden hazard that weeds can bring about in the absence of modern chemical controls. Under the old practice of shifting cultivation in tropical areas, the rejuvenating forest fallow after 2 to 5 years of cropping interrupted the cycles of insects, diseases, and weeds. Many of the worst weeds are scarce plants within the native vegetation. Yet the seeds that fall on bare soil grow promptly. Thus, if open-field cultivation is substituted for shifting cultivation with bush fallow or mixed cultures, weeds must be watched for and avoided. Weeds can reduce the yields of rice, maize, and many other crops substantially unless practical control measures are applied promptly.

C . INCENTIVES Most cultivators in the underdeveloped countries are inclined to be cautious, even conservative, to new ways of soil management. The reasons are not hard to find. Few of them have savings or even any easy way to save. Should the new combination of practices that might be suggested to them fail, they would suffer greatly. Thus they prefer the trusty old familiar system. Actually, most cultivators work hard. Some travelers fail to understand why they see so many men resting and think them lazy. People on low diets rest when additional work adds nothing significant to their incomes. Schultz (1964) explained that, considering their skill, resources, and markets, most peasant farmers work as near to the margin of economic return for added inputs as the farmers in advanced countries. Yet Penny’s studies in Indonesia suggest that some cultivators at a subsistence level will not change when they have opportunities. He gives several specific examples (Penny, 1966). Nevertheless, Gunnar Myrdal (1 968) points out the great importance of improving the use of labor in the crowded less developed countries and notes that improved technology, rather than decreasing labor, actually increases it. Although mechanization of field operations may reduce some labor requirements in fields, the use of fertilizer, plant protection, improved storage and processing, adequate market roads, and the like, increase the opportunities for labor. In many newly developing countries where additional farm products are urgently needed for food and for economic growth, the marketing system is not conducive to increased production. Commonly, losses are

130

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

so great between the rural village and the primary market that village prices are low. Few local markets near the cultivators grade their products according to quality. Special taxes may even further reduce the price. Many cultivators lack opportunities for education and for learning new skills. Many lack access to credit at reasonable rates and to such inputs as improved seeds, fertilizers, and tools at prices they can afford. In remote areas, until many cultivators adopt improved combinations of practices, reliable vendors will not have established sales outlets for new items. Yet local demonstrations near the homes of cultivators under conditions similar to theirs are convincing to many of them. The advantage in yield must be clear. Even then some cultivators hesitate to invest. The risk may seem too great if substantial purchases are required. Much of this risk can and has been overcome by a qualified government service that guarantees them at least their traditional yields. Then too, some advisors fail to inform cultivators of the great advantage of combinations of practices to have good harvests from their soils. Although occasionally one practice alone may give a large return, advisors must be cautioned to think of several at the same time, such as fertilizers, improved seeds, water control, and pest control. All these difficulties that the cultivator sees must be fairly well overcome before he has the essential incentive to go ahead. Obviously we are speaking about the bare minima of the several essential practices for a start, not the ideal. Actually nearly all countries have the soil and the labor for ample food supplies. But these will not be forthcoming until cultivators have genuine incentives. D. WATERCONTROL Some system for water control is important on nearly all soils. The system is so much influenced by field patterns that special emphasis must be given it in early steps toward the improvement of farming. For the best results the roots of plants should not lack adequate supplies of either moisture or oxygen during the growing season. The requirements of both vary among crops. The ideal arable soil is porous enough to allow the water added by rain or by irrigation to enter the rooting zone. The body of the soil holds enough water to support plants between wettings and allows any excess to pass beneath the rooting zone. If too much water accumulates in the soil, the roots lack oxygen and crops do poorly. The excess can be taken away through a drainage system, either in open ditches or through tile.

POTENTIALLY ARABLE SOILS OF THE WORLD

131

I . Runof Control By far the greater part of the crops grown in the world are watered directly from the rain that falls. The productivity of soils that have a variable supply of rainfall and are reasonably permeable can be greatly improved by one of several kinds of terraces. These earth ridges are laid out in such a way that water moves only slowly down the slope and thus has more time to penetrate the soil. Enormous areas of arable soil in the world could benefits from this practice i f i t is well done. If poorly done, with terraces on old irregular field lines, waterlogging or erosion, or both, are likely to result in waste. Many potentially excellent soils in south Asia, for example, fall into this category. They are moderately permeable, their slopes are long and gentle, and the rains come unevenly in the monsoon climate. By making terraces at slight angles to the contours along with protected waterways for the excess water of heavy rains that cannot soak into the soil, the total amount of moisture stored for plant roots is greatly increased. Not only do the roots have more water in the rainy season but also enough water is commonly stored in the soil to support an additional crop during the drier season. Then too, wells near the lower parts of slopes have additional water for irrigating fruit and vegetable gardens. On many soils this practice for runoff control alone has more than doubled production. Then, as the water system is supplemented by fertilizers, improved varieties, and pest control, yields may be higher by severalfold than those under the current traditional systems (Fig. 4). These opportunities are widespread in many of the less developed countries low in capital. A good system of terraces requires little or no capital from outside the local community. It does require skilled technicians to lay out the system. The earth can be moved in a great variety of ways. It can be excavated and carried on the head in baskets. Bullocks or other draft animals can be used with simple equipment. Tractors and machine graders are commonly used in advanced countries where labor is scarce, but machines are not necessary. Unhappily some giving technical assistance have emphasized earth-moving machines rather than precise designs of terraces. But we repeat, the job must be done well. In areas with small irregular parcels a reasonable scheme for consolidating fragmented holdings is essential to make a start toward modern farming. The terraces must be made at the proper angles to the contour, which depends on the permeability of the soil, its relief, its water-holding capacity, and the character of the rainfall. Otherwise water breaks over the terraces to make gullies

132

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

in some places or water accumulates to waterlog the soil in other places. On many gently sloping soils under mixed cultures of crops, trees, shrubs, or palms not requiring tillage, the cover itself may slow down the

FIG.4. Water control for paddy rice in the foreground. The hills in the center background are under shifting cultivation. Brush and wood are piled in strips up-and-down the slopes, covered thinly with sods, and burned for crops of maize or potatoes. Some of the slopes have low terraces. The soils were called Yellow Latosols. (Possibly Haplustults in the new system.) They have been farmed for along time by Khasi tribal people. Near Shellong, Assam, India.

runoff enough for water to enter the soil. Yet on well-covered moderate or strong slopes unless the soils are highly permeable, at least some terraces are necessary to have the optimum amount of water enter the soil and be held in the rooting zone (Fig. 5). 2. zrrigation Irrigation is an old water-control practice and plays a great role in farming in many crowded areas today. With an effective irrigation system and a good source of water, responsive soils of the deserts and semideserts can be farmed efficiently. The productivity of many other soils can be improved by supplemental irrigation during regular dry seasons or during irregular periods of low rainfall. Such irrigation has been increasing in normally well-watered parts of the United States and Europe

POTENTIALLY ARABLE SOILS OF T H E WORLD

133

where the soils are permeable. Here too, the matter of design is critical because in these normally well-watered areas a heavy rain may fall immediately after the soil has been irrigated. If only a little additional

FIG.5 . Well-made terraces in Algeria for runoff control to promote reforestation on steeply sloping soil barely moist enough for trees. (The soil is probably in the Xeralf suborder.)

water can soak into the soil, without provisions for runoff control and drainage, roots can be waterlogged and the crops fail. For best results with irrigation, soils with irregular surfaces need to be smoothed for sprinkler irrigation or leveled for border irrigation. Such smoothing or leveling is not practical with soils that are thin over rock or have lower layers of hardpans, laterite, caliche, gravel, or other materials that cannot be made into suitable surface soils for cropping. Thus carefully made detailed soil surveys are a prerequisite for planning irrigation systems. Costly systems that include large dams and canals to provide irrigation water are seldom economic unless an excellent job is done with the layout on cultivators’ fields, including leveling of the surface, terraces or bunds where needed, and artificial drainage as required. The aim in irrigation should be a complete water system that avoids deficiencies of either water or air for pIant roots. If these deficiencies of water and air are overcome,

134

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

it becomes practicable to make large investments in fertilizers, machines, improved varieties, and pest control for high production. Unhappily, this principle of excellent design adapted to the local kind of soil has not always been appreciated. We have seen many expensive irrigation projects fail to produce much, simply because the surface of each field was not leveled. In trying to wet the “high spots” cultivators waterlogged the rest of the field. Adjacent fields not irrigated at all produced almost as well. We should like to emphasize this simple fact: Most land surfaces that appear level to the casual observer are not level enough for effective distribution of irrigation water from canals. Irrigation in arid and semiarid regions commonly raises a serious problem of salts. The water that falls as rain is relatively free of soluble salts. But water drawn from wells or held in impoundments has a certain amount of soluble salts acquired from its contact with soils and rocks. The amount may be small or large. In addition, soils of arid regions commonly have underlying layers rich in various salts that may not be obvious at the immediate surface. After penetrating deeply during irrigation, water moves upward by capillary action, bringing dissolved salts from such lower layers to the surface. Many irrigated areas that lack proper natural or artificial drainage have great patches of white, salty surface soils, commonly bare except for stunted plants. These problems are greatly aggravated by the construction of highways, railroads, and even irrigation canals that lack essential deep culverts in the natural drainageways. With land leveling to avoid excess water and with good natural drainage or artificial drainage, one can flush these salts out of the soil into the deep drainage. Of course, this is most easily done if the irrigation water itself is not salty and if the soils are permeable. In contrast to runoff control, most irrigation schemes require the use of a great deal of capital from outside the local community. 3. Drainage Artificial drainage systems have been installed in many of the most productive arable soils in the world in both humid and dry regions. The casual traveler in much of the United States and of western Europe may be unaware of the drainage systems. Permeability of the soil has a great deal to do with the design of these systems. In soils of low permeability the drains must be more closely spaced than in those of high permeability. Between the drains, open or closed, the water table is an arc, not a straight line. To be effective in preventing waterlogging, the highest part of this arc between drains must be beneath the normal rooting zone of the plants. Thus, the design of drainage systems requires careful examination

POTENTIALLY ARABLE SOILS OF THE WORLD

I35

of the soil and knowledge of the amount of water that needs to be removed by the drains, especially during periods of high rainfall or snowmelt. Many naturally wet, highly organic soils can be improved by drainage. For these the design must take account of both oxidation of organic matter and normal settling of the soil as it is prepared for cropping and as use goes forward. Many otherwise responsive organic soils have been injured for use by having drains both too far apart and too deep. Excessive oxidation of organic matter is encouraged near the drains, resulting in uneven settling. Organic soils in warm countries that are drained and cropped are bound to disappear in time. E. CONSOLIDATION OF FRAGMENTED HOLDINGS The consolidation of fragmented holdings offers one of the greatest potentials for increasing farm production in millions of villages at relatively low cost. Very little outside capital is needed, but first-class public service with technical assistance is required. In old farm villages the land now used by one cultivator’s family may consist of several small, scattered tracts. It is not uncommon for a total holding of only 3 to 5 hectares to consist of 5 to 15 unconnected fragments. Some fragmentation has resulted partly from the building of roads, canals, and the like. But mainly it has resulted from the accumulated effects over generations of splitting land areas for inheritance, marriage dowries, and sale. Effective systems of water control, rotations with pasture, fencing for protection against animals, and the use of large machines are all well-nigh impossible where the village lands are split up into these tiny holdings. Then too, much extra work is required for servicing scattered holdings in contrast to larger blocks. Schemes for consolidating fragmented holdings have been underway for many years in Europe, together with laws to avoid further splitting of small farms. Most western European countries now have this problem under control (Binns, 1950; International Institute for Land Reclamation and Improvement, 1960; Jacoby, 1959). Some effort has been put forward in the less developed countries to correct the problem through consolidation and through prevention of further splitting, but not nearly enough in relation to the enormous potentials. First of all, an individual or team leader undertaking this work must be highly competent, and the cultivators of a farm village must be confident of his integrity. With incompetently drawn schemes, some cultivators end up with a smaller productive base than they had before. A few such bad schemes make the cultivators of other villages extremely cautious.

136

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

Yet one of the best schemes the senior author has seen was developed for Nawdika Village near Hazaribagh in India (Figs. 6 and 7). Both consolidation and contour terraces leading into grassed waterways were

FIG.6. Small parcels of the better soils of the Nawdika Village near Hazaribagh, India, have been consolidated. Contour terraces now conserve the water. A man is standing in the grassed waterway that leads any excess water from heavy rains away harmlessly. This alone doubled production. With fertilizers, improved varieties, and related practices, yields will likely be 8 times the previous ones.

included in one scheme. On the soil suitable for cultivation, the contour terraces became field boundaries, with side boundaries according to the size of each consolidated holding. Since the village lands included two highly contrasting kinds of soil, the fragments in these two parts were consolidated separately. One part is suitable for intensive cropping and the other part only for perennial trees and forage. These developments alone doubled the yields in the village and provided a potential for at least two further doublings with improved varieties and fertilizers to make full use of the greater amount of available moisture in the soil. Similar schemes would be equally beneficial on scores of millions of hectares in India and in other countries too. In areas of small, scattered holdings, such consolidation is nearly always essential for well-functioning runoff control schemes. It also permits vastly improved efficiency of irrigation and drainage by allowing

POTENTIALLY ARABLE SOILS OF THE WORLD

137

canals, ditches, and tilelines to be put in places where they are most effective. Despite the enormous potential benefits, the consolidation of frag-

FIG.7 . Small parcels of poorer soils of Nawdika Village, near Hazaribagh, India, have been consolidated. They are useful here only for small fruit trees and hand-cut grass. The outlet for the grassed waterway shown in figure is in right-center foreground.

mented holdings should not be undertaken lightly. Accurate detailed soil surveys are essential. Without them the risks of failure are very high. These may need to be supplemented by topographic surveys, which can be made easily with the Kelsh plotter or other modern devices using stereo pairs of aerial photographs of good quality. Most village lands include more than one kind of soil with significantly different production potentials under the sets of practices most likely to be used by the cultivators. Thus, total hectares is not a fair guide. The hectares belonging to a cultivator must be weighted, piece by piece, up or down, in reference to a standard for productivity in the village so that when the scheme is completed, each cultivator has a production potential similar to what he had before but not necessarily exactly the same number of hectares. In Europe strong efforts were made to get all of a cultivator’s holdings

138

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

into one block. This is not always possible without injustice. Where kinds of soil are contrasting, this would be a bad goal in many farm villages in the less developed countries. Instead, the soils with approximately similar responses to management should be placed in two or three groups so that many cultivators end up with two or three parcels instead of 15, but not simply with one. Since structures for water control are commonly necessary, they should be taken into account in planning the scheme along with access roads and the like. In many villages with scattered holdings, schemes for consolidation may be complicated by land tenure in achieving the goal of increasing both production and efficiency of soil use and the parallel goal of being fair to all village families. For these reasons programs for land reform, equity of taxation, fair tenant-landlord contracts, and credit may also be needed to make the scheme both productive and acceptable to the families in the village. The consolidation of fragmented holdings offers prospects for greatly improved production in millions of villages. Without it the more familiar inputs of fertilizers, water control, improved seeds, and so on cannot possibly be fully effective. Little outside capital is required. But no aspect of technical assistance requires greater skill in soil science, water control, and the economics of land use. And certainly none requires a greater command of those arts of communication that lead to confidence.

F. SOILEROSION CONTROL The susceptibility of different kinds of soil to accelerated erosion under cultivation varies widely. The principal factor is the relative ability of the whole soil to absorb rainfall rapidly. The most erodible soils are those with permeable surface horizons underlain at a depth of 10 to 40 centimeters by very slowly permeable horizons or layers. After the upper soil becomes saturated, an additional heavy rain causes the soil to flow, even on gentle slopes. Yet highly permeable tropical soils, even very clayey ones having nonswelling inactive clay, may suffer little erosion under cultivation even on strong slopes (Fig. 1). The senior author has seen extremely heavy rains penetrate such soils so rapidly that the outcoming air made a froth on the wet surface of the soil. Thus there is no direct relationship, considering all kinds of soil, between percent of slope and the hazard of erosion under cultivation. The old permeable tropical soils are at one extreme. They are highly permeable. At the other are the Vertisols or “Black cotton soils” that have high erosion hazards under cultivation in the absence of suitable terraces for control of runoff, even with only 2 percent slope.

POTENTIALLY ARABLE SOILS OF THE WORLD

139

Then too, length of slope is a factor. As running water accumulates near the upper part of a long slope, it rushes off at high velocities near the lower part. At these velocities the water has a high carrying capacity for earthy material. In practice, erosion control in fields depends on a combination of plant cover that intercepts the rain and of terraces or diversions to remove the excess water that cannot soak into the soil. Other things being equal, highyielding maize intercepts the raindrops and less soil erosion results than with low-yielding maize. Designs for the structures vary with the length and percent of soil slope, the character of the rainfall -especially during seasons when the soil is nearly free of plants-and, above all, with the morphology of the local kind of soil. Basically erosion and sedimentation are normal processes. It is important to avoid accelerated erosion on arable soils. Yet runoff control to prevent either normal or accelerated erosion may be needed on upland soils lacking potential productivity for crops. The essential combinations of engineering works and the planting of trees, shrubs, and grasses to reduce floods, mountain torrents, and sedimentation may return few onsite benefits. Still the benefits of such works for the protection of low-lying soils and cultural facilities may be highly practicable. Commonly such undertakings need to be planned for whole catchment basins with large or small structures to store water for preventing floods or to save it for irrigation or for producing hydroelectric power. Where such expensive works are planned, the control of erosion, landslides, and sedimentation may be vital to the life of the storage space back of the dams. Commonly, people assume that the sediment in a stream originates more or less uniformly from the soils over a whole catchment basin. But this is not common. Much sheet erosion merely moves soil material from upper slopes to lower ones. The sediment in many streams comes from only 5 to 15 percent of the catchment area- from a few cutbacks along the stream and its branches, deep gullies that have cut into loose earthy materials, recently deposited volcanic ash, or recent earth slides. Yet in catchment areas having little relief, a high percentage of the sediment may come from fields. Fresh, unprotected road cuttings and construction sites are especially susceptible to erosion if left without protective cover during rainy seasons. After severe forest fires or on freshly deposited volcanic ash, erosion and sedimentation can be nearly catastrophic to lower lying soil and streams. Seeding to adapted plants must be done as soon as possible. Some kinds of earthy material, such as ash, are so low in nitrogen or phosphorus, or both, that fertilization may be essential to have an effective cover soon.

140

CHARLES E. KELLOGG A N D ARNOLD C. ORVE,DAL

G . SOILBLOWING The material of some soils and earthy deposits, if barren of plants and subject to strong winds during dry periods, can be blown away. Normally the large sand grains are moved only a small distance at a time. But these moving sands add to the destruction of plants already weakened by drought. If the very fine material gets into the upper air, it may move long distances, even hundreds of kilometers. In deserts and semideserts, soil blowing continues until enough coarse material -coarse gravel and cobbles -accumulates on the surface to protect the soil beneath. In extreme situations, where earthy material lacks such coarse fragments, blowing may continue until rock or some hard layer of the soil is reached. Large areas in the Sahara, for example, are made up mostly of wind-swept crofite calcaire. The cobbly surfaces are called “desert pavement.” Under intensive use of soils for trees or palms, some people mulch the surface with small stones to protect the soil. Soils with coarse blocky structural aggregates within easy reach of tillage machines can be partially protected if this blocky material is lifted to the surface. Deep tillage of soils lacking such structural aggregates, however, usually worsens the hazard of blowing because the tillage buries any organic residues below the surface soils. Methods of tillage that leave a mixture of organic residues on the surface and in the upper surface soil during dry periods lacking growing plant cover are commonly used to reduce blowing. During the past few years modern chemical weed killers have greatly reduced the need for tillage on nearly all soils with good granular structure. Excessive tillage tends to destroy granular structure. For intensive cropping many kinds of windbreaks can be grown. Even in near deserts windbreaks can be established with drought-resistance and salt-tolerant trees or shrubs, provided there is enough water to be had by pumping from ground water or from special basins designed to trap runoff to get the plants well started. As an extreme measure, sand dunes can be smoothed and stabilized by using asphaltic or oil sprays. IV.

New Potentially Arable Soils

In addition to the great potentials for higher yields on the arable soils now in use and for reduction in wastes of farm products in harvesting, storage, and marketing, the total hectares of available arable soils in the world could be a bit more than doubled. That is, the amount of arable soil being farmed could be increased from around 1.4 billion hectares to about

POTENTIALLY ARABLE SOILS OF THE WORLD

141

3.2 billion by adding roughly 1.8 billion new hectares of potentially arable soils that are not now used for crops.* For some countries data are more abundant and more reliable than for others. Yet at least some reliable data are available for nearly every large part of the world. As these data have increased over the years, so have our estimates of potentially arable soils. Most of the 1.8 billion hectares that could be developed for farming are well or moderately well watered. That is, additional irrigation was not assumed beyond the existing potential from wells and streams. No additional estimates were made for irrigation water by desalinization of seawater. Of course, with water available at reasonable cost, additional soils could be made arable and other tropical soils in wet-dry areas could have 2 to 4 crops each year. Thus the world has the physical and biological potential for very much more food. A bit over one-half of the unused potential lies in the Tropics. Of this, roughly one-third has abundant rainfall for crops, one-third has seasons of abundant rainfall alternating with seasons that are relatively dry, and one-third is semiarid or semidesert with short rainy seasons or scanty and erratic rains. As agricultural research continues, doubtless these estimates will appear low to soil scientists of the year 2000. Many have assumed that the effect on farming of accelerating science and technology is to increase yields per acre and to reduce the onfarm labor required. They have these effects along with parallel improvement in the nonfarm service and industrial sectors of agriculture. In the United States today, for example, only about one-fifth of the people employed in agriculture work on farms. But today’s science and technology have an even more dramatic effect on farming. They make possible highly effective new systems of soil use on many kinds of soil that are unresponsive to the only possible systems of a few decades ago. These effects are especially great with well-watered soils of the Tropics. Doubtless these trends will continue. A. SOILSOF

THE

WORLD

The estimates of arable soil, already mentioned and shown in Table I, were made originally for a recent study of the President’s Science Advisory Committee, “The World Food Problem,” 1967. They are based on data about soils and related resources accumulated by the Soil Survey *After the development of these estimates, we noted that with far fewer data the late distinguished agricultural economist and geographer, Dr. 0. E. Baker ( 1 923) some 45 years ago estimated 2.590 billion hectares of potentially arable soils in the world, of which less than one-half was used for crops at that time.

142

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

TABLE I Estimated Area of Potentially Arable Soils in the World Both Cultivated and Not Cultivated Total area Potentially arable of map unit soils in map unit (hectares, (percent) (hectares, millions) millions) Prairie soils and Degraded Chernozems Chernozems and Reddish Chestnut soils (inclusions of darkgray and black soils of the subtropics and Tropics) Dark gray and black clay soils of the subtropics and Tropics (inclusions of Chernozems, Reddish Chestnut soils, and hydromorphic soils) Chestnut, Brown, and Reddish Brown soils Sierozems, Desert, and Red Desert soils (inclusions of Lithosols, Regosols, ans saline soils) Podzols and weakly podzolized soils (inclusions of Bog and Half-Bog soils) Gray-Brown Podzolic soils and Brown Forest soils Latosols, Red-Yellow Podzolic soils (inclusions of hydromorphic soils, Lithosols, and Regosols) Red-Yellow Mediterranean (including Terra Rossa) soils mostly mountainous (including many areas of Rendzina soils) Soils of mountains and mountain valleys (many Lithosols) Tundra soils

122.3

80.0

91.9

381.5

74.0

282.3

500.0 1,203.8

50.0 50.0

250.0 601.9

2,798.2

0.5

14.0

1,294.5 605.2

10.0

65.0

129.4 393.3

3,214.0

43.0

1,382.0

I 1 1.8

15.0

16.8

2,465.4 459. I 13,155.8

0.6 0.0

14.8 0.0 3,182.4

over many years, especially by the World Soil Geography Unit in the last 25 years, and synthesized on maps at the scale of 1 : 1,000,000. To give some suggestions of the distribution of soils in the world, a highly generalized soil map is included as Fig. 8. Estimates of potentially arable soils were related to a similar map in 1964 (Kellogg, 1964). Those were somewhat lower than the current estimates more carefully made for “The World Food Problem.” Since more people are familiar with the older great soil group names, such as Podzol and Chernozem, than with those in the new and less familiar system of soil classification (Soil Survey Staff, 1960, 1967), we have used the old names. Table I1 gives the names of the approximate equivalents in the new classification.

FIG.8. A small-scale soil map of the world.

I44

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

TABLE 11 Names of Great Soil Groups Shown on the Soil Map or Discussed in the Text and Their Approximate Placement in the Orders of the New System of Soil Classification (Soil Survey Staff, 1960, 1967) Great soils groups Alluvial soils Brown soils Brown Forest soils Chernozem Chestnut soils Dark Gray and Black soils of subtropics and Tropics Degraded Chernozem Desert soils Gray-Brown Podzolic soils Gray Wooded soils Hydromorphic soils Latosols Lithosols Organic soils Podzols Prairie soils Reddish Brown soils Red Desert soils Reddish Chestnut soils Red-Yellow Mediterranean soils Red-Yellow Podzolic soils Regosols Rendzina Saline soils Sierozem Terra Rossa Tundra Weakly podzolized soils

Orders in new classification Entisols Aridisols, Mollisols lnceptisols Mollisols Mollisols Vertisols Alfisols, Mollisols Aridisols Alfisols (mainly suborder Udalfs) Alfisols (Aquic taxa of various orders) Oxisols, Ultisols, lnceptisols (Lithic subgroups of several orders) Histosols Spodosols Mollisols (mainly suborder Udolls) Aridisols, Alfisols Aridisols Mollisols, Alfisols Alfisols (mainly suborder Xerall's) U1tisols Entisols (mainly suborder Psamments) Mollisols (suborder Rendolls) Aridisols Aridisols Alfisols (mainly suborder Xeralfs) lnceptisols Spodosols, Inceptisols, Alfisols

Potentially arable soils are those that give an acceptable production of cultivated crops adapted to the environment. In making the estimates, the average level of agricultural technology used in the United States was assumed. The figures in hectares include all arable soils, both cultivated and uncultivated. Part of the uncultivated soils need practices for water control - runoff control, irrigation, drainage, or some combination of these-clearing of trees, stone removal, and so on. It was assumed that the cost of these measures, however, would not be excessive in relation to anticipated returns. All arable soiIs have potential for grazing. Those with adequate rainfall can be used for forests.

POTENTIALLY ARABLE SOILS OF THE WORLD

I45

1. Soil Descriptions This section gives descriptions of the soils shown on the map in Fig. 8. Since no area o f a single kind of soil is large enough to be shown by itself on such a small-scale map, the map units are made up of several great soil groups. The principal kinds of soil are named for each map unit, but others in it are not named. Some areas of Alluvial soils, Lithosols, and Regosols, for example, occur in all map units. Also, Bog and Half-Bog soils occur in several of them. The areas of soils shown in Table I are based on measurements of units shown on this map. The figure for any map unit, therefore, includes areas of both the named kinds of soil and the inclusions. From studies of maps of much larger scale and of detailed soil surveys of sample areas, the areas of inclusions were estimated to get a total of potentially arable soils for each map unit. Prairie soils and Degraded Chernozems make up only about 122 million hectares or slightly less than 1 percent of the land surface of the world. Yet their high responsiveness makes them important. A high proportion are already used for food production. They occur mainly in the United States and the Soviet Union. Prairie soils in the United States are used especially for high-yielding corn (maize), although forages and other cereals are grown successfully too. In the Soviet Union and eastern Europe, where Degraded Chernozems are more extensive than Prairie soils, the soils are used mainly for small grains. Soil moisture for crops is commonly abundant, although shortages occur now and then. Since the soils are cold in winter, double cropping is not possible. Prairie soils, developed under tall grass vegetation in moderately temperate regions, are usually adjacent to regions that were originally forested. They have dark brown or dark grayish brown surface soils grading into brown subsoils that are somewhat richer in clay than the surface soils. Although somewhat leached, Prairie soils are high in fertility; they are intermediate between Gray-Brown Podzolic soils and Chernozems. Degraded Chernozems have a gray, leached layer between a black or dark brown surface soil and a brown subsoil. Chernozems and Reddish Chestnut soils (some inclusions of dark gray and black soils of subtropics and Tropics). A high proportion of this group is used for farming: but soil moisture, although enough for sustained annual production of wheat and other small grains, is commonly too low for high yields in some years. Yet considerable corn and sunflowers are grown on them, and sugar beets are important on those with the best supply of moisture. The group of Chernozems and Chestnut soils and inclusions provide an estimated 282 million arable hectares. The principal areas are in the United States, the Soviet Union, eastern Europe, and Argentina. The Chernozems are black soils rich in humus to a depth of some 45 to 90 centimeters. Beneath this dark material is a layer of calcium carbonate accumulated from leaching of the soil above. In their natural state, Chernozems are covered with grass and the soils are granular and high in fertility. Chernozems were the great wheat-producing soils of the world. Now corn, sunflowers, and sugar beets are also grown. Despite their granular structure and high fertility, yields usually are not high because of limited soil moisture. Nevertheless, substantial progress has been made through better tillage practices. improved varieties, and strategic use of modest amounts of nitrogen and phosphatic fertilizers. In fact, we may be on the threshold of grain yields twice those of two decades ago. The Reddish Chestnut soils have about the same general fertility as the Chernozems and about the same soil moisture limitations, but, since they are commonly warm for longer

146

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

periods, a wider range of crops can be grown. Sorghum and cotton are common crops in addition to small grains. With irrigation crop production can be increased on both Chernozem and Chestnut soils. Dark gray and black soils of the subtropics and Tropics (inclusions of Chemozems, Reddish Chestnut, and hydromorphic soils) cover nearly 500 million hectares. About onehalf of the map unit is estimated to be potentially arable. The principal areas are in Africa, South America, Australia, and India. These soils occur in several other places, including the southern part of the United States, in areas too small to be shown on the soil map. Dark gray and black soils of this group go under a variety of names, such as black cotton soils, Grumosols, Regur, “self-mulching” soils, “cracking” clays, and tropical black clays. In the new United States system, they are called Vertisols (turning soils). The mature uncultivated Vertisols have a distinctive microrelief called “gilgai.” Most areas of soils in this group are in the warm or tropical regions having alternating wet and dry seasons. Some of them are moist for several months each year. Most are nearly flat. The soils of this group have poor structure, are high in clay, are plastic, and swell and shrink. When dry, vertical cracks extend downward 30 centimeters to 1 meter. Although these soils are moderate to high in fertility, they are hard to work. In fact, they can scarcely be cultivated at all with only hand tools. For this reason, some areas are still available for settlement by people using power machinery and modern techniques for drainage, irrigation, tillage, and fertilization. Chestnut, Brown, and Reddish Brown soils cover nearly 1,204 million hectares. Substantial areas are on all continents. About 50 percent is estimated to be potentially arable. These are the soils of the warm to cool temperate grasslands. The Chestnut soils have a somewhat drierclimate than Chemozems. The Brown soils are still drier but are not desertic. The Reddish Brown soils, which extend into the Tropics in some places, are drier than the Reddish Chestnut soils but not so dry as the Red Desert soils. The surface soils have moderate supplies of organic matter, and commonly at about 20 to 30 centimeters there are layers of accumulated calcium carbonate leached from the soil above. Although moderately high in fertility, insufficient soil moisture is a serious limitation common to all soils of this group. Crop yields are generally low. Yet, with methods to conserve moisture and to prevent soil blowing, many areas are used for producing small grains and some other crops. More hectares are used for grazing than for cropping. With irrigation, the productive capacity of these soils can be increased severalfold and the range in crops that can be grown substantially enlarged, especially in warm areas. Under irrigation, nitrogen and phosphatic fertilizers are needed for sustained high yields. Such micronutrients as iron and zinc may also be needed. Sierozems, Desert, and Red Desert soils (inclusions of Lithosols, Regosols, and saline soils) cover nearly 2,800 million hectares, slightly more than 2 1 percent of the land surface of the world. Large areas are found on all continents, especially in Africa, Asia, and Australia. Dryness is common to all the soils and irrigation is essential for crop production. Only about 0.5 percent of the large area is estimated to be potentially arable, yet this amounts to about 14 million hectares. Our estimate is based in part on judgment about the availability of irrigation water at reasonable cost and does not include water that may become available from desalinization of seawater or from other expensive sources. Even if unlimited water supplies at reasonable cost were to be assumed, the estimated proportion of potentially arable soils would not be high, although it would be several times the present estimate. A substantial proportion of the areas shown on the map include soils too shallow, stony, steep, cold, sandy, or saline to be considered as potentially arable under the assumptions used in this paper.

POTENTIALLY ARABLE SOILS OF THE WORLD

147

The Desert soils have scanty vegetation and most are gray. The Red Desert soils are similar except for their reddish color; they are mainly in the tropical and subtropical deserts. The Sierozems are gray soils of dry temperate regions that have somewhat more vegetation and organic matter than the Desert soils. Most of these soils, except for the very sandy ones, are well supplied with the primary plant nutrients except nitrogen, but the balance is not always favorable. Because irrigation is expensive, only the soils reasonably free of limitations other than dryness should be selected for crop use; and even these need to be managed well to make production economically feasible. With good soil selection and good management, including heavy fertilization, adapted and responsive crop varieties, and control of weeds, insects, and diseases, production can be very high. This is especially true in the Tropics where, with the moisture limitation removed, two to four crops can be grown on the same hectare each year. Hence, at reasonable cost, modest increases in irrigated hectares in tropical and subtropical deserts could substantially increase the world food supply. Podzols and weakly podzolized soils cover nearly 1,295 million hectares. The large areas are in northern North America, northern Europe, and Siberia. Between I29 and 130 million hectares are estimated as potentially arable. Podzols are leached soils of the humid cool-temperature to cold forested regions. On the soil surface is a mat of acid, peatlike organic matter from a few to 30 centimeters thick. Under it, the surface layer of humus-rich mineral soil is only 1 or 2 centimeters thick. Directly beneath is an ashy gray, leached layer from which these soils get their name. The subsoil is a brown layer richer in humus and iron oxide than the gray layer above it. The weakly podzolized soils also have at the surface a mat of acid peatlike organic matter, but the horizons (or soil layers) below are less clearly developed than in the Podzols. Included in the southern fringes of the large areas in North America, and presumably in Eurasia too, are some Gray Wooded soils, which resemble the Podzols but are less acid and have clayey subsoils (Fig. 9). Also included throughout the entire area of Podzols, as shown on the map, are innumerable small areas of Bog and Half-Bog soils. In addition to the limitation of short freeze-free seasons, many local kinds of soils are too stony, hilly, sandy, or swampy for crop use. Crops have been produced on the least cold fringes of the large areas of Podzols and associated podzolized soils for a long time. Liming and the application of fertilizers containing nitrogen, phosphorus, and potash are necessary for satisfactory yields. With development of plant varieties that mature in shorter growing seasons and practices, including liming and fertilization, designed to make the best use of the solar energy available, cropping can be pushed onto many millions OF hectares now in forest. The major proportion of areas delineated as Podzols and podzolized soils, however, is not potentially arable in the foreseeable future. Gray-Brown Podzolic soils and Brown Forest soils: This group of soils, although not extensive, is highly important. As shown on the map, this group of soils occupies about 605 million hectares or between 4 and 5 percent of the land area of the world. About 393 million hectares are estimated to be potentially arable, most of which are already under cultivation. Gray-Brown Podzolic soils in their natural state have thin organic-rich surface layers underlain by grayish brown or yellowish brown leached layers over brown subsoils. These brown subsoils are richer in clay than the surface soils from which part of the clay has come. The less extensive Brown Forest soils are similar, but their subsoils are not enriched with clay and the soils are somewhat less acid and more permeable. The Gray-Brown Podzolic soils like other podzolic soils are leached, acid, and relatively poor in most plant nutrients in available form. They also are low in organic matter, especially

148

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

after a few years of cropping. These soils commonly have an abundance of soil moisture and are highly responsive to management. With lime and fertilizer they support a wide range of crops, for both food for people and feed for livestock. The soils are too cold in winter for plant growth. Commonly only one crop can be grown during the year.

FIG.9. New clearing of Gray-wooded soil in northern Saskatchewan. (In the region of Podzol soils on the soil map.) It was on these soils and their close relatives that much of the research on soils and plants and their relationships was conducted in the 19th and early 20th centuries. From the fall of Rome nearly to the French Revolution, grain yields in Europe were low, between 350 and 600 kg. per hectare. With the adoption of crop rotations they nearly doubled, and elimination of the fallow year gave more hectares for harvest. The application of chalk and farm manures was found to be beneficial and became a general practice. By 1850, wheat yields had risen to about 850 kg. per hectare in France, about 1200 in Germany, and somewhat over 1200 in the United Kingdom. By 1906, they had gone to 1200 kg. in France, 1800 kg. in Germany, and beyond that in the United Kingdom. Now they stand at about 2100 kg. in the United Kingdom. Somewhat more than half the increases in the United Kingdom and Germany came before the common use of fertilizers after 1850 (Ignatieff and Page, 1958). This remarkable increase in production on Gray-Brown Podzolic soils has been attained through improved technology, much of which grew out of research of the past 160 years or so. On no other great group of soils has a comparable amount of research been done. By expanding the methods of research first learned on the Gray-Brown Podzolic soils, the technology for doubling or tripling yields has already been developed for more kinds of soils and is in early stages for others. The limit has not yet been reached on any kind of soil, not even on the Gray-Brown Podzolic ones. Latosols and Red-Yellow Podzolic soils (inclusions of hydromorphic soils, Lithosols, and

POTENTIALLY ARABLE SOILS OF THE WORLD

149

Regosols) cover nearly 32 14 million hectares or between 24 and 25 percent of the land area of the world. The humid and wet-dry Tropics and subtropics are covered mainly by these soils. The estimated potentially arable hectares come to 1382 million of which today less than one-fourth are under cultivation; in the vast humid and wet-dry tropical regions of South America and Africa lie the great areas of unused or but little used potentially arable soils. Latosols, the principal component of this group, are extensive under the rain forest of the humid Tropics and under both woodland and anthropic savanna of the seasonal wet-dry Tropics. The name Latosol covers a wide group of soils, including Red, Reddish Brown, Reddish Yellow, and Yellow Latosols (Kellogg, 1950). Included also are relatively young soils from basic rocks and volcanic ash. The minerals in the Latosols are highly weathered. The soils have been subject to strong leaching with warmer water than in temperate regions. Yet in the humid Tropics abundant plant nutrients are held in the living trees. These nutrients make a continuing cycle from the soil into the plants and back to the soil as the twigs, fruits, leaves, and old wood decay. On many of the mature Latosols after the forest is removed and crops are grown without fertilizer 1 to 5 years, the supply of plant nutrients is too low for economic production. But under shifting cultivation the nutrient supply can be restored by 6 to 12 years of forest fallow. In the Reddish Brown Latosols plant nutrients are not so scanty, but they become so after a few years of cropping without fertilizers, compared to Chernozem or well-managed Gray-Brown Podzolic soils. Latosols vary widely in the amount of available phosphorous for crop plants. Some are exceedingly low indeed (Fig. 3). The small-scale soil map includes soils with wide differences in native fertility. Over the years they receive varying amounts of volcanic ash, dust from dried foam blown in from the sea, and dust from the deserts. Thus, speaking generally, the Latosols in Africa, on the whole, have higher initial amounts of calcium and other mineral plant nutrients than those in South America. Yet there are exceptions in both continents. Apparently enough calcium and other material to be beneficial moves south from the Sahara Desert and north from the Kalahari Desert.* Many of the Latosols under forest are higher in organic matter than they appear to be because the organic matter is brown, not black. Some of the Reddish Brown Latosols have as much organic matter as the nearly black soils of the American Midwest. Yet decomposition is rapid in warm humid climates. Unless measures are taken to maintain good cover, the organic matter drops to low levels with cropping. Most Latosols are permeable to water and air. Most are free of physical barriers for crop penetration, are easy to maintain in good structure or tilth, and have at least moderate moisture-holding capacity. Perhaps as much as 75 percent of the total area of Latosols and associated soils has smooth relief suitable for the use of farm machinery. Yet some ofthose inherently most productive are on strong slopes and require measures for runoff control to have good supplies of moisture. With important local exceptions, most Latosols are responsive to modern management systems and can be highly productive of many food and industrial crops. Although practical ways have not yet been found for efficient farming on the most infertile Latosols, such as some in Brazil, at a high level of productivity, research can be expected to find ways (Figs.

* Interestingly a common tree of the Africa rainforest, Chlorophora excelsia, accumulates calcium from the soil, even acid soil, and from dust on the leaves. In the odd tree are stones of calcium carbonate up to several kilograms in weight that local people use traditionally to make whitewash for their walls-people living hundreds of kilometers from any other source of limestone.

150

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

10 and 1 I). Already systems are in use for making some Latosols, such as many in Hawaii and northern Queensland, highly productive, perhaps too high for use as a standard for estimating productivity of most Latosols during the next two decades. But certainly these examples of success justify optimism about the technical feasibility of high production on millions of hectares of Latosols.

FIG.10. View of stunted cover (cerrodo) on uncultivated "dark red Latosol," probably Ustox in the new classification, near Jatai, Brazil. The stunting is probably due to mineral nutrient deficiences, especially of calcium, and seasonal dryness. Scattered throughout most areas where Latosols occur, especially in the wet-dry Tropics, are some that contain laterite. This iron-rich material of Ground-Water Laterite soils in the wet-dry Tropics is soft when formed under some 40 to 100 centimeters of soil. Yet if the climate becomes very dry or if the soil is removed through natural or accelerated erosion, the laterite hardens irreversibly. We have seen road cuttings in laterite that are now like ferroconcrete (Alexander and Cady, 1962). Once laterite hardens, it weathers only very slowly. It is found in spots under many kinds of soil developed from materials deposited above it, such as alluvium and volcanic ash. Soils with hard laterite have severe limitations for crop production because this material commonly restricts root penetration and because the surface soil is subject to serious erosion if unprotected from heavy showers. Yet some soils with soft laterite are used with reasonable success by local cultivators who have learned not to plow but to maintain a mixed culture of palms, food crops, trees, and the like. Curiously, the cultivators in Kerala of southern India discovered many centuries ago that these soils could be used for mixed cultures without plowing. We should also point out that there are small areas of soils containing hard laterite now in places having a different climate from that under which the soils formed. That is, climates do shift but once laterite is formed, it changes extremely slowly. Such soils, although dramatically conspicucius locally, make up less than 10 percent of the Latosols.

POTENTIALLY ARABLE SOILS OF THE WORLD

151

FIG. 1 I . Closer view of trees shown in Fig. 10.

Red-Yellow Podzolic soils differ from Latosols mainly in having a subsoil higher in clay than the surface soil. The clay is somewhat more active than the clay in Latosols although it is less active than that in Gray-Brown Podzolic soils. Because the subsoil of Red-Yellow Podzolic soils is less permeable than that of Latosols on comparable slopes, the erosion hazard is greater. The largest single area of Red-Yellow Podzolic soils is in the southeastern part of the United States. The steadily increasing success in developing ways to manage these soils for sustained high production of a variety of crops suggests that corresponding or even greater success should be possible with Latosols on which double cropping or even triple cropping is technically possible over extensive areas. After all, soils in the Tropics are always warm enough for plant growth, except in the high mountains, and this is a unique advantage over all other regions, including the subtropical southeastern part of the United States. Red-Yellow Mediterranean (including Terra Rossa) soils, mostly mountainous (including many areas of Rendzina) cover slightly less than I12 million hectares. Of this total, somewhat less than 15 million hectares are estimated to be potentially arable because

152

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

many of the soils are too steep or shallow for ordinary cultivation. The map shows areas of this group only around the Mediterranean Sea, but there are other small areas in southern Australia and southern California. The Red-Yellow Mediterranean soils are limited to the warm temperate and subtropical areas having dry summers and moist winters. Temperatures are a bit cool for vigorous plant growth when moisture is abundant, and moisture is scarce when the temperatures are high and more favorable for growth. The soils are widely used for olives, fruits, and vines. Terra Rossa soils, the best known members of the Red-Yellow Mediterranean soils, have developed over hard limestone. Many are on hills and mountains and tend to be thin over hard rock. In the Mediterranean basin many formerly cultivated areas have been made thinner by erosion over the centuries. Intermingled with Red-Yellow Mediterranean soils are Rendzina-like soils that have grayish to nearly black surface soils and no red color. Most of these are thin over soft limestone. Red-Yellow Mediterranean soils are not strongly leached; most of them are calcareous from the surface down. Yet, like the Latosols, Red-Yellow Podzolic, and Gray-Brown Podzolic soils, they need fertilizers, especially nitrogen and phosphorus. Soils of mountains and mountain valleys (many Lithosols) cover about 2465 million hectares, mainly in western parts of North and South America and in Europe and Asia. Most of the soils are too steep and stony for feasible cultivation. Many are thin over hard rock too, and many are too cold for crops. Nearly 15,000,000 hectares ;ire estimated to be potentially arable. These are mainly soils in valleys in tropical and temperate regions where temperatures are warm enough for cropping. The potentially arable soils vary greatly, depending on climate, vegetation, and parent material. Many are stony. Most soils of mountain valleys are young and have weakly developed soil horizons. Water for irrigation is available in many mountain valleys. Tundra soils cover about 459 million hectares across the northernmost parts of North America and the Soviet Union. Tundra soils consist of dark brown peaty material over gley horizons. Most Tundra soils are underlain by permafrost and are too cold for food crops. In favored places where the soils slope toward the sun, grasses and a few short-season food crops can be grown, but such areas are so small that no Tundra soils have been included in our estimates of potentially arable soils. Alluvial soils occur in innumerable small and irregular areas in all parts of the world. Even the largest areas, shown by the symbol “A” on the soil map, are too sinall to delineate at such a small scale. Alluvial soils are inclusions in all the units on the soil map, and they were considered in estimating arable hectares in the map units that ire shown. A high proportion of the estimated arable hectares in the unit “Soils of mountains . . . ” are in fact Alluvial soils. Altogether about 590 million hectares of Alluvial soils are distributed among the map units. Although not all Alluvial soils are potentially arable, they supply food to perhaps as much as 25 percent of the world’s population. The Alluvial soils include only those that are now occasionally flooded or that have been subject to flooding until recently, mainly in elongated areas along streams. The group does not include soils developed in old alluvium that now strongly reflect the longtime influences of climate and vegetation. The fertility of Alluvial soils is directly related to the nutrient content of soils and rocks in the watershed from which the alluvium came. For example, Alluvial soils from erosion products of sandstones and shales usually are relatively low in fertility but those from limestone, basalts, and other nutrient-rich rocks are highly fertile. Along most of the large streams, Alluvial soils of mixed origin are relatively well supplied with plant nutrients.

POTENTIALLY ARABLE SOILS OF THE WORLD

153

The foremost management problem with Alluvial soils is water control, including protection from flooding and drainage because many of these soils in humid areas are naturally wet. In dry regions, low areas of Alluvial soils are likely to be highly saline or alkalinelimitations that must be removed by drainage and leaching before the soils can be made productive under irrigation. Even though unleached and relatively fertile, Alluvial soils respond well to fertilizers. High yields usually require nitrogen and phosphorus, and some may require potassium. Alluvial soils in arid regions also may require correction of deficiencies in zinc and iron. Lirhosols and Regosols are listed as inclusions in three map units. In addition small areas are included in all map units; and Lithosols make up the most extensive component of one map unit, “Soils of mountains . . .” Lithosols are commonly steep and stony soils shallow over rock. They are not considered to be potentially arable although a few hectares have been developed for cropping by making stone terraces with much hand labor. Extensive areas are suitable for grazing or forestry. Regosols, like Lithosols, are widely distributed. They are young soils from sand, silt, glacial drift, and the like; a high proportion are sands with a low water-holding capacity that imposes a severe limitation on their use for cropping. The sandy ones are highly subject to blowing in the absence of good plant cover. Locally, there are Regosols well suited to cropping, but the proportion is low. Organic soils: Bog and Hay-Bog soils are wet or seasonally wet organic soils in millions of small areas wherever plant remains accumulate faster than they decompose. They are common in the cool temperate and cold humid regions of the world but some occur in the humid Tropics and subtropics. Usually they occupy wet swampy depressions or flats. Organic soils are roughly divided into “peats” and “mucks.” Peats commonly are raw and poor in mineral matter, although some have considerable sand, volcanic ash, or other mineral matter that was blown or washed into them. Mucks commonly are well decomposed. The material is finely divided and the plant remains are not easily identified. Mucks commonly have more mineral matter and are less acid than peats. A high proportion of organic soils are too cold for feasible crop production; they warm up later in spring than adjacent mineral soils. In temperate to tropical climates some mucks are used for cropping and, with proper fertilization and water management, can give high yields of vegetables and of some other crops, including hay. With drainage and cultivation, the organic matter of Bog or Half-Bog soils of warm regions is likely to decompose more rapidly over the years than it can be replaced under practical farming methods.

B.

WHY

So MANY POTENTIALLY ARABLESOILS ARE U N U S E D

FOR CROPS

For a great many natural, cultural, economic, historical, political, and technological reasons the population of the world is not distributed in the same way as the soils suitable for cultivation. As one example, the Mediterranean region and the dry parts of northern Africa and the Middle East were much more moist during the Pleistocene (the long glacial period) that ended some 10,000 years ago. During the long moist period before the end of Pleistocene much human population

154

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

was probably concentrated in this general region. As the glaciers receded, many people moved north into what is now Europe; others stayed behind. As the soils became drier, irrigation became necessary. Thus, the world now has a heavy concentration of population in areas not highly suited to farming, except with the heavy cost for additional water; whereas elsewhere areas of well-watered soils of high potential are unused for crops. Until recent times, most heavy transport in the world was by water. With only a few exceptions, most of today’s large cities are on the margins of oceans, large lakes, or navigable streams. Water traffic was supplemented by canals and, more recently, by railroads and heavy truck highways that permit expansion from the port cities into remote areas. Since the lower reaches of the Congo River are not navigable, people in this great basin did not join the world system of trade and culture until the end of the 19th century. A world map of commercial farming approximates a world map of facilities for heavy transport. In North America, for example, it was nearly 200 years after the first settlements before people used for crops the soils that were most productive with the techniques then available. In both Canada and the United States, farming began on the poorer soils. In many of the less developed countries people moved from the seaport towns by foot, short wagon roads, or canoes. Now there are air routes to the interior of these areas, yet the essential highways and railroads for heavy transport are lacking.

1 . Associated Economic Enterprises Then too, few commercial farming areas are productive enough to be able to support all costs for transport and other facilities out of farm income alone. We are mindful that some 200 million people still get their food by primitive shifting cultivation with little connection with the trade of the world. Nor do they share the many advantages of a modern society. Most prosperous commercial farming is associated with other economic activities, including, perhaps, mining, forestry, fishing, and especially good manufacturing based on coal, oil, or hydroelectric power. Thus, to establish new viable farming areas within which farmers may aspire to good standard of living, besides responsive soil, the area should have potential for other economic enterprises to share the costs of the infrastructure. Education and health are especially important. Skilled people are needed in all sectors of modern agriculture, not only in farming but also in the agricultural services and industries that supply machines, tools, and chemicals, and in those that process, store, and market farm products. In addition, adequate political institutions and

POTENTIALLY ARABLE SOILS O F T H E WORLD

155

security of persons are essential. These conditions are lacking in many important areas of abundant resources for farming and related enterprises. Perhaps most people in the advanced countries take for granted an open literate society and scarcely know how to communicate with people of a closed illiterate society who lack experience with technology based on modern science. Attempts at rapid change in these societies can lead to disaster. Only with great patience, skill, and tact can those minimum changes be successfully promoted that are necessary for people to adopt new techniques for improved health and farming and to build for themselves reliable institutions. Many have lightly oversimplified the handicaps to farm development in new areas into a simple slogan of “capital requirements.” Ultimately, of course, some capital is needed, but capital cannot be useful until better farming systems, adapted to both the resources and the people, have been worked out in ways that can be facilitated by good institutions, sound planning, reasonable political stability, and the security of persons.

2. Some Special Problems of the Humid Tropics As already pointed out, some of the most potentially productive arable soils are in the humid Tropics (Kellogg, 1965). Intermingled with the highly responsive soils are others not suitable for farming. The importance of reconnaissance soil surveys to locate promising areas and of detailed soil surveys for planning farm operations of any size can scarcely be overemphasized. The record of highly wasteful failures in the Tropics without such soil surveys is long. Many of the best undeveloped tropical soils are covered with heavy tropical rain forests. For some reason these soils are the least well understood of any. Many, many years ago someone said that these soils are “thoroughly leached and the organic matter burned out of them.” Judged by yields of earlier days, they were no more leached of plant nutrients than many of the currently most productive arable soils of the eastern part of the United States and of northwestern Europe. And they are higher in organic matter, although it is brown rather than black. Latosols are low in plant nutrients for crop production. In many tropical soils even calcium or phosphorus may be very low indeed, especially in the old landscapes that have not received desert or volcanic dusts. It seems reasonable that the greater dust explains why the bulk of the soils under the rain forest in Africa below the Sahara are more productive than the bulk of those just south of the Amazon in South America. The stunted growth of an uncultivated seasonally dry Latosol near Jatai, Brazil, is shown in Figs. 10 and 11. Roughly, similar views can be seen in Africa

I56

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

due mainly to fire, or fire and phosphorus deficiency, rather than to extreme calcium deficiency. It is extremely difficult to generalize about such large regions since the local kinds of soil are contrasting. Yet there are many examples in Hawaii, Puerto Rico, Queensland, Ghana, and in the Congo basin where very high yields have been obtained of many highly important crops. An enormous potential exists within these soils, which modern science and technology could develop to a very high level with a modern agricultural system. a. Contrasting Present Use of Tropical Soils. The labor incomes of cultivators using tropical soils today contrast greatly. Partly these are related to highly contrasting local kinds of soil. The soils range in age from very young to extremely old. They have had drastic changes in climate and associated changes in vegetation and erosion cycles during their formation. In the Tropics probably more contrasting local kinds of soil will finally be recognized in detailed soil surveys than in the rest of the world. Yet even on the same kinds of soil contrasts are very great. Roughly 200 million people in tropical and subtropical regions get their food mainly under some system of “shifting cultivation.” The system has local modifications and many synonyms, including milpa, djum, ladong, ray, slash-and-burn, kaingin, and dozens of others. It was long ago used widely in northern Europe. In essence, shifting cultivation refers to the practice of cutting trees and shrubs (or, less commonly, tall grasses, such as Pennisetum purpureum); burning: growing crops for 1 to 5 years and then allowing the natural cover to return to regenerate the productivity of the soil for 5 to 15 years (Kellogg, 1962b; Nye and Greenland, 1960) (see Figs. 12-15). Shifting cultivation requires a great deal of labor input for the harvest. Some emotional people have called it wasteful. Yet only differences in cultural attitudes can explain why some farmers are praised for growing sweetclover or alfalfa to enrich the soil and others are censured for using trees for the same purpose. Actually the systems of shifting cultivation, like other farming systems, can lead to soil deterioration or not, depending on soil selection and how well each system is carried out. Where medicine and other public health measures were introduced into remote villages, populations increased. As a consequence of the urge for more food, and without technical assistance for improved farming, natural fallows were shortened and soil productivity fell. If the crop part of the cycle is unduly lengthened, grasses come into the fallows rather than trees. During the dry seasons the grass is likely to be burned to round up game, to give temporary tender pasture, or just to

POTENTIALLY ARABLE SOILS OF THE WORLD

157

FIG. 12. A corridor has been cut and is ready for burning in the first stage of the modified ancient Bantu system of shifting cultivation near Yangambi in the central Congo. The soils were classified as Reddish-Yellow Latosol and would now be Haplorthox in the new system.

FIG. 13. A corridor in the second stage of the modified Bantu system of shifting cultivation after clearing and burning. Maize is the first crop. Cassava and bananas follow. The soil is like that in Fig. 12.

158

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

FIG.14. A corridor in the third stage of the modified Bantu system of shifting cultivation. Cassava and bananas are growing together. Following the final harvest of cassava, the forest will return under the shade of the second-growth cassava. The soil is like that in Fig. 12.

see it burn. The result is an infertile soil for crops. Only the grasses and tough sunloving “savanna” trees remain. They hold only a small amount of nutrients from leaching during the wet season. This kind of cover is called “anthropic” or man-made savanna (Fig. 16). Most of the savanna in the tropics is of this kind. The former forests may have been cut many centuries earlier. Carried to this point, fire is a great enemy of soil fertility. Where the fires are controlled, which is exceedingly difficult, the trees do return and rejuvenate the soil (Fig. 17). Anthropic savanna is made worse by invasion of Imperata cylindrica, called cogon grass, spear grass, and other local names. It is nearly impossible to eradicate this grass with simple tools, but it can be destroyed with heavy tractor-drawn plows and the soil reclaimed. Shifting cultivation can be improved considerably with additions of fertilizer, especially if placed a bit to the side and a bit below the seeds as they are sown. Good compost made from green plants with emphasis on those rich in leaf protein can be very helpful, especially where ferti-

POTENTIALLY ARABLE SOILS OF THE WORLD

159

FIG. 15. A periodic clearing under shifting cultivation on Red to Reddish Brown Latosol. The bamboo suggests that this area is a bit overused in crops, partly because of its nearness to a new road. Bamboo is a poor fallow tree. The area will be burned and seeded to cotton after the first rain. After one crop the wild plants will be allowed to return. The soil is probably a Rhodustult in the new system. A few miles south of Damra, Assam, India.

lizers are too costly or unavailable. Compost can also be improved by adding animal manure to the mixture. The next steps beyond shifting cultivation are mixed cultures of crops and forest trees, including perhaps cassava, bananas, maize, palms, and other crops. Much cacao is grown in only partially cleared forest (Fig. 18). Elsewhere in the Tropics and on kinds of soil similar to those used under shifting cultivation, some of the most efficient farming of the world is carried on with very high yields of both food and industrial crops. Many tropical soils that produce poorly as they are found in nature respond enormously to management. An easily corrected deficiency of phosphorus is a common example. Figure 3 shows a tropical soil so low in phosphorus that, until this deficiency is corrected, little if any useful crop or even wild vegetation can be grown. Then too, the response curve to phosphatic fertilizer is commonly sigmoidal rather than parabolic. LeMare ( 1 968) gives a recent example. There are many others, including soils of the southeastern part of the United States. That is,

160

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

FIG.16. Anthropic savanna, with termite mound on sandy Red Latosol, probably a Haplustox in the new system, near Kamind, northwest of Lubumbashi (ex Elisabethville), Congo.

unless an adequate amount of phosphatic fertilizer for the local kind of soil is used, the response to smaller applications is uneconomic. Then

POTENTIALLY ARABLE SOILS OF THE WORLD

161

FIG. 17. Natural forest returns quickly on anthropic savanna if protected from fire. This area has been protected for only 9 years. In a few more years the savanna grasses and associated trees will die and the green forest cannot be burned unless first cut down. In Low Congo on Yellow Latosol, probably Haplorthox in the new system.

too, many, but by no means all, tropical soils tend to fix added phosphate in unavailable forms, which gives added emphasis to precise placement in relation to the seed or plant. Many very old tropical soils of rainy regions are extremely low in exchangeable calcium, especially where they receive little dust from deserts or volcanoes, as in some of the upland areas of central Brazil. Under modern farming systems these and similar problems have been easily overcome by following designs of proper combinations of practices that fit the local kind of soil with suitable materials available from indu stry . As pointed out several years ago (Kellogg, 1950), we can recognize in general terms three levels of management of those tropical soils that are used for farming: ( 1 ) Management systems of the indigenous farmers without the benefit of modern science and without the facilities of modern industry. This management is not always bad- not by any means-although it may look so to one familiar only with the methods of farmers in temperate regions. The soil researcher going to the Tropics can learn a great deal by studying the native practices that have been worked out by trial and error over hundreds of generations. (2) Management systems involving the application of modern scientific knowl-

162

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

FIG. 18. Mixed culture of abaca with good thinned, gallery forest of Terminalia superba on Red Latosol, probably Haplorthox in the new system. Near Gimbi in the Low Congo.

POTENTIALLY ARABLE SOILS OF THE WORLD

I63

edge about plant growth, water control, use and function of organic matter, plant breeding, and so on, but with only the minimum of such products of modern industry as machinery, electric power, and chemical fertilizers. (3) Management systems involving the application of all modem scientific knowledge and full use of the products of modern industry. Yields and the ratio of output to input may vary enormously on similar soils under these three contrasting levels of management. Even where yields are similar, labor income may vary enormously. Perhaps of even greater general importance is the fact that large areas in t h e Tropics that cannot be farmed economically at all under the first set of systems can be under the other two, especially the third one.” We must think rather of the soil potentialities with the new tools and methods that science and industry can make available . . . . Certainly none of us would think of estimating the potentialities for production of soils in temperate regions on the basis of nonscientific management. Suppose we tried that-suppose we estimated the potential productivity of the soils in, say, Denmark, Holland, Belgium, and the eastern part of the United States on the basis of what could be done without scientific methods and without the products of modern industry?

Travelers in the Tropics commonly associate poor crops with “poor” soil and are thus led to great errors. Even the most responsive soils give low yields with inadequate management. Experience has shown that with skilled research and demonstration, gradual improvements in primitive indigenous soil management systems can be made. The very important research and experience of the Institut National pour 1’Etude Agronomique du Congo has recently been summarized and illustrated in considerable detail by Jurion and Henry (1 967). Settlers on well-selected newly developed arable soils in the Tropics need not go through all the steps between primitive and fully developed modern systems. The beginning level of soil management can be at whatever level the settlers have both the skills to handle and the industrial and agricultural services to support. b. Land Clearing. Up to now one factor holding back development of tropical soils with dense rain forest has been the great difficulty of clearing the land. Hand clearing is slow, arduous work. Clearing with bulldozers is expensive. They uproot the trees and leave the soil in poor condition. Recently new methods have been developed and demonstrated in heavy tropical forest that reduce the costs of clearing from around 750 to 1200 dollars per hectare to about 160 dollars per hectare, not including costs for road construction and transport to the development area. A A huge, curved, hard steel blade- the Rome blade- kept razor-sharp, is pushed by a heavy crawler-type tractor. Trees are sheared off at ground level (Fig. 19). The few huge trees are first split by a “stinger” on the blade (Fig. 20). The roots are left in the soil.

I64

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

Next the logs and brush are piled and burned (Fig. 2 1). In areas lacking dry seasons, oil must be used to start the burning. Then a special heavy cutting disk, with each disk held by a heavy spring

FIG. 19. Cutting tropical trees. (By courtesy of Rome Plow Co.)

so that it can “jump over” the stumps, breaks the soil and cuts the surface roots in place, partly to reduce “sprouting.” Except for any needed terraces or ditches for water control, the soil is ready for planting after two diskings. This method is a vital breakthrough that greatly reduces the high cost oT clearing and, at the same time, leaves the soil in excellent condition. In many places that are opened up with roads, valuable timber can be harvested and shipped out before the clearing. c . Planning for Growth of Farm Size. It has already been pointed out that before planning the details of a new area for farming, a highly detailed soil survey is necessary to be sure that all the soils to be cleared are responsive to management. It should be added here that the big trees near the margins of clearings may need to be taken out since they are likely to

POTENTIALLY ARABLE SOILS OF THE WORLD

165

fall in storms and damage homes or crops. A considerable part of the soils suitable for farming can be left uncleared, except for the removal of huge trees near the margins, so that as the community develops and cultivators

FIG. 20. Splitting large tree with "stinger" on the blade. (By courtesy of Rome Plow Co.)

increase their skills, holdings can be enlarged. In the homesteading in the United States the error was made of granting small tracts of land in such a way that increases in size of holdings could be made only by purchase from other settlers. d. Storage and Processing. Storage and processing of farm products require early attention. In warm humid climates many food crops, including maize and other grains, are subject to serious deterioration after harvest unless dried and carefully stored. Refrigeration is very helpful indeed. In the southern part of the United States, refrigeration gave an enormous boost to the production of livestock and to certain food crops. The advantage is even greater in the humid Tropics. Weeds are likely to be a hazard in moist tropical areas within a short

166

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

time after clearing and cultivation. The seeds are not subject to freezing or drying and, once started, weeds are difficult to control without modem herbicides.

FIG.21. Piling logs and brush. (By courtesy of Rome Plow Co.)

3 . Soil Exploration for New Settlement Areas Large areas of uniform soil are rare. In order to make orderly and rapid progress without waste, soil surveys are needed. First, relatively smallscale maps of about 1 :500,000can be constructed, to be followed by detailed surveys of the likely places. With the help of a good library, a schematic exploratory soil map can be made by highly skilled soil scientists from study of existing data on soils and related factors. Each kind of soil in the world results from a unique combination of factors as expressed in the following equation: Soil = J (1) climate, (2) plants and related animals, (3) parent earthy material, (4)relief, and ( 5 ) age of land form.

POTENTIALLY ARABLE SOILS OF T H E WORLD

I67

The first two of these are the active factors of soil formation. They act on the earthy material. The processes are conditioned by the relief. And the full development of the soil takes considerable time. With reasonable estimates of these five soil-forming factors and good knowledge of their interrelationships, the kinds of soil can be estimated. With reasonably good planimetric maps, the soil boundaries can be sketched roughly in relation to streams, roads, and other local features. Aerial photographs can also be useful, especially with stereoscopic pairs, although such photographs are less helpful in heavy tropical rain forest than in most other landscapes. In a second step, a reconnaissance soil survey is made in the field. With the schematic-exploratory map, field tours can be planned for soil examinations in representative areas. At least part of the field examinations should be supplemented with laboratory determinations of exchangeable cations, total and available phosphorus, and other physical and chemical characteristics important to the response of the soils to management. After considerable “scouting” to examine the most significant major kinds of soil and to describe and name them within some standard system of soil classification, lines can be laid to examine the soils in more detail. Lines for traverse can be established at precise locations on the map. In heavy bush, of course, these need to be cleared. Preferably they should be laid out at approximately regular intervals. From the observations made along the traverses, soil boundaries can be noted as they are crossed and additional soil descriptions prepared. A useful reconnaissance map was made recently in this way in a heavy rain forest of the southwest region of the Ivory Coast (Carroll et al., 1967), based partly on methods developed by the late C. F. Charter in Ghana (1949). With such a reconnaissance soil map at a scale between I : 100,000 and 1 :200,000,areas suitable for settlement can be plotted along with tentative locations for access roads, bridges, airstrips, and the like. The final step is to make detailed soil surveys of these selected places, preferably at a scale of around 1 :20,000. These soil maps should be suitable for operational planning of roads, boundaries of individual holdings, local water-control structures, and the like. In almost any area some individual spots should be avoided, at least at the start. Then too, some areas of suitable soil need to be set aside, unallocated, so that the community can expand later if this turns out to be desirable. The kinds of soil shown on this detailed soil map should be interpreted in terms of the probable yields of suited crops under alternative systems of management as defined in physical terms. Such interpretations can be had by examining results on similar kinds of soil elsewhere in the world where growing conditions are approximately the same.

168

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

4 . What Can the New Settlers Do? Equally important as careful examinations of the soils are detailed appraisals of the skills, goals, and potentials of the prospective settlers. Otherwise projected housing and management systems may be inappropriate for them. The soil management systems must include, after clearing, proper fertilizers, suitable varieties of crops, water control, and practices to control diseases, insects, and weeds, and, perhaps, protection from animals. Full advantage should be taken of the skills of the settlers, but if systems are set up beyond these skills, as modified by reasonable additional training and demonstration, they may fail. Then too, people vary somewhat in the way they like to live. As Nash (1966) points out, some may like to live in small villages; others may want to have their homes on their individual holdings. Other contrasting preferences can be expected.

5 . Communication Communication between those giving technical assistance and the cultivators is critical to success. In addition to excellence in diagnosing potentials of soils and technical problems of their use, a scientist must be able to make suggestions that fit the cultural ways of the people. This means he must have studied the culture of the people. Even in one area, great social and communication gaps may exist between the cultivators and the professional and administrative people of the local government. A scientist needs the confidence of both. Several gaps in communication need to be reco,,iized and overcome: (1) Snow (1959) discusses those between natural scientists and social scientists and between scientists and humanists within any one established culture; (2) those between any two established cultures, such as Arabian and Western; and (3) McLuhan (1962) illustrates these between a modern established culture and a primitive illiterate society. LevyBruhl (1 966) shows that the thought processes of primitive people differ from those of people in modern cultural systems. Clearly successful technical assistance requires a broad understanding of the people to be served and of their goals, hopes, and fears. Most people want to improve their own culture, not to change to another system. Anything that tears down their existing culture helter-skelter defeats the purpose of assistance. Besides technical skills, the scientist needs foresight to see what is possible ahead. He needs the skills of communication and patience. Perhaps most of all, he needs the gift of appreciation of other cultures. Simple tolerance is not enough. Nearly all progress is best achieved by starting with people where they

POTENTIALLY ARABLE SOILS OF T H E W O R L D

I69

are and by gradually expanding their opportunities, skills, and education. If a large and suggen change is attempted, the society may collapse rather than progress confidently. Guy J . Pauker (1968) has published an interesting study of Indonesia that shows the relationships among traditional farming methods, political instability, security of persons, and efforts to introduce new combinations of practices that could increase production substantially. Some changes are required in many societies if poor cultivators are to achieve better living standards; and there may be conflicts between the religious views and customs of the society and the ordinary application of modern science and technology to improve production. Certainly the changes suggested in the beginning should be limited to those that are essential for making progress toward more rewarding farming and agricultural systems. 6 . Institutional Development

Institutional development for appropriate local, provincial, and national services for education, health, marketing, credit, advisory services, and the like must be well thought out before new settlement and revised as the settlements grow. Some of the principal needs have been mentioned in broad terms already. The details of institutional development extend beyond the scope of this paper. REFERENCES Alexander, L. T.. and Cady, J. G . 1962. “Genesis and Hardening of Laterite in Soils.” U S . D e p t . Agr. Tech. Bull. 1282. Baker, 0. E. 1923. Geogruph. Rev. 13,l-26. Bauer, P. T. 1965. “Economic Analysis and Policy in Underdeveloped Countries.” Routledge & Kegan Paul, London. Binns, B. 0. 1950. “‘[he Consolidation of Fragmented Agricultural Holdings.” Food Agr. Orgun. U . N . . F A 0 Agr. Studies 11. Bogue. D. J . 1967. In ”Alternatives for Balancing World Food Production Needs,” Chapter 5 . Iowa State Univ. Press. Ames. Iowa. Broekmeijer. M. W . J . M. 1966. “Fiction and Truth about ‘The Decade of Development.”’ Sijthoff, Leyden, Holland. Carroll, P. H., Malgren, R. C.. and party. 1967. “Soil Survey of the Southwest Region, The Republic of the Ivory Coast.” 2 vols. Development and Resources Corp., New York. Charter, C. F. 1949. ”Methods of Soil Survey in Use in the Gold Coast.” Bull. Agr. Congo Belge 40, 109- 120. Ignatieff, V.. and Page, H. J . 1958. “Efficient Use of Fertilizers.” F o o d A g r . Orgun. U . N . , F A 0 Agr. Studies 43. International Institute for Land Reclamation and Improvement. 1960. “A Priority Scheme for Dutch Land Consolidation Projects.” Wageningen. Holland. Jacoby. Erich H. 1959. “Land Consolidation in Europe.“ Intern. Inst. for Land Reclamation and Improvement, Wageningen, Holland.

170

CHARLES E. KELLOGG A N D ARNOLD C. ORVEDAL

Johnson, Harry G. 1967. “Economic Policies Toward Less Developed Countries.” The Brookings Institution, Washington, D.C. Jurion, F., and Henry, J. 1967. “De I’agriculture itinerante a I’agriculture intensifiee.” I.N.E.A.C., Brussels. Kellogg, Charles E. 1950. Trans. 4th Intern. Congr. Soil Sci., Amster,dam, 1950, Vol. I , pp. 266-276. Roy. Trop. Inst., Amsterdam. Kellogg, Charles E. I962a. “Interactions in Agricultural Development.” In “Science, Technology, and Development- U.S. Papers Prepared for the U .N. Conference on the Application of Science and Technology for the Benefit of the Less-Developed Areas, Geneva, 1963,” Vol. 111. Agriculture, pp. 12-24. U.S. Govt. Printing Office, Washington, D.C. Kellogg, Charles E. 1962b. Soil Sci. 95, 221-230. Kellogg, Charles E. 1964. Yearbook Agr. ( U . S . Dept. Agr.) p. 57-68. Kellogg, Charles E. 1965. Proc. Agr. Res. fnsr., 1965 pp. 135-149. Natl. Res. Council, Washington, D.C. LeMare, P. H. 1968. J . Agr. Sci. 10,265-297. Levy-Bruhl, L. 1966. “How Natives Think.” Washington Square, New York. McLuhan, M. 1962. “The Gutenberg Galaxy.” Routledge & Kegan Paul, London. Martin, Kirk, and Knapp, John, eds. 1967. “The Teaching of Development Economics.” Cass. London Moseman, A. H., ed. 1964. “Agricultural Sciences for the Developing Nations.” Am. Assoc. Advance Sci., Washington, D.C. Myrdal, G. 1968. “Asian Drama, an Inquiry into the Poverty of Nations,” 3 vols. Random House, New York. Nash, M. 1966. “Primitive and Peasant Economic Systems.” Chandler, San Francisco, California. . Nye, P. H., and Greenland, D. J. 1960. Commonwealth Agr. Bur. (GI.Brit.), Tech. Commun. 51,l-153. Pauker, G. J. 1968. “Political Consequences of Rural Development Programs in Indonesia.” The Rand Corporation. Penny, D. H. 1966. “The Economics of Peasant Agriculture: The Indonesian Case,” Bull. of Indonesian Economic Studies. Australian Natl. Univ., Canberra. President’s Science Advisory Committee. 1967. “The World Food Problem,” Vols. 1 and I t . The White House, Washington, D.C. Schultz, T. W. 1964. “Transforming Traditional Agriculture.” Yale Univ. Press, New Haven, Connecticut. Snow, C. P. 1959. “Two Cultures.” Macmillan, New York. Soil Survey Staff.1960. “Soil Classification: A Comprehensive System.” U.S. Dept. Agr., Washington, D.C. Soil Survey Staff. (1967). “Supplement to Soil Classification System.” U S . Dept. Agr., Washington, D.C. United Nations. 1963. “Science and Technology for Development: Report of the United Nations Conference on the Application of Science and Technology for the Benefit of the Less Developed Areas,” Vol. 111, Agriculture. United Nations, New York.

GENETIC MODIFICATION OF PROTEIN QUALITY IN PLANTS Oliver E. Nelson Purdue University, Lafayetie, Indiana*

Page

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

1. Introduction 11. The Genetic

IV. V. VI. VII.

The Storage Proteins of Seeds ............................................................ The opaque-2 andjoury-2 Mutations in Maize The Prospects of Improvements in Other Plants .. Summary .......................................................................................... References .......................................................................................

I.

171

178

190

I9 I

Introduction

The rapidly increasing population of our world has caused serious examination of mankind’s ability to feed itself in the years to come (President’s Science Advisory Committee, 1967) Even today, in developing countries where sufficient calories are produced to nourish adequately the population, a segment of the population may be undernourished or malnourished owing to faulty distribution of available supplies or to unavailability of the proper types of food. This is true particularly for groups consuming small amounts of animal proteins for whom a large proportion of the protein as well as the calories in the diet come from a cereal grain. In such instances, the low nutritive quality of the cereal proteins coupled with the low protein content of most cereals may result in protein malnutrition especially in preschool children where protein needs are high in relation to caloric requirements. It has recently been realized that severe protein malnutrition resulting in the kwashiorkor syndrome in children may permanently affect development of mental functions (National Academy of Sciences - National Research Council, 1966; Champakam et al., 1968). The indications that permanent mental impairment may result from protein malnutrition have made it imperative that all possible means of making good quality proteins available as cheaply as possible be investigated. *Present address: University of Wisconsin, Madison, Wisconsin. 171

I72

OLIVER E. NELSON

Obvious steps toward this objective by increasing production of conventional sources of animal protein and exploiting more fully our marine fisheries are helpful, but such sources of protein are expensive for those most in need of it. Less conventional and less expensive means of making larger amounts of good quality protein available have been summarized (United Nations Economic and Social Council Report, 1967). These proposals include the improvement of protein quality in economically important plants by genetic means; the greater use of oil-seed meals as direct cources of protein in human diets: the production and marketing of fish protein concentrates (FPC) made from fish not otherwise commercially useful; intensified research to render the production of singlecell protein economical and the product palatable; and the use of synthetic amino acids or protein concentrates to bolster the nutritional value of cereals. The use of extractable leaf protein is still another alternative (Stahmann, 1968). All these proposals are receiving serious consideration, and it is probable that most will contribute to protein supplies for human consumption in the years to come. The proposal that will particularly interest most readers of this review is the genetic improvement of protein quality in crop plants. Such improvement where possible has the special advantage of being integral to a traditional food plant. Those requiring improved sources of protein need not be induced to adopt new foods, and no economic infrastructure for processing, fortification, and distribution is essential. Seed proteins are the major source of protein for man. Altschul (1965) has pointed out that the cereals produce more than 100 million tons of protein annually, and most of this is consumed by humans. Another 20 million tons is available from the legumes and oilseeds. The total annual production of animal protein reaches the same level (20 million tons). Howe et al. (1965) have indicated that when sufficient cereal is consumed to satisfy caloric requirements, the amount of protein ingested would be adequate if the quality were comparable to animal protein. This holds both for children and adults. It should be borne in mind that in considering the protein quality of a plant (seed, tuber, leaf, etc.), we are interested in the overall amino acid content of the total complex of plant proteins. Of special consequence is the content of the essential amino acids-those that humans or monogastric animals are unable to synthesize for themselves given a source of nonessential nitrogen. The essential amino acids must be included in the diet, either free or as components of the proteins consumed. For adult humans, the essential amino acids are lysine, tryptophan, phenylalanine, methionine, leucine, isoleucine, threonine, and valine. The

GENETIC M O D I FI C A T I O NOF PROTEIN QUALITY I N PLANTS

173

nutritive value of a protein or group of proteins depends on the content of the essential amino acids relative to the requirements for those amino acids of the species consuming the protein. The essential amino acid in shortest supply, relative to the requirement for it, becomes limiting for growth. Altschul ( 1965) has reviewed the role of amino acid composition as related to the nutritive value of proteins. Detailed discussions of the role of the amino acids in nutrition will be found in the review edited by Munro and Allison ( 1 964). In the past, attempts to find meaningful variation in protein quality (the content of the limiting amino acid) in any crop plant have been unsuccessful, and the protein composition of a species has been considered to be relatively immutable. Recent research has been established that large increases in the content of the limiting amino acids in maize seeds, lysine and tryptophan, are obtained in seeds homozygous for either the opaque-2 or thefloury-2 mutations (Mertz et al., 1964; Nelson et al., 1965). The basis for the changes effected by these mutations and the prospects of genetic improvement of other crop plants with respect to protein quality is the subject of this review. The general problem of increasing protein quality in plants has been reviewed recently by Becker (1963), Munck ( 1 964), and Schuphan ( I 966). II.

The Genetic Control of Protein Structure

A realistic assessment of the prospects of increasing the protein quality of any crop plant depends on an understanding of the genetic control of protein structure. The nature of this control and the mechanisms of transscribing genetic information and translating such information into the amino acid sequence of proteins have been elucidated over the past fifteen years. The information contained in a gene is a linear codescript consisting of a unique sequence of trinucleotide codons (DNA). The information is transcribed into a complementary sequence of trinucleotide codons of RNA (the messenger RNA) that is translated in turn into the amino acid sequence of a particular protein. Each codon specifies a particular amino acid in the protein sequence. The amino acid sequence (the primary sequence) of a particular protein in a given organism is invariate and suffices to explain the secondary structure (the helical configuration), the tertiary structure (the folding of the protein), and the quarternary structure (aggregation into polymeric units) of the protein. This picture has been derived from the combined and concerted efforts of biochemists, biophysicists, and geneticists. It represents a tremendous advance in our knowledge of biological systems. The pertinent literature is voluminous, but numerous reviews exist (Helinski

174

OLIVER E. NELSON

and Yanofsky, 1966; Holley, 1965; Korner, 1964; Mahler and Cordes, 1966; Schweet and Heintz, 1966). The 20 amino acids that are the usual components of protein are coded by trinucleotide codons. As there are 4 nucleotide components of DNA, there are 64 (43)possible codons. Three codons, UAG, UAA, and UGA, are nonsense triplets-do not code for any amino acid -in Escherichiu coli, the organism in which most of our information concerning codon assignments has been obtained (Garen, 1968). The remaining 61 possible triplets all code for amino acids. Since a given amino acid can be specified by more than one codon, the genetic code is said to be degenerate. Sadgopal ( 1968) has reviewed the codon assignments and the evidence that the code is universal, i.e., applies to all organisms. The transcription of the information coded in the DNA sequence of the gene is mediated by a DNA-dependent RNA polymerase that produces a linear RNA polymer complementary in base sequence to one of the two DNA strands of the helix (adenosine, uridine, guanosine, and cytidine are present in the RNA at the positions where thymidine, deoxyadenosine, deoxycytidine, and deoxyguanosine, respectively, are present in the DNA). Not all genetic sequences are transcribed into messenger RNA to be translated into protein. The RNA components of the ribosomes and the transfer RNA molecules are transcribed from the genetic information but apparently are not translated into protein sequences. They are functional RNA molecules. The messenger RNA’s bind to the protein-synthesizing organelles, the ribosomes. Concomitantly, free amino acids are activated and coupled to transfer RNA molecules by amino acid-activating enzymes that are specific both for the amino acid and the transfer RNA. A transfer RNA charged with its activated amino acid binds to a messenger RNA trinucleotide codon with a sequence complementary to the anticodon sequence of the transfer RNA, when the codon is in the proper position relative to the ribosome. The activated amino acid is then in proper position for the formation of a peptide bond between its amino group and the carboxyl group of the amino acid at the end of the growing polypeptide chain. The transfer RNA is then released from the ribosome. As the ribosome moves down the messenger RNA, each codon sequentially arrives in the proper position relative to the ribosome to bind the proper transfer RNA molecule and allow the formation of the peptide bond. A protein is thus synthesized amino acid by amino acid from its N-terminal end toward its C-terminal end. When the synthesis of the polypeptide chain is complete, the ribosome is at the end of the messenger RNA and drops off. If the proper transfer RNA with its activated amino acid is not

GENETIC MODIFICATION OF PROTEIN QUALITY I N PLANTS

175

available to match a particular codon during protein synthesis, synthesis of the polypeptide chain ceases at that point. Thus the deficiency of a particular amino acid may terminate the synthesis of various polypeptides at positions where that amino acid is specified by the code. Most of the evidence supporting this view of protein synthesis has come from experiments with microorganisms, but cell-free protein-synthesizing systems can be extracted from both plants and animals. The results from heterologous cell-free protein-synthesizing systems (ribosomes and message from one organism and transfer RN A’s from another) confirm the universality of the code. I t is probable that protein synthesis in all organisms follows the described course. There is little evidence to the contrary although it has been suggested that synthesis of the storage proteins of wheat endosperm is effected by another system since a protein body fraction obtained by relatively low speed centrifugation (50,000 g ) was found to incorporate amino acids (Morton and Raison, 1964). A comparable fraction from developing maize endosperms incorporated amino acids only when bacterial contamination was present (Wilson, 1966). Since the amino acid sequence of every protein synthesized by an organism is specified in its genetic information, conservation of a given set of amino acid sequences is strongly favored. A change at one position in the amino acid sequence of a protein can be effected by a point mutation as first shown by Ingram ( 1 957) for sickle cell hemoglobin in man. Many amino acid substitutions affect the functionality of the protein in which they occur, causing the mutation responsible for the substitution to be sublethal or lethal. Unless a substitution (mutation) offers a selective advantage, it is unlikely to increase in frequency except fortuitously in genetic isolates. The addition or deletion or a nucleotide from the DNA codescript that is the gene shifts the “reading frame” since there are no commas or spacers separating adjacent codons (Crick et al., 1961). Thus, from the point of addition or deletion, a whole new set of codons is generated resulting in a completely different amino acid sequence for the protein produced by the mutant. Such changes have a high probability of rendering the protein nonfunctional. A mutation usually affects the amino acid sequence of only one polypeptide chain. It can render the protein partially or completely nonfunctional, thus being lethal if the function of the protein is an essential one. Even if the mutation is not lethal, its effect on the overall amino acid composition of the protein coded by the locus is likely to be small, and its effect on the overall amino acid composition of the bulked proteins of the

176

OLIVER E. NELSON

organism will be negligible since a given protein usually constitutes only a small proportion by weight of the total proteins of the organism. There are exceptions to this generalization, notably the storage proteins of seeds which are to be discussed. Changes that might affect the amino acid sequence of' more than one protein can be envisioned. Assume that a transfer RNA recognizing one or more of the codons for a particular amino acid (e.g., lysine) were to be changed by a mutational event so that it recognized a glutamine codon, for example. Lysine would then be inserted during protein synthesis at some positions where glutamine was specified. Ritossa et ul. (1966) have shown that in Drosophila rnelunoguster there is a 13-fold redundance of DNA coding for transfer RNA. They point out that mutation to give a transfer RNA with an altered anticodon is unlikely to be an effective suppressor of a mutation producing an altered codon in a structural cistron because of the apparent large number of sites coding for transfer RNA. Thus this type of suppression operative in bacterial systems is unlikely in higher organisms if the transfer RNA redundancy found in Drosophila is typical. For the same reason, it is unlikely that an altered transfer RNA could appreciably affect the amino acid composition of the proteins being synthesized. In view of the limited possibilities of affecting overall amino acid composition by a mutation in a structural gene coding for a given protein and the further restriction that no change in sequence inactivating an essential protein is admissible, we can understand the mechanisms tending to conserve the sets of amino acid sequences (proteins) synthesized by an organism. In addition to the structural genes for proteins, there are also regulatory genes whose function is to determine when, and to what extent, a particular protein is to be synthesized. In a multicellular organism, each tissue or organ has its own unique spectrum of proteins appropriate to the metabolic functions of that organ. Within an organ, not all proteins are present in equal amounts, nor is the ratio of any two proteins invariate. It may vary with the stage of development, degree of nutrition, hormonal influences, etc. The synthesis of some proteins is controlled at the transcriptional level as was first suggested for the inducible enzyme, pgalactosidase, in Escherichiu coli (Jacob and Monod, 1961). Neither for microorganisms nor higher organisms do we understand to what extent the control of protein synthesis lies at the level of transcription and to what extent at the level of translation (Vogel and Vogel, 1967). The general question as to the conditions under which specific genetic information is transcribed or translated is of paramount importance in develop-

GENETIC M O D I F I C A T I O N O F PROTEIN QUALITY I N PLANTS

177

mental biology and the study of the regulation of metabolic functions. Instances are known in which the synthesis of a protein has been released by a mutation from the genetic controls ordinarily in force. Examples of this are the constitutive mutants of P-galactosidase and alkaline phosphatase of E . coli (Jacob and Monod, 1961; Echols et al., 1961). Conversely, mutations exist that severely restrict the amount of a protein synthesized without altering the structure of the protein [e.g., the thalassemia mutations affecting hemoglobin production in man (Nance, 1963; Zuckerkandl, 196411. The genetic control of protein synthesis has been reviewed briefly in order to provide a background for the discussion that follows. The amino acid sequence of each protein is genetically determined; viable mutations result most frequently in the substitution of one amino acid for another at a specified position in a polypeptide chain. The net effect is negligible in terms of the overall amino acid composition of the organism. Obviously, special circumstances are required in order to change drastically the overall amino acid composition of an organism or a tissue. It will be shown that such conditions probably exist only in the seeds of higher plants where large amounts of storage proteins that have no metabolic function are present. Ill.

The Relative Constancy of l e a f Protein Composition

The use of leaf protein concentrates as a supplemental source of protein is being intensively investigated (Stahmann, 1968). The overall amino acid balance of leaf proteins is generally good although the content of sulfur-containing amino acids (cysteine and methionine) is lower than desirable. Significantly, the amino acid content of bulked leaf proteins for a number of species has been found to be relatively constant (Lugg, 1949; Kelley and Baum, 1953; Stahmann, 1963; Gerloff el al., 1965). This relative constancy probably reflects the fact that most of the proteins in leaf tissues are metabolically active, a large proportion of the leaf protein being associated with the chloroplasts. Nonaqueously isolated chloroplasts from both bean and tobacco leaves contain approximately 70 percent of the total nitrogen of the leaves (Stocking and Ongun, 1962). The necessity that amino acid substitutions caused by mutations not disrupt the function of an essential enzyme or otherwise metabolically active protein is a strongly conservative factor limiting the possible changes in sequence. A protein with a given function can undergo changes in sequence through geological time owing to mutational events, and these changes may be fixed in newly evolving species. Nevertheless, proteins

178

OLIVER E. NELSON

carrying out the same biochemical function in species as unlike as man and baker’s yeast may still show considerable similarity in amino acid sequence as shown for cytochrome c (Margoliash and Smith, 1965). Where the evolutionary relationship is closer, as for man and the macaque monkey, the cytochrome c’s differ by a single amino acid residue. Much the same relationship is shown for the hemoglobins since the hemoglobins of man and the gorilla differ by only a single residue (Buettner-Janusch and Hill, 1965). The pressures of natural selection obviously tend to conserve the sequences of functional proteins. In tissues or organs where the majority of the proteins are functional as in leaf tissues, there appears to be small probability of significantly altering the overall amino acid composition. IV.

The Storage Proteins of Seeds

In contrast to leaf tissue, seeds may contain a considerable proportion of their protein as reserve or storage proteins that serve as a source of nitrogen for the germinating embryo. Since a plant is capable of synthesizing all the amino acids given a source of nitrogen, there is no necessity that the storage proteins have any particular amino acid composition. They can vary considerably from species to species within a family. The subject of seed proteins has been extensively reviewed (Brohult and Sandegren, 1954; Danielson, 1956; Altschul er al., 1966). In this review, we will consider principally the cereal proteins that furnish a major portion of the protein consumed in large areas of the world and the legume proteins because of their present and potential value as protein supplements. T. B. Osborne initiated systematic studies of seed proteins in 1891 and continued these studies until the 1920’s (Osborne, 1924). His classification of seed proteins according to their solubility is still used since no functional criteria can be applied. Seed proteins are classified as albumins (water soluble), globulins (insoluble in water but soluble in saline solutions), prolamines (soluble in relatively strong alcohol), and glutelins (soluble in dilute alkaline solutions but not in water, saline, or alcoholic solutions). Another method of extracting cereal proteins with an alkaline medium containing copper, sulfate, and sulfite ions has been employed by Mertz and Bressani (1 957). The extract can be fractionated to yield acid-, alcohol-, and alkali-soluble fractions. Only in seeds of the Gramineae are large quantities of prolamines and glutelins found. In seeds of other species, globulins constitute the bulk of the storage proteins (Osborne, 1924; C. R. Smith et af., 1959). The protein constituents of cereal grains have been intensively investi-

GENETIC M O DIFICAT IO NOF PROTEIN QUALITY I N PLANTS

179

gated. Mosse ( 1968) has recently reviewed the subject. The prolamines constitute a major percentage of the seed protein in wheat, Triticum vulgare (45 percent); rye, Secale cereale (40 percent); barley, Hordeum vulgare (40 percent): maize, Zea mays (50 percent); sorghum, Sorghum vulgare (60 percent): and millet, Panicum miliaceum (60 percent). Among the commonly grown cereals, only rice, Oryza sativa, with ca. 8 percent of the seed protein being alcohol-soluble and oats, Avena sativa, with I2 percent alcohol-soluble protein do not possess a major prolamine fraction. The prolamine fractions from all cereals are relatively rich in proline and glutamine, accounting for the name coined by Osborne for these proteins. They are all low in lysine content with zein (the prolamine fraction of maize) containing only 0. I g./100 g. of protein. The prolamine fractions of wheat, rye, and barley all contain larger percentages of lysine than does zein, but in no prolamine fraction is the lysine content high (Mosse, 1 966). In early investigations in Osborne’s laboratory, neither tryptophan nor lysine could be detected in zein (Osborne and Harris, 1903; Osborne and Leavenworth, I9 13). Osborne and Mendel (1 9 14, 19 16) showed that rats of all ages went into a rapid decline and eventually died if placed on a diet in which zein was the sole source of protein. If both typtophan and lysine (0.5 percent) were added to the diet, the rats grew at nearly normal rates. These early chemical and nutritional experiments with zein exemplify the problem of protein quality in wheat, rye, barley, maize, sorghum, and millet. Approximately half of the protein synthesized in these seeds is the prolamine fraction. The prolamines have a negligible to low content of lysine, which is an essential amino acid. In addition, another essential amino acid may be lacking as tryptophan is in zein. Thus, when the overall amino acid compositions of the total proteins of the seeds from these species are considered, all have low contents of lysine. Howe et al. ( 1 965) have shown in trials with weanling rats that lysine is the first limiting amino acid of rice, wheat, millet, barley, sorghum, rye, and oats. For corn, lysine and tryptophan are colimiting as might be expected from the above discussion. The additions of sufficient quantities of lysine plus the second limiting amino acid to these cereals improves the quality of the proteins so that they are equivalent to the milk protein, casein. For. all these cereals except maize, threonine appears to be the second limiting amino acid. Without supplementation, oats have the highest protein quality. The glutelin fraction of all the cereals tested has a considerably higher lysine content than the prolamine fraction. Mosse (1968) gives a lysine content of 3.2 g./IOO g. protein for corn glutelin and 3.2 g. for wheat glutelin. Jimenez (1966) also found a lysine content of 3.4 g. for corn glutelin.

180

OLIVER E. NELSON

The major proteins of the legumes are globulins. Altschul et al. ( 1966) have discussed the properties of globulin preparations from a

number of legumes. Some of these globulins, such as glycinin from the soybean, Glycine max, are not homogeneous although once considered to be so. The amino acid compositions of legume seed proteins are quite different from those of cereal proteins. Lysine is found in moderate to ample supply depending on the species. T h e sulfur-containing amino acid, methionine, has been shown to be the limiting amino acid in all legume proteins (Evans and Bandemer, 1967). In addition, most legumes have growth-inhibitory substances that are destroyed by heating. The trypsin inhibitor of soybeans is one example of such a growth-inhibitory substance. We are most interested in this review in the seed proteins of the cereals and the legumes, but with growing interest in protein sources for food and feed and with the availability of automatic amino acid analyzers, the amino acid compositions of many seed proteins have been reported. VanEtten and his colleagues ( I 96 1 , 1963, 1967) have indicated that some species (e.g., Daucus carota and Marah gilensis) produce proteins of good quality. Crumbe abyssinica and Lesquerella species do also, but the toxic thioglycosides present in seeds of the Cruciferae must be removed before their use as protein sources. White et al. (1955) have reported that the seed proteins of Chenopodium quinoa and C . pallidiculae, have excellent amino acid composition and a nutritive quality equal to milk protein for supporting rat growth. These species are grown for food in the Peruvian Andes. V. The opaque-2 and floury-2 Mututions in Muize

The opaque-2 ( 0 2 )and floury-2 (JE2) mutations in maize have been shown to effect drastic changes in the amino acid composition of seeds homozygous for either of the mutant genes (Mertz et al., 1964; Nelson et al., 1965). These mutations, which condition similar phenotypesfloury endosperms imparting a dull, opaque appearance to the kernel were isolated many years previously by Jones and Singleton (a)and Mumm (Jlz) (Emerson et al., 1935). The identification of the mutations’ biochemical effects came in the courze of the search for a mutation that would block the ability of the seed to form zein. Frey et al. ( 1 949) and Bressani and Mertz (1958) had recognized the desirability of reducing the amount of zein synthesized. It was hoped that if zein were not formed, larger amounts of the other fractions would be. These are considerably higher in lysine content than zein, thus resulting in an enhanced lysine content for the mutant seeds. The o2 andf12 mutants were two of

GENETIC MODIFICATION OF PROTEIN QUALITY IN PLANTS

181

the first four mutants to be analyzed, the other two being opaque-] and floury-I. I t had been hypothesized by the author that mutations giving the opaque phenotype were the most likely to be deficient in zein. The changes in amino acid composition in seeds or seed components effected by the mutations have been detailed (Nelson et al., 1965; Mertz et af.,1966; Nelson, 1968). Table I presents the amino acid comTABLE I Amino Acid Composition (Grams per 100 g. Protein) for Defatted Whole Kernels of opaque-2,floury-2, and Normal Maize- 1967 Crop Amino acid Lysine Tryptophan Histidine Arginine Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Cystine Valine Methionine lsoleucine Leucine Tyrosine Phenylalanine % Protein

or

+

5.0 1.3 3.5 7.2 8.8 3.8 4.7 17.1 8.4 s.l 6.7 2.0 5.2 2.2 3.4 9.3 4.2 4.4 10.5

3.0 0.7 2.6 4.9 9.2 4. I 5.6 22.6 9.6 4.7 9.2 1.7 s.7 1.3b 4.2 14.6 5.2 5.8 9.0

j?2

4.8

-

2.9 6.3 10.5 4. I 5.2 18.5 8.8 4.7 8.0 I .6 5.7 2.7 4.0 12.0 4.6 5.2 17.0

“ N o analysis available. Previous results show tryptophan values for floury-2 equal to those for opaque-2. *Lower than expected. Some may have been destroyed during hydrolysis.

position of proteins from whole seeds of normal maize, and seeds homozygous for either the oy or theJ2 mutation. The contents of lysine, tryptophan, and arginine are higher in the mutant stocks than in normal while the contents of glutamate (glutamic acid glutamine), alanine, leucine, tyrosine, and phenylalanine are lower in the mutants than in normal. In addition, the content of glycine is higher in 0 2 .The content of methionine in j&is higher than in normal or of. The greater nutritive value of the mutants compared to normal is largely due to the increased amounts of

+

182

OLIVER E. NELSON

lysine and tryptophan present. The reduced quantity of leucine in the mutants may also be useful since large quantities of leucine antagonize the use of isoleucine and valine by rats (De Muelenaere et al., 1967). It can be shown that the differences of the mutants from normal in amino acid composition on a whole-seed basis can be attributed entirely to differences in the storage proteins of the endosperm. Table I1 presents TABLE I1 Amino Acid Composition (Grams per 100 g. Protein) for Defatted Endosperms of opaque-2, floury-2, and Normal Maize" Amino acid

W64A+

W64A02

x

Lysine Tryptophan Histidine Arginine Aspartic acid Glutamic acid Threonine Serine Proline Glycine Alanine Valine Cystine Methionine Isoleucine Leucine Tyrosine Phenylalanine % Protein

1.6 0.3 2.9 3.4 7.0 26.0 3.5 5.6 8.6 3.0 10. I 5.4 1.8 2.0 4.5 18.8 5.3 6.5 12.7

3.7 0.7 3.2 5.2 10.8 19.8 3.7 4.8 8.6 4.7 7.2 5.3 0.9 * 1.8 3.9 11.6 3.9 4.9 11.1

3.3 0.8 2.2 4.5 8. I 19.1 3.3 4.8 8.3 3.7 8.0 5.2 1.8 3.2 4.0 13.3 4.5 5.1 13.6

"Data from Nelson er al. (3965) Science 150, 1469-1470. Copyright 0 1965 by the American Association for the Advancement of Science. *Other analysis of opaque-2 stocks have given cystine values equal to or greater than normal.

the amino acid compositions of nornal and mutant endosperms. The amino acids whose quantities are changed in the mutants on a whole seed basis are also found to be altered in amount in the endosperms with the relative differences between the rnutJds and normal being more pronounced. Table 111 gives the amino acid compositions of embryos from normal or opaque-2 seeds. There is little indication that the amino acid composition is different in the normal and the mutant embryos. The amino acid compositions of leaves and pollen from normal and opaque-2 plants have also been found to be similar (Nelson, 1968). Clearly, the differences

183

GENETIC MODIFICATION OF PROTEIN QUALITY IN PLANTS

in amino acid content between normal and the opaque2 mutant are confined to the endosperm proteins. Since opaque-2 maize seeds have greatly increased quantities of lysine and tryptophan, the two limiting amino acids in maize, one would expect much increased nutritive value for o2 protein as opposed to the protein of common maize. This expectation has been realized in a number of TABLE 111 Amino Acid Composition (Grams per 100 g. Protein) for Defatted Embryos of opaque-2 and Normal Maize" Amino acid

opaque-2

Lysine Histidine Ammonia Arginine Aspartic acid Glutamic acid Threonine Serine Proline Glycine Alanine Valine Cystine Methionine lsoleucine Leucine Tyrosine Phen ylalanine

5.9 2.9 2.1 9.2 9.2 13.9 3.1 5.0 5.3 5.5 5.8 4.4 0.9 I .5 2.5 5.6 2.2 3.6

Normal 6.1

2.9 2.2 9.1 8.2 13.1 3.9 5.5 4.8 5.4 6.0 5.3 I .o 1.7 3. I 6.5 2.9 4. I

"Data from Mertz et al. (1966) Advan. Chem. Ser. 57, 228-242.

experiments with 02 maize. Mertz er at. (1965) reported that weanling rats on a diet where or maize was the sole source of protein (10 percent) grew 3.5 times as rapidly as the weanling rats on a diet where common maize supplied the same level of protein. A third group of rats on a diet that supplied 10 percent protein as soybean oil meal grew at the same rate as did the group fed on 0.) maize. Similar estimates of the nutritive superiority of o2 maize in supporting the growth of weanling rats have come from other laboratories (Mertz, 1966: Piva el al., 1967). Experiments with weanling pigs have shown that o2 maize supports a growth rate 3.5 times more rapid than does common maize (Pickett, 1966). Cromwell et a f . ( 1 967) reported that adding lysine to common maize did not increase the growth rate of weanling pigs: the addition of

184

OLIVER E. NELSON

tryptophan did increase growth rate somewhat. The addition of both lysine and tryptophan (to the level present in o2maize) increased growth rate to a rate slightly below that of pigs fed on o2 maize. In a free-choice preference test between o2 maize and common maize supplemented with soybean meal to the same protein level as the o2 maize, young pigs ate about 85 percent o2 maize and 15 percent corn-soy mixture. Finishing pigs (ca. 130 pounds at the start of the experiment) fed on o2 maize (1 1.2 percent protein) gained weight approximately as rapidly as those fed on a maize-soy mixture ( 1 3 percent protein) and significantly more rapidly than those fed on common corn (8.9 percent protein) or common maize soy ( 1 1.2 percent protein) (Pickett, 1966). The nutritional value of o2 maize has been tested in humans. Bressani ( 1966) in nitrogen balance experiments with Guatemalan children demonstrated that o2 maize protein had a nutritive value as high as that of milk protein when fed a t protein levels of 1.8-2.0 g./kg./day. Children fed at the same protein level on common maize were in negative nitrogen balance (i.e., were excreting more nitrogen than was supplied in their diets) indicating the poor quality of common maize protein. Harpstead et al. ( 1 968) and Pradilla et al. ( 1 968) have reported the recovery of children from kwashiorkor on diets in which part of the nitrogen was supplied by o2 protein and the remainder by nonessential sources of nitrogen such as glycine and diammonium citrate. Clark et al. (1967) in tests with adults found that 5 of 6 subjects tested were in nitrogen balance when consuming 250 to 300 g. of o2 maize daily. For the sixth subject, a large man, 350 g. of o2 corn were needed to establish equilibrium. The investigations cited have all been carried out with meal made from the whole seeds of o2 or common maize. When the endosperm meal alone was used in feeding trials with weanling rats, those fed on o2 endosperm meal (7.5 percent protein) gained an average of 34.7 g. in 28 days while those fed on normal endosperm meal at the same protein level gained an average of 1.4 g. (Wichser, 1966). The much greater difference in nutritive value between the endosperm meals of o2 and common maize than noted with whole seed meals might have been predicted from Tables I, 11, and 111, which show that all the differences in amino acid composition between o2 and normal maize noted in whole seeds are accounted for by the differences in endosperm composition. The addition of the germs tends to bring the composition of the whole seeds of the two genotypes closer together. Clearly, the use of 0 2 maize will have more impact and be more valuable relative to common maize in those maize eating cultures where the maize is degermed before cooking. Discarding the germs in the preparation of meal markedly reduces the protein quality, however.

+

GENETIC MODIFICATION O F PROTEIN QUALITY I N PLANTS

185

The nutritive value off12 maize has not been as thoroughly investigated as has that of o2 maize. Veron ( 1967) found that when fed at a level of 10 percent protein,fle maize was superior to common maize but not equal to oe maize for weanling rats. However, Cromwell ef al. ( I 968) reported that in soybean meal-supplemented diets for young chickens 3. was superior to either normal or or maize. The opaque phenotype characteristic of the or andflz mutants is a definite asset in breeding programs since it eliminates considerable chemical analysis. The transfer of either mutation into a desired background can be followed by the opaque phenotype with the assurance that improved nutritive value is also being transferred. In genotypes conditioning the production of maize with a higher percentage of protein, the or mutation still effects its characteristic changes in amino acid composition (Nelson, 1966). This is most important since o2 maize with 10 percent protein does not support optimal growth for children or young pigs, but o2 maize with 14 or 15 percent protein should do so. Such protein levels have been attained. If such selections are equal in yield to o2 maize of lower protein content, they will be most valuable since no protein supplementation would be needed. The Jz, mutation is especially interesting in this connection since the mutation itself results in higher protein production (Nelson, 1968). It may be possible to attain the desired protein levels in many f12 stocks without additional selection. The effects of the or andj& mutations on the production of the endosperm proteins has been investigated by Jimenez ( 1966, I968), Mosse ( I966), and Mosse et al. ( 1966). The results from the two laboratories are are not directly comparable since Mosse and his collaborators extracted their protein fractions from entire seeds. Jimenez extracted the proteins from the endosperms alone. Therefore, Mosse et al. ( 1966) found a larger amount of albumins plus globulins in all genotypes than Jimenez owing to the inclusion of the germ (embryo and scutellum) proteins which largely fall into these classes. This accounts also for the differences found in the amino acid compositions of the water- and salt-soluble fractions by the two groups. They agree on the pronounced reduction in the amount of alcohol-soluble proteins synthesized by the two mutants. Additionally, Jimenez found a large increase in the amount of the endosperm albumins and globulins synthesized by the mutants even though these classes of proteins make up a relatively small proportion of the total endosperm protein. In view of the great reduction in zein synthesis, it is interesting that no difference exists in the amino acid compositions in the zeins from normal and o2 maize. Starch gel urea electrophoresis of the zeins does not

186

OLIVER E. NELSON

reveal any bands in o2 that are not present in normal (Mosse, 1966; Jimenez, 1968). Jimenez found that 3 of the 13 bands found in normal were absent in o2 zein preparations. The starch gel urea electrophoresis patterns of normal and onalbumin and globulin fractions were identical as were the amino acid compositions of these fractions. Both groups find a slightly different amino acid composition for the glutelins from the mutant stocks as compared to normal with higher lysine contents in the mutant preparations. It is the conclusion of both Mosse and Jimenez that the the change in the relative proportions of the solubility fractions (less alcohol-soluble, more water-, salt-, and alkali-soluble proteins) is the factor contributing most significantly to the change in amino acid composition of the mutant seeds. The changes effected by the f l n mutation are qualitative as well as quantitative. Jimenez (1968) finds one band of the 22 present in the globulin fraction and one of the 13 present in the zein fraction to be altered in electrophoretic mobility on starch gel urea electrophoresis, although the possibility that these differences may be conditioned by different genetic backgrounds and not by the flz mutation cannot be rigorously excluded. In our initial investigations, we were searching for mutations that would block the ability of the endosperm to synthesize the alcohol-soluble fraction. The two mutations, opaque-2 and floury-2, that were identified as affecting the synthesis of this fraction, do not completely block synthesis. In the mutant endosperms, some zein is still synthesized, and this zein corresponds to that synthesized by normal endosperms on the basis of amino acid composition and electrophoretic mobility (excepting one band inJlz). It is a reasonable assumption that the normal allele at the on locus is concerned with the regulation of zein synthesis. Apparently no proteins are altered, but the proteins of the prolamine fraction (apart from the 3 missing bands) are present in much reduced quantities. The increase in synthesis of the other fractions would be secondary to the reduction in zein synthesis which is suggested as the primary effect of the o2mutation. The increase in the components of the other solubility classes is not uniform. Jimenez ( 1968) has shown, for example, that the increase in the albumin fraction of 02 is largely an increase in one protein. There is no definitive evidence that the primary effect of the o2mutation is on zein synthesis. It is possible but less likely that the enhancement of the other fractions is primary, and the reduction in zein synthesis is secondary. The observations that the total increase in the other fractions does not equal the decrease in zein and that the quantity of free amino acids is higher in o2 suggests that the reduction in zein synthesis is pri-

GENETIC MODIFICATION OF PROTEIN QUALITY IN PLANTS

187

mary. The embryos in 0 2 seeds are larger than in normal controls and in some lines may contain more protein (Bauman, unpublished data). The effect of the& mutation may be more complex than that of 0 2 .The synthesis of zein is suppressed almost as much as with og In addition, several proteins with altered mobilities are produced. The mode of inheritance is semidominant and each successive dose o f j & in the endosperm adds an additional increment of lysine (Bates, 1966). VI.

The Prospects of Improvements in Other Plants

Since the amino acid composition of maize can be drastically improved by mutations that act to suppress the synthesis of the nutritionally poor prolamine fraction, the prospect of improving other cereal grains by the detection of homologous mutations exists if the grain contains a sizable prolamine fraction. Only rice and oats among commonly grown cereals do not. These two species rank at the top of the cereal grains with respect to lysine content expressed as grams per 100 g. of protein. For these two species, mutations homologous to o2 or fr, would have relatively little effect on amino acid composition. For the other cereal grains, such mutations would be expected to effect considerable improvement. In polyploid species such as the hexaploid Triticum vulgare, one of the genomes could contain a recessive gene homologous to o2 without any effect on amino acid composition. The difficulty of obtaining such a mutation in all the genomes is formidable. On the other hand, the presence of a semidominant mutation such asJ2 in one of the genomes would effect a change in amino acid composition and should be identifiable. Such genotypes would then constitute the starting point for a second cycle of nutritional improvement. With barley and sorghum, mutations leading to larger quantities of the limiting amino acid should be detectable. Recent reports of lines containing larger quantities of lysine have been made for each species. Virupaksha and Sastry (1968) have reported a wide range of lysine content in the whole seeds of various accessions of sorghum. One accession, Cernum 160, had a lysine content of 3.14 g. per 100 g. of protein, which is high for sorghum. This variety had a low pro1amine:glutelin ratio, so the change in composition here is brought about by a change in the relative proportions of the storage proteins. Only one sample of the variety was analyzed, and there is no report of inheritance studies with this variety. Hagberg and Karlsson ( 1968) have reported a barley accession that contains 4.6 g. of lysine per 100 g. of protein, which is above the usual levels for barley - 3.6 g. per 100 g. of protein (Mosse, 1968; Eggum, 1968). Again, this represents the analysis of a single sample, and it is not

188

OLIVER E. NELSON

yet known whether the accession grown in another year would again have a high lysine content. Assuming that one could detect mutations that suppress prolamine synthesis and significantly change the prolamine: glutelin ratio in any of the cereals rich in alcohol-soluble proteins, the results might not be as striking as with the o1 and j12 mutations in maize as Mertz (1969) has pointed out. The prolamines of cereals can be divided into 3 groups on the basis of their amino acid composition: the first group contains the gliadin of wheat, hordein of barley, and secalin of rye; the second contains the zein of maize, panicin of millet; the third contains the avenin of oats (Mosse, 1968). The prolamines of the second group contain very low amounts of lysine. To this group can be added kafirin of sorghum with a lysine content of less than 0.2 g. per I00 g. of protein. The consequence of suppressing prolamine synthesis with consequent compensatory synthesis of other protein fractions would be expected to be greater in terms of lysine content for maize, millet, and sorghum than for cereals of the first group, where the lysine content of the prolamine fraction is ca. 1 g. per 100 g. of protein. The possibility of raising the methionine content in the proteins of the legumes deserves careful study since methionine is the limiting amino acid for all legumes. For obvious reasons, the proteins of the soybean seed have been intensively investigated. As in the other leguminous seeds, the bulk of the seed protein is globulin in nature and was once thought to be a homogeneous protein (glycinin). Ultracentrifugal studies have revealed that 4 components (2, 7, 1 1 , and 15 S) are present (Naismith, 1955). Wolf and Sly ( 1 967) have shown that other methods of fractionation will separate the components. Roberts and Briggs ( 1 965) reported that the 7 S component that comprises 30 percent of the total protein has an extremely low methionine content-0.19 g. per 100 g. of protein. For comparison, the entire globulin fraction has a methionine content of 1.4 g. per 100 g. of protein. The 7 S fraction also differs appreciably from the total in its content of threonine and glycine. If the synthesis of the 7 S fraction could be blocked or suppressed genetically, and compensatory synthesis of the other fractions resulted, the methionine content would be raised substantially. Wolf et al. (1961) have reported that the relative proportions of the 7 S and 1 1 S fractions were quite different in Clark soybeans grown in Illinois and Hakuhou No. 1 soybeans grown in Japan. In this instance, it is not clear whether variety, location, or both are responsible. The possibility that lines with a low quantity of the 7 S fraction exist, or could be induced, should be investigated. In this review, the principal concern has been the enhancement of

GENETIC MODIFICATION OF PROTEIN QUALITY IN PLANTS

189

biological value of cereal and legume seed proteins. It appears that this is most likely to be achieved by altering the relative proportions of storage proteins that have different amino acid compositions, but changes in nutritional value may arise through other circumstances. In the cereals, the proteins of the germ are much superior to those of the endosperm in nutritive value. Tables I 1 and I11 demonstrate this point for maize. If a greater proportion of the protein were germ protein, the biological value of the protein would be increased. I t is possible that the differences in lysine content in different races of Mexican maize (Tell0 et al., 1965) may be explained by varying germ :endosperm ratios. Potato tubers from different varieties may have a 2-fold range in the content of such essential amino acids as lysine and methionine (Nehring and Schwerdtfeger, 1957). Reissig ( 1958) has shown that potato tubers have a substantial portion of their total nitrogen as nonprotein nitrogen (free amino acids). The true protein fraction as distinguished from crude protein (the total nitrogen content x 6.25) has a good amino acid balance. The proportion of nonprotein nitrogen varies in different varieties (40-54 percent). The content of essential amino acids expressed as an EAA index (Oser, 195 1 ) was much higher in the protein fraction (EAA index 83 to 89 percent) than in the nonprotein fraction (EAA index 31 to 43 percent). The protein content was highly correlated with the length of the growing season- the later the variety, the greater the percentage of nitrogen that was present in the protein fraction and the higher the EAA index. Within maturity groups, there were still differences between varieties as to the percentage of nitrogen present as true protein. The biological value of potatoes could be raised by selection for lines that could synthesize larger quantities of protein within a given maturity season. Toxic substances present in seeds can be important deterrents to the use of their proteins. Liener (1966) has reviewed the subject of both proteinaceous and nonproteinaceous substances in seeds that present problems because of their toxicity. The legumes as a group contain an array of antinutritional factors - trypsin inhibitors, hemagglutinins, and goitrogenic substances. Since these compounds can be destroyed by the proper heat treatments, no program to lower their concentration in the seeds seems justified. An exception is Lathyrus sativus, cultivated on 5 million acres in India. Consumption of this legume can result in permanent paralysis apparently caused by P-N-oxalyl-a,P-diaminopropionic acid. Although the toxic substance can be extracted by thorough cooking, the cooking water being discarded, or by soaking in cold water and steeping in hot water (Mohan er al., 1966), it would obviously be desirable to identify strains lacking the toxin or having a very low content. The use of meals remaining after oil is extracted from the seeds of a

190

OLIVER E. NELSON

number of cruciferous plants is limited by the presence of thioglycosides that are enzymatically hydrolyzed to yield goitrogenic isothiocyanates. As the hydrolytic enzymes are present in the meal, the meal can be moistened to enable hydrolysis to occur. The isothiocyanates can then be removed by steam distillation. Lines of Brassica carrzpestris much lower in thioglycosides have been identified (Josefsson and Appelqvist, 1969). Possibly such strains can be used in breeding programs to achieve varieties sufficiently low in thioglycosides to be used without hydrolysis and distillation. In order to utilize the proteins of cottonseed meal in animal or human diets, it is necessary to remove the toxic pigment gossypol by solvent extraction. Strains of cotton low in gossypol can be selected (Rhyne et al., 1959). Meal from these strains is equal in nutritive value to that of solvent-extracted commercial meal (F. H. Smith et al., 1961). Gossypol is eliminated altogether in genotypes lacking the pigment glands in which gossypol is produced (McMichael, 1960). If mutant plants produce fibers equal to normal plants in quality, the introduction of the mutation into all commercial varieties would make the use of unextracted cottonseed meal feasible. One other possibility deserves a brief mention. The cereal grains contain only low quantities of free amino acids. The production of any amino acid is evidently regulated to correspond closely to the demand for it in protein synthesis and other reactions. The mechanism(s) of such regulation has not been intensively investigated in higher plants, but much is known about the regulation of amino acid synthesis in microorganisms where one or more enzymatically mediated reactions may be key reactions from a regulatory standpoint. A mutation may cause the loss of sensitivity to the usual signals repressing enzyme synthesis or a loss of sensitivity by the enzyme to the usual factors restricting its activity (Sheppard, 1964; Calvo and Calvo, 1967). In either case, the effect could be an oversynthesis of a particular amino acid in terms of the requirements for protein synthesis and hence some quantity of that amino acid present as the free amino acid. No mutation of this type has ever been identified in-higher plants, but the possibility should be considered.

VII.

Summary

Serious attention should be given to the identification and utilization of mutant genes that raise the concentration of the limiting amino acids in both cereals and legumes that are important sources of protein for humans and livestock. The improvement of the nutritional quality of

GENETIC MODIFICATION OF PROTEIN QUALITY I N PLANTS

191

traditional foods has many advantages, particularly in developing nations where it may be difficult to reach large segments of the population with nutritional supplements. The factors tending to enforce a relatively uniform amino acid composition for a species have been emphasized as an essential background for those contemplating research in improving protein quality. Considering the restrictions on change in amino acid composition, there still exist opportunities to effect improvement in all the cereals where large quantities of alcohol-soluble proteins are synthesized. The possibility of improvement in the legumes appears less good, although a recent intriguing report of heterogeniety in methionine content in different fractions of soybean storage proteins may indicate that progress could also be made here. From theoretical considerations involving the genetic control of protein synthesis, it is probable that the most probable avenue to important changes in amino acid composition involves changing the relative proportions of metabolically inert storage proteins that have quite different amino acid compositions. This view has been reinforced by a study of the effects of the o2 and f12 mutations in maize. These mutations enhance markedly the nutritional value of maize seed proteins. Other possibilities of effectively changing the amino acid composition of cereal grains toward improved nutritional quality involve increasing the germ:endosperm ratio or mutations that relieve the constraints ordinarily regulating the amount of an essential amino acid synthesized. The supply of readily available plant protein may also be increased by the selection of strains lacking or low in toxic substances that must be destroyed or extracted before the seed proteins of cotton and many species of the Cruciferae can be utilized.

REFERENCES Altschul. A . M. 1965. “Proteins-Their Chemistry and Politics.” Basic Books. New York. Altschul, A. M., Yatsu, L. Y . , Ory, R. L., and Engleman, E. M. 1966. Ann. R e v . Plant Physiol. 17, I 13-1 36. Bates, L. S . 1966. Proc. High Lysine Corn Conf., Purdue Univ., 1966, pp. 6 1-66, Corn lnd. Res. Found., Washington, D.C. Becker. G . 1963. Zuechrer 33,3 13-322. Bressani. R . 1966. Proc. H i g h Lysine Corn Conj:. Purdue Univ., 1966, pp. 34-39. Corn Ind. Res. Found., Washington, D.C. Bressani, R., and Mertz, E. T . 1958. Cereal C h e m . 35,227-234. Brohult, S . , and Sandegren, E. 1954. In “The Proteins” ( H . Neurath and K. Bailey, eds.), Vol. 2, pp. 487-5 12. Academic Press, New York.

192

OLIVER E. NELSON

Buettner-Janusch, J., and Hill, R. L. 1965. Science 147,836-842. Calvo, R. A., and Calvo, J. M. 1967. Science 156, 1107-1 109. Champakam, S., Srikantia, S. G., and Gopalan, C. 1968. A m . J . Clin. Nutr. 21,844-852. Clark, H. F.. Allen, P. E., Meyers, S. M., Tuckett, S. E., and Yamamura, Y. 1967. A m . J. Clin. Nutr. 20,825-833. Crick, F. H. C., Barnett, L., Brenner, S., and Watts-Tobin, R. J. 1961. Nature 192, 12271232. Cromwell, G. L., Pickett, R. A,, and Beeson, W. M. 1967. J. Animal Sci. 26, 1325-133 I . Cromwell, G. L., Rogler, J. C., Featherston, J. R., and Cline, T. R. 1968. Poultry Sci. 47, 840-847. Danielson, C. E. 1956. Ann. Rev. Plant Physiol. 7, 2 15-236. De Muelenaere, H. J. H., Chen, M.-L., and Harper, A. E. 1967. J . Agr. Food Chem. 15, 3 10-3 I7 and 3 18-323. Echols, H., Garen, A,, Garen, S.,andTorriani, A. 1961.J. Mol. B i d . 3,475-438. Eggum, B. 1968. “Aminosyrekoncentration og Proteinkvalilet.” Stougaards Forlag, Copen hagen. Emerson, R. A., Beadle, G. W., and Fraser, A. C. 1935. Cornell Univ., Agr. Expt. Sta. Mem. 180. Evans, R. J., and Bandemer, S. L. 1967. J . Agr. Food Chem. 15,439-443. Frey, K. J., Brimhall, B., and Sprague, G. F. 1949.Agron.J. 41,399-403. Garen, A. 1968. Science 160,149-159. Gerloff, E. D., Lima, 1. H., and Stahmann, M. A. 1965. J . Agr. Food Chem. 13,139-143. Hagberg. A,. and Karlsson,. K.-E. 1969. Proc. IAEA-FA0 Panel, Riisturlga Sweden, I968 pp. 17-2 I . Intern. Atomic Energy Agency, Vienna. Harpstead, D. D.. Pradilla, A., and Sarria, D. 1968.Agron.Abstr. p. 66. Helinski, D. R., and Yanofsky, C. 1966. In “The Proteins” (H. Neurath, ed.), 2nd ed., Vol. 4, pp. 1-93. Academic Press, New York. Holley, R. W. 1965. In “Plant Biochemistry” (J. Bonner and J. E. Varner, eds.), 2nd ed., pp. 346-360. Academic Press, New York. Howe, E. E., Jansen, (3. R., and Gilfillan, E. W. 1965.Am.J.Clin. Nutr. 16,3 15-320. Ingram, V. M. 1957. Nature 180,326-328. Jacob, F., and Monod, J. 196 1. Cold Spring Harbor Symp. Quant. B i d . 26, 193-2 I I . Jimenez, J. R. 1966. Proc. High Lysine Corn C o n j , Purdue Univ., 1966, pp. 74-79. Corn Ind. Res. Found., Washington, D.C. Jimenez, J. R. 1968. Ph.D. Thesis, Purdue University. Josefsson, E., and Appelqvist, L. -A. 1969.J. Sci. FoodAgr. (in press). Kelley. E. G., and Baum, R. R. 1953. J. Agr. Food Chem. 1,680-683. Korner, A. 1964. In “Mammalian Protein Metabolism” (H. N. Munro and J. B. Allison, eds.), vol. 1, pp. 178-242. Academic Press, New York. Liener, 1. E. 1966. Advan. Chem. Ser. 57, 178-194. Lugg, J . W. H. 1949. Advan. Protein Chem. 5, 229-304. McMichael, S. C. 1960. Agron. J . 52, 385-387. Mahler. H. R., and Cordes, E. H. 1966. “Biological Chemistry,” pp. 752-809, Harper, New York. Margoliash, E., and Smith, E. L. 1965. In “Evolving Genes and Proteins” (v. Bryson and H. J . VOgel, eds.), pp. 221-242. Academic Press, New York. Mertz, E. T. 1966. Proc. High Lysine Corn C o n j . Purdue Univ.,1966, pp. 12-18. Corn Ind. Res. Found., Washington, D.C. Mertz, E. T. 1969. Agr. Sci. Rev. 6, 1-6. Mertz, E. T.. and Hressani, R. 1957. Cereal Chem. 34,63-69.

G E N E T I C M O D I F I C A T I O N O F PROTEIN Q U A L I T Y I N PLANTS

193

Mertz, E. T.. Bates, L. S.. and Nelson, 0 . E. 1964. Science 145, 279-280. Mertz, E. T., Veron, 0 . A,. Bates. L. S., and Nelson, 0. E. 1965. Science 148, 1741-1742. Mertz, E. T.. Nelson, 0 . E., Bates, L. S., and Veron, 0 . A. 1966. Adtmn. C h e m . Ser. 57, 228-242. Mohan, V. S., Nagarajan, V . , and Gopalan. C . 1966. lndiun J . M e d . R e s . 54, 410-414. Morton, R. K . . and Raison. J. K. 1964. Eiochem. J . 91, 528-539. Mosse, J . 1966. Federation Proc. 25, 1663- 1669. Mosse, J. 1968. In “Progres en chimie agricole et alimentaire.” pp. 47-8 I . Hermann, Paris. Mosse, J., Bandet, J . , Landry, J . , and Moureaux, T. 1966. Ann. Physiol. Vegetale 8, 331344. Munck, I_. 1964. Hereditas 52, 151-165. Munro. H. N.. and Allison, J. B., eds. 1964. “Mammalian Protein Metabolism,” Vols. I and 2. Academic Press, New York. Naismith, W. E. F. 1955. Biochim. Biophys. Acta 16, 203-2 10. Nance. W. E. 1963. Science 141, 123-130. National Academy of Sciences-National Research Council. 1966. “Pre-School Child Malnutrition. Primary Deterrent to Human Progress,” Publ. 1282. Natl. Acad. Sci. Natl. Res. Council, Washington, D.C. Nehring, K., and Schwerdtfeger, E. 1957. Z . Le6ensm.-Untersuch.-Forsch. 105, 12-2 I . Nelson, 0 . E. 1966. Federation Proc. 25, 1676- 1678. Nelson, 0. E. 1969. Proc. I A E A - F A 0 Panel, RiistLinga, Sweden, 1968, pp. 41-54. Intern. Atomic Energy Agency. Vienna. Nelson. 0 . E.. hlertz. E. T.. and Bates. L. S. 1965. Sciencc 150,1469- 1470. Osborne, T. B. 1924. “The Vegetable Proteins.” Longmans, Green, New York. 0 s b o r n e . T . B.. and Harris. I . F. 1903.5.A m . C h e m . Soc. 25,853-855. Osborne, T. B.. and Leavenworth, C . S. 19 13. J . B i d . C h e m . 14,48 1-487. Osborne. T . B.. and Mendel. L. 19 14. J . Biol. C h e m . 18, I - 16. Osborne, T. B., and Mendel. I-. 1916. J . B i d . C h e m . 25, 1-12. Oser. 8 . I.. 1951. J . Am. Dietet. Assoc. 27, 396-402. Pickett. R. A. 1966. Proc. High L?sine Corn Con$, Purdue Univ., 1966, pp. 19-22. Corn Ind. Res. Found., Washington. D.C. Piva, G . , Salamini, F.. and Santi, E. 1967. Aliment. Aniniale II,3-1 I . Pradilla, A., Linares, F.. and Harpstead, D. D. 1968. Agron. Abstr. p. 68. President’s Science Advisory Committee. 1967. “Report of Panel on the World Food Supply.” Vols. I and 11. U.S. Govt. Printing Office, Washington, D.C. Reissig, H. 1958. Zuechter 28,5 1-60. Rhyne, C . L., Smith, F. H., and Miller, P. A. 1959.Ayron.J . 51, 148-152. Ritossa, F. M.. Atwood, K. C., and Spiegelman, S. 1966. Genetics 54, 663-676. Roberts, R. C., and Briggs, D. R. 1965. Cereal C h e m . 42,7 1-83. Sadgopal, A. 1968. Advan. G e n e t . 14,325-404. Schuphan, W. 1966. Qualitas Plant. Mater. Veyetcibiles 13,346. Schweet, R.. and Heintz, R. 1966.Ann. R e v . Biochem. 35,723-758. Sheppard, D. E. 1964. Genetics 50,6 1 1-623. Smith. C. R., Jr., Earle, F. R., Wolff, 1. A,, and Jones, Q. 1959. J . A g r . Food C h e m . 7, 133- 136. Smith, F. H., Rhyne, C . L., and Smart, V. W. 1961.J . A g r . Food C h e m . 9,82-94. Stahmann. M. A. 1963. Ann. R e v . Plant Physiol. 14,137-158. Stdhmann, M. A. 1968. Econ. Botany 22,73-79. Stocking, C . R., and Ongun, A. 1962.A m . J . Botany 49,284-289. Tello, F., Alverez-Tostado, M. A., and Alvarado, G . 1965. Cerecrl Chenr. 42.168-384.

194

OLIVER E. NELSON

United Nations Economic and Social Council Report. 1967. “International Action to Avert the Impending Protein Crisis,” No. E14343. United Nations, New York. VanEtten, C. H., Miller, R. W., Wolff, I. A., and Jones, Q. 1961. J. Agr. Food Chem. 9, 79-82. VanEtten, C. H., Miller, R. W., Wolff, 1. A., and Jones, Q. 1963. J. Agr. Food Chem. 11, 399-4 10. VanEtten, C. H., Knolek, W. F., Peters, J. E., and Barclay, A. S. 1967. J. Agr. Food Chem. 15,1077-1089. Veron, 0.A. 1967. Ph.D. Thesis, Purdue University. Virupaksha, T. K., and Sastry, L. V. S. 1968. J . Agr. Food Chem. 16,199-203. Vogel, H. J., and Vogel, R. H. 1967.Ann. Rev. Biochem. 36,579-538. White, P. L., Alvistur, E., Dias, C., Vinas, E., White, H. S., and Collazos, C. 1955. J . Agr. FoodChem. 3,531-534. Wichser, W. R. 1966. Proc. High Lysine Corn Con$, Purdue Univ., 1966, pp. 104-1 16. Corn Ind. Res. Found., Washington, D. C. Wilson, C. M. 1966. Plant Physiol. 41,325-327. Wolf, W. J . , and Sly, D. A. 1967. Cereal Chem. 44,653-668. Wolf, W. J., Babcock,G. E.,and Smith,A. K. 1961. Nature 191,1395-1396. Zuckerkandl, E. 1964.J. Mol. Eiol. 8,128-147.

THE EXTRACTION, CHARACTERIZATION, A N D SIGNIFICANCE OF SOIL POLYSACCHARIDES G. D. Swincer, J. M. Oades, and D. J. Greenland Waite Agricultural Research Institute, University of Adelaide, South Australia

.....

I. Soil Carbohydrates 11. The Significance of

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

111. Studies on Soil Polysaccharides .............................................................

B. Extraction of Polysaccharides from Soils C. Purification of Soil Polysaccharides .................................................. D. Fractionation of “Purified” Soil Polysaccharides ................................ E. Properties of “Purified” Soil Polysaccharides F. The Origin, Synthesis, and Decomposition of IV. Methods for the Analysis of Complex Polysaccharide Materials .................. A. Introduction ................................................................................. B. Analytical Methods ....................................................................... C. Separation Methods ... V. Summary and Conclusions References .........................................................................................

Page I95 I96 199 199 199 206 210 214 218 222 222 223 227 229 230

I. Soil Carbohydrates

Although much is known about the nature and function of many polysaccharides synthesized by individual organisms, there is little information relating to the polysaccharides produced in an environment such as the soil which in a unique way brings together a great variety of biological forms. The comparative neglect of soil polysaccharides is perhaps surprising when it is realized that soils not only support the majority of higher plants but are the chief habitat for microorganisms. The amount of polysaccharide material added to soils as plant residues or synthesized in them by microorganisms must be enormous. Evidence that at least some of the polysaccharides produced in soils are capable of improving the stability of soil aggregates and therefore of encouraging the maintenance of an agriculturally favorable structure has provided the main stimulus for the study of these compounds. The slowness of progress can be attributed largely to the technological difficulties inherent in the study of any system as complex as the soil. 195

196

G . D. SWINCER, J . M. OADES, AND D. J . GREENLAND

The carbohydrates of soil are composed of a wide range of monosaccharides. Hexoses, pentoses, various deoxy and 0-methyl sugars, uronic acids, and amino sugars have been identified in hydrolyzates of numerous soils and soil extracts (Mehta et al., 196 1 ; Gupta, 1967). The presence of such a variety of components makes precise measurement of total soil carbohydrates very difficult, and this difficulty is aggravated by the low stability of most of the carbohydrate monomers under conditions so far found necessary for their release from polymeric compounds. However, the quantitative determinations that have been made indicate that carbohydrates constitute between 5 percent and 25 percent of the soil organic matter. Free monosaccharides constitute less than 1 percent of the soil carbohydrates, and extracted polysaccharides have rarely accounted for more than 20 percent (Mehta et al., 1961; Gupta, 1967). Approximately another 10 percent may consist of cellulose (Gupta and Sowden, 1964). More recently, techniques have been developed that enable almost complete extraction of carbohydrates from soil (Swincer et al., 1968a,b). The composition of carbohydrates removed by vigorous extraction procedures is similar to that of materials removed by less efficient methods, and the reason for differences in the ease of extraction of polysaccharides from different soils would appear to be physical rather than chemical. I I . The Significance of Soil Polysaccharides

The main stimulus for the study of soil polysaccharides has been the repeated indications of their influence on soil physical conditions. The polysaccharides undoubtedly also affect other soil properties such as cation exchange capacity (due to the uronic acid units), the retention of anions (due to amino groups, but only in acidic soils), carbon metabolism, biological activity (e.g., by acting as an energy source for heterotrophs), and the complexing of metals. Interest in the relationship between the physical properties and the polysaccharide components of soils was aroused by several reports which indicated that microbially produced gums could bind soil particles into stable aggregates (Winogradsky, 1929; McCalla, 1943, 1945; J . P. Martin, 1945a, 1946; Geoghegan and Brian, 1946, 1948; Haworth et al., 1946; Swaby, 1949). More recent work (Clapp et al., 1962; Harris et al., 1963; J. P. Martin and Richards, 1963; J . P. Martin et al., 1965) has confirmed the earlier observations. The presence in the soil of organisms that produce aggregate-stabilizing gums when cultivated in the laboratory (J. P. Martin, 1945a,b; Forsyth and Webley, 1949; Forsyth, 1954; Bernier,

SOIL POLYSACCHARIDES

I97

1958b; Webley et al., 1965) suggests that these gums will also occur in soils, and so would be expected to increase the stability of natural soil aggregates. As discussed below, mixtures of polysaccharides with properties which suggest that such microbial gums are included among them, have been isolated from a wide range of soils. In some instances it has been shown that the extracted polysaccharides are able to stabilize soil aggregates (Rennie et al., 1954; Dubach et ul., 1955; Whistler and Kirby, 1956; Mehta et al., 1960). Statistically significant correlations have also been demonstrated to exist between estimates of polysaccharide content and the degree of aggegation of the soil (Rennie et al., 1954; Chesters et al., 1957; Toogood and Lynch, 1959; Acton et al., 1963b; Watson and Stojanovic, 1965; Webber, 1965). The correlations obtained were not particularly close, but this is not unexpected since the estimates of polysaccharide content were somewhat crude (Acton et al., 1963a; Griffiths, 1965). In fact Griffiths ( I965), who reviewed this subject critically, made the point that satisfactory methods for the quantitative estimation of microbial polysaccharides in soils have not been developed, and consequently it has not been possible to evaluate accurately the contribution of these materials to aggregation. I t is also probable that only certain of the polysaccharides present in soils are involved in aggregate stabilization, since Oades ( 1967b), who developed and used quantitative methods of quite high precision, found that the correlation of aggregate stability with both composition and total amount of neutral sugar constituents was no better than that with other organic materials. This is consistent with the observation that bacterially produced gums differ considerably in their effectiveness (Clapp et al., 1962; J. P. Martin and Richards, 1963; J . P. Martin er al., 1965) and some plant products such as cellulose exert no direct influence on aggregate stability (Griffiths and Jones, 1965). It has also been shown that although the composition of polysaccharides in a soil under old pasture was similar to the composition of polysaccharides in the same soil type which had been under a wheat-fallow rotation for 40 years, there were differences between the treatments with respect to the amounts of carbohydrates present in the soil and the ease with which these materials could be removed from the soil (Swincer et al., 1968b). Thus the distribution of the polysaccharides within and around aggregates is probably important (Williams et al., 1967), so that only a portion of an effective polymer may actually be controlling stability. Mehta et al. (1 960) showed that artificial aggregates stabilized by adding dextrans or soil polysaccharides to dispersed soils lost their stability

198

G . D. SWINCER, J . M. OADES, AND D. J. GREENLAND

when treated with dilute (0.01 M ) sodium periodate and sodium borate (pH 9.6). The periodate oxidizes sugars containing cisglycol groups (Bobbit, 1956), and the partly oxidized polymers so produced are readily degraded in alkaline solution into various nonpolymeric fragments (Whistler and BeMiller, 1958). Polymers cleaved in this way can no longer act as bridges between the soil particles forming an aggregate. The natural aggregates examined by Mehta et al. (1960) did not lose their stability when treated in the same way, thus indicating that other agents were stabilizing the aggregates of these particular soils, possibly in addition to the polysaccharides. In soils of lower organic matter content, treatment with dilute periodate has been shown to produce a marked decrease in aggregate stability (Greenland er al., 1961, 1962; Clapp and Emerson, 1965; Deshpande et al., 1968). Furthermore, Harris er al. (1963) and Watson and Stojanovic (1965) have shown that the increase in the stability of aggregates in certain soils incubated with added organic materials is largely associated with the production of periodate-sensitive materials. This evidence together with the other data discussed above leaves little room for doubt that polysaccharides in soils exert an important influence on the stability of their physical structure. It is, however, also clear that other organic and inorganic materials can stabilize soil aggregates, and, where such materials are present, the polysaccharides may be of little additional benefit. The polysaccharides are probably of greatest importance in cultivated soils of relatively low total organic matter content (Greenland er al., 1962). In most of the studies discussed above, the relationship considered was that between polysaccharides and aggregate stability as determined by wet sieving. Aggregate stability is important in most soils except very sandy ones, since it is only by virtue of aggregate formation that a satisfactory continuity of pores in the soils is maintained, whereby adequate air and water movement can occur for optimum plant growth. In some instances more direct measurements have been made of the changes that occur in the physical properties of soils when polysaccharides are added or removed. Greenland er al. ( 1 962) showed that the permeability of beds of aggregates could be reduced by periodate treatment; and Clapp and Emerson ( 1965) showed that clays slaked and dispersed more readily after this treatment. Allison (1947), however, found that production of microbial gums in submerged soils could lead to an undesirable reduction in permeability because the microbial products were blocking some of the more important coarse pores.

SOIL POLYSACCHARIDES

199

Ill. Studies on Soil Polysaccharides

A. INTRODUCTION

The polysaccharide materials that have been extracted from soils, purified, and analyzed, often in considerable detail, have usually represented only a small proportion of the whole carbohydrate fraction of the soil organic matter. Thus, although the main stimulus for these investigations was derived from the observation that polysaccharides exert a favorable influence on soil physical properties, the polysaccharides exerting this influence could very well be those most firmly held by the soil colloids, and which were therefore neither extracted nor analyzed. However, the techniques that have been used and the results obtained undoubtedly form an important basis for understanding the nature, origin, and function of the majority, if not all, of the carbohydrate polymers present in soils.

B. EXTRACTION OF POLYSACCHARIDES FROM SOILS 1 . Introduction

Ideally an extractant for soil polysaccharides should, in order of priority: (a) be nondegradative, (b) give a sufficiently complete extraction for the materials extracted to be representative of the total, (c) be equally effective for all soils, and (d) extract selectively carbohydrate materials. None of the many extractants used fulfils these requirements. The main aim has often been simply to isolate from the soil a sample of relatively pure polysaccharide material that can be used for chemical and physicochemical characterization. 2. Assessment of Extraction Yield and Polymer Degradation a. Extraction Yield. Both colorimetric and gravimetric methods have been used to measure the proportion of the soil carbohydrates brought into solution under different sets of extraction conditions. Gravimetric methods (Rennie et a f . , 1954; Chesters et al., 1957; Salomon, 1962; Acton et af., 1963a,b) are of very limited value because they can be applied only to purified polysaccharide materials; there is a distinct possibility both of losses of polysaccharide materials during purification and of incomplete removal of contaminants (Acton er a f . , 1963a; Dormaar, 1967). In any case, for determination of the proportion extracted, the results must be related to values of “total soil carbohydrate” obtained by another method.

200

G. D. SWINCER, J. M. OADES, A N D D. J . GREENLAND

When preceded by complete hydrolysis into monosaccharides and removal of interfering compounds from the hydrolyzate, colorimetric methods can give useful estimates of the carbohydrate content of the soil, either before or after extraction, and of crude extracts. The colorimetric methods are normally applicable only to single classes of monosaccharides, such as the hexoses, pentoses, uronic acids, or hexosamines, and even within a particular class none of the methods gives the same color intensity for equimolar concentrations of the different individual monosaccharides. Moreover, optimum hydrolysis conditions vary at least from one class of sugars to another, and probably also from one individual sugar to another. Clearly, therefore, the most precise determinations of the proportion of the total carbohydrates extracted would require the measurement of the amount of each individual sugar in both the soil and the extract. Individual sugars, or the different classes of sugars, have been measured in soils and purified soil extracts (Graveland and Lynch, 1961; Thomas and Lynch, 196 1; Ivarson and Sowden, 1962; Sowden and Ivarson, 1962; Gupta et al., 1963; Gupta and Sowden, 1965; Cheshire and Mundie, 1966). Only Parsons and Tinsley (1961), Lynch et al. (1957, 1958), and Swincer et al. (1 968a,b) have attempted to relate the amounts extracted with the amounts originally present in the soil or left in the soil residue after extraction. The results of Parsons and Tinsley are probably not accurate, particularly with respect to uronic acids, as the authors themselves admit. Although Lynch et al. (1 957, 1958) claimed to have measured the recovery of the original carbohydrates in various extracts by separating and estimating seven different sugars, the yields reported were undoubtedly too high as a result of the low values for “total soil carbohydrate” that must have accompanied the very mild hydrolysis conditions used. b. Polymer Degradation. The detection and evaluation of damage to the polysaccharide molecules during the extraction process is by no means easy. Positive and conclusive evidence against changes in the carbohydrate polymers during extraction is virtually unobtainable because it is not yet possible to know the properties of these molecules before isolating them from the soil. Any information that can be obtained must be either indirect or of a negative kind. Whitehead and Tinsley ( 1964) made a useful indirect assessment of the likely degradative effect of their extraction procedure on soil polysaccharides by subjecting several other natural polymers (starch, alginic acid, chitin, cellulose, gluten) to the same treatment. Bernier (1 958a) compared the viscosities of polysaccharides extracted

SOIL POLYSACCHARIDES

20 1

from soils by different reagents, and considered that the preparations with higher viscosity were those that had suffered least degradation during extraction. This is a worthwhile approach provided samples are free from salt and ash. Swincer et a f . (1968a) preferred to use the results from analysis of extracts on Sephadex (3-25 as an indication of degradative treatments. When reagents caused extensive fragmentation of polysaccharides by hydrolysis or other means, a large proportion of the extracted materials were “low molecular weight” according to gel filtration on Sephadex (3-25. Thus, it was found that all methods involving high temperatures, even methods based on water, led to extensive polysaccharide degradation, and 0.5 N NaOH at 20°C. was the most efficient, single, nondegradative extractant. Again it should be emphasized that such experiments cannot provide conclusive evidence against degradation (e.g., by a relatively large average polymer size or the absence of specific degradation products), even though they can give indications of whether damage has occurred and thus help to rule out unsuitable extraction treatments. 3 . Extractants of Soil Polysaccharides Extractants that have been used for obtaining polysaccharides from soils include water, aqueous buffers and complexing reagents, dilute mineral acids, organic reagents, and alkalis. The yields of soil polysaccharides that have been obtained by various workers using a number of different procedures are nearly all based on gravimetric determinations and, therefore, provide only an approximate guide to the efficiency of each extractant. a. Water. Water has been a popular extractant for soil polysaccharides because of the simplicity of extraction, the relatively low simultaneous extraction of humic materials, and the ease of subsequent purfication. Most of the procedures involve extraction at temperatures of at least 70”C., either with continuous shaking or by use of Soxhlet’s apparatus (Mortensen, 1961; Keefer and Mortensen, 1963; Mortensen and Schwendinger, 1963; Thomas et al., 1967). Forsyth ( I 950) found that autohydrolysis of an extracted polysaccharide occurred in distilled water at 1 OO’C., 90 percent of the arabinose and considerable amounts of ribose being released in 24 hours. Furthermore, Swincer et al. ( 1968a) considered hot water extraction to be degradative on the basis of analysis of the extract on Sephadex (3-25. What quantitative information there is shows that hot water extracts up to about 2 percent of the soil organic matter, which could represent up to one-quarter of the total carbohydrates.

202

G. D. SWINCER, J. M. OADES, A N D D. J. GREENLAND

b. Aqueous Buffers and Complexing Reagents. Buffer solutions (Bernier, 1958a) and complexing agents, such as sodium phyrophosphate (Bernier, 1958a; Lynch et al., 1958; Mortensen and Schwendinger, 1963) and disodium EDTA (Barker et af., 1965, 1967), have been used in attempts to obtain polysaccharides with a minimum of alteration from their natural state. In general, yields were low (less than 5 percent of the soil carbohydrate), and it is impossible to know whether the material extracted was representative of the total. For this reason such mild reagents are probably of limited value for investigating the whole soil polysaccharide fraction, in spite of their ideality in terms of the risk of hydrolysis. A possible exception is the use of metal complex reagents such as Schweitzer's reagent (cuprammoniurn hydroxide) for extracting cellulose. Schweitzer's reagent was first applied to soils by Daji (1932), and its usefulness has been confirmed by Gupta and Sowden ( 1964). However, it might well be replaced to advantage by one of the more recently developed metal complex reagents with high cellulose dissolving power (Jayme and Lange, 1963). c . Dilute Mineral Acids. Although inorganic acids are relatively inefficient reagents for extracting humic materials from soil (Mortensen and Himes, 1964), they are more effective for the polysaccharide fraction. Barker et al. ( 1 965, 1967) chose 0.6 N H2SO4as the extractant for routine isolation of larger quantities of polysaccharide materials from an English muck soil. Other reagents were superior in terms of yield, but the acid extracts were more easily purified. Furthermore, treatment with acid apparently hydrolyzed bonds between polysaccharides and humic materials, thus simplifying the separation of polysaccharides from these materials. Acid hydrolysis of the polysaccharides themselves was prevented by carrying out the extraction at 3°C. Black et af.( 1955) also used dilute mineral acid (0.09 N HCl) to extract a polysaccharide complex from peat. The yield was low compared with those obtained by other workers with different reagents: furthermore, it is quite likely that some degradation occurred under the conditions used ( 1 hour at 60" to 70°C.). H F at 20°C. was the most efficient acid extractant used by Swincer et af.( 1 968a), but the extracts are not easily analyzed; thus, HCI was preferred and released 10 to 20 percent of the carbohydrate from a range of great soil groups, being most efficient in soils rich in calcium carbonate or free iron and aluminum oxides. Another acid extractant that has been tried is a suspension in water of the H+ form of a cation exchange resin (Amberlite 1 R-120) (Barker et al.,

SOIL POLYSACCHARIDES

203

1965, 1967). This treatment extracted slightly less carbohydrate material than 0.6 N H&O, under the same conditions, and the mineral acid was preferred for this and other reasons. d. Organic Reagents. Parsons and Tinsley ( 1961) extracted polysaccharides from a variety of soils by double reflux with 0.2 N lithium bromide in anhydrous formic acid. The proportions of soil organic matter (3.5 to 17.5 percent) extracted as polysaccharides are some of the highest on record. However, after extraction, a considerable amount of carbohydrate still remained with each soil residue, and the authors themselves admitted that the yields quoted were probably overestimates because some of the C 0 2 evolved in the decarboxylation method for determining uronic acids almost certainly arose from nonuronide materials. Deuel et al. (1960) have established that ready decarboxylation is a general property of the colored humic substances. I t is possible that the reducing sugar content was also overestimated due to the presence of other reducing substances in hydrolyzates of the isolated polysaccharide materials. Whitehead and Tinsley ( 1964) extracted high proportions of the organic matter from several soils by refluxing with a mixed reagent consisting of 0.4 M boric acid, 0.4 M oxalic acid, and 0.2 M lithium chloride in dimethyl formamide (DMF). The polysaccharide contents of the preparations were not determined and attempts to separate the dissolved polysaccharides from the nonpolysaccharide components were not successful. Although exposure to high temperatures was not prolonged, hydrolysis of soil polymers almost certainly occurred as appreciable degradation was observed when the procedure was applied to such natural polymeric material as starch, alginic acid, chitin, and gluten. Swincer et al. ( I968a) also considered this procedure to be degradative, based on analysis of the extract by Sephadex (3-25. DMF (without dissolved acids or salts) was also among the extractants examined by Barker et al. (1965, 1967). Extraction was for 24 hours at room temperature, and the carbohydrate yield was very low. However, the yield was increased almost 30-fold if the soil was hydrogen ion saturated before extraction with DMF. Of the other organic extractants examined, N-methyl-2-pyrrolidone was only slightly more efficient than DMF (without the pretreatment), but 8 M urea was quite superior. Not enough data were given to allow adequate assessment of the potential of the different reagents as extractants for soil polysaccharides. The only soil examined was a muck, and relative yields could be quite different in mineral soils where many of the polysaccharides are likely to be strongly associated with inorganic colloids. e. Alkalis. I n spite of some objections and criticisms, dilute NaOH,

204

G . D. SWINCER, J. M . OADES, A N D D. J. GREENLAND

the classical extractant for soil organic matter, has been used frequently for isolating soil polysaccharides (Forsyth, 1947, 1950; Rennie et al., 1954; Dubach et af., 1955; Chesters et al., 1957; Muller et af., 1960; Graveland and Lynch, 196 1 ; Acton et al., 1963a; Dormaar, 1967; Oades and Swincer, 1968; Swincer et al., 1968a,b). This extractant is more efficient than most others, an observation which has been made repeatedly in studies of the extraction of other soil organic materials (e.g., Bremner and Lees, 1949; Choudri and Stevenson, 1957; Evans, 1959). In contrast to most other workers, Whistler and Kirby ( I 956) reported that hot water and cold 0.5 N NaOH extracted comparable amounts of “purified” polysaccharide from an American soil. However, the amount of polysaccharide in their initial crude 0.5 N NaOH extract may have been much higher than they reported, as considerable losses could have occurred during the purification treatment. Large amounts of humic materials are always extracted with the polysaccharides. The separation of these materials from the polysaccharides has proved to be particularly difficult, and is a significant objection to the use of this extractant. Another objection to the use of NaOH has been the possibility of damage either by hydrolysis of the polymers or by autoxidation (Bremner, 1954; Tinsley and Salam, 1961 ; Dubach and Mehta, 1963). There is good evidence that some organic materials are, indeed, altered under alkaline conditions (Bremner, 1950; Choudri and Stevenson, 1957; Evans, 1959) although Kononova (196 1 ) discusses other evidence which, she says, shows that extraction with alkaline solutions does not change the nature of humic substances essentially. As these experiments were concerned only with overall effects, the conclusions reached do not necessarily apply directly to the polysaccharide materials, which constitute only a rather small proportion of the extracted organic matter. Bremner (1 950) found that the oxygen taken up during alkaline extraction of soils went into the acid-insoluble (humic acid) fraction rather than into the fulvic acid fraction which contains most of the carbohydrates (Acton et af., 1963a; Dormaar, 1967; Swincer et al., 1968a). Some alkaline degradation of the polysaccharides might be expected. Neuberger and Marshall ( 1966b) and Horton and Wolfrom (1 963) state that glycosidic linkages between monosaccharides are ordinarily stable to alkali, but that in certain circumstances a slow stepwise degradation may take place from the reducing end of a polysaccharide molecule. Saccharinic acids are formed and eliminated and new reducing groups are exposed until this “peeling” reaction is intercepted by a “stopping” reaction in which release of the saccharinic acid is inhibited by an unfavor-

SOIL POLYSACCHARIDES

205

able glycosidic linkage. The rate of the “peeling” reaction, which can occur in the absence of oxygen, is influenced by the position at which the penultimate sugar residue is attached (Kenner and Richards, 1957; Whistler and BeMiller, 1958). If this linkage is (1 +. 3) the terminal residue is readily detached, but ( I +. 4) and ( 1 + 6) links are much more stable, and ( 1 -+2 ) links are resistant to the action of alkali even under very vigorous conditions. Thus under mild alkaline conditions, some shortening of polysaccharide chains is possible even though there is no random fragmentation such as occurs in hot mineral acids. The presence of oxygen in an alkaline solution of a polysaccharide can lead to additional damage. I n this case the individual sugars may be oxidized without changes in polymer size (by hydrolysis of glycosidic bonds). In order to reduce this risk, Choudri and Stevenson (1957) advocated the addition of stannous chloride to the soil sample before extraction with NaOH; other workers (e.g., Bremner, 1950) have carried out alkaline extraction of soil organic matter under an atmosphere of nitrogen. Tinsley and Salam ( 196 1) mentioned two other undesirable features of NaOH as an extractant of soil organic matter. First, silica is dissolved from the mineral matter in the soil and contaminates the organic fractions separated. This can be overcome by concentration of a solution of the partly purified polysaccharide; this results in polymerization and precipitation of the silica (Swincer et al., 1968a). Secondly, NaOH is known to dissolve some protoplasmic and structural components of fresh organic tissues. The problem is eliminated if, before extraction, the free organic residues are first removed from the soil by flotation sieving (Roulet et al., 1963a) or densimetric fractionation with heavy liquids (Ford et al., 1969). $ Sequential Extraction Treatments. No single reagent has given complete extraction of soil polysaccharides, and for maximum extraction multiple treatments are necessary. This is so with other organic materials, e.g., humic materials (Choudri and Stevenson, 1957; Bremner and Harada, 1959) and organophosphates (J. K. Martin, 1964). Studies on ten great soil groups showed that a double treatment, using I N HCI followed by 0.5 N NaOH, removed more than SO percent of the soil carbohydrate in all except two cases (Swincer er d.,1968a). A third treatment using acetic anhydride containing 2.5 percent concentrated H & 0 4 at 60°C. for 2 hours brought the total yield up to 80 percent or more except in an ando soil and a red earth. Extraction with NaOH can be increased by ultrasonic dispersion. Thus the sequential procedure should give almost complete extraction in many cases and although certain soils

206

G. D. SWINCER, J . M . OADES, AND D. J . GREENLAND

may require special approaches this procedure would appear to have general applicability (Fig. 1). These results were obtained on soils after the removal of partly decomposed plant remains by a densimetric procedure (Ford et d . , 1969). SOIL Ultrasonic dispersion in a

LIGHT FRACTION - partly decomposed plant and animal remains

SOIL RESIDUE

1

I N HCI 20"

1 N HCI EXTRACT - add EDTA, neutralize with NaHCO,. Concentrate for gel filtration on Sephadex and Biogel

SOIL RFSIDUE Ultrasonic dispersion in

SOIL RESIDUE Acetic anhydride + 2.5% coric. H2SO4 60"

SOIL RESIDUE

0.5 N NaOH extract - pass through H+ Dowex 50 and Polyclar AT columns. Concentrate for gel filtration on Sephadex and Biogel REACTION MIXTURE -dilute tenfold with water and extract acetylated polysaccharides in CHCI,. Concentrate the extract for gel filtration on Sephadex LH-20

Hydrolyze to determine residual carbohydrate content

FIG.1. Procedure for isolation of polysaccharides from other soil materials. The light fraction usually contains from 10 to 50 percent of the total soil carbohydrate. Of the remainder, 10 to 20 percent occurs in the acid extract, 30 to 50 percent in the NaOH extract, 20 to 30 percent in the acetic anhydride extract, and 10 to 20 percent may be left in the final residue. Details of this procedure are described by Swincer et af. (1968a,b).

C . PURIFICATION OF SOILPOLYSACCHARIDES

I . Introduction After removal of solids by centrifugation, the first step in the recovery of polysaccharides from an extract depends on the particular extractant used. The specific primary treatments that have been applied will be discussed very briefly before consideration is given to the more general procedures used for subsequent purification.

SOIL POLY SACCHARIDES

207

2 . Primary Purijication Treatments With water extracts, the initial purification treatment is simplified because the problem of removing solutes contributed by the extractant does not arise. In fact some workers (Duff, 1952a; Whistler and Kirby, 1956) have omitted the “primary treatment” and have applied the more general procedures for purifying polysaccharides directly to the concentrated extract. However, Mortensen and co-workers (Mortensen, 196 1 ; Keefer and Mortensen, 1963; Thomas et al., 1967) favored an initial acidification step (adjustment to pH 2 with N HCI) to precipitate humic materials. To neutralize acid extracts, Black et al. ( 1 955) used 40 percent NaOH while Barker et al. (1 965, 1967) used NaHC03. The latter authors examined polysaccharides both in the resulting precipitate and in the supernatant. The precipitated polysaccharides were brought into solution with 0.3 N HCI before the application of further purification treatments. To avoid precipitation of polysaccharides during the neutralization of acid extracts Swincer et al. (1968a) added EDTA before NaHCOa. The EDTA prevented precipitation of di- and trivalent cations in the acid extract and consequently hydroxides of metals did not occur as precipitates associated with polysaccharides. Parsons and Tinsley (1 96 I ) used diisopropyl ether to precipitate organic materials from anhydrous formic acid extracts of soil and Whitehead and Tinsley ( 1 964) treated DMF extracts with diethyl ether in the presence of acetic acid. The precipitates were dissolved in water and formic acid, respectively, in readiness for further purification treatments. Where alkaline solutions have been used for polysaccharide extraction, the first purification step has been acidification to pH 2 to 3 to precipitate humic acids. The acidification treatment removes a large proportion of the organic components of the extract, as well as some of the inorganic materials, but the precipitate may also contain significant amounts of carbohydrate material (Lynch et al., 1957; Acton ef al., 1963a; Dormaar, 1967). Dormaar ( 1967) reported that some humic acid precipitates contained up to 9 percent of the total soil carbohydrates, and Acton et al. obtained values as high as 12 percent. This association of carbohydrates with the humic acid fraction is probably due to coprecipitation involving metal cations. Swincer et al. (1968a) found that 8 percent of the soil carbohydrate was precipitated with the humic acid fraction obtained by direct addition of acid to the NaOH extract, but only 1 percent of the soil carbohydrate was present in the humic acid fraction obtained by passage of the NaOH extract through a column of coarse H+ Dowex 50. The humic acid fraction was

208

G . D. SWINCER, J. M. OADES, AND D. J. GREENLAND

trapped in the resin column, and the fulvic acid solution emerging from the column was free of metal cations and was easily concentrated without the formation of precipitates. Cellulose extracted from soils with Schweitzer’s reagent was recovered by precipitation with 5 volumes of 80 percent ethanol (Gupta and Sowden, 1964). The precipitate was washed successively with water, N HCl, and water before being dried. This product was reported to be reasonably pure without further treatment.

3 . Secondary Purification Treatments The purification procedures commonly applied to soil polysaccharide extracts after the appropriate initial treatment include: (a) dialysis to remove salts, solvents, and other low molecular weight materials; (b) charcoal filtration to remove colored compounds; (c) precipitation of the polysaccharides from aqueous solution by the addition of acetone, alcohol, or other reagents; and (d) various deproteinization procedures. Dialysis in Visking tubing against distilled water has been the most commonly used purification treatment, having been applied in all cases of extraction with hot water (Duff, 1952a; Whistler and Kirby, 1956; Clapp, 1957; Mortensen, I96 1 ; Keefer and Mortensen, 1963; Mortensen and Schwendinger, 1963; Thomas et al., 1967) and acid (Black et al., 1955; Barker et al., 1965, 1967) as well as to phosphate buffer extracts (Bernier, 1958a), a sodium pyrophosphate extract (Mortensen and Schwendinger, 1963), and an acidified NaOH extract (Muller et al., 1960). Losses of carbohydrate during dialysis of various extracts against distilled water are considerable (Clapp, 1957; Muller et al., 1960). Swincer et al. (1968a) showed that much of this material could be recovered from the dialysis water as low molecular weight fragments, but not as monosaccharides. However, significant amounts were sorbed on dialysis tubing and could not be removed by washing with water. Forsyth (1 947) introduced the use of acid-washed animal charcoal for the sorption of organic materials from classical fulvic acid extracts. Charcoal will remove most of the organic materials and all the color from such extracts and has been used to purify polysaccharides not only from the fulvic acid fraction of NaOH extracts (Dubach et al., 1955; Muller et al., 1960; Swincer et al., 1968a), but also from a hot water extract (Whistler and Kirby, 1956) and phosphate buffer extracts (Bernier, 1958a). However, recoveries of polysaccharides from charcoal are poor. Whistler and Kirby (1 956) measured a recovery of 14 percent and Swincer et al. (1968a) reported a 12 percent recovery. Thus while relatively pure polysaccharides can be obtained by this procedure, losses are very high. Swincer et a f . (1968a) increased recoveries to 30 percent

SOIL POLYSACCHARIDES

209

by using 8.5 N acetic acid as eluant, and even up to 65 percent by pretreating the charcoal with stearic acid as described by Dalgleish (1955), but results obtained were difficult to reproduce. Polyclar AT (cross-linked polyvinyl pyrrolidone) which is a strong adsorbent of polyphenolic compounds (Sanderson and Perera, 1966) proved to be most useful for separating colored materials from polysaccharides in “fulvic acid” solutions. Recoveries of carbohydrates from columns of Polyclar AT were greater than 95 percent, and almost complete removal of colored compounds occurred when the pH was lowered to 2 (Swincer et al., 1968a). Very often polysaccharide materials have been recovered from aqueous solution by precipitation with excess ethanol or acetone (Forsyth, 1947, 1950; Duff, 1952a; Rennie et al., 1954; Black et al., 1955; Chesters et al., 1957; Bernier, 1958a; Mortensen, 1961; Salomon, 1962; Acton et al., 1963a; Dormaar, 1967). Polysaccharides have been purified by precipitation from 0.5 N lithium chloride solutions at pH 7 by the addition of cetyl trimethylammonium bromide (Parsons and Tinsley, 196 I ) , from formic acid solution with diisopropyl ether containing 1 percent acetyl chloride (Whitehead and Tinsley, 1964),and from alkaline solution with copper sulfate (Forsyth, 1950). The purity of some preparations was increased by redissolving the precipitate in water and repeating the step, sometimes with an alternative precipitant such as ether (Black et al., 1955) or a mixture of alcohol and acetic acid (Duff, 1952a). Although they have been widely used, these precipitation methods are far from satisfactory, particularly when not preceded by a desalting step. They are not specific for polysaccharides, as some noncarbohydrate materials are also precipitated, and they leave polysaccharides in the supernatant (Acton et al., 1963a; Dormaar, 1967). In fact, Acton et al. ( 1963a) found that the precipitate obtained by one purification procedure (Rennie et al., 1954; Chesters et al., 1957) had an ash content of more than 60 percent and a carbohydrate content of less than 20 percent, measured by an anthrone method (Brink et al., 1960). They also found that there was much more carbohydrate in the supernatant than in the precipitate. Much of the protein material that invariably remains with soil polysaccharide preparations, even after several different purification treatments, has been removed either by the emulsification procedure of Sevag et al. (1938) (Bernier, 1958a; Mortensen, 1961; and Thomas et al. 1967) or by calcium sulfate precipitation and sorption on Fuller’s earth in N acetic acid (Bernier, 1958a). However, amino acids were always detected in hydrolyzates of the “deproteinized” polysaccharides. Roulet et al. (1963b) used gel filtration and ion exchange chroma-

210

G . D . SWINCER, J . M . OADES, AND D . J. GREENLAND

tography with the primary purpose of purifying the humic substances in acid extracts. The best separation of high molecular weight carbohydrates (measured as uronic acids) from “nitrogenous compounds” (as well as from humic substances) occurred with Sephadex G-75, but no pure fractions were recovered. The “nitrogenous substances” were mainly of low molecular weight, but they occurred in all fractions. Oades and Swincer (1968) and Swincer et al. (1968a,b) have confirmed the partial separation of nitrogenous materials from carbohydrates by means of gel filtration. This separation occurred because materials of low molecular weight obtained by gel filtration invariably contained higher proportions of amino acids than larger polymers. Thus materials separated as the high molecular weight fraction on Biogel P- 100 contained the least proportions of amino acids. A more direct application of gel filtration to the purification of soil polysaccharides has been made by Barker et af. (1 965, I967), who used Sephadex G- 100 to remove humic materials from the high molecular weight polysaccharides. Limitations of the technique were that good separations could be achieved only with acidic extracts, and the problem of separating humic materials from the smaller polysaccharides remained unsolved. Other purification procedures that have been used to advantage with soil polysaccharides are desalting by ion exchange (Mortensen, 196I), and removal of clay minerals by treatment with a mixture of 0.3 N with respect to both H F and HCI for 10 minutes at 40°C. (Mortensen and Schwendinger, 1963). In summary, it can be said that in spite of numerous intensive attempts to purify the polysaccharides extracted from soils, it is doubtful if any of the resulting preparations have been completely free from noncarbohydrate materials. The possibility of carbohydrate-humic and carbohydrate-protein links will be considered later.

D. FRACTIONATION OF “PURIFIED” SOILPOLYSACCHARIDES 1 . Introduction The large number of different sugars in the hydrolyzates of “purified” polysaccharides and viscosity and ultracentrifuge studies (Bernier, I958a; Ogston, 1958) which indicate a wide range of molecular size and shape show that soil polysaccharides are complex mixtures. Their fractionation has thus proved to be even more difficult than fractionation of other polysaccharide mixtures. Most of the well-established methods of polysaccharide and polymer chemistry have been used, but in spite of a

SOIL POLYSACCHARIDES

21 I

great deal of effort, fractions of distinctly different composition or physicochemical properties have been obtained only rarely. Even during the extraction of polysaccharides from the soil there seems to be very little fractionation of the original soil carbohydrate components, at least in terms of monosaccharide composition; the same spectrum of sugars in very similar proportions has been found in hydrolyzates both of whole soils and of various organic fractions (including “purified” polysaccharide preparations) extracted from the soils (Lynch et al., 1957, 1958; Parsons and Tinsley, 1961 ; Acton et al., 1963a; Swincer et al., 1968a,b). 2. Fractionation Methods The fractionation methods most commonly tried can be listed as fractional precipitation procedures, electrophoresis, ion exchange chromatography, and gel filtration. a . Fractional Precipitation. Fractional precipitation of soil polysaccharide preparations from aqueous solution has been attempted using ethanol (Whistler and Kirby, 1956; Bernier, 1958a) and quaternary ammonium compounds (Bernier, 1958a; Streuli et al., 1958; Parsons and Tinsley, 1961), but in terms of quantitative sugar composition the fractions obtained were essentially identical. The poor results obtained by such producedures with soil polysaccharides are probably due, largely, to the complexity of the preparations as these methods have been used successfully on polysaccharides obtained from other sources (e.g., Erskine and Jones, 1956; Whistler and Lauterbach, 1958; J . E. Scott, 1960). b. Electrophoresis. Electrophoresis has been one of the less commonly used fractionation methods in polysaccharide chemistry, and separations have usually been only partly successful, even with quite simple mixtures (e.g., Northcote, 1954; Foster et al., 1956; Hocevar and Northcote, 1957; Barrett and Northcote, 1965). Several attempts to resolve the constituents of soil polysaccharide preparations using free-boundary and continuous-flow paper electrophoresis with either phosphate or borate buffers have confirmed the heterogeneity of the preparations without giving definite fractionations (Whistler and Kirby, 1956; Clapp, 1957; Bernier, 1958a: Mortensen, 196 1 ; Thomas, 1963; Mortensen and Schwendinger, 1963; Barker el al., 1965). Although this work has not led to the development of useful fractionation procedures, it has contributed toward an understanding of the behavior of soil polysaccharides; therefore, the results are discussed more fully in Section 111, E, 5 .

212

G. D. SWINCER, J. M. OADES, AND D. J . GREENLAND

c . Ion Exchange Chromatography. Anion exchange chromatography on columns packed with DEAE-Sephadex A-50 (Barker et al., 1965, 1967: Finch et al., 1967) or DEAE-cellulose (Muller et al., 1960: Roulet et al., 1963b: Thomas et al., 1967) has given the most well defined fractionations reported for isolated soil polysaccharides. The success of Barker et al. and Finch et al. is somewhat surprising as the charges on the Sephadex materials would not be very accessible to the larger polysaccharide molecules. In fact, Roulet et al. found that much soil polysaccharide material was excluded with both SE-Sephadex A-50 and DEAE-Sephadex A-50, and therefore this material was not fractionated on a charge basis. The charged celluloses appear to be the more logical choice as more charges are present on accessible surfaces. However, the low recovery (62.7 percent) reported by Thomas et al. ( 1 967) gives reason for some caution; more than a third of the polysaccharide was not eluted from the column. This contrasts with recoveries of 100 percent claimed by Muller et al. ( 1 960), also using DEAE-cellulose, and by Barker et al. (1965) using DEAE-Sephadex. Elution of soil polysaccharides from the DEAE-cellulose columns was always with phosphate buffers and an increasing gradient of sodium hydroxide. Five fractions of different composition were recovered in each case. Muller et al. and Roulet et al. observed an increase in uronic acid content and a decreasing trend in neutral sugar content from fraction 1 to 5 , the final relatively small fraction being very rich in uronic acids. Thomas et al. also found differences in uronic acid content, but the pattern was less regular; fraction 4, the largest, also had the highest concentration of uronic acids. All three groups noticed that the same neutral sugars were present in all fractions but only Roulet et al. were able to detect distinct differences in the relative amounts. Elution of a soil polysaccharide preparation of high molecular weight from columns of DEAE-Sephadex A-50 using phosphate buffer solution and a stepwise increase in chloride concentration yielded four fractions differing in pentose, glucose, and uronic acid contents (Barker et al., 1965, 1967; Finch et al., 1967). The potential of this approach, which is proving successful for the fractionation of polysaccharides of different origin, was recognized in 196 1 by Mehta et al., who suggested that it should help considerably in attempts to isolate homogeneous soil polysaccharides. d. Gel Filtration. Gel filtration is a relatively new technique with great potential for the fractionation of soil organic matter. Fractionation is basically in terms of molecular size. Both Roulet et a!. (1963b3 and

213

SOIL POLYSACCHARIDES

Barker et al. ( 1 967) have reported useful separations of extracted soil carbohydrates. The fractions obtained were not analyzed for differences in composition. Gel filtration using various grades of Sephadex and Biogel was the main technique used in studies on soil polysaccharides by Oades and Swincer ( 1968) and Swincer et al. ( 1968a,b) (Fig. 2). This work showed, as anticipated, that extracted polysaccharides possess a wide range of molecular sizes and that the technique of gel filtration can be used to obtain molecular size fractions which differ considerably with respect to the groups of compounds represented and to their monosaccharide composition. Generally, low molecular weight materials were rich in amino acids and such monomers as glucose, ribose, and glycerol. The materials of molecular weight 4,000 to 100,000 contained most of the extracted uronic acids.

20

25

30

35

Effluent fraction no.

FIG. 2. Fractionation of an I N HCI extract of the Urrbrae soil with Sephadex ( 3 - 2 5 . Conditions: Urrbrae fine sandy loam-a red-brown earth; Sephadex G-25. c o m e grade: column dimensions, 50 cm. x 2.2 cm. diameter; sample, 4 ml. of N HCI extract after neutralization in the presence'of EDTA and concentration: eluant, distilled water: effluent volume fraction, 5.0 ml.; high molecular weight polysaccharides: fraction numbers I2 to 18; low molecular weight polysaccharides: fraction numbers I9 to 38.

214

G. D. SWINCER, J. M . OADES, A N D D. J . GREENLAND

Polymers of molecular weight greater than 100,000 contained appreciable quantities of deoxyhexoses but few amino acids.

E. PROPERTIES OF “PURIFIED”SOILPOLYSACCHARIDES

1. Zntroduction Many soil polysaccharides have been well characterized, and the more important information is discussed below. However, the significance of the data is limited because the analyzed materials, (a) rarely represented more than a small fraction of the total polysaccharide content of the soils, (b) consisted of a heterogeneous complex of polysaccharides, and (c) contained significant amounts of noncarbohydrate impurities. In the following sections, emphasis is placed on the results obtained rather than on the details of technique. A general discussion of the methodological aspects of polysaccharide characterization is given in Section IV. 2 . Monosaccharide Composition Most workers have hydrolyzed their polysaccharide preparations with 1 N HzS04 and examined the component monomers by paper chromatography. Both column chromatography (Forsyth, 1950; Duff, 1952a,b, 196 1 ; Barker et al., 1 967) and paper electrophoresis (Mortensen, 196I ; Parsons and Tinsley, 1961) have been used as supplements to paper chromatography. Usually the sugars have been determined colorimetrically either after elution of the individual spots from the paper strips or after collection of a large number of small effluent fractions from the columns. In some cases (Mortensen, 196 1 ; Parsons and Tinsley, 1961; Thomas et al., 1967) the color intensity of the spots developed on the paper strips by treatment with a specific reagent has been used to estimate the relative amounts of the different components. Quant.itative data are difficult to obtain with this procedure although Parsons and Tinsley (196 1) achieved reasonable precision by spraying with 2-aminobiphenyl and measuring the color intensity of the spots with a reflectance spectrophotometer. Gas-liquid chromatography, which has several advantages over other chromatographic methods, was used by Oades and Swincer (1968) and Swincer et al. ( I 968a,b) to determine the neutral sugar composition of various polysaccharide preparations obtained from soil extracts. This quantitative procedure enabled differences in the monosaccharide composition of different soil extracts and molecular size fractions to be detected easily (Fig. 3).

215

SOIL POLYSACCHARIDES

Although many workers have detected (and sometimes measured) one or two uronic acid spots on paper chromatograms of soil polysaccharide hydrolyzates positive identification of the compounds has not been made. Quantitative uronic acid determinations have often been by direct

Soil under old pasture

60 40-

20-

0-

Ga GIMA RI X F Rh INHCL extract

6o

:1

Ga GI MAR1 X F Rh

Ga GIM A RI X F Rh

Go GIMA R i X F Rh

0 5NNoOH extract Acetylation extract Sail res!due after sequential extraction

Ih

Soil under fallow-wheat rotation

tt

dARiXFRh IN HCL extract

Go

IARiXFRh

G

iIM ARI X F Rh

0 5N NaOH extract Acetylatian extract Sail residue after sequent la1 extraction

FIG. 3. Neutral sugar profiles of soil polysaccharide preparations. Quantitative sugar analyses by gas-liquid chromatography of alditol acetates obtained from HzSO, hydrolyzates of the polysaccharide preparations. Sequential extraction procedure was done as outlined in Fig, 1. G a = galactose, GI =glucose. M = rnannose, A = arabinose, Ri =ribose, X = xylose, F = fucose, Rh = rharnnose. Soil was Umbrae fine sandy loam, a red-brown earth (Swincer et al., 1968a,b).

216

G . D.

SWINCER, J .

M.

OADES,

A N D D. J. GREENLAND

methods that do not involve prior hydrolysis of the polysaccharides to their component monomers. Both decarboxylation with HCI (Forsyth, 1950; Duff, 1952a; Parsons and Tinsley, 196 I ) and the carbazole colorimetric method (Bernier, 1958a; Mortensen, 1961; Muller et al., 1960; Barker et al., 1965, 1967; Thomas et al., 1967; Oades and Swincer, 1968) have been used. In most analyses of soil polysaccharides, amino sugars have been reported as either not detectable or present only in trace amounts. However, the true picture may be quite different because in many cases hydrolysis was with 1 N H2S04which is known to cause rapid destruction of amino sugars at high temperatures. Furthermore, results obtained after the establishment of optimum conditions for the liberation of amino sugars indicated that polysaccharides isolated from a wide range of soils contained significant amounts (3.1 to 7 percent) of these components (Parsons and Tinsley, 196I ; Swincer et al., 1968a,b). The identity of the individual amino sugars of isolated soil polysaccharides was determined by Thomas ( 1 963), who reported the presence of glucosamine, galactosamine, and N-acetyl glucosamine after paper chromatography of sugars in a 1 N HzS04hydrolyzate. Swincer et al. ( 1968a) also identified glucosamine and galactosamine by paper chromatography and also by ion exchange column chromatography. A number of minor components of extracted polysaccharides have been identified. Duff ( 1952b) separated from a soil polysaccharide three sugars with high Rf values (on paper chromatograms) and showed that they contained methoxyl groups. Two of the sugars were methylaldoses (Duff, 1954). Subsequently by a combination of partition chromatography and elution from a highly active vegetable charcoal, seven separate “high R i ’ fractions were isolated (Duff, I96 1). 2-O-Methyl-~-rhamnoseand 4-O-methyl-~-galactosewere identified. Some other workers have also detected “highRi’ sugars in hydrolyzates of soil polysaccharies but the compounds were not identified (Dubach et al., 1955; Whistler and Kirby, 1956; Muller et al., 1960; Keefer and Mortensen, 1963; Thomas, 1963; Barker el al., 1967; Thomas et al., 1967). Barker et al. obtained some evidence that their unidentified sugars were methyl ethers of glucose and xylose. 3 . Glycosidic Linkages

Owing to the complexity of soil polysaccharide preparations, structural investigations by the classical methods have not been possible. The only structural information obtained so far has come from two brief enzymological studies (Thomas, 1963; Barker et al., 1967).

SOIL POLYSACCHARIDES

217

Thomas ( 1 963) obtained evidence that his polysaccharide preparation did not contain the p-( 1 + 4)-glycosidic linkage of cellulose, the methyl ester linkage of pectin, and the ( 1 + 4)-glycosidic linkage of polygalacturonic acid. Barker et af. ( 1967) studied the behavior of a polysaccharide preparation of high molecular weight toward an a-L-rhamnosidase and concluded that 25 to S O percent of the L-rhamnose residues were present as aL-linked terminal groups.

4 . Noncarbohydrate Components Amino acids have been found in hydrolyzates of soil polysaccharides whenever the appropriate tests have been applied (Duff, 1952a; Whistler and Kirby, 1956; Bernier, 1958a: Muller et af., 1960; Mortensen, 1961; Parsons and Tinsley, I96 1 : Thomas et al., 1967; Oades and Swincer, 1968; Swincer et al., 1968a,b). The fact that complete deproteinization of the preparations has never been reported suggests that at least a portion of the polysaccharides extracted from a wide range of soils may be in the form of glycoproteins or proteoglycans. In addition, an association of the colored humic substances with some of the polysaccharides seems likely. In the rare cases of complete removal of colored materials from soil polysaccharides-e.g., by Sephadex gel filtration (Barker et af., 1965, 1967) or by charcoal filtration (Forsyth, 1947, 1950; Whistler and Kirby, 1956; Swincer et af., 1968a)-a considerable amount of polysaccharide material was lost as well suggesting the possibility of chemical linkage to the colored “impurities.”

5 . Physicochemicaf Properties

In the absence of any very informative structural investigations of the polysaccharide materials isolated from soils (Section 111, E, 3) the only indication of the nature of the materials in their polymeric state has been provided by a few physicochemical studies. The approaches used include free-boundary and continuous paper electrophoresis, viscosity measurements, ultracentrifuge studies, and titration experiments. The main observation has been the extreme heterogeneity of all soil polysaccharide fractions so far examined. Free-boundary electrophoresis in borate and phosphate buffer has usually shown a single broad peak with several superimposed ridges (Whistler and Kirby, 1956; Clapp, 1957; Bernier, 1958a; Thomas, 1963; Barker et al., 1965). Boundaries became diffuse in most cases, indicating a heterogeneous mixture of polysaccharides. Thomas ( 1963) used free boundary electrophoresis to examine polysaccharide fractions eluted

218

G. D. SWINCER, J. M. OADES, A N D D. J. GREENLAND

from a DEAE-cellulose column. Fractions with high electrophoretic mobility showed high uronic acid contents. Using continuous-flow paper electrophoresis in phosphate and borate buffers, Mortensen (1 96 1) and Mortensen and Schwendinger ( 1963) obtained peaks corresponding to uronic acid components, but separation was not complete. Complex formation in borate buffer increased the electrophoretic mobilities of the polysaccharides, but at the same time the separation of the uronic acid components was obscured even more. Viscosity and ultracentrifuge measurements (Bernier, 1958a; Ogston, 1958; Swincer et al., 1968a) have shown that the polysaccharides purified from phosphate buffer and sodium hydroxide extracts of soils were polydisperse and anisometric. With one particular preparation, molecular weights of 124 X lo3 and 130 X lo3 were obtained, respectively, by viscosity and sedimentation procedures. Ogston suggested that the average particle approximated to an elongated, stiff, and little solvated rod with an axial ratio of 80. In another instance of the use of viscometry, Mortensen (1 96 I ) obtained a rough estimate of 450,000 for the molecular weight of a soil polysaccharide that had been obtained by hot water extraction. Gel filtration has shown that a continuum of molecular sizes occurs in soil extracts from molecules of oligosaccharide size to materials of molecular weight above 200,000 (Section 111, D, 2, d). According to Mortensen ( 1 96 l ) , temperature and salt effects on the shape of titration curves indicated that his soil polysaccharide was probably a hydrogen-bonded polyuronide containing contaminants. Minima in viscosity curves suggested an isoelectric point near pH 4.9.

6. Aggregate Stabilizing Ability The properties of extracted polysaccharides in terms of aggregate stabilizing ability have been examined by several authors (Rennie et al., 1954; Dubach et al., 1955; Whistler and Kirby, 1956; Bernier, 1958a) and generally an increase in stability has been reported. Apart from this little information concerning the polysaccharides themselves is obtained from such studies.

F. THEORIGIN, SYNTHESIS A N D DECOMPOS~TION OF SOILPOLYSACCHARIDES 1. Origin and Synthesis Most soil polysaccharides are undoubtedly originally derived from plant materials, but it seems clear that very few unchanged plant poly-

SO I L POLY SACC H A RID ES

219

saccharides exist in the soil at any one time. Almost as quickly as the plant materials are added to most soils, they are decomposed by the microbial population (Kononova, 196 I ; Russell, I96 1 ; Simonart, 1964; I. L. Stevenson, 1964; Burges, 1967). Even a chemically resistant polymer such as cellulose is quite rapidly degraded by the complex of endoand exoenzymes possessed by soil fungi, bacteria, and actinomycetes (Siu, 195 1 ; Mayaudon and Simonart, 1959; Gascoigne and Gascoigne, 1960; Alexander, I96 1 ). The other plant polysaccharides, such as starch, hemicellulose, pectic substances, gums, and mucilages, are even less stable. It would, therefore, be reasonable to assume that the polysaccharide materials present in soils are largely the products of microbial metabolism. A possible exception is the mucilaginous materials secreted by the root cap cells of many plants (Samtsevich, 1965, and other references given at the end of this section). Evidence that the polysaccharides that have been isolated from soils are, in fact, microbial products is given by their relatively low content of glucose and xylose units and the abundance of mannose, rhamnose, and hexosamines, sugars not often abundant in higher plants. However, the strongest and most direct evidence has been provided by a study using substrates labeled with radioactive carbon. Keefer and Mortensen ( 1 963) examined the distribution of radioactivity among different sugars in a polysaccharide extracted from a soil after incubation with labeled glucose or alfalfa tissue. All sugars isolated were found to be labeled although there were considerable variations in specific activity. Clearly soil organisms are able to synthesize from glucose or plant tissue, polysaccharides containing each of the sugars examined (glucose, arabinose, galactose, fucose, mannose, rhamnose, xylose, and uronic acids). In addition, changes in the specific activity of the different sugars with time indicated that not only the soil polysaccharides but also their component sugars undergo continual degradation and resynthesis. It follows that noncarbohydrate materials may sometimes act as precursors for soil polysaccharides. The complex exocellular and capsular polysaccharides synthesized by many bacteria isolated from the soil and often capable of binding soil particles into stable aggregates have been analyzed in terms of component sugars (J. P. Martin, 1945b; Forsyth and Webley, 1949; Forsyth, 1954; Webley et al., 1965), but as with most other microbial polysaccharides, very little information exists relating to the more detailed structures. However, studies with pure cultures are not likely to throw much light on the situation in the soil itself. The complexity of the natural soil environment almost certainly ensures that at least the exocellular poly-

220

G. D . SWINCER, J . M. OADES, A N D D. J . GREENLAND

saccharides produced within it differ considerably from those obtained in the laboratory from either pure or mixed cultures, even using the same carbon substrates. Many of the polysaccharides that persist in soils are probably derived from the tissues of recently dead microbial cells, particularly the cell walls (Whistler and Kirby, 1956; Whitehead and Tinsley, 1963; Swincer et al., 1968a,b). It is well established that hexosamines and other sugars are important constituents of the polymers that make up the cell walls of all bacteria (Salton, 1960, 1964; Weidel and Pelzer, 1964; Rogers, 1966) and actinomycetes (Stacey and Barker, 1960), while chitin, a polysaccharide composed predominantly, if not entirely, of N-acetylglucosamine residues (Foster and Webber, 1960) is the main constituent of fungal cell walls (Stacey and Barker, 1960; Barker, 1963). The influence of the nature of the available plant substrates on the kind of polysaccharides synthesized in soils is not understood. The kind of crop growing on the soil would be expected to have some effect. There is some compositional evidence (Swincer et al., 1968b) that mucilages (F. M. Scott et al., 1958, 1963; Dawes and Bowler, 1959; Estermann and McLaren, 196 1 ; Jenny and Grossenbacher, 1963; Dart and Mercer, 1964; Jones et al., 1966; Samtsevich, 1965; Morre et al., 1967; Rovira and McDougall, 1967) and other organic substances (Rovira, 1956, 1962; Harmsen and Jager, 1962; Rovira and McDougall, 1967) exuded by growing roots are no less important as precursors of soil polysaccharides than decaying plant tissues. 2 . Decomposition The high proportion of soil organic matter represented by polysaccharides of microbial origin is an indication that some at least of these materials are relatively resistant to decomposition. The materials isolated from a soil by Whistler and Kirby ( 1 956) were found to be more resistant to decomposition than many plant polysaccharides. Furthermore, the polysaccharides produced in pure culture by certain soil microorganisms (e.g., Chromobacterium violaceum and Azotubacter indicus) are not attacked very rapidly when added to soils (J. P. Martin, 1945b, 1946; J. P. Martin and Richards, 1963; J. P. Martin et al., 1965). However, the polysaccharides produced by many other soil microorganisms are subject to extremely rapid breakdown, up to 75 percent of their carbon having been converted to C o nafter incubation for 4 weeks (J. P. Martin and Richards, 1963; J. P. Martin et al., 1965, 1966). In contrast only about 10 percent of the carbon of the polysaccharide from Azotobacter indicus was lost in the same time. Many soil organisms must be

SOIL POLYSACCHARIDES

22 I

able to produce enzymes capable of attacking a wide range of polysaccharides, probably including the cell-wall and cytoplasmic polysaccharides of soil organisms, which, except for chitin, have not been studied in this respect. Chitin is readily decomposed by a large number of soil organisms (Bremner and Shaw, 1954: Alexander, 1961). The relative persistence of microbial polysaccharides in soils is complicated by the possibility that resistance to breakdown may be enhanced by their adsorption on the surfaces of clay particles (Finch et al., 1967; Lynch and Cotnoir, 1956; Lynch et af., 1956; Greenland, 1956) or by their reaction with metal ions (J. P. Martin et af., 1966). Martin et al. found that, in general, copper exerted the greatest inhibitory effect on the decomposition in soil of plant gums and the capsular polysaccharides of soil bacteria, but with some polymers the zinc, and with others, the iron salts were most resistant. It is also possible that some polysaccharides are mechanically shielded from microbial attack, being present in parts of the soil that are inaccessible to soil organisms (Rovira and Greacen, 1957). The influence of soil type and such environmental variables as soil temperature, moisture content, aeration, pH, nutrient status, and agronomic treatment upon the decomposition (and synthesis) of the soil polysaccharides is at best only poorly understood. Presumably these factors operate with polysaccharides very much as they do with other organic materials. For example, highly aerobic conditions favor rapid metabolic activity and the conversion of a large part of the organic materials to COr; decomposition is much slower under anaerobic conditions. Harris et af. ( 1963) demonstrated that the aggregate stabilizing materials (presumably polysaccharides) that were synthesized in a sucrose-amended soil under both aerobic and anaerobic conditions decomposed rapidly in the presence of oxygen, but persisted for as long as oxygen was excIuded from the system. Clearly cultivation, which improves soil aeration, would be expected to encourage decomposition of the pol ysaccharides. Additional consequences of aeration could be (a) increased production of polysaccharide decomposing enzymes as a result of general stimulation of the microbial population, and (b) the exposure to microbial attack of polysaccharides previously protected by mechanical shielding (Rovira and Greacen, 1957). Temperature also exercises an important control over microbial activity, the metabolic rate of the microbial population as a whole increasing with temperature up to about 37°C. J . P. Martin and Craggs ( 1946) found that aggegation caused by the indigenous microflora reached

222

G. D. SWINCER, J. M. OADES, A N D D. J. GREENLAND

a maximum earlier, but was destroyed more rapidly as the incubation temperature was increased through the range 10” to 50°C. It was suggested that the microbial population at low temperatures produced more or “better quality” aggregating substances (probably polysaccharides) whereas the high temperatures favored rapid decomposition of the soilbinding agents. Similar observations were made by Harris et al. ( 1 966). In spite of various protective mechanisms, most of the polysaccharides produced in normal well-aerated soils are unlikely to persist for very long. Some of their decomposition products may act as precursors for the synthesis of new polysaccharides, but inevitably much of the carbon is lost as COr. The high level of polysaccharides present in many soils must be due largely to an adequate supply of plant materials. The timing of the addition of plant substrates in relation to other environmental factors is likely to have an important bearing on the composition and quantity of the resultant microbial polysaccharides. Studies on one red-brown earth revealed that seasonal variations in the composition of soil polysaccharides were very small although differences in the ease of extraction of the polysaccharides during the year were larger (Oades and Swincer, 1968). Such evidence as is available indicates that differences in the polysaccharides from different Great Soil Groups and in their monosaccharide composition are also small (Forsyth, 1954; Parsons and Tinsley, 1961; Lowe, 1968). IV. Methods for the Analysis of Complex Polysaccharide Materials

A. INTRODUCTION Progress toward a full understanding of soil polysaccharides has been severely limited by inadequacies in both analytical techniques and separation methods. The availability of suitable analytical techniques is of fundamental importance not only for the characterization of polysaccharide preparations, but also for the development of satisfactory purification procedures. Unless each step can be followed quantitatively there can be no certainty regarding the effectiveness of a particular extraction, purification, or fractionation method. Quantitative methods are necessary for tracing the source of any loss of polysaccharide material or of any damage to the polymers during the separation procedure. Quantitative physical and chemical analyses are rarely very simple, even with polysaccharide materials that are pure and homogeneous, and with complex mixtures containing both carbohydrate and noncarbohydrate materials the difficulties become quite formidable. Many of these difficulties have been investigated in some detail because important natural

SOIL POLYSACCHARIDES

223

polysaccharides from sources other than soils often occur in complex mixtures containing other carbohydrates and a variety of impurities. The work on glycoproteins (Neuberger and Marshall, 1966a,b; Neuberger et al., 1966; Gibbons, 1966; Spiro, 1966), mucopolysaccharides (Jeanloz, 1963; Schmid, 1964; Brimacombe and Webber, 1964; Davidson, 1966), plant gums and mucilages (Smith and Montgomery, 1959); and microbial polysaccharides (Stacey and Barker, 1960; Barker, 1963; Rogers, 1966) is most relevant to studies on soil polysaccharides. B. ANALYTICAL METHODS

I . Quantitative Analysis of Component Sugars a. General. Most techniques used for the quantitative estimation of a given sugar either in a polysaccharide or in some other carbohydratecontaining macromolecule or in a mixture of these polymers require the liberation of the monosaccharide by hydrolysis of a glycosidic linkage. This is due to the fact that, in general, the only groups present in the sugar moieties of intact polysaccharides which can be recognized by physical and chemical methods are the hydroxyl groups, and these are common to all types of sugars. Sometimes there are carboxyl, acetamide, phosphonic acid, sulfonic acid, and methoxyl groups, all of which can be measured without hydrolysis of the polymer. However, the parent sugars cannot be identified unless they are recovered intact, either as the free monosaccharides or as suitable stable derivatives. In certain cases use can be made of quantitative colorimetric procedures that do not require prior release of the sugars as a separate step; color formation occurs concurrently with the liberation of the sugars, and consequently there is no risk of destruction of the monosaccharides during hydrolysis. These procedures are of particular significance in the measurement of acid-labile sugar components (e.g., uronic acids, by the carbazole method of Dische, 1947) or for the routine assay of some of the polysaccharide materials that contain only a few components (e.g., neutral sugars, by the orcinol method of Winzler, 1955). However, these procedures are not very specific, and for this reason the results are very dependent upon the purity and sugar composition of the polysaccharide material. b. The Problem of Hydrolysis. In order to liberate sugars from complex polysaccharides, hydrolysis by acid is employed at present almost invariably. However, in the free state, all commonly occurring sugars are more or less unstable in hot acid, and the glycosidic bonds joining different sugars show different stabilities to hydrolysis by acid, some common linkages being particularly stable.

224

G. D. SWINCER, J . M . OADES, A N D D. J . GREENLAND

The acid stabilities of the various classes of monosaccharides vary greatly, the approximate order being: hexosamines > hexoses > deoxyhexoses > pentoses > uronic acids. The particular acid used has some bearing on this order: hexosamines are more quickly destroyed in HzS04 than in HCl of equal concentration whereas the reverse applies for the other groups. With amino sugars the exclusion of oxygen may greatly reduce the extent of destruction (e.g., see Walborg and Ward, I963), and the presence of heavy metals may lead to increased destruction (Hartree, 1964). An additional complication is the possibility of undesirable side reactions, such as “acid reversion” [the acid-catalyzed reaction of the reducing group of a liberated monosaccharide with the primary or even secondary hydroxyl groups of another sugar molecule to give disaccharides or oligosaccharides (Pigman, 1957; Whelan, 1960; Overend er al., 1962)], or interaction of the free sugars with amino acids (Francois et al., 1962; Gottschalk, 1966). This last point may be particularly significant in the analysis of soil polysaccharides which have always been found associated with polypeptide materials. Neuberger and Marshall ( 1 966a) consider that it is at present impossible to say whether these possible sources of error affect seriously the analytical results. The reactions being bimolecular, their effect can be minimized by carrying out the hydrolysis at low substrate concentration. Glycosidic bonds in polysaccharides are hydrolyzed at rates that are dependent largely on the nature of the sugar supplying the anomeric carbon atom of the linkage (Wolfrom and Thompson, 1957; Overend er al., 1962; Adams, 1965). Thus furanoside linkages are more labile than pyranoside linkages (Haworth and Hirst, 1930; Shafizadeh, 1958; Reichstein and Weiss, 1962); alpha glycosidic bonds are usually more stable than beta (Wolfrom and Thompson, 1957; Overend et al., 1962); with pyranoside linkages, pentoses and 2-deoxyhexoses allow easier hydrolysis than ordinary aldohexoses; and increased resistance to hydrolysis is conferred by uronic acid groups (Smith and Montgomery, 1959) and also by amino sugars (Gottschalk and Ada, 1956; Johansen et al., 1960). In practice, no perfect conditions for hydrolysis have been devised whereby it is certain that all glycosidic linkages of polysaccharides are cleaved and, at the same time, all the monosaccharides itre still intact at the end of the acid treatment. However, by dividing the sugars into groups, a set of reasonably satisfactory conditions can be derived for all groups except the uronic acids and possibly the pentoses. The optimum hydrolysis conditions for each group vary from one polysaccharide to another and must be ascertained by preliminary experiments in each

SOIL POLYSACCHARIDES

225

particular case. In general it can be said that HCI (4N to 8 N ) has been most satisfactory with hexosamines; H z S 0 4( 1 N or 2 N ) with hexoses; and more dilute acids, such as 0.1 N H2S04or HCI, with deoxyhexoses. The uronic acids and to a lesser extent, the pentoses represent a special problem which at present has no adequate solution. Perry and Hulyalkar ( 1965) state that even in the most favorable cases recoveries of polymerbound uronic acids rarely exceed 70 percent and, where the more acidlabile uronic acids are involved, the yields are considerably below this figure. Little is known about the best conditions for the release of pentoses. Although losses are often observed, it seems that reasonable recoveries are obtainable as long as optimum conditions are established (Saeman el al., 1954; Ivarson and Sowden, 1962; Cheshire and Mundie, 1966). The N-acetylamino sugars also represent a problem, although this is not usually very difficult to solve. Under strong acid conditions the polysaccharides containing these sugars are de-N-acetylated rapidly, and subsequent hydrolysis of glycosidic bonds is inhibited by the resulting free amino groups which, being positively charged in acid solution, confer some protection on the adjacent bonds (Moggridge and Neuberger, 1938; Johansen et at., 1960). However, by diluting the acid, it is possible to reduce the rate of de-N-acetylation relative to that of glycosidic splitting, and thus to recover a high proportion of the original N-acetylamino sugars for the purpose of identification. Once the sugars have been identified, they can be determined quantitatively either (a) by measuring the total N-acetyl content of the polysaccharide material, or (b) (if it is certain that in the polymer all amino sugars are N-acetylated) by determining the free (de-N-acetylated) amino sugars following their complete release by hydrolysis with 4-8 N HCI. A technique with important possibilities for measuring losses of sugars during hydrolysis is the isotope dilution method employed by Frdncois et al. ( 1 962). c . Analysis of Polysaccharide Hydrolyzates. Methods for the separation, identification, and determination of the monosaccharides in hydrolyzates of polysaccharides have been reviewed thoroughly elsewhere (Percival, 1963; Bishop, 1964; Davidson, 1966; Neuberger and Marshall, 1966a; Northcote, 1966; Spiro, 1966). Chromatographic techniques are of primary importance for the separation, detection, and preliminary identification of sugars. However, chromatographic behavior alone does not allow unequivocal identification of an individual sugar; this requires either isolation of the pure sugar in crystalline form or conversion of the sugar to a characteristic crystalline derivative.

226

G . D. SWINCER, J. M. OADES, AND D. J. GREENLAND

Satisfactory removal of the hydrolyzing acid prior to chromatography is often difficult. Volatile acids are removed quite simply by evaporation, although with HCI, particularly in the presence of heavy metals, destruction of some sugars can occur. Strong mineral acids are probably best removed with strong anion exchange resins in the carbonale, bicarbonate, or acetate forms. However, precipitation with barium hydroxide or barium carbonate is still used regularly for the removal of HzS04 in spite of the danger of selective adsorption of sugars by the precipitate. An unusual approach with definite possibilities for the removal of H a S o l is selective extraction of the sulfate with an immiscible organic liquid (Becker and Shefner, 1964). Paper chromatography which has been the most useful routine method for rapid preliminary characterization of sugar mixtures has not proved completely adequate for precise quantitative work. For this reason it is being gradually superseded by such methods as column chromatography with ion exchange resins and gas-liquid chromatography, which offer several distinct advantages. Ion exchange chromatography has been applied successfully not only to the charged sugars (amino sugars and uronic acids), but also, by formation of their borate complexes, to neutral sugars. Fully automated procedures have been developed, and the method can be used as a preparative technique for milligram quantities of sugars. Gasliquid chromatographic analysis of sugars has been developed only very recently, and it is almost certain that further advances will be made. This technique offers the possibility of rapid and precise quantitative analysis of very small samples containing a complex mixture of sugars (Oades, 1967a). Except for gas-liquid chromatography, most of the methods used for quantitative determination of the separated sugars are colorimetric. Because of the relatively low specificity of some of the color reactions, special care is required to ensure that all interfering substances are taken into account. Allowance also has to be made for the variations in color yield from sugar to sugar. Different methods are needed at least for each class of monosaccharide (e.g., hexoses, pentoses, uronic acids, hexosamines). Specific enzyme assays have been worked out for a few monosaccharides. As further assays are developed, this approach is likely to become useful for the analysis of complex hydrolyzates since there is normally no need for prior separation of the sugars. 2 . Physicochemical Analysis. Most of the methods generally used on high polymers for the determination of molecular size, shape and flexibility can be used with polysaccharides. They may be listed as: (a) hydrodynamic methods (e.g.,

SOIL POLYSACCHARIDES

227

sedimentation analysis by ultracentrifugation, viscometry, streaming birefrigence measurements, determination of diffusion constants); (b) methods based on the colligative properties of the molecules (e.g., osmometry, isothermal distillation); (c) methods involving measurements which depend directly on the physical size of the molecules (e.g., light scattering); (d) end-group determination by chemical assay; and (e) techniques that are fundamentally separative rather than analytical in nature (e.g., gel filtration, ultrafiltration through membranes of graded pore diameter, electrokinetic ultrafil tration, density gradient centrifugation, free-boundary electrophoresis). The application of these methods to polysaccharides has been reviewed in considerable detail (Greenwood, 1952, 1956; Whistler and Smart, 1953; Whistler and Corbett, 1957; Banks and Greenwood, 1963; Horton and Wolfrom, 1963; Gibbons, 1966). For thorough characterization of a polysaccharide, as many methods as possible should be combined. Many of the methods yield reliable quantitative information only with preparations that are homogeneous (i.e., “consisting of molecules having identical structure but not necessarily the same molecular weight” Banks and Greenwood, 1963), and satisfactory interpretation of the results often demands, in addition, a narrow distribution of molecular weight. For example, the diffuse sedimentation boundary produced during ultracentrifugation of a very polydisperse sample allows computation only of an approximate value for the average molecular weight. These points are of obvious significance with respect to the analysis of soil polysaccharide preparations which have proved particularly difficult to purify and fractionate. Clearly, it is imperative either to ensure that homogeneous polysaccharides have been prepared before attempting physicochemical characterization or to use only the limited range of techniques that can be applied satisfactorily to heterogeneous preparations. Undoubtedly for the most complete and precise information homogeneous polymers must be obtained, but the problems involved both in isolation of the required fraction and in assessment of its homogeneity are extremely difficult. In fact, Banks and Greenwood ( 1 963) pointed out that “it is doubtful if any polysaccharide has been examined by sufficient methods to prove unambiguously that it is homogeneous.” The techniques that can be applied most satisfactorily to the characterization of polysaccharides in heterogeneous mixtures are those based on separation methods.

C. SEPARATION METHODS A wide variety of methods have been used for the isolation of individual polysaccharides from biological materials (Whistler and Smart, 1953;

228

G. D. SWINCER, J. M. OADES, A N D D. J . GREENLAND

Pigman and Platt, 1957; Bouveng and Lindberg, 1960; J . E. Scott, 1960; Banks and Greenwood, 1963; Barker, 1963; Horton and Wolfrom, 1963: Jeanloz, 1963; Kertesz, 1963; Brimacombe and Webber, 1964; Whistler, 1965; Northcote, 1966). Initial extraction is usually the most critical step because it is often at this point that the most drastic treatments are needed. Separation of polysaccharides from cellular material is rarely easy, and the problem is especially formidable with polysaccharides that are associated intimately with an insoluble matrix as in cell walls. There is an obvious relationship here with the problem of separating polysaccharides from the mineral matrix of soils. With almost all the procedures sufficiently powerful to solubilize such polysaccharides, there is a definite risk of degradation (Northcote, 1966). Any modification of the structure of the molecules or the molecular weight distribution, or both, may invalidate many of the subsequent analyses. Once the polysaccharides have been brought into solution, they can be purified and fractionated by a variety of techniques. Fractional precipitation or dissolution of polysaccharides (and polysaccharide acetates or nitrates), either by changing the solvent composition, or pH, or temperature, has been widely used. With complex mixtures this approach is only of limited application, except for the removal of extraneous material, because of the tendency to coprecipitation and occlusion of other polymers. Moreover, in most cases fractional precipitation merely subdivides the polymolecular system into fractions on a molecular weight basis: each individual fraction represents a narrow molecular weight range, but still remains a mixture of polysaccharide types. Gel filtration is assuming a place of primary importance in the study of heterogeneous polydisperse systems. It gives very efficient separation of polydisperse materials into fractions covering a limited range of molecular weight and also provides a simple and effective method for removing low-molecular-weight impurities (Granath and Flodin, 196 I ; Anderson et al., 1965; Anderson and Stoddart, 1966; Granath and Kvist, 1967). Separation techniques based on the ionic properties of the polysaccharides offer the best prospects for isolation of homogeneous materials. These methods are amenable to all soluble polysaccharides as even those polysaccharides with no readily ionizable groups are generally slightly charged, particularly in alkaline solution, due to ionization of the hydroxyl groups, while complex formation with certain ions, notably borate ions, increases the negative charge on a polysaccharide. The techniques that have been applied successfully include selective precipitation with metal ions or quaternary ammonium salts, and ion exchange chromatography, particularly with charged celluloses. These procedures can generally be

SOIL POLYSACCHARIDES

229

made very sensitive to small differences in the net charge of the polymers and they are often capable of resolving mixtures of closely related polysaccharide species. The ion exchange procedures in particular have given some excellent separations (e.g., Jermyn, 1962; Antonopoulos et al., 1967) and the technique offers considerable scope for further refinement. In addition, some very good separations of mucopolysaccharides have been achieved recently by a combination of fractional precipitation and column chromatography (e.g., Antonopoulos et at., 1964; Pearce and Mathieson, 1967). Other procedures with distinct possibilities for the fractionation of mixtures of polysaccharides are density gradient centrifugation (e.g., Charlwood, 1966; Franek and Dunstone, 1967) and selective precipitation with antisera (e.g., Heidelberger et af., 1955), neither of which are based directly on differences in polymer size or charge. V. Summary and Conclusions

Carbohydrates represent 5 to 25 percent of soil organic materials. They consist of a wide range of monosaccharides, such as hexoses, pentoses, deoxy- and 0-methyl sugars, uronic acids, and amino sugars. Such monosaccharides exist in polymeric molecules of various sizes and degrees of complexity, which are associated more or less strongly with inorganic colloids in soils. Large proportions of carbohydrates in many soils are present in partly decomposed plant and animal remains. Glucose, presumably in the form of cellulose, is dominant in such materials. Plant litter and roots, either living or dead, are the main primary source of soil carbohydrates, but the composition of soil polysaccharides, apart from obvious plant remains, would suggest a microbial origin, either wholly or in part, e.g., plant materials which have been modified by the soil flora and fauna. Polysaccharides have been extracted from soils by many different chemical reagents, and recently methods have been devised that enable most of the carbohydrates to be isolated from other soil materials. The extracted polysaccharides show a continuum of molecular sizes and contain a wide range of neutral and charged monosaccharides, amino acids, and other unidentified nitrogenous and acid components. Carbohydrates from different soils are similar in chemical composition suggesting that the microbial population of different soils is qualitatively similar. Many methods have been used to fractionate extracted soil polysaccharides usually with limited success. The most successful methods have been based on gel filtration and chromatography on charged supports

230

G . D. SWINCER, J . M. OADES, AND D. J . GREENLAND

such as cellulose. However, fractions obtained are still complex and contain a range of different components. This complexity is not surprising in view of the wide range of substrates, organisms, and metabolic products of organisms that are subjected to chemical extraction and fractionation procedures. Generally the “turnover” of sugars in soil carbohydrates appears to be rapid, but some microbial polysaccharides are resistant to breakdown by soil organisms. The subject is complicated because of interactions with metal cations and sorption on colloid surfaces. The composition of soil polysaccharides suggests that in soils they may carry charged sites and take part in exchange reactions and act as energy sources for heterotrophic organisms. However, the main stimulus for the study of soil polysaccharides has arisen from repeated indications of their favorable influence on soil physical conditions. Much work has been directed toward this aspect, and it has been shown that microbially produced soil polysaccharides are capable of stabilizing soil aggregates against dispersion in water. N o specific fraction has yet been definitely identified as particularly active, but it is suggested that the larger polysaccharides produced by microorganisms in coarse pores of aggregates are likely to be the most effective. The mechanisms by which these polymers react with inorganic colloids is not understood, but the complex preparations obtained from soils are sorbed from aqueous solution by clay materials, and further work on the fractionation of carbohydrate preparations followed by studies of the sorption of these “purer” characterized fractions will undoubtedly prove to be worthwhile. Methods for the isolation of polysaccharides from other soil materials in good yield are now available and methods for the analysis of the extracted polysaccharides have been developed by carbohydrate chemists. Combinations of these techniques in the future will enable new information about the composition, origin, and function of soil carbohydrates to be obtained. Particularly useful information should arise from the cooperation of chemists and microbiologists using techniques involving isotopically labeled materials. REFERENCES Acton, C. J., Paul, E. A., and Rennie, D. A . I963a. Can. J . Soil Sci. 43,14 I - I 50. Acton, C. J . , Rennie, D. A., and Paul, E. A. 1963b. Can. J . Soil Sci. 43, 201-209. A d a m , G . A. 1965. Methods Carbohydrate Chem. 5,269-276. Alexander, M. 196 I . “Introduction to Soil Microbiology,” Wiley, New York. Allison, L. E. 1947. Soil Sci. 63,439-450.

SOIL POLYSACCHARIDES

23 1

Anderson, D. M. W.. and Stoddart, J. F. 1966. Carbohydrate Res. 2,104-1 14. Anderson, D. M. W., Dea, I . C. M.. Rahman, S., and Stoddart, J. F. 1965. Chem. Commun. pp. 145- 146. Antonopoulos, C. A.. Gardell, S., Szirmai, J . A., and D e Tyssonsk, E. R. 1964. Biochim. Biophys.Acta 83,1-19. Antonopoulos, C. A., Fransson, L - k , Heinegird, D., and Gardell, S. 1967. Biochim. BiophyS.ACta 148,158-163. Banks, W., and Greenwood, C. T. 1963.Advan. Carbohydrate Chem. 18,357-398. Barker, S. A. 1963. Comprehensive Biochem. 5,246-26 1. Barker, S. A., Finch, P.. Hayes, M. H. B., Simmonds, R. G., and Stacey, M. 1965. Nature 205,68-69. Barker, S. A., Hayes, M. H. B., Simmonds, R. G., and Stacey, M. 1967. Carbohydrate Res. 5,13-24. Barrett, A. J., and Northcote, D. H. 1965. Biochem.J. 94,617-627. Becker, M. J., and Shefner, A. M. 1964. Nature 202,803. Bernier, B. 1958a. Biochem. J . 70,590-598. Bernier, B. I958b. Can. J . Microbiol. 4, 195-204. Bishop, C. T. 1964.Advan. Carbohydrate Chem. 19,95. Black, W . A. P.. Cornhill, W. J., and Woodward, F. N. 1955. J . Appl. Chem. (London), 5,484-492. B0bbit.J. M. 1956.Advan. Carbohydrate Chem. 11,1-41. Bouveng, H. O., and Lindberg, B. 1960.Advan. CarbohydrateChem. 15,53-89. Bremner, J. M. 1950.J. SoilSci. 1, 198-204. Bremner, J. M. 1954. J. Soil Sci. 5 , 2 14-232. Bremner, J . M., and Harada, T . 1959. J.Agr. Sci. 52, 137- 146. Bremner, J. M., and Lees, H. 1949. J.Agr. Sci. 39,274-279. Bremner, J. M., and Shaw, K. l954.J. Agr. Sci. 44,152-159. Brimacombe, J. S.,and Webber, J . M. 1964. “Mucopolysaccharides,” Elsevier. Amsterdam. Brink, R. H., Dubach, P., and Lynch, D. L. 1960. Soil Sci. 89, 157-166. Burges, A. 1967. I n “Soil Biology”(N. A. Burges and F. Raw, eds.), pp. 479-492. Academic Press, New York. Charlwood, P. A. 1966. Brit. Med. Bull. 22, 121-126. Cheshire, M. V., and Mundie, C . M. 1966. J. Soil Sci. 17,372-381. Chesters, G . , Attoe, 0. J., and Allen. 0. N . 1957. Soil Sci. Soc. Am. Proc. 26,466-469. Choudri, M. B., and Stevenson, F. J. 1957. Soil Sci. Soc. Am. Proc. 21,508-5 13. Clapp, C . E. 1957. Ph.D. Thesis, Cornell University, New York. Clapp, C. E., and Emerson, W. W. 1965. Soil Sci. Soc. Am. Proc. 29,127- 134. Clapp. C. E., Davis, R. J., and Waugaman, S. N. 1962. Soil Sci. Soc. A m . Proc. 26,466-469. Daji, J. A. 1932. Biochem. J. 26, 1275-1280. Dalgleish, C. E. 1955. J. Clin. Pathol. 8,73-78. Dart, P. J., and Mercer, F. V. 1964.Arch. Mikrobiol. 47,344-378. Davidson, E. A. 1966. Methods Enzymol. 8,52-60. Dawes, C. J . , and Bowler, E. 1959. Am. J . Botany 46,561-565. Deshpande, T. I-., Greenland. D. J.. and Quirk, J . P. 1968. J . Soil Sci. 19, 108- 122. Deuel, H . , Dubach, P., and Mehta, N. C. 1960. Sci. Proc. Roy. Dublin Soc. A l , I I S 121.

Dische, 2. 1947. J . B i d . Chem. 167, 189-198.

232

G . D. SWINCER, J. M. OADES, A N D D. J. GREENLAND

Dormaar, J. F. 1967. Soil Sci. 103,4 17-423. Dubach, P., and Mehta, N . C. 1963. Soils Fertilizers 26,293-300. Dubach, P., Zweifel, G., Bach, R., and Deuel, H. 1955. Z . Pjianzenernaehr. Dueng. Bodenk. 69,97- 108. Duff, R. B. 1952a.J. Sci. FoodAgr. 3, 140-144. Duff, R. B. 1952b. Chem. & lnd. (London) p. 1104. Duff, R. B. 1954. Chem. & Ind. (London) p. 15 13. Duff, R. B. 1961. J . Sci. Food Agr. 12,826-83 I. Erskine, A. J., and Jones, J . K. N. 1956. Can. J . Chem. 34,821-826. Estermann, E. F., and McLaren, A. D. 196 I.Plant Soil 15,243-260. Evans, L. T. 1959. J . Soil Sci. 10, I10-1 18. Finch, P., Hayes, M. H. B., and Stacey, M. 1967. Intern. Soc. Soil Sci., Trans. Commun. I 1 and I V , Aberdeen, 1966, pp. 19-32. Intern. SOC.Soil Sci. Ford, G. W., Greenland, D. J., and Oades, J. M. 1969. J . Soil Sci. 20, (in press). Forsyth, W. G. C. 1947. Biochem. J . 41, 176-181. Forsyth, W. (3. C. 1950. Biochem. J . 46, 141-146. Forsyth, W. G. C. 1954. Trans. 5th Intern. Congr. Soil Sci., Leopoldville, 1954 Vol. 3. pp. 119-122. Forsyth, W. G . L a n d Webley, D. M. 1949. J . Gen. Microbiol. 3,395-399. Foster, A. B., and Webber, J. M. 1960. Advan. Carbohydrate Chem. 15,37 1-393. Foster, A. B., Newton-Hearn, P. A., and Stacey, M. 1956.5. Chem. Soc. pp. 30-36. Francois, C., Marshall, R. D., and Neuberger, A. 1962. Biochem.1. 83,335-341. Franek, M. D., and Dunstone, J. R. 1967. J . Biol. Chem. 242,3460-3467. Gascoigne, J. A., and Gascoigne, M. M. 1960. “Biological Degradation of Cellulose.” Butterworth, London and Washington, D.C. Geoghegan, M. J . , and Brian, R. C. 1946. Nature 158,837. Geoghegan, M. J., and Brian, R. C. 1948. Biochem. J . 43, 14. Gibbons, R. A. 1966. In “Glycoproteins” (A. Gottschalk, ed.), pp. 19-95. Elsevier, Amsterdam. Gottschalk, A. 1966. In “Glycoproteins” (A. Gottschalk, ed.), pp. 96-1 I I. Elsevier, Amsterdam. Gottschalk, A., and Ada, G. L. 1956. Biochem. J . 62,681-686. Granath, K. A., and Flodin, P. 1961. Makromol. Chem. 48, 160-171. Granath, K. A., and Kvist, B. E. 1967. J . Chromatog. 28,69-81. Graveland, D. N., and Lynch, D. L. 1961. Soil Sci. 91, 162-165. Greenland. D. .I.1956. J . Soil Sci. 7, 319-334. Greenland. D. J., Lindstrom, G. R., and Quirk. J. P. 1961. Natrirr 191, 1283-1284. Greenland, D. J., Lindstrom, G. R.,and Quirk, J. P. 1962. Soil Sci. Soc. A m . Proc. 26, 366-37 I . Greenwood, C. T. 1952. Advan. Carbohydrate Chem. 7,289-332. Greenwood, C. T. 1956. Advan. Carbohydrate Chem. 11,336-393. Griffiths, E. 1965. Biol. Rev. Cambridge Phil. Soc. 40, 129-142. Griffiths, E., and Jones, D. 1965. Plant Soil 23, 17-33. Gupta, U. C. 1967. In “Soil Biochemistry” (A. D. MacLaren and G . H. Petersen, eds.), pp. 91-1 18. Arnold, London. Gupta, U. C., and Sowden, F. J. 1964. Soil Sci. 97,328-333. Gupta, U . C., and Sowden, F. J . 1965. Can. J . Soil Sci. 45,237-240. Gupta, U. C., Sowden, F. J., and Stobbe, P. 1963. Soil Sci. Soc. Am. Proc. 27, 380382.

SOIL P O L Y S A C C H A R I D E S

233

Harmsen, G. W., and Jager, G. 1962. Nature 195, 1 I 19- 1 120. Harris. R. F., Allen, 0. N., and Chesters, G. 1966. Plant Soil 25, 361-371. Harris, R F., Allen, 0. N., Chesters, G.. and Attoe, 0. J. 1963. Soil Sci. Soc. Am. Proc. 27,542-545.

Hartree, E. F. 1964. Anal. Biochem. 7, 103-109. Haworth, W. N.. and Hirst, E. L. 1930. J . Chem. Soc. pp. 2615-2635. Haworth, W. N.. Pinkard, R. W., and Stacey, M. 1946. Nature 158,836-837. Heidelberger, M.. Dische, Z . , Neely, W. B., and Wolfrorn, M. L. 1955. J . Am. Chem. Soc. 77,351 1-3518.

Hocevar, B. J., and Northcote, D. H. 1957. Nature 179,488-489. Horton, D., and Wolfrom, M. L. 1963. ComprehensiveBiochern. 5,188-232. Ivarson, K. C., and Sowden, F. J. 1962. SoilSci. 94,245-250. Jayrne. G., and Lang, F. 1963. Methods Carbohydrate Chem. 3 , 7 5 4 3 . Jeanloz, R. W. 1963. Comprehensive Biochem. 5,262-296. Jenny, H., and Grossenbacher. K. 1963. SoilSci. Soc. A m . Proc. 27,273-277. Jerrnyn, M. A. 1962. Australian J . B i d . Sci. 15,787-79 I. Johansen, P. G., Marshall, R. D., and Neuberger, A. 1960. Biochem. J . 77,239-247. Jones, D.. Morre, D. J.,and Mollenhauer, H. H. 1966.Am. J . Bottrm 53.62 I Keefer, R. J., and Mortensen, J. L. 1963. Soil Sci. Soc. Am. Proc. 27,156- 160. Kenner, J., and Richards, G. N . 1957. J . Chem. SOC. pp. 3019-3024. Kertesz, Z. I . 1963. Comprehensive Biochem. 5, 233-245. Kononova, M. M. 1961. “Soil Organic Matter.” Pergamon Press, Oxford. Lowe, L. E. 1968. Can. J . Soil Sci. 48, 2 15-2 17. Lynch, D. L.. and Cotnoir, L. J., Jr. 1956. Soil Sci. Soc. A m . Proc. 20,367-370. Lynch, D . L., Wright, L. M., and Cotnoir, L. J., Jr. 1956. Soil Sci. Soc. A m . Proc. 20,6-9. Lynch, D. L.. Wright, L. M., and Olney, H. 0. 1957. Soil Sci. 84,405-41 I . Lynch, D . L., Olney, H. O., and Wright, L. M. 1958. J . Sci. Food Agr. 9,56-60. McCalla, T. M. 1943. Soil Sci. Soc. Am. P roc. 7, 209-2 14. McCalla, T. M. 1945. Soil Sci. 59,287-297. Martin, J. K. 1964. New Zealand J . Agr. Res. 7, 723-749. Martin. J. P. 1945a. Soil Sci. 59, 163- 174. Martin, J. P. 1945b. J . Bacteriol. 50, 349-360. Martin, J. P. 1946. Soil Sci. 61, 157-166. Martin, J. P., and Craggs, B. A . 1946. J . Am. Soc. Agron. 38, 332-339. Martin, J. P., and Richards, S. J. 1963. J . Bacteriol. 85, 1288-1294. Martin, J. P., Ervin, J. O., and Shepherd, R. A . 1965. Soil Sci. SOC.Am. Proc. 29,397-400. Martin, J. P., Ervin, J . 0.. and Shepherd, R. A. 1966. Soil Sci. Soc. Am. Proc. 30,196-200. Mayaudon.J.,andSimonart, P. 1959. PlantSoil11,181-192. Mehta, N . C., Streuli. H., Muller, M., and Deuel, H. 1960. J . Sci. Food Agr. 11, 40-47. Mehta, N . C., Dubach, P., and Deuel, H. 1961.Advan. Carbohydrate Chem. 16,335-355. Moggridge, R. C. G . ,and Neuberger, A . 1938. J . Chem. Soc. pp. 745-750. Morre, D. J., Jones, D. D., and Mollenhauer. H. H. 1967. Planta 74,286-30 I . Mortensen, J . L. 1961 Trans. 7th Intern. Congr. Soil Sci.,Madison, Wise., 1960 Vol. 2, pp. 98- 104. Elsevier, Amsterdam. Mortensen, J. L., and Himes, F. L. 1964. In “Chemistry of the Soil” ( F . E. Bear, ed.), 2nd ed., pp. 206-24 1. Reinhold, New York. Mortensen, J. L., and Schwendinger, R. B. 1963. Geochim. Cosmochim. Acta 27,201-208. Muller, M., Mehta. N . C., and Deuel, H. 1960. Z . Pjanzenernaehr. Dueng. Bodenk. 90, 139- 145.

234

G. D. SWINCER, J. M . OADES, A N D D. J. GREENLAND

Neuberger, A., and Marshall, R. D. 1966a. In “Glycoproteins” (A. Gottschalk, ed.), pp. 190-234. Elsevier, Amsterdam. Neuberger, A,, and Marshall, R. D. 1966b. In “Glycoproteins” (A. Gottschalk, ed.), pp. 235-272. Elsevier, Amsterdam. Neuberger, A,, Marshall, R. D., and Gottschalk, A. 1966. In “Glycoproteins” (A. Gottschalk, ed.), pp. 158- 189. Elsevier, Amsterdam. Northcote, D. H. 1954. Biochem. J . 58,353-358. Northcote, D. H . 1966. Brit. Med. Bull. 22,180-184. Oades, J. M. 1967a. J . Chromatography 28, 246-252. Oades, J. M. 1967b. Australian J . Soil Res. 5, 103-1 IS. Oades, J. M., and Swincer, G. D. 1968. Trans. 9th Intern. Congr. Soil Sci., Adelaide, 1968 Vol. 3, pp. 183-192. Intern. Soc. Soil Sci. and Angus & Robertson, Sydney, Australia. Ogston, A. G. 1958. Bi0chem.J. 70,598-599. Overend, W. G., Rees, C. W., and Sequeira, J. S. 1962.J . Chem. Soc. pp. 3429-3440. Parsons, J. W., and Tinsley, J. 196 I .Soil Sci. 92,46-53. Pearce, R. H., and Mathieson, J. M. 1967. Can.J . Biochem. 45,1565- 1576. Percival, E. 1963. Comprehensive Biochem. 5,l-66. Perry, M. B.,and Hulyalkar, R. K. 1965. Can. J . Biochem. 43,573-584. Pigman, W. 1957. In “The Carbohydrates” (W. Pigman, ed.), pp. 1-76 Academic Press, New York. Pigman, W., and Platt, D. 1957. In “The Carbohydrates” (W. Pigman, ed.), pp. 709-732. Academic Press, New York. Reichstein, T., and Weiss, E. 1962.Advan. Carbohydrate Chem. 17,65-120. Rennie, D. A., Truog, E., and Allen, 0. N. 1954. Soil Sci. SOC.A m . Proc. 18, 399-403. Rogers, H. J. 1966. Brit. Med. Bull. 22, 185-189. Roulet, N., Dubach, P., Mehta, N. C., Muller-Vonmoos, M., and Deuel, H. 1963a. Z . Pflanzenernaehr. Dueng. Bodenk. 101,2 10-2 14. Roulet, N., Mehta, N. C . , Dubach, P., and Deuel, H. 1963b. Z . Pjanzenernaehr. Dueng. Bodenk. 103, 1-9. Rovira, A. D. 1956. Plant Soil 7, 178-194. Rovira, A. D. 1962. Soils Fertilizers 25,167-172. Rovira, A. D., and Greacen, E. L. 1957. Australian J . Agr. Res. 8,659-673. Rovira, A. D., and McDougall, B. M. 1967. In “Soil Biochemistry” (A. D. MacLaren and G. H. Petersen, eds.), pp. 417-471. Arnold, London. Russell, E. W. 1961. “Soil Conditions and Plant Growth,” 9th ed. Longmans, Green, New York. Saeman, J. F., Moore, W. E., Mitchell, R. L., and Millett, M. A. 1954. Tappi 37,336-343. Salomon, M. 1962. Soil Sci. SOC.A m . Proc. 26,5 1-54. Salton, M. R. J. 1960. “Microbial Cell Walls.” Wiley, New York. Salton, M. R. J. 1964. “The Bacterial Cell Wall.” Elsevier, Amsterdam. Samtsevich, S. A. 1965. Soviet Plant Physiol. 12,73 1-740. Sanderson, G. W., and Perera, 9.P. M. 1966. Analyst 91,335-336. Schmid, K . 1964. Chimia (Aarau) 18,321-327. Scott, F. M., Hammer, K. C., Baker, E., and Bowler, E. 1958. A m . J . Botany 45,449-46 1 Scott, F. M., Bystrom, 9.G., and Bowler, E. 1963. Science 140,63-64. Scott, J. E. 1960. Methods Biochem. Anal. 8, 145-197. Sevag, M. G . , Lackman, D. B., and Smo1ens.J. 1938.5. B i d . Chem. 121.325-436.

SOIL POLYSACCHARIDES

235

Shafizadeh, F. 1958. Advan. Carbohydrate Chem. 13,9-61. Simonart, P. 1964. Ann. Inst. Pasteur 107,7-20. Siu, R. G . H . 195 1 . ”Microbial Decomposition of Cellulose.” Reinhold, New York. Smith, F., and Montgomery, R. 1959. “The Chemistry of Plant Gums and Mucilages.” Reinhold, New York. Sowden, F. J., and Ivarson, K. C . 1962. Soil Sci. 94,340-344. Spiro, R . G . 1966. Methods Enzymol. 8,3-26. Stacey, M., and Barker, S. A. 1960. “Polysaccharides of Microorganisms.” Oxford Univ. Press, London and New York. Stevenson, 1. L. 1964. In “Chemistry of the Soil” (F. E. Bear, ed.), 2nd ed., pp. 242-291. Reinhold, New York. Streuli, H., Mehta, N. C., Muller, M., and Deuel, H. 1958. Mitt. Gebiete Lebensm. Hyg. 49,396 (quoted by Mehta et al., I96 I ). Swaby. R. J. 1949. J . Gen. Microbiol. 3,236-254. Swincer, G. D., Oades, J. M., and Greenland, D. J. 1968a. Australian J . Soil Res. 6,21 1224.

Swincer, G. D., Oades, J. M., and Greenland, D. J . 1968b. AustrulianJ. Soil Res. 6,225235.

Thomas, R. L. 1963. Ph.D. Thesis, Ohio State University. Thomas, R. L., and Lynch, D. L. 1961. SoilSci. 91,312-3 16. Thomas, R. L., Mortensen, J. L., and Himes. F. L. 1967. Soil Sci. Soc. A m . Proc. 31,568570.

Thompson, A,, Wolfrom, M. L., and Pacsu, E. 1963. Methods Carbohydrate Chem. 2, 2 15-220. Tinsley, J., and Salam, A. I96 1. Soils Fertilizers 24, 8 1-84. Toogood, J. A., and Lynch, D . L. 1959. Can. J . Soil Sci. 39, 15 1-156. Walborg, E. F., and Ward, D. N. 1963. Biochim. Biophys. Acta Walhorg. E. F.. and Ward. D. N. 1963. Biochim. Biophys. Acra 78, 304-312. Watson, J. H., and Stojanovic, 9 . J. 1965. Soil Sci. 100, 57-62. Webber, L. R. 1965. SoilSci. Soc. Am. Proc. 29,39-42. Webley, D. M., Duff, R. B., Bacon, .I.S. D.. and Farmer, V. C . 1965. J . Soil Sci. 16, 149157.

Weidel, W., and Pelzer, H. 1964.Advan. Enzymol. 26,193-232. Whelan, W. J. 1960. Ann. Rev. Biochem. 29,111- 130. Whistler, R. L. 1965. Methods Carbohydrate Chem. 5, Arts. 19-54. Whistler, R. L., and BeMiller, J. N. 1958. Advan. Carbohydrate Chem. 13,289-329. Whistler, R. L., and Corbett, W. M. 1957. I n “The Carbohydrates” (W. Pigman, ed.), pp. 641-708. Academic Press, New York. Whistler, R. L.. and Kirby, K. W. 1956.J.Am. Chem. SOC. 78,1755-1759. Whistler, R. L., and Lauterbach, C. E. 1958. Arch. Biochem. Biophys. 77,62-67. Whistler, R. L., and Smart, C. L. 1953. ”Polysaccharide Chemistry.” Academic Press, New York. Whitehead, D. C., and Tinsley, J. 1963. J . Sci. FoodAgr. 14,849-857. Whitehead, D. C., and Tinsley, J. 1964. Soil Sci. 97,34-42. Williams, 9. G., Greenland, D. J., and Quirk, J. P. 1967. Australian J . Soil Res. 5,77-92. Winogradsky, S . 1929. Ann. Inst. Pasteur 43,549-633. Winzler, R. J. 1955. Methods Biochem. Anal. 2,279-3 I I . Wolfrom, M. L., and Thompson, A. 1957. In “The Carbohydrates” (W. Pigman, ed.), pp. 188-240. Academic Press, New York.

This Page Intentionally Left Blank

FRAGIPAN SOILS OF THE EASTERN UNITED STATES R. B. Groosman and F. J. Carlisle Soil Conservation Service, United States Deportment of Agriculture, Lincoln, Nebraska ond Hyattsville, Maryland

I. 11.

Introduction .. .. .. ... ... . . . . ... .. . ... .. . Horizons of Fragipan Soils ........ B. C.

Horizon Sequences .................................................... Fragipan Expression . ..... . .

111.

IV. Properties of Fragipans ........... ......... ... .. ... ......................... A. Composition ..............

........................................ C. Consistenc V.

r Regime ..................................

................. C . Movement of Low-Tension Water ................................................. v1. Genesis of Fragipans .... ., . ., . .. . .. . .. . ... ........... A . Bonding of the Fragipan B. Development of Fragipan Soils ...... . . . ..... ................................... . .. . VII. Fragipans and Soil U s e ...................................................................... A. Plant Growth ............. ......... ._......... B. Engineering Manipulation . ...... ... . ... . .. . ................... .. .... ... . . ... ......... VIII. Classification of Fragipan Soils ....,..... ... ..... ... ....................... ..... . . ......... 1x. Unresolved Problems ........ ................................................................ X. Summary . ..... ... .... . ... References ........ .. . ...... ...... .. .. . . . ... .. . ... ....... .... .... . . ... ... ....... ... ............... Appendix.. .. ...... .......

Pft@’ 237 240 240 24 1 242 244 246 246 249 25 1 254 254 254 255 256 256 259 263 263 264 265 269 271 212 216

I. Introduction

Twenty years ago Winters and Simonson ( I 95 1) very adequately reviewed fragipans in three pages of this publication. Since then a large number of pertinent papers have appeared, and it seems appropriate to consider the subject anew. This review is restricted to fragipans in soils of the eastern United States. 237

238

R. B. GROSSMAN A N D F. J. CARLISLE

As Figure 1 illustrates, fragipans are prominent features of some soils. Despite their prominence in the soil profile, the definition of fragipans is vague. The vagueness arises from the nature of the defining properties.

FIG. I . Block diagram of a soil with a fragipan that occurs in northeastern Wisconsin. The fragipan horizons are indicated by the suffix x. (Modified from Olson and Hole, 19671968.)

The principal defining property is the type of failure exhibited by moist soil material when pressure is applied. Yield is abrupt; a moist piece of soil breaks into fragments rather than deforming. This has been described as brittle consistence. The failure by some fragipans has been likened (Daniels et al., 1966) to the crumbling of a dry graham cracker. This defining property does not lend itself to laboratory measurement. The natural organization of the soil fabric must be kept, which raises difficulties with the measurement. Also, the failure observed depends sensitively on the water content. Resultantly, field observations are subject to large variability. A body of standardized measurements on the defining property of fragipans does not exist. Many descriptions of fragipans give no indication of the degree of brittleness. Winters and Simonson ( 1 95 1) listed some of the early publications on soils with fragipans. Olson (1962) presented a comprehensive review. As indicated in Section VIII, fragipans were recognized as pans long before they were given their current name. The concept of a pan in soil morphology involves restricted root penetration, occurrence below the

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

239

soil surface, and for most kinds of pans-although fragipans are an exception- the pronounced accumulation of a substance, such as silicate clay, iron oxides, silica, or carbonates. Historically, the concept also has involved the notion of an unusual or extreme expression of a feature or attribute and something apart from the normal or expected sequence of horizons. Reasons for considering fragipans to be genetic soil horizons have been discussed by the Soil Survey Staff ( 1960) and by Carlisle et al. (1 957). Fragipans parallel the soil surface, have upper boundaries at moderate depths (20 to 100 cm.), have a constant spatial relationship to other soil horizons in a given kind of soil, occur in geological materials of widely different origins, and may transgress differing parent materials in a local area. Their range in thickness (20 to 200 cm.) is within generally accepted limits for soil horizons. Cline (1952) states, “The consistency of depth at which it [the fragipan] occurs . . . indicates considerable control by some thing or things related to the land surface, and this more than any one thing is probably responsible for its designation as a genetic horizon.” Fragipans are subsoil horizons unless the soil has been truncated. As subsoil horizons they are subject to accumulation of illuviated substances. They are beneath the depth of maximum root concentration; organic matter accumulation is not high. Mechanical disturbance of the soil fabric is weak. Consequently, fragipans retain the marks of translocation of substances and reorganization of the fabric. Carlisle et al. (1957) have defined fragipans as follows: “Compact horizons (of high bulk density) which are hard to extremely hard when dry and firm to very firm when moist and display the property of ‘brittleness’ when both dry and moist. The term ‘brittleness,’ as used here, embodies a type of physical behavior that is characteristic of but not exclusively associated with fragipan materials. The term is used to characterize a condition in which a fragment of the material sustains increasing pressure without detectable deformation until a critical pressure is reached, at which point the material suddenly shatters.” This definition differs from that in the Soil Survey Manual (Soil Survey Staff, 195 1) by putting more emphasis on the retention of brittleness when moist. The definition in the 7rh Approximation (Soil Survey Staff, 1960) closely parallels that of Carlisle et al. ( 1 957). Fragipans have a number of characteristics, which, though not necessary to the definition, form part of the general description applicable to most but not all fragipans. They include high silt, very fine sand, or fine sand contents: moderate or low clay content: low organic matter content: medium to high bulk density when moist; slow or very slow satu-

240

R. B. GROSSMAN A N D F. J. CARLISLE

rated hydraulic conductivity; well expressed mottling; presence of bleached cracks or fracture planes that form a coarse polygonal pattern on a horizontal plane; weak pedological structural expression within the polyhedrons outlined by the bleached cracks; clearly identifiable and planar upper boundaries; presence of bodies of moved clay; and few roots, with those present largely restricted to the cracks between large polyhedrons. 11. Horizons of Fragipan Soils

A. TERMINOLOGY A N D HORIZONNOMENCLATURE Fragipans occur in soils with argillic, cambic, or spodic horizons. The cambic and spodic horizons, if present, occur above the fragipan. The argillic horizon may occur partly or wholely above or be coextensive with the fragipan. An argillic horizon is “an illuvial horizon in which layer-lattice silicate clays have accumulated by illuviation to a significant extent” (Soil Survey Staff, 1967). Cambic horizons may show several kinds of alteration, but domination by accumulation of mineral substances is excluded. The marks of cambic horizons in soils with fragipans include the mottling indicative of translocation of iron and associated with periodic saturation, higher chroma and redder hue associated with coatings of iron oxides released by mineral weathering, and sufficient development of soil structure largely to obliterate the original organization (stratification, for example) of the parent material (Soil Survey Staff, 1967). Cambic horizons may contain illuvial clay but do not meet the requirements for an argillic horizon. Spodic horizons contain appreciable amounts of precipitated amorphous materials “composed of organic matter and aluminum, with or without iron . . .” (Soil Survey Staff, 1967). Figure 1 depicts a soil with a spodic horizon above the fragipan. The suffix x is currently used to designate the fragipan (Soil Survey Staff, 1962); an example would be Bx. The suffix rn has been used earlier (Soil Survey Staff, 1951) but is now restricted to strongly cemented or indurated horizons having a consistence that is not appreciably affected by moistening. The proposals by the International Society of Soil Science ( 1 968) on horizon nomenclature use x and rn in these senses. Many soils with fragipans have two pairs of A and B horizons. Each pair of A and B horizons is referred to as a sequum. Soils with two pairs are referred to as bisequal. The soil shown in Fig. 1 is bisequal. The lower pair of horizons is designated with a prime accent. An example would be B’x. The fragipan, if present, commonly occurs in the lower sequum if the soil is bisequal. Daniels el al. (1968) have suggested the notation Be for a B horizon

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

24 1

containing discontinuous eluvial parts. The Be notation may find use as a designation for lower eluvial horizons that cannot be designated A'2 or B & A.

B. HORIZONSEQUENCES The uppermost panel of Fig. 2 illustrates the most general statements that may be made about the position of the fragipan. All fragipans occur beneath an eluvial horizon unless the soil has been eroded severely. In Wetter

Eluvial

A'2 Lxl

ond /or I

Ilc

-

IId

A2

IA2

B

IB

Bx

and/or 1

cx

1

Btx

I

FIG.2. Horizon sequences of soils with fragipans in the eastern United States. Elements within brackets indicate characteristics that are not essential to the definition of the sequence. The fragipan horizons are indicated by the suffix x: the suffix t denotes silicate clay sufficient for recognition of an argillic horizon.

some soils, more often the wetter ones, the fragipan occurs immediately beneath the eluvial horizon. In many soils, however, there is an intervening B horizon, which may be a spodic horizon, a cambic horizon, or an argillic horizon. Some soils with fragipans are bisequal. The fragipan may occur in the lower eluvial horizon and not in the underlying B horizon, in the lower B horizon but not in the second eluvial horizon, or in both. These variations form the basis for the four classes of horizon sequences shown in the set of panels designated Ia through Id; these are illustrated by soil descriptions in the Appendix. These four classes are

242

R.

B. GROSSMAN A N D F. J. CARLISLE

subdivided to give the eight classes, IIa through IIh. Accumulation of clay and occurrence of the fragipan in the lower eluvial horizon are the criteria for these subdivisions. Further subdivisions could be made. Panel IIf, for example, could be subdivided into soils with argillic, spodic, or cambic horizons above the fragipan. Brackets are placed around the horizon designations or elements of the horizon designations that are not necessary to the definition. In Fig. 2 the suffix t denotes an accumulation of silicate clay sufficient for recognition of an argillic horizon. This is a more limiting use of the t suffix than is specified by the Soil Survey Staff ( 1 962). The horizon symbols in Fig. 2 are only for convenience in discussing this subject; it is not suggested that any of these symbols should replace horizon designations in current use. Soils with horizon sequences Ha, Ilc, and IIe have fragipans in horizons where the parent material has been weakly altered. Many of the soils with fragipans in the northeastern United States developed in Wisconsin glacial till belong in one of these three classes. In contrast are fragipans in soils that have horizon sequences IIb, IId, IIf, IIg, and IIh. In these soils, alteration of the parent material of the fragipan by eluviation, illuviation, or both, has been strong. Hence, influence of the parent material in determining the properties of the fragipan is less important. Fragipans in soils south of the Wisconsin glacial advance tend to exhibit stronger alteration of the parent material.

C. FRAGIPAN EXPRESSION Fragipan expression involves thickness and depth to its upper boundary. These are considered in subsequent paragraphs of this section. It also involves properties, such as consistence, structure, and pore arrangement, that determine mechanical impedance to root penetration and rate of movement of low-tension water. Objective measurement and integration of these properties has not been achieved. Within a local association of soils, however, evaluation of relative degree of fragipan expression has validity. Maximum degree of expression of the fragipan is commonly reported for the soils of intermediate wetness (Neeley, 1965; A. E. Thomas, 1967; Redmond and Engberg, 1967; Grossman et al., 1959a; Nettleton et al., 1968a). This intermediate wetness is described approximately by the terms somewhat poorly and poorly drained (Soil Survey Staff, 195 1) and by the Aquic subgroups of the soil classification system (Soil Survey Staff, 1967). The relationship is widespread enough to suggest implication in the genesis of the fragipan.

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

243

Excluding severely eroded soils, the depth to the fragipan ranges from about 10 to 150 cm., with the upper boundary for most between 25 and 100 cm. In landscapes that have been affected by severe erosion, depth to the fragipan often is closely related to past land use and to the slope. The area of occurrence of Grenada soils in Grenada County, Mississippi (A. E. Thomas, 1967) is an example (see Appendix for description). Concepts have centered on the less eroded soils. Consequently, in many areas fragipans are shallower than much of the literature would suggest. On a given landscape the depth is frequently relatable to wetness of the soil. Two of many examples have been selected. Neeley ( 1965), studying fragipans in soils of New York developed in medium-textured glacial till, reported a range of 75 to 30 cm. to the top of the fragipan from the drier to the wetter soils. The soils range from sequence IIe through IIc to I Ia of Fig. 2. A. E. Thomas ( 1 967), working with soils developed in loess in Mississippi, reports a range from 50 to 10 cm. Sequences involved are IIf and IIb. In some associations, the fragipan is not shallower in the wetter soils. Soils of the Lebanon series and related soils in the Missouri Ozarks (Krusekopf, 1942; Scrivner, 1960) have a fragipan immediately beneath and perhaps partly within the loess that covers a buried soil (sequence 1Ig or JIh). Fragipans in this soil association are no shallower in the somewhat wetter soils that occur in slight depressions than in the associated better drained soils. Thickness ranges from about I5 to 125 cm. In some local associations in the northeastern states, thickness remains fairly constant over a considerable range in wetness (Carlisle, 1954; Neeley, 1965). These fragipans are weakly altered and bear a strong imprint of the compact glacial till parent material. The drier soils have sequences IIc and IIe of Fig. 2; the wetter, sequence IIa. Constant thickness over a wide range in wetness is not restricted to fragipans with strong parent material influence. Bailey (1964) reports rather constant thickness for the fragipans in strongly illuvial horizons (sequence Ild, Fig. 2 ) of some soils developed from limestone residuum and mixed sedimentary rocks in Kentucky. In certain loess-derived soils of the middle and lower Mississippi Valley, the fragipans become thicker in the wetter soils of the association (A. E. Thomas, 1967; Bailey, 1964). The fragipans in these soils are strongly affected by illuviation, eluviation, or both. The drier of these soils have sequences IId, IIf, and IIg; the wetter soils, sequence IIb. In other associations of loess-derived soils in the same general area, the fragipan thins with increasing wetness (Grossman el al., 1959a). Nettleton et al. ( 1968a), in their study of fragipans in certain soils of North Carolina de-

244

R. B. GROSSMAN A N D F. J . CARLISLE

veloped from Coastal Plain sediments (sequence IIh, Fig. 2), also report thinning of the fragipan with increasing wetness. 111. Occurrence of Fragipan Soils

The occurrence of soils with fragipans may be considered at several scales. Looking at the country as a whole, fragipans have been recognized in some soils in all states east of the Mississippi River and in adjacent Minnesota, Missouri, Arkansas, Oklahoma, Louisiana, and eastern Texas. Soils with fragipans are the principal soils in some parts of that area, and they are of minor extent in other parts. They have been reported in the western states (Cline, 1952; Whittig et al., 1957) but seem to be of very minor extent and importance there. Figure 3 shows areas in the United States where one or more of the principal kinds of soils have

FIG. 3. Areas in the United States where one or more of the principal kinds of soils have fragipans. The most extensive soils in the areas delineated are: A, Altisols: I , Inceptisols; S , Spodosols: U and U 2 , Ultisols (generalized from Soil Conservation Service, 1969, except U 2 delineations, which are from Bartelli, 1968). Dashed line shows the approximate eastern boundary of the prairie (generalized from Shantz and Zon, 1914).

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

245

fragipans. They are extensive and may be among the principal soils in areas where the dominant soils are Spodosols, Inceptisols, Alfisols, or Ultisols in the warm humid and the cold humid, central and eastern parts of the country. Fragipans apparently do not occur in soils of the humid prairies, the Great Plains, or the semiarid and arid areas of the west. The following relationships are evident in the broad-scale occurrence of fragipan soils: ( 1 ) Fragipans are restricted to areas where the excess of precipitation over evapotranspiration is sufficient at some time of the year for movement of water down through the soil. (2) They occur in both warm and cold climates. (3) Fragipans seemingly are absent in soils of the extensive natural grasslands of the humid prairies and the Great Plains. (4) Fragipans occur in Spodosols, lnceptisols, Alfisols, and Ultisols. Spodosols with fragipans are so common as to suggest a genetic connection. Other relationships may be evident at the larger scale of soil association maps for counties where fragipans are important. The occurrence of fragipan soils in relation to composition of soil parent materials apparently is complex when viewed generally. But local relationships may be evident, especially if the soils are not old. The surveys of Tompkins County, New York (Neeley, 1965) and of Franklin County, New York (Carlisle, 1958) illustrate the influence of lime content of parent material on occurrence of soils with fragipans. In these areas soils with well expressed fragipans are prevalent in areas of low-lime glacial till and absent in areas of high-lime till. The Franklin County, New York, survey illustrates the influence of texture of soil materials on the occurrence of soils with fragipans. I n that area fragipans are present in Spodosols developed in relatively well-graded glacial till and are absent in Spodosols developed in poorly graded glaciofluvial materials of roughly comparable mineralogical composition. In Tate County, Mississippi (J. S. Huddleston, 1967) the distribution of soils with fragipans on the countywide scale is related in part to the loess distribution pattern and to the pattern of geological erosion. Other illustrations of these kinds of local relationships may be found in published soil surveys of areas shown in Fig. 3. For a multicounty area in southern New York and northern Pennsylvania, Denny and Lyford ( 1963) showed the distribution of soil associations in which soils with fragipans are extensive. They concluded that within the region studied many of the soil differences are primarily related to kind of parent material and its hydrologic characteristics. A third and still larger scale is the detailed soil map showing delineations dominated by a particular kind of soil. This is the scale commonly used in planning farming practices for individual fields and farms. To

246

R. B. GROSSMAN A N D F. J. CARLISLE

obtain a sense of the pattern of occurrence of fragipan soils on this scale, one should mentally walk across the detailed map observing the pattern of mapping unit delineations in relation to slope, topography, and other landscape features. Practical considerations do not permit delineation on the usual soil map of the pattern of soil occurrence within distances measured in tens of feet. This scale of observation, however, may be very fruitful in understanding the genesis of the soils. Daniels and associates, for example, have worked at this scale to study certain soils with fragipans of North Carolina that occur on the less dissected part of a pre-Wisconsin geomorphic surface formed in Coastal Plain sediments (among other papers, Daniels et af., 1966; Daniels and Gamble, 1967; Nettleton et af., 1968a,b). Lyford et al. (1963) have reported a study at a similar scale on soils with fragipans in central Massachusetts that are developed in Wisconsinage glacial till. Less extensive studies on such a very local scale have been reported by Bornstein et al. ( I 9 6 3 , Carlisle (1954), Lyford and MacLean ( 1 966), Olson ( 19621, and Denny and Lyford ( I 963). IV. Properties of Fragipans

A. COMPOSITION 1 . Texture Most fragipans are loamy as the term is defined by the Soil Survey Staff (1 967). They may be skeletal, but not fragmental. Few strong fragipans are sandy. Although some fragipans exceed 35 percent clay, no fragipans have been reported with over 60 percent clay. Fragipans in weakly illuvial horizons (IIa, IIc, IIe, Fig. 21, such as are common in the northeastern United States, contain less than 35 percent clay, and most contain less than 25 percent clay (Jha and Cline, 1963). Maximum clay percentage does not appear to be strongly related to the kind and proportions of clay minerals. Fragipans with montmorillonite predominating may exceed 35 percent clay (for example, Hutcheson et al., 1959). The fine earth of fragipans, after allowance for the clay percentage, usually is high in material from 0.2 10 0.02 mm. and low in the range above 0.25 mm. A high content of silt and very fine sand may promote development and expression of fragipans (Jha and Cline, 1963; Carlisle, 1958; Grossman and Cline, 1957). 2. Chemical Chemical properties of fragipans do not seem unique. Fragipans are low in organic matter, have low or moderate levels of extractable iron

FRAGIPAN SOILS O F THE EASTERN UNITED STATES

247

relative to the clay percentage, are rarely if ever calcareous, lack appreciable soluble salts, and have at most moderate exchangeable-sodium levels. Exchange capacity ranges widely, depending on the amount and kind of clays. Base saturation ranges from low to high. Fragipans in strongly illuvial horizons of soils weakly influenced by the Wisconsin glacial advances may have very low base saturation. Base saturation usually rises with increasing wetness. Fragipans in soils with low base saturation in their lower parts may have more extractable magnesium than calcium, and some extractable sodium may be present. Hutcheson et a f . (1 959) and Pettiet ( 1 964) studied fragipans in such soils. They suggest that high extractable magnesium may make clay more susceptible to movement and rearrangement, thereby fostering fragipan formation. Presence of magnesium may be a factor in some soils with fragipans but many fragipans do not have particularly high extractable magnesium. Soil pH values range from 4 to 7. Aluminum extractable with a neutral salt follows the common pattern. Values commonly increase rapidly as the soil pH drops below 5. Soluble silica and aluminum determinations have been reported by a number of workers, either in connection with studies on bonding or incidental to characterization of. the clay mineralogy (Alexander, 1955; Comerma, 1964; Jha and Cline, 1963; Knox, 1957; Olson and Hole, 1967-1 968; Pettiet, 1964; Yassoglou and Whiteside, 1960). No consistent pattern has been observed.

3. Mineralogy Fragipan expression seems largely unrelated to the mineralogy of the sand and silt other than for the control that presence of carbonate may impose. Daniels er af. (1966) found no relationship between fragipan expression and the percentage of feldspar in the very fine sand fraction for certain soils of North Carolina having strongly eluvial fragipans. In some landscapes fragipans are restricted to the older geomorphic surfaces. These soils may have lower weatherable minerals than the associated soils on the younger geomorphic surfaces. T h e difference in mineralogy, however, would not appear responsible for the pattern of fragipan occurrence. Fragipans are common on the Ozark Plateau and on the Coastal Plain. Parent material sources high in quartz and low in weatherable minerals are common to these areas. The high proportion of quartz does not appear to have directly contributed to the prevalence of fragipans in these areas (see Section VI, A). The clay mineralogy of fragipans is similar to the clay mineralogy of horizons in comparable positions in associated soils without fragipans. In the northeastern United States, related to the prevalence of paleozoic

248

R. B. GROSSMAN AND F. J. CARLiSLE

sedimentary rocks as sources, 2:1 lattice clays other than montmorillonite predominate, with mica (illite) a quantitatively important component (among others, Knox, 1957; Jha and Cline, 1963; F. P. Miller, 1965). Moving westward into the northern Middle West, Yassoglou and Whiteside ( 1960) have reported illite, chlorite, and interstratified minerals with some interlayered montmorillonite in the clay fraction of several soils with fragipans in Michigan. Olson ( 1 962) indicated a somewhat similar suite of clay minerals in some soils with fragipans of northeastern Wisconsin. Moving southward, the fragipan in Lebanon-like soils of Missouri studied by Scrivner (1 960) contains randomly stratified montmorillonite, illite, and Vermiculite with some kaolinite; the mineralogy is allied with the surficial loess rather than the limestone residuum beneath the loess. Soils with fragipans of the middle and lower Mississippi Valley developed in loess contain appreciable montmorillonite (Anderson and White, 1958; Glenn, 1960; Grossman et af., 1959b; Hutcheson et af., 1959; Pettiet, 1964). In these soils the fragipans occur in horizons that have undergone moderate or strong illuviation (sequences IIb, IId, IIf, IIg of Fig. 2). The montmorillonite shows evidence of interlayers that do not extend fully on treatment with glycols. Hutcheson et al. (1959) suggested that in the fragipan vermiculite forms at the expense of montmorillonite. Glenn ( 1 960) and Pettiet ( 1964) reported appreciable amorphous material. In the middle Atlantic states, Nikiforoff et al. ( 1948) found significant amounts of kaolinite in the fragipan of a soil that occurs in Maryland. Nettleton et af. (1968b) studied the clay mineralogy of soils in North Carolina with fragipans in highly eluvial horizons (sequence IIg of Fig. 2). Kaolinite or a vermiculite-chlorite intergrade were the most abundant components, with lesser amounts of montmorillonite. Evidence was presented for considerable amorphous material, and the kaolinite had imperfect crystallinity. 4 . Bulk Density

Bulk densities of the moist fine-earth fabric usually exceed 1.4 g./cc. and mostly exceed 1.6 g.lcc. Differences between the moist and dry natural-clod bulk densities are not large. The coefficient of linear extensibility (Grossman er al., 1968) usually does not exceed 0.04. The fragipan rarely has a lower bulk density than overlying horizons. Differences from horizons below are variable. In soils with a strong influence of parent material on the fragipan (horizon sequences Ila, IIc, IIe of Fig. 2 ) and a

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

249

parent material with a high bulk density, the fragipan may differ little in bulk density from the parent material (examples in Soil Survey Staff, 1968b). If the bulk density of the parent material is moderate, the fragipan horizon may have the highest bulk density in the soil (for example, Jha and Cline, 1963). Fragipans in eluvial horizons that immediately overlie strongly expressed illuvial horizons (horizon sequence IIh) often have the maximum bulk density of the soil (examples in Nettleton et al., 1968a; Scrivner, 1960; Yassoglou and Whitesite, 1960). Rutledge and Horn (1965), Yassoglou and Whiteside (1960), and Pettiet ( 1964) have discussed estimates of the pore-size distribution of fragipans based on water retention determinations. A large body of information is available for the computation of pore-size distribution in the Soil Survey Investigations Reports series published by the U.S. Department of Agriculture. Medium- and fine-textured fragipans can be shown to have a high proportion of the total porosity filled with water held at energies above 15 bar; this porosity must consist of very small pores. The air-filled porosity at 1/3 bar can be shown to be small, but this is also common to many medium-textured soil materials.

B. MORPHOLOGY I . Macroscopic

Many fragipans have roughly vertical planes that delineate large prisms or blocks. Nikiforoff (1 955) vividly described the feature in the Beltsville soil of Maryland: “Throughout its thickness the pan is split into large irregular blocks ranging from about 1 f.i to 2 feet in horizontal diameter. Planes of cleavage are marked by strong bleaching of the walls of fissures which produces conspicuous streaks on the exposures of the hardpan. In vertical planes, these irregular light colored streaks are roughly parallel, whereas in horizontal planes, they form a striking polygonal network . . . . Walls of the cracks are bleached laterally for . . . a few millimeters to more than an inch . . . . Beyond these bleached zones there are yellow to orange oxidized zones so that on cuts the bleached streaks are enclosed between rust-colored bands.” The pattern may be much less distinct in other soils, particularly if the fragipan is coarse textured. The fragipans studied by Daniels et at. ( I 966) and Nettleton et al. (1 968a) lack the polygonal pattern. Information on the composition and organization of the periphery of the large prisms and blocks relative to the interiors was reported by Carlisle (1954), Gile (1958), Jha and Cline (1963), F. P. Miller (1965), Pettiet (1964), and Vanderford and Shaffer (1966). The study by F. P. Miller (1965) is particularly detailed.

250

R. B. GROSSMAN A N D F. J. CARLISLE

Structure of the large prisms or blocks is variable. Trends follow those for horizons other than fragipans: eluvial horizons usually are massive or platy; illuvial horizons usually are blocky. For example, the fragipans described by Yassoglou and Whiteside (1960) and by Nettleton et al. (1968a) occur in horizons that overall are eluvial; much of the fabric has massive or platy structure. In contrast stand the more clayey, strongly illuvial fragipans (Bailey, 1964; Rutledge and Horn, 1965). They have subangular or angular blocky structure of moderate expression and medium size. Fragipans that have undergone weak alteration may be strongly influenced by the structure of the parent material. If the parent material is platy (for example, Lyford et al., 1963), the fragipan may be platy. If it is crudely blocky (for example, Neeley, 1965; Carlisle, 1954), the fragipan usually has a crude blocky structure. If massive (for example, Jha and Cline, 1963), the fragipan may tend to massive structure. Large pores that are continuous vertically over the thickness of the fragipan usually are widely spaced. Vertical planar surfaces within the large structural units extend over distances that are a small fraction of the thickness of the fragipan, usually having dimensions of a few centimeters. Some fragipans have platy structure. The plates usually overlap, and consequently vertical pores between peds are tortuous. Fragipans are often described as vesicular, implying that the pores within peds are not interconnected. Lyford et ad. (1 963) have provided an apt description: “The pores in the peds are seldom continuous. They branch and rebranch but tend to end up inside the ped rather than continue from one side to the other. Many of the pores observed in broken peds are short and bulbous; they do not empty at the surface.” Evidence for translocation of silt or clay is common in fragipans. Field descriptions of fragipans generally refer to clay films on ped surfaces and within pores. The clay films may be thin to thick. Continuous clay films on ped surfaces are not reported in fragipans. Clay films, however, may be strongly expressed in pores. Rearrangement of silty material has been observed in many fragipans. Carlisle ( 1958) described the feature as follows: “Gray silty material, which does not occur in the layers above the fragipan, coats the upper surfaces of rock fragments and impregnates the uppermost part of the peds within the fragipan.” Descriptions of fragipans from outside New York and New England, including fragipans developed in Wisconsin glacial till, do not place as much emphasis on rearrangement of silty material.

2 . Thin-Section Observations A partial list of observations include Jha and Cline (1963), Carlisle (1954), Calhoun (19681, F. P. Miller (196.9, Nettleton et at. (1968b),

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

25 1

Olson ( I 9621, Pettiet ( I964), Yassoglou and Whiteside (1 960), Grossman et al. (1 959c), Horn and Rutledge ( 1965), Nikiforoff et al. (1 948), and McCracken and Weed ( 1 963). The descriptions have one unifying feature. Some of the clay connects between sand and silt grains, and some of it shows optical evidence of preferred orientation; the fabrics would be plasmic and sepic (Brewer, 1964). The clay may be scarce and strongly concentrated at contact points of sand and silt grains with interstices only partially filled; or the clay may be abundant and form a continuous medium within which the sand and silt grains are set. A significant portion of the clay apparently has moved over distances measured at least in millimeters; some may have come from horizons above. Bodies of moved clay are present in strongly eluvial fragipans as well as in fragipans that are illuvial B horizons. Amounts of moved clay are reported by Calhoun (1968), Carlisle (1 954), Horn and Rutledge (1 9 6 3 , F. P. Miller (1965), Nettleton et ai. ( 1968b), and Soil Survey Staff (1968a). Horn and Rutledge ( 1965) attach particular importance to the sepic micromorphology of fragipans in determining brittleness. The fragipan is viewed as “cellular.” Plasma separations that impart rigidity delimit small volumes with lower strength. Several workers have emphasized the importance of closeness of packing of the sand and silt grains. Some fragipans, however, do not show particularly close packing (Horn and Rutledge, 1965). Packing is generally closer in coarser-textured fragipans and may be a more important factor in determining the rigidity in the coarser textures. Horizons above the fragipan may show closer packing of the sand and silt grains and yet have markedly lesser rigidity than the fragipan (Grossman et al., 1959~). C. CONSISTENCE Consistence is the principal defining property of fragipans. But the classes of soil consistence are defined qualitatively and evaluation of consistence is subjective. This results in vagueness in the definition of fragipan and leads to difficulty in achieving uniform application of the fragipan concept. The soil consistence test (Soil Survey Staff, 195 1) involves compressing a piece of soil about 2 to 4 cm. across between thumb and forefinger until failure occurs. It is a kind of unconfined compression test, a subject reviewed recently by Gill and Vanden Berg (1967). If the piece of soil is uniform with no predetermined planes of weakness and the cracks that form at failure are near the center of the bearing surface and parallel the direction the force is applied, then cohesional forces largely determine the resistance to rupture. Tensile strength is therefore measured. Usually, however, the cracks form at an angle to the axis of corn-

252

R. B. GROSSMAN A N D F. J. CARLISLE

pression; hence some frictional force contributes to the resistance to rupture. Moreover, the area of contact between finger and soil is appreciable relative to the dimensions of the piece of soil. This tends to cause shearing stresses. Rearrangement may occur before rupture. The rearrangements commonly are such as to blunt the point of the crack, leading to an increase in strength of the soil material. Often the piece of soil has predetermined planes of weakness. Those cracks which most nearly parallel the axis of compression probably determine the strength. 1 . Induration

Fragipans appear cemented when dry, but soften when wetted. Cementation implies little reduction in hardness on moistening (Soil Survey Staff, 195 1). Fragipans, therefore, are not cemented. The expression “reversibly indurated” has been used, but it is not very satisfactory as it seems to be a contradiction in terms. Most fragipans slake when dry pieces are placed in water (Anderson and White, 1958; Knox, 1957; Comerma, 1964; Jha, 1961; Nikiforoff et al., 1948; Olson, 1962; R. M. Smith and Browning, 1946). One of the fragipans studied by Knox ( 1957) did not slake; silica was implicated in the bonding. Field observations of slaking by fragipans are of common occurrence, but few are published. Nettleton et al. (1968b) described the slaking of fragipans in ditch banks. Olson (1962) studied the breakdown of fragipan material on exposure.

2. Resistance to UnconJined Compression When dry, most fragipans are at least hard-“moderately resistant to pressure; can be broken in the hands without difficulty but is barely breakable between thumb and forefinger” (Soil Survey Staff, 195 1). When moist, fragipans are usually at least firm-“crushes under moderate pressure between thumb and forefinger, but resistance is distinctly noticeable” (1951). Grossman and Bartelli (1957) used a hand dynamometer to determine the force applied at the point of discrimination between several consistence classes by a group of soil scientists. The lower limit was 5 kg. for hard consistence and 3 kg. for firm consistence. Grossman and Cline ( 1 957) report a median value of 17 kg./cm.2 and a range of 4 to 25 kg./cm.2 for the resistance to rupture when dry of a number of fragipan horizons from soils of New York. Grossman (1954) measured a resistance to rupture of 4 kg./cm.2 at a water content near 1/3 bar retention for a fragipan having a resistance of 20 kg./cm.2 when dry. Knox ( I 954) assembled information on the resistance to rupture of

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

253

various particulate materials. Molding sands with 5 to 10 percent bentonite have crushing strength similar to the median value for the fragipans from New York studied by Grossman and Cline ( 1 957). 3 . Miscellaneous Tests

Yassoglou and Whiteside ( 1 960) used a cone penetration test to characterize the hardness of moist fragipan material from certain soils of Michigan. The fragipan had higher resistance than other horizons, and the fragipan subhorizon judged hardest by field examination offered the highest resistance. Rutledge and Horn ( 1 965) employed a needle penetrometer to study fragipans in soils of Arkansas. Penetration at several moisture tensions was somewhat lower for the fragipan than for the horizon immediately above. Grossman et al. (1959a) used a “dropshatter” test to characterize a fragipan soil from Illinois developed in loess. The fragipan resisted shatter more than horizons above, but less than the C horizon beneath. 4 . Brittleness

There are very few laboratory measurements of brittleness. Grossman ( 1954) compared the abruptness of rupture under unconfined compression of two fragipan materials. At a moisture content near that retained against 5 bars, the fragipan with less than 5 percent clay failed abruptly; the one with 15 percent clay showed some plastic deformation. The concept of brittleness as applied to fragipan soil material has not been clearly formulated. The reference state with respect to moisture tension for medium textures probably should be about 1/3 bar. Many moist, coarse-textured soil materials exhibit brittleness. Even some moist clayey materials exhibit brittleness if there are numerous well expressed planes of weakness in the test specimen; an example would be the B horizon of some Oxisols. Most fragipans, however, have incomplete, weak, and commonly widely spaced structural planes. Their brittleness is usually not the result of rapid failure along structural planes. Rather, it would appear to be related to the bonding by clay-size material of the sand and silt grains (the s-matrix as defined by Brewer, 1964). For coarse-textured fragipans, much of the clay occurs as braces between sand and silt grains. Amounts of clay are insufficient for local plastic movement. The leading point of a crack is not subject to blunting. Rather, the stress remains concentrated, leading to rapid propagation of the crack. Such an explanation is not applicable to medium and fine textures. In some of these instances, the highly sepic micromorphology may contribute to brittleness (Rutledge and Horn, 1965).

254

R.

B. GROSSMAN A N D F. J. CARLISLE

V. Fragipans and the Soil Water Regime

A. WATERTABLES Several investigators have reported on the water tables of soils with fragipans (Gile, 1958; Nettleton, 1965; Spaeth and Diebold, 1938; Lyford, 1964; J. H. Huddleston and Olson, 1967). Many data exist in mimeographed reports; these often are available from state soil survey organizations. Some of the studies employ lined wells placed at depths above, within, and below the fragipans. In principle, this arrangement permits detection of a perched water table above the fragipan. Other studies employ unlined wells which do not permit detection of a perched water table. Interiors of the large structural units may be considerably below saturation while free water is present between them. Lined wells that terminate within these structural units may not indicate the presence of this free water. Much of the area of fragipan soils now supports trees. Removal of water by transpiration by trees usually is greater than when the land is cleared. The water table regime under trees may underestimate the wetness of the soil when it is cultivated or used as sites for construction. On the other hand, measurements in cleared areas may indicate shallower water tables than are actually present over most of the area. Soils with fragipans are subject to the rather well defined seasonal pattern in water table depth found over much of the eastern United States. Water tables rise in late fall and remain high until transpiration by plants becomes appreciable. Often the drop in water table occurs about when the trees leaf. Water-holding capacity of the fragipan is usually low (Section V, B), and small additions of water may raise the water table markedly. Soils are placed in wetness classes mainly on inference based on the depth to and expression of mottles or low chroma parts. A significant practical question is the extent to which the water table regime of soils with fragipans differs from associated soils in the same wetness class.

B. WATER-HOLDING CAPACITY Laboratory estimates of field capacity or maximum water retention must be viewed cautiously. The determinations may be on fragmented samples resulting in a significant overestimate of water retained against low tension. As previously indicated, under field conditions the interior of the large structural units of the fragipan may not contain free water, even though free water is present in the cracks between the structural units (Nikiforoff et al., 1948; Carlisle, 1954; Alexander, 1955). Conse-

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

255

quently, maximum water retention in place would be below that calculated from the total porosity of the moist fabric. R. M. Smith and Browning ( 1946) commented on the low water content of fragipan material after wetting in the laboratory, even under vacuum. They suggested that the fragipan material may have appreciable porosity that does not fill readily with water. Comer and Zimmerman ( 1 969) reported that for a 3-year period the water content in the fragipan of a wet soil in Vermont ranged only from 19 to 23 percent by volume. The soil is not subject to recharge by upward water movement from a regional water table; recharge is by downward moving water with perhaps a lateral component of movement. The constancy of the water content over this period suggests a high degree of isolation from both withdrawal of water by plants and additions from precipitation. The effective contribution of the fragipan to the water holding capacity of the soil at any point in time over this period would appear to have been small. C. MOVEMENT OF LOW-TENSION WATER Laboratory measurements of the saturated hydraulic conductivity have been reported by R. M. Smith and Browning ( 1 946), Grossman et al. ( 1 959a), Yassoglou and Whiteside ( 1 960), and Pettiet (1 964). Values range from 0.01 to 1 inch per hour. Large numbers of field percolation determinations have been made. In some areas, these are required by law in planning small-scale sewage disposal systems. Hill ( 1966), Alexander (1955), and J . H. Huddleston and Olson (1967) have discussed procedures and present values for soils with fragipans. Horizons above the fragipan are usually quite pervious unless altered by man’s activities. Infiltration usually is not limited by a horizon above the fragipan until near-saturated conditions prevail. Fragipans have a lower saturated hydraulic conductivity than do horizons above. Consequently, low-tension water accumulates at the top of the fragipan and moves laterally. Fragipans do not necessarily have lower saturated hydraulic conductivity than the horizons beneath (for example, Yassoglou and Whiteside, 1960; Alexander, 1955). If the underlying soil materials are pervious, then the saturated hydraulic conductivity of the fragipan may be a minimum for the profile. There may be several reasons for the low saturated hydraulic conductivity of fragipans. Lack of vertical continuity of interped pores and isolation of pores within peds may have importance. O’Neal (1952) discussed the relationship between perviousness and vertical continuity

256

R. B. GROSSMAN A N D F. J. CARLISLE

of pores. Some fragipans have low total porosity. Those with moderate porosity tend to have appreciable clay (for example, Rutledge and Horn, 1965). The area that actually conducts low-tension water in the field may be quite small, limited to the periphery of the large structural units. Hydrology studies of watersheds reflect the complexity of the total natural condition. Minimum discharge rates are indicative of the integrated amounts of water moving through the soils of the area. Comer and Zimmerman ( 1 969) reported that the minimum discharge rate for a watershed in Vermont where wet soils with fragipans occupy 44 percent of the area is a magnitude lower than for a contiguous watershed where such soils occupy 22 percent of the area. They suggested that the combination of low permeability of the fragipan and the high water-holding capacity of the horizons above the fragipan is largely responsible for the lower minimum discharge rate of the watershed with the greater proportion of wet fragipan soils. VI. Genesis of Fragipans

This section has been written on the underlying assumption that clay is the bonding agent; moreover, it is assumed that this clay ranges widely in mineralogy and surface chemical properties. A. BONDINGO F THE FRAGIPAN 1 . Silicate Clay as the Bonding Agent Several workers have proposed that silicate clay is the principal bonding agent (R. M. Smith and Browning, 1946; Nikiforoff et al., 1948; Carlisle, 1954; Knox, 1957; Jha and Cline, 1963; Yassoglou and Whiteside, 1960; Comerma, 1964; Hutcheson and Bailey, 1964). Dispersing agents for clay have been shown by Comerma (1964), Knox (1957), and Jha and Cline ( 1 963) to disaggregate the fragipans of certain soils more completely than treatments designed to remove silica, hydrous iron oxides, or hydroxy aluminum compounds. There is no evidence from the large number of particle-size analyses using standard dispersing treatments that the clay in fragipans resists disaggregation. Close packing of the sand and silt is thought to contribute to the effectiveness of the clay as a bonding agent. Nikiforoff et al. ( 1 948) placed principal emphasis on the close packing and related interlocking of the sand and silt. Hutcheson and Bailey ( 1 964) also emphasized closeness of packing. They write, “. . . [we] visualize pan horizons as brick and mortar structure, i.e., silt particles acting as bricks held in a dense mass by clay mortar.” The thin-section observations by Jha and Cline ( 1 963) fit

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

257

the “brick and mortar” model. Other workers have placed greater emphasis on the clay bridges between sand and silt grains (Knox, 1957; Yassoglou and Whiteside, 1960; Grossman and Cline, 1957) with proportionately less on a dense, continuous filling of clay in the interstices. More would seem to be involved than disposition of the clay. In some horizons of clay accumulation, bridges of clay between sand grains are prominent (Soil Survey Staff, 1960, p. 41). Yet, when moist, these horizons do not necessarily exhibit the brittleness and rigidity of fragipans. Several writers (Carlisle et a f . , 1957; Yassoglou and Whiteside, 1960; Grossman and Cline, 1957) have commented on the dual role of clay. At low clay contents the bridging of clay between sand and silt grains lends rigidity. At higher clay contents, volume changes with moisture promote formation of cracks that reduce the rigidity when moist. 2. Other Bonding Agents a. Silica. The earlier literature contains suggestions that silica is the bonding agent (Marbut, 1935; Krusekopf, 1942; Winters, 1942). At the time, total analyses of soils received more emphasis than they do now. Total analyses of fragipans low in clay, with the sand and silt dominated by quartz, indicate high proportions of silica, which would be consistent with the idea of a siliceous bonding agent. Studies were then current on silica cementation of indurated horizons in certain soils of western United States (for example, Nikiforoff and Alexander, 1942). These studies were employed to support bonding by silica in fragipans. There is similar informal conjecture today. The argument runs that as silica does cement some soils, perhaps small amounts -less than that detectable by methods employed to date-may play a similar role in fragipans. Knox ( 1 957) presented the only experimental data for the implication of silica. Extraction of silica with various reagents (Section IV, A, 2) has not shown a consistent maximum in the fragipan. Baker (1967) determined the mineral stability by solubility investigations for the strongly eluvial fragipan horizon of certain soils developed in loess over cherty limestone residuum in Missouri. Kaolinite and quartz were stable. The concentration of silica was well below that supportable by opal. McKeague and Cline ( 1 963) suggested that concentration of silica by evaporation may be important in the surface adsorption of silica from solution on soil particles. Fragipans as a rule, however, are not subject to frequent and pronounced desiccation. R. W. Miller ( 1967) stressed the importance of small differences among horizons in controlling the translocation and deposition of silica. H e studied soils from Utah that are not subject to high

258

R. B. GROSSMAN A N D F. J. CARLISLE

precipitation, and one of the soils was formed in materials containing volcanic glass. Such differences from soils with fragipans of the eastern United States must be considered in the application of research on silica solubility in soils of the western United States. Calcite apparently adsorbs little or no silica (McKeague and Cline, 1963), and presence of carbonate appears to increase the adsorption of silica (R. W. Miller, 1967; McKeague and Cline, 1963). The scarcity of fragipans in calcareous soil materials and the influence of carbonates on silica solubility relationships might be related. b. Aluminum and Iron. Hydrated oxides of iron or aluminum have been suggested as bonding agents. Knox ( 1 957) and Comerma ( 1 964) found no evidence for their implication. Alexander (1955) failed to find a maximum in extractable aluminum in the fragipan. Anderson and White (1958) presented evidence for iron oxides contributing to the rigidity in a fragipan. Horn and Rutledge (1965) suggested that segregations of iron oxides in association with silicate clay are important in determining rigidity. Nettleton et al. ( I 968b) suggested that amorphous aluminum compounds may play a role in the hydrogen bonding of fragipans. c. Water Films. Surface tension effects associated with water films lend rigidity to moist soil material (Knox, 1954; Fountaine, 1954). For spherical particles, cohesion increases proportional to the reciprocal of the radius. As a point of reference, the cohesion for spheres with a radius of 10 p ranges roughly from 0.1 to 1 kg./cm.2, depending on the assumptions about packing and whether the pores are partially or entirely filled (Knox, 1954). This range compares with a crushing strength of 4 kg./cm.2 for a moist fragipan soil material of silt loam texture (Grossman, 1954). Water films may contribute appreciably to the rigidity of some moist fragipan material. It is doubtful, however, that the greater rigidity when moist of fragipans than of otherwise comparable soil materials is due to water films.

3 . Mechanisms of Bonding Knox ( 1 954) concluded from a review of the literature that chemical forces rather than mechanical interlocking are probably responsible for the strength of fragipan material. Nettleton et al. (196%) propose that amorphous clays adhere strongly to the disordered surface of quartz grains by hydrogen bonding. Acidity of the soil material fosters this bonding. Surface negative charge of the clay is low and hydroxyl groups exert a strong influence. Infrared analyses indicate the presence of hydrogen-bonded hydroxyls. These are lost on heating to 300°C. Heating to this temperature leaves the fragipan material soft and loose. Brittle-

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

259

ness was restored by rehydration under moderate relative humidity, Knox (1954) studied the effect of heating to as high as 500°C. on the strength of fragipan material. The strength was not reduced appreciably. His observations would appear to be at variance with those of Nettleton et al. ( 1968b). The fragipans investigated, however, differ markedly. Knox ( 1 954) studied fragipans with strong influence of the parent material (sequences IIa, IIc of Fig. 2); illite dominates the clays. The fragipan studied by Nettleton et al. (1968b) occurs in a strongly eluvial horizon (sequence IIh, Fig. 2) where accumulation of amorphous clay through soil development would be more likely; kaolinite is a prominent clay mineral. The relationship between the properties of the exchange complex and attraction between clay particles has received much attention in the field of soil physical chemistry. Nettleton et al. (1 968b) apply concepts from this field to bonding of fragipan material. Many fragipans have low pH and contain appreciable aluminum extractable with a neutral salt. Calcareous fragipans, moreover, are rare. Applying the ideas presented by Emerson and Dettmann ( 1960), both the low pH through the increase in positive charge and consequently stronger electrostatic attraction, and the presence of trivalent aluminum, would increase attraction between clay particles. Carbonate would reduce the attraction because the resulting pH leads to low positive charge and to precipitation of the aluminum. Such ideas, however, do not provide a general explanation for the rigidity of fragipans. Many kinds of soil horizons are acid and have high extractable aluminum. Moreover, some fragipans have pH values near neutrality.

B. DEVELOPMENT OF FRAGIPAN SOILS I. Inheritance of Properties Fragipans in some areas of the Coastal Plain south of Wisconsin glaciation occur on the older geomorphic surfaces but not on the younger surfaces (Daniels et al., 1966; Nikiforoff et al., 1948; Nikiforoff, 1955). This is evidence that these fragipans may be relicts of an older environment. Restriction of fragipans to the older, more stable parts of the landscape is not limited to the Coastal Plain. This is the pattern of occurrence in parts of the Ozark Plateau, for example. Investigators of fragipans in areas of Wisconsin glaciation do not agree on the extent to which fragipans may be relict features. Denny and Lyford ( 1 963) wrote for the area of Wisconsin glaciation along the southwestern New York-Pennsylvania border: “The soils are relatively young, are in equilibrium with the present environment, and contain few, if any, features acquired during past weathering intervals.” In contrast, Olson and Hole ( 1967- 1968)

260

R. B. GROSSMAN A N D F. J. CARLISLE

placed emphasis on the importance of changes in climate in the development of fragipans in soils in northeastern Wisconsin. Some fragipans bear a strong influence of the parent material. Not only composition, but organization of the soil fabric, may be largely determined by the parent material (Section IV, B, 1). In some soils parent material has such a strong influence that the distinction between glacial till and the fragipan may be difficult to establish (Section IX). 2 . Catastrophic Development

Processes common to a periglacial environment have been implicated in the development of fragipans. Fitzpatrick (1956) offered evidence that three features common to some fragipans can be produced by freezing wet soil. These features are platy structure, discontinuous spherical or vesicular pores, and a sheathing of fine material around pebbles. Yassoglou and Whiteside (1960) and Jha and Cline (1963) presented evidence against a periglacial origin being applicable generally and the main agent in fragipan formation. Fragipans occur in areas thought not to have been subject to periglacial influence. This would seem a convincing argument against a periglacial origin for all fragipans. Some fragipans, however, may have strong relict influence of a periglacial environment. The fact that the upper surface of fragipans occurs at predictable depths and generally conforms to the land surface does not rule out relict periglacial influence. As Nikiforoff (1955) and Lyford et al. (1 963) pointed out, the upper boundary of the fragipan could be controlled largely by the lower limit of obliteration of periglacial influence by soil development. Nikiforoff (1955) explored in detail the possibility that the gross prismatic structure of many fragipans has its origin in a periglacial environment. The large prisms would be formed by frost wedges or would be the result of dessication and contraction of soil material having an originally high water content. Jha and Cline (1963), studying a fragipan in lacustrine sediments, suggested that desiccation cracks which formed soon after drainage of the lake may have been the origin of the polygonal pattern. Olson and Hole ( 1 967- 1968) suggested that the large prisms are related to desiccation cracks formed during a dry period after close of the Pleistocene. The top of the fragipan often coincides with discontinuities in composition, organization, or both of the soil material. Evidence for changes in composition are given in the studies by Nikiforoff et af. (19481, Scrivner ( 1960), Rutledge and Horn ( 1 9 6 3 , Bailey ( 1964), Vanderford and Shaffer ( I 966), Beavers ( 1 960), and Calhoun (1968). An example of a

FRAGIPAN SOILS OF THE EASTERN U N I TED STATES

26 1

change in organization would be the shift from ablation till (let down by ice wastage) to denser till thought to have been compacted by the weight of ice. Such changes, either in composition or organization, although not directly responsible for the fragipan themselves, may have lead to conditions favorable to formation of a fragipan.

3. Incremental Development Fragipans occur beneath eluvial horizons. The overlying B horizon, if present, commonly is a type of cambic horizon which appears to have gone through a stage of active clay removal. If an overlying B horizon is an argillic horizon, it commonly is one in which accumulation of illuvial clay does not greatly overbalance removal. Tavernier and Smith (1957) pointed out that Gray-Brown Podzolic soils with fragipans commonly have either heavy bleached silt coatings on the ped surfaces of the B horizon above the fragipan or a distinct bleached horizon of eluviation between the B and the fragipan. Where the latter horizon occurs, it tongues into or interfingers with the overlying B horizon. Tavernier and Smith interpreted the tonguing or mingling of eluvial and illuvial horizons in these soils as evidence of destruction of the B horizon. Fragipans are therefore subject to the accumulation of substances from the horizons above. The substances may move as particles or in solution. Much speculation has centered on whether a precipitated substance is the bonding agent. Wetting and drying is probably the principal agent responsible for mobilization and translocation of silt and clay into the fragipan. Illuvial silt would reduce porosity and perhaps increase mechanical strength. The role of illuvial clay is less clear-cut. Many investigators have suggested that illuvial clay reduces porosity. Addition of clay, however, may increase volume change with change in moisture content and lead to more large planar pores. Furthermore, illuvial clay coats planar pores and thereby increases the prominence of structural planes of weakness. Wetting and drying may also cause translocation of substances within the fragipan. The mottled color pattern and low-chroma parts of many fragipans strongly suggest translocation of hydrous iron oxides. Once the hydrous iron oxide coatings have been removed, the silicate clay may be more subject to movement. In strongly eluvial horizons, removal of silicate clay and hydrous iron oxide coatings may have so weakened the fabric that wetting and drying can lead to extensive rearrangement to produce a denser fabric (Daniels et al., 1966; Nettleton et al., 1968b; Yassoglou and Whiteside, 1960; Pettiet, 1964). Strongly eluvial subsoil horizons tend to be zones of accumulation of free water. This free water

262

R. B. GROSSMAN A N D F. J. CARLISLE

commonly extends throughout the fabric; it is not restricted to vertical planes. Such high water contents may weaken the fabric and make it prone to rearrangement. The subject of incremental development leads to the question of time required to form a fragipan. Fragipans have formed since the close of the Pleistocene. Jha and Cline ( 1 963) reported that a fragipan formed in lacustrine sediments deposited in a lake that drained seven to eight thousand years ago. On the other hand, fragipans are not found in very youthful deposits. Lyford er al. ( 1 963) offer evidence that the fragipans in certain soils of Massachusetts must be atleast 500 years old. Many fragipans show the marks of illuviation of clay, its removal, or both. Such alteration does not occur rapidly. Fragipans are not necessarily restricted to old soils, but neither do they form in a matter of hundreds of years.

4 . Weak Disturbance Many fragipans have undergone only weak disturbance. Strong influence of the organization of the parent material implies weak disturbance as does strong expression of pedological features such as gross polygonal structure, a pronounced pattern of accumulation and depletion of hydrous iron oxides, and moved clay bodies. Fragipans are not subject to frequent and large changes in volume. Many fragipans occur deep enough and contain so few roots that intense desiccation is rare. Interiors of the large structural units of some fragipans resist wetting (Section V, B). Many fragipans have a low potential for volume change because of moderate or low clay contents. Some fragipans do have fairly high clay contents. The latter, however, tend to occur at appreciable depths, where frequent and intense desiccation would be less likely. Fragipans do not occur in shallow, clayey horizons unless the soil has been severely eroded. Disturbance by tree throw and by mass movement is greater above than within the fragipan. Lyford et al. (1 963) showed for an area of soils developed in glacial till on steep slopes that the base of small gravity movements coincided with the upper boundary of the fragipan. R. M. Smith and Browning ( 1 946), studying soils with fragipans that occur on steep slopes in West Virginia, suggested that the base of the small slip scars common to the area was at the top of the fragipan. The base of the zone of disturbance due to tree throw commonly coincides roughly with the top of the fragipans (Lyford and MacLean, 1966; Denny and Lyford, 1963; Olson and Hole, 1967-1 968; Mueller and Cline, 1959). Estimates of the rapidity of mixing by tree throw point to its importance. Denny and Lyford ( 1 963) concluded that for soils of the northeastern United States

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

263

much of the upper 2 feet has been disturbed by tree throw. The study by Mueller and Cline (1959) tends to substantiate this conclusion. Milfred e? al. ( I 967) emphasized the importance of tree throw in northeastern Wisconsin. Fragipans commonly occur at shallower depths in the wetter soils of a local association. Trees may have rooted less deeply in the wetter soils because of the shallower depth to free water. Disturbance related to the presence of the tree roots, such as the volume change resulting from water withdrawal and recharge, root ramification, tree throw, and root movement resulting from tree sway, would all have extended to shallow depths. In this view, the upper boundary of the fragipan would be largely determined by the control exercised by the water table regime on the depth of frequent disturbance. Presence of a fragipan may increase disturbance of horizons above. Mixing by tree throw, for example, is probably more frequent above the fragipan because the fragipan limits the depth of support roots (Section VII, A). Higher rates of disturbance above the fragipan may increase the perviousness of these horizons. Low-tension water consquently may move more rapidly through the horizons above and accumulate at the top of the fragipan. Accumulation of this low-tension water would affect development of the fragipan. Weak disturbance within the fragipan and stronger disturbance above the fragipan are two sides of the same coin. Properties of the soil are a reflection of both. VII. Fragipans and Soil Use

A. PLANTGROWTH Fragipans unfavorably influsnce growth by restricting rooting, either through mechanical impedance or by creating saturated conditions. Soils with shallow fragipans usually have high water tables over extended periods. Effects of mechanical impedance and of the shallow water table on plant growth often are confounded. Improving the plant nutrient status of soils has become increasingly feasible. One result has been to increase the relative importance of a physical limitation such as presence of a fragipan. This is true particularly in the southeast part of the country. The Coastal Plain of the Southeast contains much of our potential new land for farming (Bartelli, 1968). In parts of the Coastal Plain, soils with fragipans are among the principal soils (Fig. 3). Depth to the fragipan partially determines its significance to plant growth. The critical depth depends on the plants. For many plants, the

264

R. B. GROSSMAN A N D F. J. CARLISLE

influence diminishes rapidly at depths greater than 50 cm. Erosion brings the fragipan closer to the soil surface and thereby increases its significance. Erosion has greatly reduced the average depth to the fragipan in certain soils, such as the Beltsville and Grenada soils (see Appendix). Thickness of the fragipan may determine its significance to soil use and management. The apparently small influence of the fragipan in the McBride soil (see Appendix) may be largely a reflection of its thinness. Trees are the principal vegetation on much of the area of soils with fragipans. The density of tree roots is low in the fragipan (Olson and Hole, 1967- 1968; Lyford and MacLean, 1966; Mueller and Cline, 1959). The influence of root distribution on tree throw has received much attention. Olson ( 1 962) suggested that the restriction of roots by fragipans is greater for the larger roots that give mechanical support to trees than for the smaller feeder roots. Mueller and Cline ( 1 959) observed that calcareous glacial till of similar bulk density to the fragipan is not as effective a barrier to rooting. Goodlett ( 1960) discussed the confounding influences of the water table regime and depth to the fragipan for an area in central Massachusetts.

B. ENGINEERING MANIPULATION Depth to the fragipan influences the kind of agricultural drainage system needed. If the fragipan is shallow, the system should be designated to remove surface water. If the fragipan is deep, then ditches or tile lines may be feasible. Thickness of the fragipan may also affect suitability of a drainage system. If the fragipan is thin, although shallow, tile or ditches may be feasible. Fragipans increase construction costs in several ways. Difficulty in digging makes small excavations more expensive. The slow movement of low-tension water in the pan causes several problems. Pervious seepage beds for septic tanks may be required (J. H. Huddleston and Olson, 1967). Accumulation of water at the top of the fragipan may incease the time required for drainage after heavy precipitation during earth-moving operations. Fragipans beneath surfaced runways or roads may lead to accumulation of water in the subgrade, which loses strength and is subject to being pumped out under repeated loading and relaxation. Lateral water movement above the fragipan causes problems where cuts are made on sloping land. Road design must provide for the interception of water moving laterally along the top of the fragipan. Cuts must be deep enough and fill thick enough so that the road surface is either well below or above the top of the fragipan. The location and manner in which people live in the United States

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

265

increase the importance of fragipans. Most of the people live in the East where soils with fragipans are extensive. Suburban living usually involves individual houses. The scale of construction for individual houses is small enough to be influenced by the presence of a fragipan. The largescale construction in the urban core would not be so affected. Many communities in the eastern United States were established in valleys where the soils lack fragipans. Today the communities are expanding out of the valleys into areas where fragipans may be prevalent (Olson, 1966). Moreover, in these upland areas the more favorable sites for construction from the viewpoint of gentleness of slope often have soils with the strongest fragipans at the shallowest depths. C. J. Thomas ( 1 966) discussed the influence of fragipans on planning the development of a community in Massachusetts. She wrote: “The soils present problems when used for commercial, industrial, or highdensity residential purposes where community sewage disposal is not available. The hardpan [fragipan] restricts water from moving downward readily. When the soil above the hardpan is saturated, water tends to move downslope along the top of the hardpan. Hardpan layers have slow to very slow permeability and individual sewage effluent disposal systems do not function satisfactorily.” T o move beyond such a qualitative description to quantiative statements, which often are required by local ordinances, leads into problems. These are well illustrated in the several studies referred to by Olson ( 1966). T o obtain reliable measurements requires expenditure of a great deal of effort. Even the percolation test, simple in principle, is complicated in application (Hill, 1966; Huddleston and Olson, 1967). The water table regime is another example. Measurements are necessary over several years (J. H. Huddleston and Olson, 1967), and even then may not indicate conditions resulting from very infrequent weather events. Furthermore, as construction often results in diversion of water from one place to another, the measurements may not be applicable to the actual soil use. VIII. Classification of Fragipan Soils

The term fragipan is a coined name from the Latin root for brittle. It was proposed by G. D. Smith in 1946 (unpublished working papers of National Conference of the Cooperative Soil Survey) to identify soils having horizons that had been called silica hardpans by Winters ( 1942), silt-pans by R . M. Smith and Browning ( 1946), and simply hardpans by a number of workers in the eastern part of the country (for example, Carr, 1909; Hearn, 1924; Howe er al., 1924). The term was adopted by the

266

R. B. GROSSMAN A N D F. J. CARLISLE

Soil Survey Staff of the U.S. Department of Agriculture ( 1 95 1, 1960) and has been in general use for about 20 years. But that is getting ahead of the story. The kinds of horizons that are designated fragipans were recognized as important characteristics of some soils fairly early in the course of the soil survey of the country. Early in this century, Carr (1 909) wrote about the compact “hardpan” in the subsoil of Volusia soils in New York and its importance to plant roots and water relations in the soil. The concept of the Volusia series at that time, however, was based in large part on geology, and the series was not characterized as having a pan (Marbut el al., 1913). A few years later, Carter and Hull (1916) described the Leonardtown soils in Maryland as having a compact hardpan in the subsoil, and subsequently the presence of the hardpan became part of the concept of that series (H. C. Smith and Rose, 1924). A substantial number of soil series were characterized as having a compact, slowly permeable hardpan layer of the same general nature during the two decades that followed. As more detailed field studies of soils have been made and as more kinds of interpretations of soil surveys have been required, the number of kinds of soils in which fragipans have been recognized has increased. At this writing, approximately 250 soil series in that part of the country lying east of the Great Plains are characterized as having fragipans. The presence or absence of fragipans apparently was not a factor in the classification of soils into groups broader than series during the first two decades of the soil survey in this country. During that period groupings of series were primarily in terms of geography, geology, or physiography (Marbut et al., 191 3; Marbut, 1935), an approach that slowly faded during succeeding decades. Concepts of soil development and of “normal soils” that were to affect strongly soil classification in this country in later years were gaining ground in the late teens. Describing the soils of Charles County, Maryland, H. C . Smith and Rose (1 924) wrote that the profile of the Leonardtown series differs from the normal mature soils of the region in the presence of the compacted zone (fragipan) in the illuvial horizon and somewhat lighter than normal color of the overlying horizon. For the same reasons, Perkins and Bacon ( 1 925) considered Leonardtown soils to be “postmature soils.” This was in accord with Marbut’s ideas about “normal,” “mature,” or “fully developed” soils that strongly influenced his soil classification (Marbut, 1928). Seemingly, soils with fragipans and other kinds of pan horizons did not have a clear place in the higher categories of his classification. They were recognized as “overdeveloped soils” in category 111,

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

267

within which groupings were based on stage in development, and in the lower two categories. Thus, Marbut (1935) discussed the Grenada series under the heading Red and Yellow soils and the Leonardtown series under Gray-Brown Podzolic soils, but these series apparently were not considered bona fide members of those great soil groups because of their fragipans. The logical inconsistencies of Marbut’s classification were corrected to a large extent by the subsequent classification scheme outlined by Baldwin et al. ( I 938). In this system, soils with fragipans seemed to have a legitimate place in the great soil group of Planosols, which had a place in the order of Intrazonal soils. The term, Planosol, was “proposed to cover those soils with claypans and cemented hardpans not included with Solonetz, Ground-Water Podzol, and Ground-Water Laterite.” Thorp and Smith ( I 949) modified the definition as follows: Intrazonal soils having one or more horizons abruptly separated from and sharply contrasting to an adjacent horizon because of cementation, compaction, or high clay content. Soils of the Grenada, Leonardtown, and other series with fragipans were classified as Planosols in this system (for example, see Fox et al., 1958; Leighty and Wyatt, 1950). Many soil scientists recognized that the Planosol group as then constituted was very heterogeneous and that most Planosols had important properties in common with those of the zonal soils with which they were associated (unpublished working papers of the I 947 National Conference of the Cooperative Soil Survey). But agreement on an alternative classification was not forthcoming. Thorp and Smith (1949) noted that an earlier proposal to elevate Planosols to the rank of a suborder of Intrazonal soils and to recognize groups of soils with “silt pans,” clay pan soils, and certain other soils as subdivisions in the great soil group category had not been unanimously accepted by the National Conference of the Cooperative Soil Survey. During the next dozen or so years, several different approaches were followed in classifying soils with fragipans. In a number of soil surveys in the southeastern part of the country soils with fragipans were classified as Planosols according to the system outlined by Baldwin et al. (1938),as modified by Thorp and Smith (1949) (for example, McNutt et al., 1959; Fox et al., 1958; Leighty and Wyatt, 1950). In the latter part of that period some writers classified only the somewhat poorly drained and the poorly drained fragipan soils as Planosols and included the better drained fragipan soils in the zonal great soil groups to which the soils belonged on the bases of their features other than the fragipan (for example, Love et al., 1959; Matthews et al., 1961). In the northeastern states soils with

268

R. B. GROSSMAN A N D F. J. CARLISLE

fragipans were not classified as Planosols but were placed in the great soil group to which they would belong on the basis of their other characteristics (for example, Cline, 1955; Taylor, 1960; Pearson and Cline, 1960; Shearin and Hill, 1962). Nevertheless, the presence or absence of a fragipan was emphasized in the series classification of soils in that part of the country. In discussing Fragipan Planosols Cline (1 952) pointed out that, unlike the claypan of the Planosol, which occurs in the position of the B horizon of the associated normal soil, the fragipan occurs beneath, and in addition to, all the essential horizons of the solums of the Red-Yellow Podzolic, Gray-Brown Podzolic, Podzol, Brown Podzolic, and probably some other zonal great soil groups. He wrote, “In one sense, climate and vegetation appear to have exerted little control over its [the fragipan] character in comparison with their influence on horizons above.” Thus, there could be no modal horizon sequence in a great group that included all soils with fragipans. Cline ( 1 952) suggested that a great group of fragipan soils might be defined for the poorly drained soils with fragipans, as the variation in horizon sequence seemed relatively minor in those soils throughout the several soil zones. The imperfectly, moderately well, and well-drained soils with such pans could then be placed in the appropriate zonal great group and be set apart in the next lower category from soils lacking fragipans. The decision to revise the soil classification system used by National Cooperative Soil Survey and the reviews of early drafts of a new system continued to focus attention on how soils with fragipans would be best classified at categorical levels above the soil series (unpublished working papers of the National Cooperative Soil Survey). Some soil scientists wanted to make claypans, fragipans, and indurated pans bases for grouping soils at the great soil group level or above; others wanted to bring them in at the series level. Most reviewers thought the initial approximation of the new classification system overemphasized the importance of Planosols by giving them a separate place high in the system and advised that pan horizons be made diagnostic in lower categories. The idea of a broad group of Planosols, subdivided in the next lower category into soils with claypans, soils with fragipans, soils with indurated hardpans, etc., survived in the 2nd Approximation prepared in 1952, but at a lower level of the system. That was the last time in the development of the current system that soils with the various kinds of pan horizons were placed together in a class resembling the old Planosol great soil group. In subsequent drafts of the system, soils with and without fragipans were placed together in classes that roughly approximated the

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

269

Podzol, Gray-Brown Podzolic, Red-Yellow Podzolic, and a few other great soil groups. These were subdivided into classes with and without fragipans, among other criteria in the next lower category, which approximated the current great group level. With some variations, and with one exception, this arrangement persisted through the 7th Approximation of the classification system (Soil Survey Staff, 1960) and to the system that is in use today. The exception was a trial in the 5th, 6th, and 7th Approximations ( 1 956-1 960) of a proposal to keep Spodosols with and without fragipans together at the great group and subgroup levels of the system and to make the distinction on the presence or absence of a fragipan in the family category. (The class of Spodosols is approximately equivalent to the Podzol and part of the Brown Podzolic great soil groups.) The rationale for the proposal was the belief that a fragipan or a comparable horizon was normal in Spodosols. Those Spodosols that lack the fragipan were thought to be too coarse textured for a fragipan to form (Soil Survey Staff, 1960). After a few years’ trial, the classification system was amended to set apart Spodosols with and without fragipans in the great group category. Use of the fragipan as a differentia within Spodosols then paralleled its use in the orders of Inceptisols, Alfisols, and Ultisols. In the current soil classification system (Soil Survey Staff, 1967) fragipans are used uniformly to set apart one great group in each of the eleven suborders of the four orders in which soils with fragipans occur. These are listed in Table I. Note that each of the four aquic suborders has a great group of soils with fragipans but the dry or seasonally dry suborders do not have such great groups. The family category of the current soil classification system is designed to group series that are similar in certain properties, such as texture and mineralogy, that have importance in the use and management of soils. The top of the fragipan is the base of the part of the soil that is considered in applying the texture and mineralogy criteria for the classification of series into families. The rationale is that plant roots do not sufficiently exploit the soil within and below the fragipan. IX. Unresolved Problems

There are difficulties in the identification of fragipans. High bulk density and consistence have not been satisfactory criteria for recognition of fragipans in loess-derived soils of the lower Mississippi Valley (personal communication from L. J . Bartelli). Brittleness when moist does not suffice for identification of fragipans and their distinction from soils with plinthite in North Carolina (personal communication from Forrest

270

R. B. GROSSMAN AND F. J. CARLISLE

TABLE I THEGRE \ T GROUPS OF SOILS HAVINGFRAGIPANS, SHOWlNG THEIRDISTRIBUTION OF THE CURRENT SOILCLASSIFICATION SYSTEM AMONGT H I ORDERSA N D SUBORDERS

Order Entisol. Vertiso' Inceptisc '

Aridis, I Molli>,' Spodo.

1

L,uhorcler

Great groups having fragipans"

Order

Suborder

-

-

Alfisols

Aqualfs Boralfs Udalfs Ustalfs Xeralfs Aquults Humults Udults Ustults Xerults

_-

Andepts Aquepts Ochrepts Plaggepts l'ropepts I I mbrepts Iquods kerrods H timods Orthods

Fragiaquepts Fragiochrepts Fragiurnbrepts

Ultisols

-

Fragiaquods Fragihumods Fragiorthods

Oxisols Histosols

-

Great groups having fragipans" Fragiaqualfs Fragiboralfs Fragiudalfs Fragiaquults Fragiudults -

Data for illuslrative profiles for seven of these are in ttre 7rh Approximation (Soil Survey Staff, 1960) as follows: Fragiaquepts-profile Nos. 30, 48; Fragiochrepts-profile No. 24: Fragiorthods-profile No. 29; Fragiaqualfs-profile No. 82; Fragiudalfs-profile Nos. 84, 98: Fragiaquults-profile No. 94; Fragiudult-profile No. 97.

Steele). These problems concern fragipans showing strong alteration of the parent material. There are also problems of identification of fragipans that show weak alteration of the parent material. Brittle zones occur in the upper substratum of some soils developed from compact glacial till. If many feet thick, the zone can be eliminated as a fragipan on scale. In some soils the brittleness decreases in expression within 2 or 3 feet. These zones are difficult to distinguish from fragipans as presently defined. Modifying the definition of fragipans to include certain morphological features, such as bodies of moved clay or expression of a gross polygonal structure, would provide a basis for excluding certain of these zones of compact glacial till. Such a definition, however, might exclude some horizons of Spodosols and Inceptisols that are currently classified as fragipans and thus raise other problems. Fragipans are now a criterion for distinguishing great groups in the Comprehensive Soil Classification System. There is some concern with the categorical level at which fragipans are recognized. The concern arises from two somewhat different problems: (1) that of distinguishing

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

27 1

fragipans from other kinds of horizons with properties affecting root growth and water movement similar to those of fragipans; and (2) that of determining the lower limit of fragipan expression. From the point of view of soil use, the first problem may not be too serious. If a horizon has similar limitations to use and management of soil as a fragipan, its recognition as a fragipan does no great immediate harm. The question of the lower limit of expression for recognition of a fragipan seems more important. The problem has similarities to an evaluation of engineering test data that results in overdesign. Fragipans are recognized which may have marginal importance to use and management. Once a fragipan is recognized in a soil, however, there is a tendency to change appreciably the interpretive ratings for many kinds of soil use from the ratings assigned to otherwise similar soils. Moreover, the part of the soil that determines the family placement may change, and this may affect the correlation of soil interpretations. The name fragipan in a sense states and brings into focus an underlying problem that affects both identification and classification. A consistence property, relative fragility or brittleness, is the defining characteristic for a horizon that importantly affects soil use and exhibits a wide range in kind and degree of soil development. Importance of the fragipan does not lie in its fragility or any single feature, but rather in a group of attributes which together make it a pan. These attributes in combination cannot be readily measured. Separation of fragipans from other kinds of soil horizons may be improved by modification in the definition and collection of numerical information on consistence. Certainly, if fragipans were shown to have a unique bonding agent, the distinction from other kinds of pans would be more definite. But this would not necessarily resolve the difficulty in specifying the lower limit of expression for the recognition of a pan. X. Summary

Fragipans are subsoil horizons that are brittle and rigid when moist, that restrict root growth, and that transmit low-tension water more slowly than the overlying horizons. They occur in soils subject to net downward water movement. Trees rather than grass were the dominant vegetation a t the time the country was settled. Some fragipans have undergone strong eluviation, illuviation, or both. Others occur in soil materials that exhibit only weak alteration of the parent material. Clay is probably the chief bonding agent. Organization of the clay relative to the sand and silt grains may be of importance. Fragipans are low in organic matter and infrequently, if ever, calcareous. Otherwise, their mineralogical and

212

R. B. GROSSMAN A N D F. J . CARLISLE

chemical characteristics range widely. Extremes in texture are excluded; high proportions of particle size separates in the 0.2- to 0.02-mm. range may be conducive to fragipan formation. A gross polygonal structure is common. Blocky structure within the large structural units is at most moderately expressed; some fragipans have platy structure. Vertical continuity of large voids is restricted. Bodies of moved clay apparently are found in all fragipans. Most fragipan material is at least firm when moist; i.e., a piece held between thumb and forefinger offers moderate or more resistance to rupture. Fragipans occur deep enough to be subject to weak physical disturbance and below maximum influence of current soil development. Some fragipans may be largely relict features; others are not. Periglacial influences may have affected fragipans in the north, but not those in the south. Depth to the fragipan largely determines its influence on plant growth. Soils with shallow fragipans commonly have high water tables which may be a factor in restricting rooting depth. The importance of fragipans to mechanical manipulation of soils is largely relatable to their slow transmission of low-tension water. In the Comprehensive Soil Classification System, fragipans are diagnostic at the great group level and they are recognized in that category in four soil orders. In those orders they are used at a lower categorical level than the argillic, cambic, and spodic horizons but at a higher categorical level than features of similar importance to soil use having less genetic implication, such as the lithic contact. REFERENCES

Alexander, E. B., Jr. 1955. M.S. Thesis, Ohio State University, Columbus, Ohio. Anderson, J. U . , and White, J. L. 1958. Soil Sci. Soc. Am. Proc. 22,450-454. Bailey, H. H. 1964. Soil Sci. Soc. Am. Proc. 28,680-683. Baker, J. C. 1967. M.S. Thesis, University of Missouri, Columbia, Missouri. Baldwin, M., Kellogg, C. E., and Thorp, J. 1938. Yearbook Agr. (U.S. Dept. Agr.) pp. 979- I00 1 . Bartelli, L. J. 1968. Trans. 9th Intern. Congr. Soil Sci., Adelaide, 1966, Vol. 4, pp. 24325 1. Intern. Soc. Soil Sci. and Angus & Robertson, Sydney. Beavers, A. H. 1960. Trans. 7th Intern. Congr. Soil Sci., Madison, Wise., 1960 Vol. 2, pp. 1-9. Intern. SOC.Soil Sci. Bornstein, J . , Bartlett, K. J . , and Howard, M . , Jr. 196.5. Soil Sci. Soc. Am. Proc. 29,201205. Brewer, R. 1964. “Fabric and Mineral Analysis of Soils.” Wiley, New York. Calhoun, F. G . 1968. Ph.D. Thesis, Ohio State University, Columbus, Ohio. Carlisle, F. J. 1954. Ph.D. Thesis, Cornell University, Ithaca, New York. Carlisle, F. J . 1958. “Soil Survey of Franklin County, New York.” U.S. Dept. Agr., Washington, D.C. Carlisle, F. J . , Knox, E. G ., and Grossman, R. B. 1957. Soil Sci. Soc. Am. Proc. 21,320321.

FRAGIPAN SOILS O F T H E EASTERN U N I T E D STATES

273

Carr, M. E. 1909. US.Depr. Agr., Bur. Soils Bull. 60. Carter. W. T.. Jr.. and Hull, J. P. D. 1916. “Soil Survey of Howard County, Maryland,” U.S. Dept. Agr., Bur. Soils Field Operations 1916, Washington, D.C. Cline, M. G . 1952. “Soil Classification in the United States, Observations and Conclusions during Sabbatic Leave,” Mimeo. Rept. Cornell Univ., Ithaca, New York. Cline, M. G . 1955. Cornell Exr. Bull. 930. Comer, G . H., and Zimmerman, R. C. 1969.5. Hydrol. 7,98-108. Comerma, J. A. 1964. M.S. Thesis, North Carolina State University, Raleigh, North Carolina. Daniels, R. B., and Gamble, E. E. 1967. Geoderma 1, I 17- 124. Daniels, R. B., Nettleton, W. D., McCracken, R. J., and Gamble, E. E. 1966. SoilSci. SOC. A m . Proc. 30,376-380. Daniels. R. B., Gamble. E. E., and Bartelli, L. J. 1968. Soil Sci. 106,200-206. Denny, C. S., and Lyford, W. H. 1963. Geol. Surv. Profess. Paper 379. U.S. Dept. Interior, Washington, D.C. Emerson, W. W.. and Dettmann, M. G. 1960.5. Soil Sci. 11, 149- 158. Fitzpatrick, E. A. 1956. 5. Soil Sci. 7,248-255. Fountaine, E. R. 1954. J . Soil Sci. 5,25 1-263. Fox, C. J., Beesley, T. E., Leighty, R. G., Lusk, E., Harmon, A. B., Smith, H. C . , Methvin, C.. and Flowers. R. L. 1958. “Soil Survey of Franklin County, Tennessee,” U.S. Dept. Agr.. Washington, D.C. Gile, L. H..Jr. 1958. SoilSci. Soc.Am. Proc. 22,560-565. Gill, W. R.,and Vanden Berg, G . E. 1967. U S . Dept. A g r . , A g r . Handbook 316. Glenn, R. C. 1960. Truns. 7th Intern. Congr. Soil Sci., Madison, Wisc., 1960 Vol. 4, pp. 523-53 I . Intern. SOC.Soil Sci. Goodlett, J. C . 1960. Harvard Foresf Bull. 28. Grossman, R. B. 1954. M.S. Thesis, Cornell University, Ithaca, New York. Grossman, R. B., and Bartelli, L. J. 1957. Soil Sci. SOC.A m . Proc. 21,661-662. Grossman, R. B.,and Cline, M. G . 1957. SoilSci. Soc.Am. Proc. 21,322-325. Grossman, R. 9.. Fehrenbacher, J. B., and Beavers, A. H . 1959a. SoilSci. Soc. A m . Proc. 23,65-70.

Grossman, R. B., Stephen, I., Fehrenbacher, J . B., Beavers, A. H., and Parker, J. M. I959b. Soil Sci. SOC.A m . Proc. 23,70-73. . Sci. Grossman, R. B., Stephen, I., Fehrenbacher. J. B., and Beavers, A. H. 1 9 5 9 ~ Soil Soc.Am. Proc. 23,73-75. Grossman, R. B., Brasher, B. R., Fransmeier, D. P., and Walker, J. L. 1968. Soil Sci. Soc. A m . Proc. 32,570-573. Hearn, W. E. 1924. A m . Soil Surv. Assoc. Bull. 5 , 18. Hill, D. E. 1966. Connecticut A g r . Expt. Sta., N e w Haven, BUN. 678. Horn, M. E., and Rutledge, E. M. 1965. Soil Sci. Soc. A m . Proc. 29,443-448. Howe, F. B., Buckman. H. O., and Lewis, H. G . 1924. “Soil Survey ofTompkins County, New York,” U.S. Dept. Agr., Bur Soils Field Operations 1920. Washington. D.C. Huddleston,J. H., and Olson, G. W. 1967. SoilSci. 104,401-409. Huddleston, J. S. 1967. “Soil Survey o f T a t e County, Mississippi,” U.S. Dept. Agr., Washington, D.C. Hutcheson. T. B., Jr., and Bailey, H. H. 1964. Soil Sci. Soc. A m . Proc. 28,684-685. Hutcheson, T . B., Jr., Lewis, R. J., and Seay, W. A. 1959. SoilSci. SOC.A m . Proc. 23, 474-478.

International Society of Soil Science. 1968. SoilSci. Sac. A m . Proc. 32, 153-154. Jha, P. P. 1961. Ph.D. Thesis, Cornell University, Ithaca, New York.

274

R. B. GROSSMAN A N D F. J. CARLISLE

Jha, P. P., and Cline, M. G. 1963. SoilSci. Soc. Am. Proc. 27,339-344. Knox, E. G. 1954. Ph:D. Thesis, Cornell University, Ithaca, New York. Knox, E. G. 1957. Soil Sci. Soc. Am. Proc. 21,326-330. Kmsekopf, H. H. 1942. SoilSci. Soc. A m . Proc. 7,434-436. Leighty. W. J.. and Wyatt. C. E. 1950. “Soil Survey of Marshall County, Kentucky,” U.S. Dept. Agr., Washington, D.C. Love, T. R., Williams, L. D., Proffitt, W. H., Epley, I. B., and Elder, J. H. 1959. “Soil Survey of Coffee County, Tennessee,” U.S. Dept. Agr., Washington, D.C. Lyford, W. H. 1964. HurvurdForest Paper No. 8. Lyford, W. H.,and MacLean, D. W. 1966. Harvard Forest Paper No. 15. Lyford, W. H., Goodlett, J. C., and Coates, W. H. 1963. Harvard Forest Bull. 30. McCracken, R. J., and Weed, S. B. 1963. Soil Sci. Soc. A m . Proc. 27,330-334. McKeague, J. A., and Cline, M. G. 1963. Can. J . Soil Sci. 43,70-96. McNutt, E. J., Green, T. W., Kahrein, R. B., Galberry, H. S., Thomas, A. E., Tyler, M. C., and Mathews, E. D. 1959. “Soil Survey of De Soto County, Mississippi,” U.S. Dept. Agr., Washington, D.C. Marbu t, C. F. 1928. Proc. 1 s t Intern. Congr. Soil Sci.,1927, Vol. 4, pp. 1-3 I . Marbut, C. F. 1935. I n “Atlas of American Agriculture” (0.E. Baker, ed.), Part 111. U.S. Dept. Agr., Washington, D.C. Marbut, C. F., Bennett, H. H., Lapham, J. E., and Lapham, M. H. 1913. US.Dept. Agr., Bur. Soils Bull. 96. Matthews, E. D., Compy, E. Z. W., and Johnson, J. C. 1961. “Soil Survey of Montgomery County, Maryland,” U.S. Dept. Agr., Washington, D.C. Milfred, C. J., Olson, G. W., and Hole, F. D. 1967. Univ. Wisconsin Bull. 85, Soil Ser. 60. Miller, F. P. 1965. Ph.D. Thesis, Ohio State University, Columbus, Ohio. Miller, R. W. 1967. Soil Sci. Soc. Am. Proc. 31,46-50. Mueller, 0 .P., and Cline, M. G. 1959. Soil Sci. 88,107- 1 I I . Neeley, J. A. 1965. “Soil Survey of Tompkins County, New York,” U.S. Dept. Agr., Washington, D.C. Nettleton, W. D. 1965. Ph.D. Thesis, North Carolina State University, Raleigh, North Carolina. Nettleton, W. D., Daniels, R. B., and McCracken, R. J. 1968a. Soil Sci. S O C . Am. Proc. 32, 577-582. Nettleton, W. D., McCracken, R. J., and Daniels, R. B. 1968b. Soil Sci. Soc. Am. Proc. 32, 582-587. Nikiforoff, C. C. 1955. Geol. Surv. Profess. Paper 267-B. U S .Dept. lnterior, Washington, D.C. Nikiforoff,C. C.,and Alexander, L.T. 1942.SoilSci. 53,157-172. Nikiforoff,C. C., Humbert, R. P.,andCady,J. G. 1948.SoiISci. 65,135-153. Olson, G. W. 1962. Ph.D. Thesis. University of Wisconsin, Madison, Wisconsin. pp. 1 13-125. Soil Sci. SOC.Am. and Am. SOC.Agron., Madison, Wisconsin. Olson, G. W. 1966. In “Soil Surveys and Land Use Planning” (L. J. Bartelli et ul., eds.), pp. 113-125. Soil Sci. SOC. Am. and Am. SOC. Agron., Madison, Wisconsin. Olwn, G. W., and Hole, F. D. 1967-1968. Wisconsin Acad. Sci. Arts Letters 56, 173-184. O’?leal, A. M. 1952. Soil Sci. Soc. Am. Proc. 22,3 12-3 15. Pcarson, C. S., and Cline, M. G. 1960. “Soil Survey of Lewis County, New York,” U.S. Dept. Agr., Washington, D.C. Perkins, S. O., and Bacon, S. R. 1925. “Soil Survey of Prince Georges County, Maryland,” U.S. Dept. Agr., Washington, D.C.

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

275

Pettiet, J. V. 1964. Ph.D. Thesis. Mississippi State University, State College, Mississippi. Porter, H. C., Derting, J . F.. Elder, J . H., Henry, E. F., and Pendleton, R. F. 1963. “Soil Survey of Failfax County. Virginia,” U . S . Dept. Agr., Washington. D.C. Redmond, C. E.. and Engberg. C. A. 1967. “Soil Survey of Arenac County. Michigan,” U.S. Dept. Agr., Washington, D.C. Rutledge. E. M.,and Horn. M. E. IY65.SoilSci. Soc.Anr. Proc. 29,437-443. Scrivner, C. L. 1960. Ph.D. Thesis, University of Missouri, Columbia, Missouri. Shantz, H . L., and Zon. R. 1924. I n “Atlas of American Agriculture” (0.E. Baker, ed.), Part IV. U.S. Dept. Agr., Washington, D.C. Shearin, A. E., and Hill. D. E. 1962. “Soil Survey of Hartford County, Connecticut,” U.S. Dept. Agr., Washington, D.C. Smith, H. C . , and Rose. R. C. 1924. “Soil Survey of Charles County, Maryland,” U.S. Dept. Agr.. Bur. Soils Field Operations 1918. U.S. Govt. Printing Office, Washington, D.C. Smith. R. M., and Browning, D. R. 1946. Soil Sci. 62, 307-3 17. soil Conservation Service. 196Y. In “National Atlas,” (U.S. Dept. Interior. Geol. Surv., Sheet N o . 86. Soil Survey Staff. 195 I . “Soil Survey Manual,” U . S . Dept. Agr. Handbook 18. Washington, D.C. Soil Survey Staff. 1960. “Soil Classication- A Comprehensive System -7th Approximation.” U.S. Dept. Agr., Washington, D.C. Soil Survey Staff. 1962. “Supplement to Soil Survey Manual,” U.S. Dept. Agr. Handbook 18. Washington, D.C. Soil Survey Staff. 1967. “Supplement to Soil Classification System (7th Approximation). U.S. Dept. Agr., Washington, D.C. Soil Survey Staff. 1968a. U S . Dept. Agr., Soil Surv. Invest. R e p f . 19. Sod Survey Staff. I968b. US.Dept. Agr., Soil Surv. Invest. R e p t . 20. Spaeth, J. N., and Diebold, C. H. 1938. Cornell Univ.,A g r . Expt. Sra. Mem. 213. Tavernier, R.. and Smith, G. D. l9S7. Advan. Agron. 9,2 17-289. Taylor, D. C. 1960. “Soil Survey of Erie County, Pennsylvania,” U.S. Dept. Agr., Washington, D.C. Thomas, A. E. 1967. “Soil Survey of Grenada County, Mississippi,” U.S. Dept. Agr., Washington, D.C. Thomas, C. J. 1966. In “Soil Surveys and Land Use Planning” (L. J . Bartelli e t a / . , eds.), pp. 60-75. Soil Sci. SOC.Am. and Am. SOC.Agron., Madison, Wisconsin. Thorp, J., and Smith, G. D. 1949. Soil Sci. 67, I17- 126. Threlkeld, G., and Alfred, S. 1967. “Soil Survey of lonia County, Michigan,” U.S. Dept. Agr., Washington, D.C. Vanderford, H . B., and Shaffer, M. E. 1966. SoilSci. Soc. Am. Proc. 30,494-498. Whittig, L. D., Kilmer, V. J., Roberts, R. C., and Cady, J . G. 1957. Soil Sci. Soc. A m . Proc. 21,226-232. Winters, t.1942. SoilSci. Sor. A m . Proc. 7,437-440. Winters, E.. and Simonson, R. W. I95 I.Advan. Agron. 3, 1-92, Yassoglou. N. J . , and Whiteside. E. P. 1960. Soil Sci. Soc. A m . Proc. 24,396-407.

276

R. B. GROSSMAN A N D F. J. CARLISLE

Appendix

BELTSVILLESOIL The soil has a fragipan in the lower part of an illuvial B horizon and is an example of sequence Ild, Fig. 2. The fragipan shows evidence of appreciable clay accumulation. The description is from the report on soils of Montgomery County, Maryland (Matthews et al., 1961). The information on soil use comes both from this report and the one for Fairfax County, Virginia (Porter et al., 1963). Surface soil AOO- 1 to f.i inch, scattered pine needles. AO- M to 0 inch, loose but felty decomposed leaf mold. A 1-0 to 2 inches, very dark gray (IOYR 3/1) silt loam; moderate, fine, crumb structure: loose; roots abundant; strongly acid; abrupt, wavy boundary. A2-2 to 13 inches, yellow (2.5Y 8/6) silt loam; weak, fine, subangular blocky structure; friable when moist and slightly sticky when wet; roots abundant; strongly acid; gradual, smooth boundary. t Subsoil B I - 13 to 2 1 inches, brownish-yellow (IOYR 6/6), light silty clay loam; few, medium, faint mottles of strong brown (7.5YR S/8);moderate, medium, and coarse, subangular blocky structure; firm when moist; roots fairly abundant; a few rounded pebbles; thin, distinct clayskins on some faces: strongly acid; gradual to clear, wavy boundary. B21-21 to 31 inches, brownish-yellow (IOYR 618) silty clay loam: few, fine, distinct mottles of red (2.5YR 5/6); weak, coarse, platy structure; firm when moist; roots few, between structural elements only: a few pebbles and a few faint clayskins; strongly acid; clear, irregular boundary. B22m-31 to 42 inches, reddish-yellow (7.5YR 6/6) clay loam; many, fine, faint (B22x) mottles of reddish yellow (5YR 6/8) and many vertical channels and horizontal streaks of gray silt and clay; compound structure-moderate, very coarse, platy and moderate, medium to coarse, subangular blocky; very compact and dense; very firm when moist; very few roots; some pebbles and a few clayskins; this is the fragipan, or hardpan; strongly acid: clear, smooth boundary. B3m-42 to 48 inches, reddish-yellow (7.5YR 6/6) clay loam; many, medium, distinct mottles of light yellowish brown (IOYR 6/4), reddish yellow (SYR 6/8), and (B3x) white (2.5Y 8/2): compound structure-moderate, very thick, platy and moderate, coarse, blocky: compact and dense; very firm when moist; practically no roots: some pebbles and some light-gray to white silt coats; strongly acid; this lower hardpan is a transition between the true subsoil and the substratum. Substratum CD-48 to 54 inches t,very pale brown (IOYR 7/3) silty clay loam; abundant, rnediurn, distinct mottles of reddish yellow (7.5YR 6/6) and light gray (IOYR 7/2); massive; very firm; strongly acid: underlain at some depth by gravel. Matthews et a / . (1961) wrote: “Because of the almost impervious fragipan, the Beltsville soils tend to be wet at times. Frequently, they are saturated near the surface, but almost dry within or below the fragipan. The moisture-supplying capacity is moderate. A few depressed areas in the uplands are ponded for short periods after long heavy rains or quick

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

277

thaws.” On the steeper slopes, due partly to accelerated erosion, the fragipans occur at shallow depths; percolation of water is slow and runoff rapid. The soil is often too wet for cultivation in the spring, and during the summer it becomes dry. Permanent pasture and hay are the better suited crops. The soils are not suitable for septic tanks. Seepage on cut slopes causes problems in road construction.

GRENADASOIL The soil has a fragipan in the illuvial horizons of the lower sequum and is an example of sequence Ilf, Fig. 2. Parent material is loess. The description and information on soil use are from the report on soils of Grenada County, Mississippi (A. E. Thomas, 1967). Ap-0 to 5 inches, grayish-brown (IOYR 5 / 2 ) silt loam with common, coarse, faint, dark grayish-brown (IOYR 4/2) mottles; weak, fine and medium, granular structure; friable: common fine roots; few worm casts; few, fine, black concretions; few iron stains; plow pan in the lower part of horizon; neutral; abrupt, smooth boundary. A2-5 to 7 inches, brown (IOYR 5/3)silt loam with few, medium, faint, grayish-brown (IOYR 5 / 2 ) mottles; weak, medium, granular and weak, fine subangular blocky structure; friable: few fine roots; few worm casts: few iron stains; few, fine, black concretions: some material from the Ap horizon in worm channels; slightly acid; abrupt, smooth boundary. B21-7 to 17 inches, strong-brown (7.SYR 5/61 heavy silt loam; moderate, medium, subangular blocky structure: friable; few fine roots; few worm casts; some material from the A2 horizon in old root channels; few fine and medium concretions; medium acid: clear, smooth boundary. B22- 17 to 22 inches, yellowish-brown (IOYR 5/6) heavy silt loam with few, medium, distinct, light brownish-gray ( IOYR 6/2) mottles; moderate, fine and medium, subangular blocky structure: friable: few fine roots: common, fine and medium, black concretions: some material from the A2 horizon in old root channels: strongly acid: clear, wavy boundary. A’2x&Bx-22 to 28 inches, light yellowish-brown (IOYR 6/4) silt loam with many, medium and distinct, coarse, light brownish-gray ( IOYR 6/2) mottles; weak to moderate, fine and medium, angular and subangular blocky structure: friable; many, fine, medium and large, black concretions; few fine roots in cracks: few fine voids and vesicles; strongly acid; clear, wavey boundary. [Consistence of this horizon does not support use of the symbol x to designate fragipan character.] B’2xg-28 to 45 inches, mottled yellowish-brown (IOYR 5 / 6 ) , light-gray (IOYR 7/1), and pale-brown (IOYR 6 / 3 ) heavy silt loam: coarse prismatic structure that breaks to moderate angular and subangular blocky structure; firm and compact, hard when dry; common, fine and medium, black and brown concretions: common black and brown coatings on peds: few fine voids; few patchy clay films: light-gray (IOYR 7/2) silt coatings on peds and in cracks: strongly acid; gradual, wavy boundary. yellowish-brown (IOYR 5 / 6 ) silt loam with many, coarse, B’3x-45 to 60 inches faint, pale-brown (IOYR 6/3) mottles; coarse prismatic structure that breaks to weak, medium and coarse, subangular blocky structure: friable: few, fine, black and brown concretions: few black and brown coatings; medium acid. Thomas ( I 967) wrote in regard to Grenada soils that are moderately eroded: “At a depth

+,

278

R. B. GROSSMAN A N D F. J. CARLISLE

of about 2 feet, these soils have a fragipan that slows the movement of water and restricts the growth of roots. Consequently, the subsoil is waterlogged during rainy periods, especially in winter and early in spring. During dry periods in summer, however, these soils are slightly droughty because the soil above the fragipan is thin.” The influence of the fragipan depends strongly on its depth. On steeper slopes erosion has been greater and the fragipan occurs at shallower depths. Recommendations for various engineering interpretations reflect the presence of the fragipan. About road construction, Thomas (1967) wrote: “In nearly level areas, side ditches should extend below the fragipan and the pavement should be at least 4 feet above the fragipan. In steeper areas, road cuts normally extend below the fragipan, but adequate underdrainage is needed where the construction changes from cut to fill. This underdrainage can be provided by excavating the fragipan and replacing it with a more permeable material.” MCBRIDESOIL The soil has a fragipan restricted to the lower eluvial horizon and is an example of sequence Ilh, Fig. 2. Parent material is glacial till. The description and comments are from the report on soils of lonia County, Michigan (Threlkeld and Alfred, 1967). Ap-0 to 8 inches, dark grayish-brown (10YR 4/2) sandy loam; weak, fine, granular structure: friable: moderately high organic-matter content; medium acid; abrupt, smooth boundary. Bir-8 to 12 inches, yellowish-brown (IOYR 5/4-5/6) sandy loam: weak, fine, granular structure; very friable: medium acid; clear, wavy boundary. A’21- 12 to 14 inches, very pale-brown (IOYR 7/4) light sandy loam: weak, thin, platy structure; slightly compact when moist; medium acid; abrupt, wavy boundary. A’22x- 14 to 17 inches, light-gray (IOYR 7/2) light sandy loam; moderately compact, brittle fragipan; weakly vesicular: medium acid; clear, wavy boundary. B’21t- 17 to 30 inches, strong-brown (7.5YR 5/6) sandy clay loam: light-gray (IOYR 7/1) coats on cleavage faces in the upper part of the horizon: moderate, coarse, subangular blocky structure: firm: very strongly acid; clear, wavy boundary. B‘22t-30 to 38 inches, strong-brown (7.5YR 5/6) heavy sandy loam; weak, coarse, subangular blocky structure: friable: very strongly acid; clear, wavy boundary. B3 -38 to 48 inches, strong-brown (7.5YR 5/6) sandy loam: weak, coarse, subangular blocky structure; friable: medium acid in upper part, grading to slightly acid in lower part; abrupt, irregular boundary. C-48 inches +, brown (7.5YR 5/4) sandy loam: massive: friable; calcareous. The soil is considered to have moderate or moderately slow permeability and moderately low available water. Threlkeld and Alfred ( 1967) wrote: “Permeability depends to a great extent on the thickness and development of the fragipan . . . .” The fragipan in this soil, probably because of its thinness, apparently has only a marginal influence on use and management. VOLUSIASOIL The soil has a fragipan in the B horizon directly beneath the upper eluvial horizon and is an example of sequence Ila, Fig. 2. Other soils in this series have a thin B horizon above the fragipan. Parent material is glacial till. The fragipan shows insufficient evidence of clay accumulation for recognition of an argillic horizon. The description and information on soil use are from the report on soils of Tomkins County, New York (Neeley, 1965).

FRAGIPAN SOILS OF THE EASTERN UNITED STATES

279

Ap-0 to 8 inches, dark grayish-brown (IOYR 4/2) channery silt loam: weak, fine, crumb structure; friable; p H 4.8; many fine roots; few flagstones on surface; abrupt, smooth boundary; 6 to 9 inches thick. A2-8 to 14 inches, light olive-brown (2.5Y 5/4) and grayish-brown (2.5 Y S/2) channery silt loam; many, coarse, distinct, yellowish-brown ( IOY R S/6-5/8) mottles: weak, thin to medium, platy structure; friable to slightly firm: pH 5.0; few fine roots: abrupt, wavy boundary; 4 to 6 inches thick. B’xlg- 14 to 28 inches, olive-brown (2.5Y 4/41 channery silt loam; few, medium, faint, gray and brown mottles; weak, coarse prisms 10 to 20 inches across, coated with grayish-brown (2.SY 5/2) and olive-brown (2.5Y 4/4) silty material; prisms break into moderate, medium, subangular blocks when disturbed; very firm: pH 5.4: diffuse boundary; 10 to IS inches thick. - B’x2g-28 to 48 inches, olive-brown (2.5Y 4/4) and olive (5Y 5/3)channery silt loam; few, common, gray and brown mottles; weak, coarse prisms I % to 2 feet across coated with thin, grayish-brown (2.5Y 512) to gray (5Y5 / l ) silt: prisms break into moderate, medium and coarse, subangular blocks; very firm: pH 5.8 at 48 inches: diffuse boundary; 20 to 30 inches thick. C-48 inches +, olive (5Y 5/3) channery silt loam; moderate, medium blocks with gray (5Y 511) surfaces; firm; pH 5.8 to 6.2; vertical streaks border prisms; boundaries disappear below 48 inches. Neeley (1965) wrote: “The pan effectively stops downward movement of water. In sloping areas water seeps downslope through permeable layers above the fragipan . . . .” Wetness is the dominant limitation. The soils are best suited for forage crops that can tolerate wetness. Surface drainage is necessary for successful production of cultivated crops in most years. Trafficability is low for extended periods. Cut banks are subject to seep and slump. The soils are good sites for ponds. The fragipan is an important factor in road construction and maintenance.

This Page Intentionally Left Blank

The Quantitative Relationships between Plant Population and Crop Yield

R. W. Willey*and S. B. Heath University of Reading, Reading, Berkshire, England

Introduction ...................................................... 11. Relationships between Plant Density and Crop Yield A. Biological Relationships ................................... B. Yield/Density Equations ................................................... C . A Further Examination o ciprocal Equations 11 I. The Relationship between Plant Rectangularity and Crop ............... IV. The Variation in Yield of the Individual Plant ............ V . Conclusions ................................................ References .........................................................................................

I.

I.

Prrge 281

286 30 1 3 14 317 319 320

Introduction

I t may often be desirable for the agronomist to define the relationships between plant population and crop yield quantitatively. Probably the simplest reason for wishing to do this is to evaluate such characteristics as optimum population and maximum yield. This can be a useful end in itself, and it can also facilitate comparisom between different cropping situations. The latter aspect is particularly useful when the factors being examined interact with population, for comparisons can clearly be misleading if they are not made between comparable points on the population response curve. Putter et al. ( 1966) have pointed out that even comparisons between calculated parameters which have little or no biological meaning can be of value, for these may still help to pinpoint essential differences. It is more usual, however, that the agronomist wishes to define the relationships between plant population and crop yield so that in any future situation he can predict yield/population curves easily and accurately from the minimum of data. For this purpose at.least, it is desirable that an equation defining these relationships should not be *Presenr address: Makerere University College, Kampala, Uganda.

28 I

282

R. W. WILLEY AND S. 8 . HEATH

merely a mathematical empiricism but should be so far as possible an accurate description of the biological processes of competition that are involved. Some biological validity of this nature clearly provides greater justification for the use of an equation and at the same time probably ensures a wider applicability. However, any attempt to derive such an equation inevitably involves some degree of compromise, for the very nature of the problem is to describe extremely complex biological processes by a mathematical function sufficiently simple to be of use to an agronomist. The object of this review is to examine some of these equations and to discuss how far they fulfill the agronomist’s needs. But first some general points must be discussed. It is important to realize that plant population should be defined not only in terms of the number of plants per unit area (i.e., plan? density) but also in terms of the arrangement of these plants on the ground (spatial arrangement or plant rectangularity). Few workers have distinguished between these two factors and many plant population experiments have in effect studied the combined effects of both; this is certainly the case where different populations have been studied at a constant row width. Since most workers seem to have assumed that the only effect of population is one of density, it would seem logical to try to define the relationship between crop yield and density first, and to incorporate the effects of rectangularity subsequently. For this reason, the rectangularity factor is ignored when the equations in-Section 11 are discussed, even though this factor has not been eliminated in many of the experiments that will be quoted. A possible means of incorporating the rectangularity factor into an equation is discussed in Section 111. Crop yield itself may also require further definition, for in some crops the grower is as much concerned with the yield of the individual plant as with yield per unit area. Many population experiments have examined yield per plant, but it has almost invariably been in terms of mean yield per plant. Yet the degree of variation in yield of the individual plants is a factor that can determine the total yield of plants within a given size grading. Some studies in which an attempt has been made to estimate this variation are briefly discussed in Section IV. Finally, it is pertinent to point out that it is not always easy to decide upon the correct unit of plant population. This was emphasized by Holliday ( 1 960b), who suggested that it should be the basic independent plant unit, whether this was a tiller in a grass crop or perhaps an individual stem in a potato crop. Accurate identification of these population units could be of particular importance in quantitative studies, for without this it may be even more difficult to produce reliable and meaningful equations applicable to a wide range of crops.

PLANT POPULATION A N D CROP YIELD

It.

283

Relationships between Plant Density and Crop Yield

BIOLOGICALRELATIONSHIPS Before examining the different yield/density equations, it is first necessary to decide what are the basic biological relationships that these equations are attempting to describe. The only real attempt to classify these relationships seems to have been that of Holliday ( 1 960b). He suggested that there were essentially two relationships: an asymptotic one where, with increase in density, yield rises to a maximum and is then relatively constant at high densities: and apavabolic one where yield rises to a maximum but then declines at high densities. For the purposes of this present review this suggested classification is a useful one to adopt, but no exact mathematical description is inferred by the terms asymptotic and parabolic. It may well be argued, of course, that these two suggested relationships are merely different degrees of expression of a single relationship. However, in the present context the important fact to realize is that the two situations exist, for it may often happen that a given yield/density equation can describe either an asymptotic situation or a parabolic situation, but not both. Also, since the mathematical description of these two situations can be quite different, it is mathematically convenient to separate them. I t is not within the scope of this review to discuss in detail which specific crops or types of yield may conform to the different biological relationships; nor is it possible to consider how these relationships may be affected by the level of supply of different growth factors. At the same time it may be of use to give an indication of some of the cropping situations in which the different relationships can occur and to illustrate the shapes of yield/density curves which need description.* A.

1. The Asymptotic Relationship Holliday ( I 960b) suggested that total crop dry matter conformed to this relationship, but more recently several workers (de Wit, 1959; Bleasdale, 1966a; Bruinsma, 1966; Campbell and Viets, 1967; Farazdaghi, 1968) have shown that at high densities decreases in this form of yield can occur. Despite these exceptions it is probably reasonable to assume that, for practical purposes, total dry matter yield often conforms to an essentially asymptotic relationship. This situation is illustrated in *The notation used to express the relationships is as follows: y = yield per unit area; p = plant density; s = space available per plant (s = l/p): d , = distance between plants within a row (intrarow spacing): d2 = distance between rows (interrow spacing); w = yield per plant; W = maximum yield attainable by a plant; w, = yield o f a plant part. Any variable or constant with the subscript p . as in wp above, refers to a plant part as opposed to a total plant.

284

R. W. WILLEY A N D S. B. HEATH

Figs. IA and I B by some data for fodder rape (Holliday, 1960a) and for Wimmera ryegrass and subterranean clover (Donald, 195 l), all of which are asymptotic to particularly high densities. Holliday (1 960b) also suggested that those forms of yield which constituted a vegetative part of the crop conformed to an asymptotic relationship. Notable exceptions may occur (see Section II,A,2) but again it is reasonable to assume that such forms of yield often are asymptotic. This situation is illustrated by some data for potato tubers (Saunt, 1960) and root yield of long beet (Warne, 1951) in Figs. 1C and 1 D, respectively.

2 . The Parabolic Relationship Holliday (1960b) suggested that reproductive forms of yield (i.e., grains and seeds) conformed to a parabolic relationship, and the examples

;lk2oF P

P

Total

-7%:

Y 10

Y

OO 5

1

2

3

OO

P

2

4

6

P

FIG. 1. Examples of the asymptotic yield/density relationship. (A) Total dry matter of Essex Giant rape; y = tons/acre, p = 106plants/acre (Holliday, 1960a). (B) Total dry matter of Wimmera ryegrass and subterranean clover; y = g./sq. lk., p = lo2plants/sq. Ik. (Donald, 195 1). ( C ) Fresh weight of potato tubers; y = tonslacre, p = lo4 parent tuberslacre (Saunt, 1960). (D) Fresh weight yields of long beet; y = poundslplot, p = plantslfoot of row (Warne, 195 1).

PLANT POPULATION A N D CROP YIELD

285

given in Figs. 2A and 2 B for grain yield of maize (Lang et al., 1956) and barley (Willey, 1965) certainly indicate that this can be so. The maize data are of particular interest because this crop usually displays a very distinct decline in yield at high densities, and as such it represents one of the more extreme forms of the parabolic relationship. The barley data, in which the density reaches a particularly high value, are also of interest because they illustrate a point seldom evident in experimental data,

C 30.

P

P

FIG.2. Examples of the parabolic yieldldensity relationships. (A) Mean grain yield of maize for all hybrids, grown at a low level -( ), medium level (----), and a high level ( - - -) of nitrogen: y = bushelslacre, p = lo3 plantslacre (Lang cf al., 1956). (B) Grain yield of barley grown with 0 -( ), 30 (----), and 60 ( - - -) units of nitrogen: y = cwtlacre, p = lo6 plantslacre (Willey, 1965). (C) Root dry weight of globe red beet; y = 10’ kg./acre, p = lo4 plantslacre (unpublished Reading data). (D) Parsnips var. AVONRESISTOR, total fresh weight yield ( - - -), graded yield > 1.5 inches in diameter (----), graded yield > 2.0 inches in diameter (-): y = tonslacre, p = plantslsq. ft. (Bleasdale and Thompson, 1966).

286

R. W. WILLEY AND S. B. HEATH

namely that the parabolic relationship must at some stage begin to flatten off along the density axis. As mentioned in Section II,A, 1, certain forms of vegetative yield may also be parabolic. A notable instance of this seems to be the root yield of globe red beet, and some example data for this crop are given in Fig. 2C. Yet a further situation can exist where yield is parabolic, and this is where yield constitutes only those plants, or parts of plants, that fall within certain size limits, i.e., where some form of “grading” is practiced. Figure 2D illustrates this situation with some parsnip data of Bleasdale and Thompson ( I 966). It can be seen that in this particular instance total yield of roots is asymptotic, but grading produces a parabolic relationship that becomes more acute as the severity of grading is increased. This situation is of considerable importance in many crops. However, it must be emphasized that “graded” yield cannot be regarded as a biological form of yield in the same sense as those forms discussed above. For this reason, the description of this particular relationship may have to remain more empirical than that of other relationships.

B. YIELD/DENSITY EQUATIONS Section II,A indicated the general form of the biological relationships that exist between crop yield and plant density. The object of this section is to describe the different mathematical equations that have been proposed to define these relationships. Some of these yield/density equations propose a relatively simple mathematical relationship directly between yield per unit area and density, but the majority propose a basic relationship between mean yield per plant and density. The general shape of this latter relationship is illustrated in Fig. 3 for both the asymptotic and parabolic yield/density situations.

I. Polymoniul Equations One of the simplest approaches to the description of yield/density relationships has been the use of two polymonial equations applied directly to the relationship between yield per unit area and density. These have been used largely as a convenient means of smoothing experimental data: they have not been seriously proposed as general yield/density equations, and little or no biological validity has been claimed for them. In these respects they are not of any major importance in the present review, but a brief description of their scope and limitations serves as a useful introduction to the use of yield/density equations, particularly where biological validity is lacking.

287

PLANT POPULATION A N D CROP YIELD

-0.8 0.8

- 0.6 0.6

100

- 0.4Y

W

W

Y

0.4

- 0.2

50

0.2

.-0

0

4

12

P

P

20

FIG.3. The relationship between yield per plant (w)and plant population ( p ) in an asymptotic (A) and a parabolic (B)yield/density situation. (A) Total dry matter of Essex tonlplant, p = loo plantslacre (HolliGiant rape, 1952 experiment; y = tonslacre, w = day, 1960a). (B) Grain yield of maize hybrid WF9 x 38-1 I at medium N ; y = bushelslacre, w= bushel/plant, p = lo3plantslacre (Lang ef a/., 1956).

Hudson ( 1 94 1) attempted to describe the relationship between grain yield and seed rate of winter wheat with a simple quadratic expression: y

=a

+ b p + cp’

(1)

where a, b, and c are constants, c being negative. The general shape of the yield/density curve described by Eq. ( I ) is illustrated in Fig. 4, where

20.

\\ \ I \ \

FIG.4. The quadratic equation (Eq. 1) (-----) and the square root equation (Eq. 2) ) fitted to grain yield of maize hybrid HY2 X OH7 at low N ; y = bushelslacre, p = l o Tplantslacre (Lang et a/., 1956).

-(

288

R. W. WILLEY AND S. B. HEATH

it is fitted to some maize data of Lang et al. ( 1 956); it is essentially a curve which is symmetrical about a maximum value of yield. Although the degree of curvature may obviously vary, this basic shape offers little flexibility in fitting yield/density relationships. It is clearly not suitable for fitting a truly asymptotic situation, and in a parabolic situation it is likely to give a good fit only where the yield/density curve is reasonably symmetrical. But even in this latter situation, the accuracy of this equation is probably restricted to a relatively narrow range of densities around the point of maximum yield. This is because of the unrealistic implications of the equation at both high and low densities. A t high densities it implies that yield must drop sharply down to zero (see extrapolation in Fig. 4), whereas at the other extreme it implies that a t zero density yield has a value, a (which in practice may turn out to be either positive o r negative). The former implication is a serious limitation on the use of this equation at high densities. The latter implication could be only a minor disadvantage if the value of u was low; in any case, if an accurate fit at low densities was particularly desirable, the omission of a from the equation would ensure that the curve passed through the origin. The disadvantage of the symmetrical nature of the quadratic curve was avoided by Sharpe and Dent (1968) by using a square root form of pol ymonial Eq. 2):

where a, b and c are constants, b being negative. This equation again gives rise to a curve where yield rises to a maximum value and then decreases at higher densities, so it still cannot describe an asymptotic situation. Compared with the quadratic, however, it can follow a slightly more gradual decline in yield at high densities, although this is accompanied by a rather steeper increase at the low densities (see Fig. 4). It still implies that a t zero density yield has a finite value a, and that at the other end of the scale yield declines to zero, although admittedly at a rather higher density than with the quadratic. The apparent lack of any biological validity must also impose limits on the use of these two equations. For example, it would seem unwise to use them where data were not sufficiently comprehensive to give a good initial indication of the general shape of any particular yield/density situation. Also there would seem little justification for using them to extrapolate data. Such extrapolation was carried out by Keller and Li ( 1949), who used the quadratic to estimate optimum density and maximum yield of some hop data, and it is of significance that when Wilcox ( 1 950), with

289

PLANT POPULATION A N D CROP YIELD

little more justification, extrapolated the same data using the Mitscherlich equation he obtained substantially different values.

2. Exponential Equations Duncan ( I 958), when reviewing experimental data on maize, proposed an exponential equation to describe the relationship between grain yield and density. He derived this by fitting a linear regression of the logarithm of yield per plant on density. The basic relationship was therefore: log w

= log

+ bp

K

(3)

or y = p K 10bp

where K is a constant and 6, negative, is the slope of the regression line (see Fig. 5A). Carmer and Jackobs (1965) used this equation in a slightly different but analogous form:

where A and K are constants. The yieldldensity curve which this type of equation produces is comparable to the polynomials in as much as yield must rise to a maximum value and then decrease at higher densities. It can give a good fit to parabolic yield/density data, but even though it is much more flexible than the polynomials at high densities, it still cannot

100

Y

50

0

P

P

FIG. 5 . The exponential equation (Eq. 3) of Duncan (1958) fitted to a parabolic (A) and an asymptotic (B) yield/density relationship. (A) The regression line of log w against p, and the fitted yield/density curve for grain yield of maize, mean of all hybrids at medium N ; y = bushelslacre, p = lo3 plantslacre, w = bushel (Lang et al., 1956). (B) The fitted yield/density curve for total dry matter of Essex Giant rape, 1952 data; y = tonslacre, p = lofiplantslacre (Holliday, 1960a).

2 90

R. W. WILLEY A N D S. B. HEATH

give a useful practical fit to data that are asymptotic. This is illustrated in Fig. 5 , where it is fitted to some parabolic maize data of Lang et al. ( 1956) and some asymptotic rape data of Holliday ( 1960a). Apart from greater flexibility, this exponential equation has further advantages over the polynomials. At high densities the yield curve does not cut the density axis but, more realistically, only gradually approaches it. Also, this curve now passes through the origin. However, as pointed out by Duncan there may still be a defect at low densities for, as estrapolation of the regression line in Fig. 5A indicates, the equation cannot allow for a leveling off in yield per plant at densities too low for competition to occur. But this is a common defect of yield/density equations, and it is discussed later when considering Holliday’s reciprocal equations (Section 11, B, 5 , b). Duncan also pointed out that, since his equation was based on a linear regression, it was possible to construct the whole yield/density curve from the yields at only two densities. He therefore suggested that in the maize crop the examination of factors that interacted with density might usefully be carried out at two densities; the use of his equation would then allow comparison of the factors at their calculated points of optimum density and maximum yield. This technique can, of course, be used with any yield/density equation derived from some linear regression on density, and its practical potential makes it of considerable interest. Its application calls for some caution, however, for a prerequisite for its use must be a reasonable assurance that the equation used is an accurate description of the particular yield/density relationship that is under study. Duncan’s justification for suggesting its use in the maize crop was the fact that his equation gave a good practical fit to the data he reviewed. This seems reasonable, but in general a better justification would seem to be the knowledge that an equation used in this way had a good deal of biological validity and was not just an empirical one. This could be particularly important, because it was pointed out by Duncan that the farther apart the two densities, the more accurately the regression line would be determined. While this may be mathematically sound, it would seem safer in practice to include a third intermediate density so that the point of calculated maximum yield is not too far from an experimental treatment. 3 . Mitscherlich Equation Mitscherlich proposed a law of physiological relations by which he described the relationship between the yield of a plant and the supply of an essential growth factor, all other factors being held constant. He assumed that as the supply of such a factor increased, yield per plant

PLANT POPULATION A N D CROP YIELD

29 1

would approach a maximum value, and at any given point the response would depend on how far the plant yield was below this maximum. This can be expressed: - --

dw df

(W-w)c

where f is the level of supply of the factor and c is a constant. On integration this gives Eq. (4):

Mitscherlich termed c his “Wirkungsfactor” and claimed that it was constant for a given growth factor and independent of other conditions. Later, Mitscherlich ( I9 19) suggested that his equation might be applied more generally to the relationship between “space” and plant growth and so serve as a yield/density equation. Thus, substituting space, s, for the growth factor,f, Eq. (4) can be rewritten:

where K is now a general “space” constant or factor. It is evident from the basic assumption about the nature of the plant’s response that this yield/density equation describes an asymptotic situation, but not a parabolic one. Kira et af. ( 1954) examined the constancy of the space factor K . Using the yield/density data of Donald ( I95 I ) for subterranean clover, they were able to define the asymptotic value of yield per plant, W. From this value, and from mean yields per plant at the other densities, they calculated a range of K values (Table I). I t is apparent that the values decreased with increase in the space available per plant and could not be regarded as constant. Kira et a f . ( I 954) obtained similar changes in K values from yield/density data for azuki bean (Phaseofuschrysanrhus), although the trend was not so clear. This change in the value of K could have interesting agronomic implications, for it may perhaps suggest that a change in density may not only change the space available to a plant, but might also bring about some change in the environment-for example, an effect on rooting depth. However, as far as the practical use of the Mitscherlich equation is concerned, a change in K is clearly undesirable, and the value of this expression as a yield/density equation becomes questionable. Kira et al. ( 1954)

292

R. W . WILLEY AND S. B. HEATH

TABLE I APPLIED TO THE MITSCHERLICH’S FORMULA

RESULTSOF AN EXPERIMENT DONALD (195 I)a

WITH SUBTERRANEAN CLOVER OF

61 days from sowing

Density (plants/sq. link) 0.25

1 .oo

5.95-5.93 15.9-1 6. I3 60.6-62.6 241-248 1247-1393

Dry weight per plant (g.1

K

15.6 15.5 15.6 15.8 (W) 14.2 0.154 13.9 0.563 10.6 1.66

I 3 I days from sowing

Dry weight per plant (g.1

182 days from sowing

K

528 562 (W) 386 0.0073 0.0178 364 0.02 13 153 0.0370 13 16 0.0430

Dry weight per plant (g.1

K

34,080 (W) 0.00020 2 1,280 0.0006 1 4,560 0.0009 1 2,020 0.00 106 0.001 17 600 160 0.00125 29 0.00123

“After Kira et al. (1954).

did in fact point out that they could stabilize K by arbitrarily reducing the value of W, but in this event the equation must lose much of its biological foundation. Despite these criticisms of the Mitscherlich equation, the basic concept of an asymptotic yield per plant is of considerable interest. This at least provides a satisfactory theoretical description of the yield/density curve at very low densities where there is no competition. As several workers have pointed out (Duncan, 1958; Kira et al., 1954; Shinozaki and Kira, 1956; Holliday, 1960a), yield/density curves are usually unable to provide such a description and their validity at low densities is doubtful. It is also of interest that Goodall ( 1 960), examining some mangold data, and Nelder (1963), commenting on some lucerne data of Jarvis (1962), both found that the Mitscherlich equation could give as good a fit as other equations. On the other hand, as would be expected from the results of the examination of his K values by Kira et al. (1954), Donald (195 1) did not obtain a good fit to his data using the Mitscherlich equation.

4 . Geometric Equations Geometric equations were put forward by Warne ( 1 95 1 ) and Kira et al. ( 1953) to describe certain yield/density relationships; the latter workers used the term “power” equation. Essentially this type of equation assumes a linear relationship between the logarithm of yield per plant and the logarithm of density.

PLANT POPULATION A N D CROP YIELD

293

Warne (195 I ) was studying the effect of density on the yield of root vegetables (beet, parsnips, and carrots), and he proposed a linear relationship between the logarithm of root yield per plant and the logarithm of distance between plants in the row where row width was constant. Since the row width was constant Warne’s equation can be written in the form log w

= log A

+ b log ( s )

or w = A (S)!’

where A and B are constants and s is the space available per plant. On a yield per unit area basis, and including density rather than space, Warne’s equation becomes y = A (p)’-”

Kira et a f . ( 1953) obtained a linear relationship between the logarithm of total yield per plant and the logarithm of density in a soybean experiment. The form of equation they proposed was log w

+ a log p = log K

or wp“ =K

(7)

where a and K are constants, a being termed the competition-density index. This equation is exactly analogous to that of Warne-the a and K of Kira el al. being comparable with Warne’s h and A , respectively. Strictly speaking, the only type of yield/density curve which this equation can describe is one where yield is still rising at the highest density. Such curves are illustrated in Fig. 6, where Kira et af.’sequation is fitted to some of Donald’s data for different harvests of subterranean clover (Donald, 195 I). I t can be seen that as the yield/density curve approaches an asymptotic shape with the passage of time (Fig. 6B), the slope of the regression line becomes steeper and the value of a (the competitiondensity index) increases and approaches a value of 1 (Fig. 6A). However, if an asymptotic shape is reached, then, to describe constant yield at the high densities accurately, the competition-density index has to take the value of 1, and this then implies that yield is constant at all densities, i.e., the yield/density curve becomes a straight horizontal line (with value K ) . Or, from Eq. (7): w p = K constant = Y

(8)

2 94

R. W . WILLEY AND S. B. HEATH

[The Japanese workers referred to Eq. (8) as the law of constant final yield (Hozumi et al., 1956)l. It is also of interest that a competitiondensity index greater than 1 implies that yield decreases with all in-

P

P

FIG.6. The geometric (“power”) equation (Eq. 7) of Kira ef al. (1953) fitted to the total dry matter yield/density data of subterranean clover (Donald, 195 1 ) at different numbers of days after sowing (0, 6 I , 13 1 , 182); (A) The regression lines of w against p ; dashed lines indicate densities at which there is no competition; w = g., p = plantslsq. Ik. (B) The fitted yield per area/density curves; y = g./sq. Ik.

creases in density. Thus, on theoretical grounds neither the truly asymptotic nor the parabolic yield/density situation can be described. In some circumstances it may be possible to obtain a reasonably satisfactory practical fit to the former situation with a value of a fractionally less than I , although this was not the case with Donald’s data in Fig. 6. Both Warne (1951) and Kira et al. (1953) emphasized the possible significance of their respective power constants b and a. in Eqs. (6) and (7). Warne said that the higher the value of the constant, the more the plant was dependent on the space available to it; whereas Kira el al. (1953) interpreted an increase in value of the constant as indicating a more thorough utilization of the space available to the plant. From the agronomist’s viewpoint, the significance of these constants is most easily appreciated by considering the succession of yield/density curves already referred to in Fig. 6. It can be seen that the greater the value of a the fkrther the yield/density curve has progressed from its initial competition-free situation. This progression is associated with a greater degree of competition, a greater degree of curvature in the relationship and, as Kira et al. said, a more efficient utilization of space. On this basis the

PLANT POPULATION A N D CROP YIELD

295

agronomist can readily understand and compare relative differences in the values of these power constants that may have arisen due to differences in time, growing conditions, or plant species. A further criticism of the geometric equation, first raised by the Japanese workers themselves (Shinozaki and Kira, 1956), is the failure of the equation to describe the leveling off in yield per plant at densities too low for competition to occur (indicated by the broken horizontal lines in Fig. 6A). This was a criticism of the exponential equations (Section 11, B, 2), and it also applies to the reciprocal equations; its importance is discussed in Section 11, B, 5 , b.

5 . Reciprocal Equations The reciprocal equations are here regarded as those based on some mathematical relationship between the reciprocal of mean yield per plant and density. These represent a very important group of equations, which at present seems to offer the best possibilities of being able to describe yield/density relationships accurately and meaningfully. Since these equations are to be discussed more fully in Section I1 C, the present section serves only to describe their general characteristics. a. Shinozaki and Kira. The first to propose a reciprocal equation were the Japanese workers Shinozaki and Kira (1 956). This equation was derived from a simple logistic growth curve and the law of constant final yield (see Section 11, C, 1, a). It was proposed by the Japanese workers because they had found that their geometric equation could not satisfactorily fit an asymptotic yield/density situation (Section 11, B, 4). The form of equation derived was

'

- _-a+bbp W

(9)

where a and b are constants. This relationship assumes a linear relationship between the reciprocal of yield per plant and density. Shinozaki and Kira (1956) tested this relationship in a number of asymptotic yield/ density situations and found that it appeared to hold true in practice. However, this form of reciprocal equation cannot describe a parabolic yield/density situation. b. Holliday. Holliday (1960b) was the first worker to propose that both an asymptotic and a parabolic form of yield/density relationship existed. Although it was pointed out earlier that his suggestion of the asymptotic form applying to total or vegetative yield and the parabolic form applying to reproductive yield may not be entirely valid, the two

296

R. W . WILLEY AND S. B. HEATH

equations which he proposed to describe the two forms of relationship are still of considerable interest. Holliday ( 1960a) arrived at his asymptotic equation from yield/density studies on rape, kale, potatoes, and perennial ryegrass in which he observed that the reciprocal of yield per plant was linear with density:

- ‘_- a + b p W

where a and b are constants. Thus, this basic form of Holliday’s asymptotic equation is identical to that of Shinozaki and Kira (1956), and it is of some interest that Holliday produced his largely from experimental observations whereas the Japanese workers derived theirs largely from mathematical considerations. Like other workers, Holliday (1960a) pointed out that his equation could not allow for a constant yield per plant at those densities where there is no competition. He therefore termed the implied yield per plant at zero density the “apparent maximum” yield per plant and designated this A , which is equal to Ila (see Fig. 7). He also suggested that a modified form of equation beginning at the density where competition first starts might be more accurate. Thus, if this density is n, and p-n = m (see Fig. 7), Holliday’s modification can in effect be written 1

--a’+bm

W

The reciprocal of a’ would now equal the true maximum yield per plant (A’). However, Holliday admitted that, although more accurate biologically, this equation might be of limited use in practice because in any given situation A ’ would have to be determined experimentally. Shinozaki and Kira ( 1 956) suggested a rather different way of allowing for a constant yield per plant at densities free of competition. They proposed the inclusion of a factor 6 in their basic relationship, so that

- l_- a + b ( p + G ) W

They suggested that an appropriate value of 6 should be chosen so that at high densities 6 would be negligible compared with p and could therefore be disregarded, whereas at low densities 6 would come into effect and

297

PLANT POPULATION A N D CROP YIELD

yield per plant would level off to a relatively constant and more realistic value. However, this factor 6 seems to have no biological meaning and is included purely to try to improve the equation’s goodness of fit at low densities. Moreover, since its value must depend on the maximum yield per plant without competition, then, like Holliday’s A ’ , it presumably could be satisfactorily determined only by experimental means in any given situation.

A’

+n+-

m

b

4

P

b

Plant d e n s i t y

FIG. 7. Diagram of yield per plant ( w ) and reciprocal of yield per plant ( I / w ) plotted against density. illustrating the derivation of the asymptotic equation (Eq. 10) of Holliday (1960a) and its modification to allow for the absence of competiton at low densities (Eq. 1 I ) ; n is the density at which competiton starts, A is the “apparent maximum” yield per plant, A ’ is the true maximum yield per plant.

Holliday ( 1960b) proposed that the parabolic yield/density situation, where the relationship between the reciprocal of yield per plant and density is no longer linear, could be described by the quadratic expression 1

-=a W

+ b p + cpz

(12)

where a, 6, and c are constants. Holliday pointed out that the use of the quadratic was empirical, but this form of quadratic equation is a great improvement on the simple quadratic which Hudson ( I94 I ) applied directly to yield per area against density (Eq. 1). I t produces a very flex-

298

R. W . WILLEY AND S. B. HEATH

ible parabolic yield/density curve which is not symmetrical about its point of maximum yield and which flattens off realistically at high densities. Some examples of this equation fitted to yield/density curves are given in Section 11, C, 2. c. De Wit. From studies of mixtures of barley and oats, de Wit and Ennik (1 958) derived a single species yield/density equation based on a linear relationship between the reciprocal of yield per unit area and row width (where distance between plants in the row was constant). This took the form 1 Y

-=

a

+ bdz

(13)

where a and b are constants, and d z is the row width. Expressed in terms of the reciprocal of yield per plant this equation becomes: -' -b W

+ap

which is exactly analogous to the reciprocal equation (Eq. 9) of Shinozaki and Kira (1956) and the asymptotic equation (Eq. 10)of Holliday (1 960a). Later, de Wit ( 1 960) proposed a slightly different version of this equation based on a linear relationship between the reciprocal of yield per unit area and space available per plant. This is represented diagrammatically in Fig. 8, and the basic relationship is:

or

where s is the space available per plant and 1/P and Q are the points where the regression line cuts the I/y and s axes, respectively. P is therefor equal to the asymptote of yield per area. De Wit's later equation is again more readily compared with the other reciprocal equations if it is transformed to show the relationship between the reciprocal of yield per plant and density as follows: 1

1 PQ

-= - + - p

w

1 P

PLANT POPULATION A N D CROP YIELD

299

It can be seen that this equation differs slightly from those of Shinozaki and Kira ( 1956) (Eq. 9) and Holliday (1 960a) (Eq. 101, in that the value of I/w at zero density IIPQ is now defined by two constants instead of one, and one of these constants, P , is the asymptote of yield per area. The possible significance of this is discussed in Section 11, C , I , b.

Space per plant

(5)

-

FIG.8. Diagram of the reciprocal of yield per area (I/y) plotted against space available per plant (s) to illustrate the derivation of the reciprocal equation of de Wit (1960). 1/P and Q are the points where the regression line cuts the I/y and s axes, respectively.

Like Shinozaki and Kira’s equation, de Wit’s can describe only an asymptotic yield/density situation; de Wit does not appear to have proposed any modification of his equation to describe a parabolic situation. d. Bleasdale and Nelder. Bleasdale and Nelder (1960) proposed a reciprocal equation which they derived from a generalization of the logistic growth curve described by Richards (1959). This was originally proposed in the following form: 1 -e a + bp’ W

where a, 6, and 0 are constants. This equation describes an asymptotic yield/density situation, but Bleasdale and Nelder pointed out that if the power on p exceeded the power on w the equation could also describe a parabolic situation. Equation 16 was therefore restated as 1 -

@ =

W

where 4 is a constant.

a

+ bp*

300

R. W. WILLEY A N D S. B. HEATH

This is the form of equation found in their subsequent references (Bleasdale, 1966b; Bleasdale and Thompson, 1966; Bleasdale, 1967); where yield is asymptotic, 0 = 4 ; and where it is parabolic, 8 < 4. In the last situation the yield curve flattens off along the density axis similarly to Holliday’s parabolic equation (Eq. 12). The use of Eq. (1 7) was advocated by Bleasdale and Nelder (1960) because they disagreed with the division by Holliday ( I 960b) of yield/density relationships into vegetative and reproductive types. They preferred the use of a single generalized yield/density equation for all situations. However Bleasdale (1 966b) and Bleasdale and Thompson ( 1966) have stated that although there are theoretical reasons for allowing 4 to have a value other than unity, it is the ratio of 8 to 4 which is important. Also, they considered that data are rarely accurate enough to enable specific values of both 0 and 4 to be determined. They therefore suggested that in practice it is sufficient to take the value of 4 as unity. Thus Eq. (17) was restated as 1 _ W e - a + bP

(18)

Equation (1 8) is subsequently referred to as Bleasdale’s simplified equation. e. Farazdughi and Harris. Farazdaghi and Harris (1968) have recently derived a yield/density equation from the same logistic growth curve used by Shinozaki and Kira (1956). However, these workers stressed that the law of constant final yield for total crop dry matter may not always hold true. [Farazdaghi (1 968) has even shown that total dry matter in the sugar beet crop can be either asymptotic or parabolic depending on the environment.] Thus they modified the law of constant final yield to

w pY= K

(19)

and derived an equation:

This equation can describe either an asymptotic or a parabolic yield/ density situation, depending on the value of y ; in the former case y = 1, and in the latter case y > 1.

PLANT POPULATION A N D CROP YIELD

301

C. A FURTHER EXAMINATION OF THE RECIPROCAL EQUATIONS In Section I1 B the general characteristics of the different yield/density equations were discussed. The object of this section is to examine the reciprocal equations in more detail. These equations are singled out for this further examination for three main reasons: first, they are the only type of equation that can realistically describe both the asymptotic and the parabolic yield/density situations, either by means of different forms of equation (Holliday, 1960b) or by a single generalized equation (Bleasdale and Nelder, 1960); second, a good deal of biological meaningfulness has been claimed for them; and third, they have probably been used more than any other type of equation in recent years. 1 . The Biological Basis of the Reciprocal Equations a . Biological Derivation. ( 1 ) Shinozaki and Kira (I 9513, Bleasdale and Nelder ( 1 960), and Farazdaghi and Harris ( 1 968). Shinozaki and Kira derived their yield/density equations from the following assumptions on the growth of a plant: (i) The growth of a plant can be described by a simple logistic growth curve.

W =

W

1

+ ke-At

where w is the weight of the plant at time t, A is the coefficient of growth, and k is the integration constant. Both W and A are assumed constant independently to time t. (ii) A in the above equation is independent of density. (iii) Final yield per unit area is constant and independent of density after the law of constant final yield (Eq. 8) of Hozumi et al. ( 1956). On combining Eqs. (8) and (21) and determining the value of k when there is no competition at time zero, when the weight per plant is wo, the reciprocal equation can be derived 1 = W

a+bp

where a = e-A'/w,and b = ( 1 - e-")/Y.

302

R. W. WILLEY AND S. B. HEATH

Shinozaki and Kira (1 956) go on to show that their reciprocal equation can be derived from more general growth curves where A and W are not independent of time, which would be an obvious objection. It was seen earlier that Farazdaghi and Harris (1968) used the same growth function as Shinozaki and Kira but did not assume the law of constant final yield (Section 11, B, 5 , e). On the other hand, Bleasdale and Nelder (1960) stated that they derived their equation from a generalization of the logistic growth curve given by Richards (1959) “with ahalogous arguments to those of Shinozaki and Kira.” The main interest in the biological derivation of these particular reciprocal equations lies in their application to the yield/density relationship of a plant part. Kira et al. (1 956) had observed that the weight of a plant part could be related to the weight of the whole plant in the following way

or log w, = log k

+ h log w

i.e., the logarithm of the weight of the plant part has a linear relationship with the logarithm of the total weight: log K is the intercept and h is the slope of the regression line. This is Huxley’s law of relative growth, or law of allometry. Bleasdale (1967) pointed out that in its original context this law was applied to the relationship between plant part and total where plant size increased with age, whereas in the present context it is used to describe the relative changes brought about by density. From this allometric relationship, Shinozaki and Kira (1956) modified their original equation (Eq. 9) to describe the yield/density relationship of a plant part

This equation is very similar to Bleasdale’s simplified equation (Eq. 18). It can describe a parabolic yield/density curve for a plant part, although, because of Shinozaki and Kira’s initial assumptions of the law of constant final yield, this equation still assumes that the total yield/density curve is asymptotic. However, Shinozaki and Kira do not appear to have tested this equation in practice.

PLANT POPULATION A N D CROP YIELD

303

Bleasdale ( 1966a, 1967) also made use of the allometric relationship, which he stated in a slightly different form as w = Kw;

(24)

Combining Eq. (24) with his simplified equation (Eq. 18) he was able to derive a similar equation to the latter to apply to a plant part

I

-=aal+

W 0,

brp

where 0, = 0 A . Thus in the situation where 0 = I , where the total yield/ density curve is asymptotic, 0,= A . In this situation 0, can therefore be estimated directly from the allometric relationship as the slope of the regression line. Bleasdale pointed out that this could allow the construction of the whole yield/density curve from only two densities: two densities would enable 0, to be estimated from the allometric relationship, and once this constant was determined Bleasdale’s simplified equation for a plant part could also be fitted on two densities. (The dangers of fitting a yield/density curve on two densities were discussed earlier under Section 11, B, 2.) The estimation of 0, in this way also has another advantage, in practice perhaps a more useful one. As will be seen later, when fitting any reciprocal equation which contains a power it may be difficult to obtain an accurate estimate of this power. Thus, if in this instance 0, can be estimated from the allometric relationship when total yield per area is asymptotic, Eq. (25) can then be fitted more accurately. Although Farazdaghi and Harris ( 1 968) suggested that their basic equation (Eq. 20) could be used directly for a plant part, they also derived a more meaningful equation using the allometric relationship (Eq. 24). This took the form

They said that this described the way plant density affected the distribution of dry matter into plant parts. It is of interest that this equation is very similar to one of the original ones of Bleasdale and Nelder (1 960) (Eq. 17), which had proved difficult to fit in practice. But Farazdaghi and Harris (1968) pointed out that if y was estimated from their basic equation (Eq. 20) by fitting the total yield/density curve, this would then allow

304

R. W. WILLEY A N D S. B. HEATH

the more complicated equation (Eq. 26) to be fitted, since only one power would have to be estimated from the regression analysis. However, it is noteworthy that A could also be determined from the allometric relationship (whether the total yield/density curve is asymptotic or not), although y still has to be determined by fitting either Eq. (20) or Eq. (26). Also, it is of interest that the only situation in which the approach of Farazdaghi and Harris ( I 968) does not entail determining at least one power by directly fitting one of their yield/density equations is again when total yield is asymptotic, for in this case y = 1 and A is obtainable from the allometric relationship. It must be emphasized, however, that to make use of the allometric relationship, or to estimate y in Eq. (20), it is necessary to have data for total plant weight as well as for plant part. In practice this may present difficulties. For example, in the cereal crop it may be difficult to obtain comparable estimates of grain yield and total dry matter since the latter may have declined from its maximum value before the maximum value of the former is achieved. A similar situation exists in the potato crop when total dry matter and final yield of tubers are considered. Also, where the plant part is not present throughout the whole life of the crop, the use of the allometric relationship may require further consideration. In view of these difficulties, the agronomist may frequently find that in practice he is not in a position to predetermine any power which he can substitute in the equations specially derived for fitting plant part data (Eqs. 25 and 26). He must then fall back on the use of either Bleasdale’s simplified equation (Eq. 18) or Farazdaghi and Harris’s basic equation (Eq. 20) and apply these directly to his plant part data. In this situation there seems little to choose between these two equations; they both involve fitting one power and they describe very similar yield/density curves. (2) De Wit (1960). De Wit’s approach to the derivation of his yield/ density equation is of interest because it differs markedly from the approaches seen in the previous section. De Wit termed his equation a “spacing formula,” and he derived it from a consideration of the space available to a plant and the plant’s ability to take up that space. He developed this formula from a consideration of two species grown on a homogeneous field of unit surface which he assumed to be divided into a number of squares of equal size. In a first model he assumed that the growth of one plant was unaffected by the growth of another. Thus if a plant is grown in each square the yields of each species in different mixtures can be represented by Fig. 9A. However, de Wit pointed out that in practice this situation would occur only where the density was so low that there was no competition or where the competitive powers of the two species were equal.

305

PLANT POPULATION A N D CROP YIELD

D e Wit developed his argument for the more practical situations where plants did compete. This situation is illustrated diagrammatically in Fig. 9B. I t can be seen that the yields of species 2 are higher, and the yields of species 1 are smaller, than would be expected from the first model.

+Species Species

2

4 P

I

tSpecies

I

Species 2-

P

FIG.9. Diagrammatic representation of the yields of each of two species (sp. I , sp. 2) grown in different mixtures: ( A ) where there i s no competition between the species: and (B) where competition exists between the species. At any point on the p axis, all the squares of the homogeneous field contain a seed either from species I or 2.

De Wit said this was because the plants of species 2 crowded the plants of species 1 out of some of the space alloted to them. H e went on to consider the situation where only one species occupied some of the squares and the rest remained empty (i.e., the relationship between a single species and plant density). He suggested that this single species would now behave in a manner similar to the dominant species in a mixed situation in that it would occupy more space than was allocated to it. Thus the relationship between the number of plants and the yield of this species would be comparable with that observed for species 2 in the example above (Fig. 9B). D e Wit assumed this relationship to be asymptotic and, using some oat data of Montgomery (1912), he derived the expression given earlier (Section 11, B, 5, c). It might be questioned whether the now arbitrary choice of the size of the square might not affect the result of the spacing formula: however, de Wit showed that this was not the case. It is of considerable interest that from this sort of approach de Wit developed an equation very similar to the other reciprocal equations. However, since de Wit has not suggested any modification of his equation to describe parabolic yield/density situations, the derivation outlined above is not examined in depth. b. Biological Validity of the Constants. It has been shown previously that for the situation where the yield/density relationship is asymptotic the equations of Shinozaki and Kira ( 1956), Holliday ( I 960a), Bleasdale

306

R. W. WILLEY AND S. B. HEATH

and Nelder ( 1960),and Farazdaghi and Harris ( 1968) all become identical and can be written 1

-=a+bp W

or, on a yield per area basis y

=-

P a+bp

Thus as density increases, y approaches the value of llb, i.e., the asymptote of yield per area = l / b . If it can be argued that the asymptote of yield per area is a measure of the potential of a given environment, then b is a meaningful factor indicative of environmental potential. In the case of Bleasdale’s simplified equation (Eq. 18), where 8 does not equal unity, the interpretation of the value of b is less obvious, although perhaps it may still give some indication of environmental potential. As the density tends to zero the value of yield per plant tends to lla in Eq. (27). It was seen earlier that this does not represent a very realistic situation for it ignores the fact that yield per plant levels off at densities too low for competition to occur (see Section 11, B, 5 , b). However, assuming that lla gives some indication of yield per plant in a competitionfree situation then, by a similar argument to that employed for b, a can perhaps be regarded as a meaningful factor indicative of genetic potential. It is interesting to pursue this reasoning with de Wit’s spacing formula _1 - _ 1_ 1 w-QP+Pp

In this instance the asymptote of yield per area is P, and yield per plant appears to approach a value Q P as density approaches unity. Thus in this equation maximum yield per plant, QP, is defined partly in terms of an environmental factor, P. This probably represents a more realistic situation. Several workers have examined the time trend in these constants. Shinozaki and Kira (1956) found that the value of b in a soybean density experiment rapidly increased with time in the period just after germination, but thereafter it fell, rapidly at first and then more slowly, toward a constant value of b. Jones (1 968) found a similar time trend in b for the dwarf bean crop. From a biological point of view it is hard to explain the

307

PLANT POPULATION A N D CROP YIELD

rapid rise in the value of b after germination. However, this may be an effect of the absence of competition affecting the fitting of the equation. Jones (1968) also found that the value of a fell throughout the season although there was a tendency to approach a constant value toward the end. Reestman and de Wit (1959) determined the course of P and Q in their equation (Eq. 14) throughout the latter part of the growing season in an experiment on sugar beet. Both P and Q increased with time and approached a constant value, Q doing so rather more quickly than P . With the exception of the changes in b at the early stages of growth, these time trends in the values of the constants are reasonably in accordance with what might be expected from their suggested biological significance. Bleasdale and his co-workers have examined the effect of variety and environment on the values of a and b. Bleasdale ( 1 966b) analyzed some yield/density data with Eq. (18) for three varieties of onions grown in the same environment. The values of a and b obtained are given in Table 11, and these suggest that the value of a depends upon the variety. HowTABLE 11 THE VALUES OF THE CONSTANTS a A N D b OBTAINED BY FITTING EQ. (18) TO THE Y I E L D ~ D E N S DATA I T Y FOR THREEONION VARIETIES“”’

Variety LANCASTRIAN RIJNBURCER SUTTONA1

b

(I

0.01 117 0.0 I90 I 0.01706

]

0.00263

“After Bleasdale ( 1966b). ”The data were fitted taking a common value of 0 = 0.8.

ever, it must be emphasized that these differences in a were for a common fitted value of b, since Bleasdale did not find significant differences between the individual values of b. In the same paper Bleasdale stated that unpublished results with other crops suggested that the value of b varied according to the soil fertility, but the value of a did not: further that a appeared to be a constant from year to year for a given variety. This hypothesis was also borne out by work of Bleasdale and Thompson ( 1966) on parsnips. This idea was further investigated for some wheat data of Willey (1965) for which one variety was grown under a number of environmental treatments. The values of a and b obtained by fitting Eq. (18) are given in Table 111. It can be seen that for all four treatments, fitted independently for a and b, a appears to be reasonably constant but b changes.

R. W. WILLEY AND

308

THE VALUESO F TO THE

TABLE 111 CONSTANTS u AND b OBTAINED BY FITTING EQ. (18) YIELD~DENSITY DATAFOR A WHEAT VARIETY“’b

THE

Control Treatment I Treatment 2 Treatment 3 ~~

S. B. HEATH

~

U

b

0.109 0.105 0.097 0.108

0.0673 0.0815 0.1092 0.0963

~~~~

“After Willey (1965). *The data were fitted taking a common value of 8=0.5. (Yield data are given inTable 1V.)

Holliday ( 1960a) examined the meaningfulness of the constants in his equations in rather a different way. It was seen earlier (Section 11, B, 5 , b) that he appreciated the significance of the constant a and termed l/a the “apparent maximum” yield per plant (A). Thus, substituting l/A for a , the form of his basic asymptotic equation becomes

which can be written

or, on an area basis

+

He termed the expression l / ( l Ab p ) the “competition function,” and it can be seen that the value of this decreases as density increases. Holliday considered that the definition of yield per plant as A [I/( I Abp)] gave a realistic description of what actually happens in practice, for it indicates how the yield of a plant at any given density is a product of the potential of the plant (A) and the forces of competition that are acting upon it [ I / ( 1 + A b p ) ] . Similarly, Holliday ( 1 960b) expressed his parabolic equation as

+

PLANT POPULATION AND CROP YIELD

+

309

+

In this case the competition function is I/( 1 A b p Ac pz). In conclusion, from the evidence presented it appears that it may be possible to ascribe some biQlogical meaningfulness to the constants in the reciprocal equations. Thus, examination of these constants may help to pinpoint genetic or environmental components of yield/density relationships. However, it seems likely that the inevitable interaction of these two components is a far more complex situation than can be described by a few simple constants. It would therefore seem desirable that these constants should be examined in much more detail before any exact biological meaning is ascribed to them. 2. Statistical Regression Analysis and Goodness of Fit In this section it is proposed first of all to consider some general points about the statistical fitting of yield/density equations. This is followed by a more detailed examination of the fitting of the reciprocal equations and the goodness of fit which they can give. It should be emphasized that it is important to consider why the regression analysis is being carried out. Where the intention is merely to fit a smooth curve to some data points, the analysis can often be satisfactorily carried out on the yield per area/density data, as has been shown by Sharpe and Dent (1 968). However, for reasons already given, for a more satisfactory description of the true yield/density relationship, it is usually more desirable to carry out the regression on the yield per plant/density data. An illustration of this is given in Fig. 10, in which two regressions have been carried out on the grain yield/density data of Pendleton and Dungan ( 1 960). The simple quadratic equation (Eq. 1) is fitted directly to the yield per area data, and Holliday’s parabolic equation (Eq. 12) is fitted to the yield/plant data. It can be seen that there is little difference between the regression lines as far as their ability to smooth the data is concerned but, at low densities particularly, the regression on yield per plant gives a much better description of the yield/density relationships. The method which has usually been used to fit the yield/density equations has been a least squares regression. The reciprocal equations describing an asymptotic yield/density situation can be fitted by a simple linear regression, but with the introduction of powers or the use of the quadratic to describe a parabolic situation, the regression becomes more complicated. One of the assumptions on which the least squares regres-

3 10

R. W. WILLEY A N D S. B. HEATH

sion rests is that the variation about any one point is the same as that about any other. This means that the variance of the yield values, whether it be yield per plant or yield per area, must be constant over all densities.

/. OO

3

6

12

9

15

18

P

FIG. 10. The quadratic equation (Eq. 1 ) fitted directly to yield per unii area (-----) and the parabolic equation (Eq. 12) of Holliday (1960b) -( ) fitted via the reciprocal of yield per plant for the wheat data of Pendleton and Dungan ( I 960) meaned over four N levels and four varieties: y = bushelslacre, p = peckslacre.

Keller and Li (1 949) found thkassumption to hold with their data for a density experiment with hops, but their range of densities was limited. Hozumi et af. ( 1 956), when considering the individual yield of plants at three densities for leaf beet and turnips, found that the standard deviation of the points increased as plant size increased (with decrease in density), as shown in Fig. 1 1. Also, Nelder (1 963) criticized the curve fitting of Jarvis ( 1 962) for some lucerne density experiments and showed that it

...

.a

50

100

150

200

W

FIG. 11. A scatter diagram of the relationship between standard deviation (S.D.) and mean plant fresh weight of leaf beet, w (in grams), where changes in the latter were brought about by changes in plant density. (After Hozumi et al., 1956).

31 I

PLANT POPULATION A N D CROP YIELD

would be unlikely that the variation of yield per plant was uniform. He suggested that a more accurate assumption in yield/density experiments would be that the variance of the logarithm of yield per plant was constant. This assumption has been adopted by Bleasdale and his colleagues when fitting yield/density data to their equations. However, this assumption involves a more complicated treatment of the least squares regression than does the more usual assumption that the variance of w is constant. The main interest in examining the goodness of fit of the reciprocal equations is in the parabolic yield/density situation since in the asymptotic situation they are identical. In this situation the most useful comparison would seem to be between one of the equations for which some biological validity has been claimed and Holliday's more empirical equation (Eq. 12). Since most workers are probably more familiar with the approach of Bleasdale and his co-workers than with the more recent one of Farazdaghi and Harris ( 1 968), Bleadsale's simplified equation (Eq. 18) is compared with Holliday's. When fitting Bleasdale's simplified equation, the assumption is made that the variance of log (w-@)is constant; thus for a guessed value of 8, approximation to the true squares estimates of constants a and b can be obtained by a weighted regression of w - & with weights 0 % (Nelder ~ ~ 1963: Mead, unpublished). The criterion of the goodness of fit is the residual sum squares divided by 02, and the best value of 8 is that which reduces this to a minimum. An example of some spring wheat data (Willey, 1965) (Table IV) fitted by the above method illustrates this in Fig. 12. For each of the four TABLE I V G R A I NYIELD DATAFOR A WHEAT VARIETY GROWN AT FOURDENSITIES U N D E R FOURENVIRONMENTAL TREATMENTS".*

Grain yield (cwt./acre) Density (1 O6 plants/acre)

Control

Treatment I

Treatment I I

Treatment 111

0.392 1.122 2.432 5.78

20.82 35.32 29.98 23.95

20.60 28.15 30.14 15.85

19.63 23.71 19.51 10.35

19.62 20.57 22.8 I 12.93

After Willey ( 1965,. *Treatments: mean effect of reducing light intensity to 75 percent or 50 percent of full daylight during period of establishment-ear initiation (Treatment I), enr initiation-flowering (Treatment lI), and flowering-harvest (Treatment I 11). "

~

312

R. W. WILLEY AND S . B. HEATH

treatments fitted, a curve shows the goodness of fit (residual sum squares divided by P ) plotted against different 8 values. From the graph it can be seen that there was quite a wide variation between the best-fitting values

’

0.2

’

0.4

’

0.6

’

0:8

’

1lO

e FIG. 12. An example of obtaining the best-fitting value of 0 in Bleasdale’s simplified equation (Eq. 18); the best fit is where the residual sum squares (s.s.)/O* is reduced to a . .-, minimum. Wheat grain yield data of Willey ( 1965); Control (---), Treatment 1 (Treatment 11 (. . -), Treatment I11 (- ) (see Table IV).

-

-

of 6 for each treatment. However, if a common value of 8 of 0.5 was taken the individual 8’s for the four treatments did not give a significantly better fit. The best values of 8 were particularly well defined in these data, but this is not always the case. It may happen that the minimum is not so sharply defined, and in these circumstances it might be better to fit the correct but more complicated least squares regression of log ( w - O ) to avoid the approximation of the simpler weighted regression suggested by Nelder ( I 963); or this could be a situation in which Holliday’s equation is of more practical use. To make the comparison between Bleasdale’s simplified equation (Eq. 18) and Holliday’s parabolic equation (Eq. 12) valid, the latter is also fitted on the assumption that the variance of log (w)is constant. A comparison is made in Fig. 13 and Tables V and VI of the goodness of fit of

313

PLANT POPULATION A N D CROP YIELD

these equations to some selected data. The data are not meant to be comprehensive, but are chosen to illustrate the fit in two different parabolic situations, i.e., a definite parabolic situation (Table V and Fig. 13C and

70

IB P

ID

90. 80.

70. Y

60.

t 50L

I.,

4

8

12

P

16

20

24

4

.

8

12

16

20

24

P

FIG. 13. Examples of Holliday’s parabolic equation (Eq. 12) (---), and Bleasdale’s simplilied equation (Eq. 18) ( ), fitted to some yield/density data where yield declines only gradually at high densities (A and B) and where yield declines quite sharply at high densities (C and D); (A) Grain yield of wheat (Pendleton and Dungan, 1960) meaned over four N levels and four varieties; y = busheislacre, p = peckslacre. (B) Grain yield of wheat, var. HILGENDORF meaned over four years (Crawford, 1964);y = bushelslacre, p = bushels/ acre. (C and D) Grain yield of maize for hybrids WF9 X 38-1 1 and HY2 x OH7, respectively, at a medium N level (Lang e t a l . , 1 9 5 6 ) , y = bushelslacre, p = lo3p l a d a c r e .

D) and a situation where there is only a slight decrease in yield at high densities (Table VI and Fig. 13A and B). The curves plotted in Fig. 13 are also represented in the respective tables. Although these data are very limited, it can be seen that there may be little difference on average between the two approaches, although one might be better than another for a particular set of data. It can also be seen that the ability of the equations to fit the data can vary considerably (compare Fig. 13A and B with Fig. 13C and D).

314

R. W. WILLEY AND S. B. HEATH

Ill.

The Relationship between Plant Rectangularity and Crop Yield

It was emphasized in Section I that yield per unit area is dependent not only on the number of plants per unit area (plant density) but also on the spatial arrangement of those plants (plant rectangularity). Plant rectanguTABLE V A COMPARISON OF THE VARIATION REMAINING AFTER FITTING EQS. (18) AND (20)" A. Dumanovic and Penick (1962): Single-Cross Maize Hybrid Grown at 5 Densities and 4 Levels of Nitrogen Equation ( I 8):

Nitrogen level (kg.lha.1

TSS

H"= 0.45, RSSl0'

Equation (20) RSS

0 50 100 150

1.01 0.961 0.916 0.886

0.0 154 0.023 I 0.0275 0.0272

0.0 152 0.0236 0.0278 0.0274

B. Lang et al. (1956): Three Maize Hybrids Grown at 6 Densities and 3 Levels of Nitrogen

Hybrid

Nitrogen level

TSS

Equation (18), Ob=0.45,RSS/02

Equation (20), RSS

HY2 X OH7

Low Medium High Low Medium High Low Medium High

I .26 0.72 I 0.608 2.30 I .04 0.218 2.83 1.48 1.02

0.00809 0.0191 0.0 128 0.0422 0.00878 0.00209 0.0236 0.0055 1 0.00767

0.00420 0.00847 0.0139 0.0341 0.00257 0.00173 0.0200 0.00770 0.00744

WF9 X OH4 I

WF9 X 38-1 1

" RSS = Residual Sum Squares; TSS = Total Sum Squares. Both equations were fitted on the assumption the variance of log (wJ is constant. The RSSIB" obtained from fitting Eq. (1 8) are directly comparable with the R S S obtained from fitting Eq. (20). V a l u e of 0 which for the set of data as a whole reduced RSS/02 to a minimum.

larity can be most easily visualized in a row crop where it can be defined as the ratio of the distance between plants within the row to the distance between the rows. In a broadcast crop it may be more generally defined as a measure of the unevenness of distribution. This rectangularity, or unevenness of distribution, is important because of the unevenness of competition which it produces; competition may be too intense between some plants and insufficiently intense between others.

315

PLANT POPULATION A N D CROP YIELD

TABLE VI COMPARISON OF THE VARIATION REMAINING AFTER FITTING EQS. (18)

Study

TSS

Donald ( 1954) I . Wimmera ryegrass grown at 5 30.2 densities for seed 23.7 2. Subterranean clover grown at 5 densities for seed Crawford (1 964) Wheat var. HILGENDORF grown at 5 0.72 densities for grain Pendleton and Dungan ( I 960) Wheat grown at 6 densities for grain, 2.33 average over 4 nitrogen levels and 4 varieties Puckridge and Donald ( 1967) Wheat grown at 5 densities for grain 19.3

Bb

Equation ( I 8) RSS/%'

AND

(20)O

Equation (20) RSS

0.90

0.0149

0.00292

0.90

0.0394

0.0536

0.85

0.0000806

0.0000947

0.75

0.0007 1 I

0.000234

0.85

0.0171

0.040 1

RSS = Residual Sum Squares; TSS =Total Sum Squares about the mean. *Value of B which reduced RSS/A2to a minimum for the particular set of data.

The extent to which rectangularity may effect the yield of a crop is clearly dependent on the plasticity of the individual plant, which in turn must be dependent on the plant species. However, the general pattern of effects is illustrated by some winter wheat data of Harvey et al. ( 1 958) reported in Table VII. The treatments of Harvey etal. were not extreme, yet it can be seen that as rectangularity increases, either by increasing seed rate or increasing row width, yield per area gradually declines. Similar effects have been shown by Wiggans (1939) for soybeans, Reynolds ( 1950) for peas, Pendleton and Seif ( I 96 1) for maize, Bleasdale (1963) for peas, and Weber et al. (1966) for soybeans. Reynolds (1 950) also showed that as rectangularity increases the optimum density may decrease (Fig. 14). It would therefore seem desirable that equations describing the relationships between plant population and crop yield should be able to describe the effects of rectangularity as well as those of density. This can be particularly important because in the many population studies where different populations have been established on constant row width, rectangularity is not constant but increases with increase in density. Goodall ( 1 960) attempted to fit the model

316

R. W. WILLEY AND S. B. HEATH

TABLE VII THE EFFECTOF Row WIDTHA N D SEED RATE ON WINTERWHEATGRAINYIELDS (cwt./acre)" Seed rate, stoneslacre Row width (inches)

5.5

II

17

4 8

43.9 43.0 41.6

43.9 42.5 41.4

43.6 41.4 38.0

12 UAfterHarvey et al. ( 1 958).

w = adpi dJ" or

where dl is the intrarow spacing and d2 is the interrow spacing. Thus d l d z is the space available per plant. Equation (28) is therefore an extension of Eq. (6).

Goodall fitted this model to some soybean data of Wiggans ( 1939) which covered a range of densities and row widths. He found a significant

30

t

'Ot

o . - - o o

10

20

30

P

FIG. 14. The effects of rectangularity on the yield/density relationship in dried peas (Reynolds, 1950): the three curves represent different row widths, 8, 16, and 24 inches: y = cwt./acre, p = wnedacre.

PLANT POPULATION A N D CROP YIELD

317

difference between bl and b2; he suggested that this was due to row orientation effects. Donald (1963) pointed out, however, that Eq. (28) has the undesirable characteristic that, if either of the power terms is greater than the other, then the optimum rectangularity at a given density would be obtained where the distance between plants was increased in one direction and decreased in the other. Berry ( 1967) criticized Goodall’s fit to Wiggans’ data, not only on account of the poor fit of log w against log d l , but also because the values of d , and dz were not overlapping, and therefore different values of b1 and 6, could be expected. Berry ( 1967) extended the simplified equation (Eq. 18) of Bleasdale and Nelder to take into account plant rectangularity (29)

Since dld2= s, this model has included an extra term proportional to the square root of density. For a given density, w is greatest where dl = d,, i.e., where recta’ngularity is 1 : 1 , since (l/dl) (I/d2) is at a minimum value. This relationship gave a satisfactory fit to Wiggans’ soybean data. Berry considered that for irregularly spaced crops, i.e., where the rectangularity is not constant, Eq. (29) might still be used as a first approximation from Bleasdale’s simplified equation. For example, it could be used where plants are irregularly spaced within the row and rectangularity is defined by the mean intrarow distance and the interrow distance.

+

IV. The Variation in Yield of the individual Plant

It was emphasized in the introduction that the variation in the yield of the individual plant has seldom been examined in yield/density studies. The analysis has been in terms of the mean yield per plant at a particular density with no consideration of the variation about this mean. Yet this variation can be of great importance wherever the size of the individual plant is an attribute of yield. For example, in Fig. 2D the effect of size grading on the marketable yield of parsnips can be seen at each plant density although the latter has little effect on total yield. Kira et al. (1 953), Hozumi et al. ( 1956), and Stern ( 1 965) attempted to examine the effect of density and time on the variation in individual plant weights by calculating the coefficients of variation at each density. Kira et al. and Stern showed that the coefficients of variation increased with time, but the evidence was not consistent as to whether density affected the value of the coefficient of variation at any one time. However, Mead

318

R. W. WILLEY A N D S. B. HEATH

(1967) has stressed that no attempt was made in these studies to test whether the shape of the distribution curves were constant, a necessary condition before coefficients of variation can be compared. Koyama and Kira ( 1956) considered the frequency distributions of pldnt weights at different densities. They found that although -the distribution for seed weight was normal, as the plants grew the distribution became more and more skew. The development of skewness was greatest at high densities. Kira et al. ( 1953) also tried using a correlation coefficient between the weight of an individual and the mean weight of the six plants nearest to it in a soybean experiment sown in a regular hexagonal arrangement. Surprisingly, the correlation coefficients were low and proved positive, apparently suggesting cooperation among plants rather than competition. Mead ( 1967) considered that for small samples the use of this form of the correlation coefficient is a biased estimator of the degree of competition in the population. Mead proposed a measure of the weight relationship between a plant and its six immediate neighbors, grown in a regular hexagonal arrangement, termed the competition coefficient. This is a general measure of the plant neighbor relationship for all the plants in the community depending only on the size of the neighboring plants but assuming regular arrangement. Mead ( 1968) analyzed the results of some experiments laid down to examine this relationship and showed that, for cabbage, carrots, and to a lesser degree, radishes, the competition coefficients are predominantly negative, showing competition between plants rather than cooperation as was indicated by Kira’s results. Mead therefore considered the competition function of more use than the correlation coefficient. Mead ( 1 966) investigated the importance of irregularity of spacing on the variation of individual plant yield within a population. He did this by examining the importance of the size and shape of the space available to the plant in determining the yield of that plant. He suggested a model in which the total ground area under an irregularly spaced crop was divided into polygons, each one being allocated to a single plant. This was done by allocating any given spot of ground to the nearest plant. The polygons could then be characterized by three parameters, the area, the extent to which the polygon was elliptical rather than circular (a measure of rectangularity), and a measure of how far the plant was from the center of the polygon. On examining the relationship between the three parameters and the root diameter of carrots grown at three densities and three row widths, he found that the proportion of the total variation in plant yield attributable to polygon variation increased with time, the largest mean proportion at a final harvest being 20 percent, or for individual plots

PLANT POPULATION A N D CROP YIELD

319

as much as 55 percent. Mead also found that the area of the polygon was more important than its shape. Furthermore, for an irregularly planted crop the variation in area did not appear to change with change in density, but at a given density it increased with increasing row width. This would seem to indicate that for maximum uniformity within a crop, the rectangularity ratio should be as low as possible. As Mead pointed out, this approach is somewhat unrealistic for it takes no account of the size of the neighboring plants or the fact that plants other than the immediate neighbors defined by the polygons might affect the plant either directly or indirectly. However, this initial approach seems to be a useful one. V.

Conclusions

This review has attempted to examine the usefulness and biological validity of the different mathematical equations that have been proposed to describe the relationships between plant population and crop yield. I t has been seen that even the simplest and most empirical of equations may be useful in certain circumstances, e.g., to smooth data over a limited range of densities. However, to describe the relationships realistically over a wide range of densities, or to construct yield/density curves from a minimum of data, it would clearly seem to be desirable to use those equations that have a better biological foundation and have proved the most satisfactory in practice. In general the reciprocal equations seem to fulfill these requirements best. The satisfactory description which they provide of the asymptotic yield/density situation has been established by many workers. The description of the parabolic yield/density situation is less certain, but at least this group still provides a choice of three basic equations (Eqs. 12, 18, 20), all of which are inherently very flexible and and all of which offer a reasonable possibility of obtaining a satisfactory and realistic fit. A study of the evidence at present available suggested that some biological meaning can perhaps be ascribed to the constants in the reciprocal equations, but further research is needed in this field. An equally worthwhile field would seem to be the incorporation of the effects of plant rectangularity and plant variability into the yield/density equations; relatively few workers have studied these factors, yet their effects on agricultural yield may be considerable.

ACKNOWLEDGMENTS The authors wish to thank Mr. R. Mead, University of Reading, for many helpful discussions concerning the statistics.

3 20

R. W. WILLEY A N D

S. B. HEATH

REFERENCES Berry,G. 1967. Biometries 23,505-5 15. Bleasdale, J. K. A. 1963. I n “Crop Production in a Weed Free Environment” (E. K. Woodford,ed.),pp. 90-101. Bleasdale, J. K. A. I966a. Ann. Appl. Biol. 57,173- 182. Bleasdale, J. K. A. 1966b. J . Hort. Sci. 41, 145- 154. Bleasdale, J. K. A. 1967.J . Hort. Sci. 42,5 1-8. Bleasdale, J. K. A., and Nelder, J. A. 1960. Nature 188,342. Bleasdale, J. K. A., and Thompson, R. 1966.J. Hort. Sci. 41,37 1-378. Bruinsma, J. 1966. Neth. J . Agr. Sci. 14,198-2 14. Campbell, R. E., and Wets, F. G. 1967.Agron. J. 59,349-354. Carmer, S. G., and Jackobs, J. A. 1965.Agron. J . 57,241-244. Crawford, W. R. 1964. N e w ZealandJ. Agr. 108,455-463. de Wit, C. T. 1959. Jaarb. Inst. Biol. Scheik. Onders. LandbGewass. Meded. 83,129-134. de Wit, C. T . 1960. Verslag. Landbouwk. Onderzoek. 66.8,I-8 1. de Wit, C. T., and Ennik, G. C. 1958. Jaarb. Inst. Biol. Scheik. Onderz. LandbGewass. Meded. 50,59-73. Donald, C. M. 195 I . Australian J. Agr. Res. 2,355-376. Donald, C . M. 1954. Australian J . Agr. Res. 5,585-597. Donald, C. M. 1963. Advan. Agron. 15,l- I 18. Dumanovic, J., and Penick, M. 1962. Savremera Polio Priverda 4,240-250. Duncan, W. G. 1958. Agron. J . 50,82-84. Farazdaghi, H. 1968. Ph.D. Thesis, Reading University, England. Farazdaghi, H.,and Harris, P. M. 1968. Nature 217,289-290. Goodall, D. W. 1960. Bull. Res. Council Israel D8, 18 1-192. Harvey, P. N., Whybrew, J. E., Bullen, E. R., and Scragg, W. 1958. Exptl. Husbandry 3, 3 1-43. Holliday, R. 1960a. Nature 186,22-24. Holliday, R. 1960b. Field Crop Abstr. 13,159- 167 and 247-254. Hozumi, K., Asahira, T., and Kira, T. 1956. J . Inst. Polytech., Osaka City Univ.D7, 15-33. Hudson, H. G. 194 I . J . Agr. Sci. 31,138- 144. Jarvis, R. H. 1962. J. Agr. Sci. 59,28 1-286. Jones, L. H. 1968.Agr. Prog. 42,32-62. Keller, K. R., and Li, J. C. R. 1949.Agron. J . 41,569-573. Kira, T., Ogawa, H., and Shinozaki, N. 1953. J . Inst. Polytech., Osaka City Univ.D4,1-16. Kira, T., Ogawa, H., and Hozumi, K. 1954. J . Inst. Polytech., Osaka Ciiy Univ. D5,l-7. Kira, T., Ogawa, H., Hozumi, K., Koyama, H., and Yoda, K. 1956. J. inst. Polytech., Osaka City Univ. D7,l- 14. Koyama, H., and Kira, T. 1956. J . Inst. Polytech., Osaka City Univ. D7,73-94. Lang, A. L., Pendleton, J. W., and Dungan, G. H. 1956. Agron. J . 48,284-289. Mead, R. 1966.Ann. Botany (London) [N.S.] 118,301-309. Mead, R. 1967. Biometrics 23,189-205. Mead, R. 1968. J. Ecol. 56,35-45. Mitscherlich, E. A. 1919. Landwirtsch. Jahrb. 53,341-360. Montgomery, E. 19 12. Nebraska Univ.,Agr. Expt. Sta., Bull. 24(after de Wit, 1960). Nelder, J. A. 1963.5. Agr. Sci. 61,427-429. Pendleton, J. W., and Dungan, G. H. 1960. Agron. J . 52,3 10-3 12. Pendleton, J. W., and Seif, R. D. 196 I . Crop Sci. 1,433-435.

PLANT POPULATION A N D CROP YIELD

Puckridge, D. W.. and Donald, C. M. 1967. Australiun J . A g r . R e s . 18,193-2 1 1. Putter, J., Yaron, D., and Bielorai, H. 1966. Agron. J . 58,103-104. Reestrnan, A. J . , and de Wit, C. T. 1959. Nerh. . I Agri. . Sci. 7,257-268. Reynolds, J . D. 1950. Agriculture (London) 56,527-537. Richards, F. J . 1959. J . Exptl. Botany 10,290-300. Saunt, J. 1960. M.Sc. Thesis, University of Leeds, England. Sharpe, P. R.,and Dent,J. B. 1968.J.Agr. Sci. 70,123-129. Shinozaki, K., and Kira, T . 1956. J . lnst. Polytech., Osaka Ciry Univ. D7,35-72. Stern, W. R. 1965. Australian J . Agr. R e s . 16,541-555. Warne, L. G. G. 195 1. J . Hort. Sci. 26,84-97. Weber, C . R., Shibles, R. M., and Blyth, D. E. 1966. Agron. J . 58,99- 102. Wiggans, R. G. 1939. J . A m . Soc.Agron. 31,3 14-32 I . Wilcox, 0.W. 1950. Agron. J . 42,4 10-4 12. Willey, R. W. 1965. Ph.D. Thesis, Leeds University, England.

32 1

This Page Intentionally Left Blank

AUTHOR INDEX Numbers in italics refer to pages on which the complete references are listed. A

Aurand, L. W., 65, 108

Abou Akkada, A. R.. 50, 97 Acton, C. J., 197, 199, 204, 207, 209, 21 I , 230 Ada, G. L., 224,232 Adams, G . A., 224,230 Ademosum, A. A., 5 , 13, 97 Adler, F. E., 83, 99 Aitken, J. N., 92, 101 Alderman, G., 59, 97 Aldrich, D. T. A., 20, 22, 23, 66, 97, 100 Alexander, E. B., Jr., 247, 254, 255, 258, 2 72 Alexander, L. T., 150, 169, 257,274 Alexander, M., 2 19, 22 I , 230 Alexander, R. H., 13, 15, 97 Alfred, S., 275, 278 Allaway, W. H., 60, 97, 100 Allcroft, R., 56, 97 Allen, 0. N., 196, 197, 198, 199, 204, 209, 2 l9,22 1,222,231,233,234 Allen, P. E., 184, 192 Allison, J . B., 173, 193 Allison, L. E.. 19R. 230 Altschul, A. M.. 172. 173. 178. 180. 191 Anderson, D. M. W., 288.231 Alvarado, G., 189, 193 Alverez-Tostado, M. A,, 189, 193 Alvistur, E.. 180, 193 Anderson, J. U., 248, 252, 2S8, 272 Anderson, M. J., 17, I03 Andrews, E. D., 57, 97 Andrews, 0. N.. 29, 3 I , 97 Andrews, R. P., 36, 52, 102 Annison, E. F.. 41, 50, 5 I , 97 Antonopoulos, C. A,, 229,231 ap Griffith, G., 22, 54, 55, 56, 97, 99, 107 Appelqvist, L.-A,, 190, 192 Armstrong, D. G., I3,39,41,43,44,97 Arnold, G. W., 84,87,88,92,97, 103 Asahira,T., 294, 301, 310, 317,320 Atloe, 0. J., 196, 197, 198, 199, 204, 209, 22 I , 231, 233 Atwood, K. C., 176, 193

B

Babcock, G. E., 188, 194 Bach, R.. 197,204,208,216,218,232 Bacon, J . S. D., 197, 2 19, 235 Bacon, S. R., 266, 274 Baert, L., 5 5 , 69, 100 Bahn, A. V., 62,104 Bailey, H . H., 243, 250, 256, 260. 272, 2 73 Bailey, P. H., 82, 97 Bailey, R. W., 48, 97 Baker, E., 220, 234 Baker, G., 60, 97 Baker, J. C., 257,272 Baker, 0. E., 141, 169 Balch, C. C., 26, 27, 28, 36, 41,44,47, 52, 76, 97, 99, 106 Balch, D. A., 47, 97 Baldwin, M., 267, 272 Bandemer, S. L., 180, 192 Bandet. J., 185, 193 Banks, W.. 227, 228,231 Barclay, A. S., 180, 194 Barker, S. A., 202, 203, 207,208, 2 10, 2 I I , 212, 213, 214, 216, 217, 220, 223, 228, 231, 235 Barnes, R. F., 14, 15, 29, 30, 31, 97, 104 Barnett, L., 175, 192 Barrett, A. J., 2 I I , 231 Barrett, J. F., 62, 97 Barrow, N . J.. 59, 84, 97 Bartelli, L. J., 240, 244, 252, 263,272, 273 Bartlett, R. J.. 246. 272 Bartlett, S.. 47,97 Bartley, E. E., 73, 107 Barton, R. A., 92, 99 Bassette, R., 65, 100 Bates, L. S., 173, 180, 181, 182, 183, 187, 191, 193 Bath, D. L., 87, 90, 97 Bath, I. H., 45, 98 Battacharya, A. N . , 77. 107

323

324

AUTHOR INDEX

Bauer, P. T., 118, 169 Baum, R. R.,177, 192 Baumann, C. A., 60, I04 Baumgardt, B. R.,5 , 1 1 , 13, 31,47,50,97, 98, 106 Beadle, G. W., 180, 192 Beardsley, D. W., 7 1 , 98 Beath, 0. A,, 59, 60, 101, 106 Beavers, A. H., 242, 243, 248, 251, 253, 255,260,272, 273 Beck, A. B., 61, 62, 98, 106 Becker, G., 173, 191 Becker, M. J., 226,231 Bedell, F., 87, 98 Beesley, T. E., 267,273 Beeson, W. M., 183, 192 Beitz, D. C., 45, 46, 47, 99 Bell, J. M., 65, 103 Bell, M. C., 82, 98 Bell, T. A,, 25, 106 BeMiller, J. N., 198, 205, 235 Beene, E. J., 3, 35, 102 Bennett, H. H., 266, 274 Bennett, W. D., 65, 98 Bensadown, A., 73, 104 Bentley, 0. G . , 1 I , 105 Bergman, E. N., 48, 98 Bernier, B., 197, 200, 202, 208, 209, 210, 211,216, 217,218,231 Berry, G., 317,320 Betts, J. E., 62, 104 Bickoff, E. M., 61, 62, 63, 98, 104 Bielorai, H., 281, 321 Bigsby, F. W., 34, 107 Binns, B. O., 135,169 Bishop, C. T., 225,231 Black, W. A. P., 202, 207, 208, 209, 231 Blake, J. T., 89, 90, 99 Bland, B. F., 32, 98 Blaser, R. E., 17, 22, 24, 92, 95, 98, I00 Blaxter, K. L., 27, 28,3 1,32,37,39,41, 43, 44, 51, 70,72, 73, 91, 97, 98, 100 Bleasdale, J. K. A., 283,285,286,299,300, 301, 302, 303,306, 307, 315,320 Blyth, D. E., 3 15,321 Bobbit, J. M., 198, 231 Bogdan, A. V., 54,100

Bogue, D. J., 110, 169 Bohman, V. R.,8 8 , 9 0 , 1 0 6 Bolin, D. W., 13, 106 Bond, J., 53, 104 Booysen, P. de V., 34,107 Bornstein, J., 246, 272 Bosman, M. S. M., 13, 98 Bouveng, H. O., 228,231 Bowler, E., 220, 231, 234 Braden, A. W. H., 61, I04 Bradley, N. W., 81, 103 Brasher, B. R.,248,273 Bratzler, J. W., 106 Bray, A. C., 59,104 Bredon, R. M., 87, 98 Bremner, J. M., 204, 205, 221, 231 Brenner, S., 175, 192 Bressani, R., 178, 180, 184, 191, 192 Brett, D. J., 48, 103 Brewer, R.,251, 253, 272 Brian, R. C., 196,232 Briggs, D. R., 188, 193 Briggs, M. H., 26, 52, 79, 98 Briggs, P. K., 37,45, 98, I05 Brimacombe, J. S., 223, 228,231 Brimhall, B., 180, 192 Brink, R. H., 209,231 Brockway, J. M., 48, 98 Broderick, G. A,, 60, 104 Broekmeijer, M. W. J. M., 110, 169 Brohult, S., 178, 191 Brook, P. J., 65, 98 Broughan, R. W., 92,105 Brouwer, E., 43, 98 Brown, D., 80, 83,98 Brown, L. D., 27, 98 Brown, R. E., 45, 46, 47, 99 Brown, R. H., 17, 22, 98 Brown, S. M., 78, 79, 98 Brown, T. H., 93,106 Brownell, J., 57, 106 Browning, C. B., 12, 103 Browning, D. R., 252, 255, 256, 262, 265, 2 75 Bruinsma, J., 283,320 Brundage, A. L., 84, 105 Bryant, A. M., 47, 98 Bryant, H. T., 95, 98

AUTHOR INDEX

Buchman, D. T., 72, 98 Buckman, H. O., 265,273 Buettner-Janusch. J., 178, 192 Bull, L. B., 58, 98 Bullen, E. R., 3 15, 3 16, 320 Burau, R. G., 57, 106 Burdick, D., 7, 98 Burges, A., 219,231 Burns, J. C., 14, 104 Bums, K. N., 56, 97 Butler, G. W., 59, 92, 98, 105 Butterworth, M. H., 5 , 6, 98 Bystrom, B. G., 220, 234 C

Cady, J. G., 150, 169, 244, 248, 251, 252, 254,256,259,260,274,275 Calhoun, F. G., 250, 251, 260, 272 Call, J. W., 89, 99 Calvo, J. M., 190, 192 Calvo, R. A., 190, 192 Cameron, G. D. T., 33, 92, 98 Campbell, C. M., 87, 98 Campbell, J. R., 53, 98 Campbell, R. E., 283,320 Campling. R. C.. 26. 27, 28, 32, 36, 52, 7 1 , 72,75,76,77,97,99,100, 106 Carlisle, F. J., 239, 243, 245, 246,249,250, 25 1,254,256,257,272 Carlson, I. T., 13, 22, 107 Carmer, S. G., 289, 320 Carr, M. E., 265, 266, 273 Carrillo, B. J., 65, 108 Carroll, P. H., 167, 169 Carter, A. H., 81, 99 Carter, W. T., Jr., 266, 273 Cary, E. E., 60, 100 Cason, J. L., 1 1 , 98 Castle, M. E., 18, 96, 99 Chalmers, M. I., 50, 5 I , 97 Chalupa, W., 28, 52, 53, 99 Champakam, S., 171, 192 Charlwood, P. A,, 229, 231 Charter, C. F., 167, 169 Chen, M.-L., 182, 192 Chenost, M . , 34, 99 Cheshire, M. V., 200, 225,231

325

Chesters, G., 196, 197, 198, 199, 204, 209, 22 1, 222,231, 233 Choudri, M. B., 204, 205,231 Christian, K. R.. 8 I , 99 Clancy, M. J., 7, 99 Clanton, D. C., 87, 101 Clapp, C. E., 196, 197, 1% 208,21 1,217, 23 I Clapperton, J. L., 91,99 Clark, B., 1 I , 105 Clark, H. F., 184,192 Clark, K. W., I 1,99 Clark, V. R., 62,99 Clarke, E. G. C., 61,99 Clarke, M. L., 6 1,99 Clarke, R. T. J., 64,99 Clifford, A. J., 53,99 Cline, M. G., 239, 244, 246, 247, 248, 249, 250, 252, 253, 256, 257, 258, 260, 262, 263, 264, 268,273, 274 Cline, T. R.,53, 99, 185, 192 Coates, W. H., 246, 250,260,262,274 Collazos, C., 180, 194 Comer, G. H., 255, 256,273 Comerma, J. A., 247,252,256,258,273 Compere, R., 84, 101 Compy, E. 2. W., 267, 274, 276 Connolly, J. O., 84, 106 Conrad, H. R., 14, 27, 31, 32, 53, 71, 99, 100 Conway, A., 94, 95, 99 Cook, C. W., 89.90, Y9 Coop, I. E., 62, 91, 99 Cooper, .I.P., 22, 68, 99 Coote, J. N., 64, 102 Corbett, J. I-., 5,48, 83,84,85,91,99,101, 103 Corbett, W. M., 227,235 Cordes, E. H., 174, 192 Cornhill, W. J., 202,207,208,209,231 Cotnoir, L. J., Jr., 221, 233 Couchman, J. F., 70, 101 Cowlishaw, S. J., 83, 99 c o x , c . P., 47, 97 Craggs, 8. A., 22 I , 233 Cramer, D. A., 92, 99 Crampton, E. W., 4, 7, 29, 35, 36, 38, 90, 99,100

326

AUTHOR INDEX

Crawford, W. R., 3 13, 3 15,320 Crick, F. H. C., 175, 192 Cromwell, G. L., 183, 185, 192 Czochanska, Z., 92, 99 D

Daji, J. A., 202, 231 Dalgleish, C. E., 209, 231 Daniels, R. B., 238,240,242,243,246,247, 248, 249, 250, 251, 252, 258, 259, 261, 273,274 Danielson, C. E., 178,192 Dart, P. J., 220,231 da Silva, J. F. C., 29,99 Davey, B. G., 55,99 Davidson, E. A., 223,225,231 Davies, W. E., 54,56,99 Davis, C. L., 45,46,47,48,99 Davis, L. E., 84,99 Davis, R. J., 196, 197,231 Davison, K. L., 64,77,107,108 Dawes, C. J., 220,231 Dea, I. C. M., 228,231 de Freitas, J., 70, I01 de Groot, T., 54,56,99 De Haan, Sj., 80,108 Dehority, B. A., 7, 11, 12, 14, 16, 25, 71, 99,100,101,102 Deijs, W. B., 57, 108 Deinum, B., 23, 24, 100 de Loose, R., 55,69, 100 Demarquilly, C., 17, 18, 29, 70, 7 I , 72,73, 100

De Muelenaere, H. J. H., 182, 192 Denny, C. S., 245, 246, 259, 262,273 Dent, J. B., 288, 309, 321 Dent, J. W., 13, 20, 22, 23, 32, 66, 97, 98, 100

Derbyshire, J. C., 74,100 Deriaz, R. E., 19,88,91,106 Dermine, P., 24, 101 Derting, J. F., 275, 276 Deshpande, T. L., 198,231 Dettmann, M. G., 259,273 De Tyssonsk, E. R., 229, 231 Dew], H., 196, 197, 198, 203, 204, 205, 208, 209, 211, 212, 216, 217, 218, 231, 232,233, 234, 235

Deuel, M., 196, 197, 19H, 204, 208, 212, 216, 217,233 de Wit, C. T., 283,298,299, 304,307,320, 32 I Dias, C., 180, 194 Dick, A. T., 5 8 , 98, 100 Dickson, G. R., 95, 100 Diebold, C. H., 254, 275 Dijkstra, N. D., 6, 100 Dische, Z., 223, 229, 231, 233 Dobie, J. B., 72, 74, 100, 106 Donald, C. M., 284, 291, 292, 293, 294, 315, 317, 320,321 Donefer, E., 4, 29, 35, 36. 38, 99, 100 Donker, J . D., 32, 92, 103 Dormaar, J. F., 199, 204, 207, 209, 232 Dougall, H. W., 54, 100 Doyle, J. J., 90, 101 Drew, K. R., 15, 100 Drysdale, A. D., 18, 96, 99 Dubach, P., 196, 197, 203, 204, 205, 208, 209,212,216,218,231,232,233,234 Dudzinski, M. L., 84,88,92,97,103 Duff, R. B., 197, 207, 208, 209, 214, 216, 217, 219,232,235 Dumanovic, J., 3 14, 320 Duncan, W. G., 289, 292,320 Dungan, G. H., 285, 287, 288, 289, 290, 309, 310, 313, 314, 315.320 Dunham, J. R., 65,100 Dunstone, J. R., 229, 232 Duval, E., 56, 101 E

Earle, F. R., 178, 193 Echols, H., 177, 192 Edlefesn, J., 90, 99 Egan, A. R., 37, 72,100 Eggum, B., 187, 192 Ehlig, C. F., 60, 100 Ekern, A., 5 I , 70, I00 Elder, J. H., 267, 274, 27.5, 276 Ellington, A., 54, 56, 99 Elliott, F. C., 25, 106 Elliott, R. C., 25, 37, 45, / 0 7 Ellis, W. C., 82, 100, 102 El Sayed Osman, H., 50, 97 El-Shazly, K., 11, 25, 100, 104

327

AUTHOR INDEX

Ely, R. A., 7, loo Emerson, R. A.. 180, 192 Emerson, W. W., 198,231, 259, 273 Eng, K. S., 87, 98 Engberg, C. A., 242, 275 Engels, E. A. N., 13, Is, 100 Engleman, E. M., 178, 180, 191 Ennik, G . C., 298, 320 Enzie, F. D., 93, 103 Epley, I. B., 267, 273 Erskine, A. J., 21 I , 232 Ervin, J. O., 196, 197, 220, 221, 233 Essig, H. W., 79, 100 Estermann, E. F., 220, 232 Evans, L. T., 204, 232 Evans, P. S., 34, 100 Evans, R. J., 180, 192 Eveleth, D. F., 13, 106 F Fabry, J., 84, IOZ Faichney,G. J.,45,46,48,52, ZOO Farazdaghi, H., 283, 300, 301, 302, 303, 304, 306, 3 1 I , 320 Farmer, V. C., 197, 219,235 Featherson, J. R., 185, 192 Fehrenbacher, J. B., 242, 243, 248, 251, 253, 255, 273 Finch, P., 202,203,207,208,2 I0,2 1 I , 2 12, 216, 217,221,231, 232 Fisher, L. J., 76, 106 Fitzpatrick, E. A,, 260, 273 Flatt, W. P., 39, ZOO Fleming, G . A,, 54, 56, I00 Flodin, P., 228, 232 Florence, E.. 83, 99 Flowers, R. L., 267, 273 Folkins, L. P., 22, 105 Fontenot, J. P., 17,22,92,98,100 Ford, G. W., 205,206,232 Forsyth, W. G. C., 196, 201,204,208,209, 214, 216, 217, 219, 222,232 Foster, A. B., 2 1 I , 220, 232 Fountaine, E. R., 258,273 Fox, C. J., 267, 273 Fox, C. W., 62, 104 Francis, C. M., 62, 100, 104 Francois, C., 224, 225, 232 Franek, M. D., 229,232 Frankinet, M., 84, 101

Fransmeier, D. P., 248, 273 Fransson, L-A., 229,231 Fraser, A. C., 180, 192 Freer, M., 26, 32, 36, 52, 86, 99, 100, 104 Frey, K. J., 180, 192 Fulkerson, R. S., 22, 104 G

Gaillard, B. D. E., 7, 8, I00 Galberry, H. S., 267, 274 Gallagher, C. H., 63, ZOO Gamble, E. E., 238, 240, 246, 247, 249, 259, 261,273 Gardell, S., 229,231 Garen, A., 174, 177, 192 Garen, S., 177, 192 Garrett, W. N., 74, 100 Garrigus, U. S., 53, 99, 102 Gascoigne, J . A,, 219,232 Gascoigne, M. M., 219, 232 Geoghegan, M. J., 196,232 George, J. M., 56, 62, 97, 102 Gerloff, E. D., 177, 192 Gibbons, R. A., 223, 227,232 Gile, L. H., Jr., 249, 254, 273 Gilfillan, E. W., 172, 179, 192 Gill, W. R., 251, 273 Gladstones, J. S., 58, ZOO Glenday, A. C., 59, 92, 98, 105 Glenn, R. C., 248,273 Gomide, J. A., 29, 99 Gonzalez Gonzalez, V., 87, 100 Goodall, D. W., 292, 315, 320 Goodlett, J. C., 246, 250, 260, 262, 264, 273, 274 Gopalan, C., 171, 189, 192, 193 Goplen, B. P., 65, 103 Gordon, C. H.. 74, 100 Gottschalk, A., 223, 224, 232, 234 Graham, N . McC., 39, 72, 73, 91, 98, 100, 101 Granath, K. A., 228, 232 Graveland, D. N., 200, 204, 232 Gray, F. V., 48, 107 Greacen, E. L., 221, 234 Green, J. O., 66, 67, 101 Green, T. W., 267, 274 Greenhalgh, J. F. D., 5, 33, 83, 85, 92, 99, 101

328

AUTHOR INDEX

Greenhill, W. L., 70, 101 Greenland, D. J., 156, 170, 196, 197, 198, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 213, 214, 215, 216, 217,218,220,221,231,232,235 Greenwood, C. T., 227, 228,231, 232 Grieve, C. M., 29, 101 Griffiths, E., 197, 232 Griffiths, T. W., 48, 101 Grimes, R. C., 47, 87, I01 Grossenbacher, K., 220, 233 Grossman, R. B., 239, 242, 243, 246, 248, 25 I , 252,253,255,257,258,272,273 Grunes, D. L., 56, 101 G ~ p t a U. , C., 196,200, 202, 208,232 H

Hagberg, A., 187,192 Hamilton, F. J., 87,103 Hamilton, H. A., 24,101 Hamilton, J. W., 59,60,101 Hammer, K. C., 220,234 Hammes, R. C., 95,98 Handreck, K.A., 10,34,60,102 Hanel, D. J., 14,107 Hanson, C. H., 62,101 Harada, T., 205,231 Hardison, W. A., 84,95,98, I05 Harker, V. G., 5 , 13,105 Harkness,R. D., 17, 18, 19,101 Harlan, J. R., 94,101 Harmon, A. B., 267,273 Harmsen, G. W., 220,233 Harper,A. E., 182,192 Harpstead, D. D., 184,192,193 Harris, C. E., 5 , 13, 17, 18, 19, 20, 24, 28, 29, 38, 66, 75, 76, 79, 80, 84, 85, 101, 103,104,105 Harris, I . F., 179, 193 Harris, L. E., 90, 99 Harris, P. M., 300,301,302,303, 304,306, 3 11,320 Harris, R. F., 196, 198, 22 1, 222, 233 Hartree, E. F., 224, 233 Harvey, P. N., 315, 316,320 Hatfield, E. E., 53,99, 102 Hawkins, G. E., 25, I01 Haworth, W. N., 196, 224, 233

Hayes, M. H. B., 202, 203, 207, 208, 210, 211, 212, 213, 214, 216, 217, 221, 231, 232 Head, M. J., 26, I01 Heady, H. F., 90, I07 Healy, W. B., 60, 101 Heaney, D. P., 28, 29, 38. 71, 101 Hearn, W. E., 265,273 Hedin, L., 56, 101 Heidelberger, M., 229, 23.3 Heinegard, D., 229, 231 Heintz, R., 174, 193 Helinski, D. R., 174, 192 Hemingway, R. G., 58,103 Hemken, R. W., 72.98 Hennaux, L., 84,101 Henry, E. F., 275,276 Henry, J., 163, I70 Hercus, J. M., 91,101 Hibbs, J. W., 3 1,32,53,99 Hight, G. K., 33,101 Hill, D. E., 255,265,268,273,275 Hill, M. K., 91, 99 Hill, R. L., 178,192 Himes, F. L., 201, 202, 207, 208, 209, 2 12,214,216,2 17,233,235 Hinders, R. G., 7 I , 101 Hiridoglow, M., 24,101 Hirst, E. L., 224,233 Hocevar, B.J.,211,233 Hodge, R. W., 90,101 Hodgson, J., 92,101 Hodgson, J. F., 60,97 Hoehne, 0. E., 87, I01 Hoffman, H., 63,100 Hogan, J. P., 29, 37, 45, 48, 50, 98, 101, 105,107 Hogue, D. E., 5,102 Hole, F. D., 238, 247, 259, 260, 262, 263, 264,274 Holley, R. W., 174,192 Holliday, R., 282, 283, 284, 287, 289, 290, 292, 295, 296, 297, 298, 299, 300, 301, 305, 308,3 10,320 Holmes, J. C., 33,101 Holt, E. C., 94,106 Homb, T., 16,101 Honkanen, E., 65,101 Hopson, J. D., 12,101

AUTHOR INDEX

Horn, M. E., 249, 250, 251, 253, 256, 258, 260,273,275 Horton, D., 204,227,228,233 Hoveland, C. S., 2 9 , 3 I , 97 Howard, M., Jr., 246,272 Howe, E. E., 172. 179,192 Howe, F. B., 265,273 Howe, W. M., 53.98 Hozumi, K., 291, 292, 294, 301, 302, 310. 317,320 Huddleston, J. H., 254,255,264,265,273 Huddleston, J. S., 245,273 Hudson. H . G., 287,297,320 Hughes, M., 8 2 , 9 7 Hughes, R., 22,107 Hull, J. P. D., 266,273 Hulyalkar, R. K., 225,234 Humbert, R. P., 248, 251, 252, 254, 256, 259,260,274 Hungate, R. E., 49, 102 Huston, J. E., 81. 102 Hutcheson. T. B., Jr., 246. 247, 248, 256. 2 73 Hutchings, R. J., 63, 104 Hutton, E. M., 64,102 Hutton, J. B., 3 I , 32,85,102 I

Ignatieff, V., 113. 148,169 Ingalls, J. R., 3,35,102 Ingleton, J. W.. 84,104 Ingram, V. M., 175,192 Irvin, H. M., 40,104 Ivarson, K. C., 200,225,233,235 J

Jackobs, J. A., 289,320 Jacob, F., 176, 177.192 Jacobson. W. C., 74,8 I , 82,100,102 Jacoby, Erich H.. 135,169 Jager, G., 220,233 Jahn, J. R., 72,102 James, W. H., 106 Jansen, G. R., 172, 179,192 Jaques, L. B., 65, 103 Jarrige, R., 7,102 Jarvis, R. H., 292,310,320 Jayrne, G., 202,233 Jeanloz, R. W., 223,228,233

329

Jenny, H., 220,233 Jensen, E. H., 88,90,106 Jermyn, M. A., 229,232 Jha, P. P., 246, 247, 248, 249, 250, 252, 256,260,262,273,274 Jimenez,J. R., 179, 185, 186,192 Johansen, P. G., 224,225,233 Johns, A. T., 64,102 Johnson,HarryG., 118, 121, 170 Johnson, J. C., 267,274,276 Johnson, R. R., 7, I I , 12, 14, 16, 25, 52, 53,71,79,99,100,101, 102, 106 Johnston, T. D., 34,102 Jones, D., 197,220,232,233 Jones, D. I. H., 22.59.97.107 Jones, G. B., 48,107 Jones, H. E. H., 62,105 Jones, J. H., 94,106 Jones,J. K. N.,2I 1,232 Jones, J. R., 5,102 Jones, L. H., 306,307,320 Jones, L. H. P., 10,34,60,97, 102, 107 Jones,Q., 178,180, 193,194 Jones, U. I., 3 I , 102 Jordan, R. M., 84.99 Josefsson, E., 190,192 Journet, M.. 7 I , 72, 73, 100 Julen. G., 68, 102 Jung. G. A., 1 I , 14. 29, 33. 47, 105 Jurion, F., 163,170 Jury, K. E., 85, 102 K

Kahrein, R. B., 267, 274 Kamstra, L. D., I I , 72, 102 Kane, E. A,, 17,23,8 1,82,102 Kapelle, D., 80. 108 Karlsson, K.-E., 187,192 Karn, J. F., 1 I , 102 Karr, M. R., 53.102 Kates, K. C., 93,103 Kaufmann, R. W.. 84,105 Kay, M., 36,52, 102 Keane, E., 76,103 Keefer, R. J ., 20 I , 207,208.2 I 6 , 2 19,233 Keller, K. R., 288, 3 10,320 Kelley, E. G.. 177,192 Kellogg, Charles E., 112, 142, 149, 155, 156, 161,170, 267.272

330

AUTHOR INDEX

Kernp, A., 55,56,57,102,108 Kernp,A. W.. 84,85,105 Kemp, C. D.,49,84,85,103,105 Kendall, W. A., 64,102 Kennedy, G. S., 57,102 Kennedy, W. K., 16,17,23,105 Kenner, J., 205,233 Kertesz, Z. I., 228,233 Kewning, J. A., 54,56,99 Kilmer, V. J., 244,275 Kim, T., 291,292,293,294,295,296,298, 299, 300, 301, 302, 305, 306, 310, 317, 3 18,320,321 Kirby, K. W., 197,204,207,208,2 I I , 2 I 6, 217.2 18,220,235 Kivimae, A., 5,102 Knapp, John, 110, 170 Knight, R., 68,102 Knolek, W. F., 180,194 Knox, E. G., 239, 247, 248, 252, 256, 257, 258,259,272, 274 Koch, J. H., 63,100 Kohler, G. O., 62,104 Kononova, M. M., 204,219,233 Korner, A., 174,192 Koyama, H., 302,3 18,320 Kozak, A. S., 53,104 Krusekopf, H. H., 243,257,274 Kubota, J., 60,100 Kvist, 8.E., 228,232 1

Laby, R. H., 63,102 Lacey, J., 70,102 Lackman, D. B., 209,234 Lager, A,, 68,102 Lambourne, L. J., 59, 84, 8 5 , 86, 87, 91, 97,100,102 Lamond, D. R.,61,62,97,104 Lancaster, R. J., 33,84,IOI, 102 Landry, J., 185,193 Lang, A. L., 285, 287, 288, 289, 290, 313, 3 14,320 Lang, F., 202,233 Lang, R. W., 33, I01 Langlands, J. P., 56,84,86,88,89,91,102 Lapham, J. E., 266,274 Lapham, M. H., 266,274 Large, R. V., 92,93,94,102,106 Lauterbach,C. W.,211,235

Leavenworth, C. S., 179,193 Lees, H., 204,231 Lefevre, C. F., 11,102 Leigh, J. H., 90, 102 Leighty, R. G., 267,273 Leighty, W. J . , 267,274 Leitch, I., 5 , 102 LeMare, P. H., 159, 170 Leng, R. A,, 48, 103 Lesperance, A. L., 88,90,106 Levy-Bruhl, L., 168, I70 Lewis, D., 4 t , 9 7 Lewis, H. G., 265,273 Lewis, R. J., 246,247,248.273 Li, J. C. R., 288,3 10,320 Liener, I. E., 189,192 Lima, I. H., 177,192 Linares, F., 184,193 Lindahl, I. L., 93,103 Lindberg, B., 228,231 Lindner, H. R., 61,103 Lindstrom, G. R., 198,232 Linton, J. H., 65,103 Little, C. O., 8 I , 103 Lloyd, L. E., 4,29,35,36,38,99 Lofgreen, G. P., 74,100 Logan, V. S., 5,105 Lombard, P. E., 87,103 Loosii, J. K., 53, 104 Loper, G. M., 62,103 Love,T. R., 267,274 Lowe, C. C., 18,103 Lowe, L. E., 222,233 Ludwig, T. C., 60,101 Lugg,J. W. H., 177,192 Lusk, E., 261,273 Lusk, J. W., 12,103 Lyford, W. H., 245, 246, 250, 254, 259, 260,262,264,273,274 Lynch, D. L., 197,200,202,204,207,209, 211,221,231,232,233,235 Lynch, J. J.. 56,102 M

McArthur, J. M., 63,64,103 McCalla, T. M., 196,233 McCarrick, R. B., 51,76,103 McClure, K. E., 52,53,79,102,106 McCracken, R.J., 238, 242,243.246.247,

AUTHOR INDEX

248, 249, 250, 251. 252. 258, 259, 261, 273,274 McCroskey, J . E., 73, 103 McCullough, M. E.. 28. 45, 46, 47, 76, 99, 103 McDonald, A. N. C., 82.97 McDonald, I., 5,83,85,91,99,101,102 McDonald, I . W.. 41, 49, 50, 52, 55,56,58, 59,60,61,8 1,91,103 McDonald, P., 74,75,103, 105 McDougall, B. M., 220,234 McGowan, M., 13, 15.97 McGuire, R.L., 81,103 Mcllroy, R.J., 24,47,103 McKeague, J. A., 257,258,274 McLaren, A. D., 220,232 MacLean, D. W., 246,262,264,274 MacLeod, N. A,, 36,52,102 McLuhan, M., 168,170 MacLusky, D. S., 92,103 McManus, W. R.,84,87,103 McMeekan, C. P., 94,103 McMichael, S. C., 190,192 McNaughton, M. J., 62,105 McNutt, E. J., 267,274 Macpherson, A., 5 8 , 103 MacPherson, H. T., 75,103 Maguire, M. F., 49,104 Mahler, H. R., 174,192 Malgren, R.C., 167,169 Marbut,C. F., 257,266,267,274 Margoliash, E., 178,192 Marshall, R. D., 204, 223, 224, 225, 232, 233,234 Marshall, S. B. M., 50.5 I , 97 Marten. G. C., 32,84,92,99, 103 Martin, C. M., 84,105 Martin, J. K., 205,233 Martin, J. P., 196, I97,2 19,220.22 I , 233 Martin, Kirk, 110, 170 Martz, F. A., 53.98 Mathieson, J. M., 229,234 Matthews, E. D., 267,274, 276 May, P. F., 87,101 Mayaudon, J., 2 19.233 Maynard, L. A., 7,90,99 Mead, R.,3 18,320 Meek, D. C., 106 Mehta, N. C., 196, 197, 198, 203, 204,205,

33 1

208, 211, 212, 216, 217, 232, 233, 234, 235 Mellin, T. N., 17,103 Melville, J., 2,103 Melvin, J. F., 70,103 Mendel, L., 179,193 Mercer, F. V., 220,231 Meredith, W. R., 18,103 Merilan, C. P., 53.98 Mertz, E. T., 173, 178, 180, 181, 182, 183, 188,191,192,193 Methvin, C., 267,273 Meyer, J. H., 74,100 Meyer, R. M., 73,107 Meyers, S. M., 184,192 Miles, D. G., 29,103 Miles, J. T., 12,103 Milford, R., 17, 18, 19, 20, 28, 29, 36, 66, 103,104 Milfred, C. J., 263,274 Miller, F. P., 248,249,250.25 1,274 Miller, P. A., 190,193 Miller, R. W., 180,194,257,258,274 Miller, T. B., 7,9,103 Millett, M. A., 225,234 Millington, A. J., 62,100 Miltimore, J. E., 63,64,103 Minson, D. J., 5 , 7, 17, 18, 19, 20, 24, 28, 29, 36, 38, 49, 66, 71, 72, 81, 82,85,10I, 102,103,104,105 Mitchell, R. L., 55,58,99, 104,225,234 Mitscherlich, E. A., 291,320 Moe, P. W., 13, 39, 104 Moggridge, R. C. G., 225,233 Mohan, V. S., 189, 193 Moir, R. J., 37, 59, 72, 100, 104 Moisio, T., 65, 101 Mollenhauer, H. H., 220, 233 Monod, J.. 176, 177, 192 Montgomery, E., 305, 320 Montgomery, R., 223, 224,235 Moore, L. A., 7, 8, 17, 23, 40, 71, 74, 75, 76,8 I , 82,100, 102, 104,107 Moore, R. M., 63,104 Moore, W. E., 225, 234 Morley, F. W. H., 62, 104 Morre, D. J., 220, 233 Morrison, F. B., 4, 104 Mortensen, J. L.. 201, 202, 207, 208, 209,

AUTHOR INDEX

332

210, 211, 212, 214, 216, 217, 218, 219, 233, 235 Morton, R. K., 175, I 9 3 Moseman, A. H., 119, 170 Mosse, J., 179, 185, 186, 187, 188, 193 Mott, G. O., 11, 29, 30, 31, 94, 99, 104 Mottershead, B. E., 29, 31, 106 Moule, G. R., 61, 104 Moureaux, T., 185,193 Mowat, D. N., 22, 68, 104 Moxon, A. L., 11,105 Mueller, 0. P., 262, 263, 264, 274 Mulham, W. E., 90, 102 Muller, M., 197, 198, 204, 208. 21 I , 212, 216,217,233, 235 Muller-Vonmoos, M., 205, 234 Munck, L., 173,193 Mundie, C. M., 200, 225,231 Munro, H. N., 173, 193 Murdoch, J. C., 37, 77, 9 9 Murphy, R. P., 16, 17, 23, 105 Murphy, W. E., 54, 56, 100 Murray, M. G., 48,98 Murray, S., 14, 105 Myrdal, G., 129, 170 N

Naga, M. M. A., 11,104 Nagarajan, V., 189,193 Naismith, W. E. F., 188, I 9 3 Nance, W. E., 177,193 Nash, M., 168,170 Nash, M. J., 5 1,70,74,76,79,80,107 Neeley, J. A., 242,243, 245,250,274, 278, 279 Neely, W. B., 229, 233 Nehring, K., 189, 193 Nelder, J. A., 292, 299, 300, 301, 302, 303, 306,310, 311, 312,320 Nelson, A. B., 87, 98 Nelson, 0. E., 173, 180, 181, 182, 183, 185, 193 Nettleton, W. D., 238, 242, 243, 246, 247, 248, 249, 250, 251, 252, 254, 258, 259, 26 1,273, 274 Neuberger, A., 204, 223, 224, 225, 232, 233,234 Neumark, H., 75,104

Newton, J. E., 62,104 Newton-Hearn, P. A., 2 1 1,232 Nijkamp, H. J., 8,43,98,100 Nikiforoff, C. C., 248, 249, 25 I , 252, 254, 256, 257, 259, 260,274 Noller, C. H., 14, 104 Norman, A. G., 7,104 Northcote, D. H., 2 I I , 225, 228, 231, 233, 234 Norton, H. W., 53, 102 Nye, P. H., 156,170 0

Oades, J. M., 196, 197,200,201,202,203, 204, 205, 206, 207, 208, 209, 210, 211, 213, 214, 215, 216, 217, 218, 220, 222, 226,232,234,235 O’Donovan, P. B., 29, 30,31, 104 Ogawa, H., 291, 292, 293, 294, 302, 3 17, 3 18,320 Ogston, A. G., 210, 218, 234 Okamoto, M., 75, 76, 107 Olfield, J. E., 62, 104 Olney, H. O., 200,202,207,2 11,233 Olson, G. W., 238,246, 247,248,25 1,252, 254, 255, 259, 260, 262, 263, 264, 265, 273, 274 Oltjen, R. R., 53, 104 O’Neal, A. M., 255, 274 Ongun, A., 177, 193 Oram, L. R. N., 63,104 Orth, A,, 77, 104 Ory, R. L., 178, 180,191, Osborne, T. B., 178, 179,193 Osbourn, D. F., 29, 30, 32, 34, 36, 77, 78, 101, 104 Oser, B. L., 189, 193 O’Shea, J., 13, 49, 104 Oslage, H. J., 80, 106 O’Sullivan, M., 57, 104 Overend, W. G., 224, 234 Owen, F. G., 71, 101 Owen, J. B., 84, 104 P

Packett, L. V., 29, 30, 31, I04 Pacsu, E., 235 Page, H. J., 113, 148, 169

AUTHOR INDEX

Paladines, 0. L., 73, 104 Parsons, J. W., 200,203,207,209,2 I I , 2 I 4, 216, 217, 222,234 Pauker, G. L., 169, 170 Paul, E. A., 197, 199, 204, 207, 209, 21 I , 230 Paulson, G. D., 60, 104 Pearce, G. R., 86, 104 Pearce, R. H., 229, 234 Pearson, C. S., 268, 274 Pelzer, H., 220, 235 Pendleton, J. W., 285, 287, 288, 289, 290, 309, 310, 313, 314, 315,320 Pendleton, R. F., 275, 276 Penick, M., 314, 320 Penny, D. H., 129,170 Percival, E., 225, 234 Perera, B. P. M., 209, 234 Perez, C. B., 53,104 Perkins, S. O., 266,274 Perold, 1. S., 5 8 , 107 Perry, M. B., 225.234 Peters, J. E., 180, 194 Peterson, P. J., 60, 105 Pettiet, J. V., 247, 249, 25 1, 255, 26 I , 274 Pfost, H. B., 73, 107 Phillips, P. H.. 73. 108 Pickett, R. A., 183, 184, 192, 193 Pigden, W. J., 5 . 22, 28, 19, 38, 71, 101, I05 Pigman, W., 224,228,234 Pilgrim, A. F., 48,107 Pinkard, R. W., 196,233 Piva,G., 183,193 Platt, D., 228,234 Playne, M. J., 74,105 Plice, M. J., 32,105 Plumlee, M. P., 29, 30.3 I , 104 Pope, A. L..60,73,104,108 Pope, G . S., 62, I05 Pope, L. S., 73,87,98,103 Porter, H. C., 275,276 Poulton, B. R., 17,103 Pradilla, A., 184,192, I 9 3 Pratt, A. D., 3 1,32,99 Prestes, P. J., 14,104 Presthegge, K., 74,75,105 Pritchard, G. I., 5 , 22, 28, 29, 71, 101, 105

333

Proffitt, W. H., 267, 274 Puckridge, D. W., 3 15, 321 Putter, J., 281, 321 Q

Quicke, G. V., 1 I , 105 Quirk, J . P., 197, 198,231,232,235 R

Rae, A. L., 92,105 Rahman, S., 228,231 Raison, J. K., 175,193 Raymond, W. F., 2, 3, 5 , 13, 14, IS, 17, 18, 19, 20, 22, 24, 27, 28, 29, 31, 38,44,45, 48, 66, 68, 69, 74, 75, 76, 79, 80, 81, 82, 84, 85, 86, 87, 94, 95, 96, 99, 101, 103, 104, 105 Reardon, T. F., 85, 86, 91, 102 Reddy, M. C., 6 5 , 1 0 0 Redmond, C. E., 242,275 Reed, W. D. C., 25, 3 7 , 4 5 , 1 0 7 Rees, C. W., 224,234 Reestman, A. J., 307,321 Reichstein, T., 224, 234 Reid, C. S. W., 63, 105 Reid, G. W., 33, 91, 92, 101, 102 Reid, J. T., 13, 16, 17, 18, 23, 39, 73, 80, 8l,84,103,104,105 Reid, R. L., 11, 12, 14, 33, 37, 45, 47, 98, 105, 106 Reid, R. S., 48, 98 Reissig, H., 189, 193 Reith, J. W. S., 58, 104, 105 Rennie, D. A., 197, 199,204,207,209,2 I 1, 2 18,230,234 Reynolds, J . D., 315, 316, 321 Rhykerd, C. L., 14, 104 Rhyne, C. L., 190, 193 Richards, C. R., 84, 106 Richards, F. J., 299, 302, 321 Richards, G. N., 205,233 Richards, S. J., 196, 197, 220, 233 Ridley, J. R., 88, 90, 106 Riewe, M. E., 94, 106 Ritossa, F. M., 176, 193 Robards, G. E., 87, 106 Roberts, R. C., 188, 193, 244,275 Roe, R., 29, 3 1, 106

334

AUTHOR INDEX

Rogers, H. H., 66, 68, 106 Rogers, H. J., 220, 223,234 Rogler, J. C., 185, 192 Rohr, K., 46,47, 107 Ronning, M., 72, 106 Rook, J. A. F., 41,44,45,76,98,106 Rose, R. C., 266, 275 Rosenfeld, I., 59, 106 Rossiter, R. C., 62, 106 Roulet, N., 205, 209, 2 12, 234 Rovira, A. D., 220,22 1,234 Rowland, S. J., 47, 97 Rudman, J. E., 20, 106 Rumsey, T. S., 14, 104 Russell, E. W., 219, 234 Rutledge, E. M., 249, 250, 251, 253, 256, 258, 260,273,275 S

Sadgopal, A., 174,193 Saeman, J. F., 225,234 Salam, A., 204, 205, 235 Salamini, F., 183, 193 Salomon, M., 199, 209, 234 Salton, M. R. J., 220, 234 Samtsevich, S. A., 219, 220, 234 Sandegren, E., 178, I 9 1 Sanderson, G. W., 209,234 Santi, E., 183,193 Sarria, D., 184, 192 Sastry, L. V. S., 187, 194 Satter, L. D., 50, I06 Saunt, J., 284, 321 Sawers, D., 5 1, 70, I00 Schaadt, H., 52, 53, 106 Schillinger, J. A., 25, 106 Schmid, K., 223,234 Schneider, 9. H., 5, 106 Scholl, J. M., 5 , 13, 97 Schukking, S., 80, 108 Schultz, I. W., 170 Schulz, E., 80, 106 Schuphan, W., 173,193 Schweet, R., 174, 193 Schwendinger, R. B., 201, 202, 208, 210, 218,233 Schwerdtfeger, E., 189, 1 9 3 Scott, F. M., 220,234

Scott, H. W., 11, 16, 100, 105 Scott, J. E., 2 1 I , 228, 234 Scragg, W., 315, 316,320 Scrivner, C. L., 243,248,249,260,275 Seay, W. A., 246,247,248,273 Seif, R. D., 3 15,320 Sevag, M. G., 209,234 Sequeira, J. S., 224,234 ShalTer, M. E., 249,260,275 Shafizadeh, F., 224,234 Shantz, H. L., 244, 275 Sharpe, P. R., 288, 309,321 Shaw, J. C., 40,104 Shaw, K., 221,231 Shearin, A. E., 268, 275 Shefner, A. M., 226,231 Shelton, D. C., 11, 12, 105, 106 Shepherd, R. A., 196, 197. 220, 221,233 Sheppard, D. E., 190, 193 Shepperson, G., 70, 106 Shibles, R. M., 315, 321 Shinozaki, N., 292, 293,294,295,296,298, 299, 300, 301, 302, 305, 306, 317, 318, 320,321 Shorland, F. B., 92, 99 Shumard, R. F., 13,106 Simmonds, R. G., 202,203, 207, 208, 210, 21 I , 2 12,2 13,214,2 16,217,232 Simonart, P., 219,233,235 Simonson, R. W., 237, 238,275 Sinclair, D. P., 33, 101 Siu, R. G . H., 21 9,235 Slack, S. T., 16, 17, 23, 105 Sly, D. A., 188, 194 Slyter, L. L., 53, 104 Smart, C. L., 227,235 Smart, V. W., 190, 193 Smart, W. W. G., 25, 45, 46, 47, 203,106 Smith, A. K., 188, 194 Smith, C. A., 19, 24, 106 Smith, C. R., Jr., 178, 193 Smith, E. L., 178, 192 Smith, F., 223, 224, 235 Smith, F. H.,190, 193 Smith, G. D., 261, 267,275 Smith, H. C., 266, 267, 273, 275 Smith, J. C., 94, 106 Smith, R. M., 252, 255, 256, 262, 265,275

335

AUTHOR INDEX

Smolens, J., 209,234 Snow, C. P., 168,170 Sonneveld, A,, 33, 106 Sowden, F. J., 196, 200, 202, 208, 225, 232, 233, 235 Spaeth, J. N., 254,275 Spedding, C. R. W., 3, 13, 24, 92, 93, 94, 102, 105,106 Spedding, D. J., 60, 105 Spiegelman, S., 176, 193 Spiro, R. G., 223, 225, 235 Sprague, G. F., 180, 192 Srikantia, S. G . , 171, 192 Stacey, M.,196, 202, 203, 207, 208, 210, 211, 212, 213, 214, 216, 217, 220, 221, 223,231, 232, 233, 235 Stahmann, M. A,, 172, 177, 192, 193 Stanley, N. W., 25, 106 Stephen, I., 248, 25 1, 273 Stephens, D. F., 73, 103 Stern, W. R., 3 17,321 Stevenson, F. J.. 204,205,231 Stevenson, 1. L., 219, 235 Stewart, D. R. M., 91, 106 Stobbe, P., 200, 232 Stocking, C. R., 177, 193 Stoddart, J. F., 228,231 Stojanovic, B. J., 197, 198, 235 Stout, P. R., 56, 57, 101, 106 Streeter, C. L., 87, 101 Streuli, H., 197, I98,2 I I , 233,235 Sullivan, J. T.. 6, 7, 9, 98, 106 Swaby, R. J., 196,235 Swift, R. W., 106 Swincer, G. D., 196, 197, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 213, 214, 215, 216, 217, 218, 220, 222,234, 235 Sykes, J. F., 74, 75, 76, 104, 107 Synge,R. L. M., 50, 51,97, 106 Szirmai, J . A., 229, 231 T

Tadmor, A., 75, 104 Tavernier, R., 26 1 , 275 Tayler, J. C.. 19.20,88.91,106 Taylor, D. C., 268, 275 Taylor, M. W., I I, 98

Tello, F., 189, 193 Terry, R. A., 8, 12, 14, 15, 19, 20, 21, 22, 25, 26, 29, 30, 31, 32, 34, 36, 37,42,45, 4 8 , 6 8 , 7 1,99, 104, 105, 106, 107 Tesar, M.,3,35,102 Theron, E. P., 34,107 Thomas, A. E., 242, 243, 267, 274, 275, 277,278 Thomas, C. J., 265,275 Thomas, G. D., 73,107 Thomas, J. W., 3, 35, 14, 75, 76, 102, 104, 107 Thomas, R. L., 200, 201, 207, 208, 209, 2 1 I, 2 12,214,2 16,2 17,235 Thompson, A., 224,235 Thompson, R., 285,286,300,307,'320 Thomson, D. J., 29, 30, 32, 34, 36, 42, 47, 104, 107 Thorne, J. L., 90, 99 Thornton, R. F., 24, 107 Thorp, J., 267,272, 275 Threlkeld, G., 275, 278 Tilley, J. M. A., 8, 12, 14, 15, 19,20,21,22, 25, 26, 3 I , 36, 37, 42, 45, 68, 7 I , 99, 106, 107 Tillman, A. D., 53,99,106 Tinsley, J., 200, 203, 204, 205, 207, 209, 21 I , 2 14,2 16,217,220,222,234,235 Tobin, J., 76,103 Toogood, J. A,, 197,235 Tookey, H. L., 63,108 Topps, J. H., 25,37,45,107 Torell, D. T., 87,90,97,98,107 Torriani, A., 177, I92 Tossell, W. E., 22, 104 Trimberger, G. W., 16, 17, 23, 105 Troelsen, J. E., 14, 24, 34, 102, 107 Truog, E., 197, 199, 204, 209, 218, 234 Tuckett, S. E., 184, 192 Turk, K. L., 16, 17, 23, 105 Turner, J. H., 93, 103 Tyler, M. C., 267, 274 Tyrrell, H. F., 13, 39, 104 U

Ulyatt, M. J., 47, 76, 98, 107 Underwood, E. J., 54, 57, 5 8 , 107

336

AUTHOR INDEX

V

Vanden Berg, G. E., 25 1, 273 Vanderford, H. B., 249, 260,275 Van der Merwe, F. J., 13, 15, 58, 100, 107 Van Dyne, G. M., 8 7 , 9 0 , 1 0 7 van Es, A. J. H., 23, 24, 43, 98, I00 VanEtten, C. H., 180,194 Van Niekerk, B. D. H., 73, 104 van Schalkwyk, A., 87,103 Van Soest, P. J., 6, 8, 9, 10, 13, 23, 24, 25, 29, 30, 32, 34, 49, 100, 107 Vercoe, J. E., 86, 104 Veron, 0. A., 181, 183, 185, 193, 194 Vetter, R. L., 13, 22, 107 Viets, F. G., 283, 320 Vinas, E., 180, 194 Virtanen, A. I., 53, 107 Virupaksha, T. K., 187, 194 Vogel, H. J., 176, 194 Vogel, R. H., 176, 194 von Kaufmann, W., 46, 47, 77, 104, 107

W

Wainman, F. W., 27, 28, 32, 43, 98 Waite, R., 39, 97 Walborg, E. F., 224, 235 Walker, J. L., 248,273 Walker, T., 36, 52, 102 Waller, G., 73, 103 Walshe, M. J., 94, 103 Walters, R. J. K., 22, 29, 55, 97, 103, 107 Ward, D. N., 224,235 Ward, G., 65, I00 Wardrop, I. D., 60, 9 7 Warne, L. G. G., 284, 292, 293, 294, 321 Warner, A. C. I., 41, 48, 107 Warner, R. G.,53,77,104,107 Watkin, B. R., 87,101 Watson,J. H., 197, 198,235 Watson, J. N., 18, 96, 99 Watson, S. J., 51, 70, 74, 76, 79, 80, 107 Watt, J. A., 75, 103 Watts-Tobin, R. J., 175, 192 Waugaman, S. N., 196, 197,231 Weaver, H. S., 84, 106 Webber, J. M., 223, 228,231

Webber, L. R., 197,235 Weber, C. R., 315,321 Webley, D. M., 196, 197, 219, 232, 235 Wedin, W. F., 13, 22, 107 Weed, S. B., 25 1,274 Weenink, R. O., 63, 102 Weidel, W., 220, 235 Weir, W. C., 87, 90, 97 Weiss, E., 224, 234 Welch, J. A., I I , 105 Weller, R. A,, 48, 107 Weston, R. H., 29, 36, 48, 1 0 7 Whelan, W. J., 224, 235 Whistler, R. L., 197, 198, 204, 205, 207, 208, 211, 216, 217, 218, 220, 227, 228, 235 White, H. S., 180, 194 White, J. L., 248, 252, 258,272 White, P. L., 180, 194 Whitehead, D. C., 54, 108, 200, 203, 207, 209, 220,235 Whitelaw, F. G., 48, 98 Whiteside, E. P., 247, 248, 249, 250, 251, 253,255,256,257,260,261,275 Whitmore, E. T., 68, 106 Whitmore, G. E., 93, 103 Whittig, L. D., 244, 275 Whybrew, J. E., 315, 316,320 Wichser, W. R., 184, 194 Wieringa, G. W., 80, 108 Wiggans, R. G., 3 15, 3 16,321 Wilcox, 0. W., 288,321 Wilkins, R. J., 89, 108 Willey, R. W., 285, 307, 308, 31 I , 312,321 Williams, 9. G., 197, 235 Williams, E. E., 53, 104 Williams, J. D., 63, 104 Williams, L. D., 267, 274 Williams, V. J., 81, 99 Wilson, A. D., 87, 106 Wilson, C. M., 175, I 9 4 Wilson, I. A. N., 93, 106 Wilson, R. F., 76, 80, 101 Wilson, R. K., 7, 13, 99, 104 Wilson, R. S., 27, 28, 31. 32, 37, 98 Winch, J. E., 22, 104 Wind, J., 57, 108 Winogradzsky, S., 196,235

337

AUTHOR INDEX

Winter, K. A., 5 , 105 Winters, E., 237, 238, 257, 265,275 Winzler, R. J., 223, 235 Wiseman, H. G., 74, 100 Wolf, W. J., 188, 194 Wolff, I. A., 178, 180, 193, 194 Wolfrom, M. L., 204, 224, 227, 228, 229, 233, 235 Woods, A. E., 65, 108 Woodward. F. N.. 202, 207, 208, 209,231 Woolfolk, P. G., 84, 105 Worker, N . A., 6 5 , 108 Wright, L. M., 200, 202, 207, 2 I I , 22 I , 233 Wright, M. J., 64,108 Wright, P. L., 73,108 Wyatt, C. E., 267,274

Y

Yamamura, Y ., 184, I 9 2 Yanofsky, C., 174,192 Yaron, D., 281,321 Yassoglou, N . J . , 247, 248, 249, 250, 25 1, 253,255,256,257,260,261,275 Yates, N . G., 24, 63, 68, 102, 107, 108 Yatsu, L. Y., 178, 180, 191 Yoda, K., 302,320 Young, M. C., 90, 99 Z

Zimmerman, R. C., 255, 256, 273 Zon, R.,244,275 Zuckerkandl, E., 177, 194 Zweifel, G., 197, 204, 208, 216, 218, 232

Subject Index

A

D

Dactylis glomerata, 17 Daucus carota, 180 Digitaria decumbens, 36

Alfalfa, 156 nutritive value, 7,45,46 Alfisol, 245 Alluvial soil, 152-153 Aluminum, 258 Avena sativa, 179 Azotobacter indicus, 220 Azuki bean, 291

E

Ensilage, 74-80 Estrogenic compounds, forage crops, 6 1-63 Feed, digestibility coefficient, 4 supplements, 25-27,37-38 voluntary intake, 27-38,7 I , 74-79 Festuca arundinacea, 17, 29, 33 Festuca pratense, 17, 5 5 Forage crop, nutritive value, 1-108 Fragipans, 237-279

6

Bacteria, soil, 2 19 Barley, 179, 187, 188,298 Birdsfoot trevoil, 34,64 Brassica campestris, 190 Bromus inermis, 32

G C

Calcium, 5 5 , 57, 65, 149, 155, 161, 245 Carrot, 3 18 Cellulose, nutritive value, 8, 13, 14,77,89-90 soil, 197,219 Chenopodium pallidiculae, 180 Chenopodium quinoa, 180 Chernozem soil, 145-146 Chestnut soil, 146 Chitin, 221 Chlorophora excelsia, 149 Chromobacterium violaceum, 220 Clay, 198,247 Clover, 16,89 Cobalt, 57,63 Cocksfoot, nutritive value, 17, 18, 20-21, 23, 24, 32, 68 Copper, 54, 58 Corn, 179 see also maize Cornilla varia, 64 Corn silage, 46, 79 Cottonseed meal, 190 Crambe abyssinica, 180 Crownvetch, 64

Genetics, plant protein, 17 I - 194 GIycine max, 180 H

Hordeum vulgare, I79 I

lmperata cylindrica, I58 Inceptisol, 245 Iodine, 59 Iron, 247, 258 1

Ladino clover, 65 Lathyrus sativus, 189 Latosols, 148-151, 155 Lead, 58-59 Leaf proteins, 177-178 Legume, 180, 189 nutritive value, 12, 34, 5 1, 5 5 , 56, 62, 7 1, 74 Lespedeza cuneata, 25 Lignin, 7, 8, 13, 14, 89 Lithosols, 148, 152, 153 L o h m spp., 17, 18 Lotus corniculatus, 34, 64

338

SUBJECT INDEX

Lucerne, nutritive value, 17, 22,25, 30, 34. 36,53,62,63,7 1-72

M

Magnesium, 56-57. 247 Maize, see also corn, 22, 179, 180-187, 188.285 Marah gilensis, I80 Marrow-stem kale, 22 Meadow fescue, 55 Medicago sativa, 17, 20 Melilorus alba, 65 Millet, 179, 188 Molybdenum, 58 Montrnorillonite, 246, 248 N

Narrow-stem kale, 22 Nitrate, 64-65 Nitrogen, 15, 20, 23, 24, 32, 36-37, 47, 50-5 I , 56 0

Oats, 179, 187, 298 Oat straw, 36 Onobrychis vicifolia, 22 Oryza sativa, 179 P

Pangola grass, 36 Panicum milliaceum, 179 Pennisetum purpureum, 156 Phalaris arundinacea, 29, 3 I , 34 Phalaris tuberosa, 3 1, 63, 69 Phaseolus chrysanthus, 29 I Phelum nodosum, 29 Phleum pratense, 17, 29, 55 Phosphate, 65 Phosphorus, 55, 155, 159 Planosol, 267 Plant density, yield, 281-32 1 Plant protein, genetics, 171- I94 Poa pratensis, 55 Podzol, 145-148, 151, 268 Potassium, 56 Potato, 189, 304

339 R

Red clover, 22, 34, 65 Reddish Chestnut soil, 145-146 Regosols, 149, 153 Rice, 179, 187 Rye, 179, 188 Rye grass, nutritive value, 7, 17, 18, 2021, 22, 23, 24, 30.43.47, 54, 55,65, 66, 70, 89, 93 S

Sainfoin, 22, 34 Secale cereale, 179 Seed, storage proteins, 178- 180 Selenium, 59-60 Sericea lespedem, 25 Sierozem soil, 146- 147 Silica, 10-1 1, 60-61, 257 Sodium, 55 Solannum malacoxylon, 65 Soil, classification, 265-269 erosion control, 138-140 fragipans, 237-279 group descriptions, 145-1 53 horizons, 240-242 polysaccharides, 195-235 potentially arable, 109-170 surveys, 123- 130 use, 112-122, 156-163 Sorghum, 179, 187 Sorghum vulgare, 179 Soybean, 180, 188, 3 16, 3 I8 Spodosol, 245, 269 Subterranean clover, 6 1-62, 69, 293 Sulfur, 52, 59 Sweet clover, 65, 69, 156 7

Tall fescue, 17, 18, 29, 63 Terra Rossa soil, 151-152 Timothy, nutritive value, 16, 17, 29, 30, 55

Trqolium pratense, 22 Trfolium repens, 17 Trifolium subterraneum, 63 Triticum vulgare, 179, 187

SUBJECT INDEX

Tropical soil, 156- I63 Turnip, 3 10 U

control, 130-135 fragipans, 254-256, 258, 264 Wheat, 179, 188, 285, 31 1 White clover, nutritive value, 17, 47, 65

Ultisol, 245 Urea, 26, 36, 5 1, 52, 53 V

Y

Yield, plant density, 281-321

Vetch, 25 2 W

Water, 24

Zea mays, 22, 179 Zein, 179, 180-187