Advances in Agronomy, Volume 49

T

I

Lgronomy

VOLUME 49

Advisory Board Martin Alexander

Eugene J. Kamprath

Cornell University

North Carolina State University

Kenneth J. Frey

Larry P. Wilding

Iowa State University

Texas A&M University

Prepared in cooperation with the A m ' c a n Society of Agronomy Monographs Committee S. H. Anderson P. S. Baenziger L. P. Bush

M. A. Tabatabai, Chairman R. N. Carrow W. T. Frankenberger, Jr. S. E. Lingle

R. J. Luxrnoore G. A. Peterson S. R. Yates

D V A N C E S IN

VOLUME 49 Edited by

Donald L. Sparks Department of Plant and Soil Sciences University of Delaware Newark, Delaware

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

This book is printed on acid-free paper.

8

Copyright 0 1993 by ACADEMIC PRESS, INC. All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Academic Press, Inc. 1250 Sixth Avenue, San Diego, California 92 101.431 1 United Kingdom Edition published by

Academic Press Limited 24-28 Oval Road, London NW1 7DX

Library of Congress Catalog Number: 50-5598 International Standard Book Number: 0- 12-000749-5

PRINTED IN THE UNITED STATES OF AMERICA 9 3 9 4 9 5 9 6 9 7 9 8

BB

9 8 1 6 5 4 3 2 1

Contents CONTRIBUTORS .......................................... PREFACE ................................................

vii ix

USEOF COMPUTER-ASSISTED TOMOGRAPHY IN STUDYINGWATER MOVEMENT AROUND PLANT ROOTS L . A. G . Aylmore

I. I1 . I11. IV . V. VI . VII . VIII . IX .

Introduction ............................................. Computer-Assisted Tomography ............................. X-Ray CATScanners ...................................... y-Ray C A T Scanners ...................................... Application of Computer-Assisted Tomography to Soil- Water Studies ................................................. Nuclear Magnetic Resonance Imaging ........................ Dual-Energy Scanning ..................................... Recent and Future Developments ............................ Summary and Conclusions .................................. References ..............................................

2 4 13

22 26 41 43 47 49 50

PHOSPHOGYPSUM IN AGRICULTURE: A REVIEW Isabel0 S . Alcordo and Jack E . Rechcigl

I . Introduction ............................................. I1. Uses of Phosphogypsum in Agriculture ....................... 111. Environmental Considerations............................... IV Conclusions ............................................. References ..............................................

.

55 65 93 100 102

NUTRIENT CYCLING AND SOILFERTILITY IN THEGRAZED PASTURE ECOSYSTEM R . J . Haynes and P . H . Williams I . Introduction ............................................. 119 121 The Pasture System and Its Effect on Soil Properties ............. I1.

I11. Nutrient Returns in Feces and Urine ......................... IV . Soil Processes and Pasture Response in Excreta-Affected Areas .... V

130 144

vi

CONTENTS V . Modeling Nutrient Cycling under Pasture ..................... VI . Summary and Conclusions .................................. References ..............................................

174 189 191

ELECTRICAL CONDUCTMTY METHODS FOR MEASURING AND MAPPINGSOILSALINITY J . D . Rhoades

I . Introduction ............................................. I1. Determination of Soil Salinity from Aqueous Electrical Conductivity ............................................. 111. Determination of Soil Salinity from Soil Paste or Bulk Soil Electrical Conductivity .................................... IV . Conclusions and Summary .................................. References ..............................................

201 204 212 246 246

BREEDING.PHYSIOLOGY. CULTURE.AND UTILIZATION OF CICERMILKVETCH (AJtragaZm cicer L.) C. E. Townsend

I . Invoduction

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

I1. Morphology and Anatomy ..................................

111. Physiology

IV . V. VI . VII .

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

Culture ................................................. Utilization ............................................... Breeding. Genetics. and Cytology ............................ Summary and Conclusions .................................. References ..............................................

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

254 254 256 265 276 289 300 301 3 09

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

ISABEL0 S. ALCORDO ( 5 5 ) , Institute of Food and Agricultural Sciences, Agricultural Research and Education Center, University of Florida, Ona, Florida 33865 L. A. G. AYLMORE (l), Department of Soil Science and Plant Nutrition, The University of Western Australia, Nedlands, Western Australia 6009, Australia R. J. HAYNES (119), New Zealand Institute for Crop and Food Research, Canterbury Agriculture and Science Centre, Christchurch, New Zealand JACK E. RECHCIGL (55), Institute o f Food and Agricultural Sciences, Agricultural Research and Education Center, University of Florida, Ona, Florida 33865

J. D. RHOADES (20l), United States Salinity Laboratory, United States Department o f Agriculture, Agricultural Research Service, Riverside, Calqornia 92501 C. E. T O W S E N D (25 3), Crops Research Laboratory, Agricultural Research Service, United States Department of Agriculture, Fort Collins, Colorado 80526

P. H. WILLIAMS (1 19), New Zealand Institute for Crop and Food Research, Canterbury Agriculture and Science Centre, Christchurch, New Zealand

vii

This Page Intentionally Left Blank

Preface Volume 49 brings together a number of plant and soil scientists who discuss some exciting and significant advances in agronomy. The first chapter deals with the use of computer-assisted tomography (CAT) in investigating water mobility around plant roots. Topics that are discussed include background information on CAT, application of CAT to soilwater studies, nuclear magnetic resonance, and future developments and uses of CAT in agronomy. The second chapter is a comprehensive review of phosphogypsum in agriculture, including its utilization as a source of sulfur and calcium and as an ameliorant of aluminum toxicity, salinity, nonsodic dispersive soils, hard pans, and hard setting clay soils. The third chapter discusses nutrient cycling and soil fertility in the grazed pasture ecosystem. Topics that are treated include the pasture system and its effect on soil properties and nutrient cycling modeling. The fourth chapter covers important advances in methods of measuring and mapping soil salinity. Sensors and procedures for measuring bulk soil electric conductivity are discussed in detail, including the use of four-electrode, electromagneticinduction, and time-domain reflectometry sensors. The morphology and anatomy, physiology, culture, utilization, breeding, genetics, and cytology of cicer milkvetch are treated in the fifth chapter. This legume is becoming increasingly useful in parts of North America as a pasture, hay, and conservation species under irrigated and dryland conditions. I appreciate the fine contributions of the authors. DONALD L. SPARKS

ix

This Page Intentionally Left Blank

USEOF COMPUTER-ASSISTED TOMOGRAPHY IN STUDYINGWATER MOVEMENT AROUND PLANT ROOTS L. A. G. Aylmore Department of Soil Science and Plant Nutrition, The University of Western Australia, Nedlands, Western Australia 6009, Australia

I. Introduction 11. Computer-Assisted Tomography A. Theory of Attenuation B. Principles of Computer-Assisted Tomography 111. X-Ray CAT Scanners A. Construction B. Detection C. Collimation D. Reconstruction E. Hounsfield Units F. Difficulties with X-Ray Scanners IV. y-Ray CAT Scanners y-Ray Scanning System V. Application of Computer-Assisted Tomography to Soil-Water Studies A. Linearity B. Structural Definition C. Water Movement to Plant Roots V1. Nuclear Magnetic Resonance Imaging VII. Dual-Energy Scanning A. Theory of Dual-Energy Scanning B. Choice of Sources C. Application of CAT to Dual-Energy Scanning VIII. Recent and Future Developments IX. Summary and Conclusions References

Adwnrrr in Agnm~iny.V d 49 Copyright 0 I99 3 by Academic Press, Inc. AU righrs of reproduction in any form reserved. 1

2

L. A. G. AYLMORE

I. INTRODUCTION An appreciation of the physical, chemical, and biological factors determining the supply, availability, and movement of water in soil/plant ecosystems, together with suitable techniques for the measurement of the forces involved, is essential to the development of an understanding of the mechanisms and dynamics of water movement in soils and their biological implications. The importance of this field of study cannot be overemphasized, particularly in semiarid and saline environments, where the availability of scarce water resources for agriculture makes it imperative that the most efficient water utilization by plants is achieved, and where limits to growth and production are most commonly set by limitations on our knowledge of such factors. A serious difficulty encountered in attempts to relate soil water to plant response is the fact that the water content in a plant root zone varies markedly in both time and space. Slatyer (1967) emphasized the importance of the soil water potential at the root/soil interface as the main soil characteristic controlling the availability of soil water for plant growth, with its value depending on both the soil water potential of the bulk soil and the potential gradient from the bulk soil to the root surface, which develops as a result of water removal by the root. Philip (1966) also suggested that the value of the water potential at the root surfaces was critical to the distribution of water potential (and to the possibility of wilting) throughout much of the plant domain. Furthermore, although the soil water tension, or matric suction, at a given depth in the root zone may correlate well with plant response in some circumstances and provide a useful basis for imgation, a clearer understanding of water availability to plants requires some means of resolving changes in soil suction or water content over the entire root zone. Unfortunately, progress in this area has been severely limited because of the difficulties associated with direct experimental measurement of soil water content or potential at the root/soil interface and in the soil immediately around the root. Until recently, techniques for the direct measurement of soil water content or potential have either been destructive (and hence lacked continuity), have perturbed the sensitive balance being examined, were too slow in their response time, or simply lacked the dimensional resolution necessary for meaningful definition of water content distributions. Although Dunham and Nye (1973) were able to measure one-off drawdowns in proximity to curtains of roots by destructive sectioning and So et al. (1976, 1978) were able to determine water potentials at the root surface by extrapolation using a collar tensiometer - potometer system, these techniques provided only very limited insights into the dy-

CAT STUDIES OF WATER MOVEMENT

3

namics of the availability of soil water for plant growth. Consequently, questions concerning the relative magnitudes of soil and plant resistances to water movement under different conditions of soil water potential and transpirational demand (Newman, 1969a,b), concerning the nature of the water driving forces (Nobel, 1974), and concerning the extent to which root/soil contact resistance (Herkelrath et al., 1977), accumulation of soil solute concentrations (osmotic potentials) (Passioura and Frere, 1967), etc., influence water availability have remained largely unresolved. Furthermore, conflicting results obtained predominantly by the indirect measuring procedures that previously were the only available techniques (Dunham and Nye, 1973; So et af., 1976, 1978) raise questions as to the validity of the physical concepts on which theoretical treatments (Molz, 1981) have been based. Similar problems had long existed in medical diagnostic radiology in seeking a method by which the interior of a section of the human body could be viewed in a nondestructive manner without interference from other regions. With advances in X-ray physics, detector technology, and mathematical reconstruction theory, a solution to the problem was essentially achieved in the early 1970s by Hounsfield (1972), who developed the technique known as computer-assisted tomography (CAT), or more simply, computed tomography (CT). (The word tomography is derived from two Greek words: tomo, meaning slice or section, and graphy, meaning to write or display.) CAT enables the three-dimensional, nondestructive imaging of the internal structure of the object under examination using measurements of the attenuation of a beam of radiation. The application of the technique to the attentuation of X-rays (colloquially referred to as CAT scanning) allowed dramatic advances in medical diagnostic capability and benefited the medical profession greatly by reducing the need for exploratory surgery to examine the internal structures of the human body. For this work Godfrey Hounsfield shared the 1979 Nobel Prize for Medicine with A. M. Cormack, who had earlier, in 1963, developed and applied a mathematical model that allowed the determination of absorption coefficients at specific points in scanned sections from the measured attenuation of collimated beams of 6oCoy-radiation. Tomographic imaging in various forms is applicable to a number of different types of energy beams, including electrons, protons, a particles, lasers, radar, ultrasound, and nuclear magnetic resonance. However, because of its convenience and versatility in medical, industrial, and scientific applications, most attention has been directed to X-ray CT. In recent years, the opportunity to use CAT scanning for nonmedical applications has blossomed, particularly in the United States, Canada, Europe, Australia, and Japan. Hopkins et al. (1981) and Davis et af.(1986) demonstrated

4

L. A. G . AYLMORE

its application to industrial problems, particularly for nondestructive testing of timber poles, plastics, concrete pillars, steel-belted automobile tires, and electronic components. Onoe e? al. (1983) described the use of a portable X-ray CAT scanner for measuring annual growth rings of live trees. The potential applications of CAT scanning in the soil and plant sciences have also attracted increasing interest over the past decade. Numerous workers (Petrovic er al., 1982; Hainsworth and Aylmore, 1983, 1986; Crestana er al., 1985; Anderson e? al., 1988, 1990; Tollner e? al., 1987; Tollner and Verma, 1989) have demonstrated that commercially available X-ray medical scanners can provide excellent resolution for some studies of the spatial distributions of bulk density and water content in soil columns, including in particular those near plant roots (Hainsworth and Aylmore, 1983, 1986; Aylmore and Hamza, 1990; Hamza and Aylmore, 1991, 1992a,b). The quantitative usefulness of such systems in soil studies has, however, been limited by the polychromatic nature of the X-ray beam and its inability to distinguish between changes in water content and bulk density in swelling soils. Furthermore; these instruments are prohibitively expensive (about $2 million) and hence have not been generally accessible to soil and plant scientists. Consequently, work in several laboratories has sought to provide experimentally more suitable systems and to reduce vastly the cost of the equipment, by the modification of “conventional” y scanning systems (Gurr, 1962; Groenevelt et al., 1969; Ryhiner and Pankow, 1969) to utilize the CAT approach (Hainsworth and Aylmore, 1983, 1988; Crestana e? al., 1986). pRays are essentially monochromatic, and the ready availability of sources providing large differentials in energy level offers the potential to distinguish quantitatively between simultaneous changes in water content and bulk density. However, the relatively low photon emission from pray sources compared with X-ray tubes requires much longer scanning times and has as yet limited measurements by this means to slow or steady-state processes. Despite these current limitations there is no doubt that the application of this exciting new technique will, with further developments, provide a major tool for soil and plant scientists and has the potential to resolve the major controversies with respect to the physics of water uptake by plant roots.

11. COMPUTER-ASSISTED TOMOGRAPHY The theory and use of the CAT technique for medical purposes has been reviewed in some detail by Budinger and Gullberg (1974), Brooks and Di Chiro (1975, 1976), and Panton (198 l), and complete reviews of various

CAT STUDIES OF WATER MOVEMENT

5

aspects of CAT scanning have been presented by Newton and Potts (198 1) and Kak and Slaney (1988). Brief reviews of CAT scanning theory as it relates to the determination of soil water content have been presented by Hainsworth and Aylmore (1983), Crestana er al. (1983, and Anderson et al. (1988). However, as the technique has only recently been introduced in soil science, an outline of the theory of CAT is given here to familiarize readers with the technique.

A. THEORY OF ATTENUATION In conventional radiography the transmission of radiation through a three-dimensional object is used to produce a two-dimensional image of the internal features of the object on a radiation-sensitivefilm. Attenuation occurs because the photons in the incident beam may be absorbed by the material and disappear, or may be deflected out of the path of the beam, leading to a decrease in the detected radiation intensity (Fig. 1). The image formation relies on the spatial variation of radiation attenuation in the object, which gives rise to a contrast in the transmitted radiation recorded on the film. The physical quantity that characterizes the attenuation of radiation by matter is called the linear attenuation coefficient (p). The three principal mechanisms of radiation attenuation in matter are photoelectric absorption, Compton scattering, and electron- positron pair production (Cullity, 1978). In photoelectric absorption, the photon collides directly with an atom of the absorber and transfers all of the energy to one of the orbital electrons, which is ejected from the atom. This is the most important process for low-energy photons (<500 keV). Because photons with energy in excess of that required to eject an electron are unlikely to be absorbed, the photoelectric absorption coefficient decreases rapidly with

Incident beam

Transmitted beam

absorbed scattered Figure 1. Attenuation of a narrow beam of radiation by absorption and scattering.

6

L. A. G. AYLMORE

increasing photon energy. Compton scattering is the predominant scattering process in which a photon collides with an atom and is deflected from its original direction with the loss of only a portion of its energy. This energy is transferred to an atomic electron, which recoils out of the atom. The absorption of photons by Compton scattering is most probable for intermediate-energy photons (500- 1000 keV). The photon continues on at a reduced energy to undergo additional Compton scattering or to be absorbed by photoelectric interaction with a second electron. Of secondary importance may be Rayleigh scattering, in which a photon may be deflected with no loss of energy and the whole atom recoils under the impact. This can occur for photons of low energy, i.e., in the region where the photoelectric effect is dominant. At very high photon energy, > 1000 keV, a photon may be absorbed in the neighborhood of an atomic nucleus or atomic electron and produce an electron-positron pair. In soil water studies, the highest photon energy used is 662 keV from a y-radiation source of I3’Cs, thus electron- positron production is not important in the attenuation process. The attenuation of a collimated beam of monoenergetic photons of intensity I,, as a result of passing through a sample of material of thickness D, yields a transmitted intensity I behind the sample, as illustrated in Fig. 2a (Anderson et al., 1988); this can be described by Beer’s Law,

I = I, exp(-pD)

(1)

where p, the linear attenuation coefficient (often referred to as attenuation coefficient),represents the fractional attenuation per unit length of the material traversed by the radiation. The value of p depends primarily on the energy of the radiation, the electron density, and the packing density of the material. Equation (1) assumes that the material is homogeneous in composition and density over the distance D. For heterogeneous materials, one can subdivide the length D into n subdivisions (of length d), each having a different linear attenuation coefficient, and can describe the attenuation over the length D as the sum of the attenuation of these small subdivisions (Fig. 2b). The transmitted radiation intensity is then

For real objects, such as soil, the attenuating material is continuously rather than discretely distributed, so EQ. (2) takes the form of an integral

C A T STUDIES OF WATER MOVEMENT

7

I : I\

EDEDI

Y

Figure 2. Schematic representation of the attenuation of monochromatic radiation of initial intensity I,, by (a) homogeneous material, (b) nonhomogeneous material consisting of discrete units with different attenuation coefficients,and (c) nonhomogeneous material consisting of a variable attenuation coefficient,p x , over the distance x from the source. (Afler Anderson et aL, 1988.)

(Fig. 2c),

where x is the distance from the radiation source and varies between 0 and D, the thickness of the sample. Equation (3) can be rearranged to yield

The logarithm of Zo/Z is effectively a sum of the attenuation coefficients along the ray path and is called a ray sum or ray projection. Obviously, a single ray sum cannot give any information about the distribution of attenuation coefficientsat discrete points within the material along the ray path. Thus interpretative difficulties can arise because the image obtained is really a two-dimensional projection of a three-dimensional object. The

8

L. A. G. AYLMORE

Figure 3. Coordinate system for calculation of the photon attenuation coefficient at a given point. Points within the object are described by fixed (x, y ) coordinates. Rays (dashed lines) are specified by their angle (4)with the y axis and their distance (r)from the origin. The S coordinate denotes distance along the way.

aim of the CAT technique is to overcome this difficulty and to reveal the spatial distribution of attenuation coefficients unambiguously.

B. PRINCIPLES OF COMPUTER-ASSISTED TOMOGRAPHY In CAT, multiple scans from different angles in a given plane provide a large number of ray sums or projections. Using these projections, a two-dimensional image of the slice is reconstructed numerically to give the distribution-of attenuation coefficients at discrete points within the slice. For image reconstruction, an (x, y ) coordinate system (Fig. 3) is used to describe points in the slice. As the slice is scanned, ray paths through the slice can be defined by 4, the angle of the ray with respect to the y axis, and r, its distance from the origin. The distance of a point from the source on any ray path is given by the coordinate S, which varies from 0 to S. The contribution of each point to the attenuation of a ray ( I , 4) with initial intensity Z, and transmitted intensity I is denoted by

Equation ( 5 ) can be rearranged to obtain the projection value, p, of the

CAT STUDIES OF WATER MOVEMENT

9

Figure 4. Schematic illustration of a parallel-beam CAT scanning procedure showing a single projection consisting of a set of parallel ray sums and the linear and rotational movementsinvolved in the collection of data prior to image reconstruction for a cross-section of the object.

ray ,(I

4) P(r, 4) = W o / ~ , +=)

I,

AX,V ) ds

(6)

In Eq. (6),p(x, y) is determined using many independent views or projections through the object. In the simplest scanning systems these are obtained using a scanning procedure involving both linear and rotational movements of the source detector system (Fig. 4). A complete set of parallel ray sums represents a single projection for that view of the crosssectional layer of the object. Ideally, p(x, y) is a continuous two-dimensional function and an infinite number of projections are required for reconstruction. Because in practice it is physically impossible to obtain an infinite number of projections, p(x, y ) is calculated at a finite number of points from a finite number of projections. If the object is confined to a circular domain of diameter d, and the image is reconstructed at points arranged rectangularly with spacing w, then there are n = d / w points along a principal diameter. Each square cell of width w is called a pixel (an acronym for picture element). It is also assumed that there are rn projections spaced equally from 0 to 180", each consisting of n ray sums at intervals w. The minimum number of

10

L. A. G. AYLMORE

rotations required for accurate image reconstruction is given by an/4 (Panton, 1981). 1. Numerical Reconstruction

In theory, if p(r, 4) is known for every line of width w passing through a pixel of dimension w, then p(x, y) can be determined if Eq. (6) can be inverted (Panton, 198I). Bracewell (1956) was the first to devise a numerical reconstruction technique for determining,u(x, y ) from Eq. (6), and with subsequent advances there are now more than a dozen different approaches available for reconstructing p(x, y ) (Budinger and Gullberg, 1974). However, these approaches can be classified broadly into three methods: ( 1) back-projection, (2) iterative reconstruction, and (3) filtered back-projection (Brooks and Di Chiro, 1975, 1976). a. Back-Projection Reconstruction In the back-projection method, reconstruction is performed by applying the magnitude of each projection to all points that make up the ray, or, in other words, the back-projected p(x, y ) value is obtained by superimposing projections together. The process can be described by Eq. (7),

+

where rj = x cos 4j y sin 4j,the distance of the ray from the origin; $ j is the jth projection angle and A 4 is the angular distance between projections (i.e., A 4 = a/m)and the summation extends over all rn projections. Back-projection, however, does not produce a good reconstruction because each ray sum or projection is applied not only to points of high density but to all points along the ray. This defect shows up most strikingly with discrete areas of high density, producing a star artifact. The star artifact causes the density function to vary from the true density function by an intolerable margin, hence this method is rarely used these days. b. Iterative Reconstruction In the iterative methods of reconstruction, the basic strategy is to apply corrections to arbitrary initial p(x, y) values in an attempt to match the measured ray projections. Because former matchings are lost as new corrections are made, the procedure is repeated until the calculated projections agree with the measured ones within the desired accuracy. Iterative methods are primarily classified according to the sequence in which corrections are made and incorporated during an iteration, as this choice

CAT STUDIES OF WATER MOVEMENT

11

has a significant effect on the performance of the method. Three such variations have been proposed: 1. Simultaneous correction: all projections are calculated at the beginning of the iteration and corrections are applied simultaneously to all points (x, y). This method has sometimes been referred to as the iterative least-squares technique (ILST) (Goitein, 1972). 2. Point-by-point correction: each point is corrected simultaneously for all rays passing through it and corrections are incorporated before moving to other points. This technique was introduced in electron microscopy by Gilbert ( 1972), who named it the simultaneous iterative reconstruction technique (SIRT). 3. Ray-by-ray correction: a given set of ray projections is calculated and corresponding corrections are applied to all points. The updated p ( x , y ) values are then used for calculating the next projection. Ray-by-ray correction was used in the original version of the EM1 scanner (Hounsfield, 1972) and was independently discovered in electron microscopy by Gordon et al. ( I 970), who named it the algebraic reconstruction technique (ART). The iterative reconstruction methods are slow and hence are not very popular.

c. Filtered Back-Projection Filtered back-projection methods are the most commonly used and are considered to be the most accurate. The basis of the analytical methods involves a fundamental relationship between the Fourier transform of the linear attenuation function p(x, y ) and the projection function p(r, 4). Such a relationship filters the projection, accounting for the portions of the projection that may pass outside a given pixel, thereby eliminating the star artifact mentioned earlier in the back-projection reconstruction method. The filtered projections are then back-projected, i.e., the reconstructed p ( x , y ) is analogous to Eq. (7) except that the p is replaced by the filtered version p *:

A number of filtering techniques (Fourier, Radon, convolution filtering, etc.) have been developed, but the performance of filtered back-projection is not greatly affected by the choice of filtering technique. Derivation of the formula for the different filters can be found in works by Brooks and Di Chiro (1976), Herman (1980), and Kak and Slaney (1988). A schematic illustration of a reconstructed image with direct back-projection and back-projection after filtration of projections (profiles) measured for a homogeneous cylinder is shown in Fig. 5. After direct back-

L. A. G . AYLMORE

12

a

b

Figure 5. Schematic illustration of a reconstructed image with (a) direct back-projection and (b) back-projectionafter filtration of profiles measured for a homogeneous cylinder.

projection of the profiles measured, image details would be smeared if no further corrections were carried out. For this reason, back-projection is preceded by a filtration or convolution process. Filtered back-projection has the advantage that the data can be processed as they are collected and the image can be built up and displayed projection by projection, thus allowing a useful saving in measurement time, particularly when data rates are low. An added advantage is that the final image is ready virtually immediately after the scans are completed and image quality can be progressively assessed. Iterative techniques need a larger data set before image reconstruction can commence, but are capable of giving better results from fewer projections than is normally required for the convolution method (Gilboy, 1984). 2. Aliasing Artifacts

Reconstruction procedures are only as good as the data on which they are based and a number of errors or artifacts can arise through the availability of insufficient data or by the presence of random noise in the measurements. An insufficiency of data may occur either through undersampling of projection data or because not enough projections are recorded. The distortions that arise as a result of inadequate data are called aliasing artifacts and generally appear as streaks, blumng, rings, or interference patterns in the reconstructed image. Aliasing distortions may also be caused by using an undersampled grid for displaying the reconstructed image (Brooks and Di Chiro, 1976; Kak and Slaney, 1988). Regions of overestimated or underestimated attenuation associated with sharp changes in attenuation and appearing as rings or oscillations are called the Gibbs effect. Moire patterns are interference patterns that can dominate the entire image. The occurrence of these effects depends largely on the relative number of projections and rays in each projection, and it has been shown (Kak and Slaney, 1988) that a balanced image reconstruction re-

CAT STUDIES OF WATER MOVEMENT

13

quires that the number of projections should be roughly equal to the number of rays in each projection.

111. X-RAY CAT SCANNERS A. CONSTRUCTION X-Ray sources were chosen for medical scanners because of their absorption characteristics in bone and body tissues and because of the high photon output from intense tube sources, providing rapid measurements within a few seconds. Any scanning pattern that provides a suitable number of ray sums to reconstruct a satisfactory two-dimensional image can be used. Whereas the reconstruction algorithms for a parallel beam system, as used in first- and second-generation commercial X-ray systems, are simpler, the time to scan across an object and then rotate the entire sourceldetector arrangement is usually too long for many (includingmedical) purposes. This time can be reduced by using an array of sources, but only at greatly increased cost. Consequently, most third-generation scanners use a fan-shaped beam with multiple detectors to minimize scanning time. Both the source and the detector array are mounted on a yoke that rotates continuously around the object over 360". This principle enables the scanner to obtain high-resolution scan images in typical scan times of 3 to 7 sec. More recently, fourth-generation fixed-detector and rotating source scanners have been developed in which a large number of detectors are mounted on a fixed ring and an X-ray tube inside this ring continually rotates around the object. In a typical third-generation scanner, the beam is collimated to form an emerging beam angle of 42" with a variable thickness of 1 to 8 mm. The detector system consists of a scintillation detector array containing 704 individual NaI detectors lined up, without a gap, over an arch of 42".The X-ray source (an oil-cooled rotating anode tube) and detector system are connected mechanically and face each other; they are located inside the gantry. A general view of the gantry and object table is shown in Fig. 6. The gantry aperture for object positioning is 70 cm in diameter. The scanning beam penetrates the object in a scan field of 5 1 cm, thus an object up to 5 1 cm in diameter can be imaged. To produce a tomogram, the tube-detector system is rotated continuously around the object. During this scan process, the object is projected 1440 times in increments of 0.25" over a rotation of 360". The intensity distribution of each projection is recorded with the detector system. The

L. A. G. AYLMORE

14

Figure 6. General View of the gantry and object table of a typical medical X-ray CAT scanner: (1) gantry, (2) gantry aperture, (3) object table with motordriven height adjustment, (4) gantry angular indication, (5) hand wheel for manual horizontal movement, ( 6 ) control panel for the object table, and (7) positioning lights for exact object positioning. (From operating manual, Siemens SOMATOM DR-H.)

geometry and measuring principle of the scanner are illustrated in Fig. 7. The operation characteristics of the X-ray tube are 96 kV/125 - 1350 mA sec or 125 kV/100- 1240 mA sec; the milliamperage depends on the scanning time. The scan time may change from 1.4 to 14 sec. The slice thickness may be 1, 2,4, or 8 mm. Details of the CAT scanner operation are given in the review articles previously cited.

-

B. DETECTION The two most commonly used X-ray detectors in medical CAT scanners are xenon gas ionization detectors and solid-state scintillation detectors, including sodium iodide, bismuth germanate, or cesium iodide (Kak and Slaney, 1988). The primary advantage of xenon gas detectors is that they are inexpensive and can be closely spaced, providing resolution down to 1 mm. Their overall efficiency is, however, lower (around 60%) compared

CAT STUDIES OF WATER MOVEMENT

1s

Figure 7. Illustration of the geometry and measuring principle of a third-generation fan beam X-ray scanner. B, Object; C, X-ray tube assembly; D, detector system; E, sense of rotation, F,single absorption profile; G, object positioning plane; H, focus; M, scan field; K, slice under examination. (From operating manual, Siemens SOMATOM DR-H.)

with scintillation crystals coupled to a photomultiplier tube or solid-state photodiode (close to 100%).

C. COLLIMATION Compton and Rayleigh scattering can lead to errors in measurement of attenuation and subsequent image reconstruction if not excluded. Although the angle of scatter is random, generally more photons are scattered in the forward direction. Although the intensity of the scattered radiation impinging on the detector is approximately constant for different rotations, its significance will vary with the degree of attenuation of the incident beam. This directional dependence leads to streaks in the reconstructed image. However, by collimating the entrance to the detector so that photons that are not traveling in a straight line between the source and the detector are excluded, this effect can be substantially reduced. Collimation largely determines the thickness of the slice scanned and spatial resolution of attenuation.

D. RECONSTRUCTION Typically a two-dimensional mapping of attenuation coefficient values into a 5 12 X 5 12 element array [p(x,y)] is constructed from the projection data acquired during a scan. The array is graphically displayed as an image

16

L. A. G. AYLMORE

on a high-resolution monitor. As in most current commercial CAT scanners, a reconstruction algorithm, called the convolution method or filtered back-projection, generates p(x, y) (Herman, 1980; Chase and Stein, 1978). Image reconstruction utilizes a specialized extensive computer system to produce the resultant array in about 6 sec. Depending on the selected mathematical scaling (zoom factor), the effective area of the 5 12 X 5 12 matrix of picture elements (pixels) comprising the image ranges from 0.1 to 1 mm2 in the actual measured field or scanning plane. Hence, an individual pixel value can represent the attenuation value of a volume element (sometimes referred to as voxel) from 0.1 X 0.1 X 1 mm3 up to 1 X 1 X 8 mm3.

E. HOUNSFIELD UNITS The attenuation coefficient values ( p )of the material within the volume of the object, which are represented by the pixel in the image matrix, are not displayed in the conventional p units of cm-'. Before the image is viewed, these coefficients are converted into an internationally standardized number scale known as Hounsfield units (H), usually expressed as follows (Newton and Potts, 1981):

H = 1 OOO(P - P w ) /pw (9) where p,., is the linear attenuation coefficientof water (cm-l). The scale is linear, and H units for air and water are defined as - 1000 and 0, respectively. Each H unit represents about 0.1% change in the attenuation coefficient of the material. The scale runs between - lOOOH and +3000H, where, for medical purposes, soft tissues are mainly in the range of -200H to +200H and values for bones range up to IOOOH. The value of 3000H corresponds to materials of high density. In practice this range is more than adequate to handle the values commonly experienced with soil constituents. Petrovic et al. (1982) listed values ranging from 450 to 800 for a fine sandy loam soil. Grevers et al. (1989) reported Hounsfield values for various materials in a polyvinyl chloride (PVC) cylinder that had been filled with uniform sand. The Hounsfield values obtained for air, water, styrene resin, glass, aluminum, sand, and the PVC cylinder were -835,80, 197,2010, 1972,783, and 821, respectively. The Hounsfield values for the soil matrix in the two sets of soil samples used in their study ranged from 9 4 5 8 to 1236H. A similar range of values have been reported by other workers. The whole range of values, i.e., from - lOOOH to +3OOOH, can be displayed in correspondingdepth-of-gray values on the system monitor. As

+

+

CAT STUDIES OF WATER MOVEMENT

17

a rule, the values in the “+” direction are recorded bright and those in the “-” direction are recorded dark. To portray the array effectively, the operator interactively selects the mean value (window center) and the range (window width) of attenuation values. Several software features assist the operator in printing out or plotting the numerical pixel values within the interactively selected region of interest (ROI). ROI can be circular, rectangular, elliptical, or irregular in shape. The size and location of the ROI can be controlled. Increasingly magnified cross-sections through a lupine plant root and surrounding soil, illustrated in Fig. 8, show the ease with which measurements near the root can be obtained and viewed. Total number of pixels, mean, standard deviation, and root mean square deviation of pixel values in the selected region, its area and volume, and the distance between two points are available instantly by manipulation of the selected region. The images can be stored on magnetic tape for subsequent analysis.

F. DIFFICULTIES WITH X-RAYSCANNERS Although conventional X-ray scanners provide excellent outputs in the form of pictorial displays and arrays of Hounsfield units associated with pixel densities, there are major limitations on their current usefulness for the quantitative determination of parameters of interest to soil and plant scientists. These arise from several sources. 1. Beam Hardening

X-ray beams are generated by impinging an electron beam on a suitable target material. The beam generated is polychromatic in nature, with the distribution of photon wavelengths being typically normal and related to the excitation energy. The subsequent attenuation of an X-ray beam as it passes through any sample depends primarily on the electron density, the packing density of the material, and the energy of the radiation. Because X-ray CT systems usually do not exceed 140 kV excitation, photoelectric absorption is the main mechanism of absorption involved. For any given wavelength, X-ray absorption can be described (Richards et al., 1960) as p = kA3Z3p (10) where k is an empirical constant of proportionality, 3, is the wavelength, 2 is the effective atomic number (electron density), and p is the specific gravity of the soil particles. The applicability of Beer’s Law [Eq. (l)] requires a monochromatic

18

L. A. G. AYLMORE

Figures 8A and 8B.

CAT STUDIES OF WATER MOVEMENT

19

Figure 8. CAT scan image showing three increasingly magnified cross-sections through a lupine plant root and surrounding soil.(A) Scale imposed on the center of the lupine root for location of pixels. (B) Magnified image showing Hounsfield unit values at the root surface and at 1.0-mm intervals away from the root surface. (C) Hounsfield unit values throughout a section of lupine root. The root diameter is about 4 mm. (After Hamza and Aylmore, 1992a,b.)

X-ray source and one predominant mechanism of X-ray absorption. However, from Eq. (lo), the shorter the wavelength, the higher the energy, and the more penetrating the radiation. Thus because dense materials will selectively absorb lower energy X-rays from polychromatic sources, there will be a progressive filtering of the beam to higher energies (i.e., lower average wavelength) as the distance traversed by the beam increases. The process is referred to as “beam hardening” (Herman, 1980). Tollner and Murphy (199 1) quoted evidence showing that a 40-cm water body caused an effective shift in wavelength from 0.21 X 1O’O m to 0.125 X 1O1O m with a Siemens system having a copper- tungsten target. The distribution would be even more skewed toward the minimum possible wavelength after passing through soil. The CT scanner reconstructs an image of a soil sample from multiple scans in a given plane, producing essentially a map of attenuation coefficients. However, depending on the total length and composition of material in each transect, the effective wavelength and hence absorption at any

20

L. A. G. AYLMORE

given location or pixel may vary, resulting in a decrease in quantitative definition. That is, the relationship between measured attenuation and density is not in fact linear, because the location of a pixel in the object can influence the attenuation measured. Attempts can be made physically and mathematically to correct for beam hardening (Brooks and Di Chiro, 1976), but such artifacts will undoubtedly occur under the conditions that exist in soils (Petrovic et al., 1982). 2. Absorption Edge Phenomena

The attenuation of the X-ray beam passing through the sample is further complicated by the nature of the absorption process. In photoelectric absorption, X-rays are absorbed by the electron cloud of an element, with each shell (e.g., K, L, M, and N) of the cloud contributingto the absorption process. A plot of the linear absorption coefficient versus wavelength exhibits a number of sharp discontinuities (Fig. 9) known as absorption edges. These absorption edges are associated with crossing energy level

h (angstroms)

Figure 9. Variation with wavelength of the energy per X-ray quantum and of the mass absorption coefficient of nickel. (From Cullity, Efements of X-ray Difraction, 0 1978 by Addison-WesleyPublishing Co. Reprinted by permission of Addison-Wesley Publishing Co., Inc., Reading MA.)

CAT STUDIES OF WATER MOVEMENT

21

thresholds associated with each shell and mark the points on the frequency scale where the X-ray possesses sufficient energy to eject an electron from one of the shells. In practice, all wavelengths shorter than that corresponding to a given absorption edge will possess sufficient energy to eject electrons from that particular shell, and those immediately shorter will be strongly absorbed. As the wavelength decreases and the energy of the photon increases above the critical excitation value, there is less likelihood of ionization and more chance that the photon will simply pass through unabsorbed. The absorption thus decreases rather rapidly on the shortwavelength side of the absorption edge. Coupled with the polychromatic nature of the X-ray beam and consequent beam hardening, absorption edges may cause a given zone to absorb differently, depending on the total length of a particular X-ray transect. Because multiple transects with different lengths traverse each location in the scanned area, the effective wavelength may vary at a location during scanning, causing the effective absorption to vary instead of being constant. Although the beam from an X-ray tube contains a continuous spectrum ranging over a wide band of wavelengths, the Ka doublet of the target material is of most use because of its great intensity. The existence of absorption edges is commonly used to reduce the polychromaticity of the X-ray beam by passing the beam through an appropriate filter. The distribution of wavelengths in the polychromatic X-ray beam is typically normal, with most photons having wavelengths near the minimum absorption edges for the target material in the X-ray tube. By choosing for the filter an element whose K absorption edge is just to the short-wavelength side of the K a line of the target material (i.e., with an atomic number one less than that of the target metal), the intensity of undesirable K/3 wavelengths can be decreased. The presence of such a filter also decreases the content of softer radiation and helps to reduce subsequent beam hardening. However, filtration is never perfect, and because of the range of materials in soils both beam hardening and absorption edge effects are to be expected. X-Rays in general must possess short wavelengths in order to penetrate materials with a high atomic number, and absorption edge problems are especially acute in dense materials. Beam hardening is of greatest concern because the soil is an absorber at least equal to the filtering element on firstand second-generation CT systems (Tollner and Murphy, 1991). 3. Other Limitations

In addition to such quantitative limitations, their expense, and problems of accessibility, commercially available medical X-ray scanners, being generally designed to monitor horizontal patients, are not conveniently constructed for soil /plant /water studies involving plants growing vertically in

22

L. A. G. AYLMORE

soil columns. Although they provide excellent resolution for some studies, detecting attenuation as low as 0.1% (Brooks and Di Chiro, 1976), their usefulness in studying soil systems has invariably been restricted by their inability to distinguish between changes in water content and bulk density in swelling soils. Furthermore, the proprietary nature of these commercial systems generally makes software modification or extensions impossible.

1V. y-RAY CAT SCANNERS Most medical CT scanners employ intense X-ray tube sources in order to complete the measurements within run times of a few seconds. The X-ray tube in a medical scanner typically can have an equivalent source strength of 15,000 Ci distributed over a few square millimeters. However, for precise quantitative imaging, pray sources are superior to X-ray tubes in almost all respects apart from the question of source brightness, which is several orders of magnitude less than X-ray tubes (MacCuaig et al., 1986). Despite this drawback, pray sources ranging from a few millicuries to 200 Ci have been used to achieve useful tomographic images in acceptable run times. Use of pray sources eliminates the beam hardening and absorption edge problems if single-photon counting is employed with energy-dispersive detectors. (Note that pray photons are indistinguishable from X-ray photons, with the different terms being used simply to indicate their origin.) pRay sources also offer additional advantages compared with X-ray tubes, including lower cost, compactness, portability, and ready access to a very wide range of photon energies. The constancy of pray photon energy over space and time also renders pray tomography a more objective method of imaging (Gilboy, 1984). With these advantages in mind, work in several laboratories has been directed to the modification of "conventional" pray scanning systems to utilize the CAT approach, to provide experimentally more suitable systems, and to vastly reduce the cost of the equipment (Hainsworth and Aylmore, 1983, 1988; Crestana et al., 1986). Single- and dual-energy pray attenuation measurements have been used for many years to monitor changes in average bulk density and water content in soil columns (Gurr, 1962; Groenevelt et al., 1969; Ryhiner and Pankow, 1969). The mass attenuation coefficients for soil and water both vary with the energy of the radiation. Consequently, solution of the two attenuation equations for measurements made at two different pray energies allows the simultaneous determination of water content and bulk density. Because bulk density often changes significantly with wetting and drying, use of the

C A T STUDIES OF WATER MOVEMENT

23

dual-energy technique greatly improves the accuracy of water content measurements over that possible when bulk density must be assumed to remain constant (Corey et af.,1971;Gardner et af., 1972).

RAY SCANNINGSYSTEM 1. Logistic System

A general view of the y-ray CAT scanning system constructed in the soil science and plant nutrition laboratories of the University of Western Australia (Hainsworth and Aylmore, 1983, 1988) is shown in Fig. 10. The main body of the system consists essentially of two platforms, the y platform, which provides vertical up and down motions for conventional y scanning, and the CAT scan platform, which provides linear translation and rotational motions required in a CAT scan procedure. The y platform supports the lead shielding for the source and scintillation detector mounted directly opposite to each other. The CAT scan platform is placed at a fixed level in the middle of the y platform such that the y platform can move independently of the CAT scan platform. Three stepper motors were used with worm drive shafts to provide the scanning motion required. A close-up photo of the CAT scan platform, where the soil column being scanned is placed, is shown in Fig. 10B. The system is capable of scanning columns up to 10 cm in diameter and 150 cm in length. A block diagram illustrating the operation of the system is shown in Fig. 11. The system consists of three subsystems: motion control system, data acquisition system, and computer system. The motion control system is linked with the computer using an interface card, which drives the platform’s stepper motors. As shown in Figs. 10 and 11, an XT-compatible personal computer (PC-XT) is used for performing the task of controlling the scanning motions and acquiring data from the radiation measurement system. Processing the data and final presentation of the results in graphic form or producing hard copies of the results are camed out by a 386-based IBMcompatible computer. The lead source housing can accommodate two sources, which may be selected by means of a rotating axle that swings the desired source in line with the collimator. Commercially available 13’Cs (0.5 Ci), 241Am(0.2 Ci), and la9Yb(up to 2.6 Ci) sources have been used with this system. The beam is collimated by a 2 X 5-mm rotatable collimator mounted on the front of the source housing and a 2 X 2-mm collimator mounted on the detector to

24

L. A. G. AYLMORE

Figure 10. The y scanning system constructed at the University of Western Australia. (A) Scanning system with associated scalar and microcomputercontrol unit. (B) Close-up of CAT scan platform. (After Hainsworth and Aylmore, 1988.)

CAT STUDIES OF WATER MOVEMENT detector electronics

b

scanner contml computer

I l l

gamma source

6I motion platform

25

motor control card

u/

printing

m data analysis computer

Storage

Figure 11. Block diagram of pray CAT scanning system constructed at the Universityof Western Australia. The system consists of three subsystems: the computer system, the data acauisition svstem. and the motion control svstem.

ensure maximum beam “sharpness” and to allow use of voxel dimensions of2 X 2 X 2 mm or 2 X 2 X 5 mm. 2. Radiation Detection

The radiation detection and measurement system consists of a scintillation detector (Model 802-3)-a monoline crystal assembly that includes a high-resolution NaI(T1) crystal (50 X 50 mm), photomultiplier tube, and preamplifier (Model 2007P). The scintillation detector is directly connected to the spectrometer, which incorporates a long-term, stable, highvoltage (HV) supply (NE 4646), an amplifier (Model 2012), single-channel analyzer (Model 2030), and dual counter (2072A) with a 207X-03 Electronics Industries Association (EIA) interface for communication with a PC. All these modules of the spectrometer are installed in an NE 4626 nuclear instrument modules bin. All components except the high-voltage power supply were supplied by Canberra- Packard Pty. Ltd. Crestana et al. (1986) also developed an inexpensive CT miniscanner platform for laboratory soil research of construction similar to the abovementioned system, but allowing a range of photon energies to be obtained by using a combination of X-ray and z41Amsources. Being custom systems, these scanners are open to modification and adaption suited to current research interests.

26

L. A. G. AYLMORE

V. APPLICATION OF COMPUTER-ASSISTED TOMOGRAPHY TO SOIL-WATER STUDIES

A. LINEARITY Knowledge of the spatial distribution of the linear attenuation coefficient in the material may be of interest but not the desired end product in the CAT process. Generally, information about the density distribution or composition of the material is desired. Thus it is necessary to obtain theoretical or empirical relationships between the linear attenuation coefficients and the parameter of interest. If a linear relationship holds, then p is the product of the mass attenuation coefficient of the material and its physical density, and the value of p for dry soil can be described (Hainsworth and Aylmore, 1983) by P&y

=P S P S

(1 1 )

where p s is the mass attenuation coefficient (cm2/g) of soil solids and ps is the bulk density of soil (g/cm3). In order to obtain experimentally the relationship between pdryand ps, the soil must be oven dry. Equation ( 1 1) can be extended to wet soil as PWt = A

P S

+P w 8 v

(12)

where pw is the mass attenuation coefficient of water and 8, is the volumetric water content of the soil. Application of Eq. (12) requires that the bulk density of soil remain constant, i.e., no swelling or shrinking on addition or removal of water. In situations in which the bulk density of the soil does not change with the addition or removal of water, 8, can be calculated from a combination of Eqs. ( 1 1) and (12) to give

4 = (P* - Pdry)/Pw

(13) Thus, in principle, the water content distribution in the soil can then be determined by first scanning the soil column when the soil is dry and rescanning it in exactly the same position when the soil is wet. The performance characteristics of medical CT scanners have naturally been optimized for X-ray absorption relevant to body tissues, which have absorptivitiesclose to that of water [p = 0.19 1 cm-* at X-ray tube voltages around 120 peak kV (McCullough, 1975)l. However, a number of workers (Petrovic et al., 1982; Hainsworth and Aylmore, 1983; Crestana et al., 1985; Brown et al., 1987; Anderson et al., 1988; Tollner and Murphy, 199 1) have demonstrated that these systems still provide essentially linear

CAT STUDIES OF WATER MOVEMENT

27

relationships between attenuation and bulk density/water content at the higher values associated with soil and similar porous materials. As might be expected, bearing in mind the dependence of the mass attenuation coefficienton the energy of the radiation, the chemical composition of the soil matrix, and the packing or bulk density, as well as differences in effects such as beam hardening, source detector geometry, and degree of electronic discrimination, substantial variations in the slope of the linear regressions have been noted with different soils and scanners. Furthermore, plots of these relationships do not always extrapolate to the origin as required by Eqs. ( 1 1) and (12). The mass attenuation coefficient of a chemical compound, or a mixture, is more sensitive to variations in the chemical composition the lower the photon energy and the heavier the elements that are subject to abundance variations. In examining the influence of chemical composition on photon attenuation by soils, Coppola and Reiniger ( 1974) had earlier demonstrated that variation of the mass attenuation coefficient with soil composition becomes significant at energies below about 200 keV and is at a maximum below 50 keV (Fig. 12). Above 300 keV any difference in the value of the attenuation coefficient was negligible. Thus in the case of an "'Am source (59.6 keV), the pray attenuation by soil was influenced rather strongly by the particular soil composition (differences as large as 200% were calculated). In contrast, the influence was essentially negligible at the energy of 13'Cs (662 keV). They also concluded that although variations in the abundance of heavier elements such as Fe can cause significant variations in p, rather large changes in the abundance of other elements, e.g., of Si, have no comparable effect on the p coefficient. Because most commercially available medical X-ray CT scanners operate at around 120 keV excitation, significant differences in attenuation resulting from variations in chemical composition could be expected to occur. The first reported attempt to use commercial medical X-ray CAT scanners to measure spatial changes in soil bulk density was by Petrovic et al. (1982), who showed that attenuation for an American Science and Engineering CT scanner was linearly related to the bulk density of soil from the surface horizon of Metea fine sandy loam over the range of 1.2 to 1.6 Mg/m3. The sensitivity to density changes for the Metea soil was about 0.02 Mg/m3, with a maximum observed deviation from predicted of 0.07 Mg/m3. Only a slightly different regression slope was observed for a mixture of glass beads and spheres and this was attributed to the differences in atomic number of the absorbing materials. Hainsworth and Aylmore (1983), using an EM1 1007 X-ray CT scanner, demonstrated that spatial changes in soil water content with time of the order of 0.006 g/cm3 could be readily resolved by the CAT technique in nonswelling soils (Fig. 13) and

28

L. A. G. AYLMORE

- .-[ -..-

-

Average earth crust Calcareous clay Norfolk sandy loam Cecil sandy loam Nipe clay

'37Cs 'N.

.01

0.1

-.

1 .o

10.0

y Energy (MeV) Figure 12. Mass attenuation coefficients for soils of various chemical compositions versus y-ray energy. [After Coppola and Reiniger, 1974.0 by Williams & Wilkins (1974).]

illustrated the scanner's potential for studies of soil/plant /water relations. Crestana et al. (1985) also obtained essentially linear calibration curves between the output from a General Electric T 8800 scanner and soil bulk density and soil water content for a sandy soil from the Ap horizon from Trieste, Italy, and a fine sandy loam from Barretos, Brazil. However, the linear calibration curves for bulk density for the two soils diverged markedly. Changing the bulk density caused a roughly parallel displacement of the calibration curves for attenuation as a function of water content. Crestana et al. (1985) used their CT scanner to observe the change in the X-ray attenuation coefficient with time at a single point in space within a homogeneous soil core as a wetting front passed. They showed that CT scanners could be used to measure the movement of water in soils at rates of 1.6 mm/sec. Similar more detailed linear relationships between attenuation coeffi-

CAT STUDIES OF WATER MOVEMENT a

29

b

0.6

*-

0.0

E 2

5J

v

c

C a,

70

70

c c

70

Distance (mm)

70

Figure 13. Half-slice three-dimensional plots showing changes in soil water content with time due to the infiltration of water into a soil column from an artificial root (alundumtube). Representations at (a) 1 min, (b) 10 min, (c) 15 min, and (d) 20 min of inEiltration are illustrated. (After Hainsworth and Aylmore, 1983.)

cients (using a Philips Tomoscan 3 10) and volume fractions of soil solids were obtained by Anderson et al. (1988) for two silt loams from Missouri. Over 99% of the variation in CT attenuation coefficients for 40 dry soil cores for each of the two soils could be accounted for by linear regression relationships with the volume fraction of soil solids (fs). Approximately 98% of the variation in CT attenuation coefficients for the 40 wet cores for each of the two soils was accounted for by regression relationships with the volume fraction of soil water (fw), after correcting for swelling effects and differences in bulk density. Differences in attenuation coefficients for the two soils were shown to be largely due to differences in Fe content. Parameter values for the Mexico silt loam compared favorably with those determined for Metea fine sandy loam used by Petrovic et al. (1982) but differed markedly from values given in the earlier work by Crestana et al. (1985), where larger sampling volumes had been used. They suggested the possibility of developing a universal relationship between X-ray CT data versus bulk density and water content if differences in the electron densities of the soils are known and the effect of soil core size on computed tomography results could be characterized.

30

L. A. G. AYLMORE

Tollner and Murphy ( 1991) also reported linear relationships among the absorption coefficientsfor solids and liquid portions in five soils ranging in texture from sand to loam and concluded that, for many applications and providing that zero swelling and shrinkage could be assumed, one calibration relationship could be applicable to a wide range of soils. For all soils tested, except for the Wilcox clay, which contained 66% clay of 2 : 1 clay mineralogy and exhibited significant shrinkage and swelling, the predicted soil density term was constant within 5% when the water term was fixed. The assumption of zero shrinkage during drying was largely met for all soils except the Wilcox clay. Equally good linearity of response between the y attenuation coefficient and the average bulk density/water content was observed by Phogat and Aylmore (1989) using a 137Cssource (Fig. 14), but again significant variation in slope between materials is evident. Furthermore, although the relationshipsbetween y attenuation and water content for both a sand from Bassendean, Western Australia, and a kaolinite-dominated sandy loam from Kulin, Western Australia, were also essentially linear, a degree of swelling resulted in the slope of the linear regression for the Kulin soil being substantially less than that for Cs radiation as measured for the nonswelling sand (Fig. 14B). Phogat and Aylmore (1989) suggested that variations in the mean and standard deviations of pixel attenuation coefficients for a scanned layer could be used to assess the structural status of the soil, because these changes reflect changes in the uniformity of the layer arising from changes in the spatial distribution of pore volume and soil matrix (soil aggregate). However, although some correction for bulk density changes on swelling can be made on the basis of appropriate calibrations, the difficulty in accounting for swelling and bulk density changes, in general, particularly when using polychromatic X-rays, remains a major impediment in studies of soil/plant/water relations. Taking into account the number of variables that may influence the regression for bulk density, it seems unlikely that a truly universally applicable relationship for attenuation at the energy levels used in medical CT scanners can be derived, and independent calibration for specific soils will remain an essential prerequisite to their use. More generally applicable relationships may, however, be possible using higher energy monochromatic prays (e.g., from lS7Cs). Differences in soil water salinity or chemical composition of practical interest in soil science have no appreciable effect on the y-ray attenuation (Coppola and Reiniger, 1974). In most of the previous studies using Xrays, values close to the standard measured absorption value for water of 0.0191 mm-l at energies around 120 keV (McCullough, 1975) have been observed, implying that one need not determine this coefficient for each soil. However, this value will of course change with the energy of the

31

CAT STUDIES OF WATER MOVEMENT

--

0.141

/ 0 Glasslubes

(A)

0 Kulin SOH

Y=0.0111+0.0617

.8

Bulk Density (g/cm3)

0.18

c.

.-cal

0

Kulin Soil

0

PureSand

(e)

0.16 -

Y=O 1167+0 0592X

a *

0.11

0.0

0.1 0.2 0.3 0.4 0.5 Volumetric Moisture Content (cm 3/cm 3) Figure 14. Linear relations between attenuation coe5cient and (A) bulk density for Kulin soil and glass tubes and (B) volumetric moisture content for Kulin soil and pure sand. Vertical lines represent standard deviations. (After Phogat and Aylmore, 1989.)

radiation, and for prays from "Am, la9Yb,and 13'Cs the coefficients are around 0.201, 0.176, and 0.083 cm-*, respectively.

B. STRUCTURAL DEFINITION Of considerableinterest is the ability of CAT scanners to characterizethe internal structure and the nature of components present in the soil. A knowledge of the size and distribution of pores is, for example, relevant to

32

L. A. G . AYLMORE

an understanding of many important processes that take place in soils, including water entry and redistribution, aeration, and root penetration. In particular, the ability to monitor root proliferation and distribution by a noninvasive technique is an essential prerequisite to detailed studies of water uptake by plant roots. The ability of the Scanner to resolve voids and objects of different sizes and spacing is obviously a function of the pixel dimensions, which for most commercial scanners is of the order of 1 X 1 mm. Spatial resolution of objects is based on the difference in attenuation of the transmitted X-rays by the object and the adjoining soil mass. Good spatial resolution can best be achieved when there is a large difference in H values between the subject (e.g,, a void) and the background (e.g., the soil mass). Most workers (e.g., Petrovic et al., 1982; Anderson et al., 1990) have found that under such conditions CT scanners are able to detect holes of the order of 1 mm in diameter but are not necessarily able to separate adjoining holes of similar separation. Although the spatial resolution as determined by the scanner and image reconstruction algorithm used by Warner et al. (1989) was approximately 0.5 mm, the practical resolution was about twice this value due to boundary effects. If the boundary of an air-filled pore 0.5 mm in diameter exactly coincides with the pixel edges, a scan image can theoretically show that pore. However, if the pore boundary falls in the middle of the pixel, the scanner will produce an attenuation value that is an average value for the materials within the pixel (e.g., soil and air). In practice, the scanner actually monitors a volume of soil (in this case voxel dimensions were 0.5 X 0.5 X 8 mm). Hence if any part of the voxel is filled with air and the rest is filled with solid, the attenuation value for the voxel will be an average between that for the solid and that for the air. Consequently, studies of porous structure have been largely limited to macropores of dimensions similar to pixel size. The CAT image of an acrylic cylinder containing various sizes of holes is shown in Fig. 15. The CAT scanner accurately depicted the location and size (in millimeters) of the air-filled holes. Holes of 2.0 mm and larger in diameter could easily be distinguished when the image was enlarged to actual size or larger. Smaller pores, down to approximately 1.0 mm, were distinguishable from the images but were often greyish in tone. Holes 0.5 mm diameter could not be distinguished with the pixel dimensions used. The scanner used in this study had a scan field of 5 10 mm with 5 12 pixels, i.e., one pixel corresponds to an object size of approximately 1 mm (0.996 mm actual). Any part of the measured field can be reconstructed with a pixel size between 1 and 0.1 mm by selecting the zoom factor between 1 and 10. For example, a zoom factor of 10 corresponds to a measured field 5 10 mm in diameter with an object size of 0.1 mm per pixel. Thus, the use

CAT STUDIES OF WATER MOVEMENT

33

Figure 15. X-Ray CAT scanning image of an acrylic cylinder containing various sizes of holes (in millimeters).

of an object size of 5 1 mm in diameter would result in a higher resolution (0.1 mm). However, large objects may have the advantage of greater statistical significance. Difficulties may also arise in high-contrast resolution because a zone of drastically changing density, such as an air-filled hole or a stone in the soil, can cause the destruction or blurring of information elsewhere in the scan due to beam hardening and other mathematical anomalies. These include lines and streaks (Fig. 16) developed in the plane of symmetry at a density gradient interface and the Gibbs phenomena (Bracewell, 1956; Brooks and Di Chiro, 1976), where large changes in attenuation from high-density to low-density areas cause the calculated values in the adjacent low-density area to be less than the actual attenuation coefficient in that region, i.e., an overshoot. The size of object required to cause this type of artifact or the extent of the perturbation will depend mainly on the attenuation differential and the characteristics of the CT scanner used. Many scanners are better able than others to compensate for this artifact (Petrovic et af., 1982). When only small differences in attenuation exist between an object and the surrounding soil, the ability to resolve the object accurately is reduced. Thus Petrovic ef al. (1982) found that objects within the scanning plane that differed in attenuation by I%, or about 10 Hounsfield units, had to be at least 6.4 mm in diameter to be detectable using a pixel size of 1 X 1 mm. Grevers et af.(1989) observed that spatial resolution in CT scans between

34

L. A. G. AYLMORE

Figure 16. Streaking artifact in a CAT image of a vertical slice through a soil column.

soil pores and the soil matrix decreased as a result of impregnation with resin because of the lower contrast in attenuation coefficient between pores and the soil matrix than in the nonimpregnated samples. The difference in Hounsfield values between the soil matrix and air-filled pores was around 1900H, whereas that between resin-filled pores and soil matrix was around 900H. Consequently, classification of the image into pore and soil matrix was more difficult and led to an underestimation of macroporosity. These CT images were, however, analyzed by applying an image-analyzing computer to grey-scale images from the scanner, which would have contributed to the difficulty in spatial resolution. The use of actual Hounsfield values for pixels from the CT scanner would undoubtedly have reduced this discrepancy. Grevers ez al. (1989) concluded that the images of a soil macropore system obtained from CT scanning compared favorably with those obtained from thin sections and had the advantage of being nondestructive and less time consuming. Similarly, Anderson et al. ( 1990) examined the influence on effective resolution of blurring at the boundary of constructed macropores by comparing measured hole diameters on the basis of 50 and 75% differencesin mean bulk density (MBD) as determined by the CT scanner. Although good agreement ( r = 0.97) was found between the CT-measured hole diameters and the wire diameters used in constructing the holes for both the 50 and 75% MBD methods, fewer holes

C A T STUDIES OF WATER MOVEMENT c

90

0

..-

-

75

+ B (0.5-1.0mm)

60

+ D (4.7-6.3mm) + E (6.3-8.0mm) + F Clod (Massive structure)

+ C (2.8-4.Omm)

Y

P

&

35

45 30

L

a 2

15

P

n

"0

10

20

30

40

50

60

70

80

Macroporosity (%)

Figure 17. Macroporosity distribution for Scans of different sizes of awegates of Kulin soil. (After Phogat and Aylmore, 1989.)

were detected using the more restrictive 50% MBD criteria as compared to the 75% MBD method. Several attempts have been made to develop quantitative characterizations of porosity and water content in soils on the basis of the distribution of pixel attenuation. Phogat and Aylmore (1989) determined average macroporosity and the spatial and frequency distribution of macroporosity for soil samples by assigning the value of zero macroporosity to pixels having y attenuation coefficients corresponding to the bulk density of a soil aggregate, 100%macroporosity to pixels with zero attenuation coefficients, and proportional values to pixels with intermediate coefficients (Fig. 17). CAT scanning of soil samples before and after a wetting and drying cycle provided a quantitative illustration of the greater reduction in macroporosity for soil wetted under flooding compared with that wetted under capillary action (Fig. 18). A similar analysis of the distribution of pixel attenuation values was used by Sawada et af. (1989) to characterize the degree and extent of dispersion of soil water content and hence to evaluate the efficiency of soil wetting agents in overcoming water repellence. The possibility of relating the hydraulic conductivity of a porous glass bead system directly to the spatial distribution and continuity of pore space as measured by the CAT scanning procedure was examined by Phogat and Aylmore (1993). A useful preliminary relationship for glass bead systems was obtained, indicating considerable potential usefulness in this approach. Because of their inherent limitations, commercially available medical X-ray scanners have thus far proved of most use in visual studies of soil structure (Anderson el af., 1988; Jenssen and Heyerdahl, 1988; Tollner and Murphy, 1991; Grevers et af.,1989),the advancement and stability of

36

L. A. G. AYLMORE 1501

,

1

150 120 90

60

30

0

15

30

45

0

15

30

45

0 60

Macroporosity (‘A) Figure 18. Changes in frequency distributions of macroporosity for York soil after wetting under flooding or capillary action, followed by drying.

wetting fronts, and the structural changes following wetting and drying (Phogat, 1991). In addition, roots, seeds, insects (Tollner et al., 1987), and pesticide granules (Cheshire et al., 1989)within soils have been successfully detected using an X-ray CAT scanner. In the author’s laboratory CAT scanning has also been used successfully in continuous monitoring of the activity of worms and dung beetles under different soil conditions as well as in studying the intensity of root nodule development. The quantitative applications of CAT to soil water studies have, however, been largely limited to statistical assessments of macroporosity distributions before and after complete wetting and drying cycles and to the measurement of water drawdowns in proximity to plant roots in nonswelling soils (Hainsworth and Aylmore, 1986; Aylmore and Hamza, 1990; Hamza and Aylmore, 1992a,b).

C. WATER MOVEMENT TO PLANT ROOTS The flow of water to plant roots is part of an overall catenary process of water transport from the soil through the plant and into the atmosphere. It is a predominantly passive process whereby water moves in response to gradients in its total water potential. Van der Honert (1948) applied an analog of Ohm’s Law to the steady-statetranspirational flux, and although he did not examine water movement through the soil to the root surface, this can be easily incorporated because the whole pathway from soil to atmosphere forms a thermodynamic continuum (Slatyer, 1967).

C A T STUDIES OF WATER MOVEMENT

37

Consequently, the pathway of water flow can be conveniently represented by a series of resistances: the flux of water (F)from the soil to the plant and out into the atmosphere under steady-stateconditions is given by the gradient in water potential through the soil/plant /atmosphere continuum divided by the total resistance. Thus

F= - Y,)/R, (14) where Ysis the average soil water potential, Yl is the average leaf water potential, and Rt is the total resistance to water flow. Thus as a first approximation, Rt can be assumed to be the sum of the soil and plant resistances in series, W

S

Rt = R, iR,

(15)

with the soil resistance & and the plant resistance 4 being defined by &=Y8-Y0/F

(16)

and

I$,=Yo-Yl',lF (17) respectively, in which Yois the average water potential at the outside surface of the root. In recent years numerous models aimed at a quantitative formulation and prediction of the processes involved in the movement of soil water and its extraction by plant roots have been developed (Hillel, 1982). Not only do these models differ in aim and level of detail, but also in approach. Two alternative approaches have generally been taken in modeling water uptake by plant roots. The first can be described as the microscale or single-root approach based on the radial flow of water to a cylindrical sink. Solutions to this approach have been attempted both by analytical means (Gardner, 1960; Cowan, 1965), usually requiring restrictive assumptions (Hillel, 1982), and by numerical means (Molz and Remson, 1970; Hillel et al., 1975). However, such models have differed widely in their quantitative prediction of water extraction. Thus, as pointed out by Hillel (1982), there are numerous theoretical models, lacking in consensus, which remain largely untested and hence unproved. It has become easier to formulate than to validate models of water extraction by plant roots (Belmans et al., 1979). There has, in particular, been considerable controversy as to the relative magnitudes of the soil and plant resistances (Newman, 1969a,b; So et al., 1976, 1978; Reicosky and Ritchie, 1976). When soil moisture is at field capacity, the influence of soil resistance on water uptake by the plant is likely to be small. As soil moisture approaches the wilting point, the

38

L. A. G. AYLMORE

influence of soil resistance increases and undoubtedly becomes limiting (Passioura, 1980). However, the extent of its influence between these two points remains a matter of considerable controversy. Some workers (e.g., Carbon, 1973) have argued that the conductivity of soil to water is frequently so small that the transport of water to the plant root may be limited, even though there may be considerable available water lefi in the soil. Others (e.g., Newman, 1969a,b; Lawlor, 1972) have argued that the properties of the plant and its aerial environment dominate water uptake unless the soil is so dry that there is virtually no water lefi. Still others (Herkelrath et al., 1977) have suggested that, under some conditions, neither soil nor plant resistance controls uptake, but that the resistance to water flow from soil to plant at the root/soil interface is the controlling factor. Because of the difficulties encountered with bulk density changes associated with swelling soils, direct studies of soil water extraction by plant roots using CAT have so far been limited to measurements of drawdowns in essentially nonswelling soils. Hainsworth and Aylmore (1983, 1986, 1989)measured the drawdowns in soil water content associated with single radish roots in a nonswelling 15% kaolinite clay and 85% sand mixture, providing the first detailed and repetitive observations of this type. Uptake of water along the radish roots was clearly shown to be nonuniform, with the plant sequentially removing water from the top to the bottom of the root as soil hydraulic resistance became a major limiting factor in the upper layers, even at the high soil water potential used (Hainsworth and Aylmore, 1989). Comparisons between the experimentally determined drawdowns and those predicted by Cowan’s (1965) analytical model indicated that such an approach was unlikely to provide a satisfactory description of drawdowns due to the restrictive assumptions required for their solution. Although the numerical model of Hillel et a). (1975) provided a reasonable approximation to the position and shape of the measured drawdowns, significant improvements in the physical concepts on which these are based are still clearly required. Subsequently, Aylmore and Hamza (1990) and Hamza and Aylmore (1991, 1992a,b) used a combination of CAT scanning and ion-specific microelectrode techniques to measure concomitantly the spatial distribution of soil water content and Na+ ion concentrations in close proximity to lupine and radish plant roots. Figures 19 and 20 illustrate the detail and accuracy with which the CAT technique was able to resolve the changes in drawdown at different depths along the roots resulting from differences in evaporative demand and electrolyte concentration, over periods up to 8 hr from commencement of transpiration. These studies demonstrated clearly the interrelations between the important parameters within the soil /plant

CAT STUDIES OF WATER MOVEMENT

-

(3

39

0.30

E0

\

(3

E0

U TopOH --O-

Y

+I,

C

a L

Mid. 0 H

Bot. 0 H

0.20

C

0 0 L

0

L

3

0.10

0

3

6

Distance from

9

1 2

root surface

1 5

(mm)

Figure 19. Water drawdowns near top, middle, and bottom sections of single radish roots after 2 hr of transpiration (T) measured by CAT scanning. (After Hamza and Aylmore, 1992a.)

system, such as matric, osmotic, and leaf water potentials. Water uptake was uniform along the length of the essentially constant-diameter lupine roots but decreased along the tapering radish roots as the diameter and hence the surface area per unit length of the roots decreased. The accumulation of Na+ at the root surfaces of both plants increased gradually with time in a near linear fashion and was slightly higher under the higher transpiration demand. These increases were not exponential, as would be expected with nonabsorption by the roots (Passioura, 1963), and this was considered to be due to back-diffusion at the relatively high water contents used. Both the potential difference and the hydrostatic pressure difference between the roots and the leaves were observed to be linear functions of the transpiration rate, implying that both plant roots acted as near-perfect osmometers under the conditions of the experiments. Plant resistances were constant with time of transpiration and increased with increasing Na+ in the treatments. Soil resistances between the root surface and bulk soil increased as the water content decreased, remaining lower at the higher solute concentrations due to the lower extraction rates. At the high water potentials used, plant resistances were always substantially higher than correspondingsoil resistances. However, in these studies the minimum soil water potential reached was - 140 Wa and major questions remain to be answered as to how the relations between the parameters measure change

0.3

0.2

PI

O-l

1

0.0

0

0.1

Radish, high T., 2 hr

0

4 I

Lupine, high T., 2 hr

I

I

I

3

6

9

I

1

12

15

0 .o

0

1

1

1

3

6

9

50 75

100

I

1

12

1s

Lupine, high T., 8 hr

Radish, high T., 8 hr 0.0

Ip-

f

1

I

1

I

I

3

6

9

12

15

0.0 0

I

I

I

I

3

6

9

12

I 15

Distance from root surface Figure 20. Effect of Na+ concentration on water drawdowns near single radish and lupine roots after 2 hr of transpiration measured by CAT scanning. (After Hamza and Aylmore, 1992a.)

CAT STUDIES OF WATER MOVEMENT

41

as the soil water diffisivity becomes a more limiting factor in the transpiration process. In drier conditions, back-diffusion would be reduced and the effect of solute accumulation at the root surface would be enhanced. The alternative approach to modeling water uptake is the macroscale or whole-root system approach in which the soil/root system is assumed to be a continuum and water movement is assumed to be essentially a one-dimensional flow (e.g., Ogata et al., 1960; Molz and Remson, 1970, 1971; Nimah and Hanks, 1973; Hillel et al., 1975). Absorption of water by the root is treated as a sink term that is distributed in a certain pattern throughout the soil. The major disadvantage of the macroscopic approach in understanding mechanisms is that it is based on spatial averaging of soil water content or potential and thus does not allow quantitative description of the magnitude of the gradients from bulk soil to individual plant roots. CAT scanning has the power to overcome these limitations by providing detailed volume distributions of water content and potential throughout a complete plant root system and by allowing the efficiency of different parts of the root system in extracting water and interaction between adjacent roots to be examined. Unfortunately, attempts in this direction (L. A. G. Aylmore and P. J. Gregory, unpublished observations; J. B. Reid, R. D. Schuller, and L. A. G. Aylmore, unpublished observations) have, as yet, been largely confounded by difficulties in obtaining uniform packing, in compensating for variations in soil bulk density, and in the accuracy of repositioning of the column in the scanner.

VI. NUCLEAR MAGNETIC RESONANCE IMAGING An alternative approach to studying soil water extraction by plant roots has involved the use of proton ('H) nuclear magnetic resonance (NMR) computerized microimaging (Woods et al., 1989). As with X-ray CAT scanning, this technique stems from recent rapid advances in medical uses of NMR imaging for clinical diagnosis and research (Wehrli et a/., 1988). NMR microimaging utilizes the interactions of the magnetic moments of some nuclei (e.g., protons) with a strong magnetic field. These interactions can be perturbed by radio waves. In a uniform magnetic field all similar protons resonate at the same frequency. If, however, a field gradient is applied, spatially different protons resonate at different frequencies. When three orthogonal gradients are applied, analysis of the different frequencies can be used to generate a proton density map or image. By changing the three gradients it is possible to control variables such as slice position, slice

42

L. A. G. AYLMORE

thickness, and resolution. If the principal source of hydrogen protons is from water, then the hydrogen proton concentration can be correlated with water content. By using different methods of generating the image, differences in the nature of the protons can be accentuated, e.g., the mobile proton density, relaxation time, or magnetic susceptibility. The images are produced by protons with motional correlation times faster than about lo-* sec. Computed tomography can be used to reconstruct a slice of the spatial distribution of water content with resolution as low as 50 X 50 pm with a slice thickness of 1.25 mm (Johnson et al., 1986; Woods et al., 1989; Wehrli et al., 1988). NMR imaging has been successfully used in the nondestructive measurement of water content in plants (Bottomley et al., 1986; Brown et d., 1986; Omasa et al., 1985; Williamson et al., 1992). Bottomley et al. (1986) used a spatial resolution of 0.6 mm with an unidentified slice definition to observe the movement of a dilute solution of CuSO, into and through roots of Viciu faba. Brown et al. (1986) were able to differentiate anatomical regions of the Pelargonium hortorum roots with a spatial resolution of 0.1 X 0.1 X 1.2 mm. Omasa et al. ( 1985)used NMR imaging with a spatial resolution of 2 mm to image root seedlings and show changes in water content in the seedlings, and Williamson et al. (1992) used NMR for noninvasive histological studies of ripening red raspberry fruits. However, Anderson and Gantzer (1989), in comparing results obtained by X-ray CT with those obtained by NMR imaging, were unable to obtain images of the soil cores with NMR. They attributed their failure to the limited range in settings of the pulse repetition time and the spin-echo time on the commercial magnetic resonance imaging (MRI) unit used in the study. Paetzold et al. (1985, 1987;Paetzold, 1986)used NMR spectroscopy to measure the water content of bulk soil, but not in specific regions. More recently, McFall et al. (1991), by using a series of reference tubes (sand phantoms) filled with acid-washed sand at various water contents to provide a rapid reference calibration curve, were able to monitor water uptake by loblolly pine seedlings from a fine sand and to compare qualitatively the relative efficiencies of fine, lateral, and taproots in water uptake. Naturally occumng soil materials, both organic and inorganic, are generally poor specimens for direct NMR study because line broadening due to chemical heterogeneity severely reduces resolution and cannot be removed by magic-angle spinning (MAS) or technologically accessible field strengths (Bleam, 199I). In addition, major difficulties arise in quantitatively measuring water which is physically bound within the soil matrix. Hence the value of this technique is likely to be limited in its usefulness for studying soil/plant /water relations.

CAT STUDIES OF WATER MOVEMENT

43

VII. DUAL-ENERGY SCANNING A major limitation of the use of single-energy X-or pray CAT scanning systems in studies involving measurements of the spatial distribution of water content in soils has been the necessary assumption of uniform or constant bulk density. Because the attenuation is a function of both the bulk density and the water content of the soil, an accurate determination of water content in soils is not possible when changes in bulk density occur during experiments (Petrovic et al., 1982; Hainsworth and Aylmore, 1983; Anderson et al., 1988; Phogat and Aylmore, 1989). Even in effectively nonswelling soils the difficulties in obtaining uniformly packed soil columns pose problems in applying Eq. (12) to monitor water contents near plant roots, because this requires the exact superposition of wet and dry scans. Any redistribution of soil as a result of wetting will also degrade the accuracy of water content determination. To monitor changes in the spatial distribution of bulk density and water content in situations in which the bulk density of the soil changes due to swelling, shrinking, or redistribution on addition or removal of water, independent estimates of attenuation associated with both bulk density and water content are required. This can only be obtained by the simultaneous use of two sources of different energies.

A. THEORY OF DUAL-ENERGY SCANNING For two pray energies the attenuation equations for Eq. (12) may be written as PLanta

=Paps

k - e t b = p&Ps

+ Pwa&

(18)

+pwbev

(19)

where subscript “a” refers to the low-energy radiation and subscript “b” refers to the high-energy radiation. Thus p-, pwb,p,, and p& are the mass attenuation coefficients for water and soil solids, respectively. Equations (1 8) and ( 19) can be solved simultaneously to give P s = [ ( ~ w b k v a a ) - ( & v a ~ w e t b ) l / [ ( & b ~ t s a )- (/?sb/&3)1

(21) - (/4vbptsa)1 Thus, ps and 8, for an individual pixel can be calculated by scanning the wet soil column with both radiation sources at a fixed position. ev

= [(p&&vcta)- (papwetb)l/[(p&pwa)

(20)

44

L. A. G . AYLMORE

B. CHOICE OF SOURCES Although the average energy of X-radiation can be adjusted in some commercially available medical systems, the range is usually relatively narrow and the polychromatic nature of the beams would greatly complicate the application of Eqs. (20) and (21). A combination of I3’Cs and ”‘Am sources has most commonly been used in conventional dual-source y scanning because of the 10-fold difference in pray energy (662 and 59.6 keV, respectively) and the long half-lives involved (Corey et al., 1971; Gardner et al., 1972: Nofziger and Swartzendruber, 1974). However, both the time required to complete successfully a CAT scan and the precision obtained depend on the transmission intensity and hence count rate. y-Ray sources generally emit much smaller photon fluxes than do X-ray sources. Whereas high-strength 137Cssources are readily available, the beam strength obtainable from “‘Am has a practical limit because of self-absorption (Miller, 1955). Scan times required with this source are not sufficiently rapid to follow the rapid changes in soil 6,, which may be associated with, for example, water extraction by plant roots or water infiltration and redistribution in the soil profile. Clearly the minimum scanning time, commensurate with adequate precision, will be generally desirable and will be particularly necessary in dual-source CAT scanning. Hainsworth and Aylmore (1988) suggested the use of 169Ybas an alternative source to because this provides a similar energy level (63.1 keV) and attenuation coefficient for water (0.176 cm-I), but emits much higher photon outputs (more than 20 times) at high activities (e.g., 1.0 to 2.0 Ci). The major disadvantage of the la9Ybsource is its relatively short half-life of 3 1 days, resulting in a working life of approximately 2 months.

C. APPLICATION OF CAT TO DUAL-ENERGY SCANNING The feasibility of applying the CAT scanning procedure to dual-source y attenuation to enable simultaneous measurements of the spatial distributions of bulk density (p,) and water content (0,) in swelling soils was examined by Phogat et al. ( 1991). The values of mean water content for slices of both swelling and nonswelling soils determined by the dual-source (137Csand 169Yb)y CAT scanning technique showed excellent agreement (R2= 0.99 1) with the values obtained gravimetrically (but expressed volumetrically) for water contents from 0 to 0.55 cm3/cm3 (Fig. 21). The differencesin water content fell within the range k0.024 cm3/cm3.Similar agreement (R2= 0.995) was obtained between mean bulk density values obtained by the two methods, with a variation in the two determinations of

45

CAT STUDIES OF WATER MOVEMENT 0.6

a

.o

h

0.5

-

5 . 5 0.4

0 ‘

0

Dual-York Dual-Kulin Cs-York Cs-Kulin

v

c

C

al

c

u

0.3

5

g

c

0.2

l-

a

g

0.1

0.0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Volumetric water content (cm3/crn3) Figure 21. Relationships between volumetric water content and water content determined by dual- and singleenergy y computerized axial tomography scanning for soils from Kulin and York, Western Australia. (After Phogat et al., 1991.)

approximately f0.015 cm3/cm3. Use of a single source (I3’Cs) markedly underestimated soil water content particularly in a structurally unstable soil (Fig. 2 1). The results for the single 13’Cs source scanning of Kulin and York soils in Figs. 14B and 21 illustrate the effects of bulk density changes during wetting on the attenuation coefficient measurement. The kaolinitedominated Kulin soil exhibits some swelling, leading to a lower estimate of the mass attenuation coefficient for water (Fig. 14B). The smectite-dominated York soil swells substantially on wetting, reducing the effective bulk density and leading to a marked underestimation of water content by the single y technique. However, the primary objective of the CAT technique is not to measure average bulk density and water content but to reveal the spatial distributions of these quantities. The results of Phogat et al. (1991) demonstrated that, using dual-source y CAT scanning, it is possible to measure the spatial distributions of 8, and ps in soils simultaneously and nondestructively with a satisfactory level of precision. Unfortunately, despite the enhanced outsource, the accumulation of statistical errors arising put from the 169Yb from the random nature of radioactive emissions still necessitated excessively large counting times to provide acceptable accuracy. The height of

L. A. G. AYLMORE

46

13.0

’37cs

16’Yb

Dual source

Gamma radiation source

Figure 22. Reconstructed half-slice three-dimensional computerized axial tomography scanning images showing the spatial distribution of water content in slices of a uniform field of water scanned with I3’Cs and I69krb at different counting times and calculated for both single I3’Cs and l69krb sources as well as for the dual-y-sourceCAT scanning procedure. (Afier Phogat et aL, 199 I .)

the surface in Fig. 22 represents 8, for each pixel in a slice through a uniform field of water, scanned with 137Csand la9Yband calculated using Eq. (20) with counting times of 0.1,2.0, and 13 sec. These images illustrate that as the counting time is increased, the variation in pixel 8, decreases and, as a result, the surface of the images becomes more uniform for all three methods. The higher photon output of the la9Yband the fact that the average attenuation from la9Ybis greater than that from 137Csreduces the relative error (the absolute error in attenuation being roughly the same) and explains the greater uniformity of the Yb scans compared to the Cs scans. At a counting time of 13 sec, 13’Cs gave a standard deviation for 8, for pixels of 0.016 cm3/cm3whereas the corresponding value for the la9Yb scan was f0.007 cm3/cm3. When the scan data of both the sources were used in Eq. (21), it yielded a standard deviation for 8, for pixels of 0.073 cm3/cm3. This variation in pixel 8, using both sources is high in spite of the very low values for scans of the individual sources (137Csand la9Yb). This multiplicative propagation of errors is most evident when, due to

CAT STUDIES OF WATER MOVEMENT

47

random emissions, a particular pixel, estimated to have a low attenuation using I3’Cs, is estimated to have a high attenuation using 16Vb, or vice versa. As a result, small variations in the Cs and Yb scans can give rise to large and unacceptable variations when Eqs. (20) and (21) are applied. Some 169 sec for an individual ray sum and, hence 112 hr to complete a dual-source scan were required to achieve average standard deviation values of pixel water content of the order of 0.025 cm3/cm3for a uniform field of water. Such large count times limit the equipment used to the study of steady-state or only slowly changing systems.

VIII. RECENT AND FUTURE DEVELOPMENTS Current medical X-ray CT systems are designed for low-radiation dosage and are high speed to freeze the motion, resulting in only moderate spatial and contrast resolution. In contrast, industrial CT systems designed for object size/weight /density flexibility and high sensitivity, and not restricted by object motion or radiation dose level, have recently been developed [e.g., at Advanced Research and Applications Corporation (ARACOR), Sunnyvale, California, and Surrey Medical Imaging Systems (SMIS), Surrey, England]. These include systems capable of inspecting objects up to 2.4 m in diameter by 5 m long and weighing up to 49,500 kg. In contrast, inspection requirements in the advanced materials, electronics, and printed circuit-board industries have also resulted in the development of X-ray CT systems that provide 25-pm spatial resolution for such objects (ARACOR). Similar work is in progress at EMBRAPA (Brazil) to improve resolution capabilities to 1 pm (S. Crestana, personal communication, 1991). Future developments may also include CT systems using backscatter radiation from the object or emission computed tomography (where the decay of in situ radioactive isotopes produces prays) to form the image, thus facilitating the measurements in inaccessible regions. Improved image and data analysis software for the y CAT scanning system at the University of Western Australia (Schuller and Aylmore, 1993) allows two- and three-dimensional visualization and quantitative analysis of scan data not offered on typical medical scanners. This software permits a three-dimensional image to be constructed from multiple contiguous scans, which can then be viewed from any angle and distance, sectioned and sliced, or stripped away by subtractive imaging using 32 colors or shades to represent variable attenuation ranges. Figure 23 illustrates the way in which both higher attenuating compacted soil layers or less atten-

48

L. A. G. AYLMORE

Figure 23. Three-dimensional image reconstructions obtained by subtractive imaging. (A) Three highdensity layers in soil core. Note curvature at edges caused by entry of coring tube into profile. (B) Lupine root in soil column.

CAT STUDIES OF WATER MOVEMENT

49

uating root material can be clearly defined in a soil column. The distortion of the soil layers by the passage of the corer is clearly shown in Fig. 23A.

IX. SUMMARY AND CONCLUSIONS Application of computer-assisted tomography to X- and pray attenuation measurements has provided an exciting new method for nondestructive imaging within a solid matrix, with considerable potential for studying soil behavior and soil/plant /water relations in space and time. However, the information provided is currently limited by the capabilities of the instrumentation available. Commercially available medical CT scanners have proved useful for visual studies of soil structure, the advancement and stability of wetting fronts, and the structural changes following wetting and drying. However, the usefulness of these systems and of single-source y CAT scanning systems in studying soil systems is invariably restricted by their inability to distinguish between changes in water content and bulk density in swelling and shrinking soils and by the associated physical relocation of soil elements that can occur. Thus their quantitative applications have been limited to the measurement of water drawdowns in proximity to plant roots in nonswelling soils and statistical assessments of macroporosity distributions before and after complete wetting and drying cycles. Though fast in operation, the quantitative usefulness of X-ray scanners is limited by the polychromatic nature of the beam and the process known as “beam hardening.” Furthermore, the proprietary nature of these commercial systems usually makes software modification or extensions impossible. In view of their substantially lower cost and superior quantitative characteristics, pray tomographic systems are likely to prove ultimately the most useful for soil and plant studies. Simultaneous measurement of the spatial distributions of water content and bulk density in soils that exhibit swelling and dispersion has been shown to be feasible using CAT applied to pray attenuation. However, the relatively dual-source (13’Cs and 169Yb) low photon emission from y sources and the propagation of statistical errors necessitate large counting times to provide acceptable accuracy and restrict the use of present y systems to the study of steady-state or only slowly changing systems. Realization of the full potential of this technique will require substantial improvements in scanning geometry and counting electronics to improve the speed and precision of measurements. Incorporation of fan beam geometry together with improved multiple-beam detection systems (MacCuaig et al., 1986) will reduce scanning times by an

50

L. A. G. AYLMORE

order of magnitude, but will inevitably increase overall instrument costs. Improved dimensional resolution will enhance structural definition in soil systems. However, current pixel dimensions of the order of 0.5 to 1 mm are quite adequate to allow meaningful resolution of many of the controversies associated with water extraction by plant roots. Reduction in scanning times to allow more rapid monitoring of changes in soil water content would seem a priority for soil and plant studies. Improved image and data analysis software allowing two- and three-dimensional visualization and quantitative analysis of scan data will also greatly enhance these activities.

ACKNOWLEDGMENTS Much of my work in this area was funded by the Australian Research Grants Committee whose support is gratefully acknowledged. I am grateful to my colleagues in the Soil Physics Section of the Department of Soil Science and Plant Nutrition, The University of Western Australia, in particular Mr. R. D. Schuller, Dr. M. A. Hamza, and Dr. V. K. Phogat, for their helpful comments on the manuscript.

REFERENCES Anderson, S. H., and Gantzer, C. J. 1989. Determination of soil water content by X-ray computed tomography and magnetic resonance imaging. Zrrig. Sci. 10,63 - 7 I . Anderson, S. H., Gantzer, C. J., Boone, J. M., and Tully, R. J. 1988. Rapid nondestructive bulk density and soil water content determination by computed tomography. Soil Sci. SOC.Am. J. 52,35-40. Anderson, S. H., Peyton, R. L., and Gantzer, C. J. 1990. Evaluation of constructed and natural soil macropores using X-ray computed tomography. Geoderma 46, 13-29. Aylmore, L. A. G., and Hamza, M. 1990. Water and solute movement to plant roots. Trans. Int. Congr. Soil Sci., 14th 11, 124- 129. Belmans, C. J., Feyen, J., and Hillel, D. 1979. An attempt at experimental validation of macroscopic-scalemodels of soil moisture extraction by roots. Soil Sci. 127, 174- 186. Bleam, W. F. 1991. Soil science applications of nuclear magnetic resonance spectroscopy. Adv. Agron. 46,9 1 - 155. Bottomley, P. A., Rogers, H. H., and Foster, T. H. 1986. NMR imaging shows water distribution and transport in plant root systems in situ. Proc. Natl. Acad. Sci. U.S.A.83, 87 - 89. Bracewell, R. M. 1956. Strip integration in radio astronomy. Aust. J. Phys. 9, 198-217. Brooks, R. A., and Di Chiro, G. 1975. Theory of image reconstruction in computed tomogra phy. Radiology (Easton. Pa.) 117, 561 -572. Brooks, R. A., and D i Chiro, G. 1976. Principles of computer assisted tomography (CAT) in radiographic and radioisotopic imaging. Phys. Med. Bid. 21,689-732. Brown, J. M., Johnson, G. A., and Kramer, P. J. 1986. In vitro magnetic resonance micros-

CAT STUDIES OF WATER MOVEMENT

$1

copy of changing water content in Pelargonium hortorum roots. Plant Physiol. 82, 1158- 1160. Brown, J. M., Fonteno, W. C., Cassel, D. K., and Johnson, G. A. 1987. Computed tomographic analyses of water distribution in three porous foam media. Soil Sci. SOC.Am. J. 51, 1121-1125. Budinger, T. F., and Gullberg, G. T. 1974. Three dimensional reconstruction in nuclear medicine emission imaging. IEEE Trans. Nucl. Sci. NS21,2-20. Carbon, B. A. 1973. Water stress in plants on a coarse soil.Aust. J. Soil Res. 11, 33-42. Chase, R. C., and Stein, J. A. 1978. An improved imaged algorithm for (JT scanners. Med. Phys. 5,497 - 499. Cheshire, J. M., Jr., Tollner, E. W., Verma, B. D., and Blum, W. M. 1989. Radiographic detection of soil-incorporatedgranular pesticides and impacts of application methods on wireworm management. Trans. ASAE 32,415-423. Coppola, M., and Reiniger, P. 1974. Influence of the chemical composition on the gammaray attenuation by soils. SoilSci. 117, 331-335. Corey, J. C., Peterson, S. F., and Wakat, M. A. 197 I. Measurement of attenuation of Cs-137 and Am-241 gamma rays for soil density and water content determinations. Soil Sci. SOC. Am. Proc. 35,2 15 - 2 19. Cormack, A. M. 1963. Representation of a function by its line integrals, with some radiological applications. J. Appl. Phys. 34,2722-2727. Cowan, I. R. 1965. Transport of water in the soil-plant atmosphere system. J. Appl. Ecol. 2, 221 -239. Crestana, S., Mascarenhas, S., and Pozzi-Mucelli, R. S. 1985. Static and dynamic threedimensional studies of water in soil using computed tomographic scanning. Soil Sci. 140, 326-332, Crestana, S., Cesareo, R., and Mascarenhas, S. 1986. Using a computer assisted tomography miniscanner in soil science. Soil Sci. 142,56-6 1. Cullity, B. D. 1978. “Elements of X-Ray Difiaction.” Addison-Wesley, Reading, Massachusetts.

Davis, J. R., Morgan, M.J., Wells, P., Shadbolt, P., and Suendermann, B. 1986. X-Ray computed tomography. 1: A non-medical perspective. Aust. Phys. 23,245-247. Dunham, R. J., and Nye, P. H. 1973. The influence of soil water content on the uptake of ions by roots. I. Soil water content gradients near a plane of onion roots. J. Appl. Ecol. 10,585-598. Gardner, W. R. 1960. Dynamic aspects of water availability to plants. Soil Sci.89, 63-73. Gardner, W. H., Campbell, G. S., and CalissendorlT, C. 1972. Systematic and random errors in dual gamma energy soil bulk density and water content measurements. Soil Sci. SOC. Am. Proc. 36,393-398. Gilbert, P. F. C. 1972. Iterative methods for the three-dimensional reconstruction of an object from projections. J. Theor. Bid. 36, 105 - 1 17. Gilboy, W. B. 1984. X- and gamma-ray tomography in NDE applications. Nucl. Instr. Methods 221, 193-200. Goitein, M. 1972. Three dimensional density reconstruction from a series of two dimensional projections. Nucl. Znstr. Methods 101, 509- 5 18. Gordon, R., Bender, R., and Herman, G. T. 1970. Algebraic reconstruction techniques (ART) for three-dimensional electron microscopy and X-ray photography. J. Theor. Biol. 29,47 1 -48 1. Grevers, M. C. J., De Jong, E., and St. Arnaud, R. J. 1989. The characterization of soil macroporosity with CT scanning. Can. J. Soil Sci. 69,629-637.

52

L. A. G. AYLMORE

Groenevelt, P. H., De Swart, J. G., and Cider, J. 1969. Water content measurement with 60 keV gamma-ray attenuation. Bull. Int. Am. Sci. Hydrol. 14,67 - 78. Gun; C. G. 1962. Use of gamma rays in measuring water content and permeability in unsaturated columns of soils. Soil Sci. 94,224-229. Hainsworth, J. M., and Aylmore, L. A. G. 1983. The use of computer assisted tomography to determine spatial distribution of soil water content. Aust. J. Soil Res. 21, 435443. Hainsworth, J. M., and Aylmore, L. A. G. 1986.Water extraction by single plant roots. Soil Sci. SOC.Am. J. 50,841-848. Hainsworth, J. M., and Aylmore, L. A. G. 1988. Application of computer assisted tomography (CAT) to gamma attenuation measurement of soil water content. Aust. J. Soil Res. 26, 105-110.

Hainsworth, J. M., and Aylmore, L. A. G. 1989.Non-uniform soil water extraction by plant roots. Plant Soil 113, 121 - 124. Hamza, M., and Aylmore, L. A. G. 1991.Liquid ion exchanger microelectrodes used to study soil solute concentrations near plant roots. Soil Sci. Soc. Am. J. 55,954-958. Hamza, M., and Aylmore, L. A. G. 1992a. Soil solute concentration and water uptake by single lupin and radish plant roots. I. Water extraction and solute accumulation. Plant Soil 145, 187- 196. Hamza, M., and Aylmore, L. A. G. 1992b. Soil solute concentration and water uptake by single lupin and radish plant roots. 11. Driving forces and resistances. Plant Soil 145, 197-205. Herkelrath, W. N.,Miller, E. E., and Gardner, W. R. 1977. Water uptake by plants: 11. The root contact model. Soil Sci. Soc. Am. J. 41, 1039- 1043. Herman, G. T. 1980. “Image Reconstruction from Projections: The Fundamentals of Computed Tomography.” Academic Press, New York. Hillel, D. 1982.“Applications of Soil Physics.” Academic Press, New York. Hillel, D., Van Beck, C. G. E. M., and Talpaz, H. 1975. A microscopic-scale model of soil water uptake and salt movement to plant roots. Soil Sci. 120,385- 399. Hopkins, F. F., Morgan, 1. L., Ellinger, H. D., Klinksiek, R. V., Meyer, G. A., and Thomp son, J. N. 1981. Industrial tomography applications. IEEE Trans. Nucl. Sci. NS28, 1717- 1720. Hounsfield, G. N. 1972. “A Method of and Apparatus for Examination of a Body by Radiation Such as X- or Gamma-Radiation,’’British Patent No. 1283915. British Patent Office, London. Jenssen, P. D., and Heyerdahl, P. H. 1988.Soil column descriptions from X-ray computed tomography density images. Soil Sci. 146, 102- 107. Johnson, G.A., Thompson, M. B., Gewart, S . L., and Hayes, C. E. 1986.Nuclear magnetic resonance imaging at microscopic resolution. J. Magn. Reson. 68, 129- 137. Kak, A. C.,and Slaney, M. 1988.“Principles of Computerized Tomographic Imaging.” IEEE Press, New York. Lawlor, D. W. 1972. Growth and water use of Lolium perenne. I. Water transport. J. Appl. Ecol. 9,79-98. MacCuaig, N., Tajuddin, A. A., and Gilboy, W. B. 1986. Industrial tomography using a position sensitive carbon fibre anode proportional counter. Nucl. Instr. Methods A242, 620-625. McCullough, E. C. 1975. Photon attenuation in computed tomography. Med. Phys. 2, 307- 320. McFall, J. S., Johnson, G. A., and Kramer, P. J. 1991.Comparative water uptake by roots of different ages in seedlings of loblolly pine (Pinus taeda L.).New Phytol. 119,551 - 560.

CAT STUDIES OF WATER MOVEMENT

53

Miller, D. G. 1955. “Americium-241 as a Photon Source for Gamma Absorptometric Technique,” U.S. At. Energy Comm. Rep. HW-3997 1. Hanford Atomic Products Operation, Richland. Molz, F. J. 1981. Models of water transport in the soil-plant system: A review. Water Resour. Res. 17, 1245- 1260. Molz, F. J., and Remson, I. 1970. Extraction-term model of soil moisture use by transpiring plants. Water Resour. Res. 6, 1346- 1356. Molz, F. J., and Remson, I. 1971. Application of an extraction-term model to the study of moisture flow to plant roots. Agron. J. 63, 72 - 77. Newman, E. I. 1969a. Resistance to water flow in soil and plant. I. Soil resistance in relation to amount of root: Theoretical estimates. J. Appl. Ecol. 6, 1 - 12. Newman, E. I. 1969b. Resistance of water flow in soil and plant. 11. A review of experimental evidence on the rhizosphere resistance. J. Appl. Ecol. 6,26 1 - 272. Newton, T. H., and Potts, D. G. 1981. “Radiology of the Skull and Brain: Technical Aspects of Computed Tomography,” Vol. 5, pp. 3853-3917. Mosby, St. Louis, Missouri. Nimah, M. N., and Hanks, R. J. 1973. Model for estimating soil water, plant, and atmospheric interrelations: I. Description and sensitivity. SoilSci. Soc. Am. J. 37,522-527. Nobel, P. S. 1974. “Introduction to Biophysical Plant Physiology.” Freeman, San Francisco. Nofziger, D. L., and Swartzendruber, D. 1974. Material content of binary physical mixtures as measured with a dual-energy beam of gamma rays. J. Appl. Phys. 45,5443-5449. Ogata, G., Richards, L. A., and Gardner, W. R. 1960. Transpiration of alfalfa determined from soil water content changes. Soil Sci. 89, 179- 182. Omasa, K., Onoe, M., and Yamada, H. 1985. NMR imaging for measuring root systems and soil water content. Seibutsu Kankyo Chosetsu 23,99- 102. Onoe, M., Tsao, J. M., Nakamura, H., Kogure, J., Kawamura, H., and Yoshimatsu, M. 1983. Computed tomography for measuring annual rings of a live tree. Proc. ZEEE 71,907908. Paetzold, R. F. 1986. NMR instrument for soil moisture ground-truth data collection. ZTC J. 1,9-13. Paetzold, R. F., Matzkanin, G. A., and De Los Santos, A. 1985. Surface soil water content measurement using pulsed nuclear magnetic resonance techniques. Soil Sci. SOC.Am. J. 49,537-540. Paetzold, R. F., De Los Santos, A., and Matzkanin, G. A. 1987. Pulsed nuclear magnetic resonance instrument for soil-water content measurement: Sensor configurations. Soil Sci. Soc. Am. J. 51, 287-290. Panton, D. M. 198 1. Mathematical reconstruction techniques in computed axial tomography. Math. Sci. 6, 87- 102. Passioura, J. B. 1963. A mathematical model for the uptake of ions from the soil solution. Plant Soil 18,225-238. Passioura, J. B. 1980. The transport of water from soil to shoot in wheat seedlings. J. Exp. Bot. 31, 333-345. Passioura, J. B., and Frere, M. H. 1967. Numerical analysis ofthe convection and diffusion of solutes to roots. Aust. J. Soil Res. 5, 149- 159. Petrovic, A. H., Siebert, J. E., and Lieke, P. E. 1982. Soil bulk density analysis in three dimension by computed tomographic scanning. Soil Sci. Soc. Am. J. 46,445 -450. Philip, J. R. 1966. Plant water relation-Some physical aspects. Annu. Rev. Plant Physiol. 17,245-268. Phogat, V. K., and Aylmore, L. A. G. 1989. Evaluation of soil structure by using computer assisted tomography. Aust. J. Soil Res. 27, 3 13 - 323.

54

L. A. G. AYLMORE

Phogat, V. K., and Aylmore, L. A. G. 1993. Computation of hydraulic conductivity of porous media using computer assisted tomography. Aust. J. Soil Res. (submitted). Phogat, V. K., Aylmore, L. A. G., and Schuller, R. D. 1991. Simultaneous measurement of the spatial distribution of soil water content and bulk density. Soil Sci. SOC.Am. J. 55, 908 -9 I 5. Reicosky, D. C., and Ritchie, J. T. 1976. Relative importance of soil resistance and plant resistance in root water absorption. Soil Sci. SOC.Am. J. 40,293-297. Richards, J. A., Sears, F. W., Wehr, M. R., and Zamansky, M. W. 1960. “Modern University Physics.” Addison-Wesley, London. Ryhiner, A. H., and Pankow, J. 1969. Soil moisture measurement by the gamma transmission method. J. Hydrol. (Amsterdam) 9, 194-205. Sawada, Y., Aylmore, L. A. G., and Hainsworth, J. M. 1989. Development of a soil-water dispersion index (SOWADIN) for testing the effectiveness of soil-wettingagents. Aust. J. Soil Res. 21, 17-26. Schuller, R. D., and Aylmore, L. A. G. 1993. Three dimensional image reconstruction from computer assisted tomography applied to gamma attenuation in soil columns. Manuscript in preparation. Slatyer, R. 0. 1967, “Plant- Water Relationships.” Academic Press, New York. So, H. B., Aylmore, L. A. G., and Quirk, J. P. 1976. Measurement of water flux in single root system. I. The tensiometer-potometer system. Plant Soil 45, 577-594. So, H. B., Aylmore, L. A. G., and Quirk, J. P. 1978. Measurement of water flux in single root system. 11. Applications of tensiometer-potometer system. Plant Soil 49,46 1-475. Tollner, E. W., and Murphy, C. 1991. Factors affecting soil X-ray absorption coe5cients with computed tomography. Trans. ASAE 34, 1047- 1053. Tollner, E. W., and Verma, B. D. 1989. X-Ray CT for quantifying water content at points within a soil body. Trans. ASAE 32,901 -905. Tollner, E. W., Verma, B. D., and Cheshire, J. M., Jr. 1987. Observing soil-tool interactions and soil organisms using X-ray computer tomography. Trans. ASAE 30, 1605- 1610. Van den Honert, T. H. 1948. Water transport in plants as a catenary process. Discuss. 3, 146- 153. Faraday SOC. Warner, G. S., Nieber, J. L., Moore, I. D., and Geise, R. A. 1989. Characterizingmacropores in soil by computed tomography. Soil Sci. SOC.Am. J. 53,653-660. Wehrli, F. W., Shaw, D., and Kneeland, J. B. (eds.). 1988. “Biomedical Magnetic Resonance Imaging, Principles, Methodology and Applications.” VCH Publ., New York. Williamson, B., Goodman, B. A., and Chudek, J. A. 1992. Nuclear magnetic resonance (NMR)micro-imaging of ripening red raspberry fruits. New Phytol. 120,21-28. Woods, R. T., Hennessy, J. J., Kwok, E., and Hammer, B. E. 1989. NMR microscopy-A new biological tool. BioTechniques 7,6 16-622.

PHOSPHOGYPSUM IN AGRICULTURE: A REVIEW Isabel0 S. Alcordo and Jack E. Rechcigl Institute of Food and Agricultural Sciences, Agricdtural Research and Education Center, University of Florida, Ona, Florida 33865

I. Introduction A. General B. World Production and Utilization of Phosphogypsum C. Physical and Chemical Properties of Phosphogypsum 11. Uses of Phosphogypsum in Agriculture A. Source of S and Ca for Crops B. Ameliorant for Aluminum Toxicity and Subsoil Acidity C. Ameliorant for Sodic Soils D. Ameliorant for Nonsodic Dispersive Soils, Subsoil Hardpans, and HardSetting Clay Soils E. Bulk Carrier for Micronutrients and Low-Analysis Fertilizers 111. Environmental Considerations A. Effects on Surficial Ground Water B. Effects on Soils C. Effects on Crop Tissues D. Effects on Ambient Atmosphere IV. Conclusions References

I. INTRODUCTION A. GENERAL Gypsum (CaSO,.xH,O) is available for agricultural use either as mined gypsum or as a chemical by-product. Gypsum by-products are produced during phosphoric, hydrofluoric, and citric acid manufacture and as a Adwnrts in Ag~~nmny, Voi. 49 Copyright 0 1993 by Academic Press, Inc. AU rights of reproduction in my form reserved.

55

56

I. S. ALCORDO AND J. E. RECHCIGL

result of pollution control systems processes, such as in the neutralization of waste sulfuric acid and in flue-gas desulfurization. Phosphogypsum is the term used for the gypsum by-product of wet-acid production of phosphoric acid from rock phosphate. It is essentially hydrated CaSO, with small proportions of P, F, Si, Fe, Al, several minor elements, heavy metals, and radionuclides as impurities. Rock phosphate deposits are found throughout the world, and on these deposits the phosphoric acid industries are built. Countries with no natural phosphate deposits import the rock to produce phosphoric acid for their industry and agriculture. Therefore, the production of by-product phosphogypsum is more widely distributed around the world than are the natural deposits of rock phosphate. Thus, among the gypsum by-products, only phosphogypsum is of worldwide importance in quantity and distribution.

B. WORLD PRODUCTION AND UTILIZATION OF PHOSPHOGYPSUM The three basic conventional processes used in wet-acid manufacture of phosphoric acid are the dihydrate, the hemihydrate, and the hemidihydrate processes. For each megagram (Mg) of P produced, the hemihydrate process yields about 9.8 Mg of dry phosphogypsum, whereas the dihydrate and hemidihydrate processes yield about 1 1.2 Mg (Kouloheris, 1980). Worldwide production of phosphoric acid, estimated at 1 1 million Mg of P annually (Lin et al., 1990), also results in the production of approximately 125 million Mg of phosphogypsum. With only about 4% of the world's phosphogypsum production being used in agriculture and in gypsum board and cement industries, about 120 million Mg of phosphogypsum accumulates annually; most of this excess is piled in stacks, and some is stored in abandoned quarries or, in certain countries, dumped into waterways. Australia produces 940,000 Mg of phosphogypsum annually, of which 200,000 Mg is used as soil conditioners or fertilizers. The rest is stockpiled on land and in abandoned quarries. The stockpile in 1990 had reached 8 million Mg. Australia discontinued the use of phosphogypsum for making plaster products in 1983 (Beretka, 1990). India produces about 2.8 million Mg of phosphogypsum annually, and it is used primarily as a soil amendment or conditioner for sodic soils (Mishra, 1980). Since 1970 phosphogypsum production in Japan has stabilized at 2.5 to 3.0 million Mg annually, almost all of which is used in the cement, gypsum board, and plaster industries. The amount of phosphogypsum being used as fertilizer ranges from 25,000 to 48,000 Mg annually. As a result, Japan has no stockpile of phosphogypsum (Miyamoto, 1980). Full utilization of

PHOSPHOGYPSUM IN AGRICULTURE

57

Figure 1. One of the 20 phosphogypsum stacks located in Florida.

phosphogypsum in Japan is made possible by the Nissan hemidihydrate phosphoric acid process that produces high-quality phosphogypsum suitable for the construction industry (Goers, 1980). In 1979, Canada produced 4 million Mg of phosphogypsum and had some 50 million Mg in large containment areas. On the basis of the 1979 production, Canada could have almost 100 million Mg of phosphogypsum in stock at this time. Collings ( 1980) reported that producers of phosphogypsum had used the material as an additive to clay soils and as fertilizer. According to Khalil et al. (1990), 14% of the total production of phosphogypsum in Iraq was reprocessed, 58% was stored or stockpiled, and 2890 was dumped into waterways. The Netherlands, with an annual production of 2 million Mg, simply discharges its phosphogypsum into its surface waters (van der Sloot and de Groot, 1985). In 1988 the wet-process phosphoric acid plants in the former USSR produced 23.5 million Mg of phosphogypsum. Phosphogypsum production is expected to increase by the year 2000 to 43.6 million Mg annually. The total quantity of the material in stock in January, 1990, was estimated at 300 million Mg. Russia has been exploring various ways to use the material in industry and agriculture. With 66 million ha of acid soil and over 100 million ha of Solonetz and Solonetz-like soils that need to be reclaimed or ameliorated, phosphogypsum is expected to play an important role in Russian agriculture. In 1988, 3.2 million Mg of phosphogypsum was used for chemical reclamation of Solonetz soils, and by the year 2000, phosphogypsum use for this purpose is expected to reach 19.2 million Mg annually (Novikov et d., 1990). Florida leads in the production of phosphogypsum in the United States. Annual production has been placed at 32 million Mg. Over 400 million Mg [Environmental Protection Agency (EPA), 19911 is stockpiled at 20 stacks (Fig. I). Annual use of phosphogypsum in agriculture is placed at

58

I. S. ALCORDO AND J. E. RECHCIGL

2% of production. For the entire United States, present annual production has been placed at 45 million Mg (Arman and Seals, 1990),and there is a stockpile of 8 billion Mg at stacks in 12 states (EPA, 1991). Because of strict environmental regulations, phosphogypsum has not been used as raw material for applicable industries to date. Research in this area, however, is extremely active, not only in the United States but also worldwide (Boms and Boody, 1980; Chang, 1990).

C. PHYSICAL AND CHEMICAL PROPERTIES OF PHOSPHOGYPSUM Table I shows the chemical analysis of phosphogypsum produced by each of the three conventional processes used worldwide in the manufacture of phosphoric acid. Table I1 gives the chemical analysis of the major constituents of phosphogypsum from several countries. Phosphogypsum has a gypsum content ranging from 85 to 93Yo (Appleyard, 1980). It has a silty feel, with particle sizes clustering around 0.05 mm in diameter. It may contain small amounts of calcium sulfite and calcium carbonate (Nifong, 1988). Phosphogypsum may be contaminated by as many as 50 elements originally present in various forms in rock phosphate (Borris and Boody, 1980). May and Sweeney (1980, 1983) detected 30 elemental impurities in Florida phosphogypsum using emission spectrographic analysis and 44 impurities using neutron activation analysis. Some of the impurities, excluding the so-called EPA “toxicity index” metals, which are presented separately, are given in Table 111. Fluoride, also present originally in rock phosphate, may be found in relatively high concentrationsin phosphogypsum. Australian phosphogypsum was reported to contain from 11 to 13 g F kg-I phosphogypsum (Beretka, 1980, 1990). Indian phosphogypsum may contain 5 to 40 g F kg-I (Mishra, 1980), which could be of environmental concern when the phosphogypsum is applied at high rates for reclamation of sodic soils. Florida phosphogypsum has 2 to 8 g F kg-’ (May and Sweeney, 1980, 1983). Leachates of certain Florida phosphogypsum samples had been reported to contain F in excess of the EPA limit of 1.4 mg F liter-’ for drinking water (Nifong, 1988). A large proportion of F in a Florida phosphogypsum sample dissolved readily in Mehlich I solution but not in water (Alcordo and Rechcigl, 1992). Aqueous equilibrium solutions of mined gypsum and phosphogypsum, which contained 16.2 and 17.0 mmol Ca liter-’, respectively, reflect only slight differences in solubility between the two types of gypsum (Shainberg et al., 1989). Florida phosphogypsum dissolved at a constant rate of 0.26 and 0.43 g 100 m1-I in water and Mehlich I solution (0.025 A4 HCl+

PHOSPHOGYPSUM IN AGRICULTURE Table I Typical Chemical Analysis of Phosphogypsum Produced by Three Different Processes in the Wet-Acid Production of Phosphoric Acid" Processb Component CaO

so3

PZO, F Si02 Crystal. H,O

Dihydrate Hemihydrate Hemidihydrate 322.0 465.0 2.5 5.0 4.0 0.5 3.0 200.0

369.0 503.0 15.0 8.0 7.0 I .o 3.0 90.0

325.0 440.0 6.5 12.0

5.0

1.o 1.O 190.0

From Kouloheris ( 1980). Analytical results given in grams/kilogram.

Table I1 Chemical Analysis of Phosphogypsum Produced in Different Countries Country" Component Australia Canada Japan CaO

so3

SiO, A1203

Mi@ Na,O K20 Total P 2 0 5 Total F Total H,O

329.0 451.0 50.0 3.4 0.6 0.4 3.5 0.5 8.8 13.0 209.0

340.0 458.0

8.4 17.0 196.6

Iraq

304.0 329.0 435.0 449.0 40.5 10.5 10.5 1.1 4.0 0.4 4.6 0.1 0.8 I .6 1.8 2.9 2.4 6.0 190.0 182.0

United States 311.0 420.0 5.7 5.7 1.4 0.0 6.1 0.1 37.0 8.0 188.0

a Highest value in a range of values (in grams/kilogram) reported for different sources of phosphogypsum from within the country or state. Sources: Australia (Beretka, 1990); Canada (Collings, 1980); Japan (Miyamoto, 1980); Iraq (Khalil et al., 1990);United States (May and Sweeney, 1983).

59

60

I. S. ALCORDO AND J. E. RECHCIGL Table I11 Some Elemental Impurities in Florida Phosphogypsum Excluding the EPA “Toxicity Index” Metals’ Element

Concentration (mg kg’)

A 1uminum Antimony Bromine Cerium Cesium Chlorine Cobalt Copper Erbium Gadolinium Gold Iron Lanthanum Magnesium Manganese Molybdenum Potassium Sodium Strontium Titanium Uranium Zinc

2000.0 0.2 <0.9 49.0 0.5 < 150.0 0.6 <82.0 <330.0 150.0

a

<0.01

930.0 39.0 <940.0 25.0 6.6 215.0 520.0 600.0 440.0 9.6 < 340.0

From May and Sweeney ( 1983).

0.0125 M H,SO,), respectively. The aqueous solution had an electrical conductivity (E,) of 0.21 S m-’ and a pH of 5.2 (Alcordo and Rechcigl, 1992). Figures 2 through 5 show the solubility curves of the major elemental constituents of a Florida phosphogypsum in water and in Mehlich I solution. The solubilities of the “toxicity index” metals are given in Table IV. Although gypsum is a neutral salt, phosphogypsum is highly acidic; its pH in water (1 : 1) ranges from > 2 to < 5, mainly due to acid impurities, such as sulfuric, phosphoric, hydrofluoric, and fluosilicic acids (Nifong, 1988). But it is the heavy metals and the radionuclides present in phosphogypsum that give rise to environmental concerns in its use in agriculture.

PHOSPHOGYPSUM IN AGRICULTURE

61

50

PG

--e-Ca

A

40

P

-0

&

a,

-> 5 -m

30

K &

v)

c.

c 0 . 20 h

0

8 10

0

1

2

3

4

5

6

Solubility run Figure 2. Solubility curves of phosphogypsum (PG) and Ca, P, K, and Mg in PG in Mehlich I. (From Alcordo and Rechcigl, 1992.) 25

PG +

I1

Ca

A P -itMg

-eK +

0

16 1

A 2

A 3

A 4

A 5

A 6

Solubility run Figure 3. Solubility curves of phosphogypsum (PG)and Ca, P, K, and Mg in PG in water. (From Alcordo and Rechcigl, 1992.)

62

I. S. ALCORDO AND J. E. RECHCIGL 80

* PG Al

A 60

Fa

'CI

\

-> Q,

0

\

u) VI

.41Q 40

-scu --+--

\

Zn &

\

CI

0

CI

Mn --t-

F

u-

0

8

20

0

A

m

1

2

3

4

5

6

Solubility run Figure 4. Solubility curves of phosphogypsum (F'G)and Al, Fe, Cu, Zn, F, and Mn in PG in Mehlich I. (From Alcordo and Rechcigl, 1992.) 25

20

'0

-> Q,

$ f

15

v)

Q

.c1

0 10

.c1

rc 0

8 5

0

du 1

m 2

m 3

m 4

c 4

5

m 6

Solubility run Figure 5. Solubility curves of phosphogypsum (PG)and Al, Fe, Cu, Zn,F, and Mn in PG in water. (From Alcordo and Rechcigl, 1992.)

63

PHOSPHOGYPSUM IN AGRICULTURE Table IV

EPA “Toxicity Index” Metal Concentration Limits and Conteuts in Florida Phosphogypsum and Phosphogypsum Leachate Concentration (mg liter-’) EPA “toxicity index” metal AS Ba Cd Cr

0.85 105.00 0.59

Pb Hi3

I .30 0.50 1.40 0.69

Se Ag a

Analysis in

Florida PG” (mg kg-’)

6.00

Percentage of total leached”

Florida PG

leachateb

EPA limit in leachatec

28 3

0.0 13 0.200 0.010 0.040

5.0 1100.0 1.o 5.0

0.001

5.0 0.2

47

13 28 4 4 6

0.00 1 0.003 0.060

1.o 5.0

From May and Sweeney (1983). From Nifong (1988) and May and Sweeney ( I 982). From the Federal Register (1980).

The United States EPA has set certain criteria to define corrosive, hazardous, and toxic waste (Federal Register, 1980). The EPA criterion for corrosivity is a pH 5 2.0 or 2 12.5. Its criterion for toxicity ofwaste is based on the kind of contaminants that are likely to leach into ground water. These are extracted from the waste material according to EPA extraction procedures (Federal Register, 1980). The hazardous nature of the waste is judged by the concentrations of specific contaminants in the extract. Table IV lists the contaminants and the EPA concentration limits in drinking water together with the mean concentrations in the leachates of Florida phosphogypsum. Radioactivity is the major concern in the use of phosphogypsum in both industry and agriculture. Radium-226, the decay product of uranium-238 in rock phosphate, is the major source of radioactivity in phosphogypsum. In the manufacture of phosphoric acid, 86% of 238Ugoes with the phosphoric acid and 8OYo of the 226Rastays with the phosphogypsum (Roessler, 1990; EPA, 1991). Radiation levels of Florida phosphogypsum range from 296 to 1406 Bq kg-’ with an average of 747 Bq kg-l (May and Sweeney, 1980, 1983). The higher the radioactivity in the raw material, the higher it is in the phosphogypsum (Table V). Some countries have established the following radiation characterization limits for monitoring phosphogypsum

I. S . ALCORDO AND J. E. RECHCIGL

64

Table V Computed and Measured Radiation in Phosphogypsum Produced from Rock Phosphate of Various Origins"

Rock origin

Measured radiation in rock (Bq kg-')

Florida

Morocco Togo Kola Natural gypsum

Radiation in phosphogypsum (Bq kg-I) Computed

Measured

1295.0 1387.5 1295.0 1 1 1.0

869.5 925.0 85 1.O 74.0

740.0 1147.0 740.0 74.0

-

37.0

-

From Kouloheris (1980).

(Kouloheris, 1980):

Radiation (Bq kg-')

< 370 370-925 >925

Regulations/characterization Free use Compulsory declaration Compulsory registration

The EPA regulations issued on December 18, 1978, placed phosphogyp sum on the list as a hazardous waste because of its radioactivity. To be excluded from the list, a solid waste material must contain less than 185 Bq of 226Rakg-', or, for a single discrete source, the total z2aRacontent must be less than 37 X 104 Bq (Federal Register, 1978). The EPA rule on phosphogypsum (Federal Register, 1989) placed the regulation of phosphogypsum stacks under the National Standards for Hazardous Air Pollutants. The final EPA rule on the removal of phosphogypsum from the stack for use in agriculture has been formulated only recently (Federal Register, 1992). The rule is based on a biennial application of phosphogypsum to agricultural land, which land is later converted to residential use after 100 years of phosphogypsum application. Under such an assumption the acceptable maximum individual risk (MIR) of lifetime exposure from indoor radon inhalation and y-radiation of 3 X

PHOSPHOGYPSUM IN AGRICULTURE

65

would not be exceeded if the phosphogypsum used contained not more than 370 Bq kg-' and the rate of application was approximately 3.0 Mg ha-'. Such a rate covers the upper 95th percentile of the application rates used in the United States. Removal of phosphogypsum from stacks for use in research and development is now also allowed but the amount cannot exceed 3 18 kg, the amount of phosphogypsum that can be contained in a 55-gallon drum. Other uses of phosphogypsum are presently prohibited without prior approval of the EPA.

11. USES OF PHOSPHOGYPSUM IN AGRICULTURE A. SOURCEOF S AND Ca FOR CROPS 1 . Sulfur Deficiency and Need for Ca Source Other Than Lime

Sulfur is essential to plant nutrition. In general, plants contain as much S as P, the usual range being from 0.2 to 0.5% on a dry-weight basis. Sulfur ranks in importance with N as a constituent of the amino acids cysteine, cystine, and methionine in proteins that account for 90% of S in plants. It is also involved in the formation of oil in crops such as peanut (Arachis hypogaeu L.), soybean [Glycine max (L.) Merr.], flax (Linum usitissimum), and rapeseed (Brassica campestris) (Tisdale et al., 1985). In the past three decades, S deficiencieshave been reported with increasing frequency throughout the world (The Sulphur Institute, 1982; Tisdale et al., 1986). Sulfur status reports for the United States and Canada (Tabatabai, 1986), Brazil (Jones, 1967), the Australasian region (McLachlan, 1979, India (Tandon, 1986; Takkar, 1986), China (Zhongqun, 1986), Indonesia (Ismunadji, 1986), Thailand (Keerati-Kasikorn, 1986), and Bangladesh (Mazid, 1986; Islam et al., 1986) indicate a growing need to meet S deficiency in these countries. The reasons given for the increasing S deficiencies worldwide are (1) the shift from low-analysis to high-analysis fertilizers containing little or no S, (2) use of high-yielding crop varieties that remove greater amounts of S from the soil, (3) reduced industrial S emission into the atmosphere due to pollution-control measures and decreased use of high-S-content fossil fuels, (4) decreased use of S in pesticides, and ( 5 ) declining S reserves in soil due to erosion, leaching, and crop removal. Increased consumption of sulfur-

66

I. S. ALCORDO AND J. E. RECHCIGL

free, high-analysis fertilizers is seen as the most important reason for the increasing S deficiency worldwide (Jordan, 1964; Moms, 1986). Calcium, with concentrations ranging from 0.2 to 1.OYo in plant tissues, is also essential to plant life. Calcium deficiency manifests in the failure of terminal buds and apical tips of roots to develop. Also, lack of Ca results in general breakdown of membrane structures, with resultant loss in retention of cellular diffusible compounds. Disorders in the storage tissues of fruits and vegetables frequently indicate Ca deficiency (Tisdale et al., 1985). The need for Ca by plants may be readily satisfied by liming materials such as calcitic and dolomitic limestone. However, lime application in large amounts on certain soils can be detrimental to plant growth. Kamprath (1971), in a review of the effect of lime on Oxisols and Ultisols, reported that lime application that raised the soil pH to 7 resulted in reduced rates of water infiltration, reduced availability of P, B, Mn, and Zn, and reduced growth of sudangrass(Sorghum vulgare var. Sudanese L.), corn (Zea mays L.), and soybean. Therefore, for certain soils that require large amounts of Ca to support commercially viable crop yields, or for crops such as peanut that need large amounts of readily soluble Ca, a Ca source other than lime may be necessary. Thus, with increasing S deficiencies worldwide and the need for a Ca source other than liming materials, phosphogypsum deserves serious consideration for agricultural applications where mined gypsum has traditionally been used. Because phosphogypsum and mined gypsum are chemically similar, studies on mined gypsum are also cited throughout this paper when they may be of assistance in evaluating phosphogypsum for a particular agricultural application. The present review highlights a scarcity of studies on phosphogypsum. Amacher and Miller (1 987), in assessing the use of phosphogypsum in agriculture, cited only eight papers on its use as a source of S and Ca for crops, all of which were conducted outside the United States. Shainberg et al. (1989), in their extensive review of the use of gypsum on soils, cited 10 studies on phosphogypsum published from 1983 to 1986, all of which were conducted in Brazil, and all pertained to its use in alleviating subsoil acidity rather than as a source of S or Ca for crops. The scarcity of studies on phosphogypsum use in agriculture worldwide could be due to the availability of other sources of S and Ca. In the United States, environmental concerns and government regulations hindered the use of phosphogypsum even for agricultural research (Federal Register, 1989). However, results of on-going environmental studies on phosphogypsum use in Florida (Rechcigl et al., 1992a) and in other southeastern states (Mullins and Mitchell, 1990) may help change existing regulations to allow for the use of phosphogypsum in commercial agriculture.

PHOSPHOGYPSUM IN AGRICULTURE

67

2. Cereal Crops

The majority of Coastal Plain soils in the southeastern United States contain relatively small amounts of extractable SO, sulfur in their sandy surface horizons (Jordan, 1964). On the basis of the results of numerous field tests of S, Reneau and Hawkins (1980) suggested that corn planted on Coastal Plain soils that are moderately well drained to well drained, low in organic matter, and fine loamy to coarser textured, with extractable soil S of 6-7 kg ha-' in the surface horizon, would probably respond to S application, therefore also to gypsum. In North Carolina, Rabuffetti and Kamprath (1977) concluded that the effect of gypsum on corn was highly dependent on the rate of N applied. At 56 or 112 kg N ha-', gypsum had no effect on corn yield or N content of grain. At 168 and 224 kg N ha-', gypsum at 30 and 60 kg S ha-' increased grain yield and N content of grain. Nitrogen had a greater effect than did gypsum on total S accumulation in grain and stover. In Florida, field trials using 1.68 to 2.24 Mg ha-' of phosphogypsum increased green corn yields by as much as 10790 in 1986 and by 5590 in 1987 (Hunter, 1989). Raw phosphogypsum was used in 1986 and pelleted phosphogypsum was used in 1987. Friesen and Chien (1986), citing studies camed out by the International Fertilizer Development Center ( 1985)in Togo, West Africa, reported that phosphogypsum at 10 to 50 kg S ha-' increased corn grain yields by 44 to 7790 over the control. In the same experiment, elemental S increased yields by 45 to 5490. In Iraq, 1.28 Mg phosphogypsum ha-' applied to corn growing on a calcareous soil increased yields by as much as 15090 over the control (Khalil et a!., 1990). Superphosphate at SO kg P ha-' also increased yields by as much as the phosphogypsum treatment, but P in phosphogypsum (0.18%)was insufficient to influence the yield. Oates and Kamprath (1985) found that gypsum was as effective as ammonium sulfate as a source of S for winter wheat (Triticum aestivum L.), Plants responded to gypsum at 22 to 90 kg S ha-' where nonfertilized plants had S concentrations of 0.6 g kg-' of dry matter and a total N : S ratio of 21 : 1. No response to gypsum was observed when there was appreciable root growth in the B horizon that contained an accumulation of SO, sulfur. Sulfate sulfur leached from the surface horizon tends to accumulate in the B horizon where large amounts of hydrated oxides of A1 and Fe are present (Beaton et al., 1974). Based on studies on leaching of SO, sulfur from various sources applied in the fall (Rhue, 1971;Rhue and Kamprath, 1973), Baird and Kamprath (1980) suggested that improved efficiency of S uptake by winter wheat should occur on sandy soils when gypsum is applied as a topdressing in early spring. In Bangladesh, Mazid (1986) reported that wheat yields from 1042 fertilization trials conducted

68

I. S. ALCORDO AND J. E. RECHCIGL

from 1981 to 1983 increased by an average of 21% due to gypsum applied at 20 kg S ha-'. Results from demonstration trials on the effect of 125 kg gypsum (16%S) ha-' on rice (0. sativa L.) carried out from 1981 to 1983 in Bangladesh showed that 97% of 3368 demonstration sites responded to gypsum (Mazid, 1986).Rice yields in gypsum-treated sites increased by 19 to 4 1% over those from sites with the recommended NPK fertilizer without gypsum. Crop responses to gypsum occurred mainly in calcareous and continuously submerged soils and were more profitable in the monsoon season than in the dry season. Studies in Indonesia indicated that ammonium sulfate, potassium sulfate, elemental S, and gypsum were equally effective as a source of S for rice (Ismunadji and Zulkamaini, 1978; Momuat et al., 1983). Chien et al. (1987), in a greenhouse study, demonstrated that response of rice to gypsum was not dependent on the method of application. Sulfur uptake and grain yields were not different whether gypsum was broadcast, incorporated, or placed deeply in the soil. 3. Grain Legumes

Peanuts possess a unique nutritional requirement in that supplemental Ca must be supplied to the "peg," a modified stem that penetrates the soil surface to form the pod or nut. Numerous experiments (Colwell and Brady, 1945; Hallock and Garren, 1968; Cox et al., 1976; Aha et al., 1989) have shown that supplemental Ca applied at flowering improved yield and quality of large-seeded peanuts. The role of Ca in reducing pod rot incidence in peanuts is also well known (Garren, 1964; Hallock and Garren, 1968; Moore and Wills, 1974; Porter el al., 1975). Walker and Csinos ( 1980)demonstrated that increasing rates of gypsum from 0.56 to 1.68 Mg ha-' resulted in corresponding reductions in pod rot in five peanut cultivars. As early as 1945, Colwell and Brady (1945) had established the superiority of gypsum over limestone in supplying the Ca requirements of peanut. Since then, the peanut-producing belt of the southeastern United States has used fine-ground (anhydrite)mined gypsum as the principal Ca source for peanut, broadcast at 0.5 to 1.0 Mg ha-' at first flowering on soils whose Mehlich I extractable Ca is <560 kg ha-' (Mehlich, 1953). This value is still the current critical soil test for Ca for the runner-type peanut in Georgia (Alva et al., 1989). Sullivan et al. ( 1974) showed that application of dolomitic limestone on peanuts, based on soil tests, increased soil pH and soil Ca levels but did not improve seed quality and yield. On the other hand, gypsum at 0.673 Mg ha-' reduced soil pH and the detrimental effects of K on fruit yield and

PHOSPHOGYPSUM IN AGRICULTURE

69

quality, improved seed germination, seedling survival and vigor, and increased yield and improved seed quality. Daughtry and Cox ( 1974) found that three commercial gypsum materials, namely, fine-ground and granular anhydrite gypsum and phosphogypsum, applied at 0.76 Mg CaSO, ha-' at flowering, produced no difference in the yield of the Florigiant peanut. Hallock and Allison ( 1 980) used similar commercially formulated fineground (Bagged LP) and granulated (420 LP Bulk) anhydrite gypsum, and granulated phosphogypsum [Texas Gulf (Tg) Gypsum], as sources of Ca for Virginia-type peanuts at 0.605 Mg ha-'. After 2 years of testing (1977 and 1978), the results indicated, in general, that granulated phosphogypsum and mined gypsum were as effectiveas fine-ground gypsum for supplemental Ca for peanuts. When fruit matured under very dry conditions, granulated phosphogypsum and fine-ground mined gypsum were superior over granulated mined gypsum. Gascho and Alva (1990) used seven gypsum materials, including phosphogypsum, as sources of Ca for Florunner peanuts. They concluded that no other source of gypsum exceeded phosphogypsum in solubility or in its beneficial effects on peanut grade and yield when broadcast at 224 kg Ca ha-' at first bloom. In Brazil, Vitti et al, (1986) found that application of 0.1 Mg ha-' of phosphogypsum to soybean on an Oxisol and on an Ultisol increased grain yield by as much as 43 and 37%, respectively. At 0.25 Mg ha-', phosphogypsum increased grain yield of beans (Phaseolus vulgaris L.) by 13%on an Ultisol and by 54%on an Oxisol. Phcsphogypsum rates used were very low, so that the positive responses of the crops could be attributed more to S or Ca as nutrients than to the ameliorative effect of phosphogypsum on subsoil acidity. 4. Sugarcane

Gypsum applied on sugarcane (Saccharurn oficinarurn L.) as a source of Ca on Ca-deficient soils in Hawaii increased cane yield as effectively as limestone and ordinary superphosphates (Ayres, 1962). In Rhodesia, application of 22 or 45 kg S ha-' of rock S, gypsum, and MgSO, to sugarcane increased yields by 44.3, 50.8, and 44.9 tons ha-', respectively (Gosnell and Long, 1969). Buselli (1988) found that mined gypsum at 22.4 Mg ha-' applied to sugarcane on an alligator clay in Louisiana increased stalk population and yield. Golden ( 1982) observed that the application of phosphogypsum at 1.12 and 2.24 Mg ha-' to sugarcane in Louisiana resulted in total increases in cane yields over a 4-year period of 18.26 and 24.64 Mg ha-', and sugar yields of 1.69 and 2.70 Mg ha-', respectively. Breithaupt (1989), using both phosphogypsum and fluorogypsum on sugarcane at rates ranging

70

I. S. ALCORDO AND J. E. RECHCIGL

from 2.24 to 22.40 Mg ha-', reported significant increases in cane and sugar yields in treated plots over the control in both plant cane and first-year stubble harvests. Both gypsum by-products were equally effective in increasing cane and sugar yields. 5 . Fruits and Vegetables

In Florida, phosphogypsum applied to citrus (Citrus sinensis) increased fruit yields with increasing rates up to 1.12 Mg ha-' on a Myakka soil (sandy, siliceous, hyperthermic, Aeric Haplaquods). It also increased juice brix, brix :acid ratio, and Ca content. In Oldsmar soils, phosphogypsum not only increased juice brix, brix:acid ratio, and Ca content but also reduced titratable acid (Myhre et al., 1990). In Brazil, pineapple [Ananas cornosus (L.) Merill, cv. Smooth Cayenne] fertilized with phosphogypsum in combination with KCl as a substitute for &SO4 gave fruit yields similar to those fertilized with K2S04. Fruits fertilized with K2S04, however, had better fruit juice quality than those fertilized with KC1 alone or in combination with phosphogypsum (Bianco et al., 1990). Use of raw phosphogypsum at 1.68 and 2.24 Mg ha-' on various vegetable crops in 1986 in Florida increased the yields of tomatoes (Lycopersicon esculenturn Mill) by 696, potatoes (Solanurn tuberosurn L.) by 19%, and watermelons (Citrullus vulgaris) by 49%.Residuals from phosphogypsum applied in 1986 at 2.24 Mg ha-' also increased the yields of potatoes by 2290 and cataloupes (Cucurnis rnelo) by 42%, with a greater number of melons weighing 1 .O kg or more each. Pelleted phosphogypsum applied to the 1987 crop did not increase the yields of potato and bell pepper (Capsicum annuurn). The phosphogypsum pellets remained intact, albeit soft, indicating only partial dissolution (Hunter, 1989). 6. Forage Crops

Thomas et al. (195 1) demonstrated conclusively that S deficiency limits nonprotein N utilization in purified diets for ruminants, and that SO, sulfur as sole source of S can correct the deficiency. Hume and Bird ( 1970) showed that an intake of 1.9 g S day-' by sheep produced the maximum protein production in the rumen, and that inorganic SO4sulfur was used as efficiently as that from cystine for synthesis of protein by rumen microorganisms. Bray and Hemsley (1969) showed that S supplement to the diet increased both crude fiber digestion and S and N retention by sheep. Moir et al. (1967 - 1968) demonstrated that narrowing the mean dietary N :S ratio of a basal ration for sheep from 12 : 1 to 9.5 : 1 increased the mean N

PHOSPHOGYPSUM IN AGRICULTURE

71

retention from 28.8 to 36.0%. Metabolic studies by Whanger ( 1968- 1969) supported the findings of Moir et al. (1967 - 1968). Rees et al. (1982) found that sheep ate substantially more S-fertilized digitgrass forage (Digitaria pentzii Stent) than forage not S fertilized. Akin and Hogan (1983) indicated that S fertilization did not affect plant anatomy of digitgrass but enhanced the fiber-digesting capability of the microbial rumen population. Application of 86 kg S ha-' using (NH&SO, on bahiagrass (Paspalum notatum Flugge) increased dry matter yield by 25%, crude protein by 1.2 percentage unit, digestibility by 3 to 4 percentage units 30 days after application, and S content by 100%. (Rechcigl, 1989; Rechcigl et al., 1989). On a larger scale, studies in Ireland (Murphy et al., 1983) showed that cattle that grazed on S-fertilized pastures gained up to 29% more weight than those that grazed on S-deficient fields. Also, for any given daily liveweight gain, S-treated areas had 21 and 19% greater stock-carrying capacity during the first year and the second year, respectively, than the untreated pastures. These studies point not only to the need for S fertilization of forage crops for yield but also to the need to achieve a desirable range of N :S ratios to assure better quality forage. In plant protein, the N :S ratio is about 15 : 1 and remains fairly constant. If either S or N is limiting, protein synthesis is restricted, but the protein already synthesized will have a N :S ratio of about 15 : 1. Excess N relative to S supply accumulates as NO, nitrogen, amides, and amino acids. Excess S leads to SO, sulfur accumulation (Stewart and Porter, 1969). Thus the wide variation in N :S ratios. Sulfur fertilization of forage crops almost invariably results in a reduced N : S ratio in plant tissue. Lancaster et al. (197 1) reported that application of S at 40 mg kg-I of soil in the form of Na,SO, reduced the N : S ratio as follows: from 32 to 9 for orchardgrass (Dactylis glomerutu L.); from 45 to 19 and 72 to 14 for first and second clippings, respectively, of sudangrass; from 36 to 5 for ryegrass (Lolium multiforum L.); from 27 to 8 for alfalfa (Medicago sativa L.); and from 33 to 16 for clover (Trifolium repens L.). On the other hand, results from an 8-year field experiment using bermudagrass [Cynodon dactylon (L.) Pers.] showed that despite S fertilization excessive N application could result in a forage crop with N: S ratios in excess of 60 : 1 (Woodhouse, 1969). In North Carolina, mined gypsum applied annually on coastal bermudagrass at the rates of 28 and 56 kg S ha-' increased forage yields in 7 of 8 years (Woodhouse, 1969). In Louisiana, Eichhorn ei al. (1990) reported that annual application of 108 kg S ha-', using gypsum, increased bermudagrass hay yield by 16% over a 4-year period, with the highest increase (2990)occurring in the fourth year. Digestible dry matter also increased by 14.5% over the same period. In Florida, Mitchell and Blue (1989) con-

72

I. S. ALCORDO AND J. E. RECHCIGL

ducted a 6-year study to evaluate the effect of gypsum applied annually on Pensacola bahiagrass at 200 and 400 N kg ha-'. They reported that at the low N rate, gypsum application did not increase dry matter yield until the fourth year, with maximum yields thereafter predicted at an annual S application between 27 and 33 kg S ha-l. At the high N rate, 10 kg S ha-I increased dry matter yield in the second year. By the fifth and sixth years, maximum dry matter yield was predicted at an annual rate of 40 to 5 1 kg S ha-'. Results also showed that S fertilization enhanced N recovery. Maximum relative forage yield was obtained at a concentration of 1.6 1 g S kg-' dry matter. In a l-year study in Oklahoma, application of gypsum at 64 kg S ha-' reduced the N: S ratio of bermudagrass forage from 1 1.6: 1 to 7.2 : 1 but did not increase yield, N uptake, or improve N efficiency (Westerman et al., 1983). To date few studies have been camed out on the use of phosphogypsum on forage crops. Paulino and Malavolta (1989) used phosphogypsum on andropogon grass (Andropogon gayanus cv. Planaltina) grown in pots with top soil taken from a Brazilian Cerrado site. Results showed that phosphogypsum, in the absence of lime, increased regrowth dry matter yields linearly up to the maximum rate of 120 kg S ha-' used in the study. Maximum protein content was attained at 63 kg S or 380 kg phosphogypsum hav1. Lime had a significant negative effect on andropogon grass. Mullins and Mitchell (1990) used phosphogypsum at 11 to 90 kg S ha-' on wheat cut for forage in Alabama. Average increases in forage yield over a 3-year period ranged from 5.4 to 9.3% for two soil series. Comparison between mined gypsum and phosphogypsum showed no difference in forage yield of wheat. Phosphogypsum applied during fall or spring had no residual effect on yield of millet [Setaria italica (L.) Beauv] or sudangrass planted for summer forage after the winter wheat crop. In Florida, use of fresh phosphogypsum as a source of Ca applied at 2.24 to 4.48 ton ha-' reduced soil pH and forage yield of ryegrass to levels below those of the control (Rechcigl and Payne, 1989). Fresh phosphogypsum can be very acidic, with pH only slightly over 2. Phosphogypsum was evaluated as a source of S and Ca for bahiagrass on a Myakka soil in a 3-year study (Rechcigl and Alcordo, 1991;Rechcigl et al., 1992a), without and with 1% dolomite or calcium carbonate needed to bring phosphogypsum pH (1 : 1) to 5.5. Annual rates of 0.2, 0.4, and 1.0 Mg ha-' were compared to single phosphogypsum application rates of 2.0 and 4.0 Mg ha-'. Results showed that phosphogypsum, with or without lime, increased the 2-year total forage dry matter yields of bahiagrass by as much as 28% at 0.2 to 0.4 Mg phosphogypsum ha-'. Phosphogypsum, across phosphogypsum rates, with dolomite gave the highest increase in dry matter yield, with 12% over the control.

PHOSPHOGYPSUM IN AGRICULTURE

73

B. AMELIORANT FOR ALUMINUM TOXICITY AND SUBSOIL ACIDITY 1. Acid Soils and Their Formation

Jackson ( 1963) grouped soil acidity according to the dominant protondonor constituents in the soil. The most important of these are the ( I ) free mineral acids such as H,SO,, (2) organic acids, and (3) active A1 and Fe. The corresponding examples of a soil dominated by each of these proton donors are (1) the sulfate soils of Thailand (Parkpian et al., 1991), (2) the virgin Spodosols of Florida (Carlisle and Fiskell, 1962; Pettry et al., 1969, and (3) the Oxisols and Ultisols of the tropics and subtropics (Kamprath, 1970, 1971). It is with the last group of acid soils that gypsum has found application as a subsoil acidity ameliorant. Acid soils, estimated at more than 800 million ha worldwide (Orvedal and Ackerson, 1972), constitute from 40 to 50% of potentially arable highly weathered soils (Sanchez, 1977). Typical pH in water is < 5.0 and pH in salt is on the order of 4.0 at the surface to a depth of 1 m (Shainberg et al., 1989). Acid soils are located primarily in the tropics and subtropics, where intense chemical weathering occurs. High temperature and rainfall are two factors that promote rapid weathering of primary as well as secondary A1- Fe - silicate minerals, releasing, in the process, basic cations as well as A1 and Fe into the soil solution. The basic cations react with anions to form highly soluble salts, carbonates, and hydroxides. Aluminum and Fe ions, in the presence of salts, hydrolyze into hydrated A1 and Fe with the release of H ions (Chang and Thomas, 1963; Jackson, 1963), acidifying the surrounding solution. In time, the hydrated A1 and Fe, by themselves or by reacting with the various soil constituents, resynthesize into amorphous or crystalline oxides and hydrous oxides (Singh and Brydon, 1969; Pariitt and Smart, 1978; Adams and Hajek, 1978). The mineral thus formed supports a characteristic level of active ionic species in equilibrium with it. The work by Hue et al. (1987) and Walthall and Day (1988) suggests that at a given pH halloysite supports the greatest equilibrium concentration of active A1 in soil solution, followed by gibbsite, kaolinite, and smectite, in that order. In areas where rainfall exceeds evapotranspiration, the highly soluble bases are leached to greater depths than are A1 and Fe, leaving behind soil horizons enriched in active A1 and Fe and compacted by their oxides and hydrous oxides. As the horizon becomes more acidic, more A1 and Fe are solubilized from primary minerals and from their secondary oxides and hydrous oxides (Magistad, 1925). This leads to increased A1 and Fe saturation of the exchange complex of the colloidal soil constituents and to subsoil infertility. When A1 saturation of the exchange capacity exceeds

74

I. S. ALCORDO AND J. E. RECHCIGL

60%, appreciable amounts of A13+start to get into the soil solution (Nye et al., 1961). At this point A1 toxicity, caused by subsoil acidity, could set in. Heavy fertilization, could induce A1 toxicity even in soils with a relatively low A1 saturation (Kamprath, 1970, 1971). Excessive cropping and use of acid-forming fertilizers, without proper liming, could only aggravate the condition (Beverly and Anderson, 1987). 2. Al Toxicity Indices for Subsoil Acidity Amelioration

Poor root penetration and proliferation frequently observed in the highly weathered acid soils of the southeastern United States (Pearson, 1966) have been attributed to physical (Bowen, 1981) and chemical factors (Rios and Pearson, 1964). The chemical factor identified as most responsible for poor root growth is excess soluble A1 (Rios and Pearson, 1964; Adams and Lund, 1966; Adams et al., 1967; Soileau and Engelstad, 1969). The phytotoxicity of excess A1 has long been recognized (Ligon and Pierre, 1932). Trivalent A1 has been reported to inhibit root growth by binding to the PO, portion of DNA in the root cell nuclei, reducing template activity and thus cell division (Matsumoto et al., 1976; Matsumot0 and Morimura, 1980; Horst et al., 1983). In legumes, it has been shown to impair the growth of root hairs and rhizobia, thus root nodule initiation and function (Munns and Franco, 1982; de Carvalho et al., 1982). Excess A1 may also adversely affect root as well as overall plant growth in nonphytotoxic ways by competing with Ca and Mg for uptake by plants (Rengel and Robinson, 1989), by precipitating with anion nutrients such as PO, (Plucknett and Sherman, 1963) and SO, (Singh and Brydon, 1969; Adams and Rawajfih, 1977) to render them less available to plants, and by supplying the soil solution with H ions, because exchangeable A1 (pH range 4.5-5.4) and hydroxy Fe and A1 (pH range > 5.2) act as buffers to keep the soil at a pH of < 5.4 (Coleman et al., 1964; Kamprath, 1970). Studies on the phytotoxicity of A1 are complicated by the hydrolysis and polymerization of trivalent A1 to form numerous mononuclear and polynuclear ionic species coexisting in the same solution. Trivalent A1 also reacts with various ligands to form several ionic species that remain labile in the solution (Cameron et al., 1986). These complexities have been somewhat surmounted with the availability of computer speciation models based on the thermodynamics of solutions (Sposito and Mattigod, 1980). Use of these models has led to a clearer picture of the phytotoxicity of the various A1 ionic species. Pavan et al. (1982) demonstrated that reduction in root growth of coffee (Cofea arabica L.) seedlings was best correlated with AP+ activity, and shoot and root weight correlated with KC1-extractable A1 and percent A1

PHOSPHOGYPSUM IN AGRICULTURE

75

saturation of the soil. Cameron et al. (1986), using barley (Hordeum vulgure L.), showed that A13+ concentration, but not total Al, correlated best with root elongation. In soybean, reduction in tap root growth was best correlated with the sum of the concentrations or calculated activities of monomeric A1 [AP+ A1(OH)2+ Al(0H)fl species (Blarney et al., 1983; Alva et al., 1986; Noble et al., 1988b). Strong correlations were also found between root growth and activities of either A1(OH)2+or Al(0H); with soybean, subterranean clover (Trifolium subterraneum L.), alfalfa, ~ u ~ L.) (Alva et al., 1986). Parker et a/. and sunflower ( H e l i ~ n t annus (1988), using wheat, confirmed previous studies on A13+as the best indicator of A1 stress on plants in the absence of toxic polymers. They also found polynuclear hydroxy A1 to be demonstrably toxic (Bartlett and Riego, 1972; Wagatsuma and Ezoe, 1985). These species have been ignored in most studies until now. However, they failed to confirm with wheat the correlation between root growth of soybean and the sum of the activities of monomeric A1 ions reported by Alva et al. (1 986) and others. Shuman et al. (1990), in a greenhouse study using sorghum [Sorghum bicolor (L.) Moench], found that the best predictors of plant height were soil solution A13+ activity ( r = -0.91), A1 saturation of the exchange complex ( I = -0.89), and 0.01 MCaC1,-extractable A1 (r = -0.78). Taking into consideration the ameliorative effect of Ca on A1 toxicity, Noble et al. (1988a) proposed the so-called Ca-A1 balance (CAB) index: CAB = [2 l o g ( a ~ )] [3 log(ac) 2 bg(a&,H)) log(a&,H)2)] as a predictor for potential A1 toxicity. The index was found to have a good correlation (R2= 0.88) with the root length of soybean. Shamshuddin et al. (1991) found that the soil solution A1 concentration, Ca- A1 ratio, activity of A13+and A1(OH)2+,and sum of monomeric A1 activities were highly correlated with corn and peanut yields growing on a Malaysian Ultisol soil. Thus, it may be concluded that for most agronomic plants A13+concentration or activity in the soil solution or A1 saturation in the exchange complex appears to be the best single measure to assess potential A1 toxicity for a given soil. For management of A1 toxicity with the use of lime and gypsum materials, the Ca-A1 balance index of Noble et al. (1988a) should be helpful.

+

+

+

+

3. Al Toxicity and Subsoil Acidity and Their Amelioration

For soils with serious subsoil acidity, surface application of lime may not be the practical answer to the problem. Metzger (1934) and Brown and Munsell (1938) reported that 10 to 14 years were required for surfaceapplied lime to increase soil pH to a depth of 15 cm. This is due primarily to the low solubility and mobility of lime in soils. The reaction of lime with

76

I. S. ALCORDO AND J. E. RECHCIGL

the soil moisture supplies the soil with both OH and HCO, . These, however, are immediately neutralized by H at the surface. Any increase in pH in the surface soil may also increase Ca adsorption due to the extra negative charges generated in the amphoteric soil constituents (Reeve and Sumner, 1972). Neutralization of the basic constituents of lime and the adsorption of Ca at the surface work to keep the ameliorative effect of lime limited to the surface (Reeve and Sumner, 1972; Pearson et al., 1973). Thus, not only is subsoil acidity not checked by surface application of lime, but it also fails to supply Ca to the deeper horizons to alleviate subsoil infertility. Adams and Moore (1983), investigating root growth in subsoil horizons of coastal Plain soils in the United States, found not only A1 toxicity in roots in the argillic horizons (Bt) but also a more prevalent Ca deficiency in both eluvial and illuvial horizons where Ca saturation was 5 17%. Heavy application of lime to overcome low solubility and downward mobility may prove deleterious to the physicochemical properties of these soils (Kamprath, 1971). Deep placement of lime to raise subsoil pH even in advanced countries has not found widespread use because of the cost involved. In developing countries, where the problem primarily exists, it is an unacceptable solution because of the prohibitive cost and the heavy equipment required. The use of gypsum and gypsum by-products, alone or in combination with other chemical or mechanical treatments to enhance their ameliorative efficiency, appears to be the most practical approach to the worldwide problem of subsoil acidity and infertility. To alleviate A1 toxicity to plant roots, excess A1 had to be precipitated out from the soil solution or complexed into something less toxic than the previous ionic forms and, ideally, leached out of the root zone. Precipitation studies of A1 and Fe from AlCl, and FeCl, solutions, respectively, had shown that A1 and Fe were easily precipitated out of the solution onto clay surfaces by OH supplied by addition of NaOH (Alcordo, 1968). In the absence of clay, similar precipitation procedures produced gibbsite or bayerite from A1 and goethite from Fe solutions (El-Swaify and Emerson, 1975). Titration of A12S04solution with NaOH, KOH, or Ca(OH), also precipitated A12S04.On aging, the precipitate gave a chemical composition similar to basaluminite, A14(OH),oS045H,O (Singh and Brydon, 1969), and, with KOH as base, to alunite, KAl,(OH),(SO,), (Adams and Rawajfih, 1977; Adams and Hajek, 1978; Sin& and Miles, 1978). Aluminum in solution, in the presence of large concentrations of SO4 ions, forms a complex ion pair, Also,+ (Pavan et al., 1982; Cameron et al., 1986; Noble et al., 1988b;Alva and Sumner, 1989). Cameron et al. (1986), Kinraide and Parker (1987a,b), and Noble et al. (1988b) showed that roots do not appear to be adversely affected by the Also,+complex ion. Fluoride is also an important element in A1 toxicity amelioration. Triva-

-

PHOSPHOGYPSUM IN AGRICULTURE

77

lent A1 in solution in the presence of F readily forms an AlF+ ion pair. Not only is this complex nonphytotoxic (Sumner, 1990), but it is also highly labile, therefore leachable (Keerthisinghe et al., 1991). Phosphogypsum and fluorogypsumcontain considerable amounts of F. Oates and Caldwell (1985) showed that amounts of A1 leached from soil columns treated with mined and gypsum by-products were in the order fluorogypsum> phosphogypsum > mined gypsum. Adsorption of SO, in acid soils plays the key role in amelioration of subsoil acidity. Chang and Thomas (1963) were the first to propose a comprehensive mechanism to account for SO, sorption in soils. Assuming a homoionic Al-saturated clay coated with hydrated oxides of Fe and A1 (R), two opposing reactions were proposed: A1 and/or Fe hydrolysis [ Eq. (l)] and ligand exchange [Eq. (2)]:

-

+ H,O + yK+ Clay-AIJOHgK + H+ Clay-RJOH), + S@Clay-R,[(OH),-#304), + zOH-

(1) (2) According to these reactions, the hydrolysis of exposed and/or adsorbed A1 and Fe at the clay edges and/or surfaces, in the presence of salt (a K salt, for example), releases H ions into the solution, causing Al-induced (or Feinduced) acid conditions. In the case of a SO, salt, the more negatively charged SO, readily replaces OH in a ligand exchange. The OH then neutralizes H in the solution, causing a so-called self-liming within the system (Reeve and Sumner, 1972; Sumner, 1990). The implication of the above reactions to subsoil acidity amelioration using gypsum or phosphogypsum is clear. Application of gypsum to ameliorate subsoil acidity supplies the system with excess SO, salt. The presence of excess salt drives the system for more hydrolysis of A1 and Fe, which, in turn, drives the system for more SO, adsorption. Whether the pH of the system is ultimately raised or lowered depends not only on the the relative rates of the opposing reactions of hydrolysis and ligand exchange but also on the supply of SO, and Al. Excess SO, over that of the sesquioxides is expected to result in increased pH at the end of the reaction. Excess A1 and/or Fe over that of SO, should return the pH to the initial equilibrium pH before gypsum application. Also, where Al and Fe exceed SO,, an endpoint pH below the initial may also result due to the exposure of new sesquioxide surfaces that are more reactive to hydrolytic reactions. But large and sustained application of gypsum to acid subsoils should eventually increase soil pH (Keng and Uehara, 1974). The ultimate effects of gypsum on the system are (1) the dissolution of precipitated A1 and Fe on the surfaces of the silicate clays, which are known to reduce permanent cation exchange capacity (CEC), and/or (2) the release of exposed A1 at the edges of the crystal lattice. Dissolution of Clay-Al,

78

I. S. ALCORDO AND J. E. RECHCIGL

precipitated A1 and Fe should unblock this permanent negative charge arising from isomorphous substitution and should increase CEC (de Villiers and Jackson, 1967). This is in addition to the pH-dependent and ionic-strength-dependentCEC generated by ligand exchange between SO, and OH and by the presence of dissolved gypsum. These increases in CEC with proper fertilization should help build up subsoil fertility. 4. Reactions of Weathered Acid Soils with Gypsum and Phosphogypsum

Kamprath et al. ( 1 956) found that soils with 1 : 1 clays adsorbed more SO, than those with 2 : 1 clays. Adsorption was found to be directly related to concentration in the solution, and decreased as pH increased from 4 to 6. For all soils studied, increasing the PO, concentration in the solution reduced the amount of SO, adsorbed by the soils. Chao ( 1964), studying 26 inorganic and organic anions, found that PO, and F reduced SO, sorption of soils used by 44 and 30%, respectively. Hydroxyl and HCO, anions increased the pH and consequently decreased SO, adsorption. The effect of NO3 and SiO, appeared to be pH independent throughout the pH range from 4 to 6. Gebhardt and Coleman (1974) stated that the affinity of SO, for soil in a ligand exchange is at least 10 times that of nonspecifically adsorbed NO, and C1. Chao et al. (1963), in an equilibration study, found that cations in SO, salt influenced the sorption capacity of soils in the order CaSO,> K2S04> (NH,),SO, > Na2S0,. With respect to the cations saturating the exchange complex, the order was A13+> Ca2+> K+. As the pH of the solution approached neutrality, SO, sorption decreased regardless of the saturating cation. The greater the exchangeable Al, sesquioxide, and amorphous materials in the soil, the greater the pH effect. Rajan (1978, 1979), working with allophanic clays, found that SO, adsorption released OH ion in a linear relationship. He proposed that on a net positive surface, SO, is adsorbed as a bidentate, forming a six-member ring and displacing either two aquo or OH ligands. On a neutral or negative surface, SO, is adsorbed as a monodentate, displacing one aquo or one OH, making the surface more negative. A similar bidentate ligand for SO, adsorption in Fe oxides such as goethite, kaganite, lepidocrocite, hematite, and amorphous ferric hydroxide was proposed by Pariitt and Smart ( 1 978). Increases in soil pH due to SO, had been observed under laboratory and field conditions (Chao et al., 1965; Bornemisza and Llanos, 1967). Couto et al. (1979) showed pH increases of up to 0.5 units when Brazilian soils were treated with SO, solution. Ritchey et al. (1980) reported increases in

PHOSPHOGYPSUM IN AGRICULTURE

79

pH by as much as 0.8 units following gypsum treatment in Dark Red Latosol (Typic Haplustox, fine, kaolinitic, isohyperthermic) in soil columns and field studies. Farina and Channon (1988) reported that gypsum increased water pH markedly in the zone of maximum SO, sorption/precipitation. Others reported no change in pH with gypsum application (Hammel et al., 1985; Oates and Caldwell, 1985; Sumner et al., 1986). Black and Cameron (1984) reported that gypsum decreased pH. Gypsum application consistently reduced A1 saturation (Ritchey et al., 1980; Farina and Channon, 1988; Alva et al., 1990, 1991). Increases in the CEC in soils treated with SO, salt had also been observed (Hue et al., 1985; Alva et a/., 1990). These increases had been attributed to specific adsorption of SO, (Gebhardt and Coleman, 1974; Alva et al., 1990; Gillman, 1991). Also, the surface charge of variable-charge minerals generally increases with increase in the ionic strength of the solution, such as when gypsum or phosphogypsum is applied to the soil (van Raij and Peech, 1972; Keng and Uehara, 1974). The leachability of CaSO, in soils is one of the key elements in the use of gypsum or phosphogypsum in subsoil acidity amelioration, which, no doubt, is related to SO, mobility. Gaston et al. (1986) wrote an excellent review of SO, mobility in acid soils. Rhue (1971) did not find any SO, movement into the subsoil in a silty clay loam (Typic Hapladult) from 56 kg SO, sulfur ha-', using gypsum, more than 200 days after application. Rhue and Kamprath (1973) attributed this to the high SO, adsorptive capacity of the soil. Relative to lime, leaching studies had consistently shown the high mobility of gypsum in soils. Rhue and Kamprath (1973) reported that SO, from gypsum, applied at the surface of a sandy loam soil (Arenic Paleudult) at the rate of 56 kg S ha-', was totally leached from the top 45 cm 180 days after application under a total rainfall of only 40 cm for the period. Reeve and Sumner (1972), in a leaching tube study, found that surface-applied gypsum was completely leached out from the top 10 cm after application of 32 cm of water. At the topsoil level, gypsum was less effective than lime in totally eliminating exchangeable Al. Gypsum, however, increased subsoil base status and reduced exchangeable A1 in the subsoil much more effectively than did Ca(OH)2. The data of Shamshuddin et al. (1991) showed that gypsum, incorporated in the top 15 cm of a Malaysian Ultisol at 2.0 Mg ha-', reduced the activity of A13+in the soil solution by 43% and reduced the sum of the activities of monomeric Al, excluding Also:, by 60% 3 months after application. Lime applied at the same rate, however, reduced all A1 ionic species much more effectively than did gypsum. Korentajer et

80

I. S. ALCORDO AND J. E. RECHCIGL

al. ( 1984)found SO, losses from powdered gypsum were three times that of powdered anhydrite. Combination of lime and gypsum application to acid soils resulted in a greater leachability of SO, (Korentajer et al., 1983). The high solubility of gypsum, however, may cause severe loss of exchangeable Mg from the surface horizon (Reeve and Sumner, 1972; Pavan et al., 1984; Syed, 1987). OBrien and Sumner (1988) noted a similar adverse leaching effect of phosphogypsum on K, but to a lesser degree than on Mg. Because high levels of K and Mg had been observed to have adverse effects on peanut yield and quality (Brady and Colwell, 1945; Hallock and Garren, 1968), Alva and Gascho ( 1991) used gypsum to leach K and Mg differentially to reduce their concentrations at the fruiting depth of the peanut. 5 . Crop Response to Gypsum and Phosphogypsum on Acid Soils

Failure of plant roots to grow into and proliferate in deeper soil horizons on acid soils due to A1 toxicity limits the capacity of the plant to take up both plant nutrients and moisture. Highly weathered soils such as the Oxisols and Ultisols, whose mineralogy is normally dominated by 1 : 1 types of clay and oxides and hydrous oxides of A1 and Fe, not only retain very little moisture in the surface horizons after a rain but also dry out very quickly during short periods of rainless days. Wolf (1975) reported that in the Cerrados of Central Brazil corn crops can wilt after only 6 days without rain, even during the wet season. Ritchey et a!. (1980) noted that soils from several field experiments conducted on a Brazilian Savannah Oxisol sampled after 3 to 4 years from the date of application of gypsum-containing ordinary superphosphate (OSP), up to 873 kg P ha-', showed increased subsoil pH, Ca and Mg content, and decreased A1 saturation at depths as great as 75 to 90 cm. Roots of corn plants fertilized with OSP were able to extract moisture to a depth of 120 cm and enabled the plants to withstand a 2-week drought. Plants fertilized with triple superphosphate (TSP) reached a depth of only 45 cm and wilted during the 2-week drought. Pavan et al. (1984) used 0- to 30-cm depth bulk soil samples from six uncultivated acid Ultisol and Oxisol soils from southern Brazil as media for coffee seedlings. Soils were treated with gypsum, CaCO,, or MgCO, . Coffee seedlings were grown for 7 months in pots. After this period, the plants were harvested for root and top growth measurements. Results showed that in Oxisol and Ultisol, the top growth of coffee seedlings was as good in gypsum-treated soils as in the C0,-treated soils, and all had better top growth than the control. Root growth in gypsum-treated Oxisol soil was not as good as in the C0,-treated soils but it was better than the

PHOSPHOGYPSUM IN AGRICULTURE

81

control. Root growth in gypsum-treated Ultisol was intermediate between the control and the C0,-treated soils. In Georgia (Hammel et al., 1985), gypsum treatment at 35 Mg ha-' incorporated to a depth of 15 cm was compared with physical mixing of the subsoil to a depth of 1 m without lime, and physical mixing with lime. The control received no physical or chemical modification. The soil used belonged to the family of Typic Hapladults. Results showed that surfaceapplied gypsum increased soybean grain yield by 26% in the second year and corn silage yield by 35% in the third year compared to the control. The average 2-year grain soybean yield in gypsum-treated plots was 38% higher than that in physically mixed plots, whereas corn silage yield in the third year was 36% higher. Physical mixing plus lime gave the highest yields for both soybean and corn. This treatment, however, may not be economical, and surface-appliedgypsum appeared to be the most practical under largescale operation. The results of the study also showed that chemical modification of the profile is more important than physical in improving crop yields in these soils. Sumner et al. (1986), using similar treatments at a lower gypsum rate of 10 Mg ha-' on alfalfa (also on a Typic Hapladult), obtained results similar to those of Hammel et al. (1985). Mean dry matter yield of alfalfa over a 4-year period was 30% higher in gypsum-treated plots than in the control. Physical modification of the profile without lime depressed the yield whereas physical mixing with lime gave the highest yield. Gypsum at 10 Mg ha-' applied to corn on a limed, strongly acidic Plinthic Paleudult increased grain yield by an average of 19% over that of the limed control over a 3-year period. Progressive reduction in the level of exchangeable A1 was accompanied by increased subsoil Ca, Mg, and SO, sulfur. Water pH increased markedly in the zone of maximum SO., sorption/precipitation. Effects of gypsum on subsoil root development of corn were striking by the fourth cropping season (Farina and Channon, 1988). In Virginia, Rechcigl et al. (1988a) compared the effect of surfaceapplied and incorporated gypsum and lime at 13 Mg ha-' on alfalfa grown in an Aquic Fragiudult. The results showed consistently that surfaceapplied gypsum produced as much dry matter as did surface-applied lime, and both had higher mean yields than did the control across two rates of N. Incorporated lime gave a higher yield over surface-appliedgypsum during the first year, although both treatments gave similar yields by the second year. Gypsum, applied at the rate of 10.0 Mg ha-' to imgated alfalfa on a fine-loamy, siliceous thermic Typic Paleudult in Alabama, increased dry matter from 7.2 to 9.9 Mg ha-', or 37.5% (Odom,1991). Published and unpublished studies on the use of phosphogypsum as an ameliorant for acid soils in Brazil, South Africa, and the southeastern

82

I. S. ALCORDO AND J. E. RECHCIGL

United States up to 1986 have been summarized by Shainberget al. ( 1989). Rates ranging from 0.5 to 6.0 Mg ha-' of phosphogypsum significantly increased the yields of apples (Malus domestica), beans, coffee, rice, wheat, and corn. Sumner (1990) evaluated gypsum and phosphogypsum applied at 5 to 10 Mg ha-' incorporated into the soil in several field experiments on a range of soils in southeastern United States. The results indicated that there were no differences between the two gypsum sources based on crop responses and soil reactions. Highly significant and economically profitable yield responses were obtained for alfalfa, corn, soybean, cotton (Gossypium hirsutum L.), and peaches (Prunus persica L.). Gypsum and phosphogypsum application enhanced root penetration and proliferation in the subsoil, where previous conditions often prevented root growth. In Louisiana, Caldwell et al. (1990) reported that by-product gypsum applied on acid Giger silt loam in 1986 increased cotton yields in 1987, 1988, and 1989. Surface soil pH decreased from 6.0 to 5.6 at 8.8 Mg by-product gypsum ha-' even when lime was applied at half the rate of gypsum. Due to high N application, subsoil pH decreased in all plots, with gypsum-treated plots showing the largest decreases. In lime-treated plots, increase in exchangeable Ca was mainly at the top 15 cm and none at 30 cm. In by-product gypsum-treated plots, increases in exchangeable Ca were noted at 30 cm during the first year, at 45 cm during the second year, and below 60 cm during the third year. Exchangeable A1 was little affected by the treatment. Although most studies have reported a positive response of crops to gypsum and phosphogypsum application, it should be noted that some greenhouse studies reported reduction in yield of top growth (Fried and Peach, 1946) and root growth (Simpson et al., 1979) of alfalfa. However, perusal of the data of Fried and Peach (1946) suggested that Mn toxicity may have been responsible for low yields of alfalfa that lime, but not gypsum, was able to correct. The study by Simpson et al. (1979) showed that the adverse effectsof gypsum on root growth appeared to be related to increased acidity in soil pH due to gypsum application. Wright et al. (1985) found that leaching gypsum into the subsoil did not improve barley root growth, even though subsoil-exchangeableA1 was reduced. Some possible reasons for these conflicting results of gypsum use on acid soils are (1) exchangeable A1 levels may not always reflect the potentially toxic levels of A1 in the soil solution phase (Adams, 1978; Rechcigl et al., 1988b; Mathews and Joost, 1989), (2) specificity of the phytotoxicity of the A1 ionic species to plants (Alva et al., 1986; Parker et al., 1988), the distribution of which is not known in simple measurement of the changes in subsoil exchangeable Al, and (3) the specific tolerance of the plant species, and even of the different genotypes of the same species (Simpson et al., 1979), to A1 toxicity (Kamprath, 1970).

PHOSPHOGYPSUM IN AGRICULTURE

83

C . AMELIORANT FOR SODICSOILS 1 . Characteristics of Sodic Soils

In regions of the world where evapotranspiration exceeds rainfall, soluble salts move upward in the soil profile from the water table, instead of downward as occurs in regions of acid soils. In this region the dominant soil clay mineral is montmorillonite, but illite and vermicullite are also common (Shainberg et al., 1989). Soils containing various soluble salts and exchangeable Na at levels that interfere with the growth of most crops are classified as saline sodic or simply sodic soils. The E, of the saturation extracts of these soils exceed 0.2 S m-* at 25°C with Na adsorption ratio (SAR) = Na/[Mg Call’* > 15. The saturation extracts of ordinary saline soils have the same E, value as a sodic soil but with an SAR < 15 (Kilmer, 1982). The characteristic physical property of sodic soils is their dispersivity in water due to Na ions in the exchange complex of the colloidal fraction, particularly the silicate clays. When placed in water of low salt concentration, aggregates from sodic soils imbibe water until they deflocculate into individual particles (Hendricks et al., 1940; Nomsh, 1954; Russell, 1973). In the field during rain, surface aggregates are first dispersed by the impact of raindrops. Some of the dispersed particles move down with the percolating water and clog the pores (McIntyre, 1958; Gal et al., 1984; Shainberg and Singer, 1985), forming a “washed-in” crust 2 to 3 mm thick. Continued raindrop impact may result in the formation of a thin skin of compacted soil particles 0.1 mm thick, almost without pores, over the soil surface (McIntyre, 1958). As clogging of the surface pores and compaction proceed, water percolation through the profile is reduced or completely stopped. In addition to physical dispersion, chemical dispersion also occurs, not only in the surface aggregates but also in the soil matrix, the extent of which depends on the exchangeable Na percentage (ESP) of the soil and the electrolyte concentration of the percolating water (McNeal and Coleman, 1966; McNeal, 1968; Frenkel et al., 1978; Agassi et al., 1981). On drying, hard crusts develop at the surface, making seedling emergence difficult (Davidson and Quirk, 1961). Poor water percolation during rain or imgation and hard surface crusting on drying are the two major problems that need to be ameliorated if sodic soils are to be used for crop production. Figure 6 graphically illustrates the relationships between swelling, disaggregation, and dispersion on the one hand and electrolyte concentration on the other hand. Figure 7 illustrates the progressive breakdown of aggregates of different ESP values with electrolyte concentration (Rowell, 1965; Rowell et al., 1969; Russell, 1973).

+

I. S. ALCORDO AND J. E. RECHCIGL

84

L

a,

Q

NaCl normality Figure 6. Measured and calculated swelling of a Na montmorillonitein NaCl solutions. (From Russell, 1973.)

For practical applications, the effects of gypsum or phosphogypsum on sodic soils are measured by the changes in the saturated hydraulic conductivity (HC = QL/H m sec-l), where Q (m3m-2) is the amount of effluent, L (m) is the length of the soil column, and H (m) is the equilibrium height of the water (Klute and Dirksen, 1986), or by changes in the infiltration rate [IR (m sec-l)] (Parr and Bertrand, 1960), by seed emergence or germination and establishment, and by crop yield. Studies reviewed in this paper on the use of gypsum and phosphogypsum on sodic and other dispersive soils are limited to these practical aspects. 2. Gypsum or Phosphogypsum for Sodic Soil Reclamation

Reclamation or amelioration of saline and sodic soils involves the conversion of Na clays into Ca clays. Sodium saturation of the exchange complex can develop large swelling pressures between clay platelets, leading to dispersion; sodium montmorillonite, in dilute solutions of low ionic strength, tends to disperse as single platelets (Norrish, 1954). Calcium

85

PHOSPHOGYPSUM IN AGRICULTURE

t

0 10 O

10

-’

10 -2

10 -3

10 -4

Electrolyte concentration (moles/liter) Figure 7. The effect of exchangeable sodium percentage on the electrolyte concentration at which different types of structure collapse occuxs. (From Russell, 1973.)

montmorillonite tends to form tactoids, packets, or aggregates of four to nine platelets, limiting the swelling potential of the clay to between these aggregates (Norrish and Quirk, 1954; Blackmore and Miller, 1 96 1 ;Shomer and Mingelrin, 1978; Greene et al., 1978). Gypsum is used worldwide to supply Ca for sodic soil reclamation or amelioration. The pioneering studies by Hilgard (1 907) on the toxicities of Na,CO,, NaCl, and Na,SO, to plants, studies by Kelley and Brown ( 1 934) on gypsum, and other studies in the United States have fully established gypsum as a sodic soil ameliorant. Related work by the staff of the U.S. Salinity Laboratory, U.S. Department of Agriculture (Richards, 1954), led to the present recommendation that for each mmol, of Na that needs to be replaced and leached from a 15-cm depth, 2.016 Mg of CaSO, or 0.358 Mg S are required per hectare (Kilmer, 1982). Theoretical models to predict gypsum and leaching requirements of Na-affected soils had been formulated on the basis of kinetic (Melamed et al., 1977; Glas et aE., 1979) or equilibrium (Dutt et al., 1972; Tanji et al., 1972) chemistry. The kinetic approach assumes a time-dependent

86

I. S. ALCORDO AND J. E. RECHCIGL

precipitation- dissolution reaction and is represented by a kinetic rate equation. The equilibrium model assumes that the solid phase and the solution phase are in equilibrium at all times. Comparisons between measured and predicted values obtained by computer simulation analysis were much more successful using equilibrium-based models (Tanji et al., 1972; Dutt et al., 1972) than using kinetic-based models (Melamed et al., 1977; Glas et al., 1979). Oster and Frenkel(l980) used the equilibrium model to derive a simple, practical relationship between the gypsum dissolved and the exchangeable Na replaced. They found that for endpoint ESP values of 0.05,O.10,and 0.15,the slopes (gypsum dissolved/Na replaced) were 1.40, 1.27, and 1.20,respectively. Also, only as reclamation approached completion in the region where gypsum was present did the gypsum that dissolved begin to replace exchangeable Na at greater depths. The solubility of dihydrate gypsum in distilled water at 25 "C is 0.01 5 1 m CaSO, (2.1g liter-I), increasing, in the presence of NaC1, to 0.056 rn (7.84g literw1)in 3 m NaCl (Marshall and, Slusher, 1966). Keren and Shainberg (1981) showed that the dissolution coefficient [K = -(In( 1 C,/C,))/t, where C, and C, are the concentrations of gypsum dissolved at time t and at saturation, respectively] increased as NaCl concentration increased. The ratios between the KNaa of gypsum in NaCl and Kw in water (KNaa/Kw) were 1.28 and 1.45 for 0.05 and 0.10 N NaCl, respectively. These values indicated increased solubility of gypsum with increased NaCl in the solution. Analytical or mined gypsum and phosphogypsum, when compressed into disks to expose the same surface area to a solvent, showed similar solubilities and solubility rates. However, when fragmented materials ( 1 .O- 2.0 and 4.0- 5.7 mm in diameter) were used, &h..ph.BYp.was found to be 10 times that of Kmiatdm. Thus, where gypsum is supplied to farm lands via flowing water (Kemper et al., 1975; Keisling et al., 1978),the use of phosphogypsum in place of mined gypsum would cut by a factor of 10 the contact time between the bed of gypsum and the irrigation water for any desired concentration of dissolved gypsum. Also, where gypsum is applied on the field at a mixing depth of 0.5 cm and irrigated at the rate of 1 cm hr-l, the use of phosphogypsum instead of mined gypsum would increase electrolyte concentration threefold for applications of less than 5 Mg ha-' and twofold at higher rates of gypsum application (Oster, 1980). 3. Use of Gypsum and Phosphogypsum in Sodic Soils

Work by Kelley and Brown (1934)on the reclamation of the so-called black alkali (sodic) and the white alkali (saline) soils in California had established gypsum as the material for sodic and saline soil reclamation. In

PHOSPHOGYPSUM IN AGRICULTURE

87

Canada, Graveland and Toogood ( 1963) concluded from laboratory studies that gypsum was an effective amendment for Alberta solonetzic soils where Na did not exceed 20-25% of the CEC, assuming that the effective solubility of gypsum in the field was one-half that of its saturated concentration. In Australia, Barley and Hutton (1956) showed that application of 4.48 and 17.9 Mg gypsum ha-' to a fine-textured saline alkali soil before growing a lowland rice crop increased the Ca :Na ratio of both soluble and exchangeable cations and reduced clay dispersion. It also resulted in better rice seedling establishment and higher yields of stubble-sown subterranean clover. In Louisiana, gypsum applied at 7 Mg ha-' to a soil normally inundated with brackish water helped to establish marshland vegetation. Under flooded conditions, gypsum increased the survival rates of marsh hay cordgrass (Spartina patens Muhl.) from 41.5 to 91.990; salt grass (Distichlis spicata L.), from 49.7 to 92.7%; and American three square (Scirpus americanus Pers.), from 33.6 to 85.0%(Sigua and Hudnall, 1991). Phosphogypsum has been effectively used in the former USSR to reclaim solonetz and solonetzic soils, with 3.2 million Mg used in 1988 for this purpose. Its use is expected to reach 19.2 million Mg by the year 2000 (Novikov et al., 1990). Mishra ( 1980), summarizing phosphogypsum research in India, which began in 1973, concluded that up to 32 Mg ha-' of Indian phosphogypsum can be used safely for reclamation of sodic soils, despite the high F content. Keren and Shainberg (1981), in a simulated rain (27 mm hr-l) experiment on a sodic soil, found that phosphogypsum at 3.4 to 6.8 Mg ha-' maintained a much higher IR than did mined gypsum.

D. AMELIORANT FOR NONSODIC DISPERSIVE SOILS,SUBSOIL -PANS, AND HARD-SETTING CLAYSOILS 1.

Al,Fe, Gypsum, and Phosphogypsum in Aggregation and Aggregate Stability

Next to organic matter, A1 and Fe in various forms have been strongly associated with soil aggregate formation and aggregate stability (Quirk and Panabokke, 1962; Aylmore and Quirk, 1967; Russell, 1971, 1973). Lutz (1936) credited stable soil structure of oxidic soils to irreversible dehydrated ferric hydroxides. Deshpande et al. (1 968) found that the apparent correlation between structure stability and iron oxide contents was really a reflection of a still closer correlation with free A1 oxides. Fruhauf et al. (l962), using viscometric measurements, showed that aggregation of kaolinite was influenced by the degree of A1 saturation and

88

I. S. ALCORDO AND J. E. RECHCIGL

the A1 :OH ratio. Alcordo (1968) showed that the viscosity of pure A1 montmorillonite or Fe montmorillonite and A1 soil or Fe soil clay suspensions of mixed clay mineralogy, prepared by adding AlCl, or FeCl, to the clay suspensions, hydrolyzing A1 or Fe with NaOH, and dialyzing the suspensions to remove excess salt, was extremely sensitive to pH. The A1 montmorillonite suspension coagulated to a thick, almost nonflowing paste at the strong acid side, indicating strong particle- particle and particle - water molecule interactions, which progressively liquified with increases in pH. Fe montmorillonite behaved similarly, but with a lesser degree of coagulation. The effect of Al and Fe on the viscosity of the soil clays was stronger in montmorillonitic than in kaolinitic clays. El-Swaify and Emerson (1975), precipitating Fe(OH), and Al(OH), on illite and kaolinite, reported that both inhibited double-layerswelling in Na illite and Na kaolinite/illite. Trihydroxy Al [Al(OH),] was found to be more effective than Fe(OH), in reducing slaking of dry clay disks and in increasing resistance to dispersion in pyrophosphate solution. Shainberg et al. (1987) found that the HC of a sandy loam soil (Typic Haploxeralfs) that was equilibrated with a SAR 20 CaC1,-NaCl solution and leached with distilled water was faster for soils treated with 20 and 40 mmol FeCl, or AlCl, kg-l than for untreated soils or for soils receiving 10 mmol AlCl, or FeCl,. The Fe treatments were more effective than A1 treatment in maintaining the HC. Analysis of the column effluent showed that more A1 from the Al-treated soil was removed than was Fe from the Fe-treated soil. The stabilizing effect of Fe polymers on the soil was related to the charge of the polymer and not to the total amount extracted. Keren and Singer (1989), using clay-sand mixtures, found that no clay was leached from the column in the presence of Al polymers in excess of 25% of clay CEC when leached with solution of SAR 20. Iron polymers produced similar effects but to a lesser degree. Both A1 and Fe treatments increased the size of stable microaggregates compared to the untreated clay. The effect of pH on the HC of sand-clay mixtures showed that the HC was reduced at pH 7.5 but not at pH 5.5 when hydroxy-A1 polymers were present. A hydroxy Fe polymer showed a similar effect but to a lesser degree. The effect of pH on charge density of the hydroxy polymers explained their effectiveness in stabilizing the system at the lower pH (Keren and Singer, 1990). An A1 polymer prepared from AlCl, * 6H,O and NH,OH, when added to imgation water flooded on a sodic soil, significantly increased the HC and the IR rate at steady state (Elamin and Suarez, 1990). Keren and Singer ( 1991) showed that reduction in HC was much less when A1(N03), was added prior to rather than after NaOH, when A1 polymers were able to interact with the clay particles to reduce swelling and dispersion. Dispersion and transport of clay from the clay-sand mixture

PHOSPHOGYPSUM IN AGRICULTURE

89

occurred only when NaOH was added prior to Al( NO,), , causing the A1 polymer to precipitate from the clay as a separate phase. These studies suggest the following mechanism by which A1 and Fe promote aggregation and aggregate stability. The trivalent A1 and Fe first saturate the exchange complex of the clay and cause the collapse of the double layer to form large tactoids. On addition of NaOH, polymerization of the trivalent ions proceeds between the platelets, and the larger the size and the charge of the polymer, the larger and the more stable the tactoids. As Shainberg et al. (1987) had shown, these microaggregatesremained stable even with a SAR 20 when leached with distilled water. A “self-liming” mechanism had been proposed to explain increases in pH and reductions of exchangeable A1 (or Fe) in acid soils on gypsum application; the mechanism proceeded by the following reactions (Reeve and Sumner, 1972; Sumner, 1990):

+ 2AP+ + 2Fe3+

3Ca(OH), [3Ca(OH),

--

2AI(OH),

+ 3Ca2+

2Fe(OH),+ 3Ca2+]

These reactions, which are similar to the Chang and Thomas (1963) model of SO, adsorption, indicate that the amounts of A1 or Fe released into the solution would depend on the amounts of gypsum applied. The A1 ions released may then be hydrolyzed into trihydroxy A1 and/or be polymerized into other charged species. The more positively charged and the larger these polymers are the greater would be their aggregating effect on the silicate clay minerals, especially on the 2: 1 types. This effect would be strongest at the acidic side, where they are strongly positively charged (Jackson, 1963; Keng and Uehara, 1974). Conversely, these polymers are expected to have a dispersive effect at the pH where they are very strongly negatively charged (Arora and Coleman, 1979).Chiang et al. (1 987) cited a study that reported that raising the pH of a Georgia kaolinite from 7 to 8.3 made the soil more dispersive than any samples, including smectites, illites, and vermiculites. Precipitation of A1 and Fe may also affect aggregation. Leaching studies (Buyeye et al., 1985; Sumner et al., 1986; OBrien and Sumner, 1988) showed that little if any measurable Al was removed from reconstituted soil profiles. Mathews and Joost (1990) also found an apparent lack of A1 movement down the profile with phosphogypsum or langbeinite (K,S04 - 2MgS0,) application. They attributed this to SO4-induced precipitation mechanism such as the formation of alunite or basaluminite

90

I. S. ALCORDO AND J. E. RECHCIGL

(Adams and Rawajfih, 1977; Sumner et al., 1986). The removal of A1 and Fe from the solution by precipitation would cause the release of more A1 from the crystal lattice to maintain equilibrium (Dahlgren et al., 1989). This new supply of A1 and Fe may again be precipitated at the faces between platelets and between tactoids to enhance further aggregation along nonelectrostatic bonds. Tama and El-Swaify ( 1978) concluded from their study of charge, colloidal, and structural stability relationships for oxidic soils that field structural behavior depends not only on electrostatic forces but also on nonelectrostatic bonds within soil aggregates. Thus, phosphogypsum with its acid impurities and higher solubility rate should promote aggregation and aggregate stability more effectively than would mined gypsum (Oster, 1980). 2. Use of Gypsum and Phosphogypsum in Nonsodic Dispersive Soils

The surface soils of the red-brown earths of Australia, which have a much lower SAR than the United States standard for sodic soils, had been observed to disperse spontaneously under field conditions (Rengasamy et al., 1984). Like the red-brown earths of Australia, some soil series in the southeastern United States have highly dispersive surface soils (Hendrickson et al., 1963; Reicosky et al., 1977; Chiang et al., 1987), despite their high kaolinite and sesquioxides contents (Perkins et al., 1982; Shainberg et al., 1989). The effect of gypsum on the physical properties of nonsodic dispersive soils has been investigated in both heavy clay and coarse-textured soils. Sedgley (1962) found that structural pores formed by 1- to 2-mm-diameter aggregates showed least consolidation on drying when flooded with a solution containing 5 cmol Ca liter-' using gypsum. Loveday and Scotter (1966) concluded on the basis of their data that the higher the clay content, the ESP, and the evaporative potential of the soil, the greater will be the seed emergence response to dissolved gypsum application. Scotter and Loveday ( 1966) explained that improved seedling emergence in gypsumtreated soils was associated with the maintenance of a higher moisture content at the top 1.25 cm of the soil surface for longer periods of time after irrigation. Bridge and Kleinig (1968) applied 10 Mg ha-' gypsum to a coarse-textured soil that had a sandy loam A horizon underlain by a sandy clay B horizon of low HC. Field measurements showed that gypsum application significantly increased the soil moisture storage before and after irrigation at all irrigation frequencies but had no affect on the bulk density of the soil profile. Laboratory measurement also showed that gypsum increased the HC of the subsoil. Agassi ef al. (1982), using five soils from Israel, showed that plots with

PHOSPHOGYPSUM IN AGRICULTURE

91

phosphogypsum simply spread over the surface maintained the highest IR, compared to those wherein phosphogypsum was mixed with the surface soil and to the control. The IR of phosphogypsum-treated plots remained high whereas that of the control approached zero as the cumulative "rainfall" exceeded 60 mm. Phosphogypsum at 3 Mg ha-' was as effective as 5 and 10 Mg ha-' at keeping IR high. Kazman et al. (1983) showed that phosphogypsum at 5.0 Mg ha-' kept a high IR on a Natanya sandy loam soil with an ESP of 11.6 under a simulated rainfall of 26 mm hr-l, even beyond a cumulative rainfall of 70 mm. At this ESP, the IR of the control approached zero at 30 mm of cumulative rainfall. Similar results had been obtained on South African Alfisols (du Plessis and Shainberg, 1985).Agassi et al. (1985) also demonstrated under field conditions that phosphogypsum spread over the surface at 5.0 to 10 Mg ha-' effectively reduced soil erosion and runoff in wheat fields at various ranges of rainfall. In Georgia, Miller (1987) reported that surface-applied ( 5 Mg ha-') phosphogypsum doubled the final IR of Typic Hapladults and reduced soil loss by 30 to 50%. Miller (1989) concluded that increases in water intake and reductions in runoff and soil loss in runoff were primarily the effects of ionic strength. The high electrolyte concentration (0.5 to 1.3 dS m-I) in the runoff maintained by phosphogypsum kept soil clays flocculated, thereby reducing crusting. 3. Effect on Subsoil Hardpans

Besides A1 toxicity, physical bamers in the soil profile can also limit root penetration and proliferation. These include natural hardpans, dense textural B horizons, and tillage pans formed by heavy farm machinery (Bowen, 1981). Various crops, planted on dense phosphatic clay on reclaimed mined land overlain with sand, were observed to develop dense fibrous roots at the sand-clay interface with only a few primary anchor roots penetrating the heavy clay zone (I. S. Alcordo, personal communication, 1989).Calcium and SO4that leach into deeper horizons may not only ameliorate A1 toxicity and subsoil acidity, but may also improve subsoil structure to permit greater root penetration and proliferation. Radcliffe et al. (1986) showed that cone index values that measured mechanical impedance were lower on gypsum-treated plots that had been cropped for years than on the fallowed plots. They concluded that gypsum increased subsoil root activity, which in turn reduced subsoil mechanical impedance. Sumner et al. (1990) presented mechanical impedance and aggregate stability data that they claimed clearly demonstrated that both mined gypsum and phosphogypsum contributed to conditions in highly weathered soils leading to improved penetration of subsoil hardpans by roots. Gypsum materials affected subsoil hardpans by influencing flocculation and aggre-

92

I. S. ALCORDO AND J. E. RECHCIGL

gation, thus leading to better root penetration and proliferation. These, in turn, improved further subsoil aggregation. The role of pH- and SO,induced A1 and Fe precipitation in subsoil aggregation and structural development should be explored to expand the potential uses of phosphogypsum. 4. Effect on Hard-Setting Clay Soils

In Australia, Sims and Rooney (1965) analyzed years of studies on the use of gypsum to improve the surface physical conditions and internal drainage of stiff red soils, heavy red clay loams, silty loam soils overlying a dense clay layer, and grey clay loams, all highly dispersive soils located at different locations or districts. They observed substantial improvement in the physical condition of these soils when gypsum, at 2.24 to 4.5 Mg ha-', was applied to the surface of fallowed land in the spring of the year before cropping. Increased yields due to gypsum treatments resulted from better seed germination, better moisture supply (which allowed the crops to withstand long periods of dry conditions), and improved plant nutrition. They also noted that between 1963 and 1965, an estimated 44,500 ha of fallow soil were treated with gypsum to improve dryland wheat yields in the Wimmera and Southern Mallee districts of Victoria, Australia. A general relationship established by Loveday ( 1974) between crust strength, HC, and seedling emergence for a range of dispersive soils led to the implementation of the recommendation on the use of gypsum on hardsetting wheat soils in Australia (Howell, 1987). According to Davidson and Quirk (196 I), the most efficient method of applying gypsum was to dissolve it in the first irrigation water to obtain flocculation rather than Ca saturation. They reported that gypsum applied in such a manner led to more friable surface soil, increased the rate of water entry, and resulted in increased yields due to better pasture establishment of subterranean clover, white clover, ryegrass, and sudangrass. Rates of application ranged from 0.56 to 4.48 Mg ha-'. Collings (1980) reported that phosphogypsum is also being used to ameliorate heavy clay soils in Canada. In Asia, most, if not all, of the lowland rice soils are heavy clay soils that are puddled during land preparation to permit hand transplanting of seedlings. Puddling physically destroys soil aggregates. Reduction during continuous flooding solubilizes the cementing agents of Fe oxides and hydrous oxides (Kawaguchi and MatSUO, 1956; Takai and Kamura, 1966; Alcordo, 1968; Harmsen and van Breemen, 1975), further breaking the finer aggregates. In areas where rainfall or irrigation water is not sufficient to support a summer lowland rice crop, rice fields are left idle because of the extreme difficulty in

PHOSPHOGYPSUM IN AGRICULTURE

93

preparing for them as a dryland crop. Fields cake and the surface cracks into large lumps. The use of gypsum or phosphogypsum to induce fine cracking at the surface of these soils on drying should be explored together with the economics of the practice. When mixed with the clay at the last puddling, gypsum or phosphogypsum should induce fine cracking at the surface, with friable soil lumps due to the ionic strength effect of the dissolved gypsum and the nonplastic property of the undissolved gypsum particles.

E. BULKCARRIER FOR MICRONUTRIENTS AND Low-ANALYSIS FERTILIZERS Micronutrients B, Cu, Mn, Zn, and Fe are applied to soils to meet crop needs in relatively small amounts. Obtaining uniform distribution of small amounts is difficult. This difficulty is surmounted by bulk blending micronutrients with granular fertilizers. From 1950 to 1980, the market share of bulk-blended fertilizers increased from zero to more than 50% of all classes of fertilizers (Harre and White, 1985). It is expected to continue to increase as finer delineation of the fertility status of agricultural lands is achieved, requiring more custom-analysisblended fertilizers. Bulk-blended fertilizers use high-analysis fertilizers such as urea for N, which for claycoated agricultural grade is 46% N, triple superphosphate with 20% P, and KC1 with 48% K. Environmental considerations such as NO3 in drinking water and eutrophication of surface waters, due to enrichment from runoff and leached N and P fertilizers, may necessitate the use of locally blended low-analysis fertilizers applied more frequently than at present. Phosphogypsum, where readily available, provides a potential bulk camer for micronutrients and low-analysis fertilizer formulations. Pelletized phosphogypsum, enriched with micro- and macronutrients, has shown promise with urea and sulfate of potash magnesia (Hunter, 1989) as pelletizing agents. Also, phosphogypsum mixed with urea at 2.3 times the weight of the latter has been found to reduce NH, loss by 85% (Bayrakli, 1990).

111. ENVIRONMENTAL CONSIDERATIONS A. EFFECTSON SURFICIAL GROUND WATER One of the main concerns for the use of phosphogypsum in agriculture, particularly for sodic soil reclamation, where large amounts are needed, is the possible contamination of surficial aquifers or ground water. This is of

94

I. S. ALCORDO AND J. E. RECHCIGL

particular concern in developing countries, where most drinking water in rural areas is obtained from surficial aquifers. Studies on the leaching potential of Florida phosphogypsum showed that its leachates contained As, Ba, Cd, Cr, Pb, Hg, Se, and Ag at concentrations of 0.013, 0.2, 0.01, 0.01, 0.001, 0.003, and 0.06 mg liter-I respectively (Nifong, 1988; May and Sweeney, 1982). These concentrations, however, were below the EPA concentration limits (Table IV). Contamination of drinking water by these metals from phosphogypsum applied at agricultural rates is not expected to be a hazard based on the results of their leaching potentials in pure phosphogypsum. In Florida, Rechcigl et al. ( 1992a) applied phosphogypsum, with 4.3 g F kg-' and 671.5, 900.2, and 1139 Bq kg-I of 226Ra,210Po,and 210Pb, respectively, to a bahiagrass pasture on an Aeric Haplaquod at 0.0,0.4,2.0, and 4.0 Mg ha-' on August 15,1990. Results showed that phosphogypsum had no effect on pH but increased the E, and F contents of runoff, and surficial ground water at depths of 60 and 120 cm collected immediately after each heavy rain during the rest of the year. The highest E, value of 0.0974 S m-I obtained from 60-cm wells 90 days after application was less than the maximum E, of potable water in the United States, which is 0.150 S m-' (American Public Health Association, 1985). Similarly, the highest value of 0.69 mg F liter-' obtained from runoff was less than the drinking water limit of 1.4 mg F liter-' (Nifong, 1988). The highest 226Ra,210Po,and 210Pbconcentrations for all water samples obtained from treated plots at a depth of 90- 120 cm were 0.085, 0.078, and 0.028 Bq liter-', respectively. These values were no different from the highest values obtained from untreated plots, which were 0.059,0.063, and 0.0 18 Bq literp1for 226Ra,210Po,and 210Pb,respectively. The 226Ravalues were less than the maximum allowed for drinking water, which is 0.1 1 1 Bq liter-' set by the 1976 EPA Interim Primary Drinking Water Regulations (American Public Health Association, 1985). The limit for the sum of 226Raand 228Rais 0.185 Bq liter-'. The EPA (Federal Register, 1991), while setting the maximum contaminant level goal (MCLG) of zero for both 226Raand 22sRain drinking water, has proposed to raise the maximum contaminant level (MCL) for each radionuclide to 0.74 Bq in all community and nontransient, noncommunity public water systems. No recommendation has yet been made by the EPA to include the radioactivities contributed by 210Poand 210Pb.

B. EFFECTSON SOILS In Alabama, Mays and Mortvedt (1986) incorporated phosphogypsum, with 0.23 mg Cd kg-* and 925 Bq kg-I 226Ra,into a soil (coarse-silty, siliceous, thermic Glossic Fragiudults) at 0, 22, and 1 12 Mg ha-' to deter-

PHOSPHOGYPSUM IN AGRICULTURE

95

mine the effects of disposing phosphogypsum on agricultural land. The soil was sampled after a corn - wheat - soybean crop sequence. Their results showed that lzaRaincreased from 34.8 to 73.3 Bq kg-I at the 0- to 15-cm depth but had no effect at the lower depths. Mullins and Mitchell (1 990) reported that application of phosphogypsum, with 777 Bq kg-', at 0, 22, and 45 kg S ha-' to winter wheat had little or no effect on l16Raand 210Po in soil. Values obtained for l16Ra ranged from 3.7 to 13.7 for the control and 6.7 to 12.2 Bq kg-' for the phosphogypsum-treated plots. Values for 210Powere below the analytical detection limit. Rechcigl et ul. (1992a) found that phosphogypsum applied up to 4 Mg ha-' on a bahiagrass pasture had no effect on the levels of soil radionuclides. The highest soil radionuclides values from the 4.0-Mg plots up to a depth of 0.9 m sampled at 15-cm depth sections were 25.9, 29.1, and 55.5 Bq kg-' compared to those of the control plots with 25.9, 24.1, and 45.1 Bq kg-I of llaRa, llOPo, and llOPb, respectively. For agricultural land use, the National Council on Radiation Protection and Measurements (NCRPM, 1984) has suggested a limit of contamination for 226Raand lloPb at 1.5 and 0.7 kBq kg-', respectively.

C . EFFECTSON CROPTISSUES According to Grunes and Allaway (1985), Se has been the cause of economically important toxicities and deficiencies in farm animals. Se deficiency in ruminants results in white-muscle disease, and SO., sulfur at 0.33% of the diet has been observed to increase the incidence of this disease (Hintz and Hogue, 1964). This could be aggravated by gypsum and phosphogypsum application to soils in light of the observation that S application can diminish Se uptake by plants (Singh and Singh, 1979; Spencer, 1982; von Uexkull, 1986), resulting in a forage low in Se but very high in S content. Molybdenum can also be adversely affected by gypsum or phosphogypsum application. Gupta (1969) reported reductions in Mo content in brussel sprouts (Brussicu oleruceu var. gemmiferu) with S treatment with no change in soil pH. Fluoride is considered a cumulative poison. Intake of small quantities over an extended period may lead to the accumulation of F in animals to toxic levels (Campbell and Lasley, 1969). Fluorides applied at 25 to 200 mg kg-' of soil weight as NaF increased F contents of rice and wheat (Singh et ul., 1979a,b). Pot experiments showed that F content of bahiagrass increased dramatically with F application in excess of 125 mg kg-', reaching a maximum of 250 for NaF and 50 mg kg-' for phosphogypsumpond F (Institute of Food and Agricultural Sciences, 1980). Sodium F

96

I. S. ALCORDO AND J. E. RECHCIGL

applied to subterranean clover at 25,50, and 100 mg F kg-l soil resulted in F contents in the plant tissue of 32.6, 5 1.8, and 72.6 mg F kg-' dry matter, respectively, across three rates of P. Fluoride content was highest at the highest rate of P of 100 kg ha-'. Application of phosphogypsum and gypsum at 2.5, 5, and 10 Mg ha-* gave similar F contents in the plant tissue, ranging from 19.1 to 22 mg F kg-' at all rates of application (Keerthisinghe et al., 1991). The highest F content in bahiagrass fertilized with phosphogypsum up to 4.0 Mg ha-' at first harvest was 13.5 mg F kg-', which decreased with subsequent harvests to less than 8 (Rechcigl et al., 1922a). These values are way below the 30 mg kg-' of diet, the longterm ingestion of which could lead to fluorosis in bovines after a period of 2 to 3 years (CEurch, 1979). Cadmium may also be present in substantial quantities in phosphogypsum. Mays and Mortvedt (1986), however, showed that Cd contents of corn, wheat, and soybean grains fertilized with 22 and 112 Mg phosphogypsum ha-' containing 0.23 mg Cd kg-I were not different from those of the control. An extensive review of radium uptake from the soil by plants has been written by Simon and Ibrahim (1990). Radionuclide uptake is influenced by the plant species, soil properties, and the chemical form of radionuclide. This uptake may be described in terms of the concentration ratio, or CR (CR = concentration in dry plant tissue/concentration in oven-dry soil). When plant tissue is expressed in fresh-weight basis, the ratio is termed "soil-to-plant-transfer factor" (Till and Meyer, 1983). It is normally assumed that a linear relationship exists between the radionuclide concentration in the soil and in the tissue of the plant growing in that soil. Report 77 of the NCRPM (1984) cited a study of 11 types of roots and leafy vegetables grown in soil contaminated with uranium tailings; a linear relationship was observed between the 226Raconcentration in the plant tissue and in the soil. Lindekin and Coles (1977) reported that 226Rauptake from soil (18.5 Bq kg- l) by broccoli (Brassica oleracea) and turnips (Brassica rapa) was lower by a factor of 2 in soil with 5.2 g Ca kg-' compared to another soil with 3.1 g Ca kg-'. Mislevy et al. (1989) showed that biomass plants growing on two phosphatic clay soils had different levels of z26Raafter growth for a year. Elephant grass (Penniseturn purpureurn L. "PI 300086") and leucaena [Leucaena leucocephala (Lam.) De Wet] contained the lowest concentrations, averaging 3.33 and 4.8 1 Bq kg-'. Desmodium (Desrnodiurn cinerascens A. Gray), with 13.7 Bq kg-', had the highest 226Racontent. On the average, plants grown on mined phosphatic clay containing 899.1 Bq kg-I had 8.5 Bq kg-' in their tissue, whereas those grown on unmined soil containing 8.9 Bq kg-' had 1.48 Bq kg-'. Thus, mined phosphatic clay,

PHOSPHOGYPSUM IN AGRICULTURE

97

which had 101 times more radioactivity than the unmined soil, increased radioactivity in the plant tissue by only six times as much. The mean CR values for mined and unmined soils for these biomass crops were 0.009 and 0.170, respectively. These data indicate not only that uptake was nonlinear but also that plants had their limit to take up l16Ra. Guidry et af. (1990), in an exhaustive study of the relationships among food plant species, soil properties, and levels of radionuclides, reported that for l16Ra, strong positive relationships exist between lz6Rain crop tissue and lz6Rain soil, soil pH, and soil organic matter content. A negative, but not so strong, relationship was found for CEC, Mg, and Ca. Food zlOPbwas also positively correlated with soil 210Pb,but there was no clear-cut relationships between the soil properties previously mentioned and food lloPb. There was no correlation between food 210Poand soil 210Po. Mays and Mortvedt (1986) showed that phosphogypsum at 0, 22, and 1 12 Mg ha-' had no effect on the concentrations of l16Rain corn, wheat, and soybean grains. Phosphogypsum-fertilized crops at 112 Mg ha-l had 0.03, 0.83, and 0.37 Bq kg-' (the controls were 0.04, 0.46, and 0.84 Bq kg-') for corn, wheat, and soybean, respectively. Mullins and Mitchell (1990) reported that 226Raand 210Poconcentrations in wheat forage fertilized with phosphogypsum were no different from those of control. Rechcigl et af. (1992a) observed that l16Ra and llOPo contents in bahiagrass tended to increase with increases in phosphogypsum application, but the differences between the treated plots and the control were not significant. There was no trend between phosphogypsum rate and uptake in the case of lL0Pb.The mean values for the highest rate of application for bahiagrass were 1.1 1 , 8.29, and 4.70 Bq kg-I (the controls were 0.33, 1.37, and 7.51 Bq kg-l) for l16Ra, 210Po,and ll0Pb, respectively. Radium-226 concentrations in annual ryegrass at 2.0 and 4.0 Mg phosphogypsum ha-' were marginally significantly higher than in the control. The highest mean radioactivity levels in ryegrass were 5.90, 5.11, and 5.90 Bq kg-I for the treated plots (the controls were 3.63, 1.89, and 3.63 Bq kg-l) for lz6Ra, 2LoPo,and ll0Pb, respectively. According to Myhre et af.(I 990), phosphogypsum up to 2.24 Mg ha-' had no effect on l16Ra concentration in citrus fruit juice. The highest concentration observed in a sample was 0.1 1 1 Bq liter1, whereas the mean was 0.064 Bq liter1. Fruit juice from trees grown on Myakka soil had the highest l16Ra. There was no relationship between z2aRacontent in fruit juice and the rate of application of phosphogypsum. Lindekin and Coles (1977) and Lindekin (1980) concluded from their study that there is little basis for concern regarding the radiological hazard from the uptake of 226Raby plants from phosphogypsum as used in agriculture. Doses due to intake of food radionuclides under several sets of conditions have been calculated by Guidry et af.(1990).

98

I. S. ALCORDO AND J. E. RECHCIGL

D. EFFECTS ON AMBIENT ATMOSPHERE Radium-226 in phosphogypsum has a half-life of 1620 years; it decays to radon-222, a gas with a half-life of 3.8 days. Radon-222 decays into a series of short-lived radioactive progenies according to the sequence 222Rn+ *‘*Po 214Pb *I4Bi-+ 2 1 4 P which ~, is governed by a half-life of 27 min (Roessler, 1988; Molinari and Snodgrass, 1990). In the decay process, an a-particle (a helium nucleus) may be emitted, resulting in a product with an atomic number four less than its immediate progenitor, or a /?-particle (electron), with or without y-radiation. When 222Rnand its progenies are inhaled, they may deposit in the respiratory system, where they can irradiate the bronchial and lung tissues. At higher rates of exposure, as is the case among uranium miners, they may cause lung cancer. Cumulative exposure to lower concentrations of u2Rn and its progenies may present a risk of lung cancer (Roessler, 1987). The concentration of radon progeny is measured by a unit called “working level” (WL; WL/ 1000 is mWL) using air grab samples according to the procedure of Rolle (1972). One WL is the concentration of radon progeny in equilibrium with 7.4 Bq radon liter-I of air (Florida Institute of Phosphate Research, undated) or any combination of short-lived radon decay products in 1 liter of air that will result in the emission of a-particles with a total energy of 130 billion eV (EPA, 1991). Measurements of radon progeny concentrations inside a weighing room (3.6 X 7.2 m) where 0.68 Mg phosphogypsum was stored was made in the morning after radon gas was allowed to accumulate overnight. The results gave a mean radon progeny concentration of 12 mWL, which was six times higher than that of the air outside. In the field, measurements of the WL at the workers’ breathing zone during a 3-hr application of phosphogypsum by hand broadcast showed a mean concentration of 0.10 mWL (Rechcigl et al., 1992a). These measurements show that workers handling phosphogypsum for agricultural application are exposed to no more radioactivity than people are normally exposed to in Florida. The average background radiation from radon decay products inside the homes in Florida is about 4 mWL. A WL reading of more than 20 mWL inside a well-sealed house calls only for a long-term study before a radon reduction measure may be undertaken. At this low level, reduction has been found to be difficult to achieve (FIPR). Radon flux (Bq m-2 sec-l) is the amount of radon gas emanating from the ground surface. It is measured using a large-area activated-charcoal canister (Hartley and Freeman, 1985) according to EPA guidelines (Federal Register, 1989). Rechcigl et al. (1992b) measured radon flux over bahiagrass plots fertilized with 0, 0.4, 2.0, and 4.0 Mg ha-’ and found no

- -

PHOSPHOGYPSUM IN AGRICULTURE

99

Table VI Radon Surface Flux on Treatment Plots of Bahigrass Fertilized with

Phosphogypsum Measured Using the Large-Area Activated-Charcoal Canister Method" Radon surface flux (lo-' Bq m-* sec-I) following phosphogypsum application Date (mo/d/yr) 0.0 Mg ha-'

0.4 Mg ha-'

2.0 Mg ha-'

4.0 Mg ha-l

10/30- 3 1/90 02/19 - 20/9 1 04/10- 11/91

1 .3ab 1h a 1.Oa

1.2a 1.3a 1.Oa

1.4a 1.7a 1.5a

1.5a 1.la I .3a

Mean

1.3a

I .Oa

I .5a

1.3a

" From Rechcigl ef al. (1992b). bMeans across phosphogypsum rates with the same letter are not different at the 0.05 level of probability.

difference in radon fluxes between the treated and untreated plots (Table VI). In the same experiment, the ambient atmospheric radon (Bq liter-') over treated and untreated plots and over two external control locations not contiguous with the treated plots were also measured 1 m above the ground using electret ion chambers (EICs); these were corrected for background y-radiation (Kotrappa el al., 1988; Matuszek, 1990; Rechcigl et al., 1992b).The results for background y-radiation are given in Tables VII and for ambient atmospheric radon in Tables VIII and IX. Table VIII shows that rates of phosphogypsum up to 4.0 Mg ha-' had no effect on ambient atmospheric radon. Although not significant, the trend in ambient atmospheric radon among locations (Table IX) was in the order of treatment plots > control location 1 > control location 2. Control locations 1 and 2 were 0.05 and 3.2 km away from the edge of the treated plots, respectively, Note, however, that y-radiation values, which had to be deducted from the radon EICs to obtain the net ambient radon effect on the EICs, showed a trend in the reverse direction. All ambient radon values were all within the annual outdoor concentration range of 6.0-21.0 X lo-' Bq liter-' obtained in a national survey conducted by EPA (Hopper et al., 1990). The EPA risk assessments for all possible pathways by which phosphogypsum and other normally occurring radioactive material (NORM) wastes could affect the individual, as well as certain well-defined populations, are given in the EPA ( 1991) report.

I. S. ALCORDO AND J. E. RECHCIGL

100

Table VII Background y-Radiation over Treatment Plots Fertilized with Phosphogypsum Using Electret Ion Chambers"

y-Radiation @R hr')following phosphogypsum application Date (mo/d/yr) Preapplication 0611 1190-07/16/90 01/16/90-08/21/90

Mean

0.0 Mg ha-'

0.40 Mg ha-'

2.0 Mg ha-'

4.0 Mg ha-'

4.30 4.58

-

4.44

-

-

-

4.44 f 0.20

5.20a 4.60a 3.95a 5.60a 4.20a 4.55a 4.15a

4.41 0.51 4.20 f 0.36 4.38 & 0.40 5.03 f 0.40 4.50 f 0.21 4.31 f 0.23 4.29 f 0.33

4.69a

4.41 f 0.11

-

Postapplication

08/24/90-09/24/90 09/24/90- 10/30/90 10/30/90- 11/21/90 11/27/90-02/13/91 02/20/91-0312919 I 04/09/9 I -051 13191 05/11/91 -01/16/91 Mean

4.00ab 3.85a 4.45a 4.10a 4.45a 4.40a 4.10a

4.30a 4.40a 5.00a 4.50a 4.50a 4.00a

4.40a 3.95a 4.15a 4.80a 4.84a 4.05a 4.30a

4.28a

4.45a

4.44a

-

Mean

-

*

From Rechcigl ef al. (1992b). Means across phosphogypsum rates with the same letter are not different at the 0.05 level of probability.

IV. CONCLUSIONS Based on this review of the literature, phosphogypsum appears to be as good as, if not better than, mined gypsum as a source of S and Ca for crops. Most of the studies showed that surface-applied gypsum or phosphogypsum ameliorated subsoil A1 toxicity, acidity, and infertility in shorter time periods than did surface-applied lime materials. Phosphogypsum may prove to be superior to mined gypsum as an ameliorant for subsoil A1 toxicity, acidity, and infertility and as a conditioner for sodic soils, hardsetting clay soils, and subsoil hardpans because of its much higher rate of dissolution compared to that of mined gypsum. Gypsum and phosphogypsum, where they are readily accessible, are potential bulk carriers for micronutrients and low-analysis fertilizers. Increasing environmental demands to prevent contamination of ground water with nitrates and to minimize applied N and P losses in runoff, losses that promote rapid eutrophication of surface waters, may require the use of

PHOSPHOGYPSUM IN AGRICULTURE

101

Table VIII Ambient Atmospheric Radon Concentrations Measured 1 m above Treatment Plotsu ~~

Date (mo/d/yr)

Radon concentration (lo-’ J3q liter1)following phosphogypsum application

0.0 Mg ha-l

0.4 Mg ha-l

2.0 Mg ha-’

4.0 Mg ha-I

~~

Preapplication 0611 1/90-07/16/90 07116/90-08/27/90 Mean

-

-

-

-

5.2ab 1.4a 10.7a 8.9a 8.9a 4.4a 3.3a

5.2a 7.8a 10.7a 10.7a 14.8a 7.4a 4.8a

4.4a 6.7a 8.5a 9.3a 6.7a 5.2a 5.9a

5.2a 7.0a 10.7a 9.3a 5.5a 5.5a 2.2a

7.0a

8.9a

6.7a

5.9a

8.1 5.2 6.7

-

Postapplication 08/24/90 - 09/24/90 09/24/90- 10/30/90 10/30/90- 11/27/90 1 1/27/90-02/13 19 I 02/20/9 I -0312919 1 04/09/91 -05/13/91 051 1719I -071 16191 Mean

“Plots were fertilized with phosphogypsum; radon was measured using electret ion chambers. From Rechcigl et al. (1992b). Means across phosphogypsum rates with the same letter are not different at the 0.05 level of probability.

low-analysis fertilizers in commercial agriculture, at more frequent application, as they are now commonly used in recreational and residential lawns and gardens. Radionuclides, heavy metal impurities, and other pollutants at concentrations found in phosphogypsum do not appear to constitute environmental hazards to surficial ground water, soil, crop tissue, and the ambient atmosphere at rates normally used in agriculture. In conclusion, based on currently available information, phosphogypsum appears to be an environmentally safe source of S and Ca for crops as well as for other uses.

ACKNOWLEDGMENTS We wish to thank the Florida Institute of Phosphate Research, Bartow, Florida, for the research grant relating to “Influence of Phosphogypsum on Yield and Quality of Bahiagrass, and on the Environment in a Typical Florida Spodosol,” which made writing this review

I. S. ALCORDO AND J. E. RECHCIGL

102

Table IX Mean Ambient Atmospheric Radon and Mean Background Radiation Readings Measured 1 m above Treatment Plots“ Location Date (mo/d/yr)

Treatment plots

External control 1

External control 2

Ambient radon (lo-’ Bq liter’) 08/24/90-09/24/90 09/24/90- 10/30/90 10/30/90 1 1/27/90 11/27/90-02/13/91 02/20/91-03/2919 1 04/19/91-05/13/91 05/17/91 47/16/91

-

Mean

7.4 10.0 9.3 8.9 5.5 4.1

5.9 7.8 7.8 5.9 7.4 7.4 1.9

6.7 8.9 10.7 5.5 1.5 2.2

7.0

6.3

5.5

4.8*

4.1

y-Radiation @R hr-’) 08/24/90-09/24/90 09/24/90- 10/30/90 10/30/90- 11/27/90 I1/27/90-02/13/91 02/20/9 1 -0312919 I 04/19/91 -0511 3/91 05/17/9 1 -07/16/9 1 Mean

4.45ab 4.17b 4.38a 5.03a 4.50b 4.35b 4.32a

4.43a 4.25b 5.15a 5.80a 5.0b 4.30b 4.67a

5.20a 5.70a 5.70a 5.50a 6.20a 6.10a 4.85a

4.41~

4.84b

5.55a

From Rechcigl et al. (1992b). Means across location with the same letter are not different at the 0.05 level of probability.

possible. We also express our appreciation of Drs. A. K. Alva and W. D. Pitman for their critical review of the paper.

REFERENCES Adams, F. 1978. Liming anL fertilization of Ultisols and Oxisols. “Mineral Nutrition of Legumes in Tropical and Subtropical Soils’’(C. S. Andrew and E. J. Kamprath, eds.), pp. 377-394. Commonwealth Scientific and Industrial Research Organization, East Melbourne, Australia. Adams, F., and Hajek, B. F. 1978. Effects of solution sulfate, hydroxide, and potassium concentrations on the crystallization of alunite, basaluminite, and gibbsite from dilute aluminum solutions. Soil Sci. 126, 169- 173.

PHOSPHOGYPSUM IN AGRICULTURE

103

Adams, F., and Lund, Z. F. 1966. Effect of chemical activity of soil solution aluminum on cotton root penetration ofacid subsoils. SoilSci. 101, 193- 198. Adams, F., and Moore, B. L. 1983. Chemical factors affecting root growth in subsoil horizons of Coastal Plain soils. Soil Sci. SOC.Am. J. 47,99- 102. Adams, F., and Rawajfih, Z. 1977. Basaluminite and alunite: A possible cause of sulfate retention by acid soils. SoilSci. SOC.Am. J. 41,686-692. Adams, F., Pearson, R. W., and Doss, B. D. 1967. Relative effects of acid subsoils on cotton yields in field experiments and on cotton roots in growth-chamber experiments. Agron. J. 59,453-456. Agassi, M., Shainberg, I., and Morin, J. 1981. Effect of electrolyte concentration and soil sodicity on infitration rate and crust formation. Soil Sci. SOC. Am. J. 45, 848 - 85 1. Agassi, M., Morin, J., and Shainberg, I. 1982. Infiltration and runoff control in the semi-arid region of Israel. Geoderma 28, 345-355. Agassi, M., Shainberg, I., and Morin, J. 1985. Infiltration and runoff in wheat fields in the semi-arid region of Israel. Geoderma 36,263-276. Akin, D. E., and Hogan, J. P. 1983. Sulfur fertilization and rumen microbial degradation of cell walls in Digitaria pentzii (Stent). Crop. Sci. 23, 85 1-858. Alcordo, I. S. 1968. “The Flow Properties of Al-Montmorillonite, Fe-Montmorillonite, and Soil Clay Suspensions and Their Changes with Anaerobic Reduction,” Ph.D. thesis. Department of Agronomy, University of Illinois, Champaign-Urbana, Illinois. Alcordo, I. S., and Rechcigl, J. E. 1992. The use of solubility curves as a method to characterize phosphogypsum and its constituents. Commun. Soil Sci. Plant Anal. 23, 2595-2611. Alva, A. K., and Gascho, G. J. I99 1. Differential leaching of cations and sulfate in gypsum amended soils. Commun. Soil Sci. Plant. Anal. 22, I 195 - 1206. Alva, A. K., and Sumner, M. E. 1989. Alleviation of aluminum toxicity to soybeans by phosphogypsum or calcium sulfate in dilute nutrient solutions. Soil Sci. 147,278-285. Alva, A. K., Edwards, D. G., Asher, C. J., and Blarney, F. P. C. 1986. Relationships between root length and calculated activities of aluminum monomers in solution. Soil Sci. SOC. Am. J. 50,959-962. Aha, A. K., Gascho, G. J., and Guang, Y. 1989. Gypsum material effects on peanut and soil calcium. Commun. Soil Sci. Plant Anal. 20, 1727- 1744. Aha, A. K., Sumner, M. E., and Miller, W. P. 1990. Reactions of gypsum or phosphogypsum in highly weathered acid subsoils. Soil Sci. SOC.Am. J. 54,993-998. Alva, A. K., Sumner, M. E., and Miller, W. P. 1991. Chemical effects of repeated equilibriation of variable-charge soils with phosphogypsum solution. Soil Sci. SOC.Am. J. 55, 357-36 I . Amacher, M. C., and Miller, B. J. 1987. “Literature Review and Analysis to Assess the Agricultural Uses of Phosphogypsum and Associated Environmental Impacts.” Department of Agronomy, Louisiana Agricultural Experiment Station, Louisiana State University Agriculture Center, Baton Rouge, Louisiana. American Public Health Association. 1985. “Standard Methods for the Examination of Water and Wastewater,” 16th Ed. American Public Health Association- American Water Works Association- Water Pollution Control Federation, Washington, D.C. Appleyard, F. C. 1980. Gypsum industry in the United States: An overview-including potential for use of chemical gypsum. In “Proceedings of the International Symposium on Phosphogypsum,” Publ. 01-001-017, pp. 57-85. Florida Institute of Phosphate Research, Bartow, Florida.

104

I. S. ALCORDO AND J. E. RECHCIGL

Arman, A., and Seals, R. K. 1990. A preliminary assessment of utilization alternatives of phosphogypsum. In “Proceedings of the Third International Symposium on Phosphogypsum” (W. F. Chang, ed.), Publ. 01-060-083, pp. 562-583. Florida Institute of Phosphate Research, Bartow, Florida. Arora, H. S., and Coleman, N. T. 1979. The influence of electrolyte concentration on flocculation of clay suspensions. Soil Sci. 127, 134- 139. Aylmore, L. A. G., and Quirk, J. P. 1967. The micropore size distribution of clay mineral systems. J. Soil Sci. 18, 1- 17. Ayres, A. S. 1962. The utility of soil analysis in determining the need for applying calcium to sugarcane. Proc. Int. SOC.Sugar-Cane Technol. 11, 162 - 17 1. Baird, J. V., and Kamprath, E. J. 1980. Agricultural use of phosphogypsum on North Carolina crops. In “Proteedings of the International Symposium on Phosphogypsum” (D. P. Boms and P. W. Boody, eds.), Publ. 01-001-017, pp. 133- 150. Florida Institute of Phosphate Research, Bartow, Florida. Barley, K. P., and Hutton, J. T. 1956. Effects of a lowland rice crop and of gypsum on a saline alkali clay. Aust. J. Agric. Res. 7, 110- 126. Bartlett, R. J., and Riego, D. C. 1972. Toxicity of aluminum in relation to pH and phospho-

rus. SoilSci. 114, 194-201. Bayrakli, F. 1990. Ammonia volatilization losses from different fertilizers and effect of several urease inhibitors, CaCl, and phosphogypsum on losses from urea. Fert. Res. 23, 147150.

Beaton, J. D., Bixby, D. W.,Tisdale, W. L., and Platuou, 3. S. 1974. “Fertilizer Sulphur: Status and Potential in the U.S.,” Tech. Bull. 21. The Sulphur Institute, Washington, D.C. Beretka, J. 1980. Properties and utilization of by-product gypsum in Australia. In “Proceedings of the International Symposium on Phosphogypsum” (D. P. Boms and P. W. Boody, eds.), Publ. 0 1-001 -0 17, pp. 565 - 58 1. Florida Institute of Phosphate Research, Bartow, Florida. Beretka, J. 1990. The current state of utilisation of phosphogypsum in Australia. I n “Proceedings of the Third International Symposium on Phosphogypsum” (W. F. Chang, ed.), Publ. 01-060-083, pp. 394-401. Florida Institute of Phosphate Research, Bartow, Florida. Beverly, R. B., and Anderson, D. L. 1987. Effects of acid source on soil pH. Soil Sci. 143, 301 -303.

Bianco, S., Ruggiero, C., Vini, G. C., and Santos, P. R. R. S. 1990. Effects of phosphogypsum and potassium chloride on the nutritional status, production and organoleptical quality of pineapple fruits. I n “Proceedings of the Third International Symposium on Phosphogypsum’’ (W. F. Chang, ed.),Publ. 01-060-083, pp. 348-361. Florida Institute of Phos phate Research, Bartow, Florida. Black, A. S., and Cameron, K. C. 1984. Effect of leaching on soil properties and lucerne growth following lime and gypsum amendments to a soil with an acid subsoil. N.Z. J. Agric. Res. 27, 195 - 200. Blackmore, A. V., and Miller, R. D. 1961. Tactoid size and osmotic swelling in calcium montmorillonite. SoilSci. Soc. Am. Proc. 25, 169- 173. Blarney, F. P. C., Edwards, D. G., and Asher, C. J. 1983. Effect of aluminum, OH, Al, and P: Al molar ratios, and ionic strength on soybean root elongation in solution culture. Soil

Sci.136, 197-207.

Bornemisza, E., and Llanos, R. 1967. Sulfate movement, absorption, and desorption in three Costa Rica soils. Soil Sci. SOC. Am. Proc. 31,356-360. Boms, D. P., and Boody, P. W. (eds.). 1980. “Proceedings of the International Symposium

PHOSPHOGYPSUM IN AGRICULTURE

10s

on Phosphogypsum,” Publ. 01-001-017. Florida Institute of Phosphate Research, Bartow, Ronda. Bowen, H. D. 198 I. Alleviating mechanical impedance. In “Modifying the Root Environment to Reduce Crop Stress” (G. F. Arkin and H. M. Taylor, eds.), ASAE Monogr. 4, pp. 2 1 -57. American Society of Agricultural Engineers, St. Joseph, Michigan. Brady, N. C., and Colwell, W. E. 1945. Yield and quality of large-seeded type peanuts as affected by potassium and certain combinations of potassium, magnesium, and calcium. J. Am. SOC.Agron. 37,429-442. Bray, A. C., and Hemsley, J. A. 1969. Sulphur metabolism of sheep. IV. The effect of a vaned dietary sulphur content on some body fluid sulphate levels and on the utilization of urea-supplemented roughage by sheep. Aust. J. Agric. Res. 20,759-773. Breithaupt, J. A. 1989. “Effect of By-product Gypsum on Yield and Nutrient Content of Sugarcane and Soil Properties,” M.S. thesis. Lousiana State University, Baton Rouge, Louisiana. Bridge, B. J., and Kleinig, C. R. 1968. The effect of gypsum on the water storage in a sandy loam soil under an imgated perennial pasture. Trans. Int. Congr. Soil Sci.. 9th 1, 3 13-322.

Brown, B. A., and Munsell, R. I. 1938. Soil acidity at various depths as influenced by time since application, placement, and amount of limestone. Soil Sci. SOC.Am. Proc. 3, 2 17-22 1.

Buselli, E. M. 1988. “Gypsum Effects on an Alligator Clay Soil and Sugarcane in Louisiana,” M.S. thesis. Louisiana State University, Baton Rouge, Louisiana. Buyeye, S. M., Fey, M. V., and Mot& C. J. B. 1985. Maize growth and acid soil responses to treatment with phosphogypsum and lime. Proc. S. Afi. SOC.Crop Sci. Congr. pp. 366384.

Caldwell, A. G., Hutchinson, R. L., Kennedy, C. W., and Jones, J. E. 1990. “Effect of Rates of Lime and By-product Gypsum on Movement of Calcium and Sulfur into an Acid Soil,”Agron. Abstr., p. 264. American Society of Agronomy, Madison, Wisconsin. Cameron, R. S., Ritchie, G. S. P., and Robson, A. D. 1986. Relative toxicities of inorganic aluminum complexes to barley. SoilSci. SOC.Am. J. 50, 1231- 1236. Campbell, J. R., and Lasley, J. F. 1969. “The Science of Animals That Serve Mankind,” 1st Ed. McGraw-Hill, New York. Carlisle, V. W., and Fiskell, J. G. A. 1962. Relationship of acidity to other soil properties in certain Weston, Leon and Blanton profiles on three. marine terraces. Proc. -Soil Crop Sci.SOC.Fla. 22,92- 106. Chang, W. F. (ed.). 1990. “Proceedings of the Third International Symposium on Phosphogypsum,” h b l . 01-060-083. Florida Institute of Phosphate Research, Bartow, Florida. Chang, M. L., and Thomas, G. W. 1963. A suggested mechanism for sulfate adsomtion by soils. SoilSci. Soc. Am. Proc. 27,281-283. Chao, T. T. 1964. Anion effects on sulfate adsorption by soils. Soil Sci. Sac. Am. Proc. 28, 58 1-583.

Chao, T. T., Harward, M. E., and Fang, S. C. 1963. Cationic effects on sulfate adsorption by soils. Soil Sci. SOC.Am. Proc. 27, 35 - 38. Chao, T. T., Harward, M. E., and Fang, S. C. 1965. Exchange reactions between hydroxyl and sulfate ions in soils. Soil Sci. 99, 104- 108. Chiang, S. C., Radcliffe, D. E., Miller, W. P., and Newman, K. D. 1987. Hydraulic conductivity of three Southeastern soils as affected by sodium, electrolyte concentration, and pH. Soil Sci. SOC.Am. J. 51, 1293- 1299. Chien, S. H., Hellums, D. T., and Henao, J. 1987. Greenhouse evaluation of elemental sulfur and gypsum for flooded rice. Soil Sci. SOC.Am. J. 51, 120- 123.

106

I. S. ALCORDO AND J. E. RECHCIGL

Church, D. C. 1979. “Digestive Physiology and Nutrition of Ruminants,” Vol. 1 , 1st Ed. Metropolitan, Portland, Oregon. Coleman, N. T., Thomas, G. W., LeRoux, F. H., and Bredell, G. 1964. Salt exchangeable and titratable acidity in bentonit-sesquioxide mixtures. Soil Sci. SOC.Am. Proc. 28, 35 - 37. Collings, R. K. 1980. Phosphogypsum in Canada. In “Proceedings of the Symposium on Phosphogypsum” (D. P. Borris and P. W. Boody, eds.),Publ. 01-001-017, pp. 583-596. Florida Institute of Phosphate Research, Bartow, Florida. Colwell, W. E., and Brady, N. C. 1945. The effect of calcium on yield and quality of large-seeded type peanuts. J. Am. SOC.Agron. 37,4 13 -428. Couto, W., Lathwell, D. J., and Bouldin, D. R. 1979. Sulfate sorption by two Oxisols and an Alfisol of the tropics. Soil Sci. 127, 108- 1 16. Cox, F. R., Sullivan, G. A., and Martin, C. K. 1976. Effect of calcium and irrigation treatments on peanut yield, grade and seed quality. Peanut Sci. 3,81-85. Dahlgren, R. A., Driscoll, C. T., and McAvoy, D. C. 1989. Aluminum precipitation and dissolution rates in Spodosol Bs horizons in the Northeastern USA. Soil Sci. SOC.Am. J. 53, 1045- 1052. Daughtry, J. A., and Cox, F. R. 1974. Effect of calcium source, rate, and time of application on soil calcium level and yield of peanuts (Archis hypogaea L.). Peanut Sci. 1,68-73. Davidson, J. L., and Quirk, J. P. 1961. The influence of dissolved gypsum on pasture establishment on imgated sodic clays. Aust. J. Agric. Res. 12, 100- 110. de Carvalho, M. M., Edwards, D. G., Asher, C. J., and Andrew, C. S. 1982. Effects of aluminum on nodulation of two Stylosanthes species grown in nutrient solution. Plant Soil64, 141-152. Deshpande, T. L., Greenland, D. J., and Quirk, J. P. 1968. Changes in soil properties associated with the removal of iron and aluminum oxides. J. Soil Sci. 19, 108 - 122. de Villiers, J. M., and Jackson, M. L. 1967. Cation exchange capacity variations with pH in soil clays. Soil Sci. Soc. Am. Proc. 31,473-476. du Plessis, H. M., and Shainberg, I. 1985. Effect ofexchangeable sodium and phosphogypsum on the hydraulic properties of several South African soils. S. Afi. J. Plant Soil 2, 179-185. Dutt, G. R., Terkeltoub, R. W., and Rauschkolb, R. S. 1972. Prediction of gypsum and leaching requirements for sodium-affected soils. Soil Sci. 114,93- 103. Eichhorn, M. M., Jr., Feagley, S. E., and Nelson, B. D. 1990. Effects ofsulfur fertilization on Coastal bermudagrass grown on Coastal Plain soil. Proc. Am. Forage G r a d Con$ Elamin, E. A., and Suarez, D. L. 1990. “Effects of Aluminum Polymers on the Physical and Chemical Properties of a Sodic Soil,” Agron. Abstr., p. 228. American Society of Agronomy, Madison, Wisconsin. El-Swaify, S. A., and Emerson, W. W. 1975. Changes in the physical properties of soil clays due to precipitated aluminum and iron hydroxides: I. Swelling and aggregate stability after drying. Soil Sci. SOC.Am. Proc. 39, 1056- 1063. Environmental Protection Agency (EPA). 1991. “Diffuse NORM Waste: Waste Characterization and Preliminary Risk Assessment” (draft). EPA-ANR-460. U.S. EPA, Office of Radiation Programs, Washington, D.C. Farina, M. P. W., and Channon, P. 1988. Acid subsoil amelioration 11: Gypsum effects on growth and subsoil chemical properties. Soil Sci. SOC.Am. J. 52, 175 - 180. Federal Register. 1978.43(243), 58957-58959 (Dec. 18, 1978). Federal Register. 1980.45(98), 33086-33087,33118, and 33122-33131 (May 19, 1980). Federal Register. 1989.54(240), 51674-51676 (Dec. 15, 1989). Part 11. EPA 40 CFR Part 61. National emission standards for hazardous air pollutants; radionuclides; final rule and notice of reconsideration.

PHOSPHOGYPSUM IN AGRICULTURE

107

Federal Register. 1991. 56(138), 33050-33051 (July 18, 1991). Part 11. EPA 40 CFR Parts 141 and 142. National primary drinking water regulations; radionuclides; proposed rule. Federal Register. 1992. 57(107), 23305-23320. (June 3, 1992). EPA 40 CFR Part 61. National emission standards for hazardous air pollutants: national emission standards for radon emission from phosphogypsum stacks. Florida Institute of Phosphate Research. Undated. “Radiation and Your Environment: A Guide to Low-Level Radiation for Citizens of Florida.” Florida Institute of Phosphate Research, Bartow, Florida. Frankel, H., Goertzen, J. O., and Rhoades, J. D. 1978. Effects of clay type and content, exchangeable sodium percentage and electrolyte concentration on clay dispersion and soil hydraulic conductivity. Soil Sci. SOC.Am. J. 42,32 - 39. Fried, M., and Peach, M. 1946. The comparative effects of lime and gypsum upon plants grown on acid soils. J. Am. SOC.Agron. 38,6 14-623. Friesen, D. K., and Chien, S. H. 1986. Sulfur fertilizers and their application: General review I n “Proceedings of the International Symposium on and S source evaluation at I=. Sulphur in Agricultural Soils” (S. Portch and S. G. Hussain, eds.),pp. 407-431. The Bangladesh Agricultural Research Council and The Sulphur Institute, Dhaka, Bangladesh, India. Fruhauf, F., Burk, W., and Hofmann, U. 1962. Der Einfluss der Austauschfahigen kationen auf die Rheologischen eigenschafien und die Korngrossen-Verteilung des Kaolinits. Kolloid Z. Z. Polym. 180, 150- 160. Gal, M., Arcan, L., Shainberg, I., and Keren, R. 1984. Effect of exchangeable sodium and phosphogypsum on crust structure -Scanning electron microscope observations. Soil Sci. SOC.Am. J. 48,872-878. Garren, K. H. 1964. Landplaster and soil rot of peanut pods in Virginia. Pfant Dis. Rep. 48, 349 - 352.

Gascho, G. J., and Alva, A. K. 1990. Beneficial effects of gypsum for peanut. I n “Proceedings of the Third International Symposium on Phosphogypsum” (W. F. Chang, ed.), Publ. 0 1-060-083, pp. 376-393. Florida Institute of Phosphate Research, Bartow, Florida. Gaston, L. A., Mansell, R. S., and Rhue, R. D. 1986. Sulfate mobility in acid soils and implications with respect to cation leaching: A review. Proc.-Soil Crop Sci. SOC.Flu. 45,67-72.

Gebhardt, H., and Coleman, N. T. 1974. Anion adsorption by allophanic tropical soils: 11. Sulfate adsorption. Soil Sci. SOC.Am. Proc. 38,259 -262. Gillman, G. P. 1991. The chemical properties of acid soils with emphasis on soils of the humid tropics. I n “Plant-Soil Interactions in Low p H (R. J. Wright, V. C. Baligar, and R. P. Murrmann, eds.), pp. 3- 14. Kluwer Academic Publishers, Dordrecht, The Netherlands. Glas, T. K., Klute, A., and McWhorter, D. B. 1979. Dissolution and transport of gypsum in soils: I. Theory. Soil Sci. SOC.Am. Proc. 43,265 -268. Goers, W. E. 1980. Nissan hemi phosphogypsum. I n “Proceedings of the International Symposium on Phosphogypsum” (D. P. Boms and P. W. Boody, eds.),Publ. 01-001017, pp. 35 -44. Florida Institute of Phosphate Research, Bartow, Florida. Golden, L. E. 1982. “Results from By-product Gypsum Outfield Tests with Sugarcane, 1975- 1982,” Report of Projects. Department of Agronomy, Louisiana Agricultural Experiment Station, Louisiana State University Agriculture Center, Baton Rouge, Louisiana. Gosnell, J. M., and Long, A. C. 1969. A sulphur deficiency in sugarcane. S. A f . Sugarcane Technol.Assoc. 43,26-29. Graveland, D. N., and Toogood, J. A. 1963. Gypsum as an ameliorating agent for solonetzic soils in Alberta. Can. J. Soil Sci. 43, 1-6.

I. S. ALCORDO AND J. E. RECHCIGL

108

Greene, R. S. B., Posner, A. M., and Quirk, J. P. 1978. A study of the coagulation of montmorillonite and illite suspensions by calcium chloride using the electron microscope. In “Modification of Soil Structure” (W. W. Emerson, R. D. Bond, and A. R. Dexter, eds.),pp. 35-40. Wiley, New York. Grunes, D. L., and Allaway, W. H. 1985. Nutritional quality of plants in relation to fertilizer use. In “Fertilizer Technology and Use” (0.P. Engelstad, ed.), 3rd Ed., pp. 589-620. Soil Science Society of America, Madison, Wisconsin. Guidry, J. J., Roessler, C. E., and McClave, J. T. 1990. “Radioactivity in Foods Grown on Mined Phosphate Lands,” Final Rep., Publ. 05-028-088. Florida Institute of Phosphate Research, Bartow, Florida. Gupta, A. C. 1969. S X Mo interactions in plant nutrition. Sulphur Inst. J. 5,4-6. Hallock, D. L., and Allison, A. H. 1980. Effect of three Ca sources applied on peanuts. I. Productivity and seed quality. Peanut Sci. 7 , 19-25. Hallock, D. L., and Garren, K. H. 1968. Pod breakdown, yield, and grade of Virginia type peanuts as affected by Ca,Mg, and K sulfates. Agron. J. 60,253-251. Hammel, J. E., Sumner, M. E., and Shahandeh, H. 1985. Effect of physical and chemical profile modification on soybean and corn production. Soil Sci. Soc. Am. J. 49, 1508151 I . Harmsen, K., and van Breemen, N. 1975. A model for the simultaneous production and diffusion of ferrous iron in submerged soils. Soil Sci. SOC.Am. Proc. 39, 1063- 1068. Harre, E. A., and White, W. C. 1985. Fertilizer market profile. In “Fertilizer Technology and Use” (0. P. Engelstad, ed.), 3rd Ed., pp. 1-24. Soil Science Society of America, Madison, Wisconsin. Hartley, J. N., and Freeman, H.D. 1985. “Radon Flux Measurements on Gardinier and Royster Phosphogypsum Piles near Tampa and Mulberry, Florida,” prepared for the U.S. Environmental Protection Agency, Eastern Environmental Radiation Facility, Montgomery, Alabama. Pacific Northwest Laboratory, Richland, Washington. Hendricks, S. B., Nelson, R. A., and Alexander, L. T. 1940. Hydration mechanism of the clay mineral montmorillonite saturated with various cations. J. Am. Chem. SOC.62, 14571464.

Hendrickson, B. H., Barnett, A. P., Carreker, J. R., and Adams, W. E. 1963. Runoff and erosion control studies on Cecil soil in the Southern Piedmont. US.,Dep. Agric., Tech. Bull. 1281. Hilgard, E. W. 1907. “Soils-Their Formation, Properties, Composition and Relation to Climate and Plant Growth in the Humid and Arid Regions.” Macmillan, New

York. Hintz, H. F., and Hogue, D. E. 1964. Effect of selenium, sulfur and sulfur amino acids on nutritional muscular dystrophy in the lamb. J. Nutr. 82,495-498. Hopper, R. D., Levy, R. A., Rankin, R. C., and Boyd, M. A. 1990. “National Ambient Radon Study.” U.S. Environmental Protection Agency, Washington, D.C. Horst, W. J., Wagner, A., and Marschner, H. 1983. Effect of aluminum on root growth, celldivision rate and mineral element contents in roots of Vigna unguiculutu genotypes. Z . Pflanzenphysiol. 109,95- 103. Howell, M. 1987. Gypsum use in the wheat belt. J. Agric., West. Aust. 28,40-43. Hue, N. V., Adams, F., and Evans, C. E. 1985. Sulfate retention by an acid BE horizon of an Ultisol. Soil Sci. SOC.Am. J. 49, 1 196- 1200. Hue, N. V., Fox, R. L., and McCall, W. W. 1987. Aluminum, Ca, and Mn concentrations in macadamia seedlings as affected by soil acidity and liming. Commun. Soil Sci. Plant Anal. 18, 1253-1267. Hume, I. D., and Bird, P. R. 1970. Synthesis of microbial protein in the rumen: IV. The .influence of the level and form ofdietary sulphur. Aust. J. Agric. Res. 21, 315-322.

PHOSPHOGYPSUM IN AGRICULTURE

109

Hunter, A. H. 1989. “Use of Phosphogypsum Fortified with Other Selected Essential Elements as a Soil Amendment on Low Cation Exchange Soils,” Publ. 0 1-034-08 I. Florida Institute of Phosphate Research, Bartow, Florida. Institute of Food and Agricultural Sciences. 1980. “Position Statement and Supporting Reports Regarding the Proposed Farmland Industries Phosphate Fertilizer Complex near the Ona Agricultural Research Center.” Institute of Food and Agricultural Science, University of Florida, Gainesville, Florida. International Fertilizer Development Center. 1985. “Fertilizer Research Program for Africa.” International Fertilizer Development Center, Muscle Shoals, Alabama. Islam, M. S., Hoassain, K. M., Sarkar, N. I., Altamash, S., and Sarkar,J. U. 1986. Experimental methodology and responses of principal crops to sulphur fertilization in Bangladesh. In “Proceedings of the International Symposium on Sulphur in Agricultural Soils” (S. Portch and Sk. G. Hussain, eds.),pp. 184-208. The Bangladesh Agricultural Research Council and The Sulphur Institute, Dhaka, Bangladesh, India. Ismunadji, M. 1986. Experience, research and farmers’ use of sulphur in food crop production in Indonesia. In “Proceedings of the International Symposium on Sulphur in Agricultural Soils” (S. Portch and Sk. G. Hussain, eds.), pp. I 17- 136. The Bangladesh Agricultural Research Council and The Sulphur Institute, Dhaka, Bangladesh, India. Ismunadji, M., and Zulkarnaini, I. 1978. Sulphur deficiency of lowland rice in Indonesia. Sulphur Agric. 2, 17- 19. Jackson, M. L. 1963. Aluminum bonding in soils: A unifylng principle in soil science. Soil Sci. SOC. Am. Proc. 27, 1 - 10. Jones, M. B. 1967. Sulphur fertilization of Brazil‘s Campos Cerrado soils. Sulphur Inst. J. 3(2), 2-4. Jordan, H. V. 1964. Sulfur as a plant nutrient in the southern United States. US.,Dep. Agric., Tech. Bull. 1297. Kamprath, E. J. 1970. Exchangeable aluminum as a criterion for liming leached mineral soils. Soil Sci. SOC.Am. Proc. 34,252-254. Kamprath, E. J. 1971. Potential detrimental effects from liming highly weathered soils to Fla. 31,200-203. neutrality. Proc.-Soil Crop Sci. SOC. Kamprath, E. J., Nelson, W. L., and Fitts, J. W. 1956. The effect ofpH, sulfate and phosphate concentrations on the adsorption of sulfate by soils. Soil Sci. SOC.Am. Proc. 20, 463466. Kawaguchi, K., and Matsuo, Y. 1956. Movements of active oxide in dry paddy soil profiles. Trans. Int. Congr. Soil Sci.. 6th R-9,533-537. Kazman, Z., Shainberg, I., and Gal, M. 1983. Effect of low levels of exchangeable sodium and applied phosphogypsum on the infiltration rate of various soils. Soil Sci. 135, 184- 192. Keerati-Kasikorn, P. 1986. Research and farmer use of sulphur in Thailand. In “Proceedings of the International Symposium on Sulphur in Agricultural Soils” (S. Portch and Sk. G. Hussain, eds.),pp. 163- 176. The Bangladesh Agricultural Research Council and The Sulphur Institute, Dhaka, Bangladesh, India. Keerthisinghe, G., McLaughlin, M. J., and Freney, J. R. 1991. Use of gypsum, phosphogypsum and fluoride to ameliorate subsurface acidity in a pasture soil. In “Plant-Soil Interactions in Low p H ’ (R. J. Wright, V. C. Baligar, and R. P. Murrmann, eds.), pp. 509 - 5 17. Kluwer Academic Publishers, Dordrecht, The Netherlands. Keisling, T. C., Rao, P. S. C., and Jessup, R. E. 1978. Pertinent criteria for describing the dissolution of gypsum beds in flowing water. Soil Sci. SOC.Am. J. 42,234-246. Kelley, W. P., and Brown, S. M. 1934. Principles governing the reclamation of alkali soils. Hilgardia 8, 149- 177. Kemper, W. D., Olsen, J., and dehlooy, C. J. 1975. Dissolution rates of gypsum in flowing Am. Proc. 39,458-463. water. Soil Sci. SOC.

110

I. S. ALCORDO AND J. E. RECHCIGL

Keng, J. K. C., and Uehara, G. 1974. Chemistry, mineralogy and taxonomy of Oxisols and Ultisols. Proc.-Soil Crop Sci. SOC.Fla. 33, 1 19 - 126. Keren, R., and Shainberg, I. 1981. Effect of dissolution rate on the efficiency of industrial and mined gypsum in improving infiltration of a sodic soil. Soil Sci. SOC.Am. J. 45, 103- 107.

Keren, R., and Singer, M. J. 1989. Effect of low electrolyte concentration on hydraulic conductivity of clay- sand- hydroxy polymer systems. Soil Sci. SOC.Am. J. 53,349- 353. Keren, R., and Singer, M. J. 1990. Effect of pH on permeability of clay-sand mixture containing hydroxy polymers. Soil Sci. Soc. Am. J. 54, 13 10- 13 15. Keren, R., and Singer, M. J. 199 I. Hydroxy-aluminum’seffect on permeability of clay- sand mixtures. Soil Sci. Soc. Am. J. 55,6 1-65. Khalil, N. F., Alnuaimi, N. M., and Mustafa, M. H. 1990. Agricultural uses of phosphogypsum on calcareous soils. I n “Proceedings of the Third International Symposium on Phosphogypsum” (W. F. Chang, ed.), Publ. 01-060-083, pp. 333-447. Florida Institute of Phosphate Research, Bartow, Ronda. Kilmer, V. J. 1982. “Handbook of Soils and Climate in Agriculture.” CRC Press, Boca Raton, Florida. Kinraide, T. B., and Parker, D. R. 1987a. Cation amelioration of aluminum toxicity in wheat. Plant Physiol. 83, 546-551. Kinraide, T. B., and Parker, D. R. 1987b. Nonphytotoxicity of the aluminum sulfate ion, AISO.,+. Physiol. Plant. 71,207-212. Klute, A., and Dirksen, C. 1986. Hydraulic conductivity and difisivity: Laboratory methods. I n “Methods of Soil Analysis” (A. Klute, ed.),Part I, 2nd Ed. (Agronomy 9), pp. 687-734. Soil Science Society of America, Madison, Wisconsin. Korentajer, L., Byrnes, B. H., and Hellums, D. T. 1983. Effect of liming and leaching on the sulfur-supplying capacity of soils. Soil Sci. Soc. Am. J. 47, 525 - 530. Korentajer, L., Byrnes, B. H., and HeUums, D. T. 1984. Leaching losses and plant recovery from various sulfur fertilizers. Soil Sci. SOC.Am. J. 48,67 1 -676. Kotrappa, P., Dempsey, J. C., Hickey, J. R., and Stief, L. R. 1988. An electret passive environmental zzzRnmonitor based on ionization measurement. Health Phys. 54,4756.

Kouloheris, A. P. 1980. Chemical nature of phosphogypsum as produced by various wet process phosphoric acid processes. I n “Proceedings of the International Symposium on Phosphogypsum” (D. P. Boms and P. W. Boody, eds.), Publ. 01-001-017, pp. 7-33. Florida Institute of Phosphate Research, Bartow, Florida. Lancaster, D. L., Jones, M. B., Oh, J. H., and Ruckman, J. E. 1971. Effect of sulfur fertilization of forage species on yield, chemical composition, and in vitro mmen microbial activity of sheep. Agron. J. 63,621 -623. Ligon, W. S., and Pierre, W. H. 1932. Soluble aluminum studies: 11. Minimum concentrations of aluminum found to be toxic to corn, sorghum, and barley in culture solutions. Soil Sci. 34,307 - 3 18. Lin, K. T., Lai, C. I., Ghafoori, N., and Chang, W. F. 1990. High-strength concrete utilizing industrial by-product. I n “Proceedings of the Third International Symposium on Phosphogypsum” (W. F. Chang, ed.), Publ. 01-060-083, pp. 456-472. Florida Institute of Phosphate Research, Bartow, Florida. Lindekin, C. L. 1980. Radiological considerations of phosphogypsum utilization in agriculture. I n “Proceedings of the International Symposium on Phosphogypsum” (D. P. Boms and P. W. Boody, eds.),Publ. 01-001-017, pp. 401 -414. Florida Institute of Phosphate Research, Bartow, Florida.

PHOSPHOGYPSUM IN AGRICULTURE

111

Lindekin, C. L., and Coles, D. G. 1977.The radium-226 content of agricultural gypsum. US. Nucl. Regul. Comrn. [Rep.] NUREG/CP NUREG/CP.0001,369-375. Loveday, J. 1974.Recognition of gypsum-responsive soils. Aust. J. Soil Rex 12,87-96. Loveday, J., and Scotter, D. R. 1966.Emergence response of subterranean clover to dissolved gypsum in relation to soil properties and evaporative conditions. Aust. J. Soil Res. 4,

55-68.

Lutz, J. F. 1936.The relation of free iron oxide in the soil to aggregation. Soil Sci. SOC.Am. Proc. 1,43-45. Magistad, 0. C. 1925. The aluminum content of the soil solution and its relation to soil reaction and plant growth. Soil Sci. 20, 181 -225. Marshall, W. L.,and Slusher, R. 1966. Thermodynamics of calcium sulfate dihydrate in aqueous sodium chloride solutions, 0- 110”.J. Phys. Chem. 70,4015-4027. Mathews, B. W., and Joost, R. E. 1989.Reduction of subsoil aluminum toxicity to alfalfa by leaching surface-applied amendments. Proc. Am. Forage G r a d Conf pp. 54-58. Mathews, B. W., and Joost, R. E. 1990. The effects of leaching surface-applied amendments on subsoil aluminum and alfalfa growth in a Louisiana Ultisol. Commun. Soil Sci. Plant Anal. 21,567-581. Matsumoto, H.,and Morimura, S. 1980. Repressed template activity of chromatin of pea roots treated with aluminum. PIant Cell Physiol. 21,95 1-959. Matsumoto, H., Hirasawa, E., Torikai, H., and Takahashi, E. 1976.Localization of absorbed aluminum in pea root and its binding to nucleic acids. Plant Cell Physiol. 17, 127- 137. Matuszek, J. M. 1990.Standardization of radon measurements: 11. Accuracy and proficiency testing. J. Res. Natl. Inst. Technol. 95, I77- 181. May, A., and Sweeney, J. W. 1980. Assessment of environmental impacts associated with phosphogypsum in Florida. I n “Proceedings of the International Symposium on Phosphogypsum” (D. P. Boms and P. W. Boody, eds.), Publ. 01-001-017, pp. 413-442. Florida Institute of Phosphate Research, Bartow, Florida. May, A., and Sweeney, J. W. 1982. Assessment of environmental impacts associated with phosphogypsum in Florida. Rep. Invest. -US..Bur. Mines 8639. May, A., and Sweeney, J. W. 1983. Evaluation of radium and toxic element leaching characteristics of Florida phosphogypsum stockpiles. Rep. Invest.-US.,Bur. Mines

8776.

Mays, D.A., and Mortvedt, J. J. 1986.Crop response to soil applications of phosphogypsum. J. Environ. Qual. 15, 78- 81. Mazid, S. A. 1986.The response of major crops to sulphur in agricultural soils of Bangladesh -An overview. In “Proceedings of the International Symposium on Sulphur in Agricultural Soils” (S. Portch and Sk. G. Hussain, eds.), pp. 18-28. The Bangladesh Agricultural Research Council and The Sulphur Institute, Dhaka, Bangladesh, India. McIntyre, D. S. 1958.Permeability measurements of soil crusts formed by raindrop impact. SoilSci. 5, 185-189. McLachlan, K.D. (ed.) 1975.“Sulfur in Australasian Agriculture.” Sydney University Press, Sydney, Australia. McNeal, B. L. 1968.Prediction of the effect of mixed-salt solutions on soil hydraulic conductivity. Soil Sci. SOC.Am. Proc. 32, 190- 193. McNeal, B. L.,and Coleman, N. T. 1966.Effect of solution composition on the swelling of extracted soil clays. Soil Sci.Soc. Am. Proc. 30, 313- 317. Mehlich, A. 1953.Determination of P, Ca, Mg, K, Na, and NH,. N.C. Soil Test Memo. Melamed, D., Hanks, R. J., and Willardson, L. S. 1977. Model of salt flow in soil with a source-sink term. SoilSci. SOC.Am. Proc. 41.29-33.

112

I. S. ALCORDO AND J. E. RECHCIGL

Metzger, W. H. 1934. Distribution of fertilizer residues in the soil after fourteen years of a fertilizer experiment with alfalfa. J. Am. SOC.Agron. 26,620-625. Miller, W. P. 1987. Infiltration and soil loss of three gypsum-amended Ultisols under simulatedrainfall. SoilSci. SOC.Am. J. 51, 1314-1320. Miller, W. P. 1989. “Use of Gypsum to Improve Physical Properties and Water Relations in Southeastern Soils,” Publ. 0 1-020-082. Florida Institute of Phosphate Research, Bartow, Florida. Mishra, U. N. 1980. Use of phosphogypsum in reclamation of sodic soils in India. In “Proceedings of the International Symposium on Phosphogypsum” (D. P. Boms and P. W. Boody, eds.), Publ. 01-001-017, pp. 223-241. Florida Institute of Phosphate Research, Bartow, Florida. Mislevy, P., Blue, W. G., and Roessler, C. E. 1989. Productivity of clay tailings from phosphate mining: I. Biomass Crops. J. Environ. Qual. 18, 95- 100. Mitchell, C. C., Jr., and Blue, W. G. 1989. Bahiagras response to sulfur on an Aeric Haplaquod. Agron. J. 81,53-57. Miyamoto, M. 1980. Phosphogypsum utilization in Japan. In “Proceedings of the International Symposium on Phosphogypsum” (D. P. Boms and P. W. Boody, eds.), Publ. 01-001-017, pp. 535-564. Florida Institute of Phosphate Research, Bartow, Florida. Moir, R. J., Somers, M., and Bray, A. C. 1967- 1968. Utilization of dietary sulphur and nitrogen by ruminants. Sulphur Inst. J. 3(4), 15 - 18. Molinari, J., and Snodgrass, W. J. 1990. The chemistry and radiochemistry of radium and other elements of the uranium and thorium natural decay series. Tech. Rep. Ser.Z.A.E.A. 310 (Vol. I), 11- 128. Momuat, E. O., Mamaril, C. P., Umar, A. P., and Momuat, C. J. S. 1983. Sulphur deficiency and how to cure it in South Sulawesi. In “The Role of Research Results of Rice and Upland Food Crops in Agricultural Development,” pp. 295-307. Central Research Institute for Food Crops, Bogor, Indonesia. Moore, L. D., and Wills, W. H. 1974. The influence of calcium on the susceptibility of peanut pods to Pythium myriotylum and Rhizoctonia solani. Peanut Sci. 1, 18-20. Moms, R. J. 1986. The importance of sulphur in agriculture- An overview. In “Proceedings of the International Symposium on Sulphur in Agricultural Soils” (S. Portch and Sk. G. Hussain, eds.), pp. 1 - 17. The Bangladesh Agricultural Research Council and The Sulphur Institute, Dhaka, Bangladesh, India. Mullins, G. L., and Mitchell, C. C., Jr. 1990. “Use of Phosphogypsum to Increase Yield and Quality of Annual Forages,” Publ. 01-048-084. Florida Institute of Phosphate Research, Bartow, Florida. Mums, D. N., and Franco, A. A. 1982. Soil constraints to legume production, In “Biological Nitrogen Fixation for Tropical Agriculture” (P. H. Graham and S. C. Hanis, eds.), pp. 135 - 152. Centro Intemacioral de Agricultura Tropical, Cali, Colombia. Murphy, M. D., Brogan, J. C., and Noonan, D. G. 1983. Sulfur fertilization of pasture improves cattle performance. Sulphur Agric. I, 2-6. Myhre, D. L., Martin, H. W., and Nemec, S. 1990. Yield, 226Raconcentration, and juice quality of oranges in groves treated with phosphogypsum and mined gypsum. In “Proceedings of the Third International Symposium on Phosphogypsum” (suppl.), (W. F. Chang, ed.), Publ. 01-060-083, pp. 1 1 -40. Florida Institute of Phosphate Research, Bartow, Florida. National Council on Radiation Protection and Measurements (NCRPM). 1984. “Exposures from the Uranium Series with Emphasis on Radon and Its Daughters,” Rep. 77. NCRPM, Bethesda, Maryland.

PHOSPHOGYPSUM IN AGRICULTURE

113

Nifong, G. D. 1988. “Environmental Aspects of Phosphogypsum” (mimeogr.). Florida Institute of Phosphate Research, Bartow, Florida. Noble, A. D., Fey, M. V., and Sumner, M. E. 1988a. Calcium-aluminum balance and the growth of soybean roots in nutrient solutions. Soil Sci. SOC.Am. J. 52, 165 1- 1656. Noble, A. D.,Sumner, M. E., and Alva, A. K. 1988b. The pH dependency of aluminum phytotoxicity alleviation by calcium sulfate. Soil Sci. Soc. Am. J. 52, 1398- 1402. Norrish, K. 1954.The swelling of montmorillonjte. Discuss. Faraday SOC.18, 120- 134. Nomsh, K.,and Quirk, J. P. 1954.Crystalline swelling of montmorillonite. Nature (London) 173,255-256. Novikov, A. A., Klassen, P. V., and Evenchik, S. D. 1990.The status and trends of phosphogypsum utilization in the USSR. In “Proceedings of the Third International Symposium on Phosphogypsum” (W. F. Chang, ed.), Publ. 01-060-083,pp. 594-601. Florida Institute of Phosphate Research, Bartow, Florida. Nye, P.,Craig, D., Coleman, N. T., and Ragland. J. L. 1961. Ion exchange equilibrium involving aluminum. Soil Sci. SOC.Am. Proc. 25, 14-17. Oates, K.M., and Caldwell, A. G. 1985.Use of by-product gypsum to alleviate soil acidity. Soil Sci. SOC.Am. J. 49,9 15 - 9 18. Oates, K.M., and Kamprath, E. J. 1985.Sulfur fertilization of winter wheat grown on deep sandy soils. SoilSci. SOC.Am. J. 49,925-921. O’Brien, L. O.,and Sumner, M. E. 1988. Effect of phosphogypsum on leachate and soil chemical composition. Commun. Soil. Sci. Plant Anal. 19, 1319- 1329. Odom, J. W. I991. “Alfalfa Response to Gypsum, Boron and Subsoiling on an Acid Ultisol,” Agron. Abstr., p. 296. American Society of Agronomy, Madison, Wisconsin. Orvedal, A. C., and Ackerson, K. T. 1972. “Agricultural Soil Resources of the World” (mimeogr.). U.S. Department of Agriculture, Soil Conservation Service, Washington, D.C. Oster, J. D. 1980.Gypsum usage in imgated agriculture. In “Proceedings of the International Symposium on Phosphogypsum” (D. P. Boms and P. W. Boody, eds.), Publ. 01-0010 17,pp. 133- 169.Florida Institute of Phosphate Research, Bartow, Florida. Oster, J. D., and Frenkel, H. 1980. The chemistry of the reclamation of sodic soils with gypsum and lime. Soil Sci. SOC.Am. Proc. 44,41-45. Parfitt, R. L.,and Smart, R. St. C. 1978.The mechanism of sulfate adsorption on iron oxides. Soil Sci. SOC.Am. J. 42,48 - 50. Parker, D. R., Kinraide, T. B., and Zelazny, L. W. 1988.Aluminum speciation and phytotoxicity in dilute hydroxy-aluminum solutions. Soil Sci. SOC.Am. J. 52,438-444. Parkpian, P., Pongsakul, P., and Sangtong, P. 1991.Characteristicsof acid soils in Thailand A review. In “Plant-Soil Interactions at Low pH” (R. J. Wright, V. C. Baligar, and R. P. Mumnann, eds.), pp. 397-405. Kluwer Academic Publishers, Dordrecht, The Netherlands. Parr, J. F., and Bertrand, A. R. 1960.Water infdtration into soils. Adv. Agron. 12,311-363. Paulino, V.T., and Malavolta, E. 1989.Effects of phosphogypsum, phosphorus, and lime on Andropogon gayanus Kunth. Proc. Int. Grassl. Congr.. J6th pp. 39-40. Pavan, M. A., Bingham, F. T., and Pratt, P. F. 1982. Toxicity of aluminum to coffee in Ultisols and Oxisols amended with CaCO,, MgCO, and CaSO4.2H2O. Soil Sci. SOC. Am. J. 46,1201 - 1207. Pavan, M. A.,Bingham, F. T., and Pratt,P. F. 1984.Redistribution of exchangeable calcium, magnesium, and aluminum following lime or gypsum applications to a Brazilian Oxisol. Soil Sci. SOC.Am. J. 48,33-38. Pearson, R. W. 1966.Soil environment and root development. In “Plant Environment and

114

I. S. ALCORDO AND J. E. RECHCIGL

Efficient Water Use” (W. H. Pierre, Don Kirkham, John Pesek, and Robert Shaw, eds.), pp. 95- 126. American Society of Agronomy, Madison, Wisconsin. Pearson, R. W., Childs, J., and Lund, Z. F. 1973. Uniformity of limestone mixing in acid subsoils as a factor in cotton root penetration. Soil Sci. Soc. Am. Proc. 37, 727-732. Perkins, H. F., Owen, V., Hammel, J. E., and Price, E. A. 1982. Soil characteristics of the plant science farm of the University of Georgia College Experiment Stations. Ga., Agric. Exp. Stn.. Bull. 287. Pettry, D. E., Carlisle, V. W., and Caldwell, R. E. 1965. Spodic horizons in selected Leon and Immokalee soils. Proc. Soil Crop Sci. SOC.Flu. 25, 160- 172. Plucknett, D. L., and Sherman, G. D. 1963. Extractable aluminum in some Hawaiian soils. Soil Sci. SOC.Am. Proc. 27, 39 -4 1. Porter, D. H., Garren, K. H., and Van Schaik, P. H. 1975. Pod breakdown resistance in peanuts. Peanut Sci. 2, I5 - 18. Quirk, J. P., and Panabokke, C. R. 1962. Pore volume-size distribution and swelling of natural soil aggregates. J. Soil Sci. 13, 7 1 - 8 1. Rabuffetti, A., and Kamprath, E. J. 1977. Yield, Nand S content of corn as affected by N and S fertilization on Coastal Plain soils. Agron. J. 69,785-788. Radcliffe, D. E., Clark, J. L., and Sumner, M. E. 1986. Effect of gypsum and deeprooting perennials on subsoil mechanical impedance. SoilSci. SOC.Am. J. 50, 1566- 1570. Rajan, S. S. S. 1978. Sulfate adsorbed on hydrous alumina, ligands displaced, and changes in surface charge. Soil Sci. Soc. Am. J. 42,39-44. Rajan, S. S. S. 1979. Adsorption and desorption of sulfate and charge relationships in allophanic clays. Soil Sci. SOC.Am. J. 43,65-69. Rechcigl, J. E. 1989. Importance of sulfur fertilization of forages. Flu. Cattleman 53 (no. 7 ) , 52. Rechcigl, J. E., and Alcordo, I. S. 1991. “Effect of Phosphogypsum on Bahiagrass Yield and Quality,” Agron. Abstr., p. 298. American Society of Agronomy, Madison, Wisconsin. Rechcigl, J. E., and Payne, G. G. 1988. “Liming Ryegrass,” Res. Rep. RC-1988-13. Agricultural Research and Education Center, University of Florida, Ona, Florida. Rechcigl, J. E., Edmisten, K. L., Wolf, D. D., and Reneau, R. B., Jr. 1988a. Response of alfalfa grown on acid soil to different chemical amendments. Agron. J. 80,5 I5 - 5 18. Rechcigl, J. E., Reneau, R. B., Jr., and Zelazny, L. W. 1988b. Soil solution Al as a measure of A1 toxicity to alfalfa in acid soils. Commun.Soil Sci. Plant Anal. 19,989- 1001. Rechcigl, J. E., Payne, G. G., and Stephenson, R. J. 1989. Influence of sulfur and nitrogen on bahiagrass. Proc. Int. Grassl. Congr., 16th pp. 27-28. Rechcigl, J. E., Alcordo, I. S., and Roessler, C. E. 1992a. “Influence of Phosphogypsum on Forage Yield and Quality, and on the Environment in a Typical Florida Spodosol Soil,” Prog. Rep. Florida Institute of Phosphate Research, Bartow, Florida. Rechcigl, J. E., Alcordo, 1. S., Roessler, C. E., and Stief, L. 1992b. Methods of measuring ambient atmospheric radon and radon surface flux associated with phosphogypsum treatment of a Florida Spodosol soil. Commun.Soil Sci. Plant Anal. 23,258 1-2594. Rees, M. C., Davis, P., and Hogan, J. P. 1982. The effect of fertilizer and supplementary sulphur on the digestion of Digitaria pentzii in sheep. Proc. Aust. SOC.Anim. Prod. 14, 681. Reeve, N. G., and Sumner, M. E. 1972. Amelioration of subsoil acidity in Natal Oxisols by leaching of surface-applied amendments. Agrochemophysica 4, I - 6. Reicosky, D. C., Cassel, D. K., Blevins, R. L., Gill, W. R., and Naderman, G. C. 1977. Conservation tillage in the Southeast. J. Soil Water Conserv. 32, 13- 19. Reneau, R. B., Jr., and Hawkins, G. W. 1980. Corn and soybeans respond to sulfur in Virginia. Sulphur Agric. 4, 7 - 1 1.

PHOSPHOGYPSUM IN AGRICULTURE

115

Rengasamy, P., Greene, R. S. B., Ford, G. W., and Mehanni, A. H. 1984. Indentification of dispersive behavior and the management of red brown earths. Aust. J. Soil Res. 22, 41 3-432. Rengel, Z., and Robinson, D. L. 1989. Aluminum effects on growth and macronutrient uptake by annual ryegrass. Agron. J. 81,208 - 2 15. Rhue, R. D. 1971. “Availability and Residual Effects of Gypsum and Elemental Sulfur on Two Soil Series in North Carolina,” M.S. thesis. Department of Soil Science, North Carolina State University, Raleigh, North Carolina. Rhue, R. D., and Kamprath, E. J. 1973. Leaching losses of sulfur during winter months when applied as gypsum, elemental S or priUed S. Agron. J. 65,603 -605. Richards, L. A. 1954. Diagnosis and improvement of saline and alkali soils. U.S.,Dep. Agric., Agric. Handb. 60. Rios, M. A., and Pearson, R. W. 1964. The effect of some chemical environmental factors on cotton root behavior. SoilSci. SOC.Am. Proc. 28,232-235. Ritchey, K. D., S o w , D. M. G., Lobato, E., and Correa, 0. 1980. Calcium leaching to increase rooting depth in a Brazilian Savannah Oxisol. Agron. J. 72,40-44. Roessler, C. E. 1987. Radiological aspects of phosphogypsum. In “Proceedings of the Symposium on Natural Radiations and Technologically Enhanced Natural Radiation in Florida” (W. S. Properzio, ed.). Florida Chapter of the Health Physics Society, Winter Haven, Florida. Roessler, C. E. 1988. Radiological assessment of the application of phosphogypsum to agricultural land. In “Proceedings of the Second International Symposium on Phosphogypsum,” h b l . 01-037-055, pp. 5-23. Florida Institute of Phosphate Research, Bartow, Florida. Roessler, C. E. 1990. Control of radium in phosphate mining, beneficiation and chemical processing. Tech. Rep. Ser.-I.A.E.A. 310(Vol. 2), 269-294. Roue, R. 1972. Rapid working level monitoring. Health Phys. 22, 233-238. Rowell, D. L. 1965. Influence of positive charge on the inter- and intra-crystallineswelling of oriented aggregates of Na montmorillonite in NaCl solutions. Soil Sci. 100, 340347. Rowell, D. L., Payne, D., and Ahmad, N. 1969. The effect of the concentration and movement of solutions on the swelling, dispersion, and movement of clay in saline and alkali soils. J. Soil Sci. 20, 176- 188. Russell, E. W. 1971. Soil structure: Its maintenance and improvement. J. Soil. Sci. 22, 137-151. Russell, E. W. 1973. “Soil Conditions and Plant Growth,” loth Ed. Longman, New York. Sanchez, P. A. 1977. Advances in the management of Oxisols and Ultisols in tropical South America. In “Proceedings of the International Seminar on Soil Environment and Fertility Management in Intensive Agriculture,” pp. 535-566. Society of the Science of Soil and Manure, Tokyo. Scotter, D. R., and Loveday, J. 1966. Physical changes in seedbed material resulting from the application of dissolved gypsum. Aust. J. Soil Res. 4,69-75. Sedgley, R. H. 1962. Effects of disruption and flocculation on pore-space changes in beds of clay aggregates. Soil Sci. 94, 357-365. Shainberg, I., and Singer, M. J. 1985. Effect of electrolytic concentration on the hydraulic properties of depositional crust. Soil Sci. SOC.Am. J. 49, 1260- 1263. Shainberg, I., Singer, M. J., and Janitzky, P. 1987. Effect of aluminum and iron oxides on hydraulic conductivity of sandy loam soil. Soil Sci. SOC.Am. J. 51, 1283- 1287. Shainberg, I., Sumner, M. E., Miller, W. P., Farina, M. P. W., Pavan, M. A., and Fey, M. V. 1989. Use of gypsum on soils: A review. Adv. Soil Sci. 9, 1 - 1 1 1.

116

I. S. ALCORDO AND J. E. RECHCIGL

Shamshuddin, J., Fauziah, I. C., and Sharifuddin, H. A. H. 1991. Effects of limestone and gypsum application to a Malaysian Ultisol on soil solution composition and yields of maize and groundnut. In “Plant-Soil Interactions at Low p H (R. J. Wright, V. C. Baligar, and R. P. Murrmann, eds.), pp. 397-405. Kluwer Academic Publishers, Dordrecht, The Netherlands. Shomer, I., and Mingelrin, U. 1978. A direct procedure for determining number of plates in tactoids of smectites: The Na/Ca montmorillonite case. Clay Miner. 26, 135- 138. Shuman, L. M., Ramseur, E. L., and Duncan, R. R. 1990. Soil aluminum effects on the growth and aluminum concentration of sorghum. Agron. J. 82,3 13- 3 18. Sigua, G. C., and Hudnall, W. H. 1991. Gypsum and water management interactions for revegetation and productivity improvement of brackish marsh in Louisiana. Commun. Soil Sci. Plant Anal. 22, 172 1 - 1739. Simon, S. L., and Ibrahim, S. A. 1990. Biological uptake of radium by terrestrial plants. Tech. Rep. Ser.-I.A.E.A. 310(Vol. I), 545-597. Simpson, J. R., Pinkerton, A., and Lazdovskis, J. 1979. Interacting effects of subsoil acidity and water on the root behaviour and shoot growth of some genotypes of lucerne (Medicagosafiva L.). Ausf. J. Agric. Res. 30,609-619. Sims, H. J., and Rooney, D. R. 1965. Gypsum for difficult clay wheat growing soils. J. Dep. Agric., Victoria,Ausf.63,401 -409. Singh, M., and Singh, N. 1979. The effect of forms of selenium on the accumulation of selenium, sulphur and forms of nitrogen and phosphorus in forage cowpea. Soil Sci. 127, 264-269. Singh, S. S., and Brydon, J. E. 1969. Solubility of basic aluminum sulfates at equilibrium in solution and in the presence of montmorillonite. SoilSci. 107, 12- 16. Singh, S. S., and Miles, N. M. 1978. Effect of sulfate ions on the stability and exchange characteristics of an aluminum-interlayered Wyoming bentonite. SoiiSci. 126, 323329. Singh, A., Chhabra, R., and Abrol, I. P. 1979a. Effect of fluorine and phosphorus on the yield and chemical composition of rice (Oryza saliva L.) grown in soils of two sodicities. Soil Sci. 127, 86-93. Singh, A., Chhabra, R., and Abrol, I. P. 1979b. Effect of fluorine and phosphorus applied to a sodic soil on their availability and on yield and chemical composition of wheat. Soil Sci. 128,90-97. Soileau, J. M., and Engelstad, 0. P. 1969. Cotton growth in an acid fragipan subsoil: I. Effects of physical soil properties, liming, and fertilization on root penetration. Soil Sci. SOC. Am. Proc. 33,915-919. Spencer, K. 1982. Effect of sulphur application on selenium content of subterranean clover plants grown at different levels of selenium supply. Ausf. J. Exp. Agric. Anim. Husb. 22, 420-427. Sposito, G., and Mattigod, S. V. 1980. “GEOCHEM: A Computer Analysis Program for the Calculation of Chemical Equilibra in Soil Solution and Other Natural Water Sys tems.” Kearney Foundation of Soil Science, University of California, Riverside, California. Stewart, B. A., and Porter, L. K. 1969. Nitrogen-sulfur relationships in wheat (Tritinrm aesfivum L.), corn (Zea mays), and beans (Phaseoh vulgaris).Agron. J. 61,267-271. Sullivan, G. A., Jones, G. L., and Moore, R. P. 1974. Effects of dolomitic limestone, gypsum, and potassium on yield and seed quality of peanuts. Peanut Sci. 1,73-77. The Sulphur Institute. 1982. “The Fourth Major Nutrient.” The Sulphur Institute, Washington, D.C. Sumner, M. E. 1990. “Gypsum as an Ameliorant for the Subsoil Acidity Syndrome,” Publ. 01-024-090. Florida Institute of Phosphate Research, Bartow, Florida.

PHOSPHOGYPSUM IN AGRICULTURE

117

Sumner, M. E., Shahandeh, H., Bouton, J., and Hammel, J. 1986. Amelioration of an acid soil profile through deep liming and surface application of gypsum. Soil Sci. SOC.Am. J. 50, 1254-1258. Sumner, M. E., Radcliffe, D. E., McCray, M., and Clark, R. L. 1990. Gypsum as an ameliorant for subsoil hardpans. Soil Technol. 3,253-258. Syed, 0. S. R. 1987. “Deleterious Effects of Gypsum Amendments on Potassium and Magnesium Status of Field Soils,” M.S. thesis. University of Georgia, Athens, Georgia. Tabatabai, M. A. (ed.) 1986. “Sulfur in Agriculture.” American Society of Agronomy/Crop Science Society of America/Soil Science Society of America, Madison, Wisconsin. Takai, Y., and Kamura, T. 1966. The mechanism of reduction in waterlogged paddy soil. Folia Microbiol. (Prague) 11, 304-313. Takkar, P. N. 1986. Sulphur in Indian soils-Forms and their distribution, deficiency and indices of availability. In “Proceedings of the International Symposium on Sulphur in Agricultural Soils” (S. Portch and Sk. G. Hussain, eds.), pp. 299-350. The Bangladesh Agricultural Research Council and The Sulphur Institute, Dhaka, Bangladesh, India. Tama, K., and El-Swaify, S. A. 1978. Charge, colloidal and structural stability interrelationships for Oxidic soils. In “Modification of Soil Structure” (W. W. Emerson, R. D. Bond, and A. R. Dexter, eds.), pp. 40-49. Wiley, New York. Tandon, H. L. S. 1986. Sulphur in Indian agriculture-An assessment. In “Proceedings of the International Symposium on Sulphur in Agricultural Soils” (S. Portch and Sk. G. Hussain, eds.), pp. 46-64. The Bangladesh Agricultural Research Council and The Sulphur Institute, Dhaka, Bangladesh, India. Tanji, K. K., Donnen, L. D., Ferry, G. V., and Ayers, R. S. 1972. Computer simulation analysis on reclamation of salt-affected soil in San Joaquin Valley, California. Soil Sci. Soc. Am. Proc. 36, 127-133. Thomas, W. E., Loosli, J. K., Williams, H. H., and Maynard, L. A. 1951. The utilization of inorganic sulfates and urea nitrogen by lambs. J. Nutr. 43, 5 15 - 523. Till, J. E., and Meyer, H. R. 1983. Radiological assessment. US.Nucl. Regul. Comm. [Rep.] NUREG/CP NUREG/CR-3332 ORNG5968. Tisdale, S. L., Nelson, W. L., and Beaton, J. D. 1985. “Soil Fertility and Fertilizers,” 4th Ed. Macmillan, New York. Tisdale, S. L., Reneau, R. B., Jr., and Platuo, J. S. 1986. Atlas of sulfur deficiencies. In “Sulfur in Agriculture” (M. A. Tabatabai, ed.), pp. 296-322. American Society of Agronomy/Crop Science Society of America/Soil Science Society of America, Madison, Wisconsin. van der Sloot, H. A., and de Groot, A. J. 1985. Environmental aspects of the utilization of pellets prepared from fosfogypsum, coal fly ash and cement. In “Contaminated Soil” (J. W. Assink and W. J. van den Brink, eds.), 1st Int. TNO Conf. Contaminated Soil, pp. 9 19 - 923. Utrecht, The Netherlands. van Raij, B., and Peech, M. 1972. Electrochemical properties of some Oxisols and Alfisols of the tropics. Soil Sci. SOC.Am. Proc. 36, 587-593. Vitti, G. C., Ferreira, M. E., and Malavolta, E. 1986. Respostas de culturas anuais e perenes. In “I Seminario Sobre o Uso do Fosfogesso na Agricultura,” pp. 17-43. Empresa Brasileira de Pesquisa Agropecuaria, Brasilia, Brazil. von Uexkull, H. R. 1986. Sulphur interactions with other plant nutrients. In “Proceedings of the International Symposium OD Sulphur in Agricultural Soils” (S. Portch and Sk. G. Hussain, eds.), pp. 212-241. The Bangladesh Agricultural Research Council and The Sulphur Institute, Dhaka, Bangladesh, India. Wagatsurna, T., and Ezoe, Y. 1985. Effect of pH on ionic species of aluminum in medium and on aluminum toxicity under solution culture. Soil Sci. Plant Nutr. 31,547 - 56 1.

118

I. S. ALCORDO AND J. E. RECHCIGL

Walker, M. E., and Csinos, A. S. 1980. Effect of gypsum on yield, grade and incidence of pod rot in five peanut cultivars. Peanut Sci. 7, 109- 113. Walthall, P. M., and Day, W. J. 1988. “Aluminum Solubility as a Function of Clay Mineralogy and Weathering,” Agron. Abstr., p. 3 10. American Society of Agronomy, Madison, Wisconsin. Westerman, R. L., O’Hanlon, R. J., Fox, G. L., and Minter, D. L. 1983. Nitrogen fertilizer efficiency in bermudagrass production. Soil Sci. SOC.Am. J. 47,8 10- 8 17. manger, P. D. 1968- 1969. Sulphur in ruminant nutrition: Metabolism and importance. Sulphur Inst. J. 4(4), 9- 12. Wolf, J. M. 1975. Soil-water relations in Oxisols of Puerto Rico and Brazil. In “Soil Management in Tropical America” (E. Bomemisza and A. Alvarado, eds.), pp. 145154. North Carolina State University, Raleigh, North Carolina. Woodhouse, W. W. 1969. Long-term fertility requirements of Coastal bermudagrass. 111. Sulphur. Agron. J. 61,705 -708. Wright, R. J., Hem, J. L., Baligar, V. C., and Bennett, 0. L. 1985. The effect of surfaceapplied amendments on barley root growth in an acid subsoil. Commun. Soil Sci. Plant Anal. 16, 179-192. Zhongqun, L. 1986. Preliminary study of soil sulphur and sulphur fertilizer efficiency in China. In “Proceedings of the International Symposium on Sulphur in Agricultural Soils” (S. Portch and Sk. G. Hussain, eds.),pp. 371 -388. The Bangladesh Agricultural Research Council and The Sulphur Institute, Dhaka, Bangladesh, India.

NUTRIENT CYCLING AND SOILFERTILITY IN THE GRAZED PASTURE ECOSYSTEM R. J. Haynes and P. H. Williams New Zealand Institute for Crop and Food Research, Canterbury Agriculture and Science Centre, Christchurch, New Zealand

I. Introduction 11. The Pasture System and Its Effect on Soil Properties A. Nature of Pastoral Systems B. Role of Fertilizer C. Effect on Soil Properties 111. Nutrient Returns in Feces and Urine A. Quantities Returned B. Form of Return c. Pattern of Return D. Role of Excreta in Nutrient Cycling IV. Soil Processes and Pasture Response in Excreta-Affected Areas A. Release of Nutrients from Feces B. Response of Pasture in the Fecal Patch C. Movement and Transformations of Nutrients from Urine D. Response of Pasture in the Urine Patch V. Modeling Nutrient Cycling under Pasture A. Models of Nutrient Cycles B. Use of Models for Fertilizer Recommendations VI. Summary and Conclusions References

I. INTRODUCTION Grazed pastures are complex ecosystemsthat are constantly modified by the activities of man and his management of domestic herbivores, such as cattle and sheep. The majority of highly productive pastures of the world Ahanre5 in A g m ~ n r y Vol. , 49 Copyright 0 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.

119

120

R. J. HAYNES AND P. H. WILLIAMS

were created by removing the existing forest vegetation and sowing improved grass and, sometimes, legume species (Tothill, 1978; Snaydon, 198 1). Fertilizer applications to such pastures have been extensive, and continued applications are necessary to maintain the highly productive pasture species in the sward (Tothill, 1978). Indeed, nutrient availability exercises considerablecontrol over pasture growth and thus the number of domestic animals that can be supported and their productivity. Increased nutrient availability, through fertilizer applications, and the subsequent increases in pasture production offer the potential for increased animal production. Withdrawal of fertilizer typically leads to a decrease in production, the steady ingress of low-fertility-demanding weed species, reversion to scrub, and eventually back to forest (Levy, 1970; Snaydon, 198 1). Grazing animals have a dominant effect on the movement of nutrients through the soil/plant/animal system and thus on the fertility of pasture soils (Wilkinson and Lowrey, 1973; Mott, 1974). This is because animals use only a small proportion of the nutrients they ingest, and 60-9996 of the ingested nutrients is returned to the pasture in the form of dung and urine (Barrow, 1987). The urine and dung patches are therefore the areas of pasture where nutrients are recycled from excreta to soil and back to pasture plants. Although excretal patches may cover only 30-4096 of the pasture surface annually, the high nutrient input stimulates herbage growth, which may represent up to 70% of the total annual pasture production (Saunders, 1984). Nutrients also leak from the system in the patch areas through gaseous and leaching losses. The more efficiently nutrients are cycled within the system, the lower the losses, the more sustainable the system, and the lower the maintenance fertilizer input that is required. An understanding of the cycling of nutrients within the soil/plant/anima1 system is therefore central to an understanding of the fertility of pasture soils. Indeed, nutrient cycling models that link the soil/plant/anima1 system together are used routinely in some parts of the world to estimate the fertilizer requirement of grazed pastures. Several previous reviews of pasture nutrient cycling have been published (Wilkinson and Lowrey, 1973; Mott, 1974; Floate, 1981, 1987; Till, 1981; Ryden, 1984; Gillingham, 1987), but most are specific to one nutrient and/or present overall nutrient budgets. In this paper we review the cycling of nutrients within pasture soils. The major emphasis is on the central role of the grazing animal in influencing soil fertility, particularly in the dung and urine patches. In addition, we review the progress that has been made in modeling the cycling of nutrients in pasture systems.

NUTRIENT CYCLING UNDER GRAZED PASTURE

121

11. T H E PASTURE SYSTEM AND ITS EFFECT ON SOIL

PROPERTIES

A. NATURE OF PASTORAL SYSTEMS Managed pastures cover about 20% of the earth’s land surface and rangelands cover another 30% (Snaydon, 1981). Of this total area, only a small proportion was originally natural grassland and most of this has now been affected by man. Natural grasslands generally occurred as a climax ecosystem in areas where growth of trees was restricted by climatic or soil conditions. For example, large areas of grassland existed where rainfaii was inadequate for forests (e.g., steppes, prairies, and pampas). The productivity of natural and seminatural grasslands is usually low because the factors that limit the distribution of trees also limit plant growth in general. Most of the highly productive pastures of the world are man-made, and were originally created by felling and/or burning natural forest. Man-made grasslands, especially those of humid temperate and tropical areas, are often highly productive. Most pastures may be considered in seral condition even though they are apparently in equilibrium with the managed environment. Rather than consisting of climax grass species, improved pastures are mixtures of largely seral species of grasses and herbs that originated in forest clearings and woodland margins (Tothill, 1978; Williams, 1981). Such pastures are botanically unstable so that even small changes in environmental conditions, soil nutrient status, grazing management, or stocking rate can lead to rapid changes in botanical composition (Snaydon, 1987). The continued existence of pastures depends on management factors such as grazing intensity and frequency, burning, intermittent plowing, resowing, fertilizer applications, and the use of herbicides. Without this management, most pastures would revert to scrub and eventually back to forest (Moore, 1964; Snaydon, 1981). The productivity and botanical composition of pastures can be rapidly and substantially altered by grazing animals. Animals defoliate, selectively graze, and trample pastures, deposit dung and urine onto pasture, and disperse seeds (Watkin and Clements, 1978). The effects may be harmful or beneficial and act simultaneously but with different emphasis in different situations. Of course, in relation to soil fertility the return of nutrients in dung and urine is of central importance.

122

R. J. HAYNES A N D P. H. WILLIAMS

B. ROLEOF FERTILIZER Fertilizer applications have greatly increased pasture production on many grassland soils that are inherently deficient in nutrients or have become so following forest clearance. Indeed, the high-yielding, improved pasture plants used are generally sera1 species that are adapted to highfertility conditions (Tothill, 1978; Snaydon, 1981) and do not perform well in infertile and/or acid soil conditions. Thus, the application of lime and fertilizers is necessary during pasture development to boost natural soil fertility to a level capable of supporting the oversown pasture plants. During development, sustained heavy grazing pressure has often been necessary to control the secondary growth and such pressure could only be achieved with adequate subdivision and highly productive pastures supplied with nutrients (Levy, 1956). Once an adequate level of fertility is attained, maintenance fertilizer applications are usually still required in order to sustain a high level of pasture production. The major reason for this is that nutrients are lost from the pasture system through the actions of the grazing animals (Williams and Haynes, 1990a). Losses are caused by transfer of nutrients from major grazing areas through deposition of dung and urine at stock camps and unproductive farm areas (e.g., raceways and yards). Losses also occur through concentration of nutrients into small volumes of soil (dung and urine patches) in quantities greater than the short-term requirements of the pasture plants in those areas. Nutrients therefore accumulate in patch areas and are subject to gaseous and leaching losses. In addition, nutrients can be “lost” from the available soil pool during cycling due to chemical fixation (e.g., phosphate adsorption) and/or immobilization into organic forms. In many parts of the world phosphate fertilizers have been required to boost pasture production, whereas on acid soils lime has also been an important factor in pasture development. Applications of K and S and trace elements (e.g., Cu, Co, Mo, and Se) have also been important on some soils to increase pasture growth and/or improve stock health. In New Zealand and Australia, phosphate applications have been particularly important in the success of clover-based pastures (Donald, 1965; Levy, 1970). High levels of available P in soils are required in order to maintain the presence and N,-fixing activity of the clovers in the pasture and hence maintain the N input into the grass/legume pasture (Haynes, 1981). N, fixation rates in the range of 100-300 kg N ha-’ are common for grass/ clover pastures (Hoglund and Brock, 1987). Despite this, in highly productive grass/clover pastures the grass component is often N deficient (Henzell, 1981) and strategic applications of small amounts of N (25 - 50 kg N

NUTRIENT CYCLING UNDER GRAZED PASTURE

123

ha-l) are sometimes applied in early spring and autumn when clover growth is weak (O'Connor, 1982). In the grass-based intensive pasture production of Europe and the United States, N fertilizers are used widely. Pure grass pasture often responds linearly up to 200-400 kg N ha-' yr-' and application rates in this range are common (Morrison, 1987). Where pastures are cut for conservation, large quantities of nutrients are removed and the optimum N rate can be greater than that under grazed swards, where N is returned to the pasture in the form of dung and urine (Baker, 1986). Where high N rates are used, applications of P and K may also be required (Prins et al., 1986).

C. EFFECTON SOILPROPERTIES 1 . Soil Organic Matter and Nutrient Status

A major effect of improved pastures is often an increase in the organic matter content of the surface soil. The rate of organic matter accumulation and the time taken to reach an equilibrium, where organic matter additions are balanced by mineralization and losses, vary considerably with initial organic matter level, soil type, climate, and management (Whitehead, 1970; Simpson, 1987). Organic C, N, s, and P do not necessarily accumulate at the same rate at a particular site and neither do they necessarily reach equilibrium at the same time (Jackman, 1964; Quin and Rickard, 1981). This is demonstrated by the results shown in Fig. 1. For Taupo soil the biggest percentage increase during development was for S, but for Oropi the biggest increase was for P. Percentage increases differed greatly for C, N, S, and P in both soils. Such differences are attributable to differences in the individual cycles of the various elements (Stewart, 1984; Williams and Haynes, 1990b). In some locations, such as in southern Australia where levels of soil organic matter were initially very low (e.g., 0.01-O.l%C and 0.0010.0 1% N), large linear increases in organic matter content have been measured for up to 50 years under subterranean clover pastures (Russell, 1986). Accumulations of organic N in the range of 40-80 kg N ha-' yr-I have been frequently recorded under such pastures (Simpson, 1987). The initial increase in organic matter occurs in the surface soil, but with time there is a deepening of accumulation in the profile (Russell, 1986). Increases have been noted to a depth of 20cm. In some situations on irrigated pastures, organic matter can also accumulate on the surface of the

R. J. HAYNES AND P. H. WILLIAMS

124

0.9

3 5m

Nitrogen

Nitrogen

rn

0.7

v

0.5

c

0.4 0

0.3 ,

0.1

I

2I

I

.

1

I

I

I

'0.02

0 0.09 . Phosphorus 0 7

r

;

'

I

g

I

p

,

L

144%

:. 0.05 m

8 0.03 0.01

0.04

0

I

10

20

30

Pasture age (years)

0.02 40 0

rn rn

10

20

30

40

Pasture age (years)

Figure 1. Accumulation of organic C, total N, organic S, and organic P in two soils (0-7.5 cm) during pasture development. Percentage increase compared with time zero shown at 30 and 50 years, respectively, for Taupo and Oropi soils (Data from Jackman, 1964.)

soil as a dense, partly decomposed mat of up to 2.5 cm in thickness (Kleinig, 1966; Rixon, 1966a). In other locations, such as most of New Zealand, the soil organic matter content was initially reasonably high (e.g., 4- 10% organic C), but during the clearing and burning of forest vegetation (and sometimes subsequent plowing of the soil) there was a marked decrease in organic matter content (particularly organic C) in the surface soil (Walker et al., 1959; Jackman,

12s

NUTRIENT CYCLING UNDER GRAZED PASTURE

1960). Such a decrease can be seen in Table I, where organic C content decreased from 10 to 4.2% following initial development. During ensuing pasture development there was, however, an increase in organic matter accumulation and a marked increase in organic C, S, P, and total N content (Table I). The major fertilizer applied to New Zealand and Australian pastures is superphosphate, which contains 9% P and 11% S. The application of superphosphate results in an increase in the total inorganic P content of the soil and therefore an increase in the extractable (available) P content (Rixon, 1966b; Quin and Rickard, 1981; Nguyen et al., 1989). Generally, the P is accumulated mainly into the alkali-extractable Al-P and Fe-P fractions (Saunders, 1959; Batten et al., 1979), which represent phosphate adsorbed to soil colloids. Organic P increases at a slower rate than inorganic P (Saunders, 1959; Walker el al., 1959; Batten et al., 1979). Thus, Walker et al. (1959) observed that although organic P made up 90% of total P in soil (0- 10 cm) under undeveloped scrub, following 25 years of pasture development the percentage of total P in an organic form fell to 5596, even though there was an increase in the actual amount of organic P in the soil (Table I). The whole of the soil organic P pool does not necessarily represent a static pool of P unavailable to pasture plants. Tate et al. ( 1991) observed a temporal pattern in labile (easily extractable) soil organic P content of pasture soils, suggesting that such a pool may contribute to pasture P nutrition. They noted net immobilization of P in early winter and net mineralization in spring concomitant with rapid pasture growth and P uptake. Unlike phosphate, sulfate is very mobile in soils and is easily leached, thus inorganic S does not generally accumulate in the soil. In pastures fertilized with superphosphate, much of the applied S is, however, accumulated into the organic matter (Walker et al., 1959; Jackman, 1964). Table I Changes in Organic Matter Accumulation (0-10 cm) during 25 Years of Pasture Development with Annual Superphosphate Applicationsu Stage of development

OrganicC

TotalN

Totals

(%)

(%)

(%I

(%)

(%)

organic P as apercentage oftotalP

Undeveloped scrub If years of pasture 25 years of pasture

10

0.31

0.036 0.038

0.043 0.052 0.058

0.047 0.070 0.103

90 75 55

a

4.2

6.3

Data from Walker ef al. (1959).

0.20 0.59

0.071

OrganicP TotalP

126

R. J. HAYNES AND P. H. WILLIAMS

In legume-based pastures, the N content of the soil increases rapidly (due to N, fixation) but the C content often increases more slowly (Walker et al., 1959; Jackman, 1964). As a consequence, there is a characteristic gradual decrease in the C :N ratio during pasture development (Walker et al., 1959; Watson, 1969). Walker et al. (1959) found that the C :N ratio of an undeveloped scrub soil was 33 : 1, but after 25 years of grass/white clover management the C :N ratio had decreased to 1 1 : 1. However, to some extent, this effect is due to the wide initial C :N ratio of the soils because the C : N ratio of pasture soils usually approaches 1O:l (Whitehead, 1986). For example, for a soil with a low initial C:N ratio (9: I), Garwood et al. (1977) found the C:N ratio increased slightly to 10: 1 as organic matter accumulated under a grass/clover pasture. Changes in management will influence the equilibrium organic matter level that is reached. Russell and Harvey (1959) showed that after 30 years of intensive dairy farming on irrigated grass/clover pastures, soils with a high initial total N content exhibited a decline in N concentration whereas sites initially low in N exhibited an increase toward the same equilibrium N level. In New Zealand, intensified use of long-established pastures has been found to cause a decline in the soil N content (Field and Ball, 1981). 2. SoilpH

Under improved pasture there is a tendency for soil pH to decline over time (Williams and Donald, 1957; Russell, 1960; Rixon, 1966a; Batten et al., 1979;Williams, 1980).The increased soil organic matter content in the surface soil results in an increase in cation exchange capacity (CEC) and an increase in H+ saturation of the exchange complex (Williams, 1980). Excretion of H+ in the pasture rhizosphere due to excess cation uptake by the N,-fixing clover (Haynes, 1983; Jarvis and Hatch, 1985) plus nitrification of ammonium in the urine patch and subsequent nitrate leaching (Helyar, 1976) are also likely to be major contributors to such acidification. The rate of decrease in surface soil pH varies in different soils and environments (Russell, 1986). An example of the decline in pH in the surface 10 cm under a subterranean clover pasture is shown in Fig. 2. The rate of decrease in pH was greatest during the first 10 years and over a period of 50 years the decrease in pH was approximately one unit (Williams, 1980). Although it may take from 25 to 50 years for the pH to decrease by one unit (Lee, 1980; Williams, 1980), such a decrease has led to severe reductions in pasture production due to the buildup of phytotoxic levels of soluble and exchangeable soil Al and Mn (Williams and David, 1976;

NUTRIENT CYCLING UNDER GRAZED PASTURE

127

U

.-0 d L

Soil pH Figure 2. Effect of age of subterranean clover pasture on surface soil (0- 10 cm) pH. (Redrawn from Williams, 1980.)

Cregan et al., 1979; Evans et al., 1988). Such acidity problems have developed mainly on unlimed pastures in southern Australia on soils with low initial organic matter contents and low buffering capacities. For New Zealand pastures it has long been recognized that regular lime applications (at 2-4 yearly intervals) are required to maintain soil pH at a suitable value (e.g., 5.6 to 6.2) for optimum pasture production (During, 1984). 3. Biological Activity

The overall biological activity of the soil generally increases under grazed pasture (Russell, 1986). Associated with the high content of soil organic matter and dense mass of pasture roots is a large microbial biomass in the pasture rhizosphere. The microbial biomass is therefore characteristically large under improved pasture (e.g., 1200,ug C g-I) and it represents a reasonably large labile pool of nutrients (e.g., commonly 150-225 kg N ha-' and 10-60 kg P ha-') (Sarathchandra et al., 1984, 1988; Perrott and Sarathchandra, 1989). Bristow and Jarvis ( 199 1) estimated mean biomass N levels under grazed pasture of 138-246 kg N ha-' and observed that biomass N comprised I 1, 3, and 5 times more N at any one time than was

R. J. HAYNES AND P. H. WILLIAMS

128

present in the mineral N component of the soil, the standing crop, or excretal returns, respectively. Although some seasonal patterns have been observed in the levels of nutrients present in the microbial biomass under pasture (Sarathchandra et al., 1988; Tate et al., 1991),it seems likely that it is the magnitudes of the various nutrient fluxes through the large labile biomass pool that are of greatest significance to pasture soil fertility. Jenkinson and Oades (1979) observed that an improved pasture resulted in a threefold increase in soil ATP content (an index of microbial biomass) compared with unimproved scrub. Similarly, results presented in Table I1 show that soil microbial biomass was higher under an improved irrigated grazed pasture than under an unimproved wilderness area. Biomass levels were further increased under the highly productive, fertilized grazed pasture. The high organic matter content and high microbial activity under pasture also results in a high level of activity of soil enzymes, such as urease, protease, phosphatase, and sulfatase (Ross et al., 1984; Sarathchandra et al., 1984, 1988). Results in Table I1 show that sulfatase and phosphatase activities were greater under improved pasture than under unimproved wilderness and higher under a highly productive pasture than under an unfertilized pasture. The high levels of enzyme activity reflect a high rate of N, S, and P turnover through soil organic pools under pasture conditions. Native/unimproved pastures harbor complex communities of invertebrate fauna (Curry, 1987a). Management practices greatly influence the

Table I1

Effect of Long-Term Irrigation and Grazing with or without Annual Superphosphate Additions on Surface Soil (0-4 cm) Properties" Imgated and grazed

fiOpefiY ~~

~

Organic C (96) Biomass C @g g-I) Sulfatase activity (u mol product g-l h r l ) Phosphatase activity (u mol product g-' h r ' )

Wilderness area

376 kg superphosphate per hectare applied annually

Control

~

~

3.97 809 0.48

4.18 857 1.3

4.30 969 1.5

7.8

10.2

11.0

Data from R. J. Haynes and P. H. Wfiams (unpublished observations).

NUTRIENT CYCLING UNDER GRAZED PASTURE

129

invertebrate communities by affecting sward structure, composition, and productivity. In intensively managed grasslands, simplified communities of species develop that are tolerant to disturbance. Some herbivores, such as earthworms, benefit from increased food supply (e.g., dung pats) in intensively managed pastures and they increase in numbers, whereas soil microarthropods, which are favored by surface accumulation of dead plant material, decline in numbers (Hutchinson and King, 1979; Curry, 1987b). Pasture improvement has been shown to result in increased earthworm numbers (Sears and Evans, 1953; Suckling, 1975)and the weight of earthworms per hectare is closely correlated with pasture production. Suckling (1959) showed that eight years of pasture improvement increased earthworm numbers on grazed hillsides as well as on the more fertile stock camp areas. Population densities of earthworms in productive, temperate grasslands range typically from 100 to 1000 m-* (Curry, 1987a), and it has been estimated that on highly productive pastures the weight of earthworms in the soil approximates with the weight of stock carried on the surface (Sears, 1949; Waters, 1955). Earthworm populations are greater under grass/ clover than comparable all-grass pastures, and applications of N fertilizer to pastures increase their numbers, suggesting that readily available organic N is an important factor limiting earthworm populations (Watkin and Wheeler, 1966). The role of earthworms in the cycling of nutrients under pasture is substantial (Keogh, 1979; Syers and Springett, 1983) because they ingest large quantities of pasture litter and return nutrients in the form of casts. A high proportion of nutrients in casts are plant available and casts have a high enzyme activity (Syers and Springett, 1983). 4. Soil Physical Properties

Improved pastures are normally associated with improved soil structure in contrast to arable cropping, wherein structural deterioration often occurs. Where pasture development results in significant accumulation of soil organic matter, associated changes that commonly occur include an increase in aggregate stability (Clement, 1961; Clarke et al., 1967; Haynes and Swift, 1990),a decrease in bulk density (Russell, 1960; Watson, 1969), an increase in total porosity and air-filled porosity at field capacity (Garwood et al., 1977), and a greater water retentive capacity over the range - 10 to - 1500 kPa (Barrow, 1969; Garwood et al., 1977; Haynes and Swift, 1990). Soil porosity increases under pasture because of the presence of the extensive, ramified root system in the surface soil. Root growth and activity increases porosity through movement of existing soil structural units, enlargement of existing pores, and cracking caused by shrinkage and swell-

130

R. J. HAYNES A N D P. H. WILLIAMS

ing of soil due to water extraction. The high-level earthworm activity under pasture also contributes to increased porosity because the burrowing action of the worms creates soil pores through a considerable depth of soil (Syers and Springett, 1983). Under pasture, not only are soil structural pores created but aggregates are stabilized through the binding actions of soil organic matter constituents (Tisdall and Oades, 1982; Oades, 1984). Organic stabilizing agents include mucilaginous glues produced by plant roots and the rhizosphere microflora; soil humic substances, which form persistent stable complexes with mineral components of the soil; and fine roots and associated mycorrhizal hyphae, which physically enmesh soil particles (Tisdall and Oades, 1982).

111. NUTRIENT RETURNS IN FECES AND URINE

A. QUANTITIES RETURNED The amounts of nutrients returned to the soil in dung and urine vary widely between farming systems. Rough estimates can usually be made from a knowledge of the amounts of herbage consumed, its approximate nutrient composition, and information on animal requirements (Barrow, 1987). The mean nutrient content of urine and feces of dairy cows from seven farms is shown in Table 111. Some of the nutrients (e.g., K) are excreted predominantly in urine whereas others (e.g., P, Ca, Mg, Cu, Zn, Fe, and Mn) are excreted mainly in feces. Other nutrients, such as N, Na, C1, and S, are excreted in significant proportions in both feces and urine. A more detailed partitioning of total nutrient intake by dairy cows is shown in Fig. 3. These data again illustrate the contrast between the partitioning of K and that of Ca, Mg, and P between urine and dung. It is also evident that substantial proportions of P and N (26 and 17%, respectively), but only 5% of K, are removed from the farm as milk. Nutrient partitioning between dung and urine can vary depending on the nutrient content of the diet (Barrow, 1987). Similarly, the nutrient content of dung and urine can vary greatly between individual animals grazing the same pasture and for individual animals on different days and at different times of the same day (Hutton et al., 1965, 1967; Paquay et al., 1970a; Betteridge et a/., 1986; Groenwold and Keuning, 1988). Much of this variability is related to differences between animals and between days in the frequency and volume of urine and feces voided (see Section III,C, I). Such variability is caused by differencesin the intake of individual

NUTRIENT CYCLING UNDER GRAZED PASTURE

131

Table I11 Mean Nutrient Content in Urine and Feces of Lactating Cows on Seven North Carolina Dairy Farmsa

Parameter

Urine content (g liter')

Feces content (% fresh weight)

Percentage excreted in feces

Total solids Total N Total P

6.1 11.5 0.2 2.5 7.95 0.17 0.56 1.18 0.00 1 0.002 0.006 0.0002

15.4 2.9 1.2 0.6 1 0.84 1.28 0.63 0.22 0.005 0.02 0.16 0.02

85 48 95 47 28 97 78 41 95 98 99 99

C1

K Ca Mg Na

cu

Zn Fe Mn

Data from Safley ef al. ( 1984).

100

-

ao -

Retention

0,

Y

Milk

m

.-cE m

60 -

Urine

c 0 c

Dung

Lc

0 Q)

40

-

CI)

m

c

E

0,

2

20 -

cr"

n --

N

(2562)

P

(237)

K

(1720)

Ca

(726)

M9

(222)

Na

(279)

Figure 3. Percentage excretion and retention of nutrient intake in lactating dairy cows. Nutrient element intake totals (grams per day) are shown in parentheses. (Data from Hutton etal., 1965, 1967.)

132

R. J. HAYNES A N D P. H. WILLIAMS

animals and physiological differencesbetween animals in the percentage of feed intake excreted and its partitioning between dung and urine (Hutton et al., 1967; Betteridge et al., 1986). Betteridge et at. (1986), for example, measured intake and fecal and urinary output of N, K, and P by steers grazing high-quality pasture. They observed that urinary excretion of N and K vaned between 81 and 137 g N day-' and 58 and 90 g K day-', respectively. Urinary N and K concentrations were higher at night than during the day. Similarly, fecal nutrient content varied greatly, i-e., 3662 g N day-', 12-46 g K day-', and 10-23 g P day-'. Notwithstanding the great variability in the amounts of nutrients excreted in dung and urine by individual animals, the general trends in the amounts and forms excreted are discussed below.

B. FORMOFRETURN 1. Nitrogen

The proportion of total N intake that is excreted and its partition between urine and feces is dependent on the type of animal, the intake of dry matter, and the N concentration of the diet (Whitehead, 1970, 1986). For sheep and cattle, fecal excretion of N is usually about 0.8 g N 100 g-I of dry matter consumed, regardless of the N content of the feed (Barrow and Lambourne, 1962; Barrow, 1987). As shown in Fig. 4, as the N concentration of sheep fodder was increased from 1.0 to 5.0%, the N excreted as dung remained constant. Similarly for dairy cows, Lantinga et al. (1987) found that feces contained an average of 132 g N cow-' day-' irrespective of whether their level of intake was 450 or 775 g N cow-' day-'. The remainder of the N is excreted in the urine and as the N content of the diet increases, so too does the proportion of N in the urine (Fig. 4). Barrow and Lambourne (1962) found that for sheep ingesting herbage containing more than 4% N, 80% of the N was excreted as urine, whereas with herbage containing 0.8% N the proportion of the excreted N present in the urine was only 43%. In most intensive high-producing pasture systems, where animal intake of N is high, more than half of the N is excreted as urine. For instance, data for Netherlands dairy cows showed that approximately 60-65% of excreted N was present in urine (Lantinga et al. 1987; Van Vuuren and Meijs, 1987), whereas for sheep grazing grass/clover pastures in New Zealand, 70-75% of the excreted N occurred in the urine (Sears et al., 1948; Sears, 1950). The average N content of feces is 2.0-2.896 N on a dry matter basis (Floate and Torrance, 1970; Whitehead, 1970). There are small amounts

NUTRIENT CYCLING UNDER GRAZED PASTURE

133

U

Q

CI v)

D

En

A

Q

.-r

s

2

I

o

1

Total excretion Fecal excretion

2

3

4

5

Nitrogen content of feed (%) Figure 4. Effect of fodder nitrogen concentration on the excretion of N by sheep. Total excretion (feces plus urine) and excretion in feces alone are shown. (Redrawn from Barrow and Lambourne, 1962.)

of mineral N in feces but the bulk of the N is in organic form. About 20-25% of fecal N is water soluble, representing the metabolic products of the animal and microbial population in the gut; about 15-25% is undigested dietary N and the remaining 50-65% is present in bacterial cells (Mason, 1969, 1979; Mason et al., 1981). Partial chemical fractionation of fecal N from cattle by Van Faassen and Van Dijk (1987) showed an ammonium content of 1 - 10’30, an a-amino N content of 20-35%, and an amide N content of 10- 15%. Much of the remaining 40-70% is presumably present mainly as hexosamines (amino sugars), because these are major components of the cell walls of bacteria (Stevenson, 1982). The concentration of N in urine varies widely because of factors such as N content in the diet and consumption of water, but is normally in the range of 8 - 15 g N liter-’ (Whitehead, 1970). As the digestible N intake increases, so too does the proportion of urine N present as urea (Topps and Elliott, 1967). Typically over 70% of the N in urine is present as urea and the rest consists of amino acids and peptides (Doak, 1952; Bathurst, 1952). There is little mineralization of dietary organic N during passage through the animal because the majority of N in both dung and urine is in organic form. Thus, about 99% of dietary N is in organic form (Floate, I970a) and about the same percentage of excreted N is in organic form. However, after

134

R. J. HAYNES AND P. H. WILLIAMS

passage through the animal, much of the organic N is in a more rapidly mineralizable N form. This is because over 60% of excreted N is usually in the form of urine and 70 - 90% of this is in urea N form. Thus, typically at least 48% of excreted N is present as urea and this is very rapidly hydrolyzed to NHt N in the urine patch (see Section IV,C,3). 2.

sulfur

Sulfur is excreted in significant proportions in both urine and feces. Barrow and Lambourne ( 1962) observed that the proportion of S excreted in the urine varied from 90% for pasture herbage of high S content (0.5% S) to 6% for herbage with low S content (0.1% S). Kennedy and Till (1981) calculated that for Merino wethers grazing a grass/legume pasture, about 5390 of the excreted S would be in the form of urine. Similarly, our unpublished results have indicated that for sheep grazing typical New Zealand pastures (0.25-0.30% S), 50-70% of excretal S is voided as urine. Although Barrow and Lambourne (1962) observed that fecal excretion of S remains constant at about 0.11 g S 100 g-I of dry matter consumed, other workers have found that fecal S content increases with increasing S intake for both sheep and cattle (Bray and Hemsley, 1969; Bird, 1971; Langlands et al., 1973; Kennedy, 1974). Most of the S excreted in feces is in organic forms (Bird, 1971; Bird and Hume, 1971). Bird (197 1) observed that after infusion of 35Sinto the rumen of sheep, 87 - 94% of fecal S was C bonded, 4-5.4% was in ester sulfate form, and only 0.5 -4.0% was present as inorganic sulfate. The C-bonded fraction consists mainly of bacterial protein (Bird, 1971; Langlands et al., 1973). Sulfur is present in urine in both organic and inorganic forms. The amount of S excreted in urine is strongly related to S intake (Bird, 1971; Kennedy, 1974). The relative proportion of the different S forms in the urine is also dependent on the S status of the diet, and the inorganic sulfate content of urine increases markedly as S intake increases (Bird, 1972; Kennedy, 1974). Thus, Bird and Hume (1971) reported that the urine excreted by sheep fed a diet supplemented with cysteine contained 84% sulfate whereas the urine of the control animals on an unsupplemented diet contained only 14%sulfate. Under intensively grazed farming systems, sulfate S often represents 50-60% of urine S. The organic S content of urine is made up of roughly similar proportions of ester sulfate and Cbonded S (Bird and Hume, 1971). It appears that a significant amount of mineralization of dietary S occurs during passage through the animal (Till, 1981). Sulfate S makes up a small but variable proportion of the total S content of herbage (1 -20%) (Blanchar, 1986).If an animal excreted 60% of its S in the form of urine and 60%

NUTRIENT CYCLING UNDER GRAZED PASTURE

135

of that were in sulfate form, then 36% of ingested S would be returned to the pasture in readily available sulfate S form. However, there are few if any direct data to show the exact extent of animal-induced S mineralization. 3. Phosphorus

Fecal P represents the predominant pathway for animal returns of P to grazed pasture. Only trace amounts of P are normally detected in the urine of ruminants, although the amount of P increases slightly when P intake increases (Braithwaite, 1976) and there can be considerable variation between individual animals (Grace, 1983). Total fecal P content is strongly correlated with total P intake (Bromfield and Jones, 1970; Blair et al., 1977) and thus P content of the diet. Rowarth et al. (1988) found a highly significant linear relationship between P concentration in pasture on offer for grazing and P concentration of the feces deposited subsequently during a grazing period (Fig. 5). The major form of inorganic P present in feces is dicalcium phosphate (Barrow, 1975). Gerritse ( I 978) identified inositol hexaphosphate and adenosine triphosphate as the main organic phosphates in pig feces. The proportion of inorganic P in feces increases as total P intake increases (Bromfield, 1961; Barrow and Lambourne, 1962), whereas organic P content remains relatively unchanged. Thus, Rowarth ( 1987) found that fecal

1.6

1.2

0.8

-am

/

0.4

0

a

U

01 0

1

I

0.1

0.2

I

0.3

I

0.4

Pasture P concentration

I

0.5

1

0.6

(%I

Figure 5. Relationshipbetween pasture P concentration and the fecal P concentrationof sheep grazing the pasture. (Redrawn from Rowarth et al., 1988.)

136

R. J. HAYNES AND P. H. WILLIAMS

inorganic P contents increased with increasing rates of P fertilizer (and increasing herbage P concentrations)and showed a similar seasonal pattern to pasture P concentrations, whereas organic fecal P concentrations were little affected by either fertilizer rate or season. Dung contains a higher content of both organic and inorganic P than does ingested pasture (Floate and Torrance, 1970; Rowarth et al., 1988), although there can be considerable conversion of plant organic P to inorganic P during passage through the animal (Bromfield and Jones, 1970; Floate, 1970a,b). Thus, Floate and Torrance (1970) found that feces contained 80% inorganic P but ingested plant material contained only 64% inorganic P. Bromfield and Jones (1970) observed that in the spring, when P content and digestibility of pasture were high, up to 80% of the ingested organic P was mineralized via passage through the animal. They found that by summer, when P content and digestibility had decreased, mineralization became slight. 4. Other Nutrients

Potassium is mostly excreted as urine, with only 10- 30%being excreted in feces (Barrow, 1987). The K content of urine can vary widely (Williams et al., 1990a) but is usually in the range 6- 1 1 g K liter-' (Hutton et al., 1967; Ledgard et al., 1982; Williams et al., 1989). The K in urine and dung is in ionic form and is therefore readily plant available. Davies et al. ( 1962), for example, reported that virtually all the K in cattle dung was water soluble and therefore readily available. Potassium commonly represents 60- 70% of the equivalent cation content of urine. The major balancing anions in urine are C1- and HCO;, which often represent 20- 50% and 40-70%, respectively of the equivalent sum of anions. Urine is also the main form in which B and I are excreted (Barry, 1983; Barrow, 1987). With diets high in S, Mo is also readily excreted in urine, but when sheep are fed diets low in S, Mo is mostly excreted in small amounts in the feces (Barrow, 1987). The feces are the main excretory pathway for Ca and Mg; the concentration of these ions in urine is usually less than 1 g liter-'. The Ca and Mg content of feces is usually in the range 1.2-2.5% and 0.3-0.896, respectively (Hutton et al., 1965, 1967; Weeda, 1977; H o g , 1981). The high concentrations of fecal Ca and Mg result in an excess content of nutrient cations over anions in dung. This imbalance is made up by carbonate (Barrow, 1987). Barrow (1975) showed that Merino sheep feces had a CaCO, content of about 1.3%, and it seems probable that much of the Mg is also present as magnesium carbonate. Certainly, a significant proportion of the total Mg and Ca content of feces is not readily soluble in water.

NUTRIENT CYCLING UNDER GRAZED PASTURE

137

Davies et al. (1962), for example, observed that only 62% of Mg in dairy cow dung was water soluble whereas other workers (R. J. Haynes and P. H. Williams, unpublished observations, 1989) found that 52% of Mg and 32% of Ca in cow dung was readily water soluble. As a result of the presence of calcium carbonate, animal feces usually have a pH in the range 7.0-8.0 (Underhay and Dickinson, 1978; Omaliko, 1984). Feces represent the major pathway for the excretion of many trace elements (e.g., Cu, Zn, Fe, Mn, Co, and Se) (Underwood, 1981; Safley et al., 1984; Barrow, 1987) and heavy metals such as Cd (Smith, 1984) and Pb (Quarterman, 1986). Most of such excretion consists of unabsorbed dietary elements.

C . PATTERN OF RETURN The amount and availability of nutrients returned to pasture in dung and urine are influenced not only by the amount and forms of nutrients in excreta but also the number of excretions per day, the size of each excretion, and the surface area covered by the excreta. In addition, the pattern of excretal return is important because the more even the pattern, the more efficiently the nutrients are likely to be recycled within the pasture system. 1. Number, Size, and Area Covered by Excretions

The reported urination and defecation events per 24-hr period for cattle and sheep are shown in Table IV. For cattle, a range of 8 - 12 urinations and 1 1 - 16 defecations per day is common. Data for sheep are limited, but a range of 18-20 urinations and 7-26 defecations per day has been reported. The number of defecations and urinations per day can be greatly influenced by grazing conditions and environmental factors. Barrow ( 1967), for instance, observed that cattle grazing in rangeland conditions defecated less than half as often per day as cattle in intensive conditions. The number of urinations per day can be greatly influenced by water intake and therefore by the water content of the herbage (Doak, 1952), the season of the year (MacDiarmid and Watkin, 1972b),and the weather (Betteridge et al., 1986). The reported mean volume of a single urination and weight of a single defecation for cattle and sheep is shown in Table IV. Each urination event by cattle and sheep has a mean volume of 1.6-2.2 liters and 0.10-0.18 liters, respectively. The mean weight per defecation is 1.5-2.7 kg for cattle and 0.03 - 0.17 kg for sheep.

Table IV

Number and Weight or Volume of Defecstions and Urinations per D a y and Surface Area Covered by Excreta Produced by Cattle and Sheep

Reference

E O0

Johnstone-Wallaceand Kennedy ( 1944) Castle et al. ( 1950) Hancock (1950) Goodall(1951) Waite ef af. (1951) Doak(1952) Hardison et al. (1956) Petersen et al. (1956) MacLusky (1960) Davies et al. (1962) Wardrop ( 1963) H o g (1968) Weeda ( 1967) Frame (1971) MacDiarmid and Watkin (1972b)

Stock type

Mean number of defecations per day

Beef COW

11.8

Daily COW Dairy COW Dairy COW

11.6 12.2 12

Daily COW Dairy COW

-

Dairy COW Dairy COW Dairy COW Dairy COW

15.4 12 11.6 12 16.1

Dairy COW Dairy COW Beef steer Dairy COW Dairy cow

Weight of single Area covered defecation by defecation (kg wet weight) (m2) 1.77

8.5 9.8

I .48 2.27 -

10.5

-

I1 13.9

2.7 1.82

-

0.06

Mean number of urinations per day

0.09 0.05 0.07

-

0.07

10.1 11

-

9.4 8

10

12.1 -

11

-

Volume of single urination (liters)

Area covered by urination (m2)

w \o

Robertson (1972) During and Weeda (1973) Richards and Wolton (1976) Weeda (1979) Williams et al. ( 1990b) Sears and Newbold (1942) Sears(1951) Doak (1952) Raymond and Minson (1955) Bromfield ( 1961) Hemott and Wells (1963) Frame(l971) Robertson ( I 972) Morton (1984) Morton and Baird ( 1990) J. S. Rowarth (personal communication, 1990) P. H. Williams and R. J. Haynes (unpublished observations, 1990)

Dairy cow Beef steer Dairy cow Beef steers Dairy cow Sheep Sheep Sheep Sheep Sheep Sheep Sheep Sheep Sheep Sheep Sheep

Sheep

Data calculated based on 20 defecations or urinations per day.

_.

-

0.07"

-

0.09" 0.06" 0.075" 0.17

-

0.03 -

0.05 0.05

-

0.008 0.0 12 0.025 -

0.02

140

R. J. HAYNES A N D P. H. WILLIAMS

Such mean values are subject to considerable variation. For example, urinations and defecations by individual cattle have been measured at 0.85-2.85 liters (Doak, 1952) and 0.45-6.79 kg (Goodall, 1951), respectively. Many factors can contribute to this variation in excretal output (Barrow, 1967). The volume of urine excreted is strongly correlated with the amount of water absorbed (Paquay et al., 1970a,b), and on hot days water intake and urinary volumes are much greater than on mild, overcast days (Betteridge et al., 1986). As a result, seasonal differences in urinary volumes can occur (Vercoe, 1962). The quantity of feces produced is greatly influenced by the quantities of feed ingested (Hutton and Jury, 1964) and consequently factors affecting feed intake will also affect fecal output. The surface area reported to be covered by a single urination event is 0.16-0.49 m2 for cattle and 0.03-0.05 mz for sheep (Table IV). The variation in area is partly due to variations in the volumes excreted and may be also partially due to difficulties in measuring the wetted area accurately. The most accurate measurements are those that involve the use of tracers that can be observed or chemically measured, for example, chalk (During and McNaught, 196l), fluorescent dye (Morton and Baird, 1990), or bromide (Williams et al., 1990b). Other factors causing variations in the surface area covered by a urination include wind and slope (During and McNaught, 1961), soil moisture content, and other soil physical conditions, such as water-repellent or compacted soil surfaces (Williams et al., 1990b). The surface area covered by individual cattle and sheep dung patches ranges from 0.05 to 0.09 mz and 0.008 to 0.025 m2 (Table IV). Although the range of measurements for cattle is small, that for sheep is much greater. This reflects the difficulty of defining the area covered by sheep dung because it is often in the form of pellets, which can be scattered over a relatively wide area (Morton and Baird, 1990). 2. Distribution of Excretal Returns

The pattern in which nutrients are returned to the pasture in the form of dung and urine is nonuniform. The pattern of return is greatly influenced by stock behavior (e.g., camping of stock in small areas of the field) and stock management (e.g., having separate fields for daytime and nighttime grazing). Excreta can also be deposited in nonproductive parts of the farm, such as raceways and stock-handling sheds. Animals generally deposit more excreta on areas where they congregate (stock camps)-beneath trees and hedges, around gateways and water troughs, on areas away from roadsides, and on ridges and hillcrests on hill

NUTRIENT CYCLING UNDER GRAZED PASTURE

141

country farms (Hilder, 1966; Gillingham and During, 1973; Hakamata, 1980). For example, fertility transfer was studied by Hilder (1966) using flat land stocked with Merino sheep. Because of stock camping, about one-third of the total dung deposited was found on less than 5% of the total area of the field. The distribution of urine appeared to follow a similar trend. On dairy farms during winter, there can be a concentration of excreta in areas of the paddock where hay or silage are fed out (MacDiarmid and Watkin, 1972b). On hill country pastures the stock tend to camp on flat areas of land and significant quantities of nutrients are transported to these areas from the steeper slopes where the sheep graze (Gillingham and During, 1973; Gillingham et d.,1980; Saggar et d.,1988; Rowarth and Gillingham, 1990). Measurements on hill country lands have shown that 60% of dung and 55% of urine are deposited in campsite areas, which occupy only 15 - 3 1% of the total land area (Saggar et al., 1988). As a result of this uneven pattern, there is a buildup of nutrients in the campsite areas and a depletion in nutrient status on the remainder of the pasture. Nutrient balance studies on hill country pastures have shown that the annual net accumulation of nutrients on campsite areas can be about 2 10 kg N ha-', 200 kg K ha-l, 30 kg P ha-l, and 15 kg S ha-' (Gillingham and During, 1973; Gillingham et al., 1980; Saggar et al., 1988; Rowarth and Gillingham, 1990). In the New 'Zealand hill country, pasture improvement by topdressing with superphosphate is slow and withdrawal of fertilizer can lead to a rapid decline in production (Williams and Haynes, 1990a). The major reason for this is thought to be the continual transfer by grazing stock of P away from the main grazing slopes to the relatively small stock camp areas (Gil1980; Rowarth and Gillingham, 1990). As stocking rate lingham et d., increases there is less tendency for animals to camp and consequently there is a more even distribution of excreta over the paddock. This leads to reduced transfer losses of nutrients and more efficient cycling of nutrients within the system. Thus, increasing the stocking rate through subdivision of paddocks and the use of rotational grazing rather than set stocking can reduce the camping effects. On dairy farms where separate paddocks are grazed during the day and night periods, a transfer of fertility has often been observed from day to night paddocks (Sears, 1950, 1956; Hancock and McArthur, 1951). On grass/clover pastures the night paddocks can become grass dominant through high N returns and conversely the day paddocks become more clover dominant because of lowering of the soil N status (Haynes, 1981). Some workers have reported that a greater proportion of dung and urine is excreted in the nighttime compared with daytime (Castle et al., 1950;

142

R. J. HAYNES AND P. H. WILLIAMS

Waite et al., 1951; Hardison et al., 1956), but others have found no difference in the amount excreted in the day and night (Goodall, 1951; Hancock and McArthur, 1951). It is likely that the major reason for more excreta being deposited on night paddocks is that cows are held in night paddocks for a longer period than in those used during the day (Hancock, 1950). In addition, night paddocks tend to be smaller than day ones and they tend to be close to milking sheds and grazed more often than those used during the day. A proportion of excreta is deposited when the animals are not on the pasture but are on nonproductive areas of the farm, such as feeding platforms, stock-handlingsheds, milking sheds, yards, and raceways. For farms that are not equipped to redistribute the effluent back onto the pasture, the nutrients in the excreta are lost from the pastoral area. On dairy farms where cows are taken off the pasture twice a day for milking, the amounts of nutrients lost in this way have been estimated as 2 - 1 1 kg N, 4 - 14 kg K, 0.5-3 kg P, and 1-2 kg S cow-' yr-l (Goold, 1980; Clarke and Warburton, 1982; Williams et al., 1990d). 3. Models of Excretal Distribution

Several workers have described the distribution of animal excreta in the field using mathematical models. Petersen et al. (1956) tested several relationships to describe the distribution of feces on a pasture stocked with free-grazing cattle. The negative binomial distribution function was in close agreement with measurements of the distribution of dung pats. The negative binomial distribution allows for the fact that an area may be covered more than once by excreta (Le., overlapping of patches) and that there will be an increased amount of dung deposited in areas of special attention (e.g., stock camp areas). It has been used successfully by a number of workers (Donald and Leslie, 1969; Richards and Wolton, 1976; Morton and Baird, 1990).The Poisson distribution has also been used with success (Hakamata, 1980, 1985), but Petersen et al. (1956) found the negative binomial distribution described data better because the Poisson distribution allowed for overlapping of excreta but not for stock camping activity. In the negative binomial distribution the proportion of total area covered by excreta (P) is described as follows:

P = 100 - 100/[(0 + K ) / K ] k where D is the mean density of excreta deposited on a field during the time period calculated from the number and size of individual excretions and the constant K is a measure of excretal patch aggregation, which varies

NUTRIENT CYCLING UNDER GRAZED PASTURE

143

principally according to stocking density. Morton and Baird (1 990) found that for sheep at a low stock density [900 and 300 stock units (su) ha-' in winter and spring, respectively] the measured K value was 2.9, but at high stock density (1 800 and 600 su ha-' in winter and spring, respectively) the K value was 5.8. The higher K value indicates a more even distribution of excreta throughout the field due to diminished opportunity for camping at higher stock densities. Using the negative binomial distribution, it has been calculated that on a typical New Zealand dairy farm with a stocking rate of three cows ha-' yr-', 23% of the pasture would be covered in excreta in 1 year (Williams, 1988). The pasture area influenced by urine is often more than twice the area actually wetted whereas that influenced by dung pats can vary from one to six times the area covered (see Sections IV,B and IV,D). Thus the area of pasture affected by excreta may be at least twice that covered (i.e., greater than 46% of the pasture area in the above example).

D. ROLEOF EXCRETAIN NUTRIENT CYCLING The fate of pasture nutrients ingested by grazing animals under two contrasting farming systems is shown in Table V in order to demonstrate the significance of the animal to nutrient cycling. The two farming systems are an intensive, rotationally grazed dairy farm and an intensive, openrange sheep farm on hill country. Pasture production and stocking intensity are much greater on the dairy farm than on the sheep farm and consequently the amounts of herbage and nutrients ingested by the animals and the amounts excreted are also greater under the dairy system. Because of the daily removal of milk from the dairy farm, the total quantity of nutrients lost from the farm is much greater than that lost from the sheep farm in wool and meat. On the hill country sheep farm there is ample opportunity for animals to camp on hill tops and other flat areas and therefore transfer nutrients away from the main grazing areas. As a consequence, transfer losses of nutrients are about three times those under the intensive dairy system, despite the cows spending about 4 hr per day away from the grazing land (in the raceways and milking shed). Some N is lost directly from the urine patch areas through ammonia volatilization (see Section IV,C,3). On the dairy farm, the cows produce large volumes of urine, which are added to small volumes of soil. As a consequence, significant amounts of urine move below a depth of 1520 cm by preferential flow down through soil macropores (Section IV, C, 2). The amount lost in this way from sheep urinations is, as yet, unknown.

R. J. HAYNES AND P. H. WILLIAMS

144

Table V Amounts of Nutrients Ingested and Excreted and Amounts Lost Directly through Actions of Grazing Animals on Two Farm Types' Lassescaused directly by animal activity Excreted

armt type^ Intensive dairy

Hillcountry

Animal Nutrient Ingested Dung Urine products

Transfer Volatilization Preferential of from loss from excreta excreta urine

N

450

102

277

71

K P S

360 45 34

38 34 14

308

-

14

38 35

11

3

15

5

3

N K P S

264 200 24 24

66 22 22 10

178 177

20 1

112 91

2 3

10

11

-

10

20

13

-

-

50

55

3 ? ?

?

'Data calculated for a typical New Zealand dairy farm (stocking rate = 3 cows per hectare, production = 450 kg milkfat ha-' yr-I) and a typical hill country sheep (stocking rate = 14 su ha-', production = 63 kg wool and 180 kg meat ha-' TI). Data sources include Gillingham (1978), Gillingham and During (1973), Hemell and Ross (1973), Hutton et al. (1967), Metson and Saunders (1 978), Rowarth ( I 987), Saggar et al. ( 1988), Wilkinson and Lowrey (1 973), Williams ( 19881, and Williams et d.( 1990b,d).

IV. SOIL PROCESSES AND PASTURE RESPONSE IN EXCRETA-AFFECTED AREAS A. RELEASE OF NUTRIENTS FROM FECES 1. Feces Composition

Feces consist of water, residues of undigested herbage, products of animal metabolism, and a large and vaned population of microorganisms and products of their metabolism. The typical composition of animal feces is shown in Table VI. The high (47 -68%) fiber content (cellulose, lignin, and hemicellulose) reflects the low digestibility of the diet of ruminant animals. The content of structural carbohydrates in dung is inversely related to the digestibility of the ingested pasture (Marsh and Campling, 1970). The 40 - 50% detergent-soluble dry matter consists predominantly of living and dead bacterial cells originating from the gut of the animal plus some

145

NUTRIENT CYCLING UNDER GRAZED PASTURE Table VI Typical Chemical Composition of Feces from Major Types of Farm Animals'

Feces source

Neutral detergent soluble

Nitrogen

Hemicellulose

Cellulose

Lignin

Ash

Dairy cattle (lactating) Cattle (fattening) Sheep (forage fed)

41 53 45

2.0 3.0 2.5

20 22 15

28 17 28

20 8 15

12 7 13

~~

~

~

~

~

Data from Smith (1973). Values given as percentage of dry matter.

water-soluble metabolic products of both endogenous and microbial origin (Mason et al., 198 1 ). The consistency of feces varies greatly with diet and is affected mainly by the water and structural carbohydrate content of the herbage. Lush spring grass, which has a high moisture and low structural carbohydrate content (and high digestibility), can result in rather liquid feces. During and Weeda ( I973), for instance, observed that in spring and autumn, periods of highest pasture production, cattle dung was liquid and covered much greater surface areas than in the summer, when feces were firm.The ash content of cattle feces can be in the range of 20-40% (Healy, 1968; Underhay and Dickinson, 1978) and a considerable portion of this can be silica arising from ingestion of soil (Healy, 1968). The nutrient content of feces was discussed in detail in Sections II1,A and II1,B. The total nutrient content of cattle feces is well documented, but that of sheep is less well known. However, observations (R. J. Haynes and P. H. Williams, unpublished observations, 1989) have indicated that for sheep and cattle grazing the same pasture, nutrient concentrations in feces are very similar. Typical application rates of the major nutrients applied to the dung patch of sheep and cattle are shown in Table VII along with the annual output per cattle beast and per hectare of farm. It is evident that very large amounts of organic matter and N, Ca, Mg, P, and K are deposited within the dung patch area. 2. Feces Degradation

Two major processes contribute to dung degradation and thus the release of nutrients. These are (1) physical breakdown, which is caused mainly by raindrop impact (and treading) and (2) biological degradation,

146

R. J. HAYNES AND P. H. WILLIAMS Table VII

Typical Application Rates of Major Nutrients Applied to Sheep and Cattle Dung Patches and Annual Output per Cattle Beast

Typical dung concentration

Application rate per dung patch (kg ha-1)"

Parameter

(%I

Sheep

Cattle

Organic matter N P S K Ca

80 2.6 0.70 0.25 1 .o 2.0 0.66

4000 130 35 13 50

32000 1040 280 100

100

800 264

Mg

33

400

Annual Annual output output per from cattle per cattle beastb hectare of farm' (kg animal-' yr') (kg ha-I yrl) 699 23 6.1 2.1 8.8 17 5.8

1748 57 15

5.3 22 43 15

a Assuming that for sheep and cattle a dung patch consists of 0.01 and 0.20 kg (dry weight) of dung and the area covered is 0.02 and 0.05 m2, respectively. Assuming cattle defecate 12 times per day. Assuming 2.5 cattle per hectare.

which is brought about by biota such as fungi, bacteria, beetles, and earthworms. The initial release of nutrients is very dependent on factors affecting the physical breakdown of dung deposits (Underhay and Dickinson, 1978; Rowarth et af., 1985). a. Physical Breakdown Physical breakdown is influenced greatly by climate and the initial consistency of the feces (Weeda, 1967). When dry weather follows dung deposition, a hard crust forms on the dung patch (Weeda, 1967; Underhay and Dickinson, 1978). The formation of such a crust partially protects the pat from the eroding effect of raindrop impact and it also inhibits rain from penetrating and rewetting the pat. The weather immediately after dung deposition is therefore a major factor affecting degradation. As a consequence, it has often been noted that in temperate climates the disappearance of dung is less rapid during summer (when a crust forms) than in winter (Weeda, 1967; MacDiarmid and Watkin, 1972b; Rowarth et al,, 1985). Under temperate conditions (annual rainfall of 1500 mm), Rowarth et af. (1985) observed a rapid exponential breakdown of fresh sheep feces (Fig. 6). Fecal samples had completely degraded within 17 days in winter but lasted over 100 days in summer. By contrast, research under dryland conditions (300- 550 mm annual rainfall) using air-dried fecal material

NUTRIENT CYCLING UNDER GRAZED PASTURE

= m

.-

100

8

r

0.

C

.=

80

0

*-\*

\

60-

.-Em 3

2

147

40

\.

-

Winter

Time after dung deposition (days) Figure 6. Relationship between the decrease in the dry weight of sheep dung pads and time under dry summer and wet winter conditions. (Redrawn from Rowarth et al., 1985.)

(Bromfield and Jones, 1970; Rixon and Zorin, 1978) has shown that decomposition is slow and occurs at a more or less linear rate. Bromfield and Jones (1970) found a 40% weight loss from fecal samples under field conditions over a 2-year period. The consistency of the excreta is another important factor influencing degradation rate. Under moist conditions (e.g., winter) the more liquid dung patches disappear more rapidly than the firmer ones (Weeda, 1967). In contrast, under drying conditions, during which a hard crust forms, initial consistency is unimportant. These effects are shown in Table VIII; in summer and spring, initial consistency had no effect on degradation whereas in autumn and winter the liquid dung broke down more rapidly. An additional interacting factor influencing the rate of degradation can be the physical form of fecal material. Sheep, for example, can excrete dung in the form of firm pellets or larger more liquid pads; the former have a much greater surface area per unit weight of material excreted. The larger surface area makes pellets more susceptible to raindrop impact and physical degradation and as a consequence pellets normally degrade more rapidly than pads (Rowarth et al., 1985). b. Biological Degradation Microbial decomposition of dung material is essential in order to release most of the fecal N and S that are present in organic combination. This

R. J. HAYNES AND P. H. WILLIAMS

148

Table VIII Effect of Dung Consistency and Rainfall on the Number of Months Required for the Major Part of the Dung to Disappeaf ~

~~

~~

~~~~

Rainfall means after Season in which dung was deposited Summer Autumn Winter Spring

deposition of dung (mm)

Dung consistencyb 2

3

3.5 1.5

3 3

2 5

6.5

3.5 4 6 6.5

1

6

5

lweek

2weeks

3.5 4.5

8.5

-

3.8 0.3

7

6.5

1.8 1.o

8.4 2.0 2.3 4.3

4 4 4

Data from Weeda (1967). Consistency scale: 1, very liquid; 5 , very firm.

may occur during the degradation of the dung pat and/or after the pat has been dispersed and dung particles have been washed into the surface soil by rainfall. Coprophagous invertebrates, notably dung beetles, dipterous larvae, and earthworms, play an important role in dung deposition through promoting aeration and microbial activity and incorporating dung into the soil (Curry, 1987a,b). In some localities, dung beetles (especially Geotups spp. and Aphodius spp.) remove large quantities of dung and permeate the pats with burrows, thus increasing aeration (Bornemissza, 1970; Holter, 1979). In the African savanna, termites are the main agents of degradation in the dry season (Omaliko, 1981). Dung flies and other macroarthropods are commonly active in burrowing and in the physical degradation of pats. Earthworms are also active in burrowing, ingestion, and mixing of dung with the soil. Indeed, dung is actively sought out by lumbricoid earthworms and the presence of dung enhances growth and reproduction of earthworms in agricultural soils (Watkin and Wheeler, 1966; Lofs-Holmin, 1983). Holter (1979) found that earthworms accounted for 50% of the &sap pearance of cattle dung pats from the soil surface in Denmark and beetle larvae accounted for another 14-20%. In some localities the absence of coprophagous fauna adapted to cope with large quantities of animal dung can result in the accumulation of dung at the soil surface and sward deterioration on intensively grazed pastures (Gillard, 1967). Rainfall is important for biological degradation because it can maintain dung moisture levels suitable for microbial activity. Thus, the formation of

NUTRIENT CYCLING UNDER GRAZED PASTURE

149

a dry crust over the pat, which inhibits rewetting, not only hinders physical degradation but also inhibits microbial decomposition. Temperature and humidity also influence the activity of microbial populations in dung. Losses of CO, and NH, to the atmosphere are most rapid in the first few weeks after dung deposition (MacDiarmid and Watkin, 1972a; Anderson and Coe, 1974). The loss of NH, occurs through volatilization, whereas the loss of CO, is mainly due to microbial respiration. Underhay and Dickinson (1978) observed that despite the presence of a vigourous coprophilous microflora, loss of organic matter from cattle dung was only about 15% over a 2-month period and the calorific value of the dung decreased by 18%over the same period. This slow decomposition has been attributed to the fact that fungal activity is initially confined to the surface layers of the dung pats (Dickinson and Underhay, 1977). The cattle dung pat has a nonporous matrix and the initial moisture content is often 400-700% of the dry matter content. Hence O2may well become limiting in the center of the pat and this restricts the spread of hyphae into the center. Loss of organic matter from dung pats, through CO, evolution, can occur more rapidly than the loss of nutrients. Thus, there can be an increase in the concentration of elements such as Ca, Mg, Fe, P, and N in dung during its decomposition (Dickinson and Underhay, 1977; Omaliko, 1984). The processes involved in the release of the major nutrients from feces are outlined below. 3. Release of Nutrients

a. Nitrogen Because the bulk of the N in feces is in organic forms it must first undergo microbial mineralization before it is released as mineral forms. The amount of N mineralized from feces is closely related to the total N content of feces, but N mineralization is slower from feces than that from the plant materials they were derived from (Barrow, 1961;Floate, 1970a,b) (Fig. 7). The slower mineralization is not due to a difference in C :N ratios because Floate (1 970a) found that the C: N ratio of sheep dung (22 : 1 to 27 : 1) was similar to that of the ingested herbage. However, a large proportion of the C content of feces consists of undigested fibrous material (cellulose, hemicellulose, and lignin), which degrades only slowly. The slow degradation of fecal material apparently results in a slow release of other nutrients present in organic form (N and S) (Barrow, 1960). Despite this, under dryland conditions Rixon and Zorin ( 1978) observed that retention of N in feces was proportionally greater than the retention of the bulk of the fecal material, with the result that increases in the total N concentration in sheep feces were observed during decomposition. Studying the

1so

R. J. HAYNES AND P. H. WILLIAMS

Y

E

-aE

Y

Plant material

0

I

m

0

Feces

0

-

*

I

l

l

I

I

1

1

2

3

6

9

12

Incubation period (weeks)

Figure 7. Net total mineral N production as a percentage of original total N from Agrostis-Festucu plant material and sheep feces incubated for 12 weeks at 30'C. (Redrawn from Floate, 1970b, with permission from Pergamon Press PLC.)

decomposition of cattle dung, Underhay and Dickinson (1978) observed a decrease in N concentrations during the first 35 days (indicating preferential loss of N), but this was followed by an increase in N concentrations during a subsequent 35-day period of decomposition. As noted previously (Section IV,A,2), NH, is lost from decomposing dung, particularly during the first week after deposition (MacDiarmid and Watkin, 1972a). Over the first 13 days of decomposition of cattle dung, MacDiarmid and Watkin (1972a) measured a loss of 4.7% of the dung N and Ryden et al. (1987) measured losses of 1.2 and 12.0%,respectively for cattle and sheep dung over a 2-week period following deposition. The balance between C and N mineralization in feces is likely to be greatly influenced by environmental factors as well as the amount and composition of the C and N content of feces. Floate (1970c), for example, observed that as the incubation temperature was reduced from 30" to 10°C and then 5"C, the losses of C as CO, from sheep feces over a 12-week period were 16,4, and 2%, respectively. In contrast, the largest amounts of mineral N were produced at 10°C. Floate (1970d) found that mineralization of N was greatest at 100%moisture holding capacity (MHC) and was almost completely inhibited at 25% MHC. However CO, production was greater at 25% MHC than at either 50 or 100%MHC. The release of mineral N from feces results in elevated concentrations of mineral N in the soil below the dung patch. The high concentrations of

NUTRIENT CYCLING UNDER GRAZED PASTURE

151

NO; that can accumulate (e.g., 90 to 130 bg N g-') (Ryden, 1986) suggest that the dung patch (like the urine patch) can be a significant source of both NOT leaching and gaseous losses of N,O and N, (through denitrification/nitrification) from grazed pastures (see Section IV,C,3). b. Sulfur The release of S from animal excreta has received little attention. Barrow (1 96 1) reported that the amount of S mineralized from fecal pellets was closely related to the initial S content of the feces. Mineralization of feces released proportionately less S than did plant litter of the same S content (Barrow, 196 1 ; Boswell, 1983). Boswell ( 1 983) observed two distinct processes in the release of S from sheep feces. An initial rapid process in which soluble and more labile S was released was followed by a protracted slower process in which more resistant material was mineralized. During the initial period the loss of fecal S is more rapid than the loss of dry matter (Kennedy and Till, 198 1 ; Boswell, 1983), suggesting preferential mineralization of S and/or leaching of S from fecal pellets. The half-life of fecal S was found by Boswell(l983) to be 1 54 days under controlled environmental conditions with adequate moisture. Fecal S does not appear to be rapidly available to pasture plants. Under controlled environmental conditions Boswell ( 1983) found that ryegrass plants recovered about 5% of the applied 35Slabel from sheep feces. Over a 12-month period under field conditions, 16% of the radioactivity in applied 35S-labeledfeces was recovered by Kennedy and Till ( 198 1) in aboveground pasture herbage. c. Phosphorus Under dryland conditions, the availability of P from dung is initially the consequence of leaching of water-soluble inorganic P (Bromfield, 196 1 ; Rixon and Zorin, 1978). Rixon and Zorin (1978) measured a 50%decrease in P concentration in fecal samples placed under bushes in the saltbush rangeland for 20 months. An 80%decrease in P concentration was measured over the same time period for fecal samples placed on irrigated pasture. Under temperate conditions, wherein physical breakdown of dung is rapid, leaching of P from fecal material is relatively unimportant. Rowarth et al. (1989, for instance, observed that the concentration of both total P and water-extractable P from dung samples remained relatively constant with time and that the major mechanism controlling movement of P from feces into the soil was the rate of physical breakdown. The availability of fecal P to plants has been investigated in several

152

R. J. HAYNES A N D P. H. WILLIAMS

short-term greenhouse experiments (McAuliffe et al., 1949; Bromfield, 1961;Gunary, 1968). In these studies ground feces were incorporated into the soil and the inorganic fecal P content was found to be as effective as a P source as readily soluble fertilizer P. The organic P content did not, however, appear to be readily available, at least in the short term. In short-term field trials (8 and 17 weeks, respectively, in spring and autumn), Rowarth et al. ( 1990) found that P uptake from fecal inorganic P was less than that for monocalcium phosphate. Nonetheless, under field conditions, During and Weeda (1973) observed that herbage yields and P uptake were greater from cattle dung than superphosphate. Yield response of herbage to dung and superphosphate persisted for 2 and If years, respectively. Similarly, McAuliffe and Bradfield (1955) found that the availability of P from superphosphate and feces was similar for a first cut of grass, but by the third cut the availability of P from feces exceeded that from superphosphate. The superiority of dung as a P source was partially ascribed by During and Weeda (1973) to the large quantity of N (and other nutrients), which stimulated initial pasture yields more than superphosphate, thereby increasing P uptake and usage. In addition, dung applications may have decreased phosphate adsorption by the soil (see Section IV,A,4), thus increasing phosphate availability. d. Other Nutrients Release of K and Na from feces is rapid because the bulk of these elements is in water-soluble form. Weeda ( I 977) observed that in the soil below dung pats, peak levels of exchangeable K were reached 1 month after application (Fig. 8). In contrast, the release of Ca and Mg was much slower (Fig. 8) and peak levels were not reached until 4 months after application. The slower release of Ca and Mg is expected due to the lower proportion of water-soluble Ca and Mg present in dung (Section III,B,4). Underhay and Dickinson (1978) observed that Ca was leached from cattle feces more rapidly than Mg, but the reason for this is unclear. The release of micronutrients from feces has received little attention, although Barrow (1987) pointed out that the alkaline conditions in the dung are likely to limit the solubility of nutrients such as Fe, Mn, Zn, and Cu and thus retard their release. 4. Effect on Soil Properties

Increases in extractable P, exchangeable Ca and Mg, and sometimes exchangeable K commonly occur in the surface soil 2.5 to 5 cm below dung patches (Davies et al., 1962; MacDiarmid and Watkin, 1972b; Dur-

153

NUTRIENT CYCLING UNDER GRAZED PASTURE

Figure 8. Changes in exchangeable cation levels in the surface 3.8 cm of soil with time below cattle dung patches (values represent differences between levels under dung pats and those in the surrounding soil). (Redrawnfrom Weeda, 1977.)

ing et al., 1973; Weeda, 1977). Increased levels of exchangeable Ca, Mg, and K in the soil (0-2.5 cm) below cattle dung patches 3 years after deposition are shown in Table IX. Over a 15-month period following cattle dung deposition, Weeda (1 977) observed a sustained rise in Truog P levels. Mean levels over that period in control and dung-covered plots were 17 and 22 pg P g-*, respectively, in the 0- to 3.8-cm soil layer and 10 and 13pg P g-l, respectively, in the 3.8- to 7.6-cm layer. Due to the high pH and high CaCO, content of feces (Section III,B,4), an increase in soil pH below the dung patch is a common phenomenon (Davies et al., 1962; During et al., 1973; Omaliko, 1984). Davies et al. (1962) recorded a soil pH increase from 6.1 to 6.7 below dairy cow dung

Table IX Effect on Soil Properties of Cattle Dung Additiona Exchangeable cations

Control Dung

9.4 10.9

0.76 0.89

5.2 5.6

12.4 17.7

11 15

1.5 2.6

'Data from During er al. (1973). Effects are evaluated 0-2.5 cm below the area of addition after a 3-year period.

154

R. J. HAYNES AND P. H. WILLIAMS

pats 10 months after application and During et al. (1973) measured a pH increase from 5.2 to 5.6 in the 0- to 2.5-cm increment 3 years after application (see Table IX). In the soil/pasture/animal system dung appears to be the main natural agent that maintains soil pH or arrests decreases in pH (During et al., 1973). Soil organic matter content (organic C and total N) was shown by During et al. (1973) to be significantly increased below dung patches even on a soil with an initially high soil organic matter level (Table IX). This increase is attributable to the very large inputs of organic matter that are deposited in the dung patch (e.g., equivalent to 20-50 tonnes per hectare (During and Weeda, 1973) (Table MI). Dung deposition plays a major role in the buildup in soil organic matter that often occurs under improved pasture (see Section II,C, 1). In the surface soil below the dung patch there is a decrease in both phosphate and sulfate adsorption capacity (During and Weeda, 1973; During et al., 1973).This is thought to be due to the significant increases in soil pH (During et al., 1973). The increased soil organic matter content may also contribute to this effect because soluble organic materials are known to be able to block some of the adsorption sites on hydrous oxide surfaces and thus decrease phosphate adsorption (Sanchez and Uehara, 1980; Yuan, 1980).

B. RESPONSE OF PASTURE INTHE FECALPATCH 1. Direct Adverse Affect

Dung pats can cover the soil surface and exclude light from the sward for several months, leading to the death of plants that are covered. MacDiarmid and Watkin (197 1) observed that if cattle dung patches remained for more than 15 days on the pasture, there was little regrowth from plants underneath the pat because pasture herbage, particularly clovers, decayed rapidly. Under solid dung pats pasture often dies, but Weeda (1967) noted that under very liquid dung patches pasture regrowth is fairly rapid because these pats disperse and disintegrate rapidly. Where plants were killed below the patch, Weeda (1967) found that although the outside 2.5 cm of the patch area was rapidly covered by tillers of surrounding grasses, the central area remained sparsely covered for 6 to 12 months and in some cases for up to 2 years. Recolonization of the bare area may come from the surrounding herbage

NUTRIENT CYCLING UNDER GRAZED PASTURE

155

or from seeds of pasture and weed species in the soil and/or dung. While recolonization is occumng, pasture production is lost from the area. Weeda ( 1967) found that in grass/clover pastures white clover was usually the first species to recolonize these areas and it tended to remain dominant for 12- 18 months. In sheep pastures, bare areas of land due to dung deposition are rarely seen. There are two important factors that contribute to this difference. First, the weight of dung produced per sheep per defecation is much less than that per cattle beast (0.0 1 versus 0.20 kg, respectively; see Table VII). Second, sheep feces are often produced as small pellets that are scattered over the pasture rather than being deposited as one large pat, as is the case with cattle. 2. Positive Pasture Response

Even though there is often a depression in pasture growth immediately below the dung pat, the pasture adjacent to the pat often shows an appreciable positive dry matter response. Cattle dung patches affect growth of herbage over an area five to six times that actually covered (Norman and Green, 1958; MacLusky, 1960; MacDiarmid and Watkin, 1972b; During and Weeda, 1973; Weeda, 1977). The positive effects of dung on surrounding pasture growth are often long-lived and may last up to 2 years (Weeda, 1967, 1977; During and Weeda, 1973; Richards and Wolton, 1976). The increased growth is usually related to increased grass growth (Norman and Green, 1958; MacDiarmid and Watkin, 1971; Weeda, 1977). Weeda (1977) found that herbage in the zone of dung application was reduced during the first If months after application due to pasture death. However, in the adjacent zone (a band 12.7 cm wide around the pat) pasture growth was stronger than in the control and this compensated for the depression in the center. In the following spring (approximately 1 year after application) herbage yields in both the patch area and the adjacent zone were significantly superior to those in an area unaffected by dung. When herbage around dung patches was cut to simulate grazing (taking account of reduced herbage utilization around dung patches; see Section IV,B,4), Weeda ( 1967) found that there was no dry matter yield increase in the patch area but there was a 38% increase in yields in the adjacent zone. Overall, there was a 28% increase in dry matter due to dung deposition over a l)-year period. In similar experiments During and Weecia (1973) measured overall dry matter increases of 27 and 30%, respectively over a 3-year period for late winter and summer dung depositions.

156

R. J. HAYNES A N D P. H. WILLIAMS 3. Chemical Composition of Herbage

The increase in herbage dry matter production around the dung pat is a response to added nutrients. This is probably caused by lateral wash of dung particles, lateral spread of roots, and the presence of stoloniferous plants, such as white clover, in the sward. The fact that the response is often long-lived, lasting up to 2 years, reflects the slow release of nutrients from the dung. The increase in dry matter production is often attributed to a response primarily to N in the dung (During and Weeda, 1973; Lantinga et al., 1987), although where other nutrients such as P, K, Mg, or Ca are limiting a response to dung application would be expected. On acid soils, the liming effect of dung is likely to stimulate production. In general, concentrations of N, P, Mg, and K are increased in herbage around the dung pat (Davies et al., 1962; During and Weeda, 1973; During et al., 1973; Weeda, 1977). There is, however, little or no increase in the concentration of Ca in herbage despite the high input of Ca (Weeda, 1977). This is presumably due to the large reserve of exchangeable Ca present in many soils. On all-grass pastures supplied with high rates of N (e.g., 200-500 kg N ha-' yr-l), dung pats appear to have little if any effect on herbage N content or grass growth around the pats (Lantinga et al., 1987), presumably because N (as well as other nutrients) is not limiting to herbage production. The recovery of individual nutrients present in dung by pasture herbage will be very much dependent on the magnitude of the herbage response and on which nutrients are limiting production. On a K-deficient site, Weeda (1977) found over a 3-year period that the apparent recovery of nutrients from dung was as follows: K, 94%; Na, 55%; Mg, 27%, P, 24%; and Ca, 3990. 4. Herbage Utilization

Rejection of herbage around cattle dung patches is a characteristic of pasture contamination by cattle (Marsh and Campling, 1970; Wolton, 1979; Wilkins and Garwood, 1986). Initially this rejection is caused mainly by the unpleasant odor of the dung (Marten and Donker, 1964; MacDiarmid and Watkin, 1972b). However, subsequently the ungrazed herbage can become mature and unpalatable and rejection is then predominantly due to this factor rather than proximity to dung (Norman and Green, 1958). The area around pats that is at least partially rejected by cattle has been estimated to be 5 - 12 times greater than the pat area itself (Greenhalgh and Reid, 1968; Bastiman and Van Dijk, 1975). The period of

NUTRIENT CYCLING UNDER GRAZED PASTURE

157

rejection may last from 2 to 3 months (Weeda, 1967) to 13 to 18 months (Norman and Green, 1958; Castle and MacDaid, 1972). Grazing pressure is a major factor affecting the extent of herbage rejection. When there is a surplus of herbage, grazing cattle reject pasture around dung pats to a large extent and the pasture becomes covered with clumps of long, rank herbage that then becomes unpalatable. As a consequence, much rejection occurs around old pats. Tiller density of grasses growing in the rejected areas is also gradually reduced, giving rise to sward deterioration (Lantinga et al., 1987). With intensive grazing caused by high stock density, the rejected area is minimized. The length of herbage around pats is maintained at an acceptable level because the upper parts of the herbage in these areas are grazed (Wolton, 1979). By about the fourth rotational grazing, the height of grazing returns to that of the surrounding pasture (Weeda, 1967, 1977). Thus, on highly productive dairy farms with rotationally grazed pastures the major source of wastage is partial rejection of herbage around freshly voided feces- not rejection around old pats (MacLusky, 1960; Weeda, 1967; MacDiarmid and Watkin, 1972b;During and Weeda, 1973). Indeed, under intensive grazing systems cows may even become accustomed to eating dung-affected herbage (Marsh and Campling, 1970; Garstang and Mudd, 1971). The net effect of dung patches on pasture growth and herbage intake by animals will depend on the balance between two major factors. These are (1) the magnitude of the positive growth response of the pasture in the vicinity around the patch, and (2) the grazing pressure that will determine the extent to which the animals use this increased pasture growth. At extremely low grazing pressures, particularly under set stocking, rejection will be marked. Nonetheless, there may be some compensation for poor utilization in the patch areas through increased grazing severity over the remaining sward. The net result could be that herbage utilization is similar to that of unfouled pastures (Wilkins and Garwood, 1986). At moderate grazing pressures there is likely to be a reduction in herbage utilization on fouled paddocks (in comparison with previously unfouled ones), because any possible increase in grazing severity on nonpatch areas cannot compensate for the rejection around dung pats (Greenhalgh and Reid, 1968; Wilkins and Garwood, 1986). Under high grazing pressures, where rejection is minimized, there will often be little or no reduction in herbage intake on fouled pastures (Boswell, 1971; MacDiarmid and Watkin, 1972a; Lantinga et al., 1987).In fact, where there is a sizeable herbage dry matter response to dung application, intake will be increased. There are no reports of rejection of herbage around sheep depositions. This is presumably for the same reasons that bare areas are not evident on

158

R. J. HAYNES AND P. H. WILLIAMS

sheep. pastures (i.e., small volumes of feces are deposited mainly as relatively dry pellets over a relatively large surface area).

C. MOVEMENT AND TRANSFORMATIONS OF NUTRIENTS FROM URINE 1.

Urine Composition

The main nutrients present in animal urine are N, K, and S (Section 111,B).Typical application rates of these nutrients per urine patch for sheep and cattle and annual output per cattle beast and per hectare of farm are shown in Table X.The very high quantities of N and K applied to urine patches of both sheep and cattle are evident. The movement and transformations of nutrients in the urine patch are discussed below. 2. Macropore Flow of Urine

Under field conditions, preferential flow of applied water through large, previously air-filled macropores (>50 pm in diameter) occurs when the application rate exceeds the rate at which water infiltrates the soil matrix (Bouma, 1981). Macropores include soil fracture planes and channels created by plant roots, earthworms, and other soil microfauna. During a urination event, the high rate of urine application can cause ponding at the

Table X Typical Application Rates of Major Nutrients Applied to Sheep and Cattle Urine Patches

Nutrient N

S K

Urine concentration (g liter1) 10 0.35 9

Application rate per urine patch (kg ha-[)= Sheep

Cattle

500

lo00

18 450

35 900

Annual output per cattle beast (kg animal-I ~ r ' )

Annual output from cattle per hectare of f m ~(kg ha-I TI)'

13

183

2.6 66

6.5 165

a Assuming that for sheep and cattle a urination consists of 0.15 and 2.0 liters applied to an area of 0.03 and 0.20 m2,respectively. Assuming cattle urinate 10 times per day. 'Assuming 2.5 cattle per hectare.

NUTRIENT CYCLING UNDER GRAZED PASTURE

159

soil surface and the initiation of preferential flow (Monaghan et al., 1989; Whitehead and Bristow, 1990; Williams et al., 1990a-c). For example, when a cow urinates, the flow rate is normally 0.2 liters sec-' (Goodall, 1951;Davies et al., 1962)and the surface application rate is approximately 10 liters m-z (Hog, 1981). Macropore flow has been shown to occur immediately following both sheep and cattle urination events. Using cow urine spiked with tritiated water, Williams et al. (1990c) detected tritium in the first 0.02 pore volume of leachate collected from undisturbed soil cores 15 cm deep. Similarly, Monaghan et al. ( 1989)measured the highest N concentrations in leachate in the first 0.1 pore volume of drainage after application of sheep urine to undisturbed soil cores. Ryden et al. (1984) reported substantial amounts of NHt nitrogen at depths between 0.66 and 1.33 m in the profile below urine patches. The immobility of NHt in soils suggests that macropore movement of urea had occurred and this had been followed by urea hydrolysis. Over a 7-month period following a urine event, Williams et al. (1989) found that the vast majority of N and K absorbed by the pasture plants originated from the surface 12-cm depth of soil in the urine patch. This coincided with the distribution of plant roots within the soil profile because 86% of roots in the surface 30 cm were found in the top 15 cm. Thus, at least in the short term, nutrients that move below 15 cm or so by macropore flow will represent a loss of nutrients from the pasture system. Furthermore, nutrients that move below 15 cm by macropore flow will be more likely to be subsequently leached to lower soil depths by percolating drainage water than are those that accumulate close to the soil surface. Considerable spatial variability exists in the extent of macropore flow in urine patches in the same field (Williams et al., 1990~).Such variability reflects spatial variability in the number of surface-connected macropores and their continuity and tortuosity, as well as other factors that influence surface flow of added urine. These include surface microtopography, surface water repellency, surface compaction, and the influence of plant cover on interception and stem flow of urine. A simple method for estimating the extent of macropore movement at a field site is to simulate urination events using potassium bromide solutions and measure the amount retained in the surface soil horizon (Williams et al., 1990b). The unrecovered bromide is that which has moved below that depth by preferential macropore flow. Because preferential flow occurs too rapidly for any significant chemical reactions to occur between urine solutes and the soil, the concentration of nutrients in the leachate is similar to that of the urine. Thus, the amount of nutrients lost due to macropore movement of urine can be calculated from the nutrient content of urine and the bromide distribution.

160

R. J. HAYNES AND P. H. WILLIAMS

Below dairy cow urinations, Williams et al. (1990b) measured losses of nutrients (to below a 15-cm soil depth) via macropore flow ranging from 0 to 46% on seven different soil types. Whitehead and Bristow (1990) measured a 37% loss of lSN-labeledurea immediately following cattle urinations under field conditions. Losses from sheep urinations are likely to be less than those from cattle because the surface application rate of urine is approximately 5 liter m-2 in comparison with 10 liter m-? for cows. However, no field data are available to substantiate the extent of such losses. 3. Urea Hydrolysis and Ammonia Volatilization

In the soil, urea undergoes hydrolysis catalyzed by the enzyme urease to form (NH,),CO,: (NH2)zCO + 2H2O

(NHJ, CO,

(1) This reaction causes localized areas of high pH close to the site of hydrolysis: C@-

+

+ H,O eHCO: + OH-

(2) Hydrolysis is rapid and Sherlock and Goh ( 1984) calculated half-lives of urine urea as 3.0 and 4.7 hr, respectively, under summer and autumn conditions. The reduction in hydrolysis rate in autumn was attributed to lower soil temperatures. Hydrolysis of urine urea is more rapid than that of pure urea when it is added to the soil under similar conditions (Doak, 1952; Sherlock and Goh, 1984). The major reason for this is that hippuric acid, a minor nitrogenous constituent of animal urine, has a stimulatory effect on urea hydrolysis (Doak, 1952; Whitehead et al., 1989). The high pH of urine (8.6) would also directly favor the hydrolysis of urea (Sherlock and Goh, 1984)because this is the optimum pH for urease activity (Vlek et al., 1980). In addition, our unpublished results have shown that the urea in animal urine hydrolyzes extremely rapidly after release from the animal, suggestingthat urease activity is already present in the voided urine. During the first 24 hr after a urination event, as urea hydrolysis proceeds, there is a rapid rise in soil pH in the urine patch (Fig. 9). The rise is greatest near the soil surface and decreaseswith soil depth (Ball et al., 1979; Vallis et al., 1982).A rise in pH of 2.5 to 3.5 pH units is not uncommon in the surface 0.5 cm of soil (Vallis et al., 1982; Sherlock and Goh, 1984). As nitrification proceeds, the soil pH decreases again (Fig. 9). Following urea hydrolysis, large amounts of NHt nitrogen accumulate in the soil. By 24 hr after a urination event, concentrations of NHt nitro-

161

NUTRIENT CYCLING UNDER GRAZED PASTURE 8-

B-

7 -

Ip 54

0

2

4

6

8

10

12

14

Time after application (days) Figure 9. Mean daily pH at four soil depth increments in a urine patch over a 2-week period. (Redrawn from Vallis ef a/., 1982.)

gen in the surface 10 cm of soil in the urine patch commonly reach 100-250,ug N g-l (Ball et al., 1979; Carran et al., 1982; Sherlock and Goh, 1984). Concentrations of NH: from 500 to 1000pg N g-I have been observed in the surface 2.5 cm (Vallis et al., 1982). Ammonium ions interact with the cation exchange complex in the soil, resulting in electrostatic binding of NHt ions to clay and organic colloids. Some NHacan also become fixed in the lattices of 2 : 1 clay minerals. Significant gaseous losses of NH, occur from the urine patch subsequent to a urination event. A necessary prerequisite for NH, volatilization is a supply of free NH, near the soil surface. The conversion of NHt ions to NH, is thus the major process regulating the potential loss of NH, from soils: NH:

+ OH- e NH,+ H,O

(3)

The equilibrium between NHt and NH, is affected by many factors, but generally the supply of NH, is favored by high soil pH, high temperatures, and evaporative loss of soil water (Haynes and Sherlock, 1986). The high concentration of NHZ and high pH in the urine patch both favor NH, volatilization losses. Temperature is a particularly important factor. Lockyer and Whitehead ( 1990) measured a positive correlation between soil temperature at a 3-cm depth during the 3 days following urine application and the extent of volatilization loss from urine patches. Losses of NH, from urine patches generally represent from 4 to 46% of urine N (Table XI) and losses of 15 to 25% are common. Hot, dry, summer conditions favor losses whereas cool, moist, winter conditions

R. J. HAYNES AND P. H. WILLIAMS

162

minimize losses. Thus Sherlock and Goh (1984) measured mean urine patch volatilization losses of 22% in summer, 25% in autumn, but only 12% in winter. Under conditions in which mean air temperatures were 1 6 T , Ryden et al. (1987) measured losses of urine N of 22%; when mean air temperatures were 8 'C losses of only 10%were recorded. In the tropical dry season Vallis et al. (1985) observed losses as high as 46% (Table XI). The characteristic pattern of NH, volatilization from urine patches (Vallis et al., 1982; Sherlock and Goh, 1984)is illustrated in Fig. 10. There is a rapid rise in NH, flux during the first 24 hr and this is followed by a more gradual exponential decline. Superimposed on this are temperatureinduced diurnal fluctuations with higher rates occumng in the daytime. Under a rotational grazing system Ryden (1985a) showed that, as expected, NH, losses were greatest from pastures during, and immediately after, grazing and that highest rates of loss were associated with high stock densities (Fig. 1 1). The bulk of such NH, would have originated from recently formed urine patches, although some could have come from dung patches (Section IV,A,3). Between grazings, NH, losses continued at a low rate. Jarvis et al. (1989) measured losses of ammonia from grazed ryegrass swards provided with 210 and 420 kg N ha-' yr-' of 0.8 and 5.9 kg N ha-' yr-l , respectively, whereas losses from a grass/clover pasture estimated to be fixing 160 kg N ha-' yr-' were 0.7 kg N ha-' yr-l. The much larger

Table XI Examples of Field Measurements of Ammonia Volatilization Losses from Urine Patches Ammonia loss Reference Ball et al. ( 1979) Carren et al. ( 1982) Vallis et al. ( 1982)

Method

Climate

Direct Direct Direct

Temperate (summer) Temperate (summer) Subtropical (summer/ winter) Sherlock and Goh (1984) Direct Temperate (summer/ winter) Vallis and Gardener Short-term I5N Tropical (dry/wet seasons) (1984) balance Short-term "N Tropical (dry/wet seasons) Vallis et al. (1985) balance Ryden et al. (1987) Direct Temperate (summer/ autumn) Whitehead and Bnstow Direct Temperate (summer)

(% of that applied)

15-18 17-36 14-28 12-25 16-32 46 9-25 18

(1990)

Lockyer and Whitehead ( 1990)

Direct

Temperate (summer/ winter)

4-27

NUTRIENT CYCLING UNDER GRAZED PASTURE

163

3.2

a

i

2.4 ln

1.6

E

.-0 .-E

0.8

4-

P P

O 3.2

m

6

2.4

c 4-

m

E

i=

1.6

0.8

Night

mDay

0 0.2 1

2

3

4

5

6

7

8

Ammonia loss (percent loss per 4 hr)

Figure 10. Rate of ammonia volatilization from urine patches over a 2-week period under temperate conditions. Losses shown (a) during Australian spring (November) and (b) summer (February). (Redrawn from Vallis et a/., 1982.)

2.0

132,

,24,

1.6

II

1.2

0.8

0.4

24 31 7 14 21 2 8 5 12 19 May June July

2

d9 l16 -23 n 27- 4l 11- 19 30

Aug.

Sept.

Oct.

Figure 11. Mean losses of ammonia from grazed swards (yearling steers in a 28-day rotation) receiving 420 kg N ha-' yr-'. Bars show grazing period and stocking rates (steers per hectare). (Redrawn from Ryden, I985a.)

164

R. J. HAYNES AND P. H. WILLIAMS

losses from the high-nitrogen ryegrass pasture were attributed to the larger numbers of stock carried and also the greater proportion of ingested N being returned in the form of urine in this high-nitrogen treatment. 4. Soil pH and Charge Characteristics

In soils rich in colloids with variable charge characteristics (i.e., they possess surfaces for which H+ and OH- are potential determiningions), the rise in soil pH in the urine patch causes an increase in surface negative charge and thus the CEC increases. This has two major effects. First, it tends to reduce gaseous losses of NH3 by decreasing the concentration of NHt in soil solution. Second, it increases the retentive power of the soil for the high concentrations of K+ added in the urine. For example, in a laboratory study Williams et af. (1988) compared the retention of K by a volcanic soil of variable charge following application of either KCl or cow urine. Because of the rise in soil pH and surface negative charge following urea hydrolysis, more K was adsorbed by the soil from the cow urine than from the KCI. Such a phenomenon is likely to reduce leaching losses of K in the period immediately following a urination event. 5. Nitrification

Following urea hydrolysis, the accumulated NHZ is nitrified. This process is carried out by the actions of chemoautotrophic nitrifying bacteria. In soils, five genera of autotrophs are known to be able to oxidize NH; to NO;: Nitrosomonas, Nitrosococcus, Nitrosospira, Nitrosolobus, and Nitrosovibrio. Oxidation of NHt to NO; occurs as follows: NH:

+ lfo,

-

NO;

+ 2H+ + H,O

-

One genus, Nitrobacter, is known to oxidize NOT to NO,: NO; + +02

NO;

The process is acidifying because, during the initial oxidation of NHZ to NO;, 2 mol of H+ are produced per mole of NH: oxidized. The rate of nitrate accumulation in the urine patch varies greatly, depending on soil and environmental conditions (Watson and Lapins, 1969; Holland and During, 1977; Thomas et af., 1988). Under warm temperate conditions, NO, is often the major form of N present in the urine patch after 3 to 5 weeks (Thompson and Coup, 1940; Ball et af., 1979), although under very dry summer conditions, Sherlock and Goh (1984) found only 20-23% of urine N had been converted to NO, after 41 days. Soil temperature is a particularly important factor influencing nitrifica-

NUTRIENT CYCLING UNDER GRAZED PASTURE

165

tion (Haynes, 1986a). Under field conditions in New Zealand, Holland and During (1977) found that nitrification of urine N was not appreciable until 7 days after urination. At temperatures of 7.5" to l O T , nitrification continued to be slow and was not complete until 60 days. By contrast, at temperatures above 15°C it was rapid once initiated and was complete within 30 days. In the Scottish upland under acidic soil conditions, Thomas et af. (1988) noted that high levels of NO; accumulated in the urine patch only from summer applications. For spring, autumn, and winter applications, levels of NOT were generally low and comprised up to only 25% of the total inorganic N in the urine patch at any one time. Under hot, moist, subtropical conditions, Vallis et al. (1982) observed that nitrification was very rapid, with over 50% of the applied urine N being recovered as NO; after 2 weeks. High NHZ concentrations and high pH, such as those that occur in the urine patch, can inhibit nitrification (Haynes, 1986a). Nitrite oxidation is depressed to a greater extent than NHZ oxidation and consequently NO; can sometimes accumulate in the urine patch (Barlow, 1972; Holland and During, 1977; Vallis et al., 1982). Vallis et af. (1982), for example, found that in subtropical Queensland concentrations of up to 45 pg g-' NO, nitrogen accumulated in the surface 2.5 cm of urine patches after 7 days, but this had virtually disappeared after 14 days. In a field study Barlow (1972) recorded maximum levels of NO; nitrogen in urine patches of 17 - 38 pg g- in the surface 2 cm of soil. Nonetheless, in many studies only traces of NO; nitrogen have been recorded (e.g., Holland and During, 1977; Ball et af.,1979). Conditions of low soil water content are apparently particularly favorable for NO; accumulation in urine patches (Barlow, 1972; Holland and During, 1977), presumably because they result in increases in NH; concentrations in soil solution. Holland and During (1 977) found that NO; in urine patches did not exceed 1 pg N g-' except in a summer application under hot, dry conditions with low moisture content, wherein it reached 9 pg-' in the surface 15 cm of soil. Accumulation of NO: in the urine patch has led to speculation that gaseous losses of N through chemodenitrification may occur (Barlow, 1972; Sherlock and Goh, 1983). Chemodenitrification is a term used to describe a range of chemical reactions of NO, ions with soils, resulting in the emission of various nitrogenous gases (N,, NO, NO,, and sometimes N,O). Chemodenitrification is generally favored by accumulation of high concentrations of NO; in association with low soil pH (i.e., pH <5.5) (Haynes and Sherlock, 1986). The high pH in the urine patch while NO; is accumulating hinders gaseous loss through chemodenitrification. Consequently, Barlow (1 972) observed that total nitrogen dioxide volatilization

166

R. J. HAYNES AND P. H. WILLIAMS

from simulated field urine patches during 25 days following application represented only 0 - 2% of urine urea N applied. Losses of nitric oxide are also likely to be low; Galbally and Roy (1978) observed rates of loss of 0.001 and 0.002 kg N ha-' day-' from ungrazed and grazed pasture, respectively. 6. Immobilization of Nutrients into Organic Forms

During transformations of nutrients, such as applied urea N and sulfate S, in the urine patch a portion can be rapidly immobilized, by the microbial biomass, into soil organic forms. Keeney and MacGregor (1978), for example, observed that 13%of applied 15Nurea in a simulated urine patch was recovered in organic form after 7 days. However, incorporation of N into organic matter is not always rapid. Using 15N-labeledcattle urine, Whitehead and Bristow (1990) found that 15Nwas slowly incorporated into the soil microbial biomass during the initial 16 days following urine application. After 16 days only 6.3% of the 15Nin the surface soil was recovered in the microbial biomass and this rose to 17% by day 28 before falling to 2.8% by day 321. At day 28 only 3.7% of I5N in the surface soil had been incorporated into humified organic matter, but by 32 1 days this had risen to 13%. Rapid incorporation of urine sulfate into soil organic forms was reported by Williams and Haynes (1990b). They found that 25% of 35Slabeled sulfate in sheep urine had been incorporated into soil organic S after 7 days, and this rose to 37% by 64 days. The rate of incorporation of N into the microbial biomass and other soil organic fractions will depend on factors influencing mineralization/immobilization turnover, for example, the C: N ratio of organic residues and environmental factors (Haynes, 1986b). Similar factors will influence the incorporation of urine sulfate into soil organic fractions. Furthermore, incorporation of urine 15N or 35S into soil organic fractions does not necessarily mean that organic N and S is accumulatinginto the urine patch area, because some unlabeled soil organic N and/or S may be being mineralized simultaneously. 7. Losses of Dinitrogen and Nitrous Oxide Gases

A major source of N, and N20 emissions from soils is the process of denitrification. Denitrification is carried out by a limited number of aerobic bacteria that can grow in the absence of molecular oxygen while reducing NO; and/or NO; to N, and N,O (Haynes and Sherlock, 1986). The process of denitrification is promoted under anaerobic conditions, high levels of soil NO,, and a readily available source of carbon, and in

NUTRIENT CYCLING UNDER GRAZED PASTURE

167

general it is positively related to soil pH and temperature. The relative proportion of N,O/N, evolved during denitrification increases as the soil becomes more aerobic (Firestone et al., 1979). Ryden ( 1986) observed that under grassland, the conditions that were conducive to high rates of denitrification (N, plus N20 evolution) were soil NO, content > 5 pg N g-*, soil temperature > 5 "C,and an air-filled porosity < 15 - 40%. The potential for denitrification losses from grazed pastures is high due to the high levels of readily oxidizable C in the surface soil and the high concentrations of NO; present in soils under urine (and dung) patches. During summer, Ryden (1984), using the acetylene inhibition technique (Ryden et al., 1979), measured mean rates of denitrification loss from an irrigated grazed pasture of 0.05 to 0.35 kg N ha-* day-', 1.7 to 16 times greater than those from a comparable cut sward. Similarly, on an unimgated site, Ryden ( 1986)showed greater rates of denitrification from grazed than cut swards. Some results from these two studies (Ryden, 1984, 1986) are summarized in Table XII. The higher rates of denitrification from grazed swards are the result of large losses from urine- and dung-affected areas (Table XII). These areas had considerably higher contents of both soil NO: and soil water than did the remainder of the sward (Ryden, 1986). Denitrification losses were high in autumn (October sampling), when soil moisture content had returned to near field capacity. However, with the onset of winter (November sampling), soil temperatures had

Table XI1 Denitrification Losses for Irrigated, Fertilized Pastures during Summer and Unirrigated, Fertilized Pastures during Autumna Rate of denitrification (kg N ha-' day-') Grazed sward Management and conditions Imgated pasture Day 3 Range over 20 days Unimgated pasture 21 October (soil temp., 11 "C) 30 November (soil temp., 6°C)

Cut sward

Areas not affected by excreta

Areas affected by excreta

0.040 0.004-0.040

0.058 0.0 13-0.064

0.199 0.032-0.40

<0.00 1

0.164

0.009

0.016

0.318 (urine) 0.386 (dung) 0.027 (urine) 0.073 (dung)

After Jarvis et al. (1987). Data from Ryden (1985b) and Ryden and Nixon (1985).

168

R. J. HAYNES AND P. H. WILLIAMS

declined to about 5°C and denitrification was much reduced. On an annual basis, Ryden (1986) estimated denitrification losses of 40 and 20 kg N ha-' yr-I from grazed and cut swards of ryegrass receiving 420 kg N ha- yr- . Sherlock and Goh (1 983) measured greater losses of N,O from simulated urine patches when sheep urine rather than aqueous urea was applied. Whereas the N,O release following urea application took 12-24 hr to peak, increased N,O fluxes were measurable immediately after the application of urine. The reason for this effect of urine is unknown. A possible explanation is that the very rapid hydrolysis of urea in urine, as compared to urea alone, results in rapid generation of CO, and thus the rapid onset of anaerobosis in microsites. This could initiate the onset of denitrification using indigenous NOT as a substrate (Sherlock and Goh, 1983). Gaseous losses of N20 can also occur during the process of nitrification (Bremner and Blackmer, 1978). During nitrification, N,O is produced under aerobic conditions but the rates of N,O evolution increase as oxygen concentrations decrease (Goodroad and Keeney, 1984). Some workers have suggested that nitrification may well be the major source of N,O production from soils, especially when they are fertilized with urea or NHt (Klemedtsson et al., 1988). The large amounts of NHZ that are typically nitrified in the urine patch make it likely that nitrification would contribute significantly to N,O losses from grazed pastures. Such losses would not have been measured in the work by Ryden (1984, 1985b, 1986) because the "acetylene block" technique inhibits the nitrification process (Berg et al., 1982). The magnitude of N,O losses mediated by nitrification on grazed pastures has yet to be quantified. 8. Leaching Losses of Nutrients

Most studies of nutrient leaching from grasslands have been concerned with leaching losses of NOT from under fertilized cut swards. Generally, such losses range from 8 to 20 kg N ha-' yr-' (Ryden, 1984; Garwood and Ryden, 1986) and, as a consequence, it has been generally considered that leaching losses from grasslands are small. However, the activity of grazing animals has a large impact on the leaching losses of NO? from grazed pastures. The primary source of nitrate in drainage waters from grazed pastures is animal urine. As already noted, preferential flow of urine down surfaceconnected macropores may result in significant amounts of nutrients moving below 15 cm in the soil profile immediately following a urination. The quantity of N applied in the urine patch far exceeds the immediate requirements of the pasture plants and consequently mineral N accumu-

NUTRIENT CYCLING UNDER GRAZED PASTURE

169

lates in the soil profile. Indeed, the mean mineral N concentration in the surface 20-30 cm of soil in the urine patch is often in the range of 100-250pg N g-l (Ball and Ryden, 1984; Ryden et al., 1984; Thomas et al., 1988). Once the bulk of this mineral N has been nitrified, leaching of NO, will occur when excess precipitation occurs and water is moving down the profile. Thus, Sherwood and Fanning (1989) found that there was considerable leaching of nitrate from urine patches deposited in autumn and early winter but negligible loss from those deposited in spring and summer. The direct impact of the grazing animal on nitrate leaching was demonstrated by Ryden et al. (1984). In this study, nitrate leaching was assessed from the hydrologic balance at the experimental site and the nitrate content of the soil and underlying chalk below grazed and cut swards. Nitrate contents of profiles below grass/clover and fertilized ryegrass swards, either cut or grazed, are shown in Fig. 12. The nitrate contents in the profile below grazed swards were greater by a factor of 5.7 to 10.2 than those below contrasting cut swards (Garwood and Ryden, 1986). The estimated leaching losses of nitrate were 2.5, 23, 29, and 162 kg N ha-’ yr-* for the cut and grazed grass/clover and ryegrass swards, respectively. Ryden et al.

Soil nitrate content ( pg N g” ) 0

1.0 0

2.0

0

2.0

4.0

0

1.0

10

12

14

16

Figure 12. Distribution of nitrate in the soil (and chalk below) from successive depths below cut and grazed ryegrass/white clover swards (a and b, respectively) and below cut and ryegrass swards receiving 420 kg N ha-I yr-l (c and d, respectively).(Redrawn from Garwood and Ryden, 1986, with permission of Kluwer Academic Publishers.)

170

R. J. HAYNES AND P. H. WILLIAMS

(1984) related these differences to the mean concentrations of NO; in the surface 30 cm of soil from the cut and grazed treatments. Mean NO; concentrations were 2.5 pg N g-l below the cut swards and 7.5 p g N g-l below the grazed swards. However, below the urine-affected areas of the grazed sward mean concentrations were 53pg N g-'. Whitehead and Bristow (1990) noted substantial leaching losses of nitrate from urine patch areas of pasture during a spring period in which exceptionally heavy rainfall was recorded. Steele et al. ( 1984) studied leaching losses from intensively grazed grass/ clover pastures receiving 0 or 172 kg N ha-' yr-l. Leaching losses of NO, were estimated to be 88 and 193 kg N ha-', respectively on the unfertilized and fertilized pastures. It is important to recognize that when an anion such as nitrate is leached, equivalent amounts of cations will also be leached as counterions for NO;. Calcium and to a lesser extent Mg2+are the major counterions for NO; leaching in urine patches (Holland and During, 1977; Hogg, 1981; Williams et al. 1990a). Thus, Steele et al. (1984) observed that Caz+was the major cation leached from grazed pasture and that the concentrations of NO; and Ca2+in leachates were closely correlated ( r 2 = 0.96***). Because K+ is the major cation in urine, it would be expected that K+ would be the major cation leached. However, following a urination event, the added K+ rapidly equilibrates with the exchangeable cations in the soil (Williams et al., 1988). Following urea hydrolysis and subsequent nitrification, both NO; and H+ ions are produced in the urine patch. The H+ ions can displace other cations on the soil exchange sites, resulting in a greater quantity of potentially leachable cations being present in soil solution. Because Ca2+is the dominant exchangeable cation in most soils, it is Ca2+that is the predominant cation displaced and subsequently leached.

D. RESPONSE OF PASTURE INTHE URINE PATCH 1. Positive Pasture Growth Response

Normally, there is a marked positive pasture growth response in the urine patch. This is commonly attributable to a response to added N and normally lasts for 2 to 3 months (Norman and Green, 1958; During and McNaught, 1961; Ledgard and Saunders, 1982; Ledgard et al., 1982; Thomas et al., 1988). Dry matter yield increases of 1.1- to 7-fold due to urine applications have been recorded (During and McNaught, 1961; Thomas et al., 1988). Response is greatest in spring and autumn (Dale, 1961) and these are also the periods when maximum responses to fertilizer

NUTRIENT CYCLING UNDER GRAZED PASTURE

171

N occur in pastures (Ball and Field, 1982). Responses to urine N in midsummer and winter are generally restricted by environmental conditions (i.e., low soil moisture and low temperature, respectively) that are not conducive to rapid pasture growth. In grass/clover pastures the yield response comes almost entirely from the grass component of the sward (Dale, 1961; Ledgard and Saunders, 1982).Clover is a poor competitor with grass for mineral N (Haynes, 1981) and the consequent vigorous grass growth shades the clover and suppresses its growth. Thus, the pasture in the urine patch becomes grass dominant in comparison with the surrounding sward (Ball et al., 1979; Ledgard et al., 1982). Ball et al. (1979) noted that in plots receiving 0, 300, and 600 kg ha-' of urine N, the contribution of clover to total pasture yield over the 53 days following application was 48, 19, and 12%,respectively. Nitrogen fixation by the clover component of the sward is also markedly depressed in the urine patch (Ball et al., 1979; Ledgard et al., 1982).This is due to the inhibitory effects of high soil mineral N on N, fixation by the rhizobium bacteria. 2. UrineScorch

Following a urination event by sheep or cattle, scorching and/or death of pasture within the areas affectedby urine can sometimes occur. Scorch can occur sporadically throughout the year in both moist and dry conditions (Richards and Wolton, 1975) and the scorched symptoms can appear within 1 day of urine application. The scorching effect is thought to occur indirectly through an effect on the root system rather than directly on the foliage (Richards and Wolton, 1975). The effect probably occurs as a result of the detrimental effects of NHt toxicity coupled with the high salt content of the applied urine. There is a positive relationship between urine N concentration and the seriousness of urine burn (Quin, 1977; Groenwold and Keuning, 1988). For dairy cows, Groenwold and Keuning (1988) noted that urine scorch was closely correlated with the urea content of urine and trials with artificial urine showed that a concentration of 10 g urea liter-' had little effect, 15 g liter-' gave appreciable damage, and 20 g liter-' caused severe damage. Due to large variations in the N concentration and ionic strength of animal urinations (Section II1,A) not all urinations cause scorch, and damage can vary among animals. Cows with concentrated urine were shown by Groenwold and Keuning (1988) to cause more sward damage than those with weaker urine. Keuning (1980) and Groenwold and Keuning ( 1988) noted that during the first few hours of the morning, when ionic strength and N concentrations in the urine were higher than later in the

172

R. J. HAYNES AND P. H. WILLIAMS

day, urinations caused more scorch damage. Increasing fertilizer N rates, and therefore increasing sward N status, increases the frequency of urine scorch (Richards and Wolton, 1975; Keuning, 1980), presumably through increasing the N content of the urine. Decreasing the ionic strength and N concentration of urine by increasing the fluid intake of cattle decreases the damage due to scorch (Groenwold and Baarslag, 1982). Recovery of the pasture from scorch depends greatly on season. Where scorch occurred in spring at the center of urine spots, Dale ( 1961) found that rapid pasture growth occurred in an annular zone around the scorch and this usually grew over the spot, thus hiding it. In contrast, where urine scorch occurred during the summer, when pasture growth was restricted, some spots remained bare for 9 to 10 months. The result of pasture plant death caused by urine scorch is ingress of weed species. Richards and Wolton (1975) found that urine patches in which ryegrass plants had been killed were colonized by species such as Plantago major, Rumex obtusifolius, Sonchus oleraceous, and Poa annua. Similarly, in ryegrass swards, Keuning (1980) found the major consequence of urine scorch was deterioration of the botanical composition of the sward, with ingress of P. annua. 3. Chemical Composition of Herbage

Herbage growing in the urine patch during the first few months following a urination event has a higher N and K content compared to the surrounding herbage (Watkin, 1957; During and McNaught, 1961; Lotero et af., 1966; Joblin and Keogh, 1979; Joblin, 1981; Williams et af., 1989). At the same time, Ca, Mg, P, and sometimes Mn contents tend to be depressed (Watkin, 1957; Joblin and Keogh, 1979; Joblin, 1981). The increases in herbage N and K concentration found in urine patches are to be expected from localized application of urinary urea and K. The high concentrationsof soil solution K in the urine patch apparently tend to depress uptake of Ca and Mg by the herbage (Saunders and Metson, 1959; During and McNaught, 1961). In addition, the much greater pasture production, induced by added urine N, will cause a dilution effect on the concentration of other nutrients such as P, Ca, Mg, and Mn. Joblin (1981) and Joblin and Keogh (1979) observed lower herbage Mn concentrations in the urine patch, which may have been a result of localized high pH in the urine patch following urea hydrolysis. Recovery of urine N by pasture herbage is characteristicallylow, ranging from 8 to 55% (Holland and During, 1977; Ball et af., 1979; Carran et al., 1982; Ledgard and Saunders, 1982; Ball and Keeney, 1983; Thomas et al., 1988).Recoveries are often in the range of 20 to 30% for highly productive

NUTRIENT CYCLING UNDER GRAZED PASTURE

173

pastures but depend greatly on the environmental factors influencing pasture growth. Where rapid pasture growth is favored, recoveries are relatively high, but where growth is limited, recoveries are likely to be low. Holland and During (1977) found recoveries of urine N by ryegrass to be 30 to 40% in winter, spring, and autumn but they were less than 20% for summer applicationswhen dry conditions limited pasture growth. Ball and Keeney (1983) noted a similar trend. The recovery of urine K appears to be similar to that of urine N. During and McNaught ( I 96 1) found that over a 20 month period the apparent recovery of applied urine N in herbage was 10% and that for K was 20 - 23%;Williams et al. (I 989) found recoveries of 40% for K and 55% for N over an 8-month period. Cation uptake and translocation are both enhanced under NO; nutrition (Haynes, 1986c), thus high rates of NO; uptake by herbage in the urine patch may favor uptake of urine K as a counterion (Carran, 1988).In contrast, when nitrification is slow and NHZ remains the dominant mineral N form in the urine patch, competition between NHZ and K+ during plant uptake may limit K and/or N uptake by the sward. 4. Implication to Nutritional Disorders of Animals

The increased K and lowered Ca, Mg, and Na concentrations in urineaffected pasture have implications with respect to the prevalence of nutritional disorders in grazing animals (Joblin, 1981; Saunders, 1984). High K :(Ca Mg) ratios have been implicated in the incidence of hypomagnesemic grass tetany in lactating cows. This disorder is caused by a Mg deficiency in the animal brought about by many factors, including decreased Mg adsorption from the gut (Grace, 1983), a decreased Na: K ratio in the gut fluid, increased ammonia from ingestion of young pasture of high N status, and decreased available energy (Martens and Rayssiguier, 1980). In herbage affected by excreta the K :(Ca Mg) ratio can increase from 0.9 - 1.5 to 1.7 -2.4 (Saunders, 1984). Ratios greater than 1.6 are thought to be related to hypomagnesemic grass tetany (Metson et al., 1966; Kemp and t’Hart, 1967). The Na :K ratio of pasture in the urine patch areas is also reduced compared with unaffected herbage; Saunders ( 1984)observed a decrease in Na: K ratio from 0.06 to 0.04 in excreta-affected areas. In addition, the high N content of urine-affected pasture may lead to the production of high ammonia in the gut. The combination of these factors suggests that if the grazing pressure is high enough to ensure that stock eat significant quantities of herbage growing in areas affected by excreta, then metabolic disorders may occur.

+

+

174

R. J. HAYNES AND P. H. WILLIAMS

V. MODELING NUTRIENT CYCLING UNDER PASTURE A. MODELS OF NUTRIENT CYCLES 1. Major Pools and Fluxes

A simplified model of nutrient cycling under grazed pasture is shown in Fig. 13. This diagram shows the major pools of nutrients and the major flows between them and demonstrates the important role of the grazing animal in nutrient cycling. The soil component includes nutrients in plantavailable form, those held in soil organic matter, and those in chemically "fixed" forms. Nutrients (mainly N, S, and P) may be released from organic form through the microbial process of mineralization, but at the same time available nutrients may be immobilized in organic form by the

E Animal products

ANIMAL

Animal transfers

Decaying herbage

/G

ORGANIC MATTER

AVAILABLE NUTRIENTS

Leaching loss

CHEMICALLY FIXED NUTRIENTS

3

Other inputs

Other losses

Figure 13. Simplified nutrient cycle for grazed pasture ecosystems.

NUTRIENT CYCLING UNDER GRAZED PASTURE

175

actions of the soil microbial biomass. Available nutrients (K, P, and most micronutrients) may also be converted into chemically fixed forms by precipitation and adsorption reactions, but may be released from fixed forms via processes such as weathering, solubilization, and desorption. Nutrients are absorbed from the available soil pool by the growing pasture. Nutrients are translocated within the plant and are used in various metabolic processes. Subsequently, nutrients may undergo retranslocation and recycling within the plant. Under a grazed system, plant tops are consumed by the grazing herbivores and such recycling is much reduced. By ingesting herbage, grazing animals encourage pasture plants to grow and therefore take up more nutrients from the soil. The proportion of above-ground herbage that is consumed is dependent on stocking rate and can commonly reach 85% of that produced. Plant nutrients can be returned to the soil organic pool through senescing shoots and roots. Grazing animals can promote senescence by treading herbage and/or pulling herbage out of the ground. As discussed in Section III,A, a large percentage of nutrients (often over 90%)ingested by the grazing animals is returned to the pasture in the form of dung and urine. Nutrients in excreta can be in organic and/or inorganic form, depending on the particular nutrient in question. Thus, returned nutrients may be in readily plant-available forms or in forms that require mineralization before they are available for plant uptake. Small amounts of nutrients are removed in the form of animal products (e.g., milk, wool, or live animals.) The animals also give rise to nutrient losses through transfer of nutrients to camp sites (Section 111,C,2) and unproductive parts of the farm such as races and stock yards. Once nutrients are returned to the soil, losses can occur through leaching, gaseous emissions, and runoff and erosion. Leaching and gaseous losses occur preferentially from excretal patch areas because of the aggregation of nutrients in these areas of the pasture (Section IV). Because of the losses of nutrients during cycling (i.e., animal products, animal transfer losses, and leaching and gaseous losses in excretal patches), fertilizer inputs are required to maintain soil fertility. Additional inputs of nutrients can occur via irrigation water, rainfall, and dry deposition. In grass/legume pastures a significant input of N occurs through biological N, fixation. 2. Types of Models

There are two principal types of models that quantitatively describe nutrient cycles- mass balance and dynamic models. Mass balance models are balance sheets of the amounts of nutrients within various compart-

176

R. J. HAYNES AND P. H. WILLIAMS

ments (pools) of the system and the amounts of nutrients entering and leaving the system. Such balance sheets are relatively easy to construct because they involve measuring the nutrient content of the various components of a system and the various inputs and losses from the system. Balance sheets also identi@ the major processes of transfer within the system and can be used to estimate the rate of net transfer of nutrients between compartments. The use of radioactive and/or heavy isotopic tracers is a valuable aid in nutrient balance work. In field experiments, labeled material (e.g., fertilizer, plant material, or excreta) can be introduced into the grazing system and its appearance and loss from various compartments can be measured (Till, 1981). Isotopic tracers are particularly useful in tracing the pathway and rate of transfer of nutrients between compartments (May et al., 1972). Construction of mass balance nutrient budgets can often identify portions of a nutrient cycle when data are lacking or imprecise. Simple input/ output models are increasingly being used to predict fertilizer requirements for agricultural systems. The major disadvantages of these models are that they tend to be site specific, interrelated processes are often combined together, and data are aggregated over time and space. Often, the balance sheet is constructed on an annual basis and this can obscure multiple recycling, which may be occurring within this time scale. In reality, nutrient cycles are dynamic systems in which there are many strong interactions between the various components. Many of the transformations within the cycles are reliant on biological processes and are therefore affected greatly by environmental factors. Thus, complex dynamic nutrient models have been constructed to model the system more realistically. Dynamic models differ from simple mass balance models in that they include equations describing the feedback control mechanisms that affect nutrient transformations. With the aid of computers to carry out calculations, dynamic models can make predictions based on small time periods (Le., days or hours). It is therefore possible to incorporate the effects of daily and weekly changes in climate into the model. Thus, dynamic models can be used to describe a nutrient cycle that exists in a range of environmental conditions. Such models can be used to investigate the sensitivity of the cycle to management and environmental variables and predict the consequences of changes in these factors (e.g., changes in stocking rate or fertilizer applications). 3. Mass Balance Data

Examples of values for the major fluxes of N, P, S, and K identified in Fig. 13 for various farm types are presented in Table XIII.From these data it is clear that annual plant uptake of nutrients generally greatly exceeds the

NUTRIENT CYCLING UNDER GRAZED PASTURE

177

annual nutrient inputs from fertilizer (and N2 fixation in the case of N). Furthermore, not all the annual additions of fertilizer nutrients are likely to be used immediately by the pasture plants, because fertilizer use efficiency is generally around 60%. Thus, if the systems are in equilibrium (i.e., the soils are not gaining or losing nutrients), then large quantities of nutrients must be cycling within the system. Indeed, nutrient returns in dung and urine are generally of the same order as fertilizer inputs. It is also evident from Table XI11 that in intensively managed systems, a large proportion of nutrients absorbed by the pasture is ingested by the animals, whereas in the sparsely stocked systems, a greater proportion of nutrients is returned via plant litter. This has implications for the rate of nutrient cycling. As noted in Section III,B, significant mineralization of organic plant P and S occurs during passage through the animal and, in addition, much of the N is excreted as urea, which is rapidly converted to the readily plant-available NHt and NO; forms. Thus, the N, S, and P returned in animal excreta are more rapidly available to the pasture plants than are the N, S, and P in decaying plant residues. As a result, the activity of grazing animals accelerates the cycling of these nutrients within the system. A more detailed balance sheet for P, S, and K on an Australian sheep farm is shown in Table XIV. Again, these data demonstrate that in the absence of nutrient recycling, fertilizer inputs are insufficient to provide plant nutrient requirements, even if there was 100% fertilizer use efficiency. The available soil pools of S, P, and K, represent about 20,60, and 18096, respectively, of annual pasture uptake (Table XIV). The Colwell method for estimating available P extracts considerably more P than other commonly used methods (e.g., Olsen and resin-extractable P), so that the 60% value can be considered an atypically high figure. If the system is at equilibrium, the average size of the available pool (as well as the total soil pool) will remain relatively constant from year to year. Hence, in order for the available pools of S and P to supply the growth requirements of pasture plants they must continually be supplied by the cycling of nutrients via animal excreta. These nutrients may cycle via organic or inorganic pathways, depending on the form in which they are excreted. The more intensively grazed the system, the larger the cycling component and the less relevant the extractable pool of soil nutrients is to the system. Thus, predicting the fertilizer requirements of pastures based on soil tests is a particularly difficult task. The proportions of S, P, and K in the root zone in organic form range from about 90% for S to 40% for P and virtually zero for K. Thus, the relative importance of the organic fractions in supplying nutrients to the available pool varies considerably between different nurients. For S, and to a lesser extent P, the cycling of nutrients through the organic fractions is

Table XIII Examples of the Major Fluxes of N, P,I(,and S (kg ha-') in Various Farm Typesa

c

Reference

Farm-type

Ball and Field (1982)

Intensively managed daily farm Improved

Lambert et al. ( 1982)

-4 01

Quin (1982) Lambert et al. ( 1982) Ineson (1987)

Parfitt ( 1980)

A Fermizer inputs

Nitrogen 448N

F G B C D E Animal Animal J K Other Plant Litter Animal product transfer H I Leaching Gaseous inputs accumulation return ingestion removal losses Feces Urine loss IOSS 96

1132

550

521

61

0

135

403

240

163

0

180

360

90

0

17

143

0

0

Phosphate 38P

0

276

89

102

190

ND

9

ND'

34

120

14

I1

280

20

30

90

140

80

40

102

41

4

ND

13

24

10

4

162

110

52

3

13

9

13

ND

ND

57

26

31

10

ND

13

ND

0

0

pasture,

sheep farm Intensively managed sheep farm Unimproved hillp a s t a , sheep farm Unimproved hill pasture, sheep farm Intensively managed dairy farm

During ( 1981)

Intensively managed dairy farm Quin and Rickard Intensively (1981) managed sheep farm Harrison (1 985) Unimproved hill pasture, sheep farm During (198 1)

U

\o

Intensively managed dairy farm Quin and Rickard Intensively (1981) managed sheep farm Till and Blair Sheep farm (1974) Williams (1988) Williams (1988)

Intensively managed dairy farm Highly productive, intensively managed dairy farm

20P

0

40

8

32

7

13

20

ND

0

0

20P

0

32

9

26

2

4

24

ND

0

0

16P

0

16

11

5

0.8

1.2

5

ND

0

0

Sulfur 22s

0

30

5

26

3

3

8

12

15

0

22s

0

36

9

29

6

ND

10

13

25

0

23s

0

25

4

21

2

ND

10

10

ND

0

Potassium 0

0

300

70

230

13

39

18

160

57

0

63K

0

615

61

554

19

25

51

459

66

0

Letters above fluxes refer to the flux diagram shown in Fig. 13. ND, Not determined.

180

R. J. HAYNES AND P. H. WILLIAMS Table XIV Sulfur, Phosphorus, and Potassium Budgets for a Wool Growing System Given Maintenance Applications of Superphosphate Pool/flow

P

S

Fertilizer (kg ha-' yr-I) Soil (kg ha-') Total Organic Availableb Plant (kg ha-') Clover Phalaris Roots Dead matter Total (kg ha-' yr-I) Domestic animals Intake (kg ha-' yrl) Body (kg ha-') Product (kg ha-' yr-I) Urine (kg ha-l yr') Feces (kg ha-' yi') Other consumersb Coleoptera larvae (kg ha-') Oligochaete, large (kg ha-') Diplopoda (kg ha-') Total (kg ha-I y r l ) b

22

28

50

310 160

250 140 10

46000

_

_

_

30

K

-

720

4.4

4.1

21

6.6 1.2 52

6.4

41 400 400

1.1 50

-

20 1.1 1 11

130 1.5

23

8

15

23

I

0.9 0.5 0.2

90

115

0.3 0.2 0.1 65

~

Data from Till (198 1). Available soil P determined by bicarbonate extraction (Colwell, 1963); S determined as phosphateextractable sulfate and K as exchangeable K.

likely to be extremely important. In a pasture soil, with its characteristically high organic matter content and high rate of organic matter turnover, this makes reliance on soil tests for fertilizer advice even more questionable. For example, the available S q - sulfur pool in the soil is being continually supplied by turnover of soil organic S. If the rate of release of S from organic matter equals that of plant uptake, then S a - sulfur in the soil can remain virtually zero without any sulfur stress in pasture occurring. In contrast to S, large amounts of K are contained in soil minerals, virtually none is in organic forms, and a very small proportion is in readily exchangeable form (Table XIV).In soils with a significant content of 2 : 1

NUTRIENT CYCLING UNDER GRAZED PASTURE

181

clay minerals, thus containing a significant amount of “clay-fixed’’ K (Sparks and Huang, 1985), the supply of K from this buffering reserve of mineral K can be extremely important. The system is often nowhere near equilibrium because there is a net supply of K from the mineral fraction. Thus, in Tables XI11 and XIV, plant uptake of K greatly exceeds the fertiliser input. Date in Table XIV also show that the annual intake of P and S by other consumers exceeds that of the domestic herbivores. This “other consumer” pool represents an extremely complex grouping of microflora and fauna whose major activity is decomposition of organic matter. The amounts of nutrients in some of the faunal pools are shown in Table XIV. Unlike the pools of nutrients held in domestic herbivores, the pools in other consumers can be of great importance due to the very short life cycles of many of these groups. As noted in Section II,C,3, the amounts of N and P held in the soil microbial biomass under a typical improved pasture may approximate 60 kg N and 30 kg P per hectare. Mass balance data can be used to construct more complex nutrient cycling simulation models. Scholefield el al. (199 l), for instance, constructed a mass balance model of the N cycle in United Kingdom pastures. The model was constructed using 10 experimental field systems; a mineralization submodel that is sensitive to soil texture, sward age, previous cropping history, and climatic zone was included. Another submodel determines the partition of soil inorganic N to either plant uptake or to processes of loss (i.e., leaching, denitrification, and ammonia volatilization). A full-output model for a fertilizer input of 400 kg N ha-’ yr-‘ in conjunction with poor (on the left) or good (on the right) drainage is shown in Fig. 14. As expected, the rates of mineralization and leaching losses were greatest from the well-drained site. In contrast, denitrification losses were greatest from the poorly drained site. Such a model is of use in estimating losses of N from a given site at a known fertilizer input and thus estimating fertilizer use efficiency as well as the environmental impact of various practices. 4. Dynamic Models

Very few dynamic models of nutrient cycles under grazed pasture have been developed. However, often parts of a nutrient cycle are modeled and then these are put together to form a model of a larger system. Many dynamic models have been developed for parts of arable farming systems but could also be applicable to the pasture system. For example, for N, dynamic models have been developed (Tanji, 1982) of various components of the N cycle, such as transformations and transport of soil N,

R. J. HAYNES AND P. H. WILLIAMS

182

Fertilizer 400 400 Atmosphere Urine 2 9 5 323

Dung 69 76

i

I

I

Volatilization

1

Plant Uptake 6 4 5 704 Denitrification

2 4 5 267

Organic

Inorganic 864 992

Mineralization 1AA

9aa

Figure 14. Full output from a model of N cycling in pastures grazed by beef cattle at a fertilizer rate of 400 kg N ha-' yr' in combination with poor drainage (values on the left) and good drainage (values on the right) on a clay loam soil.(Redrawnfrom Scholefield et a/., 1991, with permission of Kluwer Academic Publishers.)

transformation and plant uptake of soil N, turnover of N through the soil microbial biomass, leaching of nitrate during the period of crop growth, and influence of fertilizer and daily climate on plant growth and N uptake. In addition, Smith and Langlands (1976) developed a dynamic model of N movement through the grazing animal. May et al. (1972, 1973) went some way toward developing a dynamic S cycling model under pasture. Based on a series of experiments in which 35S-labeledgypsum was applied to a pasture grazed with Merino sheep, they proposed a simplified site-specific S cycling model. The model was developed for a closed system that precludes losses of S via leaching, animal transfers, and animal products. It does, however, allow for the effects of seasonal and climatic changes on the S cycle, differential grazing by sheep, foliar absorption of sulfate from fertilizer application, adherence of gypsum particles to ingested plant material, and increased plant S content after fertilizer applications. May et al. (1972) used simulation runs to estimate some transfer rates within the model that could not be directly measured by other techniques.

NUTRIENT CYCLING UNDER GRAZED PASTURE

183

A dynamic model of the P cycle in a semiarid ungrazed grassland in North America was developed by Cole et al. (1977), and Blair et al. ( 1977) developed a dynamic model of P cycling for grazed pastures in New South Wales, Australia. The structure of the latter model is presented in Fig. 15, and equations describing flow rates are given in the study by Blair et al. ( 1977). The model brought together information from a variety of sources and it was used to simulate a grazing system over a 20-year period. A comparison of outputs with the results of a long-term grazing experiment showed that the model produced biologically valid data. By changing parameters in the simulation runs the model was used by the authors to identify sensitive areas of the P cycle and highlight priority areas for further research. In the light of over a decade of further research it would be useful to revise and update this model because it has already been shown to model adequately the grazed pasture system. It could also be used as the basis for dynamic models of the cycling of nutrients other than P.

B. USEOF MODELSFOR FERTILIZER RECOMMENDATIONS Phosphorus is the major limiting nutrient to pasture production in Australasia and consequently several computer-based models have been developed to aid in estimating the P requirements of individual paddocks. In Australia several systems have been formulated based on the construction of a fertilizer response curve for a particular farm system and then economic criteria are used to determine optimum fertilizer rate (Bennett and Bowden, 1977; Helyar and Godden, 1977; Cameron, 1987). The response curve used is the exponential Mitscherlich function:

Y = A[ 1 - B exp(- CP)] where Y is pasture yield, P is the quantity of elemental P applied as fertilizer, and A, B, and C are constants such that A is pasture yield at that site with P nonlimiting, B is the relative response to applied nutrient, and C is the curvature coefficient of the response curve. On-farm factors and the experience of local advisors and farmers are used to estimate values for the parameters A, B, and C for the particular paddock or farm being considered. A is determined mainly by annual rainfall, soil type, pasture species, and management practices; C is determined by soil type, fertilizer source, grazing management, and pasture species and B is usually derived from soil test data and/or previous fertilizer use history. Once the response function has been estimated, economic theory is applied to obtain the point of maximum profit (where marginal

-

SOIL MOISTURE MODEL PASTURE GROWTH MODEL

ANIMAL INTAKE AND GROWTH MODEL

Figure 15. Schematic model of the structure of a dynamic phosphorus cycle for a pasture grazed with sheep. (Redrawnfrom Blair er aL, 1977.)

NUTRIENT CYCLING UNDER GRAZED PASTURE

185

costs equal marginal returns) on the curve. This point is then compared with B and the current requirement is calculated by difference. An additional factor (a maintenance requirement curve) was added to the basic model by Helyar and Godden (1977). This takes into account the fact that pasture yields will approach a steady state with repeated applications of fertilizer at a given rate. At that equilibrium level, fertilizer applications just balance nutrient losses from the system. Such models have considerable advantages over conventional systems where soil test values are simply used to give a fertilizer recommendation. A major drawback with relying almost entirely on soil test values is the large spatial variability in values within a given paddock. Friesen and Blair (1984), for example, found that within a paddock the coefficient of variation for P soil test values ranged from 2 1 to 90%. Such spatial variability is particularly marked in grazed paddocks because the P is returned to the pasture in discrete areas via dung pats. In addition, as noted in Section V,A,3, the pool of cycling P may contribute as much P to the annual pasture growth as the available P pool measured by a soil test at any one time. In conventional recommendation systems that rely on soil test values, the same recommendation is given for all farms with a common soil type and soil test value in a given locality. Because the models take account of on-farm factors, particularly relating to the number and type of grazing animals, site-specific fertilizer recommendations are given. In addition, the use of models allows the advisor to put an economic value on pasture and thus make an economic evaluation of the optimum fertilizer rate. An alternative approach to P fertilizer advice has been taken by Cornforth and Sinclair (1982), who developed a simplified model of the P cycle to calculate the maintenance P requirement of a well-developed pasture maintained at a steady level of production by a constant annual application of fertilizer P. Maintenance requirements are determined by estimating the amount of P required to replace losses in the soil (through incorporation into unavailable inorganic and organic forms) by animal transfer and animal products (Fig. 16). The model and a full list of information required for its use are presented in Table XV. The soil loss factor (SLF) refers to P that is effectively lost from the cycling pool when inorganic and organic P compounds that are unavailable to plants accumulate in the soil. Soils are grouped into three P loss categories: low, medium, and high. The animal loss factor (ALF) refers to loss of P in animal products and in excreta not returned to the main pasture area. P losses increase with land slope because of increased camping and tracking. Stocking rate (SR)refers to a standard stock unit (su), which corresponds to an ewe that is 55 kg at mating producing one lamb

186

R. J. HAYNES AND P. H. WILLIAMS P lost via animals (produce and excreta)

t

P lost in soil (inorganic and organic forms)

Figure 16. Simplified P cycle that forms the basis for estimating maintenance P requirements for grazed pastures, according to Cornforth and Sinclair (1982).

per year. Various classes of stock can be expressed as stock units (e.g., Jersey cow = 6 su, Angus beef cow = 4.8 su). Potential carrying capacity (CC) is defined as the number of stock units that would be carried per hectare when pasture utilization is 90% and pasture production is maintained at 95% of the maximum achievable with unlimited fertilizer P.

Table XV Model for Calculating Phosphorus Requirements for Pasture Maintenance" Information required SLF, soil P loss factor (soil group) ALF, animal P loss factor (stock type, landform) SR, stocking rate CC, potential stock carrying capacity PU, pasture utilization Recent fertilizer history soil P status (Olsen P) Model Overall maintenance P requirement (kg ha-') = log,, (loo/( 100 - (8550X SR)/(CC X PU)]) X CC X (0.005 X CC 0.275) (PU X ALF X 0.0301 SLF X 5.79)

+

+

Compiled from Cornforth and Sinclair (1982).

NUTRIENT CYCLING UNDER GRAZED PASTURE

187

Pasture utilization (PU) is the pasture eaten in the course of a year expressed as a percentage of the total grown in the year. It can vary from 90% in intensive rotationally grazed systems to less than 50% in extensive high-country farms. There are three main parts to the calculations of P loss for a specific calculation. First, the P loss is calculated for the system when pasture production is at 90% of maximum. Second, the required amount of pasture for the specific situation is determined based on stocking rate and pasture utilization. Finally, the total P loss associated with this required level of production is deduced. This involves application of the Mitscherlich equation to calculate P loss at the required pasture yield from P loss at 90% of maximum yield. The overall equation for maintenance P is shown in Table XV. The model predicts maintenance fertilizer P requirements only. Soils may contain more or less P than is appropriate for the required yields so that short-term adjustments to maintenance rates can be made based on soil test values. The Cornforth-Sinclair (1982) model has the advantage of being a simplified mechanistic model of P cycling that incorporates the main on-farm factors that are known to influence the P economy of pastures. It is well suited to the permanent pasture land of New Zealand, which has more or less reached a steady state. The fact that it takes into account the role of the animal is particularly important on hill country pastures because on these pastures stock camping has a large influence on soil fertility. The model has been computerized and is used routinely for pastoral fertilizer advice in New Zealand. However, the model is not well-suited for mixed-crop farming, where pasture is alternated with arable land, because in such situations steady-state conditions do not apply. In addition, the model does not yet have the facility to provide an economic analysis of the value of fertilizer applications. Maintenance S recommendations for New Zealand pastures are also made based on a model similar to that described above for P (Sinclair and Saunders, 1981). The information required for its use is shown in Table XVI. The model allows for inputs of S from the atmosphere, imgation water and plant uptake from the subsoil. Outputs of S from the cycle include leaching, immobilization of S in allophanic soils, removal of animal products and loss via transfer of excretal S to unproductive areas of the farm. Information additional to that required for the P model includes distance from the coast, and a sulphate leaching index value based on a combination of sulphate adsorption capacity, annual rainfall and soil drainage status. As with the P model, the fertilizer recommendation can be modified on the basis of a soil sulfate test. The K recommendations are made on the basis of a balance sheet of the

188

R. J. HAYNES AND P. H. WILLIAMS Table XVI Model for Calculating Sulfur Requirements for Pasture Maintenan& Information required A, stocking rate B, animal S loss factor (stock type, landform, sulfate leaching index) C, end-of-season leaching (sulfate leaching index) D, immobilization (important only in allophanic Soils) E, distance from coast F, rainfall G, fraction of rainfall sulfate leached (calculated from sulfate leaching index) H, sulfur added in imgation water (S concentration in water, depth of water applied) Model Overall maintenance S requirement = (AB) C D - {[ 12 - (OSE)](F- 1OOO)( 1 Q1-H

+

+

Compiled from Sinclair and Saunders ( 1981).

inputs of K from the soil and outputs of K from the farm via animal products, transfer of excretal K to unproductive areas, and leaching (Campkin and Cornforth, 1984). Because of the importance of soil K, soil test results are used in the calculation of the fertilizer requirements, rather than as a modifying factor as in the P and S models. Potassium supply is determined by a measurement of both plant available K (exchangeable K) and an index of long-term supply of K by nonexchangeable(fixed) sources. This model is depletive rather than being a true maintenance model because a supply of exchangeable and nonexchangeable K is used to partially offset the losses of K that occur. Ultimately, this supply will deplete the amount of K in the soil. The model tends to underestimate the contribution of soil K to the system but also underestimates losses of K (Williams et al., 1990d). The above discussion has shown how simplified mass balance nutrient budgets can be used to estimate maintenance fertilizer requirements of pastures. The simplicity of such models means that the amount of information required as inputs is minimized and that the principles of the models are readily understood by advisors and farmers. However, simplicity also brings limitations, and dynamic models such as that developed by Blair et al. (1977) for P are likely to be more adaptable to different farming

NUTRIENT CYCLING UNDER GRAZED PASTURE

189

and environmental conditions and to situations where steady-state conditions are not occurring. Such dynamic models would allow on-farm input parameters to be varied and their effects on production and fertilizer requirements to be predicted. In conjuction with economic data related to farm inputs and outputs, these models could become extremely powerful tools in linking farm management practice and fertilizer advice together.

VI. SUMMARY AND CONCLUSIONS The improved grass and herb species used in highly productive pastures are generally adapted to high-fertility soil conditions. As a consequence, fertilizer applications are often required during pasture development in order to boost natural soil fertility to a level capable of supporting these species. Once an adequate level of fertility has been reached, maintenance fertilizer requirements are generally required to sustain a high level of production. Nonetheless, annual nutrient uptake by the pasture is often considerably greater than the maintenance fertilizer application needed to sustain high production. The reason for this is that large quantities of nutrients are cycled within the ecosystem through the actions of the grazing animals. By ingesting herbage, animals encourage pasture plants to grow and therefore take up more nutrients from the soil. Approximately 60-90% of the nutrients in herbage ingested by the animals is returned to the pasture in the form of urine and dung. Urine and dung patches are therefore the areas where nutrients are recycling from excreta to soil and back to pasture plants. Although excretal patches may cover only 30-40% of the pasture surface annually, the high nutrient input stimulates herbage growth that may represent 70% of the annual pasture production. Thus, nutrient transformations in the excretal patch areas are of central importance to the fertility and productivity of grazed pastures. As well as being responsible for the cycling of nutrients within the system, animals are also the major agents for nutrient losses. Recycled nutrients are returned to small volumes of soil (in urine and dung patches) in quantities that generally exceed the immediate needs of the pasture plants. Consequently, nutrients accumulate in these soil volumes and leaching and gaseous losses occur preferentially from these areas of pasture. Grazing animals also give rise to nutrient losses from the main pasture area through transfer of nutrients in excreta to camp sites (around hedges, troughs, and ridges and on hill crests on hill country farms) and unproductive parts of the farm (raceways and stockyards). In addition, small quantities of nutrients are removed in animal

190

R. J. HAYNES AND P. H. WILLIAMS

products (milk, fiber, and live animals). These animal-induced losses are the major reason that maintenance fertilizer applications are required in order to sustain pasture production. Nutrients are partitioned differently between dung and urine, with K being excreted mainly in urine; P, Ca, and Mg are excreted principally in dung and N and S are excreted in both forms. Nutrients returned in excreta can be in inorganic and organic forms, depending on the particular nutrient being considered. For some nutrients (e.g., P and S) significant mineralization of ingested organic forms occurs during passage through the animal, whereas much of the N is excreted in the readily available organic urea form. Major nutrient transformations that occur in the urine patch have been documented. Urea hydrolysis (with concomitant NH, volatilization losses) has been extensively studied and subsequent nitrification has been followed. Subsequent losses of N through NO? leaching and gaseous emissions of N, and N 2 0 (through nitrification and denitrification) are known to occur. However, the extent of such losses and factors affecting them (particularly gaseous emissions) are not well known. Leaching losses of SO;- are also likely to occur. The release of nutrients from dung has been studied and K and Na have been shown to be released rapidly whereas Ca, Mg, and P, which are present in less soluble forms, are released more slowly. Under moist humid climates, physical degradation of dung pats is the limiting factor to nutrient release, but under dryland conditions, leaching of nutrients from pats becomes the major mechanism of release. A major area where knowledge is lacking relates to the pathways and rates of movement of nutrients deposited in urine and dung through various soil pools and back to the plant. Nutrients deposited in inorganic form may well cycle through labile organic pools and subsequently become available for plant growth. An understanding of these pathways is important because, if cycling efficiency can be increased, losses will be decreased and hence environmental pollution will be lowered and maintenance fertilizer requirements will also be lowered. Thus, the more efficiently that nutrients are recycled within the system the more sustainable the system will be. From our present knowledge of the major nutrient inputs and losses for the grazed pasture system and an understanding of the pathways of nutrient flux within the system and some key measurements, simple mass balance nutrient models for various pasture systems can be constructed. Such simple models have been used to calculate site-specific maintenance fertilizer requirements of pastures based on the amount of nutrient required to replace losses in the soil (e.g., through fixation and leaching) and losses by animal transfer and in animal products.

NUTRIENT CYCLING UNDER GRAZED PASTURE

191

Little work has concentrated on the development of dynamic models of nutrient cycling under grazed pasture. Development of such models is hampered by uncertainty regarding the pathways and rates of cycling of nutrients from excreta through various soil pools and back to the pasture plants. Such uncertainty makes the development of rate equations for various fluxes through the soil compartments rather speculative. Such models would, however, be extremely powerful tools both from the standpoint of farm management and fertilizer advice and for developing management options to maximize cycling, minimize losses, and produce more sustainable pastoral systems.

REFERENCES Anderson, J. M., and Coe, M. J. 1974. Oecologia 14, 11 1- 125. Baker, R. D. 1986. In “Grassland Manuring” (J. P. Cooper and W. F. Raymond, eds.),Br. Grassl. Soc. Occas. Symp. 20, pp. 15-27. British Grassland Society, Hurley, Maidenhead, England. Ball, P. R., and Field, T. R. 0. 1982. In “Nitrogen Fertilisers in New Zealand Farming” (P. B. Lynch, ed.), pp. 45-64. New Zealand Institute of Agricultural Science, Wellington, New Zealand. Ball, P. R., and Keeney, D. R. 1983. Proc. Int. Grassl. Congr., 14th pp. 342-344. Ball, P. R., and Ryden, J. C. 1984. Plant Soil 76,23-33. Ball, P. R., Keeney, D. R., Theobald, P. W., and Nes, P. 1979. Agron. J. 7, 133- 175. Barlow, E. W. R. 1972. Master of Rural Science thesis. University of New England, Armidale, Australia. Barrow, N. J. 1960. Aust. J. Agric. Res. 11, 33 1- 338. Barrow, N. J. 1961. Aust. J. Agric. Res. 12, 644-650. Barrow, N. J. 1967. J. Aust. Inst. Agric. Sci. 33,254-262. Barrow, N. J. 1969. Aust. J. Exp. Agric. Anim. Husb. 9, 437-444. Barrow, N. J. 1975. Aust. J. Soil Res. 13,63-67. Barrow, N. J. 1987. In “Managed Grasslands” (R. W. Snaydon, ed.), pp. 181-186. Elsevier, Oxford, England. Barrow, N. J., and Lambourne, L. J. 1962. Aust. J. Agric. Res. 13,461 -471. Barry, T. N. 1983. In “The Mineral Requirement of Grazing Ruminants” (N. D. Grace, ed.), N. Z. Soc. Anim. Prod. Occas. Publ. 9, pp. 48-55. New Zealand Society of Animal Production, Hamilton, New Zealand. Bastiman, B., and Van Dijk, J. P. F. 1975. Exp. Husb. 28,7-17. Bathurst, N. 0. 1952. J. Agric. Sci. 42,476-478. Batten, G. D., Blair, G. J., and Lill, W. J. 1979. Aust. J. Soil Res. 17, 163- 175. Bennett, D., and Bowden, J. W. 1977. In “Reviews in Rural Science 111” (G. J. Blair, ed.),pp. 77 - 8 I. University of New England Press,Armidale, Australia. Berg, P., Klemedtsson, L., and Rosswall, T. 1982. Soil Biol. Biochem. 14,301 -303. Betteridge, K., Andrewes, W. G. K., and Sedcole, J. R. 1986. J. Agric. Sci. 106,393-404. Bird, P. R. 1971. Aust. J. Biol. Sci. 24, 1329-1339. Bird, P. R. 1972. Aust. J. Biol. Sci. 25, 1073- 1095. Bird, P. R., and Hume, I. D. 1971. Aust. J. Agric. Res. 22,443-452.

192

R. J. HAYNES AND P. H. WILLIAMS

Blair, G. J., Till, A. R., and Smith, R. C. G. 1977. In “Reviews in Rural Science III” (G. J. Blair, ed.), pp. 9- 19. University of New England Press, Armidale, Australia. Blanchar, R. W. 1986. In “Sulphur in Agriculture” (M. A. Tabatabai, ed.),pp. 455-490. American Society of Agronomy, Madison, Wisconsin. Bornemissza, G. F. 1970. J. Aust. Entomol Soc. 9,31-41. Boswell, C. C. 1971. Master of Philosophy thesis. University of Reading, Reading, England. Boswell, C. C. 1983. Ph.D. thesis. University of New England, Armidale, Australia Bouma, J. 1981. Agric. Water Manage. 3,235-250. Braithwaite, G. D. 1976. J. Dairy Res. 43, 501-520. Bray, A. C., and Hemsley, J. A. 1969. Aust. J. Agric. Res. 20, 759-773. Bremner, J. M., and Blackmer, A. M. 1978. Science 199,295-296. Bristow, A. W., and Janis, S. C. 1991. J. Sci. Food Agric. 54,9-21. Bromfield, S. M. I96 1. Aust. J. Agric. Res. 12, I 1 1 - 123. Bromfield, S. M., and Jones, 0. L. 1970. Aust. J. Agric. Res. 21,699-71 1. Cameron, I. H. 1987. In “Temperate Pastures: Their Production, Use and Management” (J. L. Wheeler, C. J. Pearson, and G. E. Robards, eds.), pp. 137- 139. Australian Wool Corporation/Commonwealth Scientific and Industrial Research Organization Canberra, Australia. Campkin, R., and Cornforth, I. S. 1984. In ‘‘ Fertiliser Recommendations for Pasture and Crops in New Zealand (I. S. Cornforth and A. G. Sinclair, eds.),pp. 16- 17. Ministry of Agriculture and Fisheries, Wellington, New Zealand. Carran, R. A. 1988. N.Z. J. Agric. Res. 31,65-69. Carran, R. A., Ball, P. R., Theobald, P. W., and Collins, M. E. G. 1982. N.Z. J. Exp. Agric. 10,377-381. Castle, M. E., and MacDaid, E. 1972. J. Br. G r a d SOC.27, 133- 137. Castle, M. E., Foot, A. S., and Halley, R. J. 1950. J. Dairy Res. 17, 215-229. Clarke, R. M., and Warburton, D. J. 1982. “Report for National Water and Soil Conservation Organization,” p. 88. Agricultural Engineering Department, Massey University, Palmerston North, New Zealand. Clarke, A. L., Greenland, D. J., and Quirk, J. P. 1967. Aust. J. Soil Res. 5,59-68. Clement, C. R. 1961. J. Br. Grassl. SOC.16, 194-200. Cole, C. V., Innis, G. S., and Stewart, J. W. B. 1977. Ecology 58, 1 - 15. Colwell, J. D. 1963. Aust. J. Exp. Agric. Anim. H u b . 3, 190- 197. Cornforth, I. S., and Sinclair, A. G. 1982. N.Z. J. Exp. Agric, 10, 53-61. Cregan, P. D., Sykes, J. A., and Hymock, A. J. 1979. Agric. Gaz. N.S. W. 90,33-35. Curry, J. P. 1987a. Grass Forage Sci. 42, 325-341. Curry, J. P. 1987b. Grass Forage Sci. 42, 103- 120. Dale, W. R. 196 I. Proc. N Z. Grassl. Assoc. 13, 1 I8 - 124. Davies, E. B., Hog& D. E., and Hopewell, H. G. 1962. Trans. Int. SOC.Soil Sci., Palmerston North, N.Z. pp. 715-720. Dickinson, C. H., and Underhay, V. H. S. 1977. Trans. Br. Mycol. Soc.69,473-477. Doak, B. W. 1952. J. Agric. Sci. 42, 162- 171. Donald, C. M. 1965. Aust. J. Sci. 27, 187-198. Donald, A. D., and Leslie, R. T. 1969. Parasitology 59, 141 - 157. During, C. 198 1, N.Z. Soil News 29,69- 153. During, C. 1984. “ Fertilisers and Soils in New Zealand Farming.” Government Printer, Wellington, New Zealand. During, C., and McNaught, K. J. 196 I . N.Z. J. Agric. Res. 4, 591 - 605. During, C., and Weeda, W.C. 1973. N.Z. J. Agric. Res. 16,423-430. During, C., Weeda, W. C., and Dorofaeff, F. D. 1973. N.Z. J.Agric. Res. 16,43 1-438.

NUTRIENT CYCLING UNDER GRAZED PASTURE

193

Evans, J., Hochman, Z., OConnor, G. E., and Osborne, G. J. 1988. Aust. J. Agric. Res. 38, 605 -6 18. Field, T. R. O., and Ball, P. R. 1981. Proc. N.Z. Grassl. Assoc. 43,64-69. Firestone, M. K., Smith, M. S., Firestone, R. B., and Tiedje, J. M. 1979. Soil Sci. SOC.Am. J. 43, 1140- 1144. Floate, M. J. S. 1970a. J. Br. G r a d Soc. 25, 295-302. Floate, M. J. S. 1970b. Soil Biol. Biochem. 2, 173-185. Floate, M. J. S. 1970~.Soil Biol. Biochem. 2, 187- 196. Floate, M. J. S. 197Od. SoilBiol. Biochem. 2, 275-283. Floate, M. J. S. 1981. In “Terrestrial Nitrogen Cycles: Processes, Ecosystem Strategies and Management Impacts” (F. E. Clark and T. Rosswall, eds.), pp. 585-601. Ecological Bulletins, Stockholm. Floate, M. J. S. 1987. In “Ecosystems of the World 17B. Managed Grasslands, Analytical Studies” (R. W. Snaydon, ed.), pp. 163- 172. Elsevier, Amsterdam. Floate, M. J. S.,and Torrance, C. J. W. 1970. J. Sci. Food Agric. 21, 1 16 - 120. Frame, J. 1971. Scott. Agric. 10,28-44. Friesen, D. K., and Blair, G. J. 1984. Aust. J. Soil Res. 22, 8 1-90. Galbally, I. E., and Roy, C. R. 1978. Nature (London) 275,734-735. Garstang, J., and Mudd, C. H. 1971. J. Br. Grassl. SOC.26, 194. Garwood, E. A., and Ryden, J. C. 1986. In “Nitrogen Fluxes in Intensive Grassland Systems” (H. G. van der Meer, J. C. Ryden, and G. C. Ennik, eds.), pp. 99-1 13. NijhofF, Dordrecht, The Netherlands. Garwood, E. A., Tyson, K. C., and Clement, C. R. 1977. “G. R. I. Technical Report No. 2 1.” The Grassland Research Institute, Maidenhead, Berkshire, England. Gemtse, R. G. 1978. J. Sci. FoodAgric. 29,577-586. Giilard, 0. 1967. J. Aust. Inst. Agric. Sci. 23, 30-34. Giilingham, A. G. 1978. W.D. thesis. Massey University, Palmerston North, New Zealand. Gillingham, A. G. 1987. In “Ecosystems of the World 17B. Managed Grasslands, Analytical Studies” (R. W. Snaydon, ed.), pp. 173 - 180. Elsevier, Amsterdam. Gillingham, A. G., and During, C. 1973. N.Z. J. Exp. Agric. 1,227-232. Giilingham, A. G., Syers, J. K., and Gregg, P. E. H. 1980. N.Z. J. Agric. Res. 23,323-330. Goodall, V . C. 1951. Proc. N.Z. Grassl. Assoc. 13,86-94. Goodroad, L. L., and Keeney, D. R. 1984. Soil Biol. Biochem. 16,39-43. Goold, G. J. 1980. N.Z. J. Exp. Agric. 8, 93-99. Grace, N. D. 1983. In “The Mineral Requirements of Grazing Ruminants” (N. D. Grace, ed.),N.Z. Soc. Anim. Prod. Occas. Publ. 9, pp. 106-111. New Zealand Society of Animal Production, Hamilton, New Zealand. Greenhalgh, J. F. D., and Reid, G. W. 1968. J. Agric. Sci. 71, 1 1 1-222. Groenwold, J., and Baarslag, A. 1982. Versl., Centr. Agrobiol. Onderzoek, Wageningen 62. Groenwold, J., and Keuning, J. A. 1988. Versl., Centr. Agrobiol. Onderzoek, Wageningen94. Gunary, D. 1968. J. Agric. Sci. 67,295-304. Hakamata, T. 1980. Nippon Sochi Gakkaishi 26,201 -207. Hakamata, T. 1985. Proc. Int. Grassl. Congr., 15th pp. 657-659. Hancock, J. 1950. N.Z. J. Sci. Technol.,Sect. A 3 2 4 22-59. Hancock, J., and McArthur, A. T. G. 1951. Proc. Ruakura Farmers’ Con$ pp. 32-38. Hardison, W. A., Fisher, H. L., Graf, G. C., and Thompson, N. R. 1956. J. Dairy Sci. 39, 1735-1741. Harrison, A. F. 1985. J. Environ. Manage. 20, 163- 179. Haynes, R. J. 1981. Adv. Agron. 33,227-261. Haynes, R. J. 1983. Grass Forage Sci. 38, 1 - 1 1.

194

R. J. HAYNES AND P. H. WILLIAMS

Haynes, R. J. 1986a. In “Mineral Nitrogen in the Plant-Soil System” (R. J. Haynes, ed.),pp. 127- 165. Academic Press, Orlando, Florida. Haynes, R. J. 1986b. In “Mineral Nitrogen in the Plant-Soil System” (R. J. Haynes, ed.),pp. 52- 126. Academic Press, Orlando, Florida. Haynes, R. J. 1986c. In “Mineral Nitrogen in the Plant-Soil System” (R. J. Haynes, ed.), pp. 303 - 378. Academic Press,Orlando, Florida. Haynes, R. J., and Sherlock, R. R. 1986. In “Mineral Nitrogen in the Plant-Soil System” (R. J. Haynes, ed.), pp. 242-302. Academic Press, Orlando, Florida. Haynes, R.J., and Swift, R. S. 1990. J. SoilSci. 41,73-83. Healy, W. B. 1968. N.Z. J. Agric. Res. 11,487-499. Helyar, K. R. 1976. J. Aust. Inst. Agric. Sci. 42,217-221. Helyar, K. R., and Godden, D. P. 1977. J. Aust. Inst. Agric. Sci. 43,22 - 30. Henzell, E. F. 1981. In “Nitrogen Fixation” (W. J. Broughton, ed.), pp. 264-289. Oxford University Press, Oxford, England. Henzell, E. F., and Ross, P. J. 1973. In “Chemistry and Biochemistry of Herbage” (G. W. Butler and R. W. Bailey, eds.), pp. 227-246. Academic Press, New York. Herriott, J. B. D., and Wells, D. A. 1963. J. Agric. Sci. 61,89-99. Hilder, E. J. 1966. Proc. Int. G r a d Congr., 10th pp. 977-981. Hogg, D. E. 1968. Proc. Inf. Congr. Soil Sci., 9th pp. 631 -638. Hogg, D. E. 1981. N.Z. J. Exp. Agric. 9,39-46. Hoglund, J. H., and Brock, J. L. 1987. In “Ecosystems of the World 17B. Managed Grasslands, Analytical Studies” (R. W. Snaydon, ed.), pp. 187- 196. Elsevier, Amsterdam. Holland, P. T., and During, C. 1977. N.Z. J. Agric. Res. 20,479-488. Holter, P. 1979. Oikos 32,393-402. Hutchinson, K. J., and King, K. L. 1979. In “Grassland Ecosystems and Man” (A. Breymeyer and G. M. van Dyne, eds.), pp. 823-852. Cambridge University Press, Cambridge, England. Hutton, J. B., and Jury, K. E. 1964. N.Z. J. Agric. Res. 7,583-595. Hutton, J. B., Jury, K. E., and Davies, E. B. 1965. N.Z. J. Agric. Res. 8,479-496. Hutton, J. B., Jury, K. E., and Davies, E. B. 1967. N.Z. J. Agric. Res. 10,367-388. Ineson, P. 1987. In “Agriculture and Conservation in the Hills and Uplands” (M. Bell and R. G. H. Bunce, eds.), Inst. Temest. Ecol. Symp. 23, pp. 44-49. Institute of Terrestrial Ecology, Grange-over-Sands, Cumbria. Jackman, R. J. 1960. Proc. N.Z. Soil Sci. SOC.4,24-27. Jackman, R. J. 1964. N.Z. J. Agric. Res. 7,445-471. Jarvis, S. C., and Hatch, D. J. 1985. Ann. Bot. 55,41-51. Jarvis, S . C., Shenvood, M., and Steenvoorden, J. H. A. M. 1987. In “Animal Manure on Grassland and Fodder Crops: Fertiliser or Waste?’ (H. G. van der Meer, R. J. Unwin, T. A. van Dijk, and G. C. Ennik, eds.), pp. 195-212. Nijhoff, Dordrecht, The Netherlands. Jarvis, S. C., Hatch, D. J., and Roberts,D. H. 1989. J. Agric. Sci. 112, 205-216. Jenkinson, D. S., and Oades, J. M. 1979. SoilBiol. Biochem. 11, 193- 199. Joblin, K. N. 1981. N.Z. J. Agric. Res. 24, 293-297. Joblin, K. N., and Keogh, R. G. 1979. J. Agric. Sci. 92,571 -574. Johnstone-Wallace, D. B., and Kennedy, K. 1944. J. Agric. Sci. 34, 190- 197. Keeney, D. R., and Madregor, A. N. 1978. N.Z. J. Agric. Res. 21,443-448. Kemp, A., and t’Hart, M. L. 1967. Neth. J. Agric. Res. 5,4- 17. Kennedy, P. M. 1974. Aust. J. Agric. Res. 25, 1015-1022. Kennedy, A. P., and Till, A. R. 1981. Aust. J. Agric. Res. 32,339-351. Keogh, R. G. 1979. Proc. Australas. ConJ Grassl. Invert. Ecol, 2nd pp. 49-51.

NUTRIENT CYCLING UNDER GRAZED PASTURE

195

Keuning, J. A. 1980. In “The Role of Nitrogen in Intensive Grassland Production” (W. H. Prins and G. H. Arnold, eds.), p. 163. Centre for Agricultural Publication and Documentation, Wageningen, The Netherlands. Kleinig, C. R. 1966. Aust. J. Agric. Res. 17, 327-333. Klemedtsson, L., Svensson, B. H., and Rosswall, T. 1988. Biol. Fertil. Soils 6 , 106- 111. Lambert, M. G., Renton, S. W., and Grant, D. A. 1982. In “Nitrogen Balances in N.Z. Ecosystems” (P. W. Gandar, ed.), pp. 35-39. Department of Scientific and Industrial Research, Palmerston North, New Zealand. Langlands, J. P., Sutherland, H. A. M., and Playne, M. J. 1973. Br. J. Nutr. 30,537-543. Lantinga, E. A., Keuning, J. A., Groenwold, J., and Deenen, P. J. A. G. 1987. In “Animal Manure on Grassland and Fodder Crops: Fertiliser or Waste?” (H. G. van der Meer, R. J. Unwin, T. A. van Dijk, and G. C. Ennik, eds.),pp. 103- 117. Nijhoff, Dordrecht, The Netherlands. Ledgard, S.F., andSaunders, W. M. H. 1982. N.Z. J. Agric. Res. 25, 541-547. Ledgard, S. F., Steele, K. W., and Saunders, W. H. M. 1982. N.Z. J. Agric. Res. 25,61-68. Lee, B. 1980. Rural Res. 106,4-9. Levy, E.B. 1956. Proc. Int. Grassl. Congr., 12th pp. 595-61 I. Levy, E. B. 1970. “Grasslands of New Zealand.” Government Printer, Wellington, New Zealand. Lockyer, D. R., and Whitehead, D. C. 1990. Soil Biol. Biochem. 22, 1 137- I 152. Lofs-Holmin, A. 1983. Swed. J. Agric. Res. 13, 31 -37. Lotero, J., Woodhouse, W. W., And Petersen, R. G. 1966. Agron. J. 58,262-265. MacDiarmid, B. N., and Watkin, B. R. 1971. J. Br. Grassl. Soc. 26,239-245. MacDiannid, B. N., and Watkin, B. R. 1972a. J. Br. Grassl. SOC.27,43-48. MacDiarmid, B. N., and Watkin, B. R. 1972b. J. Br. Grassl. SOC.27,48-54. MacLusky, D. S. 1960. J. Br. Grassl. SOC.15, 181-188. Marsh, R., and Campling R. C. 1970. Herb Abstr. 40, 123- 130. Marten, G. C., and Donker, J. D. 1964. J. Dairy Sci. 47,773-776. Martens, H., and Rayssiguier, Y. 1980. In “Digestive Physiology and Metabolism in Ruminants” (Y. Ruckebush and P. Thivend, eds.),pp. 447-468. Avi, Westport, Connecticut. Mason, V. C. 1969. J. Agric. Sci. 73,99 - 1 11. Mason, V. C. 1979. Z. Tierphysiol, Tieremaehrg.Futtermittelhd. 41, 140- 149. Mason, V. C., Kessank, P., Ononiwu, J.C., and Narang, M.P. 1981. Z. Tierphysiol, Tierernaehrg. Futtermittelkd. 45, 174 - 184. May, P. F., TU, A. R., and Cumming, M. J. 1972. J. Appl. Ecol. 9,25-49. May, P. F., Till, A. R., and Cumming, M. J. 1973. J. Appl. Ecol. 10,607-626. McAuliffe, C., and Bradfield, R. 1955. Soil Sci. 79,475-486. McAuliffe, C., Peech, M., and Bradfield, R. 1949. Soil Sci. 68, 185- 195. Metson, A. J., and Saunders, W. M. H. 1978. N.Z. J. Agric. Res. 21,341-353. Metson, A. J., Saunders, W. M. H., Collie, T. W., and Graham, V. W. 1966. N.Z. J. Agric. Res. 9,410-436. Monaghan, R. M., Cameron, K. C., and McLay, C. D. A. 1989. N.Z. J. Agric. Res. 32,

-

237 244.

Moore, C. W. E. 1964. In “Grasses and Grasslands” (C. Barnard, ed.), pp. 182-205. Macmillan, London. Momson, J. 1987. In “Ecosystems of the World 17B. Managed Grasslands, Analytical Studies” (R. W. Snaydon, ed.), pp. 61 -70. Elsevier, Amsterdam. Morton, J.D. 1984. “The Spatial Distribution of Excreta and Its Contribution to Soil Fertility,” Intern. Rep. N.Z. Minist. Agric. Fish. Ministry of Agriculture and Fisheries, Wellington, New Zealand.

196

R. J. HAYNES AND P. H. WILLIAMS

Morton, J. D., and Baird, D. 1990. N.Z. J. Agric. Res. 33,285-294. Mott, G. 0. 1974. In “Forage Fertilisation” (D. A. Mays, ed.),pp. 323-339. American Society of Agronomy, Madison, Wisconsin. Nguyen, M. L., Rickard, D. S., and McBride, S. D. 1989. N.Z. J. Agric. Res. 32,245-262. Norman, M. J. T., and Green, J. 0. 1958. J. Br. Grassl. SOC.13, 39-45. Oades, J. M. 1984. Plant Soil76,319-337. O’Connor, M. B. 1982. In “Nitrogen Fertilisers in New Zealand Agriculture” (P. B. Lynch, ed.), pp. 65-76. New Zealand Institute of Agricultural Science, Wellington, New Zealand. Omaliko, C. P. E. 1981. J. Range Manage. 34,360-362. Omaliko, C. P. E. 1984. Trop. Ecol. 25, 214-220. Paquay, R., De Baere, R., and Lousse, A. 1970a. J. Agric. Sci. 74,423-432. Paquay, R., De Baere, R., and Lousse, A. 1970b. J. Agric. Sci. 75,25 1-255. Parfitt, R. L. 1980. N.Z. J. Exp. Agric. 8,215-217. Perrott, K. W., and Sarathchandra, S. U. 1989. N.Z. J. Agric. Res. 32,409-413. Petersen, R. G., Lucas, H. L., and Woodhouse, W. W. 1956. Agron. J. 48,440-444. Prins, W. J., Den Boer, D. J., and Van Burg, P. F. J. 1986. In “Grassland Manuring” (J. P. Cooper and W. F. Raymond, eds.),Br. Grassl. Soc.Occas. Symp. 20, pp. 28-45. British Grassland Society, Hurley, Maidenhead, England. Quaterman, J. 1986. In “Trace Elements in Human and Animal Nutrition” (W. Mertz, ed.), pp. 28 1 - 3 17. Academic Press, Orlando, Florida. Quin, B. F. 1977. N.Z. SoilNews25, 142-144. Quin, B. F. 1982. In “Nitrogen Balances in N.Z. Soils” (P. W. Gandar, ed.), pp. 95- 102. Department of Scientific and Industrial Research, Palmerston North, New Zealand. Quin, B. F., and Rickard, D. S. 1981. Proc. Int. Grassl. Congr., 14th pp. 323-326. Raymond, W. F., and Minson, D. J. 1955. J. Br. Grassl. SOC.10,282-296. Richards, I. R., and Wolton, K. M. 1975. J. Br. Grassl. Soc. 30, 187-188. Richards, I. R., and Wolton, K. M. 1976. J. Br. Grassl. SOC.31,89-92. Rixon, A. J. 1966a. Ausf. J. Agric. Res. 17, 303-316. Rixon, A. J. 1966b. Aust. J. Agric. Res. 17, 317-325. Rixon, A. J., and Zorin, M. 1978. Soil Biol. Biochem. 10,347-354. Robertson, B. T. 1972. N.Z. Fert. J. 40, 16-18. Ross, D. J., Spek, T. W., Cowling, J. C., and Whale, K. N. 1984. Aust. J. Soil Res. 22, 3 19- 330. Rowarth, J. S. 1987. Ph.D. thesis. Massey University, Palmerston North, New Zealand. Rowarth, J. S., and Gillingham, A. G. 1990. Proc. N.Z. Grassl. Assoc. 51, 17-20. Rowarth, J. S., Gillingham, A. G., Tillman, R. W., and Syers, J. K. 1985. N.Z. J. Agric. Res. 28,497 - 504. Rowarth, J. S., Gillingham, A. G., Tillman, R. W., and Syers, J. K. 1988. N.Z. J. Agric. Res. 31, 187-193. Rowarth, J. S., Tillman, R. W., and Gillingham, A. G. 1990. N.Z. J. Agric. Res. 33, 473-477. Russell, J. S. 1960. Ausf.J. Agric. Res. 11,902-926. Russell, J. S. 1986. In “Australian Soils: The Human Impact” (J. S. Russell and R.F. Isbell, eds.), pp. 374-396. University of Queensland Press, St. Lucia, Queensland, Australia. Russell, J. S., and Harvey, D. L. 1959. Aust. J. Agric. Res. 10,637-650. Ryden, J. C. 1984. Proc. -Fert. Soc. 229. Ryden, J. C. 1985a. In “Cdour Prevention and Control of Organic Sludge and Livestock Farming” (V. C. Nielson, J. H. Voorburg, and P. Hermite, eds.), pp. 33-42. Elsevier, Amsterdam.

NUTRIENT CYCLING UNDER GRAZED PASTURE

197

Ryden, J. C. 1985b. In “Denitnfication in the Nitrogen Cycle” (H. L. Gotterman, ed.), pp. 121 - 134. Plenum, London. Ryden, J. C. 1986. In “Nitrogen Fluxes in Intensive Grassland Systems” (H. G. van der Meer, J. C. Ryden, and G. C. Ennik, eds.), pp. 59-73. NijhoE, Dordrecht, The Netherlands. Ryden, J. C., and Nixon, D. J. 1985. Grassl. Res. Inst. Final Annu. Rep. 1984-1985. pp. 21-22. Ryden, J. C., Lund, L. J., Letey, J., and Focht, D. D. 1979. SoilSci. Soc. Am. J. 43, 110- 118. Ryden, J. C., Ball, P. R., and Ganvood, E. A. 1984. Nature (London) 311,50-53. Ryden, J. C., Whitehead, D. C., Lockyer, D. R., Thompson, R. B., Skinner, J. H., and Garwood, E. A. 1987. Environ. Pollut. 48, 173- 184. Safley, L. M., Barker, J. C., and Westerman, P. W. 1984. Trans. ASAE 27, I 150- 1 153. Saggar, S., MacKay, A. D., Hedley, M. J., Lambert, M. G., and Clark, P. A. 1988. In “Towards the More Efficient Use of Soil and Fertiliser Sulphur” (R. E. White and L. D. Currie, eds.), Proc. Workshop, Massey Univ., pp. 262-278. Massey University, Palmerston North, New Zealand. Sanchez, P. A., and Uehara, G. 1980. In “The Role of Phosphorus in Agriculture” (F. E. Khasawneh, E. C. Sample, and E. J. Kamprath, eds.),pp. 47 1-514. American Society of Agronomy, Madison, Wisconsin. Sarathchandra, S. U., Perrott, K. W., and Upsdell, M. P. 1984. Soil Biol. Biuchem. 16, 177-183. Sarathchandra, S. U., Perrott, K. W., B o a , M. R., and Waller, J. E. 1988. Biol.Fertil. Soils 6,328-335. Saunders, W. M. H. 1959. N.Z. J. Agric. Rex 2,445 -462. Saunders, W. M. H. 1984. N.Z. J. Agric. Res. 27,405-412. Saunders, W. M. H., and Metson, A. J. 1959. N.Z. J. Agric. Rex 2, 1211 - 1231. Scholefield, D., Lockyer, D. R., Whitehead, D. C., and Tyson, K. C. 1991. Plant Soil 132, 165- 177. Sears, P. D. 1949. Sheep FarmingAnn. pp. 1-19. Sears, P. D. 1950. J. Br. Grassl. Soc. 5,267-280. Sears, P. D. 1951. N.Z. J. Sci. Technol.,Sect. A 33.4, 1 -29. Sears, P. D. 1956. Proc. Int. G r a d Congr., 7th pp. 92- 103. Sears, P. D., and Evans, L. T. 1953. N.Z. J. Sci. Technol.,Sect. A 35A (Suppl. I), 42-52. Sears, P. D., and Newbold, R. P. 1942. N.Z. J. Sci. Technol.,Sect. A MA, 36 -6 1. Sears, P. D., Goodall, V. C., and Newbold, R. P. 1948. N.Z. J. Sci. Technol., Sect. A MA, 231 -250. Sherlock, R. R., and Goh, K. M.1983. Soil Biol. Biochem. 15,615-617. Sherlock, R. R., and Goh, K. M. 1984. Fen. Res. 5, 181 - 195. Sherwocd, M.,and Fanning, A. 1989. I n “Management Systems to Reduce the Impact of Nitrates” (J. C. Germon, ed.), pp. 32-42. Elsevier, London. Simpson, J. R. 1987. In “Temperate Pastures: Their Production, Use and Management” (J. L. Wheeler, C. J. Pearson, and G. E. Robards, eds.), pp. 143- 154. Australian Wool Corporation/CommonwealthScientific and Industrial Research Organization, Canberra, Australia. Sinclair, A. G., and Saunders, W. M. H. 198I. Proc. N.Z. Fert. Manu. Res. Assoc. Tech. Conf: 18, 3- 14. Smith, H. A. 1984. In “Biochemistry of Essential Ultratrace Elements” (E. Frieden, ed.), pp. 341 -366. Plenum, New York. Smith, L. M. 1973. In “Alternative Sources of Protein for Animal Production,” pp. 147- 173. National Academy of Sciences, Washington, D.C.

198

R. J. HAYNES AND P. H. WILLIAMS

Smith, R. G., and Langlands, J. P. 1976. In “Reviews in Rural Science 11” (T. M. Sutherland, J. R. McWilliam, and R. A. Len& eds.),pp. 167- 173. University of New England Press, Armidale, Australia. Snaydon, R. W. 1981. In “Grazing Animals” (F. H. W. Morley, ed.), pp. 13- 3 1. Elsevier, Amsterdam. Snaydon, R. W. 1987. In “Ecosystems of the World 17B. Managed Grasslands, Analytical Studies” (R. W. Snaydon, ed.),pp 8 1-87. Elsevier, Amsterdam. Sparks, D. L., and Huang, P. M. 1985. In “Potassium in Agriculture” (R. D. Munson, ed.), pp. 201 -275. American Society of Agronomy, Madison, Wisconsin. Steele, K. W., Judd, M. J., and Shannon, P. W. 1984. N.Z. J. Agric. Res. 27, 5 - 1 1. Stevenson, F. J. 1982. In “Nitrogen in Agricultural Soils” (F. J. Stevenson, ed.), pp. 67- 122. American Society of Agronomy, Madison, Wisconsin. Stewart, J. W. B. 1984. In “Current Perspectives in Microbial Ecology: Proceedings of the Third International Symposium on Microbial Ecology” (M. J. Mug and C. A. Reddy, eds.), pp. 442-446. American Society of Microbiology, Washington, D.C. Suckling, F. E. T. 1959. N.Z. J. Agric. Res. 2,488-543. Suckling, F. E. T. 1975. N.Z. J. Exp. Agric. 3, 351-436. Syers, J. K., and Springett, J. A. 1983. In “Earthworm Ecology” (J. E. Satchell, ed.),pp. 67 - 83. Chapman & Hall, London. Tanji, K. K. 1982. In “Nitrogen in Agricultural Soils” (F. J. Stevenson, ed.), pp. 721 -772. American Society of Agronomy, Madison, Wisconsin. Tate, K. R., Speir, T. W., Ross, D. J., Parfitt, R. L., Whale, K. N., and Cowling, J. C. 1991. Plant Soil 132,219-232. Thomas, R. J., Logan, K. A. B., Ironside, A. D., and Bolton, G. R. 1988. Plant Soil 107, 173-181. Thompson, F. B., and Coup, M. R. 1940. N.Z. J. Sci. Technol. 22, 175- 180. Till, A. R. 1981. In “Grazing Animals” (F. H. W. Morley, ed.),pp. 33-53. Elsevier, Amsterdam. Till, A. R., and Blair, P. F. 1974. Trans. In?. Congr. Soil Sci., 10th 2, 153- 157. Tisdall, J. M., and Oades, J. M. 1982. J. Soil Sci. 33, 141- 163. Topps, J. H., and Elliott, R. C. 1967. Anim. Prod. 9,219-227. Tothill, J. C. 1978. In “Plant Relations in Pastures” (J. R. Wilson, ed.),pp. 385-402. Commonwealth Science and Industrial Research Organization, Canberra, Australia. Underhay, V. H. S., and Dickinson, C. H. 1978. J. Br. Grassl. Soc. 33, 189- 196. Underwood, E. J. 1981. “The Mineral Nutrition of Livestock.” Commonwealth Agricultural Bureaux. Slough, England. Vallis, I., and Gardener, C. J. 1984. Aust. J. Exp. Agric. Anim. Husb. 24,522-528. Vallis, I., Harper, L. A., Catchpoole, V. R., and Weier, K. L. 1982. Aust. J. Agric. Res. 33, 97- 107. Vallis, I., Peake, D. C. I., Jones, R. K., and Md=own, R. L. 1985. Aust. J. Agric. Res. 36, 809- 8 17. Van Faassen, H. G., and Van Dijk, H. 1987. In “Animal Manure on Grassland and Fodder Crops: Fertiliser or Waste?” (H. G. van der Meer, R. J. Unwin, T. A. van Dijk, and G. C. Ennik, eds.), pp. 27-45. Nijhoff, Dordrecht, The Netherlands. Van Vuuren, A. M., and Meijs, J. A. C. 1987. In “Animal Manure on Grassland and Fodder Crops: Fertiliser or Waste?” (H. G. van der Meer, R. J. Unwin, T. A. van Dijk, and G. C. Ennik, eds.), pp. 17-25. Nijho5, Dordrecht, The Netherlands. Vercoe, J. M. 1962. Proc. Aust. SOC.Anim. Prod. 4, 160- 162. Vlek, P. L. G., Stumpe, J. M., and Bymes, B. H. 1980. Fert. Res. 1, 191-202. Waite, R., MacDonald, W. B., and Holmes, W. 1951. J. Agric. Sci. 41, 163- 173.

NUTRIENT CYCLING UNDER GRAZED PASTURE

199

Walker, T. W., Tharpa, B. K., and Adams, A. F. R. 1959. SoilSci. 87, 135-140. Wardrop, J. C. 1963. Br. J. Anim. Behav. 1,23-31. Waters, R. A. S. 1955. N.Z. J. Sci. Technol.,Sect. A 3 6 4 516-525. Watkin, B. R. 1957. J. Br. Grassl. SOC.12,264-277. Watkin, B. R., and Clements, R. J. 1978. In “Plant Relations in Pastures” (J. R. Wilson, ed.), pp. 273 -289. Commonwealth Scientific and Industrial Research Organization, Canberra, Australia. Watkin, B. R., and Wheeler, J. L. 1966. J. Br. Grassl. SOC.21, 14-20. Watson, E. R. 1969. Aust. J. Agric. Res. 20,447-456. Watson, E. R., and Lapins, P. 1969. Aust. J. Exp. Agric. Anim. Husb. 9, 85-91. Weeda, W. C. 1967. N.Z. J. Agric. Res. 10, 150- 159. Weeda, W.C. 1977. N.Z. J. Agric. Res. 20,471-478. Weeda, W. C. 1979. N.Z. J. Exp. Agric. 7,289-293. Whitehead, D. C. 1970. “The Role of Nitrogen in Grassland Productivity,” Bur. Pastures Field Crops, Bull. 48. Commonwealth Agricultural Bureaux, Hurley, Maidenhead, England. Whitehead, D. C. 1986. In “Nitrogen Fluxes in Intensive Grassland Systems” (H. G. van der Meer, J. C. Ryden, and G. C. Ennik, eds.), pp. 47-58. Nijhoff, Dordrecht, The Netherlands. Whitehead, D. C., and Bristow, A. 1990. J. Appl. Ecol. 27,667-678. Whitehead, D. C., Lockyer, D. R., and Raistrick, N. 1989. Soil Biol. Biochem. 21,803-808. Wilkins, R. J., and Ganvood, E. A. 1986. In “Grazing” (J. Frame, ed.), Br. Grassl. SOC. Occas. Symp. 19, pp. 19-3 1. British Grassland Society, Hurley, Maidenhead, England. Wilkinson, S. R., and Lowrey, R. W. 1973. In “Chemistry and Biochemistry of Herbage” (G. W. Butler and R. W. Bailey, eds.),pp. 247-315. Academic Press, New York. Williams, C. H. 1980. Aust. J. Exp. Agric. Anim. Husb. 20, 561 -567. Williams, 0. B. 1981. I n “Grazing Animals” (F. H. W. Morley, ed.), pp. 1 - 12. Elsevier, Amsterdam. Williams, P. H., 1988. Ph.D. thesis. Massey University, Palmerston North, New Zealand. Williams, C. H., and David, D. J. 1976. Aust. J. SoilRes. 14,85-93. Williams, C. H., and Donald, C. M. 1957. Aust. J. Agric. Rex 8, 179-189. Williams, P. H., and Haynes, R. J. 1990a. N.Z. J. Ecol. 14,49 - 57. Williams, P. H., and Haynes, R. J. 1990b. Trans. Congr. SoilSci., 14fh 4,276-281. Williams, P. H., Hedley, M. J., and Gregg, P. E. H. 1988. N.Z. J. Agric. Res. 31,431 -438. Williams, P. H., Hedley, M. J., and Gregg, P. E. H. 1989. N.Z. J. Agric. Res. 32,415-421. Williams, P. H., Gregg, P. E. H., and Hedley, M. J. 1990a. N.Z. J. Agric. Res. 33, 15 1 - 158. Williams, P. H., Gregg, P. E. H., and Hedley, M. J. 1990b. N.Z. J. Agric. Res. 33,489-495. Williams, P. H., Hedley, M. J., and Gregg, P. E. H. 1990~.Aust. J. Soil Res. 28, 857- 868. Williams, P. H., Gregg, P. E. H., and Hedley, M. J. 1990d. N.Z. J. Agric. Res. 33,661 -668. Wolton, K. M. 1979. In “Changes in Sward Composition and Productivity” (A. H. Charles and R. J. Haggar, eds.), Br. Grassl. Soc. Occas. Symp. 12, pp. 13 1 - 135. British Grassland Society, Hurley, Maidenhead, England. Yuan, T. L. 1980. Soil Sci. SOC.Am. J. 44,95 1-955.

This Page Intentionally Left Blank

ELECTRICAL CONDUCTIVITY M~THODS FOR M~ASURING AND MAPPING SOIL SALINITY J. D. Rhoades United States Salinity Laboratory, United States Department of Agriculture, Agricultural Research Service, Riverside. California 92 50 1

I. Introduction 11. Determination of Soil Salinity from Aqueous Electrical Conductivity A. Principles of Aqueous Electrical Conductivity B. Soil Water Salinity C. Soil Extract Salinity 111. Determination of Soil Salinity from Soil Paste or Bulk Soil Electrical Conductivity A. Principles of Soil and Paste Electrical Conductivities B. Determining Soil Salinity From Saturated Paste Electrical Conductivity C. Determining Soil Salinity From Bulk Soil Electrical Conductivity IV. Conclusions and Summary References

I. INTRODUCTION The diagnosis, management, and reclamation of saline' soils are evaluated from spatial information on soil salinity levels. Salinity problems are especially prevalent and serious in irrigated lands, with about one-third of the imgated land in the United States being seriously salt affected. In some countries the ratio is nearly one-half (Postel, 1989). Many of the surface water and groundwater supplies associated with these lands are being Soil salinity refers to the presence of excessive levels of dissolved, or readily dissolvable, inorganic solutes in soils. Ahanre1 in Agwmmy, Val. 49

20 1

202

J. D. RHOADES

depleted and salinized by their consumption for irrigation and/or by the return to them of salt-laden drainage water. Sufficient leaching and drainage are required to keep salinity within irrigated soils from exceeding tolerable levels, if crop production and profitability of irrigated agriculture are to be sustained. However, it is these very processes that often lead to the pollution of our water resources. Currently, programs are being implemented to reduce leaching and restrictions are being legislated to control the discharge of saline drainage water from irrigation projects. Concomitantly, reuse of saline drainage water is increasing as disposal is being limited and as the availability of fresh water supplies is decreasing. With less leaching and drainage and greater use of saline water for irrigation, the soil salinity hazard potential increases. The proper management of irrigated agriculture, especially under the above-described conditions, requires periodic information on the soil salinity status of fields, farms, projects, and hydrogeologic areas. Only with this information can the appropriateness and effectiveness of farm practices, of land use plans, of water quality plans, and of irrigation project operations be assessed with respect to leaching/drainage adequacy, salt balance, irrigation sustainability, water use efficiency, and environmental protection. Practical methods for measuring, monitoring, and mapping soil salinity are essential to meet these burgeoning needs. In addition, practical procedures are needed for locating representative measurement/ monitoring sites in order to map the distribution and extent of salt-affected soils, to delineate areas of under- and overirrigation and areal sources of salt loading, and to monitor and assess salinity trends. Ideally, it would be desirable to know the concentrations of the individual solutes in the soil water over the entire range of field water contents and to obtain this information immediately in the field. Practical methods are not available at present to permit such determinations, although determinations of total solute concentration (i.e., salinity) can be made in situ using electrical or electromagnetic signals from appropriate sensors. Such immediate determinations are so valuable for salinity diagnosis, inventorying, monitoring, and irrigation management needs that, in many cases, they supplant the need for soil sampling and laboratory analyses. However, if knowledge of a particular solute(s)concentration is needed (such as when soil sodicity or a specificion toxicity is to be assessed), then either a sample of soil, or of the soil water, is required to be analyzed. Of course, the methods to accomplish this require much more time, expense, and effort than do the instrumental field methods. Thus, a combination of the various methods should be used to minimize the need for sample collection and chemical analyses, especially when monitoring solute changes with time and characterizing the salinity conditions of extensive areas.

MEASURING AND MAPPING SOIL SALINITY

20 3

Assessing soil salinity is complicated by its spatially variable nature. Numerous samples (measurements) are needed to characterize just one field. Furthermore, soil salinity is dynamic in nature due to the influences of varying soil/crop/irrigation management practices, water table depth, soil permeability, evaporation and transpiration rates, rainfall amount and distribution, and salinity of the perched groundwater. Thus, soil salinity information needs to be updated as conditions change. When the need for repeated measurements and extensive sampling requirements is met, the expenditure of time and effort to characterize and monitor the salinity condition of a large area with conventional soil sampling and laboratory analysis procedures becomes impractical. However, rapid instrumental field techniques for measuring soil electrical conductivity, for inferring salinity from it, and for locating spatial position on the landscape, coupled with use of data logging equipment, statistics, and computer-assisted mapping techniques, offer us the potential to meet our soil salinity assessment needs in this regard. The additional use of geographic information systems and remote sensing technology further increases this potential. Soil salinity has been customarily defined and assessed in terms of laboratory measurements of the electrical conductivity of the extract of a saturated paste of a soil sample (EC,), because electrical conductivity is a practical index of the total concentration of ionized solutes in an aqueous sample and the saturation percentage (SP) is the lowest water/soil ratio for the practical laboratory extraction of readily dissolvable salts in soils (U.S. Salinity Laboratory Staff, 1954). But it can also be determined from the measurement of the electrical conductivity of a soil water sample (EC,). This latter measurement can be made either in the laboratory on a collected sample or directly in the field using in situ, imbibition-type salinity sensors. Alternatively, salinity can be indirectly determined from measurement of the electrical conductivity of a saturated soil paste (EC,) or of the electrical conductivity of the bulk soil (EC,). EC, can be measured either in the laboratory or field using simple and inexpensive equipment. EC, can be measured in the field, using electrical-type probes placed in contact with the soil, or remotely, using electromagnetic induction devices. The latter two measurements require more expensive, but very cost effective, equipment. From EC, and EC,, soil salinity can be derived in terms of either EC, or EC, . The appropriate method to use depends on the purpose of the determination, the size of the area being evaluated, the number and frequency of measurements needed, the accuracy required, and the available equipment/manpower. This paper reviews the various electrical conductivity methods for measuring soil salinity together with compatible ways for mapping it, including establishing the locations of measurement sites. Advantages and limita-

204

J. D. RHOADES

tions of the alternative methods are discussed and a practical integrated mobile system for measurement/monitoring/mappingis described. For earlier treatises on the instrumental field methodology of soil salinity measurement and assessment, see Rhoades (1976, 1978, 1984, 1990a,b, 1992a,b), Rhoades and Oster (1986), Rhoades and Corwin (1984, 1990), and Corwin and Rhoades ( 1990).

11. DETERMINATION OF SOIL SALINITY FROM AQUEOUS ELECTRICAL CONDUCTIVITY A. PRINCIPLES OF AQUEOUS ELECTRICAL CONDUCTMTY Electrical conductivity is a numerical expression of the inherent ability of a medium to cany an electric current. Because the EC of an aqueous solution is closely related to the total concentration of dissolved electrolytes (ionic solutes) in the solution (water itself is a very poor conductor of electricity), it is commonly used as an expression of the total dissolved salt concentration of an aqueous sample, even though it is also affected by the temperature of the sample and by the mobilities, valences, and relative concentrations of the individual ions comprising the solution. Furthermore, not all dissolved solutes exist as charged species; some combine to form ion pairs, and some of the ion pairs are neutral and do not contribute to electrical conductivity. The determination of EC generally involves the physical measurement of the resistance (R),expressed in ohms, of a material. The resistance of a conducting material (such as a saline solution) is inversely proportional to its cross-sectional area (A) and directly proportional to its length (L). The magnitude of the resistance measured therefore depends on the characteristics (dimensions) of the conductivity cell used to contain the sample and the electrodes. Specific resistance ( R , ) is the resistance of a cube of the sample 1 cm on edge. Practical cells are not of this dimension and measure only a given fraction of the specific resistance; this fraction is the cell constant ( K = R/R,). The reciprocal of resistance is conductance (C). It is expressed in reciprocal ohms, i.e., mhos. When the cell constant is applied, the measured conductance is converted to specific conductance (i.e., the reciprocal of the specific resistance) at the temperature of measurement. Herein, specific conductance is called electrical conductivity, EC: EC = l/R,

=K/R

(1)

Electrical conductivity has been customarily reported in micromhos per

MEASURING AND MAPPING SOIL SALINITY

20s

centimeter (pmho/cm), or in millimhos per centimeter (mmho/cm). In the International System of Units (SI), the reciprocal of the ohm is the siemen (S) and, in this system, electrical conductivity is reported as siemens per meter (S/m), or as decisiemens per meter (dS/m). One dS/m is equivalent to one millimho/cm. Electrolytic conductivity (unlike metallic conductivity) increases with temperature at a rate of approximately 1.9%/1"C.Therefore, EC needs to be expressed at a reference temperature for purposes of salinity expression; 25 "Cis most commonly used in this regard. The best way to correct for the temperature effect on conductivity is to maintain the temperature of the sample and cell at 25 f 0.5"Cwhile EC is being measured. The next best way is to make multiple determinations of sample EC at various temperatures both above and below 25"C, then to plot these readings and interpolate the EC at 25°C from the smoothed curve drawn through the datapairs. For practical purposes of agricultural salinity appraisal, EC can be measured at one known temperature other than 25 "Cand then adjusted to this latter reference using an appropriate temperature coefficient (A). These coefficients are usually based on sodium chloride solutions, because their temperature coefficients closely approximate those of most surface waters and groundwaters. Potassium chloride solutions are not generally used for this purpose because they have a lower temperature coefficient of conductivity than is typical of most natural waters or soil extracts. Another limitation in the use of temperature coefficients to adjust EC readings to 25°C is that they vary somewhat with solute concentration. The lower the concentration, the higher the coefficient, due to the effect that temperature has on the dissociation of water. However, for practical needs, these limitations may be ignored and the value off; may be assumed to be single valued. It may be estimated as follows:

+

f ; = 1 0.019(t - 25)

(2)

or

f ; = (0.0004)t' - (0.0430)t

+ 1.8 149

(3) The latter relation was derived from the data given in Table 15 of Handbook 60 (U.S. Salinity Laboratory Staff, 1954). In turn, the EC at 25°C (EC25) is estimated by multiplying the EC measured at temperature t (EC,) by the temperature coefficient as follows: EC25 = EC,f;

(4)

Because of differences in the equivalent weights, equivalent conductivities, and variations in the proportions of the various solutes found in soil extracts and water samples, the relationships between EC and total solute concentration and osmotic potential are only approximate. They are still

206

J. D. RHOADES

quite useful, however. These relationships are as follows: total cation (or anion) concentration, millimoles charge/liter = l OEC,, , in dS/m; total dissolved solids, milligrams/liter = 640ECZ5,in dS/m; and osmotic potential, 100 kPa at 25°C zz 0.4ECZ,,in dS/m.

B. SOILWATER SALINITY Theoretically, the electrical conductivity of the soil solution (EC,) is a better index of soil salinity than is EC,, because the plant roots actually experience the soil solution; they extract their nutrients from it, absorb other solutes from it, and consume it through the process of transpiration. However, EC, has not been widely used as a means for measuring or expressing soil salinity for several reasons. First, it is not single valued; it varies over the irrigation cycle as the soil water content changes (Rhoades, 1978). Thus, EC, does not lend itself to simple classifications or standards unless it is referenced to a fixed water content, such as field capacity. Second, and probably most importantly, EC, has not been widely adopted for routine appraisals of soil salinity because methods for obtaining soil water samples at typical field water contents are not very practical. Samples of soil solutions may be obtained from soil samples in the laboratory by means of displacement, compaction, centrifugation, molecular adsorption, and vacuum or pressure extraction methods (Richards, 1941). Displacement methods have been described by Adams ( 1974); combination displacement/centrifugation methods, by Gillman (1976), Mubarak and Olsen (1976, 1977), and Elkhatib d al. (1986); a combination vacuum/displacement method, by Wolt and Graveel ( 1986); a simple field-pressure filtration method, by Ross and Bartlett (1 990); and adsorption techniques, by Davies and Davies (1963), Yamasaki and Kishita (1972), Gillman (1976), Dao and Lavy (1978), Kinniburgh and Miles (1983), and Elkhatib et al. (1987). Comparisons of the various methods have been made by Adams et al. ( 1980), Kittrick ( 1983), Wolt and Graveel ( 1986), Menzies and Bell ( 1988), and Ross and Bartlett ( 1990). Two means of measuring EC, in undisturbed soils exist. One is to collect a sample of soil water using an in situ extractor and then to measure its EC; the second is to measure EC, “directly” in the soil using in situ, imbibition-type “salinity sensors.” Soil water samples are usually collected in the field using vacuum extractors. The suction method, first proposed by Briggs and McCall ( 1904), is useful for extracting water from the soil when the soil water suction is less than about 0.1 MPa. Although the available range of soil moisture for crops extends to 1.5 MPa of soil suction, most water uptake by plants takes

MEASURING AND MAPPING SOIL SALINITY

207

place within the range of 0-0.1 MPa. Therefore, the suction method is applicable for many salinity monitoring needs. Although different extraction devices have been used, the most commonly used is the porous ceramic cup. Early vintage extractor construction and performance have been described in a bibliography assembled by Kohnke et al. (1940). Reeve and Doering (1965) described in detail the more modern equipment and procedures for its use. These procedures have been used at the U.S. Salinity Laboratory with good success for salinity appraisal purposes. Wagner (1965) used similar devices to estimate nitrate losses in soil percolate. Other improved and specialized versions have since been developed for various purposes, including a miniature sampler that eliminates sample transfer in the field (Hams and Hansen, 1975), samplers that shut off automatically when the desired volume of sample is collected (Chow, 1977), samplers that function at depths greater than the suction lift of water (Parizek and Lane, 1970; Wood, 1973), and samplers that minimize “degassing” effects on solution composition (Suarez, 1986, 1987). Soil water has also been extracted using cellulose acetate hollow fibers (Jackson et al., 1976; Levin and Jackson, 1977), which are thin-walled, semipermeable, and flexible. Claimed advantages include flexibility, small diameter, minimal chemical interaction of solutes with the tube matrix, and compositional results comparable with those from samples obtained from ceramic extraction cups. Pan-type collectors have also been used to collect soil percolate (Jordan, 1968). Additionally, large-scale vacuum extractors (15 cm wide by 3.29 m long) have been built and used to assess deep percolation losses and the chemical composition of soil water (Duke and Haise, 1973). Ceramic “points,” which absorb water on insertion into the soil, have also been used to sample soil water with some success (Shimshi, 1966). However, only very small samples are obtained with these points and there are potential errors due to vapor transfer and chromatographic separation. Tadros and McGarity (1976) have analogously used an absorbent sponge material. Various errors in sampling soil water can occur with the use of any of the above types of extractors. Included are factors related to sorption, leaching, diffusion, and sieving by the cup wall; sampler intake rate; plugging; and sampler size. Nielsen et al. ( I 973), Biggar and Nielsen (1976), and van De Pol et al. ( I 977) used soil water extractors to determine salt flux in fields and have demonstrated that field variability in this regard is very large. They concluded that soil water samples, being point samples, can provide only indications of relative changes in the amount of solute flux, but not quantitative amounts, unless the frequency distribution of such measurements is established. Because the composition and concentration of soil water are not homogeneous through its mass, water drained from large

208

J. D. RHOADES

pores at low suctions (as collected by vacuum extractors) may have a composition very different from water extracted from micropores. A point source of suction, such as a porous cup, samples a sphere of different-sized pores, depending on distance from the point, the amount of applied suction, the hydraulic conductivity of the medium, and the soil water content. Although vacuum extractors are versatile and easily usable and provide for in situ sampling of soil water, they are, as evident from the above discussion, not without limitations. The different suction-type samplers and other methods for sampling soil solution and various errors associated with them have been critically reviewed by Rhoades (1978), Rhoades and Oster (1 986), Litaor (1 988), and Grossman and Udluft (199 1). When the total concentration of salts in the soil water is sufficient information, i.e., when specific solute analyses are not needed, in situ devices capable of directly measuring EC, may be used advantageously. Kemper (1959) developed the first in situ salinity sensor. It consisted of electrodes imbedded in porous ceramic to measure the electrical conductivity of the solution imbibed within the “ceramic cell.” When placed in soil, these devices imbibe and come to diffusional equilibrium with the soil water. Richards (1966) improved the design of the soil salinity sensor to shorten its response time and to eliminate external electrical current paths. This unit is now produced commercially. In this unit (Fig. I), the salinitysensitive element is an approximately 1 -mm-thick ceramic disk containing platinum screen electrodes on opposite sides. This gives a short diffusion path and thus lowers response time. Another feature of the design is a preloaded spring. After the salinity sensor is placed in the soil, the spring is released to ensure good contact of the ceramic plate with the soil. A thermistor is incorporated in the sensor so that the EC may be adjusted for temperature effects. An oscillator circuit system has been developed for automated salinity sensor measurements and data logging (Austin and Oster, 1973). This permits linear readings to be obtained with lead lengths of up to several hundred meters. Salinity sensors have been used primarily in agricultural research when continuous monitoring of soil salinity in soil columns, lysimeters, and field experiments is required (Oster and Ingvalson, 1967; Rhoades, 1972; Oster et al., 1973, 1976; Ingvalson et a/., 1970). The accuracy of the commercial ceramic sensor has been found to be 20.5 dS/m (Oster and Ingvalson, 1967). Reliability of commercial sensors was determined by removing them from field and lysimeter experiments after 3 to 5 years of continuous operation and comparing their calibrations relative to original ones (Oster and Willardson, 1971; Wood, 1978). About 68% of the tested sensors had calibrations within 14% of the original calibrations after 5 years. Shifts varied in direction and magnitude, and some complete failures occurred.

MEASURING A N D MAPPING SOIL SALINITY

209

Figure 1. Commercial meter and salinity sensor showing ceramic disk in which platinum electrodes are imbedded; lead wires are sheathed in plastic housing.

Response times of commercial salinity sensors have been evaluated in field situations (Wesseling and Oster, 1973; Wood, 1978). In the matric potential range of -0.05 to -0.15 MPa, 90% of the response of these sensors to a step change in salinity will occur within 2 to 5 days. Thus, it may be concluded that salinity sensors are not well-suited for measuring short-term changes in salinity because of their relatively long response time (at least several days). At lower matric potentials, response times are longer. Desaturation of the ceramic occurs at matric potentials more negative than -0.2 MPa, significantly reducing the conductance of the ceramic salinity sensor (Ingvalson et al., 1970). Hence, this type of sensor is not accurate in “dry” soils. Salinity sensors constructed of porous glass have been developed; these remain saturated with soil water to 2 MPa matric potentials (Enfield and Evans, 1969), but they are fragile and are not available commercially. Soil disturbance during installation can result in errors associated with modified water infiltration in the backfilled hole used to install salinity sensors. Special precautions during their installation must be taken to avoid this. Although obviously not without limitations, salinity sensors may be used

210

J. D. RHOADES

Figure 2. Variations in in situ soil water electrical conductivity and tension in the root zone of an alfalfa crop during the spring of the year. (After Rhoades, 1972.0 by Williams & Wilkins, 1972.)

advantageously for continuously monitoring electrical conductivity of soil water at selected depths over relatively long periods of time, as illustrated in Fig. 2. They are not well-suited for measuring short-term changes of salinity, especially in dry soils. Many units may be needed because of their small sampling volume and because of the substantial heterogeneity of soils. These numbers can be minimized if the sensors are primarily used to follow changing salinity status at a specific location over time. They are simple in principle, easily read, and sufficiently accurate for intermediateterm salinity monitoring. They are, of course, not practical for mapping purposes for obvious reasons.

C . SOILEXTRACTSALINITY Because present methods of obtaining soil water samples at typical field water contents are not very practical, aqueous extracts of the soil samples are usually made in the laboratory at higher than normal water contents for routine soil salinity diagnosis and characterization purposes. Because the absolute and relative amounts of the various solutes are influenced by the water/soil ratio at which the extract is made (Reitemeier, 1946), the water/soil ratio used to obtain the extract should be standardized to obtain results that can be applied and interpreted generally. As stated earlier, soil

MEASURING AND MAPPING SOIL SALINITY

21 1

salinity is most generally defined and measured on aqueous extracts of so-called, saturated soil pastes (U.S. Salinity Laboratory Staff, 1954). This water content and the water/soil ratio (the so-called saturation percentage) vary with soil texture but are used not only because they are the lowest ones for most soils for which sufficient extract can be practically removed from a soil sample for the compositional analysis of major salt constituents, but also because they are related in a reasonably general and predictable way to soil water contents and ratios under field conditions. For these same reasons, crop tolerance to salinity is also most generally expressed in terms of the electrical conductivity of the saturation extract (EC,) (Maas and Hoffman, 1977; Maas, 1986, 1990). EC, is typically determined as follows. A saturated soil paste is prepared by adding distilled water to a sample of air-dry soil (200-400 g) while stirring and then allowing the mixture to stand for several hours to permit the soil to imbibe the water and the readily soluble salts to dissolve fully, so as to achieve a uniformly saturated and equilibrated soil water paste. At this latter point, which is sufficiently reproducible, the soil paste glistens as it reflects light, flows slightly when the container is tipped, slides freely and cleanly off a spatula, and consolidates easily when the container is tapped or jarred after a trench is formed in the paste with the broad side of the spatula. The extract of this saturation paste is usually obtained by suction using a funnel and filter paper. The EC and temperature of this extract are then measured using standard conductance meters/cells/thermometers; EC,, is calculated from Eq. (4) to give EC,. For more details on these procedures, see Rhoades ( 1982, 1993). To eliminate some of the subjectivity of the saturation extract method, Longenecker and Lyerly ( 1964)proposed wetting the sample by capillarity using a “saturation table.” Beatty and Loveday ( 1974) and Loveday ( 1972) advocated predetermining the amount of water at saturation on a separate soil sample using a similar capillary wetting technique and then adding this amount to all other samples of the same soil. Allison (1973) recommended slowly adding soil to water, rather than water to soil, when making pastes to speed preparation of the saturated paste. All of these modifications offer advantages over the standard procedure under certain situations. Other extraction ratios, such as 1 : 1, 1 :5, etc., are easier to use than that of the saturation paste, but they are less well-related to soil properties and more subject to errors from peptization, hydrolysis, cation exchange, and mineral dissolution. Sonnevelt and van den Ende ( 1971) recommended a 1 :2 volume extract. This method is a compromise between the saturation paste extract and the higher dilution “weight” extracts. The water contents of the 1 :2 volume pastes of sandy and clayey soils are higher and lower, respectively, relative to the saturation paste extract. For purposes of moni-

212

J. D. RHOADES

toring, when relative changes are of more concern than the absolute solute concentration(s), these quicker, simpler methods of “fixed extraction ratios” may be used to advantage in place of the saturation extract. Of course, the relations given in Handbook 60 (U.S. Salinity Laboratory Staff, 1954) to predict exchangeable sodium percentage from the sodium adsorption ratio apply only to saturation paste extract, as do most of the other indices/criteria/standards used to express/interpret soil salinity/sodicity/ toxicity and plant response (salt tolerance and plant growth data). Once soil extract samples are obtained, laboratory chemical analyses can be camed out to determine, in addition to the electrical conductivity of the extract (EC,), the concentrations of the individual solutes, i.e., Na+, Ca2+, Mg2+,K+, C1-, S@-, HCO:, C e - , and NOT. Methods for such analyses are given elsewhere (Rhoades, 1982). More details about the methods for measuring the electrical conductivity and total dissolved solid contents of aqueous samples and extracts are given by Rhoades (1993).

111. DETERMINATION OF SOIL SALINITY FROM SOIL PASTE OR BULK SOIL ELECTRICAL CONDUCTIVITY

A. PRINCIPLES OF SOILAND PASTE ELECTRICAL CONDUCTIVITIES A model of the electrical conductivity of mixed soil/water systems that has been shown to be very useful for purposes of salinity appraisal is illustrated in Fig. 3. This method assumes that the electrical conductivity of a soil (or a soil paste) containing dissolved electrolytes(salts) in the soil “solution” can be represented by conductance via three pathways (or elements) acting in parallel: (1) conductance through continuous soil solution pathways (a liquid element), (2) conductance through alternating layers of soil particles and the soil solution that envelopes and separates these particles (a solid- liquid, series-coupled element), and (3) conductance through or along the surfaces of soil particles in direct and continuous contact with one another (a solid element). In most imgated soils and pastes, the solid element is insignificant and, for all practical purposes, the model reduces to a two-component model (Rhoades et al., 1989a). This model is mathematically represented by Eq. (5):

where EC, and EC, are the specific electrical conductivities of the soil water in the fine pores (series-coupled pathway) and in the large pores (continuous pathway), respectively; ,8 and ,8 are the corresponding

MEASURING AND MAPPING SOIL SALINITY

213

air

.-.

.-.

A Figure 3. Schematic representation and model of electrical conductivity in soil. (A) The three paths that current can take in unsaturated soil. (B) Simplified soil model consisting of the three conductance elements (a-c) in parallel. (After Rhoades eta/., 1989a).

series-coupled and continuous pathway volumetric contents of soil water; 0, is the total volumetric content of soil water; 0, is the total volumetric content of soil particles; and EC, is the average specific electrical conductivity of the soil particles. The soil water in the continuous pathway, dWc (=0, - Ow), is envisioned as the “mobile” water phase. It can be different in electrolyte composition (i.e., EC,) than that in the “immobile” water phase (i.e., EC,), which is associated with the fine and intraped pore water (i.e., the immobile water, 0,). At equilibrium, EC, and EC, would be the same, but during transient-state periods, such as immediately after irrigation or rainfall, they would likely be different. This model assumes that EC, is independent of 0, and EC,, which appears to be the case for most practical purposes (Shainberg et al., 1981; Bottraud and Rhoades, 1985; Rhoades et al., 1990b). For conditions of EC, greqter than about 2 dS/m and for soils with typical values of EC, (I 1.5 dS/m), the product 0,EC, is so much larger than the product B,EC, that the latter can be neglected. Equation ( 5 ) then simplifies to

[

1

EC, -I-(0, - O,)EC, 0, For such cases, the relation between EC, and EC, in Eq. (6) is linear for EC,

=

(”

+ -

J. D. RHOADES

214

any value of O,(Ow - 0,) beyond some threshold level and the y intercept depends on EC,, O,, and 0,. Because the ratio [(O, O,)z/O,] is typically close to the value 1, the intercept of Eq. (6) is approximately equal to EC,. The earlier EC, model of Rhoades et al. (1976) is analogous to this limiting case version of Eq. (9,as discussed elsewhere (Rhoades et al., 1989a). At low levels of EC,, the relation between EC, and EC, is curvilinear, as described in Eq. (5). The first term of the equation determines the shape of the EC, - EC, curve. Over the remainder of the EC, range, EC, and EC, are linear, with Ow - 0, representing the slope, as described above. So, although Eq. ( 5 ) describes the full relation between EC, and EC,, Eq. (6) can be used for conditions of EC, 2 2-4 dS/m (which corresponds approximately to EC, 1 1- 2 dS/m). A typical set of data illustrating the appropriateness of the abovedescribed model and of generalizationsusing the model is shown in Fig. 4 for Waukena loam soil. The solid line is that described by Eq. ( 5 ) and the

+

-

.c

0

h c

4-

L - - - _ - - - _ - - - _ _ - - _ _ - _ :

9,-

Waukena loam

12

16

20

Electrical Conductivity of Soil Water, ECw,dS/m Figure 4. The electrical conductivity of Waukena loam soil as a function of the electrical conductivity and volumetric content of soil water. The measured data points (0)are shown and the solid line is the “fit” of these combined data by Eq. (5). (After Rhoades et al., 1989a.)

MEASURING AND MAPPING SOIL SALINITY

3 a

I .o

1

I

-

0.6

-

0.4

-

I

- 0.0209

EC,= 0.023%C 0.8

I

215

-

-

0.2

I

1

I

I

Clay Content, % Figure 5. Correlations between EC, and clay percentage for a number of soils from the San Joaquin Valley of California. (After Rhoades et af.,1989a.)

circles represent experimental data. EC, represents the EC of the equilibrating water or the water expressed from the soil by pressure filtration. The soil had been extensively leached with waters of different salinities (EC,), therefore EC, was essentially equal to EC, and to EC, under the conditions of this experiment. The data and model relations also show that the ability to determine accurately EC, (or EC,) from EC, decreases as 0, decreases. This is so because the required accuracy of measurement of EC, becomes limiting as the EC, =f(EC,) relation flattens at low values of 0,. At very low values of 0, (ZO.l), it is not possible to determine EC, (or EC,) from EC, at all (see Rhoades et al., 1976). To use Eq. ( 5 ) or (6) to assess soil salinity (EC, or EC,) from EC,, the values of EC, , Ow, and 0, must be known. EC, and 0, can be estimated using Figs. 5 and 6, respectively. The means of obtaining these relations are described elsewhere (Rhoades et al., 1989a). The value 0, can be measured in the field using time domain reflectometry (TDR) methods or it can be adequately estimated, for many practical purposes, by “feel.” The TDR method is described later. The value 0, can be estimated from bulk density (pB) as 0, = &/2.65, where 2.65 is a reasonable estimate of the average particle density of most mineral soils. Equation (5) may be solved for EC,, with the assumption that EC, = EC,, by arranging it in the form of a quadratic equation and solving for its

J. D. RHOADES

216

Volumetric Content of Soil Water, 8, Figure 6. The volumetric content of soil water in a series path as a function of the total water content for various soils. (After Rhoades el a/., 1989a.)

positive root:

EC,

=

-b+dFXG

+

where a = (O,)(Ow- Om), b = (0, Om)z(EC,) (O,EC,), and c = O,EC,EC,. If EC, is desired, it can be obtained from

(ECwOw)= (EC,O,

(7)

2a

+ EC,O,)

+ (Ow - O,)(O,EC,)

= EC,(SP/lOO)p,

-

(8) where SP is the gravimetric water content of the saturation paste expressed as a percentage and p~ is the bulk density of the soil (Rhoades et al., 1989a). The relations described above give the opportunity to distinguish between the “mobile” and “immobile” contents of soil salinity. Because the soil water that plants draw on to meet transpiration requirements is primarily 0, and because the salinity effect is primarily one of reducing water availability to the plant, the salt in the mobile water phase, EC,, ideally should be used as a basis for relating crop responses to soil salinity. In as much as EC, is the dominant contributor to EC, in saline soils, it may be

MEASURING AND MAPPING SOIL SALINITY

217

determined from Eq. (6) for a given value of EC, using an intercept term estimated from the relations of Figs. 5 and 6. This procedure is applicable 2 dS/m. For conditions of lower salinity, the intercept term where EC, of Eq. [ 5 ] must be used. The appropriate y intercept value, and thus EC,, may be determined using successive approximation techniques. Because of the linear relationship [Eq. ( 6 ) ] that exists at a fixed water content (such as field capacity) between EC, and EC, at significant values of EC, (or EC,), soil salinity (either EC, or EC,) may be empirically related to EC, by an expression of the following type: EC, (or EC,) = m(EC, - EC:) (9) where EC: is [(O, 8,)2/8s](EC,), or -EC,, and is predicted as previously described (i.e., using Figs. 5 and 6), and rn is the slope of the EC, (or EC,) =f(EC,) relation at field capacity. A typical linear relationship between EC, and EC, of this type is shown in Fig. 7. Analogous relations

+

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Electrical Conductivity of Soil, ECa (dS/m)

Figure 7. The relation between bulk soil electrical conductivity and the electrical conductivity of the saturation paste extract (soil salinity) for Dateland soil at field capacity content. (After Rhoades, 1981.)

218

J. D. RHOADES

have been developed between an EC, and EC, (Rhoades, 1980). With such calibrations, one can predict EC, (or EC,) from EC, for field soils of various types, provided they are at or near field capacity water content at the time of EC, measurement. Simple calibration procedures have been developed in order to obtain the calibration relations of Eq. (9) appropriate to field soils with their natural structures, pore size distributions,and water retention properties (Rhoades, 1976, 1980, 1981; Rhoades and Ingvalson, 1971;Rhoades and van Schilfgaarde, 1976; Rhoades et al., 1977). Numerous satisfactory field calibrations (as in Fig. 7) have been obtained for many soils around the world and they have been found to be very similar for soils of similar textures (Rhoades and Ingvalson, 1971;Halvorson and Rhoades, 1974; Rhoades, 1976, 1979, 1980, 1981; Halvorson et a/., 1977; Rhoades et al., 1977;Yadau et al., 1979;Loveday, 1980; van Hoorn, 1980; Nadler, 1981; Bohn et al., 1982). It has been shown that the calibrations [slopes and intercepts of Eq. (9)] can be predicted from soil properties such as texture (or percentage clay content) and saturation percentage (or field capacity) water content (Rhoades, 1981; Rhoades et al., 1989a). For more information on methods for calibrating EC, =f(EC,), see Rhoades (1976) and Rhoades et al. (1977). The effect that varying the soil water content (i.e., departure from Calibration water content) has on salinity determined from Eq. (9) depends on whether or not salt loss from the soil occurs with the change of water content (Rhoades et al., 1981). Immediately following an irrigation, salt loss occurs as the water drains to field capacity; hence, EC, is very sensitive to changes in 0, during such times. Afler the rapid drainage ceases and the soil is at field capacity, further major losses of soil water in cropped soils occur mainly through evapotranspiration. Almost all of the salt in the water taken up by the plant root system is excluded from entering by the root membranes; the salt is left behind in the remaining water. Likewise, no salt is lost through evaporation. Hence, the salt concentration (or electrical conductivity) of the remaining soil water is increased proportionately as 0, is reduced by evapotranspiration. Because of this inverse proportional relationship between EC, and Ow, the product EC,0, found within a given soil volume at field capacity will not change appreciably as 0, is reduced below field capacity; thus, the product EC,0, may be considered (ignoring salt precipitation) as approximately constant at water contents of field capacity or less. However, changes in 0, do affect EC, through its influence on the partitioning of 0, and Om, as shown in Fig. 6. As 0, decreases below field capacity due to evapotranspiration, EC, will show an approximately linear decrease according to the relationship where (Y

A EC, = aA0,K (10) is a factor related to the relation between 0,, and , 8 and K =

MEASURING AND MAPPING SOIL SALINITY

219

EC, 0, = a constant. For typical soils the error in EC, caused by A 0, is not large, with reasonable deviation in 0, from field capacity water content. Experimental data supporting these conclusions about the relatively insignificant effects of A 0, on AEC, and AEC, during and following an irrigation are given by Rhoades et al. (198 1). The appropriateness of using 0, as a reference for water content in establishing EC, =f(EC,) calibrations is supported by the results of Bottraud and Rhoades (1985). It is apparent from the above discussionsthat EC, is primarily a measure of the total dissolved salt in a soil on a volumetric basis. Hence, given the value of (EC, 0,) at field capacity, an estimate of EC, at any lower water content can be calculated from Eq. (8). Hence, the limits of EC, (or osmotic potential) over an irrigation cycle can be ascertained (estimated) from EC, (and hence from EC,). For many such practical applicationsthis procedure can be used in place of Eqs. (5) and (6) to estimate EC, from EC,. An equation analogous to Eq. (5) established for bulk soil electrical conductivity exists for saturated soil pastes, as follows:

where EC, is as defined previously, EC, is the electrical conductivity of the saturated paste, 0, and 0, are the volume fractions of total water and solids in the paste, respectively, ,8 is the volume fraction of water in the paste that is coupled with the solid phase to provide a series-coupled electrical pathway through the paste, EC, is the average specific electrical conductivity of the solid particles, and the difference e, - e, is e,, which is the volume fraction of water in the paste that provides a continuous pathway for electrical current flow through the paste (a parallel pathway to 0,). Assuming the average particle density ( p , ) of mineral soils to be 2.65 g/cm3 and the density of saturation soil paste extracts (p,) to be 1.00, 6, and 0, for saturated pastes can be directly determined from SP as follows:

and

The saturation percentage of many mineral soils can be adequately estimated in the field for purposes of salinity appraisal from the weight of a paste-filled cup of known volume (Rhoades et al., 1989b). Figure 8 may be used for this purpose; for details of the relations inherent in this figure, see Wilcox (195 1).

220

J. D. RHOADES

CUP

GRAMS PASTE Figure 8. Theoretical relation between saturation percentage (SP) and weight (in grams) of 50 cm3of saturated paste, assuming a particle density of 2.65 g/cm3.(After Rhoades et al., 1989b.)

B. DETERMINING SOILSALINITY FROM SATURATED PASTE ELECTRICAL CONDUCTMTY EC, can be determined from measurements of EC, and SP [using Eqs. (1 1) - ( 13), if values ofp,, Ow, and EC, are known. These parameters can be adequately and simply estimated, as demonstrated by Rhoades et al. (1989b,c). For typical arid land soils of the southwestern United States, ps may be assumed to be 2.65 g/cm3; EC, may be estimated from SP as EC, = 0.019SP - 0.434, and the difference 0, - Om may be estimated from SP as Ow -,8 = 0.0237(SP).0-M57 The measurement of EC, and SP

MEASURING AND MAPPING SOIL SALINITY

22 1

can be easily made using an EC cup of known geometry and volume. The method is suitable for both laboratory and field applications, especially the latter, because the apparatus is inexpensive, simple, and rugged and because the determination of EC, can be made much more quickly than that of EC, . In this method a saturated soil paste is made as described previously and is then placed in a conductivity cup of known volume. From the weight of the paste, SP is determined; from the conductance, EC, is determined. Then EC, is obtained from Fig. 9 given EC, and SP, using the curve corresponding to the SP value, or else it is calculated using the following equation:

EC,

=

-b+-

2a

+

+

- O,EC,, where a = O,(Ow - OA, b = (0, 0,,,J2EC, (Ow - O,)O,EC, and c = - O,EC,EC,. The values of EC,, O,, Ow, and 0, are estimated from SP using Eqs. ( 12) and ( 13) and the relations given above. Sensitivity analyses and tests have shown that the estimates used in this method are generally adequate for salinity appraisal purposes of typical mineral arid-land soils of the southwestern United States (Rhoades et al., 1989~).For organic soils or soils of very different mineralogy or magnetic properties, these estimates may be inappropriate. For such soils, appropriate values for p,, EC,, and 0, will need to be determined using techniques analogous to those of Rhoades et al. (1989b). The accuracy requirements of these estimates may be evaluated using the relations given by Rhoades et al. (1 989c). It should be noted that EC,O, is not equivalent to ECJ, because different amounts of soil are involved in the two measurements. The relation between these two products is

Ecw~w/Pb= EC,~,/p, (15) Data to support this are given by Rhoades (1981) and Rhoades et al. (1990b). The ratio OJp, is equivalent to SP/lOO (see Rhoades et al., 1989a,b).

C. DETERMINING SOILSALINITY FROM BULKSOILELECTRICAL CONDUCTIVITY Soil salinity can be determined from bulk soil electrical conductivity by essentially one of three ways. Here we discuss these alternative methods of salinity appraisal, but first the various instrumental means of measuring EC, will be briefly reviewed.

J. D. RHOADES

222

€

10

80

\

I I

I

v)

U

90 I ' 100 I I I

6

aJ

0

w

I

6

I

1 4

I I I I

2 c

I

I

I

I

w

1 2 I

*0

X

C

0 c

I

I

I

I

I

I-

2a

c Q v) v-

0 >I

zw ,-

c 0

a

U

C

0

0

Q 0

I-

L

c 0 Q,

iii 0

2

4

6 8

10 12 14 16 18 20

Electrical Conductivity of Saturation Paste, EC,,,

dS/m

Figure 9. Relations between electrical conductivity of saturated soil paste (EC,), electrical conductivity of a saturation extract (EC,),and saturation percentage(SP),for representative and-land soils. (After Rhoades et al., 1989b.)

MEASURING AND MAPPING SOIL SALINITY

22 3

1. Sensors for Measuring Bulk Soil Electrical Conductivity

Three types of soil conductivity sensors presently exist that are capable of measuring bulk soil electrical conductivity. Two are field-proven, portable sensors that are now commercially available: (1) a four-electrode sensor and (2) an electromagnetic induction sensor. A third sensor, based on time domain reflectometry technology, has shown good promise and utility in certain experimental applications. Each method has its own advantages and limitations. a. Four-Electrode Units Bulk soil electrical conductivity can be measured using four electrodes inserted into the soil, a combination electric current source/resistance meter, and connecting wire. Such a surface array of electrodes and a generator/meter unit are shown in Fig. 10. The current source-meter unit may be either a hand-cranked or a battery-powered type. Units designed for geophysical purposes generally read in ohms and, if used for general soil salinity appraisal, should measure from 0.1 to 1000 ZZ. A commercially

Figure 10. A “fixed-amy” four-electrode apparatus and commercial generator/meter. (After Rhoades, 1992a.)

2 24

J. D. RHOADES

available unit designed specifically for soil salinity appraisal is battery powered and reads directly in dS/m (Fig. 10). Electrodes used in surface arrays can be made of stainless steel, copper, brass, or almost any other corrosion-resistant metal. Array electrode size is not critical, except that the electrode must be small enough to be easily inserted to a depth of 5 cm or less. Electrodes 1.O to 1.25 cm in diameter by 45 cm long are convenient for most surface array purposes, although smaller electrodes are preferred for determining EC, within soil depths of less than 30 cm. Any flexible, well-insulated, multistranded, 12- to 18gauge wire is suitable for connecting the array electrodes to the meter. For hand-carried survey or traverse work, the array electrodes may be mounted in a board with a handle (see Fig. 10) so that soil resistance measurements can be made relatively quickly for a given interelectrode spacing (Rhoades, 1976).These “fixed-array” units save the time involved in spacing the electrodes. A mobilized, tractor-mounted version of a fixedarray unit, which makes automated on-the-go measurements and which includes data-logging and satellite-based, site-positioning equipment, has recently been built (Rhoades, 1992a,b);this unit is capable of making both faster and more widely spaced readings then can be accomplished manually, while simultaneously providing the x and y coordinates of each measurement site (see Fig. 11). A four-electrode salinity probe, in which the electrodes are incorporated into a probe (Rhoades and van Schilfgaarde, 1976), is used for small soil volume measurements. Conveniently sized current source-meter units have been designed for use with the four-electrode salinity probe (Austin and Rhoades, 1979).Commercial versions of both the four-electrodeprobe and the meter are made by Martek Instruments.* The newest version of the Martek SCT system, which reads directly in EC, corrected to 25°C and incorporates a data logger and a timer, is shown in Fig. 12. b. Electromagnetic Induction Unit Soil electrical conductivity can be measured remotely using electromagnetic induction (EM) methodology. The basic principle of operation of the EM soil electrical conductivity meter is shown schematically in Fig. 13. An EM transmitter coil located in one end of the instrument induces circular eddy-current loops in the soil. The magnitude of these loops is directly proportional to the electrical conductivity of the soil in the vicinity of that loop. Each current loop generates a secondary electromagnetic field that is proportional to the value of the current flowing within the loop. A fraction

* Mention of trademark or proprietary products in this manuscript does not constitute a guarantee or warranty of the product by the U.S. Department of Agriculture and does not imply its approval to the exclusion of other products that may also be suitable.

MEASURING AND MAPPING SOIL SALINITY

225

Figure 11. An automated mobile (tractor-mounted) “fixed-array’’four-electrode system. (After Rhoades, 1992a.)

Figure 12. Two commercial four-electrode soil conductivity probes (small and standard sizes) and generator meter/data logger. (After Rhoades, 1992a.)

226

J. D. RHOADES

T

- TRANWITTER Coil -

R RECEIVER COIL

INDUCED CURRENT FLOW IN GROUND

Figure 13. Diagram showing the principle of operation of an electromagnetic induction soil conductivity sensor. (After Rhoades and Corwin, 1981.)

of the secondary induced electromagnetic field from each loop is intercepted by the receiver coil, and the sum of these signals is amplified and formed into an output voltage that is linearly related to depth-weightedsoil EC,, ECZ. Figure 14 shows a commercially available EM soil salinity sensor (Geonics EM-38) in the vertical and horizontal (coils) positions. This device was designed to meet the general-purpose needs of soil salinity appraisal. The EM-38 device contains appropriate circuitry to minimize instrument response to the magnetic susceptibility of the soil and to maximize response to EC,. It has an intercoil spacing of 1 m, operates at a frequency of 13.2 kHz, is powered by a 9-V battery, and reads EC:directly. The coil configuration, frequency and intercoil spacing were chosen to permit measurement of EC: to effective depths of approximately 1 and 2 m when placed at ground level in horizontal and vertical configurations, respectively. Other available “geophysical” units are capable of deeper measurement. Mobilized, automated EM measurements (including data logging) can be made “on the go” using the recently built (Rhoades, 1992a,b) EM-sensing system shown in Fig. 15. As with the mobile, fourelectrode system, this system also incorporates synchronized, satellitebased, site-positioningequipment (including data logging). c. Time Domain Reflectometry Unit Both the electrical conductivity and dielectric constant (hence water content) of the soil can be measured using time domain reflectometry methodology. With the TDR method, the apparent dielectric constant of

MEASURING AND MAPPING SOIL SALINITY

227

Figure 14. The Geonics EM-38 electromagnetic induction soil conductivity -sensor. (After Rhoades, I992a.)

Figure 15. An automated mobile electromagnetic induction soil conductivity sensing system. (After Rhoades, 1992a,b.)

228

J. D. RHOADES

the soil, E, is obtained using Eq. (16) by measuring the transit time, t, of a voltage pulse applied to a parallel transmission line (dual-rod probe) of length L embedded in the soil of electrical conductivity EC,, where c is the velocity of light in a vacuum:

(16) The signal is attenuated in proportion to EC, so that the transmitted voltage, VT, is reduced according to: E = (ct/2L)2

VR = VTexp(-2arL) (17) where a! is an attenuation coefficient, which increases linearly with EC, as

a = 60EC,/ & (18) The TDR equipment capable of making the above measurements is shown in Fig. 16. Soil water content has been shown to be correlated with E (Topp et al., 1980, 1982, 1984) and EC, and EC, have been shown to be corre-

Figure 16. A TDR probe and meter. (After Rhoades and Oster, 1986.)

MEASURING AND MAPPING SOIL SALINITY

229

lated with VJV, (Dalton et af.,1984; Dasberg and Dalton, 1985; Dalton and van Genuchten, 1986), as measured by TDR techniques. The use of TDR for measuring soil salinity is relatively new (it has been used mostly in laboratory studies); the practical attributes of the method for general field use in salinity appraisal cannot be fairly judged at this time. However, it offers the potential advantage of measuring both water content and soil electrical conductivity simultaneously. This is its chief attraction. Of course, it has not be mobilized like the four-electrode and EM methods have been and is likely less well-suited for mapping purposes. For a detailed discussion of the theory and applicationsof TDR in salinity assessment, see Dalton et af. (1990). 2. Procedures for Measuring Bulk Soil Electrical Conductivity

a. Large-Volume Measurements For the purpose of determining soil salinity of entire root zones, or some fraction thereof, it is desirable to make the measurement of EC, within a depth of 1 to 1.5 m. This is usually accomplished with the four-electrode equipment by configuring the surface array of electrodes in a straight line with the spacing between the two current (outer) electrodes selected depending on the desired depth@).The relative spacing between the inner electrode pairs can also be varied. The electrodes are often spaced in the so-called Wenner array, with equal spacings between all of them (Rhoades and Ingvalson, 1971). When using the Martek SCT meter, each of the inner pair of electrodes is preferably placed inward from its closest outerpair counterpart a distance equal to 1OYo of the spacing between the outer pair. In both of the above arrangements, as well as for others, the inner pair of electrodes is generally used to measure the electrical potential (or resistance) while current is passed between the outer pair. The effective depth of current penetration for either configuration (in the absence of appreciable soil layering) is approximately equal to about one-third the outer electrode spacing, y; “average” soil salinity is measured to approximately the depth y / 3 (Rhoades and Ingvalson, 1971; Rhoades, 1976; Halvorson and Rhoades, 1976). Thus, by varying the spacing between current electrodes, one can measure soil salinity to different depths and within different volumes of soil. An advantage of this “surface-array” method is the relatively large volume of soil that is measured compared to that measured by insertion four-electrode probes (discussed later) or by using customary soil samples. The volume of measurement is about ( ~ y / 3 where ) ~ , y is as defined above.

230

J. D. RHOADES

Hence, effects of small-scale variations in field soil salinity can be minimized by these relatively large-volume measurements. For measurements taken in the Wenner array (electrodesequally spaced) using geophysical-typemeters, which measure resistance, bulk soil electrical conductivity is calculated (in dS/m) as EC, = 159.2f,/aRt (19) where a is the distance between the electrodes in cm, R, is measured resistance in ohms at the field temperature t, f , is the previously described temperature compensation factor used to adjust the reading of EC, to a reference temperature of 25"C, and 159.2 is the numerical equivalent of 1000/2~.For measurements made with the Martek SCT meter, a factor (cell constant) is supplied in its instruction manual for each spacing of outer electrodes; this factor is entered into the meter and the correct soil EC, reading is directly displayed in the meter readout. Relatively large volumes of soil are also measured with the electromagnetic induction technique. The volume and depth of measurement are influenced by the spacing between coils, the current frequency, and the orientation of the axes of the magnets/coils with respect to the soil surface plane. The effective depths of measurement of the Geonics EM-38 device are about 1 and 2 m when it is placed on the ground and the coils are positioned horizontally and vertically, respectively. The effective width of the measurement extends out about 0.5 m to the sides and ends of the unit. The EM-38 device does not provide a linear measure of EC, with depth, rather a depth-weighted value EC,*is obtained as stated earlier. The soil depth intervals of 0 to 0.3, 0.3 to 0.6, 0.6 to 0.9, and 0.9 to 1.2 m contribute about 43,2 1, 10, and 696, respectively, to the ECZreading of the EM unit when it is positioned on homogeneous ground in the horizontal position (Rhoades and Corwin, 1981). Thus, the weighted bulk soil electrical conductivity read by the EM device in this configuration is approximately EC:=

0.43EC,o-o.3

+ 0.21EC,o.34, + O.lOEC,o.,_o.,

+ 0.06EC,0.9-,.2 + 0.2EC,,,,

(20)

where the subscript designates the depth interval in meters. Corresponding percentages in the vertical position are 17, 21, 14, and 10, respectively.* Recent studies show that these proportions do not hold for nonhomogeneous profiles (Rhoades el al., 1990a).

* The relative contributions (R) to the secondary EM field (or ECg from all material below a depth Z can be theoretically calculated from Rv = 1/(4Z2 ])In,and R, = (4Z2 I)*'z - 22, for the vertical (V)and horizontal (H) dipoles, respectively (McNeil, 1980).

+

+

23 1

MEASURING AND MAPPING SOIL SALINITY

It is desirable to be able to determine soil EC, within various depth intervals so that soil salinity can be calculated within the various parts of the root zone as needed for making assessments and management decisions. Because the proportional contribution of each soil depth interval to EC:, as measured by the EM unit, can be varied by changing the coil orientation or, as shown by Rhoades and Corwin (198 I), by raising the unit above the ground to various heights, it is possible to estimate EC, by depth within the soil from a succession of EM measurements made either at various orientations, or at heights above ground, or both. The EC, values within different discrete soil depth intervals have been shown to be correlated with a succession of EM, readings made above ground as follows: ECa,0-0.3

=POEM,

+PIEM1

+P2EM2

+hEM3

(2 1a)

+P4EM4

+

EC4~.3-o,,5=~oEMo+ 71EM1 + 72EM2 + 73EM3 + 74EM4 * . * (21b) where EM represents the reading obtained with the EM-38 unit held in the horizontal position and 0, 1, 2, 3, and 4 represent height above ground in increments of 30 cm. The values of the coefficients reported by Rhoades and Corwin (1981) for Eq. (2 1) have not been widely tested and likely vary for different profile types (i.e., EC, depth patterns). Another series of empirical equations and coefficients have been developed to estimate EC, within discrete soil depth intervals from just two measurements made with the magnetic coils of the EM-38 instrument positioned at ground level, first horizontally and then vertically (Corwin and Rhoades, 1982, 1984). For a given depth increment, x l to x2, equations (as revised by Rhoades et al., 1989d) of the form

+

ECE::-x2 = kHEMa2’ kVEMt2’k k 3 (22) have been used in this regard, where EM, and EM, are the readings obtained with the EM-38 device positioned at the soil surface in the vertical and horizontal positions, respectively; x l to x2 is the soil depth increment in centimeters and k,, k,, and k3 are empirically determined coefficients for each depth increment. These four readings are more practical to obtain than those required in Eq. (21). Equation (22) is also more easily solved than is Eq. (21) and the results are almost as accurate for the two depth intervals 0 - 30 and 30 - 60 cm. The following improved and more general relationship has been developed more recently for both different soil depths and types of EC, profiles (Rhoades, 1992a): In EC, = Po PI In EM, P3(ln EM, - In EM,) (23)

+

+

where Po, PI, and p3 are empirical coefficients. In the earlier uses of Eq.

J. D. RHOADES

232

(22), two profile types were distinguished based on EMH/EMv ratiosregular (EM, > EMH) and inverted (EM, > EM,). Equation (22) and its manner of use suffers at least three deficiencies. It does not correct for nonlinearity in the EM,-EC, and EM,-EC, relations that occur at high values of EC, (see Corwin and Rhoades, 1990), it forces near-uniform profiles into either regular or inverted types, and it does not recognize the colinearity that exists between EM, and EMV (see Lesch et al., 1992). Equation (23) minimizes these deficiencies by separating soil profile types into three classes (regular, uniform, and inverted), by including the curvilinear forms of the uniform EM,-EC, and EMV-EC, relations and their differences to identify the three profile types, and by using the difference In EMH - In EMVin place of EM, as the second variable in the relation to minimize the colinearity problem. The theoretical relation between In EM, and In EM,-In EMV for uniform EC, profiles is shown in Fig. 17. The fitted curve (In EMH - In 0.00836EMH) describes a theoretiEMv = 0.04334 0.03058 In EM, cally uniform EC, profile. Profile types may be classified based on deviation from this relation. For the practical purposes of solving Eq. (23), the profile types have been classified as follows (Rhoades, 1992a): sites having values of In EM, - In EM, within +5% of the theoretical value (i.e., 0.04334 0.03058 In EM, 0.00836EMH) are designated “uniform”; those with measured values >5% of the theoretical are designated “inverted” and those with measured values < 5% of the theoretical are designated “regular.” Empirically determined values of the coefficients for Eq.

+

+

+

+

3.0

I

1.5 h

-1.5

-3.0 -0.2

0.0

0.2

0.4

In(EMH)-ln(EMv)

Figure 17. Theoretical relation between In EM, and the difference In EM, - In EM, for uniform EC, profiles. (After Rhoades, 1992a.)

23 3

MEASURING AND MAPPING SOIL SALINITY Table I

New Relations for Predicting Soil Electrical Conductivity" Depth (cm)

n

rz

- In EM,) - In EM,) - In EM,)

650 626 200

0.756 0.753 0.688

- In EM,)

73 70 24

0.806 0.815 0.810

56

0.908 0.811 0.906

Equation for electrical conductivity For regular profiles

0-30 30-60 60-90

In EC, = 0.414 In EC, = 0.836 In EC, = 0.674

+ 0.985 In EM, + 2.336(1n EM, + 1.262 In EM, + 1.307(ln EM, + 1.089 In EM, - 0.446(1n EM,

0-30 30-60 60-90

In EC, = 0.478 In EC, = 0.699 In EC, = 0.477

+ 1.209 In EM, + 0.41 l(ln EM, + 1.234 In EMH- 0.623(1n EM, + 1.053 In EM, - 0.691(1n EM,

0-30 30-60 60-90

In EC, = 0.626 InEC,=0.881 In EC, = 0.563

+ 1.239 In EMH+ 0.325(1n EM, - In EMV) + 1.2161nEMH- 1.318(lnEMH-hEM,) + 1.206 In EM, - 1.641(ln EM, - In EM,)

For uniform profiles

- In EM,) - In EM,)

For inverted profiles

55

21

" Predictions are for conductivity within different soil depth increments from the electromagnetic measurements made with the EM-38 device, placed on the ground in the horizontal (EM,) and vertical (EM,) configurations. After Rhoades (1992a). (23) based on these classification criteria observed in a variety of California soils are given in Table I. For more discussion of the theory and calibration of EM and EC,, see Corwin and Rhoades ( 1990). For more information on the effects of EC, depth patterns on EM readings, see Rhoades et al. (1990a). For more information about the basis of Eq. (23), see Rhoades (1992a). b. Small-Volume Measurements Sometimes information on salinity distribution within a small, localized volume of soil is desired, such as that within different sections of the seedbed or under the furrows. For such conditions, the insertion fourelectrode salinity probe (Rhoades and van Schilfgaarde, 1976) is recommended. In the standard-sized probe (see Fig. 12), four annular rings are molded in a plastic matrix that is slightly tapered so that it can be inserted into a hole made to the desired depth with a Lord-type coring tube (or one of similar diameter). In the smaller unit (see Fig. 12), the probe is simply pushed into the soil to the desired depth. In either sized unit, the probe is attached to a shaft (handle) through which the electrical leads are passed and connected to a meter. Burial-type units are also available in which the leads from a handleless probe are brought to the soil surface (Rhoades,

234

J. D. RHOADES

1979b). A multiple-depth version has been built by Nadler et al. (1982). The volume of sample under measurement with any of these probe sensors can be varied by changing the spacing between the current electrodes and the overall diameter of the probe. The standard-sized Martek SCT Probe has a spacing of 6.6 cm and measures a soil volume of about 2350 cm3. The Martek “bedding” probe measures a soil volume of about 25 cm3. When using meters that display resistance, EC, (in dS/m) is calculated as follows: EC, = kf,/R, (24) where k is an empirically determined geometry constant (cell constant) established for the probe in units of 1000 cm-’, R , is the resistance in ohms at the field temperature, andf, is the factor used to adjust the reading to a reference temperature of 25°C [see Eqs. (2) and (3)]. With the Martek unit, EC, is directly displayed by the meter either at field temperature or at 25 “C. For more detailed descriptions of the four-electrode and EM equipment and their manner of use for salinity appraisal, see Rhoades (1990b, 1992a) and Rhoades and Miyamoto ( 1990). To determine soil EC, with a TDR unit, the parallel pair of wave guides (see Fig. 16) is inserted into the soil and the time and voltages displayed on the meter are determined. EC, and dielectric constant are then calculated using Eqs. (16)-( 18). For a detailed description of this equipment, its calibration, and its use for salinity appraisal, see papers by Dalton and colleagues (Dalton and Poss, 1990; Dalton el al., 1990). 3. Salinity Interpretation

Soil salinity, in terms of either EC, or EC,, can be determined from bulk soil electrical conductivity by essentially one of three ways. Each has its own advantages and disadvantages. These alternative ways will now be described. a. “Soil-Type Calibration” Technique Soil salinity (EC, or EC,) can be determined from measurement of EC, made at a reference soil water content using a calibration in the form of Eq. (9) established for the particular soil in question, or at least sufficiently similar to it. These soil-type calibrations should be established so as to apply to field soils with their natural structures, pore size distributions,and water retention properties. As discussed in Section III,A, various methods have been developed in this regard and numerous satisfactory field calibrations (9> 0.9) have been obtained for many soils around the world and used subsequently to diagnose and map salinity; very similar calibrations

MEASURING AND MAPPING SOIL SALINITY

235

have been obtained for soils of similar textures. Because water content (as well as salinity) affects soil electrical conductivity, the calibrations and determinations of EC, are made preferably when the soil is near field capacity. However, measurements and salinity appraisals can be made at lower water contents that exceed a certain minimum level. This minimum water content is about 10%on a gravimetric basis, but it may be somewhat higher for very sandy soils (Rhoades et al., 1976). For irrigated soils, measurements and calibrations ideally should be made after irrigation when the soil water content is at field capacity. This water content is sufficiently reproducible for such practical calibrations. Under dryland conditions, calibrations/measurements should be made in early spring or on fallow land, in order to take advantage of the relative uniformity of soil water that exists under such conditions. Given the value of EC, or (EC,) at field capacity, one can readily estimate EC, at lower field water contents using Eq. (8); hence the limits of EC, (or osmotic potential) occurring over the irrigation cycle can be ascertained and used for the practical prognosis of crop response to varying in situ soil water salinity. For many such practical applications, the “compensation” phenomenon [implied in Eq. (8) and discussed earlier] precludes the need to be able to measure EC, per se, or to be too concerned about measurements of EC, having to be made exactly at calibration water content. b. “Model/Field-Estimates” Technique Soil salinity can be determined for any soil and water content in excess of a threshold value from measurement of EC,, estimates of soil clay percentage and percent water content relative to field capacity using Eqs. (7) and (8) (i-e.,the bulk soil electrical conductivity model and the relation between EC, and EC,) and the empirical relations given below to estimate pb, EC,, and 0,. The advantage of this model/field-estimates method is that it accounts for the site-to-site variabilities in soil properties (clay and water content, in particular) that occur within typical fields or in other areas of interest, Essentially, it generates a specific calibration between EC, and EC,, or EC,, for the particular soil condition encountered at each site of EC, measurement in the field (area) under evaluation. Such “specific” calibrations are more generally accurate than are the “average soil-type” calibrations (the previously discussed method) when applied to the whole of the area, assuming it to be the same as the calibration soil-type (Rhoades et al., 1990b). Field tests of this method have shown it to be sufficiently accurate for the practical purposes of salinity diagnosis and mapping, to be faster than conventional soil sampling and laboratory methods (mea-

236

J. D. RHOADES

surement of EC, per se, either directly or as estimated from EC,), and to be generally more accurate than the soil-type calibration method (Rhoades et al., 1990b). The disadvantage of the method is that one must estimate (by feel) the clay percentage and relative water content of the soil at each site and depth of measurement. This requires that the soil be probed at every measurement site/depth and time taken to make these estimates. The relations used in the practical application of the soil EC, model/ field-estimates method to appraise soil salinity are (after Rhoades et al., 1990b)as follows: SP = 0.76(9/oC) pb =

+ 27.25

1.73 - 0.0067SP

(25) (26)

e, = pb12.65

(27)

0,

(29)

e,

= Ode( FC/ 100)

= 0.639

e,

+ 0.01 1

(30)

and

EC, = 0.0 19SP - 0.434

(31) where %C is clay percentage as estimated by “feel” methods, 8,, is the corresponding estimated volumetric water content at field capacity, and FC is the percent water content of the soil relative to that at field capacity, also as estimated by “feel.” Given the above assumptions, estimates, and measurement of EC,, EC, is calculated from the solution of Eq. (7). Then EC, is determined from Eq. (S), assuming that EC, = EC, and, therefore, that EC,O, = ECwceWc EC,O,. Alternatively, EC, can be obtained graphically using data given by Rhoades ( 1990a) or Rhoades and Miyamoto ( 1990). Sensitivity analyses and results of field tests have shown that the estimates and assumptions described above are generally adequate for practical salinity appraisal purposes of typical mineral, arid-land soils of the southwestern United States (Rhoades et al., 1989c, 1990b); i.e., that EC, can be estimated in the field sufficiently accurately for most salinity appraisal purposes from the accurate measurement of EC, and from reasonable field estimates of %C and FC made by “feel.” For organic soils or soil of very different mineralogy or magnetic properties, these estimates may be inappropriate. For such soils, appropriate estimating procedures will have

+

MEASURING AND MAPPING SOIL SALINITY

237

to be developed using techniques analogous to those used by Rhoades et af. (1989b). The accuracy requirements of these estimates may be evaluated using the relations given by Rhoades et af. (1989~). If more accurate determinations of EC, or EC, than can be obtained by the estimation procedures described above are required, then quantitative measurements of S,, EC,, pb, etc. should be made using appropriate methods and used in place of the above-described estimates. An alternative model procedure for determining salinity from EC, at various water contents is that of Nadler (1982). This procedure “curve-fits” what amounts to an f(0) relation using moisture release data established for the particular soil in question and an empirical “effective porosity” relation based on A& To date, the method has been successfully applied only to disturbed soil samples, it requires considerable laboratory effort to establish the empirical fit, it only applies to the “fitted” soil, and its applicability to field soils is not generally good (unpublished data). c. “Regression Model/Ground-Truthing” Technique A statistical/ground-truthing approach, which establishes a predictive (regression) relation between EC, and EC, (or EC,), can be used to determine salinity in an area of land that is relatively homogeneous with respect to soil conditions other than salinity. This approach includes analyzing for salinity a small subset of soil samples collected from within the “homogeneous” area at a fraction of the sites where EC, (or EM) has been measured. It is the basis of the integrated-instrument method of salinity mapping discussed in Section III,C,5. To date this technique has been used primarily with EM methods, which provide depth-weighted measurements of EC, (actually, surface-array, four-electrode techniques do, as well), thus it will be described herein in terms of EM. However, the principle applies equally well to other instrumental techniques for measuring EC,. In this approach, numerous EM readings are obtained within the sampling area (usually a field) under evaluation on a uniform (centric systematic) grid basis. Based on the observed field pattern of EM readings, a relatively small number of EM measurement sites are chosen for soil sampling using a statistical model/ procedure described in detail elsewhere (Lesch et af., 1992). Soil samples are collected at these sites and the salinities of these soil samples are determined by any accepted method of salinity appraisal [the method of Rhoades et al. (1989b)is recommended for this purpose]. A multiple linear regression relation between the EM readings and measured soil salinities (EC, , EC,, or some other expression) within each soil depth of interest is

J. D. RHOADES

238

then established, i.e.,

+

log(EC,) = Bo B,[log(EM,)] + &[log( EM,) - log(EM,)] f 3 3 (32) and subsequently used to predict the salinities at the vast number of unsampled sites/depths in the area where the remainder of the EM measurements were made. This “single-step” method eliminates the need to first convert EM (essentially EC,Y, to EC, within a particular soil depth using general relations [such as those in Table I, which are based on Eq. (23)] and thence to EC,, as is required in soil-type calibration and model/ field-estimates methods. Experimental results suggest this method works well for landscapes that are relatively homogeneous in all factors affecting EC, conditions other than salinity, such as individual fields uniformly managed or sections of natural landscapes that have similar soil types and properties (certain dryland landscapes, for example). This approach substitutes easily acquired EM field measurements for the more difficult procedures of soil sampling and their laboratory analysis for soil salinity. It very substantially reduces the number of soil samples needed to map accurately and intensively the spatial salinity patterns within fields. Larger areas of land are mapped by joining adjacent areas on a field-by-field basis. This method is more practical than the geostatistical procedures that are traditionally used for salinity mapping purposes (Webster, 1985, 1989), because it reduces the intensive soil sampling generally required to obtain the accurate variogram estimates needed in these procedures (unpublished data). The major limitation of the method is the requirement that the fields be under uniform management and that soil water, bulk density, and clay content be reasonably homogeneous. If needed, however, larger fields (or areas) can be subdivided into smaller more homogeneous units and the method applied analogously to each sufficiently homogeneous subunit. Alternatively, practical measurements in addition to EM, such as location coordinates, elevation, etc., can be made and incorporated into the regression relation [as coefficient B, in Eq. (32)] to adjust for some of the “other” factors influencing the salinity prediction (Lesch et al., 1992). Disadvantages of the method are the need to enter the field a second time after the EM readings have been taken to acquire the soil samples and to locate the selected sample sites. The latter locations are not difficult to establish when numbered markers have been left in the field at the sites of each EM measurement. This need for reentry is not a major factor when large areas are being mapped. A soil sampling team is usually sequenced 1 day after the EM measurement operation; the statistical calculations used to select the sampling sites can be made at fieldside by another team member. This method

MEASURING AND MAPPING SOIL SALINITY

239

is especially appropriate where very rapid, mobile instrumental systems are being used to map intensively extensive areas of land. 4. Comparisons of the Different Methods of Measuring Soil Salinity

Only a few direct comparisons of the various instrumental and conventional methods of measuring soil salinity have been made to date. Salinity measurements were made by Yadau et al. (1979) using four methods in a field experiment in India: (1) porous-matrix salinity sensor (EC,), (2) vacuum-cup soil water sampler (EC,), (3) soil samples (EC,), and (4) four-electrode bulk soil conductivity sensor (EC,; surface Wenner array method). These investigatorsfound a better linear correlation between EC, and EC, (r = 0.93) than between EC, and EC, (r = 0.78) or between EC, and EC, (I = 0.78). They concluded that, for purposes of diagnosing the salinities of the soils of an extensive area, the four-electrode technique is preferred because it is more rapid, simple, and practical. Loveday (1980) compared the four-electrode (surface Wenner array) technique with soil sample extracts (EC,) in a survey of 50 field sites in Australia. The water contents of the soils at the time of measurement were not controlled and were not generally at field capacity. In spite of this, he obtained relatively high correlations between EC, and EC,, though variance was high. He attributed this to field variability effects and concluded that the fourelectrode method was good for gross survey work but not accurate enough for predictive purposes. However, Loveday (1 980) used simple soil-type calibrations (i.e., as described in Section III,C,3,a) and only two 5-cm-diameter soil samples were used to estimate the salinity of the relatively large volume of soil included in the Wenner measurement. One must question that such small samples represent “ground truth,” and hence Lovedays’ (1980) conclusion. Indeed, it has been found that the so-called groundtruth of soil salinity, as typically determined using small-volume soil Samples, used to test the credibility of instrumental techniques of salinity appraisal are usually not very representative of the larger volume of soil involved in the instrumental measurements (Rhoades et al., 1989d, 1990b; Lesch et al., 1992). Loveday (1980) also concluded from his results that EC,-EC, calibrations found in the United States were probably universally applicable to soils of similar texture. van Hoorn (1980) compared salinities measured using extracts of soil samples with those obtained using both fourelectrode surface array and fourelectrode salinity probe methods in large experimental tanks. van Hoorn (1980) concluded that for survey work either the Wenner method or the four-electrode probe could be used, but that the accuracy of the latter is much greater. Nadler and Dasberg

240

J. D. RHOADES

(1980) evaluated soil salinity measurements made in small salinized field plots with in situ ceramic porous matrix sensors, four-electrode salinity probes, a four-electrode Wenner array, and soil sample extracts (1 : 1). They found good correspondence between “expected” salinity and both “soil extract” salinity and “four-electrode probe” salinity, but not with “porous matrix salinity sensor” salinity. Nadler and Dasberg ( 1980) attributed the latter discrepancy to lag-time problems. They also concluded that the Wenner array method could be used more reliably under drier soil conditions than could the four-electrode probe, which requires better electrode/soil contact for accurate measurements. I have obtained good results with the use of four-electrode and EM methods of soil salinity appraisal in numerous locations in the United States and elsewhere, and have found these techniques to be useful in varied applications, including salinity diagnosis, mapping, monitoring, saline seep and water table encroachment identification, irrigation scheduling and control, and leaching fraction assessment. Experience in recent years has been less satisfactory with regard to the suitability of porous matrix salinity sensors, because of their small sampling volume and lagtime response to changing soil salinity situations. However, for some applications they may still be the preferred technique. In fact, I recommend that the various techniques be used complementarily; the EM sensor is most suitable for surveying large landscapes to isolate areas of similarity and difference, and the four-electrode probe is more suited to acquiring more detailed information on EC, (and salinity) within various regions of the root zone, such as below the furrow, within the bed, with distance from drip emitters, etc. The fewest appropriate number of soil samples can then be taken from the different areas for detailed chemical analysis of the salinity composition, if desired, using the salinity variability information obtained with the instrumental readings. This “combined-use” approach greatly facilitates the tedious, time-consuming and costly aspects of soil sampling. Whether the soil samples are reacted with water, or whether soil water per se is isolated from the soil for detailed analysis, is a matter of need and practicality. For practical reasons, aqueous extracts are generally used, although ideally one would prefer an analysis of the actual soil water. One should always be aware of the limitations of each of the alternative methods for measuring soil salinity and take them into account. The most appropriate one@) should be used according to the specific needs and objectives of each particular situation. Again, the overall task of measurement and monitoring of soil salinity can be greatly facilitated through the combined use of the various methods. The EM and four-electrode instrumental methods can be used for most of the field characterization needs; laboratory analyses can then be carried out on only the minimum appro-

MEASURING AND MAPPING SOIL SALINITY

24 1

priate number of soil samples collected in accordance with the findings of the field instruments. The areas requiring separate sampling are most easily determined from EM findings; the depths to be sampled and numbers of samples to be taken from within each sampling area/depth are most accurately determined from four-electrode probe findings. 5. Determination of Locations of Measurement Sites

For purposes of salinity mapping, sample sites must be associated with geographic location (x and y coordinates). If rapid methods of salinity measurement are to be used to best advantage, then an equally rapid means of determining sample site location must also be used. For this purpose, the LORAN system used in marine and aviation navigation can be employed with success for some types of salinity mapping (see Rhoades et al., 1990b, 1990~).The LORAN-C system, which broadcasts pulsed radio signals at a frequency of 100 kHz, is operated by the United States Coast Guard in cooperation with several other countries as an aid to navigation. The coverage is very good regardless of terrain and is not limited to line-of-sight transmission because of the use of the lowfrequency radio waves. The low-frequency radio band is propagated by means of the ground wave, so that the radio waves closely follow the surface of the earth. The receiver location is calculated by measuring the time delays of the received signals from three different transmitter locations and applying the principle of triangulation. Thus, the LORAN-C receiver is essentially a precise time-difference measuring instrument that processes the received information to determine a position fix. The position fix is given directly in terms of latitude and longitude coordinates expressed in degrees, minutes, and seconds. With local calibration the repeatability of position determination can be as good as 10 m or better (Rhoades et al., 1990~).This “accuracy” is good enough for regional surveys, but not for more detailed mapping. The global positioning system (GPS) provides a more accurate and equally practical means to establish sample-site positions for measurement and mapping purposes. The GPS is a satellite-based radio navigation system operated by the United States Department of Defense (DOD). When fully operational (1993) it will consist of 2 1 satellites in circular orbit at a 20,000-km altitude and will provide worldwide, 24-hr coverage. The system is analogous in concept to LORAN, but utilizes line-of-sight reception of signals from multiple satellite-based transmitters of known position. A GPS receiver unit obtains/deciphers coded and synchronized signals emitted by several (usually at least three) GPS satellite transmitters in terms of time of measurement, distance from the transmitter, and position

242

J. D. RHOADES

of the receiving antenna. Distances are determined by measuring the difference in travel time of the radio signal between the various satellites and the receiver by means of accurate, synchronized clocks contained within both the transmitters and receivers. The receiver (sample) position is calculated by “triangulating” the range distance from three or more satellites of known position. The GPS includes five control stations evenly spaced around the earth near the equator. They track each satellite, determine their exact positions, and transmit correction factors to the satellites and, in turn, back to the receivers. The suitability of GPS for soil survey purposes has been shown by Long et al. (1991) to meet the accuracy requirements for detailed soil surveys (530.5 m). Under the mountainous and forested conditions of western Montana, positional accuracy was found to average between 3 and 4 m in the open, and between 5 and 6 m under closed forest canopies (Gerlach and Jasumbach, 1989). Accuracy is increased to between 2 and 5 m by averaging multiple readings taken over 10 sec and by postprocessing the data obtained by the mobile-receiver data to correct it for “drift” using analogous data collected over the same time period with a fixed-base reference receiver (base station). Accuracies of receiver position to within 2 - 5 m of true are made possible through use of this so-called “differential-mode” operation (two receivers-one mobile and one stationary) and postprocessing technique. This is the procedure that !use in my mobile salinity assessment systems to establish the coordinates of EC, measurement and soil sampling locations (Rhoades, 1992a,b). These corrections can also be done in real time by incorporation of a portable personal computer/software into the system. 6. Examples of Use of Instrumental Systems for Assessing and Mapping Soil Salinity

Both near-surface and subsurface salinity have been successfully mapped using several of the methods described above by making the instrumental measurements spatially within the area of interest and displaying the data in terms of contour, or three-dimensional, maps or as “profiles” of EC, or EC,, depending on purpose and preference. A detailed evaluation of the suitability of the various instrumental field methods for measuring EC, and of the model/field-estimates method for converting EC, to EC, and for mapping soil salinity was evaluated in a 15-square-milestudy in California (Rhoades et al., 1990b). In this project, the instrumental measurements were made manually at predetermined sites that were located by use of LORAN navigation techniques; the area was traversed on foot. Contour maps were made of measured and predicted salinities using data collected at about 1000 locations. EC, was

MEASURING AND MAPPING SOIL SALINITY

24 3

Predicted (4P, full, II)soil salinities 40 I41 64

4012516

40I0868

4009220 243744

245340

246936

w

248532

SCALE 1:998 k

i

250128

251724

I

Figure 18. Contour map of predicted soil salinities, EC, in dS/m; 400-m grid basis. (After Rhoades et al., 1990b.)

predicted from EC, as measured by both four-electrode (Wenner array and insertion probe) and EM methods. The values of EC, (both measured and predicted) were plotted using SURFER (1986) software at both 200- and 400-m grid spacings. This resulted in 1000 and 273 equally spaced grid nodes, respectively. Only contoured maps for the 400-m grid spacing are shown in Figs. 18 and 19, because the 200-m grid spacing resulted in too many contours to represent clearly in maps of such scale. The corresponding values of measured and predicted EC, at each of these nodes were determined with the SURFER software and were used to calculate the differences. Visual comparisons of the measured and predicted salinity maps showed them to be essentially the same, irrespective of the means of measuring EC,. The correct picture of salinity distribution within the study area was obtained using any of the three instrumental measurement methods. The actual salinity levels were also similar, as well (see Rhoades et al., 1990b). Sample variability due to size differences in the volumes of soil used to measure salinity was concluded to be appreciable and to result in an

J. D. RHOADES

244 4014164

P-

243744

24!

243936

251 724

248532

SCALE 1:998 b

I

I

Figure 19. Contour map of measured soil salinities, EC, in dS/m; 400-m grid basis. (After Rhoades ef a/., 1990b.)

underestimate of the accuracy of the instrumental/model appraisal method under typical field conditions (Rhoades et al., 1990b). Where the differences were large, the salinity levels were so high as to make them agriculturally unimportant. The accuracies of the calculated values are deemed sufficiently adequate for practical salinity mapping purposes. These findings demonstrate that soil salinity appraisal and mapping can be appropriately accomplished without need for soil samples and laboratory analyses using field measurements of soil electrical conductivity (by any of the three methods tested -four-electrode probe, four-electrode surface array, and electromagnetic induction) and estimates of soil water content, bulk density, and surface conductance made in the field by “feel” methods. More recently, soil salinity has been characterized in even more detail than that described above in field evaluations of the mobilized versions of the four-electrode and EM techniques and of the newer approach (Section III,C,3,c) used with them for cohverting EC, (or EM) readings to EC,. The areas were traversed with the mobile systems, collecting readings at a periodicity that provided a gridlike pattern of required/desired intensity (between 10 and 25 m). These measurement sites were marked and their

245

MEASURING A N D MAPPING SOIL SALINITY

exact positions were established using the GPS technology described above. A small number of these sites was subsequently chosen for soil sampling (ground truthing) based on the observed EC, field pattern. A second trip was made into the fields and a relatively small number of soil samples was collected using rapid, tractor-mounted augering/coring equipment. The salinities of these soil samples were then determined using the rapid method of Rhoades et al. (1989b). The salinities at the remaining nonsampled sites were predicted from the corresponding EM readings through use of the multiple linear regression relation [Eq. (32)] established for each field. The salinity contour maps obtained using this method were nearly identical to those obtained by conventional soil sampling. An example is shown in Fig. 20, after Lesch et al. (1992). This approach provides a very practical and cost-effective means to reduce substantially the number of soil samples needed to map accurately in detail the spatial salinity patterns that occur at the field scale. The suitability of such detailed spatial data to assess the adequacy/appropriatenessof irrigation/drainagesystems is demonstrated by Rhoades (1992a,b).

Field S2A

a

Predicted In(ECe) based on 36 samples

b Observed In(ECe) based on 206 samples

300

388

304

304

ne 152

152

76

76

0

0 0

70

140

210

200

358

0

70

140

216

288

Figure 20. Comparison of contour plots of predicted (a) and measured (b) values of In EC, (0-30 cm soil depth). ( A h Lesch et al., 1992.)

358

246

J. D. RHOADES

IV. CONCLUSIONS AND SUMMARY Measurement of the electrical conductivity of saturated soil paste extracts can be accurately predicted from measurements of the electrical conductivity and saturation percentage of the soil paste, with substantial savings in time. The measurement of bulk soil electrical conductivity (EC,), made using both four-electrode and electromagnetic induction techniques, can be used to even greater advantage for salinity appraisal. Soil salinity can be determined from EC, directly in the field without requiring laboratory analysis. Alternatively, limited ground-truth techniques involving limited laboratory analysis can be used to predict salinity from the more easily and quickly made measurements of EC,. Both of these salinity assessment techniques have been shown to be reliable and practical. Commercial instrumentation is available for measuring EC, using both four-electrode and electromagnetic induction methodology. Such instrumentation has now been mobilized and integrated with GPS technology, permitting both EC, (and hence salinity) and measurement site location to be determined on the go. These equipment and methodologies have been shown to be useful for purposes of measuring, monitoring, and mapping field salinity, for detecting the presence of shallow water tables, and for assessing the adequacies of leaching and drainage practices for soil salinity control.

REFERENCES Adams, F. 1974. Soil solution. In “The Plant Root and Its Environment” (E. W. Carson, ed.), pp. 44 1 -48 1. University Press of Virginia, Charlottesville, Virginia. Adams, F., Burmester, C., Hue, N. V., and Long, F. L. 1980. A comparison of columndisplacement and centrifuge methods for obtaining soil solutions. Soil Sci. SOC.Am. J. 44,733-735.

Allison, L. E. 1973. Oversaturation method for preparing saturation extracts for salinity appraisal. Soil Sci. 116,65-69. Austin, R. S., and Oster, J. D. 1973. An oscillator circuit for automated salinity sensor measurements. Soil Sci.Soc. Am. Proc. 31, 327 - 329. Austin, R. S., and Rhoades, J. D. 1979. A compact, low-cost circuit for reading four-electrode salinity sensors. Soil Sci. Soc. Am. J. 43,808-810. Beatty, H. J., and Loveday, J. 1974. Soluble cations and anions. Tech. Commun.Commonw. Bur. Soil Sci. 54, 108- 117. Biggar, J. W., and Nielsen, D. R. 1976. Spatial variability of the leaching characteristics of a field soil. Water Resour. Res. 12,78 - 84. Bohn, H. L., Ben-Asher, J., Tabbara, H. S., and Marwan, M. 1982. Theories and tests of electrical conductivity in soils. Soil Sci. Soc. Am. J. 46, 1143- 1146.

MEASURING AND MAPPING SOIL SALINITY

247

Bottraud, J.-C., and Rhoades, J. D. 1985. Referencing water content effects on soil electrical conductivity-Salinity calibrations. Soil Sci. SOC.Am. J. 49, 1579- 1581. Briggs, L. J., and McCall, A. G. 1904. An artificial root for inducing capillary movement of soil moisture. Science 20,566- 569. Chow, T. L. 1977. A porous cup soil-water sampler with volume control. Soil Sci. 124, 173- 176.

Corwin, D. L., and Rhoades, J. D. 1982. An improved technique for determining soil electrical conductivity- Depth relations from above ground electromagnetic measurements. Soil Sci. SOC. Am. J. 46,517-520. Corwin, D. L., and Rhoades, J. D. 1984. Measurement of inverted electrical conductivity profiles using electromagneticinduction. Soil Sci. SOC. Am. J. 48,288 - 29 1. Corwin, D. L., and Rhoades, J. D. 1990. Establishing soil electrical conductivity-Depth relations from electromagneticinduction measurements. Cornmun. Soil Sci. Plant Anal. 21,861-901.

Dalton, F. N., and Poss, J. A. 1990. Soil water content and salinity assessment for irrigation scheduling using timedomain reflectometry: Principles and applications. Actu Hortic. 278,381-393.

Dalton, F. N., and van Genuchten, M. T. 1986. The timedomain reflectometry method for measuring soil water content and salinity. Geodenna 38,237-250. Dalton, F. N., Herkelrath, W. N., Rawlins, D. S., and Rhoades, J. D. 1984. Time domain re5ectometry: Simultaneous measurement of soil water content and electrical conductivity with a single probe. Science 224,989-990. Dalton, F. N., Poss,J. A., Heimovaara, T. J., Austin, R. S., Alves, W. J., and Russell, W. B. 1990. Principles, techniques and design considerations for measuring soil water content and salinity using timedomain reflectometry. U.S.Salinity Lab. Res. Rep. 122. Dao, T. H., and Lay, T. L. 1978. Extraction of soil solution using a simple centrifugation Am. J. 42, 375method for pesticide adsorption-Desorption studies. Soil Sci. SOC. 377.

Dasberg, S., and Dalton, F. N. 1985. Field measurement of soil water content and bulk electrical conductivity with timedomain reflectometry. Soil Sci. SOC.Am. J. 49, 293297.

Davies, B. E., and Davies, R. I. 1963. A simple centrifugation method for obtaining small samples of soil solution. Nature (London) 198,216-217. Duke, H. R., and Haise, H. R. 1973. Vacuum extractors to assess deep percolation losses and chemical constituents of soil water. Soil Sci. SOC. Am. Proc. 37,963-964. Elkhatib, E. A., Berknett, 0. L., Baligar, V. C., and Wright, R. J. 1986. A centrifuge method for obtaining soil solution using an immiscible liquid. Soil Sci. SOC.Am. J. 50, 297299.

Elkhatib, E. A., Hern, J. L., and Staley, T. E. 1987. A rapid centrifugation method for obtaining soil solution. Soil Sci. SOC.Am. J. 51,578 - 583. Enfield, C. G., and Evans, D. D. 1969. Conductivity instrumentation for in situ measurements of soil salinity. Soil Sci. SOC.Am. Proc. 33,787-789. Gerlach, F. L., and Jasumbach, A. E. 1989. “Global Positioning Systems Canopy Effect Study.” Technology Development Center, Forest Service, US. Department of Agriculture, Missolula, Montana. Gillman, G. P. 1976. A centrifuge method for obtaining soil solution. Div. Soils Div. Rep. (Aust., C.S.I.R.O.)16. Grossman, J., and Udluft, P. 1991. The extraction of soil water by the suction-cup method: A review. J. Soil Sci. 42, 83-93. Halvorson, A. D., and Rhoades, J. D. 1974. Assessing soil salinity and identifying potential

248

J. D. RHOADES

saline-seep areas with field soil resistance measurements. Soil Sci. SOC.Am. Proc. 38, 576-581. Halvorson, A. D., and Rhoades, J. D. 1976. Field mapping soil conductivity to delineate dryland saline seeps with fourelectrode technique. Soil Sci. Soc. Am. J. 40, 57 1 - 575. Halvorson, A. D., Rhoades, J. D., and Ruele, C. A. 1977. Soil salinity-four-electrode conductivity relationships for soils of the Northern Great Plains. Soil Sci. SOC.Am. J. 41, 966-971. Harris, A. R., and Hansen, E. A. 1975. A new ceramic cup soil-water sampler. Soil Sci. SOC. Am. Proc. 39, 157-158. Ingvalson, R. D., Oster, J. D., Rawlins, S. L., and Hoffman, G. J. 1970. Measurement of water potential and osmotic potential in soil with a combined thermocouple psychrometer and salinity sensor. Soil Sci. Soc. Am. Proc. 34,570-574. Jackson, D. R., Brinkley, F. S., and Bendetti, E. A. 1976. Extraction of soil water using cellulose-acetatehollow fibers. Soil Sci. Soc. Am. J. 40,327 - 329. Jordan, C. F. 1968. A simple, tension-free lysimeter. Soil Sci. 105, 81-86. Kemper, W. D. 1959. Estimation of osmotic stress in soil water from the electrical resistance of finely porous ceramic units. Soil Sci. 87,345 - 349. Kinniburgh, D. G., and Miles, D. L. 1983. Extraction and chemical analysis of interstitial water from soils and rocks. Environ. Sci. Technol. 17, 362-368. Kittrick, J. A. 1983. Accuracy of several immiscible displacement liquids. Soil Sci. Soc. Am. J. 47, 1045- 1047. Kohnke, H., Dreibelbis, F. R., and Davidson, J. M. 1940. A survey and discussion of lysimeters and a bibliography on their construction and performance. M i x . Pub1.US.,Dep. Agric. 372. Lesch, S. M., Rhoades, J. D., Lund, L. J., and Corwin, D. L. 1992. Mapping soil salinity using calibrated electromagneticmeasurements. Soil Sci. SOC.Am. J. 56, 540-548. Levin, M. J., and Jackson, D. R. 1977. A comparison of in situ extractors for sampling soil water. SoilSci. SOC.Am. J. 41,535-536. Litaor, M. I. 1988. Review of soil solution samplers. Water Resour. Res. 24,727-733. Long, D. S., DeGloria, S. D., and Galbraith, J. M. 1991. Use of the global positioning system in soil survey. J. Soil Water Conserv. 46,293-297. Longenecker, D. E., and Lyerly, P. J. 1964. Making soil pastes for salinity analysis: A reproducible capillary procedure. Soil Sci. 97,268-275. Loveday, J. 1972. Moisture content of soils for making saturation extracts and effect of grinding. Div. Soils Tech. Pap. (Aust., C.S.I.R.O.)12a. Loveday, J. 1980. Experiences with the 4-electrode resistivity technique for measuring soil salinity. Div. Soils Div. Rep. (Aust., C.S.I.R.O.)51. Maas, E. V. 1986. Salt tolerance ofplants. Appl. Agric. Rex 1, 12-26. Maas, E. V. 1990. Crop salt tolerance. ASCE Manuals Rep. Eng. 71,262- 304. Maas, E. V., and Hoffman, G. J. 1977. Crop salt tolerance-Current assessment. J. Irrig. Drain. Div., Am. SOC.Civ. Eng. 103(IR2), 115 - 134. McNeil, J. D. 1980. Electromagnetic terrain conductivity measurement at low induction numbers. Tech. Note-Geonics Ltd. T N 4 . Menzies, N. W., and Bell, L. C. 1988. Evaluation of the influence of sample preparation and extraction technique on soil solution composition. Aust. J. Soil Res. 26,45 1-464. Mubarak, A., and Olsen, R. A. 1976. Immiscible displacement of the soil solution by centrifugation. Soil Sci. SOC. Am. J. 40,329-331. Mubarak, A., and Olsen, R. A. 1977. A laboratory technique for appraising in situ salinity of soil. SoilSci. SOC.Am. J. 41, 1018-1020. Nadler, A. 1981. Field application of the fourelectrode technique for determining soil solution conductivity. Soil Sci. SOC.Am. J. 45,30-34.

MEASURING AND MAPPING SOIL SALINITY

249

Nadler, A. 1982. Estimating the soil water dependence of the electrical conductivity soil solution/electrical conductivity bulk soil ratio. SoilSci. SOC.Am. J. 46,722-726. Nadler, A., and Dasberg, S. 1980. A comparison of different methods for measuring soil salinity. Soil Sci. Soc. Am. J. 44,725 - 728. Nadler, A., Magaritz, M., Lapid, Y., and Levy, Y. 1982. A simple system for repeated soil resistance measurements at the same spot. Soil Sci. Soc.Am. J. 46,661 -663. Nielsen, D. R., Biggar, J. W., and Fah, K. T. 1973. Spatial variability of field-measured soil-water properties. Hilgardia 42,2 15- 259. Oster, J. D., and Ingvalson, R. D. 1967. In situ measurement of soil salinity with a sensor. Soil Sci. SOC.Am. Proc. 31,572-574. Oster, J. D., and Willardson, L. S. 1971. Reliability of salinity sensors for the management of soil salinity. Agron. J. 63,695 -698. Oster, J. D., Willardson, L. S., and Hoffman, G. J. 1973. Sprinkling and ponding techniques for reclaiming saline soils. Trans. Am. SOC.Civ. Eng. 16, 89-91. Oster, J. D., Willardson, L. S., van Schilfgaarde, J., and Goertzen, J. 0. 1976. Imgation control using tensiometers and salinity sensors. Trans. ASAE 19,294-298. Parizek, R. R., and Lane, B. E. 1970. Soil-water sampling using pan and deep pressurevacuum lysimeters. J. Hydrol. 11, 1 - 2 1. Postel, S. 1989. “Water for Agriculture: Facing the Limits,” Worldwatch Pap. 93. Worldwatch Institute, Washington, D.C. Reeve, R. C., and Doering, E. J. 1965. Sampling the soil solution for salinity appraisal. Soil Sci. 99,339-344. Reitemeier, R. F. 1946. Effect of moisture content on the dissolved and exchangeable ions of soils of arid regions. Soil Sci. 61, 195-214. Rhoades, J. D. 1972. Quality of water for irrigation. Soil Sci. 113,277-284. Rhoades, J. D. 1976. Measuring, mapping and monitoring field salinity and water table depths with soil resistance measurements. Soils Bull. 31, 159- 186. Rhoades, J. D. 1978. Monitoring soil salinity: A review of methods. Establishment of water quality monitoring programs. Proc. Annu. Am. Water Resour. Cont2, 150- 165. Rhoades, J. D. I979a. Salinity management and monitoring. Rep. -Univ. Calg. Water Resour. Cent. 45,73 - 87. Rhoades, J. D. 1979b. Inexpensive four-electrodeprobe for monitoring soil salinity. Soil Sci. SOC.Am. J. 43,8 17- 8 18. Rhoades, J. D. 1980. Determining leaching fraction from field measurements of soil electrical conductivity. Agric. Water Manage. 3,205 - 2 15. Rhoades, J. D. 1981. Predicting bulk soil electrical conductivity vs saturation paste extract electrical conductivity calibrations from soil properties. Soil Sci. SOC.Am. J. 45, 43-44. Rhoades, J. D. 1982. Soluble salts. Agron. Monogr. 9, 167- 178. Rhoades, J. D. 1984. Principles and methods of monitoring soil salinity. In “Soil Salinity and Imgation-Processes and Management,” Vol. 5, pp. 130- 142. Springer-Verlag, Berlin. Rhoades, J. D. 1990a. Sensing soil salinity problems: New technology. Proc. Natl. Zrrig. Symp. Irrig. Assoc. ASCE, 3rd pp. 422-428. Rhoades, J. D. 1990b. Determining soil salinity from measurements of electrical conductivity. Commun. Soil Sci. Plant Anal. 21, 861-901. Rhoades, J. D. 1992a. Instrumental field methods of salinity appraisal. SSSA Spec. Publ. 30 (in press). Rhoades, J. D. 1992b. Recent advances in the methodology for measuring and mapping soil salinity. Proc. Int. Symp. Strategies Utilizing Salt Afected Lands, ISSS Meet. Rhoades, J. D. 1994. Salinity: Electrical conductivity and total dissolved salts. Agron. Monogr. 9 (submitted).

250

J. D. RHOADES

Rhoades, J. D., and Convin, D. L. 1981. Determining soil electrical conductivity-depth relation using an inductive electromagneticsoil conductivity meter. Soil Sci. SOC. Am. J. 45,255-260. Rhoades, J. D., and Corwin, D. L. 1984. Monitoring soil salinity. J. Soil Water Conserv. 39, 172-175. Rhoades, J. D., and Corwin, D. L. 1990. Soil electrical conductivity: Effects of soil properties and application to soil salinity appraisal. Commun. Soil Sci. Plant Anal. 21,837-860. Rhoades, J. D., and Ingvalson, R. D. 1971. Determining salinity in field soils with soil resistance measurements. Soil Sci. Soc. Am. Proc. 3554-60. Rhoades, J. D., and Miyamoto, S. 1990. Testing soils for salinity and sodicity. In “Soil Testing and Plant Analysis,” 3rd Ed., SSSA Book Ser. 3, pp. 299-336. Soil Science Society of America, Madison, Wisconsin. Rhoades, J. D., and Oster, J. D. 1986. Solute content. Agron. Monogr. 9,985- 1006. Rhoades, J. D., and van Schilfgaarde,J. 1976. An electrical conductivity probe for determining soil salinity. SoilSci. SOC.Am. J. 40,647-651. Rhoades, J. D., Raats, P. A. C., and Prather, R. J. 1976. Effects of liquid-phase electrical conductivity, water content, and surface conductivity on bulk soil electrical conductivity. Soil Sci. SOC.Am. J. 40,65 1 -655. Rhoades, J. D., Kaddah, M. T., Halvorson, A. D., and Prather, R. J. 1977. Establishing soil electrical conductivity- salinity calibrations using four-electrode cells containing undisturbed soil cores. Soil Sci. 123, 137 - 14I. Rhoades, J. D., Corwin, D. L., and Hoffman, G. J. 1981. Scheduling and controlling imgations from measurements of soil electrical conductivity. Proc. ASAE Irrig. Schedul. ConJ: pp. 106-115. Rhoades, J. D., Manteghi, N. A., Shouse, P. J., and Alves, W. J. 1989a. Soil electrical conductivity and soil salinity: New formulations and calibrations. Soil Sci. SOC.Am. J. 53,433-439. Rhoades, J. D., Manteghi, N. A., Shouse, P. J., and Alves, W. J. 1989b. Estimating soil salinity from saturated soil-paste electrical conductivity. Soil Sci. Soc.Am. J. 53, 428433. Rhoades, J. D., Waggoner, B. L., Shouse, P. J., and Alves, W. J. 1989c. Determining soil salinity from soil and soil-paste electrical conductivities: Sensitivity analysis of models. Soil Sci. SOC.Am. J. 53, 1368- 1374. Rhoades, J. D., Lesch, S. M., Shouse, P. J., and Alves, W. J. 1989d. New calibrations for determining soil electrical conductivity-depth relations from electromagnetic measurements. Soil Sci. SOC.Am. J. 53, 74-79. Rhoades, J. D., Corwin, D. L., and Lesch, S. M. 1990a. Effect of soil EC,-depth profile pattern on electromagnetic induction measurements. US.Salinity Lab. Rep. 125, I 108. Rhoades, J. D., Shouse, P. J., Alves, W. J., Manteghl, N. A., and Lesch, S. M. 1990b. Determining soil salinity from soil electrical conductivity using different models and estimates. Soil Sci. Soc. Am. J. 54,46-54. Rhoades, J. D., Lesch, S. M., Shouse, P. J., and Alves, W. J. 1990~.Locating sampling sites for salinity mapping. Soil Sci. SOC.Am. J. 54, 1799- 1803. Richards, L. A. 1941. A pressure-membrane extraction apparatus for soil solution. Soil Sci. SOC.51,377-386. Richards, L. A. 1966. A soil salinity sensor of improved design. Soil Sci. SOC.Am. Proc. 30, 333-337. Ross, D. S., and Bartlett, R. J. 1990. Effects of extraction methods and sample storage on properties of solutions obtained from forested Spodosols.J. Environ. Qual. 19, 108- I 13. Shainberg, I., Rhoades, J. D., and Prather, R. J. 1980. Effect of low electrolyte concentration

MEASURING AND MAPPING SOIL SALINITY

2s1

on clay dispersion and hydraulic conductivity of a sodic soil. Soil Sci. SOC.Am. J. 45, 273-277. Shimshi, D. 1966. Use of ceramic points for the sampling of soil solution. Soil Sci. 101, 98- 103. Sonnevelt, C., and van den Ende, J. 1971, Soil analysis by means of a 1 :2 volume extract. Plant Soil 35, 505 - 5 16. Suarez, D. L. 1986. A soil water extractor that minimizes COz degassing and pH errors. Water Resour. Rex 22,876-880. Suarez, D. L. 1987. Prediction of pH errors in soil water extractors due to degassing, Soil Sci. SOC.Am. J. 51,64-67. SURFER. 1986. Reference manual. Golden Software, Golden, Colorado. Tadros, V. T., and McGarity, J. W. 1976. A method for collecting soil percolate and soil solution in the field. Plant Soil44,655-667. Topp, G. C., Davis, J. L., and Annan, A. P. 1980. Electromagnetic determination of soil water content: Measurement in coaxial transmission lines. Water Resour. Res. 16, 574- 582. Topp, G. C., Davis, J. L., and Annan, A. P. 1982. Electromagnetic determination of soil water content using T D R I. Applications to wetting fronts and steep gradients. Soil Sci. SOC.Am. J. 46,672-678. Topp, G. C., Davis, J. L., Bailey, W. G., and Zebchuk, W. D. 1984. The measurement of soil water content using a portable TDR hand probe. Can. J. Soil Sci. 64,3 13 - 32 1. Topp, G. C., Yanuka, M., Zebchuk, W. D., and Zegelin, S. 1988. Determination of electrical conductivity using time domain reflectometry: Soil and water experiments in coaxial lines. Water Resour. Res. 24,945 -952. U.S. Salinity Laboratory S M . 1954. Diagnosis and improvement of saline and alkali soils. U S . Dep. Agric. Handb. 60. van De Pol, R. M., Wierenga, P. J., and Nielsen, D. R. 1977. Solute movement in a field. Soil Sci. Soc. Am. J. 41, 10- 13. van Hoorn, J. W. 1980. The calibration of four-electrode soil conductivity measurements for determining soil salinity. Proc. Int. Symp. Salt Affected Soils pp. 148- 156. Wagner, G. H. 1965. Changes in nitrate N in field plot profiles as measured by the porous cup technique. Soil Sci. 100,397-402. Webster, R. 1985. Quantitative spatial analysis of soil in the field. Adv. Soil Sci. 31,505-524. Webster, R. 1989. Recent achievements in geostatistical analysis of soil. Agrokem. Talajtan 38,519-536. Wesseling, J., and Oster, J. D. 1973. Response of salinity sensors to rapidly changing salinity. Soil Sci. Soc. Am. Proc. 37, 553-557. Wilcox, L. V. 195 I . A method for calculating the saturation percentage from the weight of a known volume of saturated soil paste. Soil Sci. 72,233-237. Wok, J., and Graveel, J. G. 1986. A rapid method for obtaining soil solution using vacuum displacement. SoilSci. Soc. Am. J. 50,602-605. Wood, J. D. 1978. Calibration stability and response time for salinity sensors. Soil Sci. SOC. Am. J.42,248-250. Wood, W. W. 1973. A technique using porous cups for water sampling at any depth in the unsaturated zone. Water Resour. Res. 9,486-488. Yadau, B. R., Rao, N. H., Paliwal, K. V., and Sarma, P. B. S. 1979. Comparison of different methods for measuring soil salinity under field conditions. Soil Sci. 127, 335-339. Yamasaki, S., and Kishita, A. 1972. Studies on soil solution with reference to nutrient availability. I. Effect of various potassium fertilizer on its behavior in the soil solution. Soil. Sci. Plant Nutr. 18, 1-6.

This Page Intentionally Left Blank

BREEDING,PHYSIOLOGY, CULTURE, AND UTILIZATION OF CICER MILKVETCH (Astragahs cicer L.) C. E. Townsend Crops Research Laboratory, Agricultural Research Service, United States Department of Agriculture, Fort Collins, Colorado 80526

I. Introduction 11. Morphology and Anatomy A. Root B. Stem and Leaf C. Flower D. Seed Coat Anatomy 111. Physiology A. Seed Germination B. Seedling Growth C. Flowering D. Vernalization E. Photoperiod IV. Culture A. Adaptation B. Soils and Soil Fertility C. Seed Scarification and Inoculation D. Establishment E. Pest Resistance F. Pollination Requirements G. Seed Production H. Weed Control V. Utilization A. Hay B. Pasture VI. Breeding, Genetics, and Cytology A. Objectives B. Cytology and Inbreeding Depression C. Breeding Methodology D. Seedling Vigor A k a in A v y , Val. 49 Copyright 0 1993 by Academic Press, Inc. AU rights of reproducuon in my form reserved.

253

254

C . E. TOWNSEND E. Forage Yield F. Seed Weight G. Date of Flowering H. Cultivars VII. Summary and Conclusions References

I. INTRODUCTION The genus Astrugulus is one of the largest genera of flowering plants in the world; it comprises about 2500 species, of which 500 are in the New World (Podlech, 1986). Species of the more advanced sections are adapted to xeric conditions, whereas those with the more primitive traits are generally mesophytic species found widespread in mountains and relatively humid regions. The astragali are found in three widely separated parts of the world. The largest number of species occurs in the Eurasian and North African regions; smaller numbers occur in western regions of North and South America (Ledingham and Rever, 1963). Cicer milkvetch (Astrugulus cicer L.), a long-lived rhizomatous perennial legume, was first introduced into the United States in 1926 and was widely used in experimental plantings in the Great Plains and western United States from 1929 to 1935 (P. R. Henson, personal communication, 1968). It is native to and widely dispersed in central and eastern Europe, including European Russia and the Caucasus (Komarov, 1965). Cicer milkvetch is slowly finding its place as a pasture, hay, and conservation species under irrigated and dryland conditions primarily in the central and northern Great Plains and adjacent Rocky Mountains, and in similar areas of western Canada.

11. MORPHOLOGY AND ANATOMY A. ROOT Cicer milkvetch has a branched taproot. Its roots do not reach the depths that alfalfa (Medicago sutivu L.) roots do, but we have traced them to depths in excess of 2 m under favorable soil moisture conditions (unpublished observations, 1976). The upper portions of the primary root may reach a diameter of 2.5 cm or more. Cicer milkvetch spreads by rhizomes

CICER MILKVETCH (Artragah cicer L.)

255

under favorable moisture conditions, but it is not as aggressive as crownvetch (Coronilluvaria L.). Rhizomes begin to appear just below the crown buds when seedlings are about 6 weeks old. A few small roots may emerge from the nodes on rhizomes, but independent plants do not develop from rhizomes. The sectioning of rhizomes, however, is an excellent method for vegetative propagation of selected plants. Considerable variability exists among plants for spread, i.e., at the end of the second growing season under spaced-plant conditions, spread ranges from about 0.2 m to more than 1 m.

B. STEMAND LEAF Cicer milkvetch has a decumbent growth habit. Its hollow, succulent stems can reach a length in excess of 1 m. Leaves are pinnately compound. The mean number of leaflets per leaf is about 25, with a range of 15 to 35 (Townsend, 1970). The leaflets are paired except for the single one at the tip of the leaf. Leaflets are lanceolate to lance-oblong in shape and have rounded tips. Average length of the leaf is about 12 cm, with a range of 7 to 21 cm. Stipules are about 8 mm long, connate at the base, and oblong to triangular-oblong in shape. A comparative electron microscopic investigation of the ultrastructure of chlorenchyma cells of plants adapted to the far north revealed that cicer milkvetch and other winter-hardy species had similarities, i.e., “their chloroplasts often had deep invaginations and swellings, the plastids were situated more compactly, the number of mitochondria was greater, the endoplasmic reticulum was more developed and many lipid drops were found in the cells” (Miroslavov and Bubolo, 1980).

C . FLOWER The inflorescence of cicer milkvetch is a compact raceme with 15 to 60 florets. Each floret consists of a calyx with five united sepals, a corolla, 10 stamens, and a pistil. The corolla consists of five petals: a large standard (banner), two lateral wing petals, and two fused petals that form the keel. There is little variation in flower color, which ranges from white to pale yellow or cream. The number of racemes per plant ranges from 3 1 to 157, with a mean of 68 (Townsend, 1971a). The reproductive part of the flower consists of 10 stamens and a simple pistil. The stamens form a diadelphous tube in which nine are fused and the one free stamen is attached to the base of the floret nearest the standard

2 56

C. E. TOWNSEND

petal. The ovary is derived from a single carpel fused along the ventral suture. The number of ovules per ovary ranges from 9 to 15, with a mean of 12 (Townsend, 1971b).

D. SEEDCOATANATOMY The seed coat anatomy of cicer milkvetch is similar to that for other small-seeded forage legumes (Miklas et al., 1987). Scanning electron microscopy shows the following: the cuticle has a distinct faveolate pattern at the strophiole; the macrosclereid cells become longer and thinner as they neared the strophiole in comparison to the rest of the seed coat; the osteosclereids are shorter and wider at the seed tip in comparison to other sites on the seed coat; and the osteosclereidsare not present at the hilum or strophiole. Scanning electron microscopy has not revealed any cellular or morphological differences between hard and nonhard seed coat structures for Oxytropis riparia Litv., Oxytropis campestris L., alfalfa, or cicer milkvetch (Solum and Lockerman, 1991).

111. PHYSIOLOGY

A. SEEDGERMINATION Those species that germinate readily over a relatively wide range of temperatures tend to be easier to establish than those with highly specific temperature requirements. Rate of germination for cicer milkvetch was much slower than that for alfalfa, especially at cooler temperatures (Townsend and McGinnies, 1972b), and cicer milkvetch was more difficult to establish than alfalfa under field conditions (Townsend and McGinnies, 1972a; McGinnies and Crofts, 1986). Total germination for cicer milkvetch increased when alternating seed temperatures ( 12 hr at each temperature) were increased from 5/2OoCto 15/25"C,but germination decreased at 20/35 "C. The optimum germination temperatures were 15/25"C and 25°C when seeds of the cultivar Lutana were germinated using 15, 20, 25, 30, and 15/25"Ctemperature treatments (Carleton et al., 1971). Therefore, an alternating temperature of 15/25"C with 12 hr at each temperature under dark conditions is recommended for rapid and complete germination (Carleton et al., 197l ; Townsend and McGinnies, 1972b). In laboratory studies, there was considerablevariability among polycross

CICER MILKVETCH (Astragulus cicer L.)

257

progenies for seed germination at 5/20"C and 15/25"C alternating temperature (12 hr at each temperature) treatments (Townsend, 1974a) (Fig. 1). Germination at 15/25"C and 5/20"C began on the third and fifth day following planting, respectively. At 15/25"C, all progenies exceeded 75% in total germination, but they differed greatly in rate of germination. Not only were there large differences among progenies for total germination at 5/20"C, but the differences in rate of germination were very striking. Progenies with rapid germination at 5/20"C also germinated rapidly at 15/25"C, but some progenies with rapid germination at 15/25"C germinated slowly at 5/20"C. Therefore, it would be more desirable to select for rapid germination at 5/2OoCthan at 15/25"C. The hard seed coat of cicer milkvetch permits the seed to remain viable for 12 years or more when stored at ambient temperature and low relative humidity (Hafenrichter et al., 1965;Townsend, 1990b). After scarification, the seed loses viability rapidly, and thus should not be scarified until just before planting (Carleton et al., 1971). Seeds of polycross progenies representing four populations were stored in an unheated and uninsulated building at Fort Collins, Colorado, and then scarified and germinated after 1 and 12 years of storage (Townsend, 1990b). Percentage germination on a daily basis through day 14 after planting and total germination at 15/25"C were similar for the four populations and for the I- and 12-year-old seeds. Progenies within populations differed significantly* for rate of germination except on day 14 after planting with 1-year-old seed. Germinability on day 14 ranged from 87 to 100%after 12 years of storage. When evaluated at 5/20"C, large differences were noted among populations and among progenies within populations for both rate of germination and total germination. This indicates that the populations and progenies within populations differed genetically for rate of germination. Thus, genetic differences among populations were expressed at 5/20°C, but not at 15/25"C. Location and year of seed production within a location significantly influenced subsequent seed germination of polycross progenies of cicer milkvetch (Townsend, 1977b). For each year, rate of germination was highest at 15/25"C, intermediate at 5/20"C, and lowest at 5/15"C. In one year, however, total germination (day 14 after planting) at 5/2OoC was equal to that at 15/25"C. Within each year of production there were significant differences among progenies for both rate of germination and total germination at each temperature treatment. One progeny (F-1) was consistently slow to germinate in all years at 15/25"C, but its total germination was high in 3 of the 4 years evaluated. At 5/20"C, the same progeny

* Level of significance was p

0.05 in all studies reported to have significant differences.

258

C . E. TOWNSEND

90-

2191 DAY 20-

80 -

GRAND MEAN 70 -

10-

6

0

'

n

I 4 th DAY 40

n

1

n

'

20-

n

14th DAY CONTROL

CONTROL

-

30 -

I c5

Figure 1. Frequency distribution of 89 polycross progenies of cicer milkvetch for percentage seed germination on four different days during treatment at 15/25"C (left) and 5/2OoC(right). Germination of progenies included in the 10%germination class ranged from 0 to 1096, etc. The control entry was a large-seeded lot (4.53 g/lOOO seeds) that gave excellent seedling emergence under both greenhouse and field conditions. (From Townsend, 1974a.)

CICER MILKVETCH ( A m q u l u s cicer L.)

259

(F-1) had an exceptionally slow rate of germination and low total germination except for one year when its total germination was equal to that of the other progenies. At 5/ 15"C, F- 1 also had a slow rate of germination and low total germination. Environmental factors during seed maturation that affect subsequent seed germination are not known, but could include temperature, soil moisture, and soil fertility. Those factors affecting seed germination could also affect the genetic composition of a synthetic cultivar. If the germination of seed from a particular parental clone was reduced, fewer seedlings from that clone would become established in the field. Abernethy (1987) evaluated several seedlots of the cultivars Monarch and Lutana to osmoconditioning at four levels of poly(ethy1ene glycol) (PEG) (0, 150, 200, and 250 g of PEG kg-* water) for 8 days at 15°C. These treatments were followed by drying at 20°C for 1.5, 24, 48, and 168 hr. Although percentage germination at 15"C following the osmoconditioning treatments was variable, osmoconditioning increased the rate of germination at all levels of PEG. Percentage germination and rate of germination were generally highest following the 1.5-hr postosmoconditioning drying period, but the effects were minimal with up to 48 hr of drying. The response of field emergence to osmoconditioningwas variable for one study, but in a second study osmoconditioned seeds produced significantly greater seedling emergence than the control. Because of variable results, osmoconditioning did not appear to be a reliable practice with cicer milkvetch seed.

B. SEEDLING GROWTH Classical growth analysis was used to compare the early seedling growth of cicer milkvetch, alfalfa, and sainfoin (Onobrychis viciaefolia Scop.) over a 10-week growth period in a greenhouse environment (Smoliak et al., 1972). Cicer milkvetch did not differ significantly from alfalfa and sainfoin for relative growth rate (RGR) and net assimilation rate (NAR), but alfalfa and sainfoin produced significantly more top growth. Alfalfa and cicer milkvetch were similar for root weight and leaf area through week 9, but by week 10, those traits were greater for alfalfa than for cicer milkvetch. Root weights for sainfoin were greater than those for alfalfa and cicer milkvetch through week 7, but were similar thereafter. Leaf areas for sainfoin were much greater than those for cicer milkvetch and alfalfa throughout the study. Leaf area ratios (LARS), however, were variable among the three species, with cicer milkvetch having the highest mean LAR. To obtain a better understanding of early seedling vigor for cicer milk-

260

C . E. TOWNSEND

vetch, growth parameters (including RGR, LAR, and NAR) were evaluated over nine growth periods within a 42-day growth period at two temperature treatments in environmental chambers (Townsend and Wilson, 1978). The two temperature treatments, 25 day/20"C night and 20 day/ 15 "C night, differed significantly for RGR, LAR, and NAR. All three growth measurements tended to be greatest at 25/20"C. Within both temperature treatments, RGR decreased with increased length of growth period until by day 42 there was essentially no difference between temperatures for RGR. Leaf area ratios also decreased with time at both temperatures. Net assimilation rate decreased with increased length of growth period at 25/20"C, but at 20/15"C the NAR was not significantly influenced by growth period. Simple correlation coefficients between shoot weight, root weight, total weight, shoot height, leaf area, and number of leaves were significant and ranged from 0.9 1 to 0.99 for each temperature treatment. The high correlations between all traits indicated that, in general, any one of them would be a reliable indicator of seedling vigor. The influence of seed weight on seedling growth parameters of cicer milkvetch was studied in environmental chambers over a 28-day growth period (Townsend and Wilson, 1981). Seed weight was positively and linearly associated with initial seedling weight, initial leaf area, initial leaf weight, and final seedling weight. Generally, improved seedling performance was associated with increased seed weight. The lack of association of seed weight with some important growth analysis traits, such as RGR, however, suggests that certain seedling vigor traits are independent of seed weight. This, in turn, suggests that there are at least two genetic components of seedling vigor: that due to seed weight and that due to seedling vigor per se. Seven seed-weight classes of cicer milkvetch ranging from 3.2 to 5.6 g/1000 seeds were compared for several seedling traits, including hypocotyl diameter, hypocotyl length, radicle length, total seedling length, seedling dry weight, cotyledon dry weight, and cotyledon area (Townsend, 1979d). Seed weight and hypocotyl diameter were significantlycorrelated (r ranged from 0.83 to 0.99) with all characters except hypocotyl length (r = 0.48 and 0.7 I , respectively). Seedling competition of sainfoin, birdsfoot trefoil (Lotus corniculutus L.), and cicer milkvetch was evaluated in monoculture and mixed culture under greenhouse conditions (Smoliak and Hanna, 1977). Parameters measured included leaf area; leaf, root, and total seedling weights; and specific leaf weight. In monoculture, leaf, root, and total seedling weights were similar for sainfoin and cicer milkvetch, but higher than was found for birdsfoot trefoil. In mixed culture, sainfoin was more aggressive than cicer milkvetch and birdsfoot trefoil, and cicer milkvetch was more aggres-

CICER MILKVETCH (Artragah ricer L.)

26 1

sive than birdsfoot trefoil. Of the mixtures evaluated, cicer milkvetch and birdsfoot trefoil were the most compatible with each other. Poor seedling establishment of cicer milkvetch is related to slow early seedling development (Smoliak et al., 1972). Cicer milkvetch seedlings grew more slowly than alfalfa and sainfoin seedlings when the roots were held at 13, 18, or 27°C. At 7"C, cicer milkvetch seedlings failed to emerge and grow. When alfalfa and cicer milkvetch were grown on coal mine spoil or on topsoil over spoil in a greenhouse study, cicer milkvetch produced slightly less top and root growth than did alfalfa (Nicholas and McGinnies, 1982).

C. FLOWERING In the seedling year under field conditions, the percentage of nonflowering cicer milkvetch plants ranged from 44% for unselected populations to 22% for populations selected for improved forage yield (Townsend, 1980b; Townsend and Ackerman, 1975). During the seedling year in the field, nonflowering plants within polycross progenies ranged from 2 to 51% (Townsend, 1980b) (Fig. 2). After going through the winter in the field, all plants in 13 of 34 progenies flowered the following spring; only 3.5% of the total population did not flower in comparison to 22% in the seedling year. On the regrowth in the second year, nonflowering plants within progenies ranged from 4 to 40%, which was similar to that in the seedling year. These field studies indicated that some plants required vernalization for flower induction.

D. VERNALIZATION Cicer milkvetch does not flower in artificial environments that induce flowering of the more commonly grown forage legumes. Pretreatment at relatively cool temperatures enhanced subsequent flowering in a growth chamber environment (Townsend and McGinnies, 1973). The pretreatment consisted of placing the clones in the dark for 3 weeks with 1 week each at 5 , 1.5, and -2°C. When field-grown plants of the cultivar Lutana were brought to the greenhouse every 2 weeks from mid-September through December, the percentage of plants flowering increased linearly (I = 0.93) with sampling date, ranging from 29 to 100% (Townsend, 1981b). Removal of the top growth of field-grown plants in mid-September reduced the subsequent percentage of plants flowering in the greenhouse

C . E. TOWNSEND

262

8

'

Y

50

v)

5a

*z

40

30

[L

W

i% 20

9 LL

z

4

10

0

LUTANA

POLYCROSS PROGENIES

Figure 2. Distribution of 34 polycross progenies of cicer milkvetch and the cultivar. Lutana for percentage of nonfloweringplants in the field during:(A) the seedling year, (B) the first growth period of the second year, and (C) the second growth period of the second year. Each bar represents one progeny, but the position of a progeny usually varies among graphs. (From Townsend, 1980b.)

from 75 to 33% and reduced the average number of racemedplant from 6.2 to 1.7. The reason for reduced flowering following the removal of top growth was not known, but reduced levels of carbohydrate reserves in the roots and rhizomes may have played a role. The percentage of plants flowering and the number of racemes/plant were always higher following vernalization at 5/2OoC in the laboratory than following vernalization in the field. Frequently, over 90% of the plants flowered and up to 17.0

CICER MILKVETCH (Astrugulus ricer L.)

26 3

racemes/plant were obtained following vernalization at 5/20°C in the laboratory. Although 8OYo or more of the plants flowered following field vernalization, the number of racemes/plant (3- 7) was especially low. The flowering response of plants from the cultivar Lutana varied considerably in the first, second, and third growth cycles after vernalization (Townsend, 1981b). Significant differences occurred among the three cycles for number of racemes/plant (13.3, 6.4, and 5.8 for the first, second, and third cycles, respectively), but not for percentage of plants flowering (97, 80, and 78%). Seedling age significantly influenced the effectiveness of vernalization (Townsend, 1982). Seedlings in early stages of development were not as receptive to the vernalization stimulus as were those in later stages, i.e., 60% or less of the seedlings at the primary-leaf or third-leaf stages flowered, whereas 80 to 98% of the seedlings in the fifth- or ninth-leaf stages flowered. Differences related to seedling ages were even more striking when compared on the basis of number of racemes/plant, i.e., over three times as many racemes were produced on older seedlings (16.7) as on the younger seedlings (4.9). Seedlings should be at the fifth-leaf stage of development or older before vernalization. Seedlings at the ninth-leaf stage of development were more receptive to the vernalization stimulus when a dark period was involved than when seedlings were maintained under continuous light. Averages of 15.0 and 4.0 racemes/plant were produced following 14- and 24-hr photoperiods, respectively (Townsend, 1982). This suggests that continuous light inhibited the perception of the vernalization stimulus or restricted metabolic reactions associated with vernalization. Polycross progenies differed significantly for response to 10 vernalization treatments, and progeny means ranged from 6.2 to 17.4 racemeslplant (Townsend, 1982). The relative performance of the progenies for flowering response to chilling in the laboratory was similar to that in the field. The progenies also differed significantly for response to photoperiod during vernalization. The average number of racemes/plant for the progenies ranged from 6.0 to 25.2 and from 0.6 to 8.6 for the 14- and 24-hr photopenods, respectively.

E. PHOTOPERIOD Photoperiod significantly influenced the percentage of plants flowering and the number of racemes/plant following vernalization (Townsend, 1981b) (Table I). Only 30% of the plants flowered under the 13-hr photoperiod whereas 97% or more of the plants flowered under the 15- and 16-hr

C. E. TOWNSEND

264

Table I Influence of Photoperiad 011 Flowering of Lutnnn Cier Milkvetch after Vernalization in Field or Laboratorf No. of racemes/plant

Photoperiod

Plants flowering

(hours)

(%)

Mean(SE)

Range

13 14 15 16

30 78 98 97

1.9 (0.6) 5.8 (0.7) 11.6 (1.0) 13.3 (1.2)

0-20 0-26 0-26 0-37

'From Townsend ( 198 1b).

photoperiods. The longer photoperiods not only increased the percentage of plants flowering, but also increased the profuseness of flowering. So much variability existed among plants for photoperiodic response to flowering that it was not possible to determine the critical photoperiod. For example, in the 13-hr treatment one plant produced 20 racemes, which was comparable to the better plants in the 14- and 15-hr treatments. Aborted racemes were observed on several plants in the 13-hr treatment, but not on plants in the other treatments. Apparently, the 13-hr photoperiod was adequate for flower initiation but was not adequate for the completion of flower development. Most mature plants of cicer milkvetch will flower adequately if they are vernalized at 5 "C night/20°C day temperatures for 6 weeks and then are grown at warm temperatures under a 15-hr photoperiod. Photoperiod influenced the amount of regrowth following a 1 August harvest (Townsend, 1988). Growth chamber studies showed that the improved growth of some plants in the third growth period (1 August - 15 September) is due to the absence of a photoperiod-induced dormancy, i.e., the growth of some plants is not reduced as much by the decreasing photoperiods as is that of other plants. When nonvernalized seedlings were grown in the greenhouse under a photoperiod adequate for flowering, seedlings grown during the spring months produced more flowers (35% of the plants flowered with an average of 2.2 racemes/plant) than seedlings grown during the autumn months (2% of the plants flowered with an average of 0.02 racemes/plant) (Townsend, 1982).

CICER MILKVETCH (Astragdus cicer L.)

265

IV. CULTURE A. ADAPTATION Cicer milkvetch is adapted to the central and northern Great Plains and adjacent Rocky Mountains of the United States, and to similar areas in western Canada that receive 400 mm or more of annual precipitation. It made outstanding growth when associated with sod-forming grasses such as smooth bromegrass (Bromus inermis Leyss.) and intermediate wheatgrass [ Thinopyrum intermedium(Host.) Barkworth and Dewey] in Kansas, Nebraska, and North Dakota (Atkins, 1953). Studies in eastern Colorado demonstrated that cicer milkvetch has potential as a rangeland species (Townsend et al., 1975). When seeded on topsoiled, stripmined lands in Montana, it showed good establishment, survival, canopy cover, and productivity (Holechek et d.,1981, 1982). Plantings of cicer milkvetch in western Oklahoma have met with mixed results, with moisture being the most limiting factor (Kneebone, 1959; Berg, 1990). Cicer milkvetch is widely recommended as a component of mixtures for the revegetation of disturbed lands in the western United States (Cook et al., 1974; Thornburg, 1982; Wasser, 1982). It is recommended for planting in the following ecological communities: the north aspect of the sagebrush- foothill (250- 330 mm of precipitation) and piiion-juniper (330-400 mm) communities; lower (330-400 mm) and upper (430560 mm) zones of the ponderosa pine and mountain brush communities; and the aspen community (2500 mm) (Cook et al., 1974). Plummer et al. (1968) noted that cicer milkvetch grew well on the most favorable sites of the piiion-juniper and big sagebrush belts of Utah. It is adapted at elevations of 1650-2300 m in northeastern California (Cornelius and Talbot, 1955) and in parts of the Pacific Northwest and Great Basin States above 1200 m (Hafenrichter et al., 1968). Cicer milkvetch is one of the most winter-hardy species of cultivated forage legumes. It was the only one of six legume species (alsike clover, Trifolium hybridum L.; red clover, Trifolium pratense L; white clover, Trifolium repens L.; alfalfa; and birdsfoot trefoil) that showed adequate initial stand establishment on an N-deficient, subalpine (3400 m) site in Colorado (Berg and Barrau, 1978). After 25 years, there is still an excellent stand of cicer milkvetch at this site (W. A. Berg and R. L. Cuany, personal communication, 1989). Of the 10 species of forage legumes seeded at five subalpine (3230- 3350 m) sites in Colorado, cicer milkvetch received high ratings for both establishment and persistence (Etra et al., 1984). In addition to the species used by Berg and Barrau (1978), Etra et al. (1984)

266

C. E. TOWNSEND

included big trefoil (Lotus uliginosis Schkuhr), flat pea (Lathyrus sylvestris L.), crownvetch, and sainfoin. Cicer milkvetch has persisted for 20 years at a 2200-m site in central Wyoming that receives 28 cm of precipitation annually (G. E. Schuman, personal communication, 1989). Cicer milkvetch was also the most winter-hardy species of forage legumes evaluated at numerous sites in western Canada (Johnston et al., 1975), and has persisted and performed well for over 35 years in Alberta, Canada (Moyer, 1989). Cicer milkvetch is well adapted in Minnesota (Marten et al., 1987; 1990; Sheaffer and Marten, 1991) and Iowa (Kephart et al., 1990), and appears to be adapted to areas of New York (R. F. Lucey, personal communication, 199I), Massachusetts (Zak et al., 1972), and Kansas (G. L. Posler and A. W. Lenssen, personal communication, 1991). It did not persist in a mixture with tall fescue (Festuca arundinacea Schreb.) in Georgia (Dobson et al., 1976). Although cicer milkvetch is native to central and eastern Europe, there are only a few reports concerning its use. Yakimova and Kolev (1981), Buyanova ( 198I), and S. T. Szabo (personal communication, 1990)stated that cicer milkvetch could be grown as a cultivated crop in the Baltic region of Russia, in Bulgaria, and in Romania, respectively; S. T. Szabo (personal communication, 1990) also stated that cicer milkvetch can be found on revegetated sites in Hungary. Haraszti and Vetter (1989)reported that cicer milkvetch is widespread in Hungary and is often the dominant species in grasslands.

B. SOILSAND SOILFERTILITY Cicer milkvetch is adapted to a wide range of soil types and tolerates slight acidity to slight alkalinity, but prefers calcareous soils (Hafenrichter et al., 1968). It does better than alfalfa on some soils with a high water table. The soil fertility requirements of cicer milkvetch appear to be similar to those of alfalfa. Cicer milkvetch showed promise for roadside stabilization in Massachusetts (Zak et al., 1972). Successful seedling establishment, however, required a soil pH between 6.5 and 7.0 and high levels of Ca, P, and K. Vickers et al. (1977) studied the effects of pH on cicer milkvetch growth in growth rooms. Lime markedly increased the growth of both tops and roots. Cicer milkvetch and Astragafusglycyphyllus L. tolerated a wide range of soil pH in Hungary, but the Ca requirement of cicer milkvetch was greater than that of A. glycyphyllus (Haraszti and Vetter, 1989). In New Zealand, pot trials using a high-country, acid-infertile subsoil showed that cicer milkvetch is intolerant to acid soils (Davis, 198I). Cicer

CICER MILKVETCH (Astruguh cicer L.)

267

milkvetch responded to added P, but foliage and soil analyses showed that both A1 toxicity and P deficiency may have reduced plant growth. Although some plants developed a few long, stout roots, not a single taproot developed. Bowman and Townsend (1990) evaluated the tolerance of cicer milkvetch to soils with pH values of 4.5,5.3,6.3, and 8.0 in greenhouse studies. The role of P, Ca, Al, Mn, and Mo in relation to cicer milkvetch tolerance to low and high pH values was investigated also. In the calcareous soil (pH 8.0) cicer milkvetch responded initially to P and after the first harvest to both N (plants were not inoculated with Rhizobium sp.) and P. When grown on the slightly acid (pH 6.3) soil, cicer milkvetch responded well to N. With the acid (pH 5.3) soil it responded to the combination of N and P, but not to N alone. Cicer milkvetch grew poorly and did not respond to the fertility treatments on the very acid (pH 4.5) soil apparently because of Mn and A1 toxicity and because of physical properties of the soil (the soil cemented on drying).

C. SEEDSCARIFICATION AND INOCULATION The impermeable seed coat of cicer milkvetch restricts water imbibition and results in poor seed germination and subsequent poor stand establishment. Hafenrichter et al. (1965) reported the initial germination and hardseed content of cicer milkvetch to be 11 and 75%, respectively. After 14 years in storage at several locations in the western United States, germination increased to over 60% and hard seeds decreased to about 10%. Carleton et al. (197 1) utilized a quick-swell test to determine the effectivenessof mechanical scarification on cicer milkvetch seeds. When 30 - 50% of the seeds imbibed water within 24 hr, they were considered to be properly scarified. After scarification the seed loses viability rapidly. Therefore, the seed should not be scarified until shortly before planting. Mechanical seed scarification equipment is available and, if properly used, it does an excellent job of scarification. Mechanical and H2S04scarification were equally effective in scarifying seeds of cicer milkvetch, but mechanical scarification is much safer (Miklas et al., 1987). The strophiole and seed tip were the two regions affected most by mechanical scarification, whereas the strophiole was the major region affected by H2S04scarification. The percentage of hard seed following mechanical scarification was significantly correlated with seed length, width, volume, and weight ( r = -0.91, -0.66, -0.88, and -0.82, respectively). Density was the only seed character that was significantly correlated (r = 0.70) with percentage hard seed following H2S04scarification. Progenies with larger seeds tended to be more easily scarified me-

268

C . E. TOWNSEND

chanically than those with smaller seeds. Consequently, weight would be the easiest seed trait to use when selecting and breeding for increased permeability following mechanical scarification. Seed should always be inoculated with the proper strain or strains of Rhizobium leguminosarum because these bacteria may not be endemic to the soil or their population may be too low to give adequate inoculation. The inoculant available in the United States is a moist peat-base powder containing the rhizobia bacteria. Comparative information is not available for cicer milkvetch on the effectiveness of the different methods of inoculation. The methods used to inoculate alfalfa seed (Burton, 1972), however, appear to be effective. Those methods are as follows: (1) mixing seed with a moist peat-base powder containing the bacteria, (2) adding water to the seed or inoculant and then mixing, or (3) coating the seed with a suspension of acacia gum, CaCO, , and rhizobia bacteria. Considerable variability exists among strains of rhizobia bacteria with regard to effectiveness on cicer milkvetch (J. C. Burton and C. E. Townsend, unpublished observations, 1970). Mature, effective nodules are rather large and elongated, frequently clustered, and have pink to reddish-brown centers. In a greenhouse study involving the N,-fixing ability of several Astragalus spp., cicer milkvetch had the second highest concentration of N and produced the greatest herbage yield (Davis, 1982a). Accessions of cicer milkvetch differed for effectiveness of N, fixation under laboratory conditions (Walsh et al., 1983). The variability was of a magnitude sufficient to be potentially valuable in a plant breeding program. Shoot dry weight was a good estimator of nodule activity and nodule dry weight in an N-free rooting medium for alfalfa, cicer milkvetch, and sainfoin (Major et al., 1979). Major et al. (1979) concluded that shoot dry weight could be used in a nondestructive manner to identify plants with superior N,-fixing ability. The amount of symbiotically fixed N, in the herbage of seedling stands of forage legumes treated with fresh dairy cow excreta was less for cicer milkvetch (7 kg ha-') than for red clover (29 kg ha-'), alfalfa (26 kg ha-'), and birdsfoot trefoil (22 kg ha-') (Russelle and Buzicky, 1988). In the year after establishment, the amount of symbiotically fixed N, in the herbage averaged 77, 7 1, 33, and 55 kg ha-' for red clover, alfalfa, birdsfoot trefoil, and cicer milkvetch, respectively. The higher yield of cicer milkvetch in the urine treatment (4.0 Mg ha-') in comparison to that of the symbiotically fixed N, treatment (2.9 Mg ha-') indicated that the N requirement for maximum production was not met by N, fixation. In some instances, the low yield of cicer milkvetch can probably be attributed to relatively slow nodulation due to the lower population of effective rhizobia bacteria in the soil, in contrast to that for the other species. Studies involving the compar-

CICER MILKVETCH (Artragoh cicer L.)

269

ative forage yields of alfalfa and cicer milkvetch alone and in mixtures with grasses indicate that the N,-fixing ability of cicer milkvetch under field conditions is similar to that of the more commonly grown legumes.

D. ESTABLISHMENT Early seedling vigor of cicer milkvetch is relatively poor when compared to that of alfalfa (Townsend and McGinnies, 1972a; McGinnies and Crofts, 1986).Good stands of cicer milkvetch, however, can be obtained if care is taken with respect to seedbed preparation, date and depth of planting, grass association, and weed control. Scarified and inoculated seeds should be planted about 2 cm deep in a firm, well-prepared seedbed with 30-cm spacings between rows under imgation and wider spacings under dryland. A depth-control device and a packer wheel to firm the soil after planting are very important. Seeding rates for pure stands of cicer milkvetch and cicer milkvetch -grass mixtures should be about 9 and 6 kg ha-', respectively (Stroh et al., 1972). Cicer milkvetch should not be planted with a companion crop because of competition. In areas where an August seeding does not permit enough plant growth for winter survival, seedings should be made in the spring. For reclamation seedings at high elevations in the western United States, nonscarified seed should be planted in late October before the first permanent snow cover of the season (Etra et al., 1984). However, in high-altitude irrigated meadows of the western United States, plantings should be made in late June or early July. In Wyoming, early spring seedings of alfalfa and cicer milkvetch produced better seedling emergence than did fall seedings in 2 of 3 years (Hart and Dean, 1986). Cicer milkvetch did not become established following frost seeding in a grass sod in Iowa (George, 1984). Planting patterns influenced the ease of establishingcicer milkvetch with cool-season grasses (Kenno et al., 1987). The planting patterns were (1) legume and grass seeded as a mixture in the same row, (2) legume and grass seeded alone in alternate rows, and (3) legume and grass seeded alone in rows perpendicular to each other. At the end of the seedling year the percentage of cicer milkvetch in the mixture was 38, 17, and 17% for the alternate, cross, and mixed seeding patterns, respectively. Legume content was 17, 23, 28, and 32% for the orchardgrass (Dactylis glomerata L.), smooth bromegrass, meadow bromegrass (Bromus biebersteinii Roem. and Schult.), and intermediate wheatgrass, respectively. By the end of the second harvest year, there was little difference among the planting patterns because the cicer milkvetch content of the forage ranged from 85 to 90% (Townsend et al., 1990). The legume content in the orchardgrass, meadow

270

C . E. TOWNSEND

bromegrass, smooth bromegrass, and intermediate wheatgrass mixtures was 86, 86, 90, and 96%, respectively. The rhizomatous trait of cicer milkvetch permits a relatively poor stand to improve substantially within 1 or 2 years under favorable conditions. Selecting seeds that are large in size or high in weight has been suggested as a means for improving the emergence of small-seeded forage species, because large seeds generally give greater seedling emergence from deeper depths of planting compared to small seeds. In greenhouse studies, seed weight did not influence seedling emergence of falcatus milkvetch (Astrugulusfulcutus Lam.), but large seeds of cicer milkvetch gave greater seedling emergence than did small seeds (Townsend, 1972a). The relationship between seed weight and seedling emergence was variable for both species at two field locations, but a significantly greater number of seedlings was obtained from large seeds than from small seeds in two of four studies. In general, there appeared to be more promise for improving seedling vigor through selection for large seeds in cicer milkvetch than in falcatus milkvetch. Seeds of falcatus milkvetch are about 20% heavier than those of cicer milkvetch. Seven seed-weight classes (3.2 to 5.6 g/lOOO seeds) of cicer milkvetch differed significantly for seedling emergence from 1.3-, 2 5 , and 3.8-cmdeep plantings in the field (Townsend, 1979d). There was no consistent trend among seed weight classes for seedling emergence at the 1.3-cm depth, but at the 2.5-cm depth seedling emergence generally increased with increasing seed weight. The 5.6-g class consistently gave the highest emergence from the 3.8-cm depth, i.e., seedling emergence of the 5.6-g class was about twice that of the 3.2-g class. Seed weight was significantly correlated ( r = 0.82 to 0.98) with seedling emergence from the 3.8-cm-deep planting. Correlation coefficients between seedling traits (hypocotyl diameter, hypocotyl length, radicle length, total seedling length, seedling dry weight, cotyledon dry weight, and cotyledon area) and seedling emergence from the 3.8-cm depth were generally significant except for those involving hypocotyl length. Because seed weight is easily obtained and was highly correlated with all seedling characters except hypocotyl length and with seedling emergence from the 3.8-cm-deep planting, there appears to be no advantage to using any of the other characters as criteria in selecting for early seedling vigor. Cicer milkvetch and alfalfa were successfully established by sod seeding after the pasture sward had been suppressed by glyphosate [N-(phosphonomethy1)glycinel (Malik and Waddington, 1990). Legume establishment was best when fall seeded after sod suppression than when spring seeded after sod suppression the previous fall. Nonscarified seed was used. Although initial seedling emergence of alfalfa (49 plants m-l of row) was

CICER MILKVETCH (Astragalus cicer L.)

27 1

twice that of cicer milkvetch (24 plants m-' of row) in the fall-seeded trial, the rhizomatous trait permitted cicer milkvetch to develop a stand equal to or slightly better than that of alfalfa 4 years later (51 versus 46% ground cover). Cicer milkvetch can become established in unexpected and relatively harsh sites. It was established on a reclaimed mine site in Wyoming from seed that had passed through the gut of pronghorn antelope (Antilocapra americana) (G. E. Schuman, personal communication, 1989). The pronghorns grazed mature seed pods and then transported the hard seeds through feces to a nearby area where the seeds germinated and developed into mature plants.

E. PESTRESISTANCE Relatively little is known about the susceptibility of cicer milkvetch to insect pests. In Canada, aphids, thrips, seed chalcids, sweetclover weevils (Sitona spp.), and grasshoppers attack cicer milkvetch (Johnston et al., 1975). Aphids, thrips, seed chalcids, and grasshoppers have been observed on cicer milkvetch in Montana (Stroh et al., 1972). In Colorado, cicer milkvetch was readily grazed by heavy populations of grasshoppers (unpublished observations, 1982). Grasshoppers feed on the flowers and developing seed pods before grazing the leaves. Blister beetles (Epicauta spp.) have also been observed feeding on the flowers. Seed chalcids are present every year in Colorado, with the infestation being much worse in some years than others. When 15 cultivars representing seven species of forage legumes were evaluated for suitability as food plants for several species of rangeland grasshoppers (lesser migratory, Melanoplus sanguinipes F. ; differential, Melanoplus diferentialis Thomas; red-legged, Melanoplus femurrubrum DeGeer; two-striped, Melanoplus bivittatus Say; Melanoplus packardii Scudder; Melanoplus gladstoni) in North Dakota, one cultivar each of alfalfa, birdsfoot trefoil, and cicer milkvetch was least preferred in greenhouse and field tests (Hewitt et al., 1982). Also, in an earlier study (Hewitt, 1969), cicer milkvetch was not preferred by M. sanguinipes. When sufficiently large populations of insects were present in Iowa, (1) there were significantly fewer alfalfa weevils (Hypera postica Gyllenhal) on cicer milkvetch and birdsfoot trefoil than on alfalfa, with no difference between cicer milkvetch and birdsfoot trefoil, (2) there were significantly more potato leafhoppers (Empoascafabae Hams) on alfalfa than on cicer milkvetch and more on birdsfoot trefoil than on cicer milkvetch, and (3) there were significantly more pea aphids (Acyrthosiphonpisum Hams) on

272

C. E. TOWNSEND

alfalfa than on cicer milkvetch, with no difference between cicer milkvetch and birdsfoot trefoil (Kephart et al., 1990). The alfalfa weevil did not feed on cicer milkvetch in Colorado, but the other insects werenot present or were so few in number that the level of tolerance could not be assessed (Townsend et al., 1990). In other studies, the blue alfalfa aphid (Acyrthosiphonkondoi Shinji) and the pea aphid were capable of reproducing on cicer milkvetch under both field and greenhouse conditions, but cicer milkvetch was a relatively poor reproductive host when compared to alfalfa (Ellsbury and Nielson, 1981). The Russian wheat aphid (Diuraphis noxia Mordvilko) did not reproduce or survive on cicer milkvetch (Kindler and Springer, 1989). As with insect pests, relatively little is known about the susceptibility of cicer milkvetch to diseases. Sclerotinia root rot, crown rot, and stem rot have been reported on cicer milkvetch in Montana (Stroh et al., 1972), but most forage legumes are susceptible to this disease complex. In Canada, several root, crown, or stem rot diseases have been noted, but they have caused little damage (Johnston et al., 1975). In Colorado, Fusarium root and stem rots may attack an occasional plant in the field, but attempts to reinoculate seedlings with the fungus were not successful (E. G. Ruppel, personal communication, 1975). Ruppel has also isolated Stemphylium spp. and Alternaria spp. from leaves and stems of mature plants in the field. The leaf spots incited by the latter two fungi have been observed only once on cicer milkvetch in Colorado and that was during a period with high relative humidity. Observations indicate that the vigorous rhizome system helps cicer milkvetch to outgrow effects of the root and crown diseases. As cicer milkvetch becomes more widely grown, pests will undoubtedly become more of a problem.

F. POLLINATION REQUIREMENTS Cicer milkvetch is a cross-pollinated species. In Colorado, seed set was excellent under open-pollination conditions, i.e., seed set ranged from less than 100 seeds to over 300 seeds per raceme with the 176- to 200-seed class being the most frequent (Townsend, 1971a). When pollinators were excluded, no seed was set (Townsend, 1971b). The pollinating mechanism must be activated (tripped) before pollination can occur. Cicer milkvetch has the simple valvular tripping mechanism like that found in the clovers (Trifolium spp.). This tripping arrangement permits repeated pollinator visits. Bumble bees (Bombus spp.) are the principal pollinators in Montana, Wyoming, and Colorado (Stroh et al., 1972; unpublished observations, 1970). In western Canada, bumble bees, honey bees (Apis mellgera

CICER MILKVETCH (A.rtragalus ricer L.)

273

L.), and leaf-cutting bees (Megachile rotunda Fab.) have been observed to work cicer milkvetch (Richards, 1986). Flowers remained receptive to pollination for 7 days. Age of flower at time of pollination influenced subsequent seed set, with those pollinated the first 4 days after opening producing more seed than the older flowers. In Colorado, flowers do not appear to remain receptive to pollination for more than 2 or 3 days (unpublished observations, 1970). Fertilization occurs 24 to 48 hr following pollination (Townsend, 1980a). Seed matures about 28 days following fertilization. A few days following fertilization the ovary begins to expand and develops into an inflated, bladderlike pod. The pod changes in color from green to reddishgreen to black with a leathery texture at maturity. The pods are subsessile and remain attached to the rachis at maturity. Seeds seldom shatter but they detach and rattle inside the pod. The number of seeds per pod ranges up to 12. Generally, mature seeds are bright golden-yellow in color, but sometimes they will have a light-greenish tinge. Seeds are flat to rounded in shape, about twice the size of alfalfa seeds, and there are about 250,000 to 285,000 seeds kg-'.

G. SEEDPRODUCTION Cicer milkvetch has high seed production potential. Yields in excess of 1000 kg ha-* have been obtained under irrigation, with the average yields ranging from 450 to 675 kg ha-' in Montana (Stroh et al., 1972). Under dryland conditions with good soil moisture, yields ranged up to 225 kg ha-'. Seed yields in Alberta, Canada, ranged from 57 to 725 kg ha-' (Richards, 1986)and from 6 14 to 858 kg ha-' in 20- and 3-year-old stands, respectively (Johnston et al., 1971). Row spacings of 45 to 90 cm are used under irrigation, with wider spacings for dryland plantings. Wide-row spacings permit cultivation for weed control and furrow irrigation. Rainfall plus irrigation should not exceed about 450 mm annually (Stroh et al., 1972). Because bumble bees are the principal pollinators, plantings should be near bumble bee habitats, for example, undisturbed grassland areas, including fence rows', ditch banks, and windbreaks. Observations suggest that one bumble bee per 25 m2 will provide adequate pollination. Seed should be harvested only from the first growth because some plants require vernalization and do not flower during the second growth period. Cicer milkvetch is an indeterminate-floweringspecies; therefore, a few immature seed pods will be present at harvest. The best harvesting procedure is to windrow the seed crop and allow it to dry thoroughly before combining.

2 74

C. E. TOWNSEND

H. WEED CONTROL Weed control measures such as clipping or mowing have been used with varying success in the establishment of cicer milkvetch (unpublished observations, 1979). Preliminary results suggest that mowing alone is not a satisfactory treatment because mowing of broadleaf weeds opens the canopy and permits vigorous growth of grass weeds, which were very competitive with cicer milkvetch seedlings. Kerr and Klingman (1960) encountered a similar problem when mowing alone was used as a weed-control practice in the establishment of birdsfoot trefoil. Successful stands of cicer milkvetch, however, were obtained by controlling weeds with mowing at a wildlife refuge in South Dakota (R. Gilbert,personal communication, 1992). Selective herbicides should improve the establishment of cicer milkvetch, but relatively little is known about the tolerance of this species to herbicides. Currently, no herbicides are registered for use on cicer milkvetch in the United States, whereas in Canada, flamprop [N-benzoylN-(3-chloro-4-fluorophenyl-~~-alanine], sethoxydim (2-[ 1-(ethoxyimino) butyll-5-[2-(ethylthio)propyl]-3-hydroxy-2-cyclohexen-1-one}, and trifluralin [2,6-dinitro-N,N-dipropyl-4-(trifluoromethyl)benzenamine] are registered. Postemergence herbicides are more suitable than preplantincorporated herbicides, particularly in semiarid regions, because soil moisture is not lost when herbicides are applied postemergence. In Canada, Moyer et al. (1979) evaluated the tolerance of cicer milkvetch to 6 preplant-incorporated and to 17 postemergence herbicides in growth chamber and field studies. Cicer milkvetch was tolerant to barban (4-chloro-Zbutynyl 3-chlorophenyl carbarnate), trichloroacetic acid (TCA), and diclofop ((+)-2-[4-(2,4-dichlorophenoxy)phenoxy]propanoic acid}, which control annual grasses, and to trifluralin, which controls some broadleaf weeds in addition to annual grasses. Cicer milkvetch had marginal tolerance to bromoxynil (3,5-dibromo-4-hydroxybenzonitrile) and bentazon [ 3-( 1-methylethyl)-(lH)-2,1,3-benzothiadiazin-4(3H)-one 2,2dioxide], which control some broadleaf weeds. Seedlings of cicer milkvetch were evaluated for tolerance to 4-(2,4-dich1orophenoxy)butanoic acid (2,4-DB) and (2,4-dichlorophenoxy)acetic acid (2,4-D) alone and in combination with diclofop (Townsend and Schweizer, 1984). In the greenhouse, seedlings in the second- through the fifth-leaf stages of growth were more tolerant to 2,4-DB alone and in combination with diclofop than were seedlings in the cotyledonary and primary-leaf stages. All rates of 2,4-D amine with or without diclofop reduced seedling growth at all stages of development. In the field, 2,4-DB with or without diclofop did not reduce seedling growth. The higher rates of 2,4-D amine alone (1.02 and 1.36 kg a.i. ha-') suppressed seedling

CICER MILKVETCH (Astruguhs ricer L.)

275

growth throughout the growing season. All combinations of 2,4-D amine plus diclofop suppressed seedling growth, with the higher rates being the most suppressive. Seedlings of cicer milkvetch had good tolerance to imazethapyr (2[4,5-dihydro-4-methyl-4-(1-methylethyl)-5-oxo-1H-imidazol-2yl]-5-ethyl3-pyridinecarboxylicacid) at 0.07 and 0.14 g ha-' and to bromoxynil at 0.28 g ha-' (S. D. Miller, personal communication, 1992). However, bromoxynil at 0.43 g ha-' caused substantial injury. Moyer ( 1989) studied the effect of several herbicides on cicer milkvetch and weeds during the year of establishment (year 1) and on the weed component of the forage for an additional 4 years. Although cicer milkvetch yields were low following the herbicide treatments in year 1, herbicides were required in year 1 for acceptable yields in year 2, i.e., weeds contributed less than 5% of the total forage from herbicide-treated plots in year 2 whereas the control treatment averaged about 50Yo weeds. After year 2, annual and winter annual weeds were of little importance. From year 3 until the end of the study, dandelion (Tarmacum spp.) was the major weed component of the forage. Trifluralin or S-ethyl dipropyl carbamothioate (EPTC) in combination with bromoxynil and 2,4-DB with sethoxydim or diclofop provided satisfactory control of both grass and broadleaf weeds. Weeds create problems in established stands of cicer milkvetch in addition to those encountered in seedling stands. When cicer milkvetch was harvested from two to seven times annually, the annuals kochia [Kochia scoparia (L.) Schrad.], pigweed (Amaranthus spp.), foxtail (Setaria spp.), and barnyard grass [Echinochloa crusgalli (L.) Beauv.] were the principal weeds in the five-, six-, and seven-cut treatments during the first 2 years of a 3-year study (Townsend et al., 1978). The other treatments were relatively free of weeds. By the third harvest year, the composition of the weed population had changed considerably, with perennial dandelion increasing in all treatments and annuals decreasing. The experimental area was essentially free of weeds at the time the study was initiated because weeds were removed by hoeing during the year of establishment. Cicer milkvetch was successfully established in Canada by a preplanting incorporation of trifluralin at 1.1 kg ha-' and a postemergence application of 2,4-DB at 0.35 kg ha-' (Malik and Waddington, 1989). After the establishment of cicer milkvetch, the residual herbicides hexazinone [3-cyclohexyl-6-(dimethylamino)-l-methyl-1,3,5-triazine-2,4(lH,3H)-dione], metribuzin [4-amino-6-(1,l-dimethylethyl)-3-(methylthio)-1,2,4-triazin-5(4H)one], and terbacil [5chloro-3-(1,l-dimethylethyl)-6-methyl-2,4(1H,3H)pyrimidinedione] were applied alone at 1 kg ha-' in October. The graminicides sethoxydim at 0.8 kg ha-' and fluazifop ((+)-2-[4-[[5(trifluoromethyl)-2-pyridinyl]oxy]phenoxy]propanoic acid} at 0.5 kg ha-'

276

C. E. TOWNSEND

were applied the following spring after plant growth had resumed. Cicer milkvetch along with alfalfa tolerated all herbicides.

V. UTILIZATION A. HAY Forage quality of cicer milkvetch is excellent when measured by factors such as crude protein (CP), in vitro digestible dry matter (IVDDM), cell wall constituents (CWCs), lignin, hemicellulose, and silica (Table 11) and compares favorably with that of other forage legumes (Seamands et al., 1972; Stroh et al., 1972; Johnston et al., 1975; Townsend et al., 1978; Gabrielsen et al., 1985; McGraw and Marten, 1986; Marten et at., 1987, 1990; Kephart et al., 1990). For livestock nutrition, of the mineral elements examined (Ca, Mg, P, and K), only P was possibly deficient (Church, 197 I). This potential deficiency was confined to the two-harvest treatment (Townsend et al., 1978). The green forage of cicer milkvetch contains about four percentage points more moisture than is contained in other forage legumes. Alfalfa, birdsfoot trefoil, and cicer milkvetch were compared for CWCs, CP concentration, and IVDDM concentration in Iowa (Kephart et al., 1990). Under a three-cut system of management, the seasonal concentrations of neutral-detergent fiber (NDF) and acid-detergent lignin (ADL) were lower in cicer milkvetch than in alfalfa and birdsfoot trefoil; seasonal concentrationsof CP and IVDDM were generally higher in cicer milkvetch than in alfalfa and birdsfoot trefoil. When harvested twice, cicer milkvetch compared favorably to alfalfa and birdsfoot trefoil for seasonal concentrations of NDF and ADL. Concentrations of CP and IVDDM were lower in twice-cut cicer milkvetch than in alfalfa, birdsfoot trefoil, and thrice-cut cicer milkvetch. Alfalfa and cicer milkvetch did not differ for seasonal CP concentrations but did for seasonal CWC concentrations, with those for cicer milkvetch being the lowest (Gabrielsen et al., 1985). Increased frequency of cutting increased forage quality. The amino acid composition of cicer milkvetch forage was compared to that of alfalfa (Kaldy et al., 1978). Methionine, isoleucine, and valine were the first, second, and third most limiting amino acids for both species. The composition with respect to other amino acids was similar for both species, and total protein content was 22.7 and 20.3% in cicer milkvetch and alfalfa, respectively. Cicer milkvetch forage does not contain toxic levels of nitro compounds

Table I1 Influence of Frequency of Harvest on Forage Yield and Concentration of Quality Factors and Minerals in Cicer Milkvetch Forage and Rhizome Nonstruct~ualCarbohydrateConcentration'

Harvest treatment (no. of cuttings)

L.S.D. 0.05

Quality and mineral content (@g) Forage yield (Mg ha-')

Crude protein

IVDDM

CWC

Lignin

Hemicellulose

Cellulose

Si

Ca

Mg

P

K

10.5 10.7 9.7 10.2 9.0 9.2

191 227 250 269 217 284

648 687 694 704 704 700

384 326 306 292 273 270

74 66 72 65 53 54

45 31 39 42 63 39

238 214 171 165 149 147

37.8 22.3 24.0 25.9 18.2 30.6

25.5 23.3 25.0 25.0 26.5 24.9

3.7 3.4 3.6 3.8 4.3 3.6

1.4 1.9 2.1 2.2 2.5 2.7

28.4 31.5 31.2 32.6 30.0 31.3

0.9

10

11

16

8

11

13

9.5

1.3

us

0.2

ns

NOUS~~Ctural CHO 216 181 164 153 I49 122

'Modified from Townsend et al. (1978). Forage yield data are the averages of 3 years; quality and mineral data are the averages of 2 years; and the total nonstructural carbohydrate data were obtained at the end of the third harvest year. IVDDM, In vitro digestible dry matter; CWC, cell wall constituent; us, not significant.

278

C. E. TOWNSEND

(3-nitropropionicacid 3-nitro- 1-propanol)(Williams et al., 1976),tannins (Davis, 1973; Sarkar et al., 1976), oxalates, alkaloids (Davis, 1973; 1982b), or Se (Davis, 1972; Johnston et al., 1975). Sheep (Ovis aries) fed a cicer milkvetch-grass mixed-hay diet for 52 consecutive days remained healthy and suffered no observable side effects (Wegert, 1977). No cases of bloat in animals grazing the forage have been reported, but excessive foam formation has been noted in in vitro studies (Cooper et al., 1966). When inoculated with a spore suspension of the nonpathogenic fungus Helminthosporium carbonum Ullstrup, cicer milkvetch produces at least five isoflavonoid phytoalexins, namely, mucronulatol, astraciceran, maackiain, cajanin, and an unknown isoflavone (Ingham and Dewick, 1980; Martin and Townsend, 1989). S. S. Martin and C. E. Townsend (unpublished observations)determined the total amounts and quantitative composition of these five stress metabolites in about 50 plant introductions representative of the native ecological and geographic range of the species. Significant differences existed among the introductions for both parameters. Other compounds, including flavonoid glycosides (such as kaempferol, quercetin, and isorhamnetin derivatives) and the coumarins (scopoletin and umbelliferone), were found in populations from Armenia and Georgia (Alaniya et al., 1984). It is important to determine the factors that control the accumulation of these compounds because of their possible role in pest resistance or as antiquality agents. An antinutritive material occurs in the forage of cicer milkvetch that affects in vitro digestibility by mixed rumen microflora (Weimer et al., 1991). The water-soluble extract showed transient ( 18 to 24 hr) and concentration-dependent microflora inhibition. It was suggested that the inhibitor acts by preventing adhesion of the fibrolytic bacteria to cellulose. There was no difference between the Monarch and Lutana cultivars in inhibition levels. The primary spring growth patterns of alfalfa, birdsfoot trefoil, cicer milkvetch, and sainfoin were evaluated on a weekly basis from 10 May to 28 June for dry matter accumulation, IVDDM concentration and accumulation, and CP concentration and accumulation (McGraw and Marten, 1986). Although there were differences among species for the traits measured on most sampling dates, no species was consistently superior throughout the study. On June 14, cicer milkvetch, alfalfa, birdsfoot trefoil, and sainfoin had reached 5 , 10, 15, and 50% flower, respectively. Leaf-to-stem ratios reached about 1.O on 10 May, 24 May, 30 May, and 10 June for alfalfa, birdsfoot trefoil, sainfoin, and cicer milkvetch, respectively. On June 21, the leaf-to-stem ratios were 0.38, 0.36, 0.41, and 0.72 for alfalfa, birdsfoot trefoil, sainfoin, and cicer milkvetch, respectively. All species reached maximum dry matter accumulation on 2 1 June.

CICER MILKVETCH (Astruguh cicer L.)

279

Figure 3. Average distribution of forage yield of Lutana cicer milkvetch among cuttings within harvest treatments for 3 years in Fort Collins, Colorado. (From Townsend et al., 1978.)

When grown under irrigation and harvested two to seven times annually for 3 years in Colorado, the two- and three-cut treatments generally yielded the most forage and the seven-cut treatment yielded the least (Townsend et al., 1978) (Table 11). Cutting height was 7.5 cm for all treatments. The contribution by the first cutting ranged from 36 to 62% for the seven- and two-cut treatments, respectively. In the four-, five-, and six-cut treatments there appeared to be a cyclic distribution of yield in which a relatively high-yield cutting was followed by a relatively low-yield cutting (Fig. 3). A similar cyclic trend had been noted previously (Townsend, 1974b). Cicer milkvetch was slow to recover after cutting. Most of the regrowth originated from axillary buds and the active axillary buds were removed in the two-, three-, and four-cut treatments. About 7 to 10 days were required for inactive buds to produce new growth. However, in the five-, six-, and seven-cut treatments considerable vegetative material remained after clipping because much of the upright growth from active axillary buds was below the cutting height. Regrowth can originate at axillary buds, crown buds, and rhizome buds. Persistence was excellent for all harvests, but total nonstructural carbohydrates in the rhizomes decreased linearly (Y = -0.97) with increased frequency of cutting. Carbohydrate reserve trends in the taproot of cicer milkvetch were similar to those in the rhizome and they were less cyclic than the reserve

280

C. E. TOWNSEND

trends in the taproot of alfalfa (Gabrielsen et al., 1985) (Fig. 4). The carbohydrate reserves in the roots and rhizomes of cicer milkvetch were high in the spring, but became relatively low after the first harvest, and remained at that level or near it until late July. Then, there was a consistent accumulation of reserves during the remainder of the growing season. Carbohydrate reserve trends in the roots and rhizomes of cicer milkvetch resemble those reported for birdsfoot trefoil (Smith, 1962) and sainfoin (Cooper and Watson, 1968). The dependence of cicer milkvetch on reserves for regrowth after harvest appears to be less than that of alfalfa and to be more closely related to residual leaf area (Gabrielsen et al., 1985). The performance of cicer milkvetch relative to other forage legumes is variable and depends on location and number of harvests (Table 111). Generally, alfalfa yielded more than cicer milkvetch, although alfalfa yields ranged from 82 to 165% of cicer milkvetch yields. Cicer milkvetch yields compared favorably with those of the other species at all locations except Bozeman, Montana. Cicer milkvetch produced more hay under a twoharvest than a three-harvest treatment in Montana (Stroh et al., 1972) and Iowa (Kephart et al., 1990), whereas in Colorado (Townsend et al., 1978) and Minnesota (Sheaffer and Marten, 1991) there was little difference between the two- and three-harvest treatments. In the high elevations of Colorado, however, cicer milkvetch should be managed to give a one-hay harvest and an aftermath for grazing (E. G. Siemer, personal communication, 1985). Cicer milkvetch was not competitive in mixtures with sainfoin, birdsfoot trefoil, or red clover when harvested for hay for three harvest years in Montana (Cooper, 1979). By the first cutting of the third harvest year, however, the yield of cicer milkvetch alone did not differ significantly from the legume mixtures or from pure stands of alfalfa, sainfoin, birdsfoot trefoil, or red clover. The relatively low yield of cicer milkvetch in the second harvest of all years contributed to its lower total yield. It was also suggested that cool night temperatures limited the growth of cicer milkvetch. Cicer milkvetch is very competitive with cool-season grasses under irrigation. When compared in mixtures with cicer milkvetch, there was little, if any, difference among cool-season grass species with regard to compatibility because by the sixth year the amount of cicer milkvetch in the mixture ranged from 76 to 83% (Townsend et al., 1990). The grass species were smooth bromegrass, meadow bromegrass, crested wheatgrass [Agropyron cristatum (L.) Beauv. ssp. pectinatum (Bieb.) Tzvel.], intermediate wheatgrass, pubescent wheatgrass [ T. intermedium ssp. barbulatum (Schur.) Barkworth and Dewey], tall wheatgrass [Thinopyrurn ponticum (Podp.) Barkworth and Dewey], and creeping foxtail (Alopecurus arundin-

CICER MILKVETCH (Astragdus cicer L.) 500 -

A

Cicer Milkvetch

_-__

-

I-

3 V

Alfalfa

400I13

-I

I-

3

V

V

J

0

x

0

z s I-

28 1

/---

/

Ioot 200 -

100 -

I

SAMPLING DATE

B -

Cicer Milkvetch

I-

3 V

Alfalfa

Y Y

0

z

I-

20

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

SAMPLING DATE

Figure 4. Seasonal variation in total nonstructural carbohydrates(TNC) in the taproots of cicer milkvetch and alfalfa cut three (A) and four (B) times annually in 1980. (From Gabrielsen et al., 1985.)

Table I11 Comparative Forage Yields for Several Species of Forage Legumes and for Cicer Milkvetch

Source

Location' Colorado

Townsend (unpublished, 1971) Townsend (19746, Townsend ef al. ( 1990)

Iowa

Gabrielsen et al. (1985) 3cut kut Kephart ef al. (1990) 2-cut 3cut

Minnesota

3cut kut Stroh el al. (1972) Cooper ef al. ( 19786,

Bozeman (a) (b) Huntley New Mexico

Utah Canada

Forage legumes (% of cicer milkvetch yield) Alfalfa

Birdsfoot trefoil

11.4 10.8 5.6 11.2 10.4

126 106 96 108 105

-

10.6 9.5

119 127

-

9.2 7.0

125 165

92 122

10.2 9.8

104 131 146 99 109

104

Crownvetch

94

-

94

-

Sheafer and Marten ( 1991) 2cut

Montana

Cicer milkvetch forage yield (Mg ha-')

Kalispell Cooper (1979) Bozeman Melton (1973) Bleak (1969) Johnston ef al. (1975) Lethbridge Stavely

8.0 16.1 11.0 5.8 6.8 11.8 10.4 7.9 15.8 1.2 5.2 2.7

-

107 124 82 100 138 75

94 104

-

132 121 102 105 106

-

-

All studies except those. in Iowa, Minnesota, Utah, and Canada were conducted under irrigation.

* Clipping studies (simulated grazing).

Sainfoin

Kura clover

Red clover

Alsike clover

White clover

CICER MILKVETCH (Astrugdus cicer L.)

283

aceus Poir.). Tall wheatgrass did not persist beyond the second harvest year. The yield of cicer milkvetch alone and of several cicer milkvetchgrass mixtures approached the yield of alfalfa alone (1 2.1 Mg ha-'). In a similar irrigated study in Montana, Scheetz and Stroh (1982) reported that reed canarygrass (Phalaris arundinacea L.) was the most compatible grass with cicer milkvetch. Smooth bromegrass was too competitive with cicer milkvetch for the first 3 years of their 5-year study, but was compatible during the last 2 years. Orchardgrass and creeping foxtail were compatible with cicer milkvetch throughout their study. Meadow bromegrass, pubescent wheatgrass, and tall fescue were initially compatible with cicer milkvetch, but they lost stand after the third year. Western wheatgrass [Pascopyrum smithii (Rydb.) Love], Russian wild ryegrass [Psathyrostachys juncea (Fisch.) Nevski], timothy (Phleum pratense L.), Kentucky bluegrass (Poa pratensis L.), and switchgrass (Panicurn virgatum L.) were not compatible with cicer milkvetch. The mean yield of cicer milkvetch alone was significantly higher than the mean yield of all grass- cicer milkvetch mixtures. Studies in the Flint Hills region of Kansas involving mixtures of cicer milkvetch and switchgrass, sideoats grama [Bouteloua curtipendula (Michx.) Torr.], and indiangrass [Sorghastrum nutans (L.) Nash] showed that cicer milkvetch was too aggressive to be seeded with these warmseason grasses (G. L. Posler and A. W. Lenssen, personal communication, 1992). Within 3 or 4 years after seeding, cicer milkvetch dominated the stands. The IVDDM concentration of cicer milkvetch ranged from 550 to 575 g kg-I throughout the season and was higher than that for the native legumes, Illinois bundleflower [Desrnanthus illinoensis (Michx.) MacMill.], leadplant (Amorpha canescens Pursh), and roundhead lespedeza (Lespedeza capitata Michx.). The IVDDM concentration of cicer milkvetch was also higher than that for the three native grass species and for the other native legumes, catclaw sensitive brier [Schrankia nuttallii (DC.) Standl.] and purple prairieclover [Petalosternonpurpurea (Vent.) Rydb.], except early in the season when the IVDDM concentrations of these species were similar to those of cicer milkvetch. The crude protein concentration of cicer milkvetch was about 150 g kg-' and its relative relationship to the other species was similar to that for IVDDM. In Minnesota, cicer milkvetch was compared with six species of forage legumes under two-, three-, or four-cut regimes for forage yield, CP concentration, IVDDM concentration, and persistence (Sheaffer and Marten, 1991). Alfalfa was the only species to yield significantly more forage than cicer milkvetch and that was at the three- and four-cut treatments. Yields of kura clover (Trifolium ambiguum Bieb.) were reduced because of slow establishment. Cicer milkvetch compared favorably with all species for CP

2 84

C. E. TOWNSEND

and IVDDM concentrations, except for kura clover, which consistently had higher IVDDM concentrations. By the third harvest year, alfalfa, cicer milkvetch, and kura clover were the only species with adequate persistence to produce measurable yields under the four-cut treatment. Alsike clover, red clover, and crownvetch persisted for only two harvest years at all cutting treatments. Birdsfoot trefoil persisted for three harvest years at the two- and three-cut treatments, but only for 2 years at the four-cut treatment. Kura clover, however, persisted the best under all cutting schedules. In Connecticut, kura clover forage also had greater IVDDM concentrations than did cicer milkvetch, but cicer milkvetch had slightly greater CP concentrations than was found for kura clover (Allinson et al., 1985). A study in Minnesota that included “droughted” and “well-watered” variables demonstrated that (1) alfalfa had the most drought tolerance, (2) birdsfoot trefoil and cicer milkvetch had intermediate drought tolerance, and (3) red clover had the least drought tolerance of the four species evaluated (Peterson et al., 1992).A severe winter resulted in the death of all birdsfoot trefoil and red clover in both water regimes. About 30% of the alfalfa and cicer milkvetch persisted in the droughted regime, whereas only about 5% persisted in the well-watered regime. The rhizomatous trait of cicer milkvetch offers an opportunity for reestablishment if the winter kill is not too severe. Cicer milkvetch was compared with single- and double-cut red clovers for dry matter production in eastern Canada (Ottawa, Ontario) (Faris and Ta, 1985).None of the 18 single-cut red clover entries differed significantly from Oxley cicer milkvetch and only one of 25 double-cut entries yielded significantly more than Oxley. In the fourteenth year following establishment on a dryland site at a 2 150-m elevation in Utah, the yields of cicer milkvetch, Ladak alfalfa, and A-169 alfalfa in mixtures with grasses were essentially identical (1.2 Mg ha-’) (Bleak, 1969). However, about 28% of the total yield of the mixtures involving cicer milkvetch and Ladak was legume, whereas 37% of the A- 169-grass mixture was legume. In mixtures with crested wheatgrass on a second dryland site in Utah, cicer milkvetch increased total forage yield over that of crested wheatgrass alone, but the increase was less than that contributed by alfalfa and falcatus milkvetch in their grass mixtures (Rumbaugh et al., 1982). In western Oklahoma, cicer milkvetch and alfalfa were evaluated for persistence, forage yield, and apparent N, fixation on two Ndeficient soils (Berg, 1990). On one site, cicer milkvetch died during the summer of the second harvest year because of drought. Cicer milkvetch was marginally adapted to the second site and its mean 5-year forage yield was 3.1 Mg ha-’. Apparent total N, fixation over six growing seasons was 945 kg N

CICER MILKVETCH (Astmgalus ricer L.)

285

ha-'. Comparable values for alfalfa were 4.5 Mg ha-' of forage and 1488 kg N ha-'. When alfalfa alone, cicer milkvetch alone, and mixtures of each legume with five cool-season grass species were evaluated under dryland in western Nebraska, forage yields were higher for alfalfa alone (1.3 Mg ha-l) and alfalfa-grass mixtures (1.4 Mg ha-') than those for cicer milkvetch alone (0.9 Mg ha-') and cicer milkvetch-grass mixtures (1.2 Mg ha-') (Schultz and Stubbendieck, 1982). Alfalfa was much more competitive with the grasses than cicer milkvetch was; by the second year of the study over 50% of the alfalfa-grass mixture was alfalfa, whereas only 20% of the cicer milkvetch -grass mixture was cicer milkvetch. Concentration of crude protein was higher for alfalfa alone (189 g kg-') than that for cicer milkvetch alone (162 g kg-') (Schultz and Stubbendieck, 1983). Concentration of IVDDM, however, was higher for cicer milkvetch alone (730 g kg-') than that for alfalfa alone (680 g kg-l). In a study of seasonal dry matter yield and digestibility of seven grass species, alfalfa, and cicer milkvetch under dryland in eastern Montana, cicer milkvetch was one of the latest species to initiate spring growth and reach peak forage production, which was 30 days (20 July) after 10% bloom (White and Wight, 1981). The estimated IVDDM concentration of cicer milkvetch forage was the highest of any species throughout the growing season. Also, the dry matter yields at the 600-g kg-' stage of digestibility for cicer milkvetch, Altai wild ryegrass [Leymus angustus (Trin.) Pilger], and Russian wild ryegrass [P.juncea (Fisch.) Nevski] were significantly higher than those for the other species. Although cicer milkvetch ranked low for regrowth, its forage ranked high for digestibility (650g kg-I). The number of harvests of cicer milkvetch per year dramatically influences forage distribution. When harvested twice, about 60 and 70% of the total yield were produced in the first harvest in Colorado (Townsend et al., 1978) and Montana (Stroh et al., 1972),respectively. In Iowa, the percentage of total yield contributed by the first harvest of a two-cut schedule ranged from 62 to 77% (Kephart et al., 1990). In Colorado, mean percentage contribution for the first, second, and third cuttings to total forage yield was 54, 34, and 12%, respectively for cicer milkvetch-grass mixtures and 42, 32, and 2696, respectively, for alfalfa (Townsend et al., 1990). When the contribution of the first cutting to total yield exceeded 55% for cicer milkvetch and cicer milkvetch- grass mixtures, the contribution of the third cutting to total forage yield was reduced from 14 to 7%. When compared to alfalfa in Colorado, cicer milkvetch yields proportionally more forage in the first harvest, about the same in the second harvest, and less in the third harvest. In Iowa, the contribution of the first, second, and

286

C. E. TOWNSEND

third harvests to total yield was 57, 22, and 2196, respectively, for cicer milkvetch and 45, 36, and 1996, respectively, for alfalfa (Kephart et al., 1990). Total yield as well as the relatively low amount of the total yield contributed by the third cutting for cicer milkvetch may be increased by breeding and selection for types that are insensitive to the shortening daylength of mid- to late summer (Townsend, 1988).

B. PASTURE There have been relatively few grazing studies of cicer milkvetch. In a high-altitude irrigated meadow in Colorado, beef (Bos taurus) gains from a mixed cicer milkvetch-Garrison creeping foxtail pasture (434 kg ha-') were 84% of that from mixed N-fertilized smooth bromegrass-Garrison creeping foxtail pastures (5 14 kg ha-') (Rumburg, 1978). Forage yields of the cicer milkvetch-grass pasture (5.6 Mg ha-') were 87% of those from the fertilized grass pastures (6.4 Mg ha-'). The grass pastures were fertilized with 112 kg N ha-' annually. In an imgated pasture study in Nebraska, a pure stand of cicer milkvetch produced significantly less beef (563 kg ha-') than did a pure stand of alfalfa (710 kg ha-'), an alfalfa-grass mixture (987 kg ha-'), a cicer milkvetch-grass mixture (908 kg ha-'), and grass alone (965 kg ha-') (Nichols et a!., 1982). The grass-alone treatments were fertilized with 280 kg N ha-'. Under irrigation in Wyoming, a cicer milkvetch-Garrison creeping foxtail mixture produced 5.9 Mg ha-' of hay and 263 kg ha-' of beef, whereas a mixture of Eski sainfoin-Garrison creeping foxtail produced 3.6 Mg ha-' of hay and 221 kg ha-' of beef (Seamands et al., 1972). By the third harvest year only 5% of the sainfoingrass mixture was sainfoin whereas 82% of the cicer milkvetch-grass mixture was cicer milkvetch. Alfalfa and sainfoin produced more forage than cicer milkvetch in grazing studies with sheep on a subirrigated pasture in Canada (Smoliak and Hanna, 1975). Although the percentage of forage consumed was similar for the three species, the animals grazed sainfoin first, alfalfa second, and cicer milkvetch last. A comparative grazing trial in New Zealand demonstrated that the preference of sheep for several legume species was as follows: alfalfa > birdsfoot trefoil > cicer milkvetch > falcatus milkvetch > A. glycyphyllus L. (B. J. Wills, personal communication, 1987). Astragulus glycyphyllus was essentially untouched. In an irrigated study in western Canada, pure stands of Kentucky bluegrass, reed canarygrass, orchardgrass, sainfoin, and cicer milkvetch were grazed by beef cattle (Russell et al., 1982). The grasses were fertilized at the standard recommended rate of 224 kg N ha-'. The five species differed only slightly in dry matter production, but orchardgrass produced the

CICER MILKVETCH (Astragalus cicer L.)

287

highest beef gains (628 kg ha-') followed by cicer milkvetch (570 kg ha-'). Sainfoin gave the highest net return followed by cicer milkvetch. Sainfoin, however, persisted poorly. Cicer milkvetch initiated growth later in the spring and ceased growth earlier in the fall compared to the other species. A rest period of 5 to 6 weeks between grazings preserved cicer milkvetch in a mixture with grasses in Canada (Wilson and Rode, 199 1). When grown in mixtures with Regar meadow bromegrass and Manchar smooth bromegrass and under simulated grazing conditions at three locations in Montana, cicer milkvetch yields (1) almost equaled birdsfoot trefoil yields at two locations (95 and 98%) and were less than birdsfoot trefoil yields at one location (80%); (2) were greater than white clover yields at one location (1 15%) and were less than white clover yields at two locations (78 and 87%); and (3) were similar to alfalfa yields (94%) at the only location where alfalfa was evaluated (Cooper et at., 1978). The varied performance of cicer milkvetch at the three locations was attributed to night temperature, with the best performance occurring at locations having the warmest temperature. In a frequency-of-clipping study in Colorado (Townsend, 1974b), the yield of Lutana cicer milkvetch (5.6 Mg ha-') compared favorably with the yield of one cultivar of alfalfa (5.4 Mg ha-'), five cultivars of crownvetch (5.6 Mg ha-'), and seven cultivars of birdsfoot trefoil (5.2 Mg ha-'). Its yield was substantially better than that of two cultivars of white clover (4.3 Mg ha-'). After two harvest years with nine clippings per year to a stubble height of 3.8 or 7.6 cm, the vigor of stands for all species except white clover was poor. A unique feature of cicer milkvetch is that no cases of bloat have been reported in animals grazing its forage. Leaf cell rupture studies with bloatsafe legumes (birdsfoot trefoil, cicer milkvetch, and sainfoin) and bloatinducing legumes (alfalfa, red clover, and white clover) demonstrated that the mesophyll cells of the bloat-safe legumes were more resistant to mechanical rupture compared to those of bloat-causing legumes (Howarth et al., 1978). This work suggested that the stronger mesophyll cells of bloatsafe legumes do not release the intracellular foaming agents as rapidly as do the mesophyll cells of the bloat-inducing legumes; this theory was supported by a subsequent study (Howarth et at., 1982). The disappearance rates of dry matter and nitrogen from green tissue in fistulated sheep were highest for alfalfa, intermediate for birdsfoot trefoil and cicer milkvetch, and lowest for sainfoin. Lees et al. (1981) determined the mechanical strength of whole leaves and isolated mesophyll cells for the same bloat-safe and bloat-causing legume species used by Howarth ef al. (1978). Birdsfoot trefoil had strong cell walls; cicer milkvetch and sainfoin had moderately strong cell walls

288

C . E. TOWNSEND

and high tissue strength; alfalfa and white clover had weak cell walls and low tissue strength; red clover had moderately strong cell walls and low tissue strength. They concluded that cell wall strength alone or in combination with tissue strength influenced cell rupture. The leaflets of cicer milkvetch and birdsfoot trefoil had thicker epidermal layers than those of alfalfa, red clover, and white clover (Lees et al., 1982). The thicker epidermal layer of cicer milkvetch contained epidermal cells smaller than those of the bloat-causing legumes. In addition, cicer milkvetch leaflets had bundle sheath extension cells that served as supporting structures by connecting secondary and tertiary veins to both epidermal layers. Other studies by Lees (1984) supported the previous findings: the bloat-safe legumes had thicker cuticles, epidermal cell walls, and mesophyll cell walls than did the bloat-causing legumes. Cell walls of sainfoin and cicer milkvetch tended to be thicker and disintegrated significantly more slowly during cellulase digestion than did those of alfalfa, red clover, and white clover (Sant and Wilson, 1982). Gas production by rumen microorganisms was greater with whole and chewed leaves from alfalfa, red clover, and white clover than from cicer milkvetch, birdsfoot trefoil, and sainfoin (Fay et al., 1980). Gas production, however, was similar for homogenized leaves of bloat-causing and bloat-safe legumes. Foam production was greater from chewed herbage and homogenized leaves of bloat-causing legumes than that from bloat-safe legumes. These findings are consistent with the theory that rates of disintegration and digestion of legumes by rumen bacteria are important determinants in pasture bloat. Scanning electron microscopy revealed that rumen bacteria invaded and colonized leaflets of bloat-causing alfalfa much faster compared to their action on the leaflets of bloat-safe sainfoin (Fay et al., 1981). Using whole leaflets, Fay et al. (1981) compared three bloat-causing legumes (alfalfa, red clover, and white clover) and three bloat-safe legumes (sainfoin, cicer milkvetch, and birdsfoot trefoil) for rates of leaching and bacterial digestion by measuring the percentage loss of dry matter. The bloat-causing legumes showed a higher dry matter loss by both leaching and bacterial digestion than did the bloat-safe legumes. The component polycross progenies of the cultivar Monarch differed significantly for palatability when grazed by sheep (Townsend, 1986). Several progenies ranked consistently high for palatability and other progenies ranked consistently low. None of the Monarch progenies, however, were significantly more palatable than Lutana. All animals remained healthy and suffered no observable side effects during the 2-year study with 74 consecutive days of grazing each year. Grazing studies with sheep in Minnesota also demonstrated significant and consistent variability among spaced plants of cicer milkvetch for palatability (N. J. Ehlke, personal

CICER MILKVETCH (Astruguh ricer L.)

289

communication, 1989).The latter studies consisted of two grazing cycles in each of 2 years. Thus, the prospects for improving the palatability of cicer milkvetch appear promising. When pure stands of alfalfa, birdsfoot trefoil, cicer milkvetch, and sainfoin were grazed in Minnesota, heifers showed a definite dislike for cicer milkvetch and some animals with white hair developed a severe photosensitization reaction (Marten et al., 1987). There was variability among years for animal response to the photosensitization reaction, but there was no difference between animal breeds. The forage was analyzed for many biological and chemical properties, such as “plant disease organisms (including Pithomyces chartarum), primary phototoxins, and other chemical components, as well as chick bioassays for phototoxins,” but nothing was found that differentiated cicer milkvetch from alfalfa. The blood of animals grazing cicer milkvetch or alfalfa contained phylloerythrin, a normal breakdown product of chlorophyll, but the levels were much higher in those animals grazing cicer milkvetch than in those grazing alfalfa. This suggested that secondary (hepatogenous) photosensitization occurred in animals grazing cicer milkvetch. Forage yield, stand persistence, and nutritive value of the forage for cicer milkvetch were excellent. In a grazing study of pure stands of alfalfa, birdsfoot trefoil, cicer milkvetch, and red clover, lambs developed a photosensitization response while grazing cicer milkvetch (Marten et al., 1990) that was similar to that developed by heifers in the previous Minnesota study. Although cicer milkvetch was the least palatable species and caused a photosensitization response in one-third to one-half of the animals in all breeds, lambs gained as well or better on cicer milkvetch compared to animals grazing the other species. Average lamb production for the 3-year study was 784, 744, 739, and 826 kg ha-’ for alfalfa, birdsfoot trefoil, red clover, and cicer milkvetch, respectively. As with heifers, phylloerythrin was present in the blood of sheep grazing cicer milkvetch or alfalfa, but, as indicated by fluorescence peaks, the concentrations were highest in animals grazing cicer milkvetch. They concluded, as in the cattle study, that secondary (hepatogenous) photosensitization occurred and that the concentration of phylloerythrin in the blood serum and skin influenced the degree of photosensitization.

VI. BREEDING, GENETICS, AND CYTOLOGY

A. OBJECTIVES Relatively little breeding and genetic research has been conducted using cicer milkvetch, but seedling emergence and subsequent stand establishment are two of the most important traits needing improvement. Little is

290

C . E. TOWNSEND

known about pest susceptibility, but cicer milkvetch appears to have resistance to several insect pests that attack alfalfa (Kephart et ul., 1990 Townsend et al., 1990). Improvement of seedling disease resistance under imgation is an important objective in Canada (S. N. Acharya, personal communication, 1992). As the species becomes more widely used, the importance of pest resistance will undoubtedly increase. Forage yield is being improved by selecting for earlier initiation of spring growth, for more rapid recovery after harvest, and for a photoperiodic response in mid- to late summer. Antiquality traits need to be investigated in order to improve palatability of the forage and to eliminate the photosensitivity reaction that some animals develop in some environments when grazing pure stands of cicer milkvetch.

B. CYTOLOGY AND INBREEDINGDEPRESSION Old World species of Astragulus have a basic chromosome number of 8 and about 25% are polyploids (Ledingham and Rever, 1963). In contrast, most New World species have a basic chromosome number of 1 1, 12, 13, or 14 with essentially no polyploidy, and only a few species have a basic chromosome number of 8. Cicer milkvetch, an Old World species, has 64 somatic chromosomes (Ledingham, 1960). The chromosome number, along with the marked inbreeding depression (Townsend, 1972b) and high fertility (Townsend, 197la), suggested that cicer milkvetch is an allooctoploid. Latterell and Townsend (1982) indicated that the species is probably an autoallooctoploidthat has been diploidized by natural selection because there was generally a low frequency of multivalents at metaphase I of meiosis. Although some meiotic irregularities, such as univalents, lagging chromosomes, and dividing univalents, were observed at anaphase I and 11, they were not believed to contribute to genotypic instability, infertility, or phenotypic variation. Chromosome pairing in polyhaploids, derived from twin seedlings, revealed considerable homology among the genomes (R. L. Latterell and C. E. Townsend, unpublished), which suggests that the species may be a segmental alloploid or, if autoalloploid, comprises no more than two divergent genomes. The polyhaploids were both male and female sterile. The process of diploidization is incomplete though evidently well advanced in this species. Townsend ( 1971b) considered cicer milkvetch to be a predominantly self-incompatible species because 43% of a 424-plant population did not set a single seed, and only 2% set in excess of 50 seeds per raceme following self-pollination under field conditions. On the other hand, Scheetz et al. ( 1972) reported that cicer milkvetch has a relatively high self-fertility po-

CICER MILKVETCH (Astrugulus ricer L.)

291

tential. Air temperatures of 27°C day/l6 and 21 "C night, 27°C day/2l0C night, and 32°C day/21"C night had no consistent effect on the selfcompatibility reaction (Townsend, 197 1b). One generation of inbreeding reduced the forage yield of cicer milkvetch by about 50% (Townsend, 1972b). Inbreeding had such a deleterious effect on plant vigor that paired S, and open-pollination progenies could be differentiated visually. Considerable variation existed among So plants for inbreeding depression. Forage yields of the S progenies ranged from 2 1 to 80%, 23 to 899'0, and 29 to 101%of their open-pollinationcounterparts for each of three harvests, respectively. The range for spread (34 to 94%) among S, progenies at the end of the second year was similar to that for yield. Mean plant heights of the SI progenies were 78, 81, and 8290 of the open-pollination progenies for each of the three harvests, respectively. Within each harvest the height of the S, progenies ranged from 59 to 107%, 62 to 1 lo%, and 63 to 100% of the open-pollination progenies. Seedling mortality in the greenhouse was greater for S , progenies than for openpollination progenies. Consequently, the reduction in vigor on inbreeding probably was greater than indicated. Scheetz et al. (1972) noted that one generation of inbreeding reduced seedling growth of cicer milkvetch 32% in greenhouse studies.

,

C. BREEDING METHODOLOGY Initially, the germplasm base available for the improvement of cicer milkvetch was exceptionally narrow (Townsend, 1970). Although significant differences existed among 20 accessions for phenotypic diversity for such traits as vigor, height, date of flowering, and spread, 11 of the 20 accessions traced to a single introduction (PI 66515) made in 1926. Additional introductions were evaluated for phenotypic diversity, but some were believed to have a common origin because they were obtained from botanical gardens in Europe and were similar phenotypically (Townsend and Ackerman, 1975). Smoliak and Johnston (1976) reported enough diversity for agronomic traits such as forage yield, seed germination percentage, speed of germination, and seedling vigor within the cultivar Oxley to make additional improvement via a breeding program. Simple recurrent selection has been a promising method for breeding cross-pollinated, self-incompatible species such as cicer milkvetch. For recurrent selection to be effective, a large portion of the genetic variance must be additive. A diallel cross analysis demonstrated significant variability for vigor score, height, date of flowering, spread, and seed weight (Townsend, 1975). When the variance for these traits was partitioned into

292

C. E. TOWNSEND

general combining ability (GCA), specific combining ability (SCA), and reciprocal effects, most of the genetic variance for all characters except vigor in the seedling year was caused by the additive component (GCA). Therefore, recurrent selection was believed to be an effective method for improving cicer milkvetch.

D. SEEDLING VIGOR Genetic variability for seedling emergence was demonstrated when polycross progenies differed significantly for seedling emergence under field conditions on three dates of counting (Townsend, 1974a) (Fig. 5). Means for seedling emergence at the first, second, and third counts ranged from 3 to 27, 9 to 30, and 12 to 33 seedlings rn-' of row, respectively. Progenies with the highest seedling emergence at the first count were also the highest at the other two counts. Therefore, only the early seedling count would be required. Recurrent selection for rapid seedling growth at 20°C day/ 15"C night in environmental growth chambers was more effective for improving subsequent seedling emergence in the field than selecting at 25/20"C (Townsend, 1979b). Of the progenies selected at 20/15"C and 25/20"C, 44 and 15%, respectively, were significantly higher than the control for seedling emergence in the field. Similar results were obtained with two intercross populations selected at the same two temperature regimes, i.e., 53% at 20/15"C and 17%at 25/2OoC.Also, 53%of the progenies from a high-seedweight population, selected for rapid seedling growth in a 20/15"C environment, were significantly better than the control for seedling emergence in the field. Seedling emergence was poor for other populations not selected for rapid seedling growth at 20/15OC, i.e., emergence from a 3.8cm-deep planting in the greenhouse, early initiation of spring growth and recovery afier harvest, and combinations of seed weight and mature plant vigor. Selection of seedlings for rapid growth in a 20/15"C environment was an effective method for improving seedling emergence of cicer milkvetch under field conditions. One cycle of simple recurrent selection and one cycle of modified simple recurrent selection were conducted within the component polycross progenies of the cultivar Monarch for improved seed germination and seedling elongation at 5/2OoC and 12/20°C temperature regimes in the laboratory and for seedling emergence in the field (Townsend, 1985). Seedling elongation of the cultivar Monarch was curvilinear at both 12/20"C and 5/2OoCtemperature treatments (Fig. 6), whereas seedling elongation of the polycross progenies was also curvilinear at 12/20°C but linear at 5/20"C.

CICER MILKVETCH (Artragah cicer L.) 40

13rd COUNT

CONTROL

I

JUNE 14

--

8

u

30

29 3 i

I

- 2 n d COUNT MAY 30

z

GRAND MEAN

0

a

-1

30

I

1st COUNT MAY I2

,GRAND

MEAN

I

SEEDLINGS PER METER Figure 5. Seedling emergence on three dates for 88 polycross progenies of cicer milkvetch seeded on 10 April, 1972, at Fort Collins, Colorado. The control entry was a large-seeded lot (4.53 g/lOOO seeds) that gave excellent seedling emergence under both greenhouse and field conditions. (From Townsend, 1974a.)

C . E. TOWNSEND

294 501-

=-22.33+10.54X-0.39X2 2

R = 40

-

E E v

I c 30 -

w z W

I

7

-/

'\$

=-15.39+6.77X-0.19Xz

R 2 = 0.997

10' 4

I

I

I

5

6

7

I

8

I

I

I

II

12

I

9 10 DAY

1

13

I

I

14

15

Figure 6. Seedling growth of Monarch cicer milkvetch through day 14 of germination at 5/2O"C(x) and 12/2OoC(0).(From Townsend, 1985.)

Mean seedling length of the selected progenies (48 mm) on day 12 at 5/20"C was about the same as that for Monarch on day 12 at 12/20"C. Seedling emergence in the field was the primary criterion for selecting parents for the next cycle. Following one cycle of selection at 5/2OoC,32, 37, and 26% of the progenies were significantly better than Monarch for seed germination, seedling elongation, and seedling emergence, respectively (Townsend, 1985). Following one cycle of selection at 12/20"C, 13, 19, and 21% of the progenies were significantly better than Monarch for seed germination, seedling elongation, and seedling emergence, respectively. Although selection for early seedling vigor under relatively cool conditions in a controlled environment improved seedling emergence under field conditions, this was not a practical procedure for improving early seedling vigor because of the time required to make the measurements. However, the procedure may have merit, if selection is visual. None of the seed or seedling traits evaluated was consistently correlated with seedling emergence. Therefore, seedling emergence in the field is the only reliable procedure to evaluate early seedling vigor of cicer milkvetch. Additional information on the inheritance of the seed germination response to temperature was provided by Townsend (1990b). Seeds of 80 polycross progenies from four populations were evaluated for rate of germination and total germination at 15/25"C and 5/2OoC.At 15/25"C, the

CICER MILKVETCH (Astrtzgulus cicer L.)

29 5

values for percentage of seed germination on a daily basis through day 14 after planting and total germination were similar for the four breeding populations. At 5/20 C, however, large differences occurred among populations and among progenies within populations. Consequently, genetic differences were expressed at 5/2OoCbut not at 15/25"C. The differences among populations can be explained by their parentage. For the two populations with the slowest rate of germination and the lowest total germination, clone F-1 was the male or female parent for about 45% of the parental plants. For the other two populations in which germination was affected less, clone F- 1 was the male or female parent for only about 10%of the parental plants. Little attention has been given in forage breeding programs to factors associated with differential germination of seeds and emergence of seedlings, and their possible effects on the genetic composition of synthetic cultivars. Linear regression was a simple and effective method of demonstrating the relative stability and adaptability of polycross progenies for seedling emergence in diverse environments, including four sites in Colorado and one each in Montana, Oklahoma, New Mexico, and Alberta, Canada (Townsend et al., 1979). Although there were large differences among environments for seedling emergence, several progenies ranked high in all environments. Such progenies were average or above average for stability and could be considered well adapted to most environments. The progeny with the highest average emergence had below average stability because it was poorly adapted to the most unfavorable environment. As expected, soil moisture conditions at planting time and subsequent precipitation greatly influenced seedling emergence. O

E. FORAGEYIELD Populations of cicer milkvetch differed significantlyin combining ability for seedling dry weight in the greenhouse and forage yield in the field (Townsend, 1976). Both greenhouse and field studies showed that the frequency of high-yielding progenies was higher in high-vigor and largeseeded populations than in a low-vigor population. The lower hay yield of cicer milkvetch when compared to alfalfa can be attributed, in part, to its slower initiation of spring growth and slower recovery after harvest. In an effort to improve the forage yield of cicer milkvetch, Townsend (198 la) conducted one cycle of recurrent selection for both early initiation of spring growth and more rapid recovery after harvest and just for more rapid recovery after harvest. The resulting 57 polycross progenies differed significantly for forage yield in each of 3 years

296

C . E. TOWNSEND

(Fig. 7). Only 59 and 85% of the polycross progenies selected solely for more rapid recovery after harvest yielded significantly more forage than Lutana in 1977 and 1978, respectively. In contrast, 71 and 100% of the progenies selected for both early initiation of spring growth and recovery after harvest yielded significantly more forage than Lutana in 1977 and 1978, respectively. Consequently, selecting simultaneously for both traits appears to be more effective for improving forage yield than selecting only for more rapid recovery after harvest. In 1979 the relative yield of Lutana was substantially higher than in 1977 and 1978 and progeny yields ranged from 8 1 to 1 17%of that of Lutana. In addition, the relative yield of Oxley was equal to that of Lutana. Only four progenies yielded significantly more forage than Lutana and two were from each group. The markedly improved performance of Lutana for forage yield in 1979 relative to the polycross progenies can be attributed, in part, to plant spread because Lutana spread more than most progenies. Cicer milkvetch differed in growth response to the decreasing daylengths of mid- to late summer (third growth period) (Townsend, 1988). Plant height on 15 September following a 1 August harvest for the component polycross progenies of the cultivar Monarch ranged from 15 to 55 cm. Two cycles of recurrent selection increased plant height and plant weight in each of three growth periods, with the greatest increase generally occurring in the third period (unpublished observations, 1990). Plant height of the selected populations in the third growth period ranged from 105 to 123% of that of Monarch. Plant weight of the selected populations ranged from 104 to 120% and from 105 to 120% of that of Monarch for the third growth period and for total yield, respectively. The photoperiod-insensitive trait has the potential for improving the forage yield of cicer milkvetch so that it is similar to that of alfalfa.

F. SEEDWEIGHT Recurrent selection was an effective method for increasing the seed weight of cicer milkvetch (Townsend, 1977a) (Fig. 8). Seed weight of the original population ranged from 2.70 to 4.80 g/lOOO seeds. In comparison, samples of the cultivars Lutana and Oxley weighed 3.64 and 3.52 g/lOOO seeds, respectively. In the first cycle of selection, seed weight on an individual plant basis ranged from 2.97 to 5.40 g/lOOO seeds with a mean of 3.97. This represented an increase in seed weight of about 13% over the original population. To help prevent the loss of plant vigor, selection was for high vigor as well as for high seed weight. In selecting parents of the second cycle, selection pressure in one subpopulation was intense for mature-plant

CICER MILKVETCH (Astrugulus ricer L.)

LL

0

s I20 I10

I00

3

-1

LI 0

$

ui

297

130 I 20 I10

100

[ ' 2 1 z 110

1979

a

F

3 loo

u.

0

8

90 80 --

POLYCROSS PROGENIES Figure 7. Distribution of 57 polycross progenies of cicer milkvetch for forage yield over 3 years, expressed as a percentage of the cultivar Lutana. Of the 57 progenies, 17 were selected for both early initiation of spring growth and recovery after clipping; the other 40 progenies were selected solely for more rapid recovery after clipping. (From Townsend, 198 la.)

C. E. TOWNSEND

298 I

,

,

I

,

,

,

,

,

I

,

,

O R I G I NAL

MEAN [PARENTS OF FIRST CYCLE] I

25

n

n

r

2 0 - SECOND CYCLE

POPULATION TWO

~

~

i

,

~

,

,

i

,

_

~

~

-

1510-

50

'

1

'

Figure 8. A comparison of the frequency distributions for seed weight in the original population and after one and two cycles of recurrent selection in polycross progenies of cicer milkvetch. (From Townsend, 1977a.)

vigor but was relaxed somewhat for seed weight. In the other subpopulation, selection pressures were reversed. Mean seed weights of subpopulations 1 and 2 were 4.47 and 4.73 g/lOOO seeds, respectively. The increase in seed weight was 12.6 and 19.1%over cycle 1 and 27.4 and 34.8Yo over the original population for subpopulations 1 and 2, respectively. Heritability estimates for seed weight were 75 and 79% for subpopulations 1 and 2,

~

i

CICER MILKVETCH (Astrugdw ricer L.)

299

respectively. Enough genetic variability appeared to remain after two cycles of selection to make additional progress. The greatest disadvantage of recurrent selection was that each cycle of selection required 3 years.

G. DATEOF FLOWERING Cicer milkvetch was evaluated in a 10-parent diallel cross for flowering date in the seedling year and in two growth periods of the second year (Townsend, 1984). Significant differences occurred among progenies for flowering in each growth period. Variance for these traits was partitioned into GCA, SCA, and reciprocal effects, and each component was significant for all growth periods. When plants from the late X early flowering crosses were classified as either flowering or nonflowering in the seedling year, flowering (vernalization not required) was partially dominant over nonflowering (vernalization required). In the seedling year, 73% of all plants flowered and the mean flowering date among progenies ranged from 12 August to 25 August (Townsend, 1984). Considerable variability existed within all progenies for date of flowering, with the earliest plant flowering on 20 July. In the second year, 98% of all plants flowered in each of the two growth periods. In the first growth period there was little variability for date of flowering because progeny means ranged only from 16 June to 18 June. There was also relatively little variability within progenies because the earliest plant flowered on 12 June. The range among progeny means was 6 days (1 to 7 August) for date of flowering during the second growth period of the second year, with the earliest plant flowering on 19 July. Because of the large amount of additive genetic variance, recurrent selection should be an effective method for developing earlier flowering types in the seedling year and in the second growth period of the second year. The relatively small amount of variability in the first growth period of the second year indicates that other germplasm will be needed to develop earlier flowering types for that growth period.

Four cultivars of cicer milkvetch have been released. Lutana was released in Montana and Wyoming (Stroh et al., 1971). It traces to 127 plants from PI 66515, introduced from Sweden in 1926. The parental plants were selected for earliness of spring growth, rapid recovery after cutting, rapid rhizome spread, and uniformity of seed maturation. Oxley

3 00

C. E. TOWNSEND

was released in Canada (Johnston et al., 1971, 1972). It traces to OA-408, which was introduced from the former USSR in 1931 and was grown for many years at the Canada Agriculture Research Substation, Manyberries, Alberta. Monarch was released in Colorado (Townsend, 1980~).It is a 40-clone synthetic selected primarily for improved seedling emergence; however, the parental clones also had excellent mature plant vigor. In Colorado, Monarch has better seedling emergence than either Lutana or Oxley. Forage yields of Monarch are equal to or significantly higher than those of Lutana. Oxley does not yield as well as either Monarch or Lutana in Colorado. Seedling emergence of the component polycross progenies of Monarch ranged from 125 to 200% of that of Lutana. After two growing seasons, average plant spread from rhizomes for the component polycross progenies of Monarch under spaced conditions ranged from 92 to 133%of that of Lutana. Oxley spreads by rhizomes much faster than Lutana or Monarch. Evans and Abernethy (1983) developed a sodium dodecyl sulfate polyacrylamide gel electrophoretic method of using seeds to identify the cultivars Lutana, Oxley, and Monarch. Windsor, a 15-clone synthetic, was released in Colorado and Wyoming in 1992 (Townsend, 1993). It traces to Monarch and was selected for improved forage yield that is due to a photoperiodic response that permits increased plant growth in mid- to late summer. Seedling emergence of Windsor is equal to that of Monarch. In addition, 17 germplasms have been developed and released (Townsend, 1979a,c, 1987, 1990a; Townsend and Hinze, 1979; Townsend and Ditterline, 1993).

VII. SUMMARY AND CONCLUSIONS Cicer milkvetch, a perennial nonbloating forage legume, is well adapted to many areas in the United States, Canada, and central and eastern Europe, including European Russia and the Caucasus. Progress has been made in the domestication of cicer milkvetch, and its use as a cultivated species should increase. It has great potential because of its strong perenniality, winter hardiness, apparent resistance to some insects that attack other species of forage legumes, nonbloating traits, relatively high forage yield, N,-fixing ability, and compatibility with cool-season perennial grasses. It is an ideal species for long-term pastures, meadows, and disturbed lands. Cicer milkvetch has relatively poor seedling vigor, but excellent stands can be obtained if properly scarified seeds are planted in a firm seedbed. When inoculated with the appropriate rhizobia bacteria, substantial

CICER MILKVETCH (Artruguhr cicer L.)

301

amounts of N, are fixed, as demonstrated by subsequent forage yields of the legume and companion grasses. Quality of its forage is excellent when measured by standard laboratory procedures. Cicer milkvetch, however, is not as palatable to livestock as the other more commonly used forage legumes. Also, in some environments animals grazing pure stands of cicer milkvetch may develop a photosensitivity reaction. The antiquality traits related to palatability and photosensitization and their environmental interactions must be identified and corrected through breeding or management. Although the appropriate practices for managing cicer milkvetch for hay or pasture have not been thoroughly investigated, it is evident that the practices used for alfalfa are not satisfactory. Because cicer milkvetch is slower to establish than most cultivated forage legumes, it needs one full growing season before cutting or grazing. The few instances where cicer milkvetch has not persisted in an otherwise favorable environment can probably be attributed to inappropriate management practices. About 50% of the total yield is produced in the first growth period of the season. Therefore, management systems that can utilize this growth characteristic need additional investigation. Relatively little effort has been devoted to the improvement of cicer milkvetch through breeding and genetics. This effort needs to be increased. Seedling vigor and forage yield have been improved somewhat, but additional improvements are needed. Little is known about its susceptibility to disease and insect pests. The germplasm base must be expanded. The agronomic potential of cicer milkvetch is excellent in a forage production system with good management.

REFERENCES Abernethy, R. H. 1987. Response of cicer milkvetch seed to osmoconditioning.Crop Sci. 27, 117- 121. Alaniya, M . D., Aneli, D. N., Patudin, A. V., and Komelin, R. V. 1984. Flavonoid glycosides of Astragafus cicer. Chem. Nat. Compd. 19, 500- 50 1. Allinson, D. W., Speer, G. S., Taylor, R. W., and Guillard, K. 1985. Nutritional characteristics of kura clover (Trifolium ambiguum Bieb.) compared with other forage legumes. J. Agric. Sci. 104, 227-229. Atkins, M. D. 1953. New legumes for the Northern Plains. Crops Soils 5(9), 16- 18. Berg, W.A. 1990. Herbage production and nitrogen accumulation by alfalfa and cicer milkvetch in the Southern Plains. Agron. J. 82,224-229. Berg, W.A., and Barrau, E. M. 1978. Management approaches to nitrogen deficiency in revegetation of subalpine disturbances. In$ Ser. -Colo. Water Resour. Res. Znst. 28, 174- 18 I . Bleak, A. T. 1969. Growth and yield of legumes in mixtures with grasses on a mountain range. J. RangeManage. 21,259-261.

302

C . E. TOWNSEND

Bowman, R. A., and Townsend, C. E. 1990. Growth response of cicer milkvetch to different soil reactions. Agron. Abstr. p. 263. Burton, J. C. 1972. Nodulation and nitrogen fixation in alfalfa. Agronomy 15,229-246. Buyanova, V. F. 198 1. Biology of some Astragalus spp. in the flora of the USSR, cultivated in Leningrad Russian-SFSR1. Rastit. Resur. 17, 420-433 [Biol. Abstr. 74, 059044 (1982)l. Carleton, A. E., Austin, R. D., Stroh, J. R., Wiesner, L. E., and Scheetz, J. G. 1971. Cicer milkvetch (Astragalus cicer L.) seed germination, scarification and field emergence studies. Bull.--Mont., Agric. Exp. Stn. 655. Church, D. C. 197 1 . The macro (major) minerals. Dig. Physiol. Nutr. Ruminants 2,4 13 -45 I . Cook, C. W., Hyde, R. M., and Sims, P. L. 1974. Revegetation guidelines for surface mined areas. Range Sci. Dep. Sci. Ser. (Colo.State Univ.)16. Cooper, C. S. 1979. Evaluation of legume mixtures for hay planting. Agron J. 71,81-83. Cooper, C. S., and Watson, C. A. 1968. Total available carbohydrates in roots of sainfoin (Onobrychis viciaefolia Scop.) and alfalfa (Medicago sativa L.) when grown under several management regimes. Crop Sci. 8,83 - 85. Cooper, C. S.,Eslick, R. F., and McDonald, P. W. 1966. Foam formation from extracts of 27 legume species in vitro. Crop Sci. 6, 2 I5 -2 16. Cooper, C. S., Welty, L. E., Laudert, S. B., and Wiesner, L. E. 1978. Evaluation of Regar meadow bromegrass in Montana. Bull. -Mont. Agric. Exp. Stn. 702. Cornelius, D. R., and Talbot, M. W. 1955. Rangeland improvement through seeding and weed control of east slope Sierra Nevada and on southern Cascade Mountains. US., Dep. Agric., Agric. Handb. 88. Davis, A, M. 1972. Selenium accumulation in Astragalus species. Agron. J. 64, 751 -754. Davis, A. M. 1973. Protein, crude fiber, tannin, and oxalate concentrations of some introduced Astragalus species. Agron. J. 65,6 I3 -6 15. Davis, A. M. 1982a. Nitrogen production by selected Astragalus species. Agron. J. 74, 454-456. Davis, A. M. 1982b. Crude protein, crude fiber, tannin, and oxalate concentrations of 33 Astragalus species. J. Range Manage. 35,32 -34. Davis, M. R. 1981. Growth and nutrition of legumes on a high country yellow-brown earth subsoil. 1. Phosphate response of Lotus, Trifolium, Lupinus, Astragalus, and Coronilla species and cultivars. N.Z. J. Agric. Res. 24, 321 -332. Dobson, J. W., Fisher, C. D., and Beaty, E. R. 1976. Yield and persistence of several legumes growing in tall fescue. Agron. J. 68, 123- 125. Ellsbury, M. M., and Nielson, M. W. 198 1. Comparative host plant range studies of the blue alfalfa aphid, Acyrthosiphon kondoi Shinji, and the pea aphid, Acyrthosiphon pisum (Harris) (Homoptera: Aphididae). US.,Dep. Agric., Sci. Educ. Adm., Tech. Bull. 1639. Etra, J., Cuany, R. L., and Thor, G. L. 1984. Persistence and performance of grass and legume strains at subalpine test plots in Colorado. Inf: Ser. -Colo. Water Resour. Res. Znst. 53, 59-77. Evans, R. K., and Abernethy, R. H. 1983. Identification of cicer milkvetch cultivars using SDS-polyacrylamide gel electrophoresis. Can. J. Plant Sci. 63, 1087- 1090. Faris, M. A., and Ta, T. C. 1985. Performance of double and single cut red clover and cicer milkvetch entries at the Ottawa Research Station. Forage Notes 29, 15 - 18. Fay, J. P., Cheng, K.-J, Hanna, M. R., Howarth, R. E., and Costerton, J. W. 1980. Zn vitro digestion of bloat-safe and bloatcausing legumes by rumen microorganisms: Gas and foam production. J. Dairy Sci. 63, 1273- 128I . Fay, J. P., Cheng, K.-J, Hanna, M. R., Howarth, R. E., and Costerton, J. W. 198 1. A scanning

CICER MILKVETCH (Astrugahs cicer L.)

303

electron microscope study of the invasion of leaves of a bloat-safe and bloat-causing legume by rumen microorganisms. Can. J. Microbiol. 27, 390-399. Gabrielsen, B. C., Smith, D. H., and Townsend, C. E. 1985. Cicer milkvetch and alfalfa as influenced by two cutting schedules. Agron. J. 77,416-422. George, J. R. 1984. Grass sward improvement by frost-seeding with legumes. Proc. Forage Grassl. Conj pp. 265-269. Hafenrichter, A. L., Foster, R. B., and Schwendiman, J. L. 1965. Effect of storage at four locations in the west on longevity of forage seeds. Agron. J. 57, 143- 147. Hafenrichter, A. L., Schwendiman, J. L., Harris, H. L., MacLauchlan, R. S., and Miller, H. W. 1968. Grasses and legumes for soil conservation in the Pacific Northwest and Great Basin States. U.S.,Dep. Agric.. Agric. Handb. 339. Haraszti, E., and Vetter, J. 1989. Phytochemical data on the Astragalus species of Hungary. Bot. Kozl. 74-75, 379-390 [Biol. Abstr. 90,080028 (1990)). Hart, R. H., and Dean, J. G. 1986. Forage establishment: Weather effects on stubble vs. fallow and fall vs. spring seeding. J. RangeManage. 39,228-230. Hewitt, G. B. 1969. Twenty-six varieties of forage crops evaluated for resistance to feeding by Melanoplus sanguinipes. Ann. Entomol. Soc. Am. 62,731 -74 I . Hewitt, G. B., Wilton, A. C., and Lorenz, R. J. 1982. The suitability of legumes for rangeland interseeding and as grasshopper food plants. J. Range Manage. 35,653-656. Holechek, J. L., Depuit, E. J., Coenenberg, J. G., and Valdez, R. 1981. Fertilizer effects on establishment of two seed mixtures on mined land in southeastern Montana. J. Soil Water Conserv. 36,24 1 - 244. Holechek, J. L., Depuit, E. J., Coenenberg, J. G., and Valdez, R. 1982. Legume establishment on strip mined lands in southeastern Montana. J. Range Manage. 35,298 - 300. Howarth, R. E., Goplen, B. P., Fesser, A. C., and Brandt, S. A. 1978. A possible role for leaf cell rupture in legume pasture bloat. Crop Sci. 18, 129- 133. Howarth, R. E., Goplen, B. P.,Brandt, S. A., and Cheng, K.-J. 1982. Disruption of leaf tissues by mmen microorganisms: An approach to breeding bloat-safe forage legumes. Crop. Sci. 22, 564-568. Ingham, J. L., and Dewick, P. M. 1980. Astraciceran: A new isoflavan phytoalexin from Astragalus cicer. Phytochemistry 19, 1767 - 1770. Johnston, A., Smoliak, S., Hironaka, R., and Hanna, M. R. 1971. Oxley cicer milkvetch. Can. J. Plant Sci. 51,428-429. Johnston, A., Smoliak, S., Hanna, M. R., and Hironaka, R. 1972. Registration of Oxley cicer milkvetch. CropSci. 12,715. Johnston, A., Smoliak, S., Hanna, M.R., and Hironaka, R. 1975. Cicer milkvetch for western Canada. Agric. Can. Publ. 1536. Kaldy, M. S., Smoliak, S., and Hanna, M. R. 1978. Amino acid composition of cicer milkvetch forage. Agron. J. 70, I3 I - 132. Kenno, H., Brick, M. A., and Townsend, C. E. 1987. Establishment of cicer milkvetch with four cool-season grasses. Crop Sci. 27,8 10- 8 12. Kephart, K. D., Higley, L. G., Buxton, D. R., and Pedigo, L. P. 1990. Cicer milkvetch forage yield, quality, and acceptability to insects. Agron. J. 82,477-483. Kerr, H. D., and Klingman, D. L. 1960. Weed control in establishing birdsfoot trefoil. Weeds 8, 157- 167. Kinder, S . D., and Springer,T. L. 1989. Alternate hosts of Russian wheat aphid (Homoptera: Aphididae). J. Econ. Entomol. 82, 1358- 1362. Kneebone, W. R. 1959. An evaluation of legumes for western Oklahoma rangelands. Okla., Agric. Exp. Stn./U.S. Dep. Agric. Bull. B-539.

3 04

C. E. TOWNSEND

Komarov, V. L. (ed.). 1965. “Flora of the USSR,” Vol. XII. Leguminosae: Astragalus (translated by the Israel Program for Translation Ltd. and published for the Smithsonian Institution and the National Science Foundation, Washington, D.C.). Latterell, R. L., and Townsend, C. E. 1982. Cytology and breeding behavior of cicer milkvetch. Proc. Int. Grassl. Congr., 14th pp. 174- 176. Ledingham, G. F. 1960. Chromosome numbers in Astragalus and Oxytropis. Can. J. Genet. Cytol. 2, 119- 128. Ledingham, G. F., and Rever, B. M. 1963. Chromosome numbers of some southwest Asian species of Astragalus and Oxytropis (Leguminosae). Can. J. Genet. Cytol. 5, 18 - 32. Lees, G. L. 1984. Cuticle and cell wall thickness relation to mechanical strength of whole leaves and isolated cells from some forage legumes. Crop Sci. 24, 1077- 1081. Lees, G. L., Howarth, R. E., Goplen, B. P., and Fesser, A. C. I98 1. Mechanical disruption of leaf tissues and cells in some bloat-causing and bloat-safe forage legumes. Crop Sci. 21, 444-448. Lees, G. L., Howarth, R. E., and Goplen, B. P. 1982. Morphological characteristics of leaves from some legume forages: Relation to digestion and mechanical strength. Can. J. Bot. 60,2126-2132. Major, D. J., Hanna, M. R., Smoliak, S., and Grant, R. 1979. Estimating nodule activity of sainfoin, alfalfa, and cicer milkvetch seedlings. Agron. J. 71,983-985. Malik, N., and Waddington, J. 1989. Weed control strategies for forage legumes. Weed Technol. 3,288-296. Malik, N., and Waddington, J. 1990. No-till pasture renovation after sward suppression by herbicides. Can. J. Plant Sci. 70,261 -267. Marten, G. C., Ehle, F. R., and Ristau, E. A. 1987. Performance and photosensitization of cattle related to forage quality of four legumes. Crop Sci. 27, 138- 145. Marten, G. C., Jordan, R. M., and Ristau, E. A. 1990. Performance and adverse response of sheep during grazing of four legumes. Crop Sci. 30,860-866. Martin, S. S., and Townsend, C. E. 1989. Isoflavan stress metabolites in North American cultivars of Astragalus cicer. Am. J. Bot., Suppl. 76(6), 194. McGinnies, W. J., and Crofts, K. A. 1986. Effects of N and P fertilizer placement on establishment of seeded species on redistributed mine topsoil. J. Range Manage. 39, 118-121. McGraw, R. L., and Marten, G. C. 1986. Analysis of primary spring growth of four pasture legume species. Agron. J. 78,704-7 10. Melton, B. 1973. Evaluations of sainfoin and cicer milkvetch in New Mexico. Res. Rep. -N. M.. Agric. Exp. Stn. 255. Miklas, P. N., Townsend, C. E., and Ladd, S. H. 1987. Seed-coat anatomy and scarification of cicer milkvetch seed. Crop. Sci. 27, 766-772. Miroslavov, E. A., and Bubolo, L. S. 1980. Ultrastructure of leaf chlorenchyma cells of some arctic plants. Herb. Abstr. 51,6078. Moyer, J. R. 1989. Weed control during cicer milkvetch establishment and yields in subsequent years. Can. J. Plant Sci. 69,2 I3 - 222. Moyer, J. R., Smoliak, S.,and Johnston, A. 1979. Tolerance of seedling cicer milkvetch to herbicides. Can. J. Plant Sci. 59, 1 155- 1 157. Nicholas, P. J., and McGinnies, W. J. 1982. An evaluation of 17 grasses and 2 legumes for revegetation of soil and spoil on a coal strip mine. J. Range Manage. 35,288-293. Nichols, J. T., Gray, A. M., Clanton, D. C., and Myran, D. 1982. Alfalfa, cicer milkvetch, cool-season grasses. Imgated pasture-animal production. Nebr., Agric. Exp. Stn., Beef Cattle Rep. pp. 3 1 - 32.

CICER MILKVETCH (Astraguh cicer L.)

305

Peterson, P. R., Sheaffer, C. C., and Hall, M. H. (1992). Drought effects on perennial forage legume yield and quality. Agron. J. 84,774-779. Plummer, A. P., Christensen, D. R., and Monsen, S. B. 1968. Restoring big game range in Utah. Utah Division of Fish and Game Pub. 68-3. Podlech, D. 1986. Taxonomic and phytogeographical problems in Astragalus of the Old World and south-west Asia. Proc.-R. Soc. Edinburgh, Sect. B: Biol. Sci. 89, 37-43.

Richards, K. W., 1986. Pollination requirements of cicer milkvetch, Astragalus cicer L. J. Range Manage. 39,457-459. Rumbaugh, M. D., Johnson, D. A., and Van Epps, G. A. 1982. Forage yield and quality in a Great Basin shrub, grass, and legume pasture experiment. J. Range Manage. 35,604609.

Rumburg, C. B. 1978. Grazing of grass and cicer milkvetch-grass pastures with yearling calves. Agron. J. 70,850-852. Russell, K. D., Hironaka, R., and Wilson, D. B. 1982. An economic analysis of alternative management techniques for beef production on imgated pastures in Alberta. Can. Farm. Econ. 16, 1-7. Russelle, M. P., and Buzicky, G. C. 1988. Legume response to fresh dairy cow excreta. Proc. Forage G r a d Conf pp. 166- 170. Sant, F. I., and Wilson, D. 1982. Use of a cellulolytic enzyme digestion technique to distinguish bloat-causing from non-bloat-causing legumes and to select for speed of mesophyll cell wall disintegration in red clover (Trifolium pratense L.). J. Agric. Sci. 98, 99- 102.

Sarkar, S. K., Howarth, R. E., and Goplen, B. P. 1976. Condensed tannins in herbaceous legumes. Crop. Sci. 16,543-546. Scheetz, J. G., and Stroh, J. R. 1982. Compatibility of ‘Lutana’ cicer milkvetch in alternate row mixtures with 12 grass species.Agron. Abstr. p. 128. Scheetz, 1. G., Carleton, A. E.,and Stroh, J. R. 1972. Fertility and inbreeding studies of cicer milkvetch (Astragalus cicer L.). Crop Sci. 12,758- 759. Schultz, R. D., and Stubbendieck, J. 1982. Herbage yield of fertilized cool-season grasslegume mixtures in western Nebraska. J. Range Manage. 35,473-476. Schultz, R. D., and Stubbendieck, J. 1983. Herbage quality of fertilized cool-season grasslegume mixtures in western Nebraska. J. Range Manage. 36,57 1-575. Seamands, W. J., Lang, R. L., Corbridge, M., and Boilsen, E. A. 1972. Lutana cicer milkvetch-Eski sainfoin imgated pasture studies. Univ. Wyoming, Agric. Exp. Stn. Res. J. 56. Sheaffer, C. C., and Marten, G. C. 1991. Kura clover forage yield, forage quality, and stand dynamics. Can. J. Plant Sci. 71, 1 169 - 1 172. Smith, D. 1962. Carbohydrate root reserves in alfalfa, red clover, and birdsfoot trefoil under several management schedules. Crop Sci. 2,75-78. Smoliak, S., and Hanna, M. R. 1975. Productivity of alfalfa, sainfoin, and cicer milkvetch on subimgated land when grazed by sheep. Can. J. Plant Sci. 55,4 15 -420. Smoliak, S., and Hanna, M. R. 1977. Seedling competition of some forage legumes in monoand mixed culture under greenhouse conditions. Can. J. Plant Sci. 57,897-903. Smoliak, S., and Johnston, A. 1976. Variability in forage and seed production and seedling growth in Astragalus cicer. Can. J. Plant Sci. 56,487-491. Smoliak, S., Johnston, A., and Hanna, M. R. 1972. Germination and seedling growth of alfalfa, sainfoin and cicer milkvetch. Can. J. PZant Sci. 52,757-762. Solum, D. J., and Lockerman, R. H. 1991. Seed coat surface patterns and structures of

306

C . E. TOWNSEND

Oxytropis riparia, Oxytropis campestris, Medicago sativa, and Astragalus cicer. Scanning Microsc. 5,779-786. Stroh, J. R., Carleton, A. E., and Thornburg, A. A. 1971. Registration of Lutana cicer milkvetch. Crop Sci. 11, 133. Stroh, J. R., Carleton, A. E., and Seamands, W. J. 1972. Management of Lutana cicer milkvetch for hay, pasture, seed, and conservation uses. Bull. -Mont., Agric. Exp. Sin. 666. Thornburg, A. A. 1982. Plant materials for use on surface-mined lands in arid and semiarid regions. U.S. Dep. Agric. SCSTP-157 and EPA-600/7-79-134. Townsend, C. E. 1970. Phenotypic diversity for agronomic characters in Astragalus cicer L. Crop Sci. 10,69 1 -692. Townsend, C. E. 1971a. Association among characters related to seed production in Astragalus cicer L. Crop Sci. 11,307-308. Townsend, C. E. 1971b. Self-compatibility studies with Astragalus cicer. L. Crop Sci. 11, 769-770. Townsend, C. E. 1972a. Influence of seed size and depth of planting on seedlingemergence of two milkvetch species. Agron. J. 64,627-630. Townsend, C. E. 1972b.Comparison of S,and open-pollination progenies ofAstragalus cicer L. for certain agronomic characters. Crop Sci. 12,793-795. Townsend, C . E. 1974a. Selection for seedling vigor in Astragalus cicer L. Agron. J. 66, 241 -245. Townsend, C. E. I974b. Productivity of several perennial forage legumes under imgation and frequent cutting. Bull. -Cola., Agric. Exp. Stn. 5624. Townsend, C. E. 1975. General and specific combining ability for several agronomic traits in diallel cross progenies of cicer milkvetch. Crop Sci. 15,341 -343. Townsend, C. E. 1976. Combining ability for seedling dry weight and forage yield in cicer milkvetch. Crop Sci. 16,480-482. Townsend, C. E. 1977a. Recurrent selection for high seed weight in cicer milkvetch. Crop Sci. 17,473-476. Townsend, C. E. 1977b. Germination of polycross seed of cicer milkvetch as aEected by year of production. Crop Sci. 17,909 - 9 12. Townsend, C. E. 1979a. Registration of C-4 and C-5 cicer milkvetch germplasm. Crop Sci. 19,298. Townsend, C. E. 1979b. Breeding cicer milkvetch for improved seedling emergence. Crop Sci. 19,613-616. Townsend, C . E. 1979c. Registration of C-9 cicer milkvetch germplasm. Crop Sci. 19,934. Townsend, C. E. 1979d. Associations among seed weight, seedling emergence, and planting depth in cicer milkvetch. Agron. J. 71,410-414. Townsend, C. E. 1980a. Forage legumes. In “Hybridization ofCrop Plants” (W. R. Fehr and H. H. Hadley, eds.), pp. 367 - 380. American Society of Agronomy/Crop Science Society of America, Madison, Wisconsin. Townsend, C. E. 1980b. Flowering characteristics of cicer milkvetch clones and their polycross progenies. Crop Sci. 20,479-483. Townsend, C. E. 1980~.Registration of Monarch cicer milkvetch. Crop Sci. 20,670-67 1. Townsend, C. E. 1981a. Breeding cicer milkvetch for improved forage yield. Crop Sci. 21, 363-366. Townsend, C. E. 198I b. Vernalization and photoperiod requirements for maximum flowering of cicer milkvetch. Crop Sci. 21,9 17-92 1. Townsend, C. E. 1982. Influence of seedling age and duration of vernalization on flowering of cicer milkvetch. Agron. J. 22, 1242- 1245.

CICER MILKVETCH (Astragah cicer L.)

307

Townsend, C. E. 1984. Inheritance of flowering date in cicer milkvetch. Crop Sci. 24, 196-200. Townsend, C. E. 1985. Recurrent selection for improved seed germination, seedling elongation, and seedling emergence. Crop Sci. 25,425 -429. Townsend, C. E. 1986. Evaluation of polycross progenies of cicer milkvetch for palatability by sheep. Crop Sci. 26,377-380. Townsend, C. E. 1987. Registration of four germplasm lines of cicer milkvetch. Crop Sci. 27, 368. Townsend, C. E. 1988. Breeding cicer milkvetch for increased forage yield during the decreasing photoperiods of mid- to late-summer. Agron. Abstr. p. 98. Townsend, C. E. 1990a. Registration of four germplasm lines of cicer milkvetch. Crop Sci. 30, 428-429. Townsend, C. E. 1990b. Changes in seed germination with time for polycross progenies of cicer milkvetch. Crop Sci. 30,694-698. Townsend, C. E. (1993). Registration of Windsor cicer milkvetch. Crop Sci. (in press). Townsend, C. E., and Ackerman, W. D. 1975. Variability for vigor, height, and flowering in introductions ofcicer milkvetch. Can. J. Plant Sci. 55,843-845. Townsend, C. E., and Ditterline, R. L. (1993). Registration of C-18, C-19, C-20, and C-21 germplasms of cicer milkvetch. Crop Sci. (in press). Townsend, C. E., and Hinze, G. 0. 1979. Registration of C-7 and C-8 cicer milkvetch germplasm. Crop Sci. 19,934. Townsend, C. E., and McGinnies, W. J. 1972a. Establishment of nine forage legumes in the central Great Plains. Agron. J. 64,699 - 702. Townsend, C. E., and McGinnies, W. J. 1972b. Temperature requirements for seed germination of several forage legumes. Agron. J. 64,809-812. Townsend, C. E., and McGinnies, W. J. 1973. Factors influencing vegetative growth and flowering in Astragalus cicer L. Crop Sci. 13,262- 264. Townsend, C. E., and Schweizer, E. E. 1984. Tolerance of cicer milkvetch (Astragalus cicer) seedlingsto herbicides. Weed Sci. 32,37-42. Townsend, C. E., and Wilson, A. M. 1978. Seedling growth of cicer milkvetch in controlled environments. Crop Sci. 18,662-666. Townsend, C. E., and Wilson, A. M. 1981. Seedling growth of cicer milkvetch as affected by seed weight and temperature regime. Crop Sci. 21,405-409. Townsend, C. E., Hinze, G. O., Ackerman, W.D., and Remmenga, E. E. 1975. Evaluation of forage legumes for rangelands of the central Great Plains. Gen. Ser. -Colo., Agric. Exp. Stn. 942. Townsend, C. E., Christensen, D. K., and Dotzenko, A. D. 1978. Yield and quality of cicer milkvetch forage as influenced by cutting frequency. Agron. J. 70, 109- 113. Townsend, C. E., Remmenga, E. E., Dewald, C. L., Ditterline, R. L., Melton, B. A., and Smoliak, S. 1979. Evaluation of seedling emergence in cicer milkvetch by linear regression. Crop Sci. 19,694-697. Townsend, C. E., Kenno, H., and Brick, M. A. 1990. Compatibility of cicer milkvetch in mixtures with cool-season grasses. Agron. J. 82,262-266. Vickers, J. C., Zak, J. M., and Odurukwe, S. 0. 1977. Effects of pH and Al on the growth and chemical composition of cicer milkvetch. Agron. J. 69, 5 I 1 5 13. Walsh, J. F., Bezdicek, D. F., Davis, A. M., and Hoffman, D. L. 1983. Nitrogen fixation capabilities of plant introduction accessions of pasture and range forage legumes. Agron. J. 75,474-478. Wasser, C. H. 1982. Ecology and culture of selected species useful in revegetating disturbed lands in the West, U.S. Dep. Inter., Fish Wildl. Sew.. Biol. Sew. Program 82/56.

-

308

C. E. TOWNSEND

Wegert, W. L. 1977. “Production and Forage Quality of Cicer Milkvetch (Astragalus cicer L.) Hay at Six Stages of Maturation,” M.S.thesis. Western State College, Gunnison, Colorado. Weimer, P. J., Buxton, D. R., and Hatfield, R. D. 1991. Inhibition of ruminal cellulolysis in Vitro by extracts of cicer milkvetch (Astragalus cicer). Proc. Znt. Symp. Forage Cell Wall Struct. Digestibility poster abstr. B-8. White, L. M., and Wight, J. R. 1981. Seasonal dry matter yield and digestibility of seven grass species, alfalfa, and cicer milkvetch in eastern Montana. Agron. J. 73,457-462. Williams, M. C., James, L. F., and Bleak, A. T. 1976. Toxicity of introduced nitrocontaining Astragalus to sheep, cattle, and chicks. J. RangeManage. 29,30-33. Wilson, D. B., and Rode, L. M. 1991. Irrigated pastures in western Canada. Agric. Can.Publ. 1862/E. Yakimova, Y., and Kolev, I. D. 1981. Phytocenotic characteristicsof species belonging to the genus Astragalus (milkvetch) in Bulgaria. Rasteniev’d Nauki 18,63-71 [Biol. Abstr. 74, 016386 (1982)l. Zak, J. M., Troll, J., Havis, J. R., Hyde, L. C., Kaskeski, P. A., and Hamilton, W. W. 1972. “A Handbook for the Selection of Some Adaptable Species for Massachusetts Roadsides,” Rep. 24-R5-2656 Roadside Dev. Department of Plant and Soil Science, University of Massachusetts, Amherst, Massachusetts.

Index content of excreta, 136- 137 phosphogypsum source, 66 Absorption edge phenomena, X-ray CAT Cereal crops,phosphogypsum use, 67-68 scanners, 20 - 2 1 Cicer milkvetch, 253-301 Acid soil adaptation, 265 -266 amelioration, aluminum toxicity indices, breeding, genetics, and cytology 74-75 breeding methodology, 29 1 -292 crop response to gypsum and phosphogyp cultivars, 299 -300 sum, 80-82 cytology and inbreeding depression, formation, 73-74 290-29 1 serious acidity, phosphogypsum, 75 - 78 flowering date, 299 weathered, reactions to gypsum and phosforage yield, 295-297 phogypsum, 78 - 80 objectives, 289-290 Aliasing artifacts, 12- 13 seedling vigor, 292-295 Aluminum seed weight, 296,298 - 299 aggregation and aggregate stability, 87-90 establishment, 269-271 toxicity flowering, 26 1 -262 indices, subsoil acidity amelioration, herbicide tolerance, 274-276 74-75 morphology and anatomy, 254-256 serious subsoil acidity, 75-78 flower, 255 -256 Ambient atmosphere, phosphogypsum efroot, 254-255 fects, 98- 100 seed coat, 256 Ammonia stem and leaf, 255 nitrification, 164- I66 pest resistance, 271-272 volatization, 16I - 164 photoperiod, 263-264 Animals, nutritional disorders, urine-afplanting patterns, 269-270 fected pasture, 173 pollination requirements, 272 - 273 Astragalus cicer L. see Cicer milkvetch seed germination, 256-259 Attenuation theory, 5-8 optimum temperature, 256-257 rate, 257, 259 B seedling emergence and seed weight, 270 seedling growth, 259-26 1 Back-projection reconstruction, 10 seed production, 273 Beam hardening, X-ray CAT scanners, 17seed scarification and inoculation, 26720 269 Beer's Law, 6 seed weight and seedling growth parameBiological activity, soil, 127- 129 ten, 260 Breeding, cicer milkvetch, see Cicer milksoils and soil fertility, 266-267 vetch utilization hay, see Hay C Pasture, 286 -289 vernalization, 26 1 -263 Cadmium, in phosphogypsum, 96 weed control, 274-276 Calcium Collimation, X-ray CAT scanners, I5

A

309

3 10

INDEX

Compton scattering, 6 Computer-assisted tomography, 1- 50, see also pray CAT scanners; X-ray CAT scanners application to dual-energy scanning, 4447

attenuation theory, 5 - 8 industrial systems, 47 potential applications, 4 principles, 8 - 13 aliasing artifacts, 12 - 13 back-projection reconstruction, 10 6ltered back-projection, 11- 12 iterative reconstruction, 10- 11 numerical reconstruction, 10- 12 recent and future developments, 47 -49 soil water studies, 26-41 attenuation coefficient differences, 28 29

bulk density, 30 - 3 1 Hounsfield values, 33-34 linearity, 26- 3 1 macroporosity, 35 m a s attenuation coefficient, 27-28 mathematical anomalies, 33-34 mean bulk density, 34-35 soil water extraction, 38 spatial distribution of soil water content,

Dinitrogen gas,losses, 166- 168 Dual-energy scanning theory, 43 water movement, 43-47 CAT application, 44 - 47 source choice, 44 Dynamic models, nutrient cycles, 18 1- 183

E Earthworms, population densities, 129 Electrical conductivity, see also Soil salinity aqueous, principles, 204 - 206 bulk soil, principles, 2 12- 2 19 correlation with clay percentage, 2 15 relation between bulk soil and saturation paste extract, electrical conductivity, 217-218

soil paste, principles, 2 19-220 two-component model, 2 12- 2 13 Electrolytic conductivity, soil salinity, 205 Electromagnetic induction soil conductivity sensor, 224,226-227 Electron-positron pair production, 6 EPA rule, phosphogypsum, 64 - 65 Excreta, see Nutrients, returns in excreta

F

38,45-46

spatial resolution of objects, 32-33 structural definition, 3 1- 36 water drawdowns, 38-40 water movement to plant roots, 36-41 Conductance, specific, see Soil salinity Cornforth-Sinclair model, phosphorus cycling, 185- 187 cows excretion and retention of nutrient, 130131

feces composition, 145 nitrogen content of excreta, 132- 134 number and size of excretions, 137- 140 Crop tissues, phosphogypsum effects, 95 -97 Cultivars, cicer milkvetch, 299 - 300 Cytology, inbreeding depression, cicer milkvetch, 290-291

D Denitrification, losses from grazed pastures, 166- 168

F w s , see also Nutrients, returns in excreta composition, 144- 145 degradation, 145- 149 rainfall and, 148- 149 nitrogen release, 149- 15 1 phosphorus release, 15 1- 152 sulfur release, 15 1 Fertilizer low-analysis, phosphogypsum as bulk carrier, 93 recommendations, using nutrient cycling models, 183-189 role in pastures, 122- 123 Filtered back-projection, 11- 12 Flower, cicer milkvetch, 255-256,261-262 Flowering date, cicer milkvetch, 299 Fluoride Al toxicity amelioration, 76-77 phosphogypsum effects, 95-96 Forage crops cicer milkvetch, 279-280,282-286 yield, 295-297, 301

INDEX

311

phosphogypsum uses, 70-72 Fourelectrode soil conductivity probes, 223-225 Fruits, phosphogypsum uses, 70

Herbicides, cicer milkvetch tolerance, 274276 Hounsfield units, 16- 17 Hydrolysis, urea, 160 - 161

G

I

pray CAT scanners, 22-25 advantages, 22 application to dualenergy scanning, 4447 improved image and data analysis software, 47 logistic system, 23 -25 radiation detection, 25 utility, 49 Geonics EM-38 device, 230 Grain legumes, phosphogypsum uses, 68-69 Grasslands, man-made, 121 Grazed pastures, see Pasture ecosystem Grazing, surface soil properties, 128 Ground water, surficial, phosphogypsum effects, 93 - 94 Gypsum agregation and aggregate stability, 87-90 by-products, see Phosphogypsum crop response on acid soils, 80- 82 nonsodic dispersive soils, 90-9 1 reactions with weathered acid soils, 78 - 80 self-liming mechanism, 89 sodic soil reclamation, S4-86

Inbreeding depression, cytology, cicer milkvetch, 290-291 Iron aggregation and aggregate stability, 87 - 90 Al toxicity amelioration, 76 Irrigation, surface soil properties, 128 Isoflavonoid phytoalexins, cicer milkvetch, 278 Iterative reconstruction, 10- 1 1

H Hardpans, phosphogypsum effect, 9 1- 92 Hard-setting clay soils, phosphogypsum effect, 92 - 93 Hay, cicer milkvetch, 276-286 amino acid composition, 276 carbohydrate reserves, 279-281 competitivenessin mixtures, 280, 283 forage yield, 279-280,282-286 isoflavonoid phytoalexins, 278 quality factor and mineral concentrations, 276-277 spring growth patterns, 278 Herbage chemical composition, 156, 172- 173 in dung application zone, 155 utilization, in grazed pastures, 156- 158

L Leaching, nutrient losses, 168- 170 Leaf, cicer milkvetch, 255

M Magic-angle spinning, 42 Magnesium, content of excreta, 136- 137 Mass attenuation coefficient, 27-28 Mass balance models, nutrient cycling, 175181

Mathematical models, excretal distribution, 142- 143 Micronutrients, phosphogypsum as bulk carrier, 93 “Model/Field-Estimates” technique, 235 237 Molybdenum, phosphogypsum effects, 95

N Nitrate, accumulation in urine patches, 165 Nitrification, ammonia, 164- 166 Nitrogen content of excreta, 132- 134 major fluxes in farm types, 176- 180 release from feces, 149- 15 1 Nitrous oxide gas, losses, 166- 168 Nonsodic dispersive soils, use of gypsum and phosphogypsum, 90-91 Nuclear magnetic resonance imaging, water movement, 4 1 -42

312

INDEX

Nutrient cycling excreta role, 143- 144 modeling, 174- 189 dynamic models, 18 1- 183 major pools and fluxes, 174- 175 mass balance data, 176- 181 types, 175-176 use for fertilizer recommendations, 183-189 soil/plant/animal system, 120 Nutrients immobilization into organic forms, 166 leaching losse~,168- 170 loss from pasture systems, 122 release from feces composition, 144- 145 degndation, 145- 149 nitrogen, 149-151 PhOSphoms, 15 1 - 152 soil property effect, 152- 154 sulfur, 15I returns in excreta, 130- 144 calcium, 136 distribution models, 142- 143 distribution of returns, 140- 142 magnesium, 136 nitrogen, 132- I34 number, size, and area covered, 137140 nutrient cycling role, 143- 144 phosphorus, 135- 136 potassium, 136 quantities returned, 130- 132 sulfur, 134- 135 trace elements, 137 status and soil organic matter, 123- 126 Nutritional disorders, animals, urine-affected pasture, 173

0 Organic matter, soil nutrient status, 123126

P Pasture, cicer milkvetch, 286-289 Pasture ecosystem, 1 19- 19I annual nutrient uptake, 189

biological activity, 127- 129 botanical instability, 121 earthworm populations, 129 fertilizer role, 122- I23 grazing animals, 121 nutrient movement, 120 modeling nutrient cycling, see Nutrient cycling movement and transformations of nutrients from urine, f 58- 170 composition, 158 dinitrogen and nitrous oxide gases loss, 166- 168 leaching ~OSS~S,168- 170 macropore flow, 158- I60 nitrification, 164- 166 nutrient immobilization into organic forms, 166 soil pH and charge characteristics, 164 urea hydrolysis and ammonia volatization, 160- 164 nature of, 12I nutrient release from feces, see Nutrients response in fecal patch, 154- 158 direct adverse affect, 154- 155 herbage chemical composition, I56 herbage utilization, 156- I58 positive pasture response, 155 response in urine patch, 170- 173 herbage chemical composition, 172173 nutritional disorders of animals, 173 positive growth response, 170 - 17 I urine scorch, 171- 172 soil organic matter and nutrient status, 123- 126 soil pH, 126-127 soil physical properties, 129- 130 Pest resistance, cicer milkvetch, 27 1 -272 pH, soil pastures, 126- 127 urine patches, 164 Phosphogypsum, 55- 102 agricultural uses acid soil formation, 73 -74 aggregation and aggregate stability, 87 90 ameliorant for aluminum toxicity and subsoil acidity, 73- 82 ameliorant for sodic soils. 83-87

INDEX bulk carrier for micronutrients and lowanalysis fertilizers, 93 cereal crops, 67 - 68 forage crops, 70- 72 fruits and vegetables, 70 grain legumes, 68-69 hardpans, 9 1 -92 hard-setting clay soils, 92-93 nonsodic dispersive soils, 90- 9 1 sodic soil reclamation, 84-86 sugarcane, 69 - 70 sulfur deficiency and need for Ca source other than lime, 65 -66 ambient atmosphere effects, 98- 100 background y-radiation, 99 - 100 crop response on acid soils, 80-82 crop tissue effects, 95-97 physical and chemical properties, 58-65 analysis from countries, 58 - 59 elemental impurities, 58, 60 EPA rule, 64-65 radioactivity, 63-64 Solubility curves, 60-62 toxicity index metals solubility, 60, 63 reactions with weathered acid soils, 7880 soil effects, 94-95 surficial ground water effects, 93-94 world production and utilization, 56 - 58 Phosphorus content of excreta, 135- 136 major fluxes in farm types, 176- 180 release from excreta, 151- 152 requirement estimations using nutrient cycling models, 183- 187 Photoelectric absorption, 3 -6 Photoperiod, cicer milkvetch, 263 -264 Pigs, phosphate content of feces, 135136 Plant roots, water movement around, see Computer-assisted tomography Pollination, cicer milkvetch requirements, 272-273 Porosity, soil, pastures, 129- 130 Potassium content of excreta, 136 major fluxes in fann types, 176- 180 requirement estimations using nutrient cycling models, 187- 188

313

R Radioactivity, phosphogypsum, 63 Radionuclide, uptake and phosphogypsum, 96-97 Radon, phosphogypsum and, 98-99, 101 Rayleigh scattering, 6 Ray sum, 7 Recolonization, fecal patch areas, 154- I55 “Regression ModellGround-Truthing” technique, 237-239 Rhizobium leguminosarum, cicer milkvetch seed inoculation, 268 Root, cicer milkvetch, 254-255

S salinity sensors, 208-210 Saturation extract method, 2 1 1 Seedling growth, cicer milkvetch, 259-261 vigor, cicer milkvetch, 292-295 Seeds

coat, anatomy, cicer milkvetch, 256 germination, cicer milkvetch, 256-259 production, cicer milkvetch, 273 scarification and inoculation, cicer milkvetch, 267-269 weight, cicer milkvetch, 296,298-299 Sheep feces composition, 145 number and size of excretions, 137- 140 phosphorus content of excreta, 135 sulfur content of excreta, 134- 135 SO,, adsorption in acid soils, 77 - 79 Sodic soils characteristics, 83-84 reclamation, gypsum or phosphogypsum, 84 - 86 use of gypsum and phosphogypsum, 8687 Soil, see also Acid soil aggregation, Al, Fe, gypsum and phosphogypsum, 87-90 cicer milkvetch, 266 - 267 extract salinity, 210-212 fertility, cicer milkvetch, 266-267 organic matter and nutrient status, 123I26

3 14

INDEX

pH and charge characteristics, urine patches, 164 phosphogypsum effects, 94-95 porosity, pastures, 129- 130 properties nutrient release from excreta effect, 152-154 pastures, 129- 130 Soil loss factor, 185 Soil salinity, 201 -246 aqueous electrical conductivity salinity sensors,208 - 2 10 soil extract salinity, 210-212 soil water salinity, 206 - 2 10 assessment, 202-203 bulk soil electrical conductivity, 221, 223-245 attenuation coefficient, 228 electromagnetic induction unit, 224, 226-227 fourelectrode units, 223-225 large-volume measurements, 229 -233 “Model/Field-Estimates” technique, 235-237 new relations, 232-233 “Regression Model/Ground-Truthing” technique, 237-239 sensors, 223-229 small-volume measurements, 233-234 “soil-type calibration” technique, 234235 surface-array method, 229-230 time domain reflectometry unit, 226, 228-229 definition, 201 instrumental system use examples, 242245 measurement method comparisons, 23924 1 measurement site location determination, 24 I - 242 salinity maps, 242-245 saturated paste electrical conductivity, 220-222 soil water content variance effect, 2 18 “Soil-type calibration” technique, 234-235 Soil water potential, 2 sample collection, 206-207 sampling errors, 207 - 208

studies, 26 volumetric content as function of total water content, 215-216 Stems, cicer milkvetch, 255 Sugarcane, phosphogypsum uses, 69-70 Sulfur content of excreta, 134- 135 deficiency forage crops, 70 phosphogypsum use, 65-66 major fluxes in farm types, 176- 180 release from excreta, I5 1 requirement estimations using nutrient cycling models, 187- 188 Superphosphate effect on organic matter accumulation, 125 surface soil properties, 128

T Time domain reflectometry unit, soil salinity, 226,228-229

U Urea, hydrolysis, 160 - 16 1 Urine, see also Nutrients, returns in excreta scorch, 171-172

V Vegetables, phosphogypsum uses, 70 Vernalization, cicer milkvetch, 26 1- 263 Volatization, ammonia, 161 - 164

W Water drawdowns, 38-40 flow, total resistance, 37 movement around plant roots, see also Computer-assisted tomography dual-energy scanning, 43-47 NMR, 4 1 -42 Weed control, cicer milkvetch, 274-276 Wenner array, 230

315

INDEX

X X-ray CAT scannen, 3 -4, 13 - 22 absorption edge phenomena, 20-2 1 beam hardening, I7 -20 collimation, 15

-

construction, 13 14 detection, 14- 15 Hounsfield units, 16- 1 7 limitations, 2 I - 22 reconstruction, I5 - 16