Advances in Agronomy, Volume 55

$. V O L U Ms sE . .- Advisory Board Cornell University Eugene J. Kamprath North Carolina State University Kenne...

Author: Donald L. Sparks

30 downloads 1489 Views 24MB Size Report

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

Report copyright / DMCA form

DOWNLOAD PDF

$.

V O L U Ms sE .

.-

Advisory Board

Cornell University

Eugene J. Kamprath North Carolina State University

Kenneth J. Frey

Larry P. Wilding

Iowa State University

Texas A&M University

Martin Alexander

Prepared in cooperation with the American Society of Agronomy Monographs Cornmiltee

M.A. Tabatabai, Chairman S. H. Anderson P. S. Baenziger W. T. Frankenberger, Jr.

D. M. Kral S. E. Lingle R. J. Luxmoore

G . A. Peterson S. R. Yaks

D V A N C E S I N

onomy VOLUME 55 Edited by

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

ACADEMIC PRESS SanDiego NewYork Boston London Sydney Tokyo Toronto

This book is printed on acid-free paper.@ Copyright 0 1995 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. A Division of Harcourt Brace & Company 525 B Street, Suite 1900, San Diego, California 92101-4495 United Kingdom Edition published by Academic Press Limited 24-28 Oval Road, London NWl 7DX

International Standard Serial Number: 0065-21 13 International Standard Book Number: 0-12-OOO755-X PRINTEDIN THE UNlTED STATESOF AMERICA 95 96 9 7 9 8 99 0 0 B B 9 8 7 6

5

4

3 2 1

Contents CONTRIBUI-ORS. .............................................

iX

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

xi

SYNCHROTRON X-RAYTECHNIQUES IN SOIL,PLANT, AND ENVIRONMENTAL RESEARCH Darrell G . Schulze and Paul M. Bertsch 1. Introduction. ..................................................... 11. Synchrotron X-Ray Sources and Synchrotron Light.. .................. III. Synchrotron X-Ray Techniques and Their Applications................. Iv. Accessing Synchrotron Facilities. .................................... References .......................................................

2 3

1s 54 57

GEOGRAPHIC INFORMATION SYSTEMS IN AGRONOMY G. W. Petersen, J. C. Bell, K. McSweeney, G. A. Nielsen, and P. C. Robert I. Introduction. ..................................................... II. Overview of GIS Technology. ...................................... III. Remote Sensing. ..................................................

Iv. Terrain Analysis and Soil-LandscapeModeling. ....................... V. Site-Specific Farming. ............................................. VI. Environmental Applications.........................................

vn.

Conclusions ...................................................... References .......................................................

68 68 74 84 87 98 104 105

USDA PLANTGENOMERESEARCH PROGRAM I. 11.

USDA Plant Genome Research Program Participants Introduction.. .................................................... Progress .........................................................

In. Plant Genome Database.. ..........................................

Iv. Future Projections ................................................ References .......................................................

V

113 115 147 154 156

vi

CONTENTS

ANALYSISOF ORGANICMATTERIN SOILEXTRACTS AND WHOLESOILSBY PYROLYSIS-MASS SPECTROMETRY M. Schnitzer and H.-R. Schulten I . Introduction...................................................... I1. Fundamentals of Pyrolysis-Mass Spectromemc Methods ............... III. Analysis of SOM by Pyrolysis-Soft Ionization Mass Spectrometry........ lv. Summary of Data Obtained on the Extractions with Organic Solvents .... V. Curie-Point Py-GUMS of Humic Acids and the Development of Novel Concepts for Their Chemical Structure .............................. VI. Analysis of Soil Organic Matter by Py-FIMS .......................... VII. Effects of Minerals on the Py-FIMSof Fulvic Acid .................... WJ . Other Applications ................................................ M . Conclusions ...................................................... References .......................................................

168 170 176 190 191 199 208 210 211 213

ROLEOF METAL-ORGANICCOMPLEXATION IN METALSORPTION BY SOILS

Robert D . Harter and Ravendra Naidu I . Introduction...................................................... I1. Organics in the Soil Solution ....................................... 111. Metals in the Soil Solution ......................................... Iv. Effect of Low-Molecular-Weight Organics on Metal Ion Reactions with Organic Surfaces .................................................. V. Effect of Organics on Reactions of Metal Ions and Complexes with Inorganic Surfaces................................................. VI. Environmental Implications......................................... VII . Summary and Research Needs ...................................... References .......................................................

219 220 223 229 236 248 254 254

DNA MARKERS AND PLANTBREEDINGPROGRAMS Michael Lee I. Introduction...................................................... 11. Assessing Genetic Diversity and Merit ............................... 111. Genome Architecture: Genetic and Physical Characterization of Crop Plant Gnomes ................................................... Iv. Analysis of Complex T raits and Phenomena........................... V. Marker-Assisted Selection ..........................................

VI. Survey of the Status of DNA Markers in Cultivar Development Programs W . Summary and Conclusions ......................................... References .......................................................

265 269 283 300 313 320 328 330

CONTENTS

vii

LONG-TERM PERSISTENCE OF ORGANICCHEMICALS IN SEWAGE SLUDGE-AMENDED AGRICULTURAL LAND: A SOILQ U A LPERSPECTIVE ~

Angus J. Beck. Ruth E. Alcock. Susan C. Wilson. Min-Jian Wang. Simon R . Wild. Andrew P. Sewart. and Kevin C.Jones I . Introduction ...................................................... 345 II. Long-Term Experiments and the Compounds Investigated.............. 348 111. Influence of Sewage Sludge Applications on the Concentration and

Persistence of Organic Chemicals in Soils.............................

353

Quality Criteria................................................... V. Conclusions ...................................................... References .......................................................

376 386 387

Iv. Implications of Sewage Sludge Application to Farmland for Soil

PEANUTBREEDINGAND GENETICS

David A. Knauft and Johnny C. Wynne I. Introduction ...................................................... 11. Diversity of Peanut ................................................ m . GeneticVariability ................................................ Iv. Genetics ......................................................... V . Breeding......................................................... VI. Research Related to Breeding ....................................... w . Summary and Conclusions ......................................... References ....................................................... INDEX.....................................................

393 394 397 403 406 416 429 430 447

This Page Intentionally Left Blank

Contributors Numbers in parentheses indicate the pages o n which the authors' conmbutions begin.

RUTH E. ALCOCK (349, Institute of Environmental and Biological Sciences, Lancaster University, Lancaster, LA1 4YQ, United Kingdom ANGUS J. BECK (345), Institute of Environmental and Biological Sciences, Lancaster University, Lancaster, LA1 4YQ, United Kingdom J. C. BELL (67), Soil Science Department, University of Minnesota, St. Paul, Minnesota 55108 PAUL M. BERTSCH (l),Advanced Analytical Centerfor Environmental Sciences, Savannah River Ecology Luboratq, University of Georgia, Aiken, South Carolina 29802 ROBERT D. HARTER (2 19), Department of Natural Resources, University of New Hampshire, Durham, New Hampshire 03824 KEVIN C. JONES (345), Institute of Envimnmental and Biological Sciences, Lancaster University, Lancaster, LA1 4YQ, United Kingdom DAVID A. KNAUFT (393), Department of Crop Science, North Carolina State University, Raleigh, North Carolina 27695 MICHAEL LEE (265), Department of Agronomy, Iowa State University,Ames, Iowa 50011 K. MCSWEENEY (67), Department of Soil Science, University of Wisconrn, Madison, Wisconsin S3706 RAVENDRA NAIDU (2 19), CSIRO Division of Soils and Cooperation Research fm Soil and Land Management, Glen Omond, SA 5064 Australia G. A. NIELSEN (67), Department of Plant and Soil Science, Montana State University, Bozeman, Montana 59717 G. W. PETERSEN (67), Department ofAgronomy and EnvironmentalResources Research Inm'tute, The Pennsylvania State University, UniversityPark, Pennsylvania 16802 P. C. ROBERT (67), Soil Science Department, University ofMinnesota, St. Paul, Minnesota 55108 M. SCHNITZER (167), Agriculture and Agri-Food Canada, Ottawa, Ontario Cana& H.-R. SCHULTEN (167), Fachhochschule Fresenius, Department of Trace Analysk, 6S193 Wiesbadm, Gennany DARRELL G. SCHULZE (l), Agronomy Department, Purdue University, West Wayette, Indiana 47907

X

CONTRIBUTORS

ANDREW P. SEWART (349, Institute of Environmental and Biological Scim e s , Luncaster Univm’ty, Lancaster, LA1 4YQ, United Kingdom USDA PLANT GENOME RESEARCH PROGRAM PARTICIPANTS (1 13), USDA, Agricultural Researcb Service, Beltsoille, Maryland 2070s MIN-JIAN WANG (349, Institute of Environmental and Biological Sciences, Luncaster Univm’ty, Luncaster, L A 1 4YQ, United Kingdom SIMON R. WILD (345), Institute of Environmental and Biological Sciences, Lu?uaster Univm‘ty, Luncaster, L A 1 4YQ, United Kingdom SUSAN C. WILSON (349, Institute of Environmental and Biological Sciences, Lancaster University, Luncaster, LA1 4YQ, United Kingdom JOHNNY C . WYNNE (393), Department of Crop Science, North Carolina State University, Raleigh, N o d Carolina 2769s

Preface Volume 55 contains eight outstanding reviews on cutting-edge advances in the crop and soil sciences. The important themes in agronomy of environmental quality and biotechnology are emphasized. Three of the reviews deal with aspects of pollutants in the soil environment, persistence, and effects on soil quality. Three of the reviews discuss important advances in molecular biology and plant breeding, including the plant genome research program of the USDA, DNA markers, and advances in peanut breeding. Three of the reviews discuss state-of-the-art techniques that have wide applicability in agronomy, including synchrotron radiation, geographic information systems (GIs), and pyrolysismass spectrometry. Chapter 1 is a comprehensive review of synchrotron radiation techniques and their use in soil, plant, and environmental research. These exciting methods are enabling scientists to obtain important mechanistic information on the interactions of contaminants with soils. Chapter 2 reviews advances in GIS technology and its use in integration of scientific data, spatial and temporal variability, and environmental assessments. Chapter 3 thoroughly reviews the USDA plant genome research program including its history and advances in improvement of important agronomic crops. Chapter 4 discusses an important technique, pyrolysis-mass spectrometry, that has been successfully used to determine the structure of soil organic matter and humic substances. Chapter 5 covers effects of metal-organic complexation on sorption of metals in soils including the role these complexes play in contaminant transport, soil genesis and fertility, and metal toxicity. Chapter 6 explores the role of DNA markers in plant improvement and in plant breeding research. Chapter 7 discusses the effects of sewage sludge applications on the concentration and persistence of organic chemicals in soils over long time periods. Chapter 8 is a thorough discussion of advances in the genetics and breeding of peanuts. Many thanks to the authors for their first-rate reviews.

DONALD L. SPARKS

xi

This Page Intentionally Left Blank

SYNCHROTRON X-RAYTECHNIQUES IN Son, PLANT,AND ENVIRONMENTAL RESEARCH Darrell G. Schulzel and Paul M. Bertsch2 %gronomy Department Purdue University West Lafayette, Indiana 47907 2Advanced Analytical Center for Environmental Sciences Savannah River Ecology Laboratory University of Georgia Aiken, South Carolina 29802

I. Introduction II. Synchrotron X-Ray Sources and Synchrotron Light A. General Description of a Synchrotron B. Properties of Synchrotron Radiation C. Hard versus Soft X-Ray Synchrotrons D. Summary III. Synchrotron X-Ray Techniques and Their Applications A. X-Ray Absorption Spectroscopy (XAS) B. Synchrotron X-Ray Fluorescence Spectroscopy C. Standing Wave and Fluorescent X-Ray Interference Techniques D. Infrared Mimspectroscopy E. Miissbauer Spectroscopy F. X-Ray Ditfraction G. Small Angle Scattering H. X-Ray Microscopy I. X-Ray Computed Microtomography W. Accessing Synchrotron Facilities References

1 AhunnrinAgra~nr~r VdvnuSI Copyright 0 1995 by Academic P m , Inc. All righm of reproduction in any form reserved.

2

D. G. SCHULZE AND P. M. BERTSCH

I. INTRODUCTION X-ray-based techniques have a long history of applications in agricultural research. The discovery of the crystalline nature of colloidal soil particles in the 1930s, for example, was a major breakthrough made possible by the then-new technique of X-ray diffraction. X-ray powder diffraction remains an essential tool today for soil mineralogy and chemistry research, while other X-ray-based techniques such as X-ray fluorescence spectroscopy, radiography, and computed tomography are important to individual researchers based on the availability of equipment and the needs of particular research projects. Commercially available X-ray instrumentation relies on specialized vacuum tubes as the X-ray source. The capability of sealed-tube X-ray sources has not increased significantly since Wilhelm Conrad Rontgen’s discovery of X rays a century ago. The introduction of rotating anode X-ray tubes in the 1960s brought about a 10-fold increase in X-ray intensity, but the basic constraints of a vacuum tube X-ray source, namely, significant intensity over only a few narrow energy ranges and a highly divergent source, remained. Beginning in the 1950s, the high-energy physics community began to build particle accelerators to study the fundamental properties of matter. One type of particle accelerator, the synchrotron, was designed to accelerate charged particles around a nearly circular trajectory so the particles could be made to strike a target at high energy. Synchrotrons produced large quantities of electromagnetic radiation, including X rays, as a byproduct of steering the particles around the ring. This radiated energy was originally considered a nuisance because it continually had to be replaced, but it soon became apparent that the synchrotron radiation had many useful properties for X-ray-based techniques. New generations of synchrotrons designed exclusively as X-ray sources have followed, and these powerful sources of X rays have become important to a wide variety of scientific disciplines. The past 5 years have seen a growing number of applications of synchrotronbased techniques to problems in the soil and environmental sciences. Synchtronbased techniques have applications in many other areas of agricultural research as well. In this review, we will attempt to highlight some of the major applications in soil, plant, and environmental research. Many of these applications represent the first use of synchrotron-based techniques in particular agricultural disciplines. We expect increased applications in the areas highlighted in the following sections, as well as entirely new applications in other areas of agricultural and environmentalresearch. The first section of this review will describe how synchrotrons work, describe in some detail the properties of synchrotron radiation, and explain some of the terminology associated with synchrotronbased research. The second section will review the applications of synchrotronbased techniques to soil, plant, and environmental research and suggest possible

SYNCHROTRON X-RAY TECHNIQUES

3

future applications. The final section will describe how one can obtain access to synchrotron sources to conduct research.

11.

SYNCHROTRON X-RAYSOURCES AND SYNCHROTRON LIGHT

Synchrotrons vary in their capabilities, so a general understanding of their differences will allow one, in a general sense, to assess the suitability of a particular synchrotron for a specific experiment. We draw on articles by Winick (1987), who gives a more general introduction, and Rivers (1990) and Smith and Rivers (1994), who give more technical descriptions of synchrotrons and the properties of synchrotron radiation.

A. GENERAL DESCRIPTION OF A SYNCHROTRON Synchrotrons are large machines. For example, the Advanced Photon Source (APS) at Argonne National Laboratory (Fig. 1) has a storage ring 1104 m in circumference, while the X-ray storage ring at the National Synchrotron Light Source (NSLS) at Brookhaven National Laboratory is 170 m in circumference. Because of their size, complexity, and high operating costs, synchrotrons tend to be national or regional facilities, and researchers must travel to them to take advantage of their unique capabilities. In a synchrotron (Fig. 2), charged particles, either electrons or positrons, are injected into a ring-shaped vacuum chamber kept at ultrahigh vacuum (- 10-9 Ton). The vacuum chamber is not a perfect circle, but consists of a series of arcs connected by straight sections. The particles enter the ring via the injection magnet and then travel around the ring at near the speed of light, being steered by a series of bending magnets. Additional magnets, quadrupoles with four individual poles and sextupoles with six (not shown in Fig. 2), focus and shape the particle beam as it travels around the ring. Energy lost by the particles as they travel around the ring is replaced when the particles travel through radiofrequency cavities, where synchronized electromagnetic fields impart energy to the particles to keep them circulating around the ring at near the speed of light. Radiation in the form of infrared, visible, ultraviolet and X ray light is emitted when the charged particles pass through the bending magnets or through insertion devices, additional magnetic devices called wigglers or undulators, which are "inserted" into the straight sections of the ring. Beam lines allow the radiation to enter experimental stations, shielded moms that house the instrumentation

Figure 1. Artist’s drawing of the Advanced Photon Source (Argonne National Laboratoryphoto).

SYNCHROTRON X-RAY TECHNIQUES Insertion Device Beamline

Experimental Stations

\

2I I u-m

Injection Magnet

Radio-frequency Cavity

5

m

Vacuum Chamber

Flgwe 2. Schematic diagram of a synchrotron X-ray source.

needed for specific experiments. The ring itself is located behind a heavily shielded concrete wall, and all pipes transmitting radiation and all experimental stations housing instrumentation are heavily shielded and electrically interlocked to protect the users from potentially lethal radiation. To the physicist, a synchrotron is designed to accept low-energy particles and accelerate them to some higher energy, while a storage ring is designed to accept particles at a given energy and circulate them at that energy for long periods of time (Winick, 1987). As the particles circulate around the storage ring, they collide with residual gas molecules in the vacuum chamber, resulting in a gradual loss of the particle beam current. Thus, X-ray intensity produced as the particles pass through the bending magnets and insertion devices drops off over time. In typical operations, particles are injected into the ring up to some maximum operating current. The particle beam current then decays over time until some lower limit is reached, at which point the remaining particles are dumped. The ring is then refilled and the cycle is repeated. The time between fills, and thus the time a continuous, stable X-ray beam is available, varies from about 3-4 to 12 h or more, depending on the operating characteristics of the particular storage ring. Refilling of the ring generally takes about an hour or less. Synchrotrons that circulate positrons typically

D. G. SCHULZE AND P. M. BERTSCH

6

have a longer beam lifetime than synchrotrons that circulate electrons, thus offsetting the additional complexity of producing positrons.

1. First, Second, and Third Genecation Synchrotron X-Ray Sources Synchrotrons originally designed for high-energy physics experiments, but retrofitted with X-ray beam lines, are considered the first generution of synchrotron X-ray sources. These machines continued to be operated for highenergy physics experiments, while the X-ray researchers had to arrange their work around a schedule dictated by the physics experiments. The X-ray research was considered “parasitic” with respect to the high-energy physics research; thus, references to parasitic synchrotron X rays occur in the early literature. Examples of first generation sources in the U.S. include the Stanford Synchrotron Radiation Laboratory, which now is operated solely as an X-ray source, and the Cornell High Energy Synchrotron Source (Table I), which still serves both the high-energy physics and the X-ray research communities. In the 1970s it became apparent that additional, dedicated synchrotron X-ray facilities were needed, and second generation synchrotrons were constructed. These second generation synchrotrons largely rely on X-ray generation by bending magnets, but they were designed and optimized specifically as X-ray sources. Second generation sources include the National Synchrotron Light Source in the U .S . Table I Sekted Firat, Second, and Third Generation SynchrotronResearch Facilities.

Acronym

SSRL CHESS

LURE HASYLAB SRS KEK NSLS

APS ALS

ESRF SRing-8 a

Facility First Generation Sources Stanford Synchrotron Radiation Laboratory Cornell High Energy Synchrotron Source Laboratoh pour, I’Utilisation du Rayonnement Electromagdtique Hamburger SynchrotronstrahlungsLabor Second generation sources Synchrotron Radiation Source Photon Factory National Synchrotron Light Source Third generation sources Advanced Photon Source Advanced Light Source European Synchrotron Radiation Facility 8 GeV Super Photon Ring

See Winick and Williams (1991) for a complete list worldwide.

Location Stanford, California Ithaca, New York Orsay, France Hamburg, Germany Daresbury, United Kingdom Tsukuba, Japan Upton, New York Argonne, Illinois Berkeley, California Grenoble, France Nishi Harima. Japan

SYNCHROTRON X-RAYTECHNIQUES

7

and others in Europe, Japan, and elsewhere (Table I). As the second generation sources were becoming operational in the 1980s, it was discovered that the insertion devices described earlier could be built and operated reliably to produce X-ray beams several orders of magnitude more powerful than a bending magnet on the same ring. Although some insertion devices have been installed on first and second generation synchrotrons, these machines often do not have the physical space for large numbers of insertion devices. Third generation synchrotrons, which have just begun operation or are still under construction, are designed specifically to accommodate large numbers of insertion devices. Third generation synchrotron sources include the Advanced Photon Source at Argonne National Laboratory, Argonne, IL (Fig. l), the Advanced Light Source (ALS) at Lawrence Berkeley Laboratory, Berkeley, CA, the European Synchrotron Radiation Facility (ESRF) in Grenoble, France, and the SPring-8 in Nishi Harima, Japan (Table I), as well as others around the world. Synchrotron radiation sources representing all 3 generations were operational or planned at 39 laboratories in 15 countries as of 1991 (Winick and Williams, 1991).

B. PROPERTIES OF SYNCHROTRON RADIATION 1. X-Ray Generation A unifying concept that applies to such diverse radiation sources as radio transmitters, X-ray tubes, and synchrotrons is that electromagnetic radiation i s produced when electric charge is accelerated (Miller, 1972). The acceleration of electrons as they travel back and forth in a radio antenna gives rise to radio waves. In an X-ray tube, electrons are accelerated to a high speed by the large potential difference between the filament and the target (e.g., 35,000 V) and then rapidly decelerated (a negative acceleration) when they strike the target, resulting in the emission of a continuum of electromagnetic radiation called the Bremsstrahlung or breaking radiation. In a synchrotron, the charged particles (either electrons or positrons) circulating in the ring experience an inward centripetal acceleration as they pass through the magnetic field of the bending magnets that steer them around the ring or a number of centripetal accelerations when they pass through the insertion devices. Thus, Bremsstrahlungor breaking radiation is generated as the particles pass through the bending magnets and insertion devices. In contrast to an X-ray tube, which emits X rays in all directions, X rays produced by a bending magnet are all emitted in a thin, fan-shaped pattern in the plane of the particle orbit and tangential to the forward direction of the particles (Fig. 3). The particles in the storage ring travel at very nearly the speed of light, and the thin, fan-shaped beam of X rays is the result of a relativistic effect explained by the theory of special relativity. If one were an observer riding along

D.G. SCHULZE AND P.M.BERTSCH

T rn>\

t7-K

flux

Figure 3. Charged particles traveling through the magnetic field of bending magnets or insertion devices experience an acceleration that results in the emission of X rays tangential to the particle trajectory. The diagram also illustrates the difference between X-ray flux, brightness, and brilliance.

with an electron or a positron experiencing an acceleration in a bending magnet, one would observe that the radiation emitted from the particle is always emitted in the shape of a torus (donut without a hole), regardless of the speed of the particle. If the particle is moving slowly (much less than the speed of light) relative to a stationary observer in the laboratory, there is virtually no difference between what the observer in the laboratory sees and what the observer traveling with the particle sees, namely, radiation emitted in a toroidal pattern. If, however, the particle is moving at very nearly the speed of light relative to the stationary observer in the laboratory, the observer traveling with the particle still sees the radiation being emitted in a toroidal pattern, but the observer in the laboratory sees a cigar-shaped pattern of X rays emanating in front of the moving particle, with very little radiation emanating to the back or sides. The reason for this is that special relativity requires that the speed at which electromagnetic radiation propagates, i.e., the speed of light, is the same constant for both the observer riding with the particle and the observer in the laboratory. The only way that this can occur is if the stationary laboratory observer sees the toroidal radiation pattern distorted into a cigar-shaped radiation pattern preceding the rapidly moving electron or positron. Figure 3 illustrates the instantaneouspattern of emitted radiation as a truncated cone. The sum of all the instantaneous, cigar-

SYNCHROTRON X-RAY TECHNIQUES

9

shaped patterns that arise as the particle travels around an arc results in the fanshaped X-ray pattern produced by a bending magnet (Fig. 3). The width of the fan in the horizontal plane is determined by the radius and degrees of arc of the bending magnet and is wide enough that portions of the fan from one bending magnet can be used to supply two or more separate beam lines. Vertically, the angle subtended by the fan is only a few hundred microradians or less (100 prad = 0.0057’); the exact value depends on the energy of the particles in the ring. Thus, even 10 or 20 m from the source, the X-ray beam is only a few millimeters high, and the instrumentation for a particular experiment may be 20-40 m or more from the X-ray source. The very low angular divergenceof the beam means that the X rays are almost parallel or highly collimated. This high degree of natural collimation has distinct geometric advantages in the design of experimental apparatus compared to the constraints imposed by the highly divergent X rays produced by an X-ray tube. In contrast to X-ray tubes, which produce relatively weak bremsstrahlung radiation and stronger, more usable X-ray fluxes at only one or two energies corresponding to the characteristic fluorescence radiation of the target element, bending magnets produce intense radiation over a broad energy range. Thus, bending magnets are “white light” radiation sources. Synchrotron bending magnets produce intense radiation from the infrared, through the visible and ultraviolet, and far into the X-ray region of the electromagnetic spectrum. The term, synchrotron light, is used to refer to radiation over this entire spectral range, not just to visible light. The use of the term X-ray light to refer to radiation in the X-ray region may seem unusual, but X-ray photons differ from visible light photons only in their energy.

2. Energy versus Wavelength X-ray energy is used more conveniently in much of the following discussion rather than X-ray wavelength, which is more familiar to diffractionists. Wavelength, A, in angstroms can be calculated from energy, E,in kilo-electron-volts (keV) by using the relation,

A

=

12.398IE.

For example, CuKa radiation (A = 1 S418 A), widely used for diffraction experiments, has an energy of 8.042 keV.

3. Definitions of X-Ray Intensity There are different ways of defining the intensity of X rays emanating from an X-ray source. The intensity units defined here use the Advanced Photon Source conventions (Rivers, 1990;Smith and Rivers, 1994)and are illustrated schemat-

10

D.G . SCHULZE AND P.M. BERTSCH

ically in Fig. 3. Other definitions are also in use; thus, caution must be exercised in comparing the definitions presented here with others from the literature. The flux is the number of photons per second per bandwidth per horizontal angle (0) integrated over the entire vertical angle (JI). Thus, flux has units of photons s-1 (0.1% bandwidth)-' mrad-1. Thus flux is relevant for experiments in which a large sample intercepts the entire beam in the vertical direction, such as a spectroscopy experiment using a bulk sample. Brightness is the flux per vertical angle (I+) or the number of photons per solid angle, i.e., photons s-1 (0.1 % bandwidth)-' mad-2. Brightness is the relevant intensity definition for experiments that use a collimator or pinhole to allow only a small spot of X rays to strike the sample, because brighmess is a measure of how many photons will pass through the pinhole. Brilliance is the brightness per source area, i.e., photons s-1 (0.1% bandwidth)-' mrad-2 mm-2. Brilliance is relevant to experiments that use mirrors to focus the entire X-ray source onto the sample because the smaller the original source of the X rays, the smaller the size of the focal spot on the sample. Thus, although two synchrotrons may have the same brightness, if the X rays emanate from a smaller source area in one synchrotron, it will have the greater brilliance. The source area is the area in space defined by the trajectory of the particle beam. The trajectory is not exactly the same for each particle in the ring, and the range of particle trajectories defines an envelope that defines the size of the particle beam. The particle beam typically is larger in the older synchrotrons. Thus, although first generation synchrotrons may produce very high flux, they may not produce particularly brilliant beams. The second generation, and in particular the third generation, storage rings were designed to produce the very tight particle trajectory necessary to produce X-ray beams with high brilliance. Rings with tight particle trajectories, which produce highbrightness, high-brilliance X-ray beams, are referred to as low emirrunce sources. Flux, brightness, and brilliance are a function of the design and operating parameters of a particular storage ring and cannot be altered by an individual user of the facility. The user, however, often has the option of choosing between one or more synchrotron facilities for a particular experiment. For example, a spectroscopy experiment that requires high flux to study low concentrations of an element in a relatively large bulk sample might be done on a more available first generation source. On the other hand, a spectroscopy experiment that requires spatially resolved data from very small areas of a heterogeneous sample will generally have to be done on a high-brilliance third generation source.

4. Energy Distribution of Bending Magnets The energy distribution of bending magnetic radiation from three synchrotron X-ray facilities in the US. is illustrated in Fig. 4 (note the log scale on the y-axis). As a point of reference, a conventional, Cu-target, sealed-tube X-ray

11

SYNCHROTRON X-RAYTECHNIQUES

0

20

40

60

80

100

X-ray Energy (kev)

ngUre 4. Brightness of throe U.S. synchrotron X-ray bending magnet sources. A P S , NSLS, and ALS refer to synchrotron soucces listed in Table 1 . BW,bandwidth. (Dataprovided by M. L. Rivers, University of Chicago.)

source used most by soils laboratories has a brightness of about 108 photons s-1 (0.1% bandwidth)-' mrad-2 (Kim, 1986), and this brightness is available only at an energy of 8.041 keV, the KCL,,2 emission line. What is immediately apparent is that synchrotron radiation from a bending magnet is 105-106 times brighter than that from a sealed-tube X-ray source and that this brightness occurs over a wide range of energies. The actual energy distribution is a function of the energy of the particles, in giga-electron-volts (GeV), and the field strength of the bending magnets (Rivers, 1990; Smith and Rivers, 1994). These parameters are a function of the design of the synchrotron and cannot be changed by the user. Note that, below 5 keV, the brightness of all three synchrotronsillustrated in Fig. 4 varies only by a factor of 10, but at higher energies, the difference becomes very large.

5. Energy Distribution of Insertion Devices Wigglers and undulators consist of multiple, alternating pairs of magnets (Fig. 5 ) that are inserted into the straight sections of the ring. Wigglers and undulators cause the particle beam to follow a sinusoidal path as it passes through the device; thus, the particles receive multiple centripetal accelerations as they pass through the device. a. Wigglers Wigglers are white light sources, like bending magnets. The brightness from a wiggler is like that from a bending magnet with the same magnetic field strength, multiplied by the number of magnetic poles in the device (Rivers, 1990; Smith

12

D. G. SCHULZE AND P. M. BERTSCH Direction of Magnetization

PerGanent Magnets

Hgure 5. X-ray generation by insertion devices. Particles traveling through the device receive multiple centripetal accelerations, resulting in the production of an X-ray beam more intense than that produced by a bending magnet.

and Rivers, 1994). Many new wigglers have 20-50 magnetic poles. The energy distribution of wiggler radiation from three U.S.synchrotron X-ray facilities is illustrated in Fig. 6 (again, note the log scale on the y-axis). Note the approximately 50-fold increase in brightness from a wiggler source compared to a bending magnet on the same storage ring (Fig. 4). Because the deflection of the particle beam by a wiggler can be relatively large, the resulting fan of X rays is often from 10 to 50 times wider than it is high, i.e., 8 is 10-50 times greater than $ in Fig. 3.

Figure 6. Brightness of wigglers on three U.S. synchrotron X-ray sources. APS, NSLS, and ALS refer to synchrotron sources listed in Table I. The brightness of two daerent NSLS wigglers, one on beam h e X17 and the other on beam line X25, is illustrated.BW, bandwidth. (Data provided

by M. L. Rivers, University of Chicago.)

SYNCHROTRON X-RAYTECHNIQUES

13

b. Undulators Undulators differ from wigglers in two ways. First, the distance between successive magnetic poles is smaller and the magnetic field is lower; thus, the deflection of the particle beam as it passes through the device is much less than that for a wiggler (the particles “undulate” rather than “wiggle”). Second, undulators often have about 3 times more magnetic poles than wigglers, up to around 150. Undulators are designed so that the X-ray waves starting at each undulation of the beam add up with the waves starting at successive undulations. The result is a constructive and destructive interference pattern, such that the smooth energy distribution characteristic of a bending magnet or wiggler is broken up into regions of very high intensity alternating with regions of lower intensity (Fig. 7). Note that the peaks of these energy distributions are very intense, with a brightness of up to 10’8 photons s-’ (0.1% bandwidth)-’ mrad-1. This is 10’0 times brighter than a conventional laboratory X-ray source! Thus, undulators are line sources that have distinct advantages for certain types of experiments. The position of the peaks in energy, however, are not fixed and can be varied by changing the gap between the magnet arrays. Rather than the fan-shaped beams of X rays produced by bending magnets and wigglers, undulators produce tightly collimated beams of X rays, making them the brightest, most brilliant X-ray sources available. In fact, the collimation of an undulator beam is typically better than that of a visible light laser. Figure 7 illustrates a spectrum from an undulator planned for the Advanced Photon Source. The smooth curves labeled n = 1, n = 3, and n = 5 illustrate the intensities and energy range available from the first, third, and fifth harmonics, respectively, by varying the undulator gap.

IQwe 7. Brightness of an APS undulator at a specific set of operating pamneters. The smooth curves labeled n = I , n = 3, and n = 5 illustrate the energy ranges accessible using the first, third, and fifth harmonics by varying the undulator operating parameters. (Data provided by M. L. Rivers, University of Chicago.)

14

D. G. SCHULZE AND P.M. BERTSCH

6. Polarization and Time Structure The radiation from bending magnets and insertion devices is highly polarized in the plane of the electron or positron orbit (Rivers, 1990; Smith and Rivers, 1994). The polarization can be exploited to reduce the scattered background and decrease the detection limit in X-ray fluorescence spectroscopy,or it can be used to design certain X-ray absorption spectroscopy (XAS) experiments. In the case of wigglers, the magnet arrays can be oriented at 90"to one another, resulting in circulatory polarized, rather than linearly polarized, light. The electrons or positrons are not distributed evenly around the storage ring, but are grouped into a number of bunches. X rays are produced only when a bunch of particles passes through the bending magnet or insertion device. The exact timing and duration of these X-ray pulses depend on the size and operating characteristics of the ring. For the APS operating with 20 bunches of positrons in the ring, the X-ray pulses will be about 80 ps long separated by intervals of 184 ns (Alp et al.. 1993a). This pulsed time structure is used to advantage in Mossbauer spectroscopy (Alp et al., 1993a,b). It is also possible to collect a difiaction pattern or an absorption spectrum with the X rays from a single bunch of particles, allowing the study of chemical reactions down to the picosecond time scale.

c. HARD VERSUS SOET X-RAY

SYNCHROTRONS

Synchrotron radiation is produced over a wide range of energies, from the infrared region with energies of <1 eV to the hard X-ray region with energies of 100 keV or more. The better the vacuum in the storage ring, the longer the particle beam is able to circulate and still produce an acceptable X-ray flux. Whenever possible, the storage ring vacuum is isolated from the beamline vacuum by beryllium (Be) windows, which are used to isolate various components of the beamline from one another. Although Be, with an atomic number of 4, is the least absorbing metal for this application, it does absorb low-energy radiation. The number and thickness of the Be windows in the beamline determine the low-energy cutoff of a particular beamline. Hard X-ray synchrotrons have most or all of the beamlines isolated from the storage ring vacuum by Be windows. In addition, whenever possible the instrumentationon a hard X-ray ring is designed to operate in air, but the N, 0, and Ar in the air also efficiently absorb and scatter low-energy radiation. Thus, beamlines on hard X-ray rings typically have lowenergy cutoffs between 2 and 5 keV. From a practical standpoint, elements from about Ca (Z = 20)and heavier in the periodic table can be studied by using X-ray absorption spectroscopy, with both the sample and instrumentation in air. Elements with Z < 20 can be studied with difficulty or not at all using a hard X-ray beamline. Sofs X-ray beamlines are designed to take advantageof the low-energy

SYNCHROTRON X-RAY TECHNIQUES

IS

synchrotron light, which is not available from a hard X-ray beamline. Soft X-ray beamlines generally have a direct vacuum connection to the storage ring, and the sample generally must also be placed in a high-vacuum environment. Thus, to perform X-ray absorption spectroscopy on an element like Si (Z = 14), one may have to go to an appropriately equipped soft X-ray beamline. Storage rings designed specifically for soft X-ray beamlines are designed for different operating parameters than those of a hard X-ray ring and, thus, produce a different energy spectrum. The Advanced Light Source and the Advanced Photon Source are sister machines in that the ALS is a third generation soft X-ray source, while the APS is a third generation hard X-ray source. Note that the brightness of the ALS bending magnets (Fig. 4) and wigglers (Fig. 7) drops off very rapidly compared to the APS and is more than lo3 times lower than the APS at 20 keV. The National Synchrotron Light Source actually consists of two storage rings, a soft X-ray ring called the vacuum ultraviolet or VUV ring and a hard X-ray ring.

D. Su-Y In summary, synchrotron light is extremely intense, is emitted over a wide range of energies, is highly collimated and highly polarized, and has a pulsed time structure. All of these properties make synchrotron light sources extremely useful for a wide variety of analytical techniques, many of which are only possible using a synchrotron source.

III. SYNCHROTRON X-RAY TECHNIQUES AND THEIR APPLICATIONS In the following section, we highlight the major synchrotron-basedtechniques with applications in the soil, plant, and environmental sciences. Some techniques are well established and widely used, e.g., X-ray absorption spectroscopy, X-ray diffraction, and the X-ray microprobe, and we only cite representative papers and point the interested reader to some of the many excellent reviews already in the literature. Other techniques, such as Mossbauer and infrared spectroscopies, are still being developed and there is little or no literature on direct applications to soil, plant, or environmental research. In this case we have given our assessment on possible applications in the agricultural sciences. What should become apparent, however, is that synchrotron-based research techniques have a broad range of applications in soil, plant, and environmental research. S ynchrotron-based techniques provide information at scales of measurement from angstroms (10- 10 m) to millimeters (10-3 m). The order of presentation in this section follows this general theme of increasing scales of measurement,

D. G. SCHULZE AND P.M.BERTSCH

16

beginning with spectroscopic techniques, which provide molecular scale information, and ending with computed tomography, which provides information up to the millimeter scale.

k

X-RAY hSORPTION SPECTROSCOPY

X-ray absorption spectroscopy(XAS) is a powerful technique that can provide detailed chemical and structural information about a specific absorbing element, whether it is a major component of a bulk solid phase (either crystalline or noncrystalline), a trace component of the bulk phase, a soluble species, or a surface-sorbed component. For most elements of the periodic table, XAS can provide specific information on the local environment of an absorber, including its coordination number and the identity of and distances to nearest, and sometimes next nearest, neighboring atoms. This type of information is not generally available from other techniques, particularly for poorly ordered mineral phases. Additionally, the energy of specific features around the absorption edge of an element often provides information on the oxidation state, coordination geometry, and transitions between core levels and partially occupied or continuum levels in the excited state. Perhaps the most attractive feature of XAS from the standpoint of soil chemistry and mineralogy and plant sciences is that, for elements with Z > 20, XAS can be conducted on wet samples, suspensions, and solutions under ambient conditions at absorber concentrations down to 100 Fg g-1 or less. It is precisely the atom-specific, noninvasive, in situ character and the ability to study such a wide range of samples that make XAS so attractive for investigating chemical speciation in the soil, plant, and environmental sciences. For almost all applications, XAS is a technique that requires a synchrotron X-ray source, as conventional sources do not produce sufficient X-ray intensity to obtain a spectrum with adequate signal to noise in a reasonable time period. There are a number of excellent reviews dealing with the principles of XAS and its application to problems in the biological, chemical, materials, and earth sciences (e.g., Sayers er al., 1971; Stem, 1974; Brown and Doniach, 1980; Calas et al., 1987; Brown et al., 1988; Hasnain, 1991; Charlet and Manceau, 1993; Fendorf et al., 1994b). For a detailed description of the principles and applications in other fields, readers are referred to these citations and the references therein. 1. Basic Principles of XAS

The attenuation of X rays by atoms of a given element varies smoothly with incident energy until a sharp increase in absorption occurs over a narrow energy range (Fig. 8). This narrow energy range, referred to as the absorption edge,

SYNCHROTRON X-RAYTECHNIQUES

17

-

I Pre-Edge Region

P M l w n W M

,....,.... .... ....,....,...., -50

0

50

160

l!h

200

250

-

300

1C 0

Relative Energy (ev)

Pre-Edge Region

XANES

EXAFS

1 . 2

Figure 8. Hypothetical XAS spectrum of a first row transition metal demonstrating the three major regions and the approximate relative energies normalized to the absorption edge (0 eV). The physical origins of the various structures axe represented at the bottom, where UMO represents unoccupied molecular orbital. The inset is a representation of the outgoing (fromthe central absorber) and backscattered (from neighboring atoms) photoelectric waves, whose constructive and destructive interference patterns give rise to the oscillations in the EXAFS region.

corresponds to the production of photoelectrons, the primary process responsible for the attenuation of X rays by matter over the 0.5- to 100-keV energy range. Photoelectron production represents the excitation of inner-core (K, L, or M) electrons by the incident X-ray photons to bound state unoccupied or continuum levels. This phenomenon occurs with high probability when the incident X-ray energy, E, is approximately equal to the binding energy of the core level elec-

18

D. G. SCHULZE AND P. M.BERTSCH

tron, Eb; thus, this is the basis for the elemental specificity of the technique. Once excited in this fashion, the absorber atom returns to the ground state through secondary processes such as X-ray fluorescence or Auger electron production. Most XAS experiments using hard X rays measure either X-ray attenuation (transmission) or fluorescence X-ray emission during data collection. The transmission mode generally is employed when the element under study is a primary constituent of the bulk phase under investigation, while the fluorescence yield mode generally is employed when the element of interest is a trace constitrient of the bulk phase or a surface-sorbed constituent. Experiments that utilize soft X rays generally cannot employ these detection modes because of the strong interaction of soft X rays with matter. Also, as a result of the large absorption cross section of soft X rays with air, these experiments are typically conducted under vacuum. Alternate detection modes for soft X-ray experiments include Auger electron and total electron yield. Since the mean free path of a 500-eV Auger electron is on the order of 20 A, the number of Auger electrons emitted is equal to the number of core holes created in the first 20 A from the surface (de Groot, 1991). Thus, as opposed to fluorescence yield, which has a probe depth of >loo0 A, Auger electron yield is a surface sensitive technique commonly employed to study low-Z elements at surfaces. Total electron yield, which is an extremely sensitive detection technique, measures all outgoing electrons, regardless of energy. The total electron yield signal appears to be dominated by secondary electrons created in the cascade process of the Auger decay electrons, although the probe depth of the technique is poorly defined, falling in the range of 20-200 A (de Groot, 1991). There are many details associated with the specific experimental conditions of an XAS experiment and even more are associated with complete data analysis required to extract the atomic level information desired. Although these are beyond the scope of this discussion, they are covered in detail in the very thorough reviews cited earlier and in a number of detailed monographs (e.g., Teo, 1986; Koningsberger and Prins, 1988). Fundamental to an XAS experiment is the monochromatization of the incident polychromatic X-ray beam. This is usually accomplished by rotating an appropriate crystal in the X-ray beam path in order to scan the energy range in the region around the absorption edge of the target element. ’I)pically, channel-cut or double crystals are used to monochromate the X-ray beam and to direct the outgoing monochromated X-ray beam in a fashion that is convenient for the experimental setup. The double crystal monochromator has the added advantage of allowing for a slight “detuning” such that most of the high-energy harmonics are rejected, a requirement in XAS spectroscopy. For the more traditional channel-cut single crystal monochromators, focusing mirrors are usually used to reject the higher energy harmonics and to increase the X-ray flux per unit area at

SYNCHROTRON X-RAYTECHNIQUES

19

the sample. In either case, the typical beam size of most traditional beam lines is in the centimeter to millimeter size range. a. The Preedge Region An X-ray absorption spectrum typically can be divided into three major energy regions, all of which have fundamentallydifferent physical origins (Fig. 8). The first region is to the immediate low-energy side (2-10 eV) of the main absorption feature, or “white line” (Le., where E C Eb), and is commonly referred to as the preedge region, although it has, by convention, been included with the second, or X-ray absorption near edge structure, region discussed in the following. Preedge absorption features, which are common for first row transition metals, are a result of transitions from a core level (e.g., Is, 2s) to empty or partially filled, bound excited states, primarily the nd molecular orbital state. The transition probabilities giving rise to this preedge region are governed by selection rules for dipolar electronic transitions. The relative intensity of the preedge feature, therefore, is related to the symmetry of the ligands around the absorber. For many first row transition metals, the preedge feature is intense when the metal coordination environment lacks an inversion center, i.e., tetrahedral coordination, such that d-p orbital mixing occurs, thereby providing allowed character to the otherwise forbidden transition. For a coordination environment having a central inversion center, i.e., octahedral coordination, d-p orbital mixing is at a minimum and this feature is very weak or nonexistent. Thus, the intensity and energy (central position) of the preedge absorption resonances provide information on the site geometry of the absorber, which is commonly related to the oxidation state. For example, Cr oxidation states can be deduced from the presence or absence of a predominant preedge feature that is characteristic of Cr(VI), which is tetrahedrally coordinated, but is nearly absent for Cr(III), which is octahedrally coordinated. b. The XANES or NEXAFS Region The second energy region extends from just beyond the preedge to approximately 50 eV above the absorption edge (i.e., where E Eb) and is referred to as the X-ray absorption near edge structure (XANES) or the near edge X-ray absorption fine structure (NEXAFS). This energy region is usually characterized by very intense resonance features, arising from electron transitions to unoccupied bound state and continuum levels and from multiple scattering of the emitted photoelectrons by atoms surrounding the absorber. The location or energy of the absorption edge is usually defined as the inflection point (precisely defined as the first derivative) or half-height on the edge step, which is the ascending limb of the main absorption feature (Fig. 8). Additional means for defining the edge position include the identification of some clearly discernible

-

20

D. G. SCHULZE AND P. M. BERTSCH

feature, such as the crest of the main absorption peak. As will be discussed in the following, this assignment tends to be more matrix dependent, as the broadness of the white line is generally more sensitive to bonding environments. The exact energy of both the preedge and absorptionedge features is related to the chemical environment of the absorber, such as oxidation state. The energies of the preedge feature and the absorption edge typically are increased by -1-3 eV for each electron withdrawn from the valence shell. This “chemical shift” is related to the decreased shielding of core electrons with increasing valence, resulting in increased binding energy of core levels. In many instances the edge position varies regularly with the oxidation state, such that semi-quantitativeestimates of metals in different oxidation states in mixed valence compounds can be derived. The multiple scattering resonances to the high-energy side of the main absorption feature are related to the geometrical arrangement of the first and more distant neighboring atoms around the central absorber. Although an incomplete understanding of the physical processes involved in the multiple scattering region has hampered the ability to derive specific bonding information from a XANES spectrum, comparisons between the “fingerprint”region of spectra from known compounds and unknown samples allow important qualitative information on bonding environmentsto be deduced. Additionally, interatomicdistances from the central absorber to surrounding atoms can be estimated from the energy positions of the multiple scattering resonances in a XANES spectrum according to

where R is the interatomic distance, c is a constant, Em,, is the energy of the multiple scattering resonance, and Eb is the energy of the bound state transition (Bianconi, 1988). Although the interatomic distances determined by this approach are less accurate than those extracted from a detailed analysis of the extended portion of the X A S spectrum (see the following), they are, nevertheless, useful for comparing the coordination environments of a given central absorbing atom in different chemical forms. c. The EXAFS Region The third energy region extends from approximately50 eV to as much as lo00 eV or greater above the absorption edge (i.e., where E > Eb) and is termed the extended X-ray absorption fine structure (EXAFS). This region is represented by low probability of electronic transitions and a very short resonance time before the excited photoelectrons leave the region of the atom from which they were emitted. The frequency oscillations in this region arise from constructive and destructive interference patterns between the outgoing and the returning photoelectric wave that has been backscattered from first and sometimes second shell neighboring atoms (Fig. 8). The frequency of the oscillations is inversely related

SYNCHROTRON X-RAYTECHNIQUES

21

to the bond distance between the absorber and neighboring atoms, sometimes extending out to several shells of ligands. The amplitude of these oscillations is related to both the identity and number of atoms surrounding the central absorber. Because the physical processes giving rise to the EXAFS oscillations are reasonably well understood and can be modeled by using a single-scattering, plane wave approximation, data analyses of an EXAFS spectrum can provide, under ideal conditions, the identity of the surrounding ligands, specific bond distances to within 20.02 A, and coordination numbers of first and second shell ligands to within 5%. The data generated in an XAS experiment represent a weighted average of all possible coordination environments experienced by the element of interest. Therefore, “ideal conditions” require that the system studied represent an isolated homogeneous phase in which the element under investigation experiences relatively small structural or static disorder and from which other spectral interferences in the EXAFS region are absent. Many applications in the soil, environmental, and plant sciences do not meet these criteria, making detailed structural analysis difficult. Nevertheless, both XANES and EXAFS spectroscopic techniques are powerful tools for investigators in the soil, environmental, and plant sciences interested in specific information regarding the chemical speciation of a wide range of elements.

2. Applications of XAS in the Soil,Plant, and Environmental Sciences a. XAS of Minerals and Other Inorganic Solids Applications of both XANES and EXAFS spectroscopy to the elucidation of cation environments in minerals and glasses began appearing in the earth sciences literature in the early 1980s, and this pioneering work has been reviewed extensively (e.g., Calas et al., 1987; Brown et al., 1988; Charlet and Manceau, 1993). More recently, numerous studies have demonstrated the usefulness of XANES and EXAFS for providing structural information on important noncrystalline components in soils (Combes et al., 1986, 1989, 1990; Manceau et al., 1990, 1992b,c; Manceau and Drits, 1993; Drits et al., 1994; Ildefonse et al., 1994; Li et al., 1994) and specific chemical speciation information on contaminants associated with sorptive phases, including soils (Hayes et al., 1987; Chisholm-Brause et al., 1989a,b, 1990a,b; Roe et al., 1991; Charlet and Manceau, 1992a,b; Dent et al., 1992; Manceau and Charlet, 1992; Bidoglio et al., 1993; Waychunas et al., 1993; Fendorfet al., 1994a,b; Bertsch et al., 1994a,b; O’Day et al., 1994a,b). Poorly ordered hydrous oxides of iron, manganese, and aluminosilicates are important high surface area phases in soils and sediments that can control the solution concentrations of many nutrients and contaminants. These phases have

22

D. G. SCHULZE AND P.M. BERTSCH

been difficult to study using traditional X-ray diffraction techniques since the very small crystallite size results in little coherent scattering and, therefore, only a few poorly defined X-ray reflections. Other XANES and EXAFS investigations have provided information on the coordination of Fe and Mn in these poorly ordered phases and have demonstrated the importance of hydrolysis conditions on the distinct local structure formed in the initial stages of nucleation (Combes et af.,1986, 1989, 1990; Manceau et af., 1990, 1992b,c). It has been demonstrated that Fe and Mn always occupy octahedral sites within these phases, in contrast to the suggestion that 6-line ferrihydrite contains a significant population of tetrahedral Fe. These studies have also indicated that Fe may have as many as five distinct local structures that resemble common oxyhydroxide phases, such as lepidocrocite and goethite, all of which are highly dependent on specific synthesis conditions. Furthermore, an A1 XANES spectroscopic investigation of imogolite and allophane has suggested that A1 may occupy only a single octahedral site in these poorly ordered phases that more closely resembles AlV’ in dioctahedral smectite than the two distinct AlVI sites of gibbsite or kaolinite, which previously have been considered structural analogues (Ildefonse et af., 1994). It is clear that XAS can provide unique structural information on the local environments of important cations comprising the ubiquitous poorly ordered phases in soils. Future XAS investigations of such phases and their precursors undoubtedly will provide invaluable information regarding their local structure and the mechanisms controlling their formation and transformation under various hydrolysis regimes and in a variety of weathering environments. b. XAS of Surface-Sorbed Species The ability to develop adequate models for predicting the fate of nutrients and contaminants in both surface and subsurface environments is highly dependant on an accurate description of the partitioning of these constituents between the solid and solution phases and ultimately to the capability to provide specific molecular level information on species distribution in both of these phases. As a result, the partitioning of constituents to sorbate phases in soils, or “sorption,” has been an area of intense research in the soil and environmental sciences. Many investigations have been concerned with the specific orientation and/or structure of adsorbates at the aqueous-mineral interface, although many of the hypotheses proposed concerning the microscopic properties of adsorbates have arisen largely from macroscopic measurements. The underlying mechanisms controlling the removal of a chemical species from the aqueous phase to a contiguous solid phase (sorption) include (Sposito, 1986): (1) precipitation, the growth of a solid phase composed of a molecular unit that repeats itself in three dimensions; (2) adsorption, an accumulation of a chemical species at the aqueous-mineral interface without the formation of a three-dimensional molecular unit; and (3) absorption, the diffusion of a chemical species from the aqueous phase into the structure

SYNCHROTRON X - W TECHNIQUES

23

Flgure 9. Representation of a mineral surface and (a) an outer-sphere surface complex where

H,O molecules are interposed between the sorbate and the surface; (b) an her-sphere surface complex where the ligated water molecules have been replaced by surface functional group; (c) a multinuclear surface complex; (d) a surface precipitate where the sorbate is arranged in a threedimensional network; (e) absorption, or solid state diffusion and substitution of the sorbate in the mineral structure; and (f) occlusion of the multinuclear surface complex within the structure of the mineral phase.

of a solid phase (Fig. 9). Additional complexity surrounds specific structural orientations and the chemical speciation of adsorbates. For example, critical information about adsorbed metals and metalloids includes the degree of innerand outer-sphere association (Fig. 9). For constituents that form inner-sphere complexes with surface functional groups, it is important to elucidate their specific molecular arrangement, such as monodentate and/or bidentate orientation. Finally, to fully understand the surface interaction and to adequately predict the behavior of a surface-sorbed constituent, it is critical to have information on its chemical speciation, such as the presence of multinuclear units or ternary complexes. Macroscopic investigations generally do not provide information to distinguish between the primary sorption mechanism(s) or to infer the additional molecular level information associated with surface adsorption (Sposito, 1986). While a number of spectroscopic methods can be used to interrogate the surface environment of monomineralic systems via probe molecules and reporter groups (Motschi, 1987; Johnston et al., 1993), few are as versatile or provide specific molecular level information as a well-planned and -executed XAS experiment. Many surface spectroscopic techniques require high concentrations of the adsorbate or the use of dry samples, often under vacuum. As described previously, XAS generally does not require such sample manipulation, and investigations have demonstrated the ability of XAS to elucidate removal mechanisms, i.e., distinguish adsorption from precipitation for a wide range of metals and metalloids sorbed to important mineral phases commonly found in soils (Hayes

24

D. G. SCHULZE AND P. M.BERTSCH

et al., 1987; Brown et al., 1989; Chisholm-Brause et al., 1990a,b; Roe et al., 1991; Charlet and Manceau, 1992a,b; Combes et al., 1992; Dent et al., 1992; Manceau and Charlet, 1992; Manceau et al., 1992a,b; Bidoglio et al., 1992, 1993; Waychunas et al., 1993; Fendorf et al., 1994a; O’Day et al., 1994a,b).

Many of these studies have also demonstrated the ability of XAS to provide specific microscopic information on the adsorption mechanism, i.e., distinguish between inner- and outer-sphere surface association, to provide direct evidence on surface orientation, i.e., to suggest mono- or bidentate surface complexation, and to provide information on the chemical speciation of the surface-sorbed component, including oxidation state, multinuclearity, and ternary complexation. In an early application of XAS to probe the local structure of the mineralwater interface, Hayes et al. (1987) examined the sorption of the Se oxyanions, selenite (SeO,*-) and selenate (Se0,2-), to the ubiquitous soil mineral goethite. Detailed analyses of the EXAFS region of the derived spectra indicate that the selenite is adsorbed primarily as an inner-sphere complex, on the basis of the evidence for the presence of three 0 atoms in the primary shell at 1.70 8, and two Fe atoms as second shell backscatterers at 3.38 8,, which is compelling evidence for the formation of binuclear (with respect to Fe), bidentate complexes with surface functional groups. In contrast, the EXAFS spectrum of sorbed selenate provided evidence for four 0 atoms in the first shell at 1.65 8, and no evidence for a significant contribution in the second shell, suggesting a primarily outersphere association. Although this study was quite informative, it must be kept in mind that the results of an XAS experiment represent a population-weighted average of all possible environments experienced by the adsorbate. Thus, to derive meaningful results over a range of conditions required to properly assess environmental significance, it is important to examine sorptive phases over a complete domain of surface coverages. It is also critical to ensure that only sorbed constituents remain in the suspension preparation (i.e., excess soluble constituents are properly washed from the suspension) and, in most instances, to combine the results of XAS experiments with extensive macroscopic experimentation (e.g., adsorption-desorption behavior and exchangeability) and, perhaps, other spectroscopic evidence. The need for careful sample preparation and diverse experimental approaches has been supported in some EXAFS investigations examining metal and metalloid sorption to oxides and aluminosilicates. A comprehensive EXAFS examination of arsenate adsorption to the poorly ordered iron oxyhydroxide phase ferrihydrite and to more crystalline goethite revealed the predominant formation of a binuclear, bidentate arsenate complex with goethite over all surface coverages examined (Waychunas el al., 1993). In contrast, arsenate adsorption to ferrihydrite was found to consist of both mono- and bidentate complexation at lower surface coverages, with bidentate and possibly some tridentate complexes pre-

SYNCHROTRONX-RAY TECHNIQUES

2s

dominating at higher surface coverages. No evidence for an As-rich surface precipitate or ferric arsenate was observed, even at the highest of surface coverages examined. Another detailed EXAFS investigation that examined Coz+ sorption to kaolinite also suggested the existence of both inner- and outer-sphere populations of the sorbate (O’Day et al., 1994a,b). At low surface coverages the inner-sphere population predominated, with a proposed bidentate Co octahedral complex, comer-shared with A1 and Si polyhedral units composed of edge, nonbridging Al-OH inner hydroxyl sites and edge Al-0-Si bridging oxygen sites. With increasing surface coverage, the geometrical arrangement of the predominant inner-sphere complex resembled edge-shared Co-A1 octahedra rather than comer-shared. Another significant finding in this investigation was the evidence for second neighbor Co atoms at 3.10-3.13 8, indicating the presence of multinuclear Co complexes, which formed at <5% monolayer coverage. The multinuclear Co species observed have geometry similar to that of the edge-shared hydroxy-bridged Co(OH), solid phase, having somewhat shorter Co-Co distances. Other EXAFS investigations of sorbed transition metals have also indicated the formation of multinuclear complexes on surfaces at concentrations well below the solubility product of known phases (Chisholm-Brause et d ,1990a,b; Charlet and Manceau, 1992a,b; Dent et al., 1992; Fendorfer al., 1994a). Charlet and Manceau (1992a,b) reported the formation of a Cr(II1) polynuclear complex on the surface of goethite at
26

D. G. SCHULZE AND P. M. BEWSCH

characteristic of the CrO,*- ion. The oxidation was also accompanied by desorption of the Cr(V1) and release of Mn(I1) into solution. Similarly, XANES investigations of Tl(1) and Ce(II1) sorption to birnessite have revealed surface-induced oxidation to Tl(II1) and Ce(IV) (Bidoglio et al., 1992, 1993). Additionally, a XANES investigation of Cr(V1) sorption to a laboratory-prepared goethite containing 0.3 wt% Fe(II), revealed partial reduction to Cr(II1) following adsorption, as evidenced by the reduction in amplitude of the preedge feature compared to that of the main absorption feature (Bidoglio et af., 1993). It was suggested that the structural ferrous iron was the source of electrons for the surface-facilitated redox reaction. d. XAS of Biomolecules Other applications of XAS relevant to plant science and soil and environmental microbiology have included XANES and EXAFS investigations of metal coordination environments in a number of important proteins. EXAFS at the FeKa edge has been used extensively to provide direct information on coordination environments and structural conformations of Fe-S clusters of electron transfer proteins, such as rubredoxin, ferredoxin, and aconitase (Cramer et al., 1978). Other Fe-containing proteins that have been investigated extensively with XAS include hemoglobin, ferritin, and transferrin (Lindley, 1991). A preliminary Fe XANES and EXAFS investigation of the native form of soybean lipoxygenase has revealed Fe(II) with a coordination sphere of the active site having 6 f 1 N/O ligands at 2.05-2.09 A, suggesting a structure consisting of 4 imidazole nitrogen atoms at longer distances and 2 oxygens (probably carboxylate) at shorter distances from the central iron atom (Van der Heijdt et af., 1991). The yellow substrate-treated form of the enzyme revealed an Fe(II1) center ligated by three imidazole nitrogens and three oxygens, suggesting the loss of one imidazole and the gain of one 0 ligand of unknown origin during activation of the enzyme by product peroxidase (Van der Heijdt et al., 1991). Other important metal-containing biomolecules or biological materials that have been investigated by XAS have included methane monooxygense (Fe) from methylococcus casulatus (DeWitt et af., 1991), purple acid phosphatase (Fe and Zn) from red kidney bean (Priggemeyer et al., 1991), spinach plastocyanin and rusticyanin (Cu) from Thiobacillusferrooxidans (Murphy et al., 1991), Mn environments in photosystem I1 (Kusunoki et af., 1991; Corrie et al., 1991), a number of Ca*+binding proteins (Korystova et al., 1991), metal-binding centers in DNA (Povey et al., 1991), nitrogenase (Mo and Fe) (Chen et af., 1993a,b), and Ca environments in milk (Irlam er af., 1985) and bone (Harris er al., 1988), to name just a few. e. XAS of Soluble Metal Complexes The chemical speciation of soluble constituents also has widespread applications in the plant, soil, and environmental sciences. XAS has been used to

SYNCHROTRON X-RAY TECHNIQUES

27

investigate soluble chlorine-metal complexes (Sandstrom, 1984; Farges et al., 1994; Brown et af., 1995), and an EXAFS investigation of Cs+ crown ether complexes, which have importance in nuclear waste isolation, has revealed a shell of 6 oxygen atoms at 3.03 A, a shell of 12 carbon atoms at 3.96 A,and 1 Br atom at 3.64 A, confirming nuclear magnetic resonance data that suggested ion pair formation between the crown ether-complexed Cs+ and B r (Kenner et al., 1995). These studies demonstrate the great utility of both EXAFS and XANES for such applications, and there will undoubtedly be many more such applications in the future.

f. XAS of Soils and Other Heterogeneous Materials Most of the initial EXAFS and XANES experiments of biomolecules and surface-sorbed components have provided specific structural and chemical speciation information not available with other techniques and have demonstrated the great potential of XAS applied to a wide range of problems in plant biochemistry and physiology and in soil and environmental chemistry. These studies, however, have largely dealt with relatively simple isolated biomolecules or, in the case of surface-sorbed components, monomineralic systems. As described earlier, the complex heterogeneous assemblage of surface functional groups and ligands and the large array of elements in plants and soils make a comprehensive EXAFS analysis of a trace component or a surface-sorbed species difficult. One strategy proposed to study coordination environments in heterogeneous soil systems is to carefully characterize the interactions of a specific absorber with individual mineral phases and gradually increase the complexity (Fendorf, et al., 1994b). XANES spectroscopy is more promising in this regard, especially if average oxidation state information or qualitative information on coordination environments is sufficient. For example, extensive S XANES experiments have been conducted on coals and petroleum products, as well as on petroleum source rock extracts (Huffmanet af., 1991; Huggins et al., 1991; Waldo et al., 1991). These studies demonstrate the utility of S K-edge XANES spectroscopy for characterizing the oxidation states and organic S functionality in coal and for following S speciation changes during desulfurization processing. Tokunaga et af. (1994a) investigated oxidation states of Se in sediments from the Kesterson Reservoir in California and followed transformations of Se in laboratory microcosms following flooding. In the highly contaminated reservoir sediments, XANES revealed that the Se was present primarily in the monoclinic elemental form. The laboratory microcosm experiments where Se(V1) and organic amendments were made to ponded soils confirmed that Se(V1) was reduced to Se(1V) and then to Se(0) within 4 days of ponding. The reoxidation dynamics of Se was also monitored via XANES spectroscopy. Cotter-Howells et af. (1994) used XAS data collected at the Pb L(III) edge along with X-ray diffraction and analytical electron microscopy data to identify pyromorphite as the major lead-bearing

28

D. G. SCHULZE AND P. M. BERTSCH

phase in some mine-waste soils. XAS spectra alone were found to be sufficiently unique to distinguish pyromorphite from other possible Pb-bearing phases. g. Spatially Resolved X A S Another strategy for applying both EXAFS and XANES spectroscopies to complex matrices, such as soils, plant tissues, and other biological samples, is to exploit the heterogeneity of elemental distributions characterized by regions of isolated phases. Such regions often have local elemental concentrations that are much higher than the average of the bulk material and may be represented by a more limited population of coordination environments. Spatially resolved XAS techniques provide the opportunity to employ such a sampling strategy, and this approach has been successfully applied to a number of systems. Among the first applications of spatially resolved or micro-XANEStechniques are the determination of Cr oxidation states in the 50-200-pm regions within olivine grains from lunar basalt (Sutton et al., 1993), contaminated soils, and simulated nuclear waste forms (Bajt et al., 1993), the determination of U oxidation states in 50300-pm regions within contaminated soils and sediments (Bertsch et al., 1994a,b), the determination of Fe oxidation states in silicates and oxides (Bajt et al., 1994). the determination Cr and Se oxidation states in soils (Tokunaga et al., 1994b), the determination of speciation in localized regions of hyperaccumulating aquatic plants (Hunter et al., 1994), the determination of Ni, Cr, and Se speciation in the bone (apatite) and protein (keratin) components of turtle shell fragments (Bertsch et al., 1994a), and the determination of Mn oxidation states in the rhizosphere of living plants (Schulze et al., 1995a). All of these studies have been conducted using beamline X-26A at the NSLS, a dedicated microprobe beamline originally configured as a white light microanalytical X-ray fluorescence instrument for trace elemental analysis (see the following). Use of the basic microprobe configuration with the addition of a monochromator has provided researchers with the opportunity to examine either small samples or small isolated regions within a sample by XAS. As will be described in the following, elemental distributions can be determined quickly and mapped out for samples prior to examination by the micro-XANES technique. h. Quantification of Oxidation States Using XAS Quantitative methods for the determinationof oxidation states and mixtures of oxidation states for Cr (Bajt et al., 1993), U (Bertsch et al., 1994a,b), Mn (Schulze et al., 1995b). and Fe (Bajt et al., 1994) have been developed, and others are currently under development. The method for the quantitative determination of Cr(V1) in glasses, minerals, cementitious waste forms, and plant tissue is based on the area of the preedge peak normalized to the average intensity in the continuum, i.e., 400-500 eV above the edge (Bajt et al., 1993; Bertsch et al., 1994b; Hunter et al., 1994) (Fig. 10). By using this method, it is possible to

SYNCHROTRON X-RAY TECHNIQUES

29

-20

-10

0 10 20 Relative Energy (eV)

30

40

Flgure 10. Mixtures of Cr(II1) and Cr(VI) standards demonstrating the regular increase in the intensity of the preedge feature with increasing Cr(V1) that can be normalized against an above edge feature in the continuum to provide quantitative estimates of the Cr(V1) content of a sample. (Unpublished data provided by Dr. Douglas Hunter.)

determine the fraction of Cr (VI) in samples having a total Cr concentration of 10 pg g-1 within 5% accuracy with little or no sample manipulation. The methods developed for U, Fe, and Mn are based on the regular shifts in the energy or intensity of the edge position with oxidation state. For example, the method developed for the determination of the fraction of U(V1) and U(1V) in a sample demonstrated that the relative position of the edge, defined as the first derivative of the edge jump, varied regularly from 0 eV for 100% U(1V) to 4.45 eV for 100% U(VI), with various proportional mixtures of the end members displaying a linear relationship with the shift in the edge energies (Bertsch et al., 1994b) (Fig. 11). The method was found to be useful for examining contaminated soils and chemically extracted soils in regions between 50 and 200 pm at bulk U concentrations as low as 1 pg g-1, also within 5% accuracy. The slope of the edge step and, thus, the edge position as defined in this study was found to be matrix independent, whereas the width of the white line was found to vary with matrix composition. Some earlier investigations on U-containing phases, including glasses, defined the edge position as the crest of the white line and reported only qualitative relationships between the edge position and oxidation state, a relationship probably influenced by the various matrices examined. By using the micro-XANES technique, evidence for regions within contaminated sediments composed of pure U(IV) and U(V1) was discernible; there was also evidence for regions having U in mixed oxidation states. Other qualitative information on

D.G. SCHULZE AND P.M. BERTSCH

30

o . O I ~ . . " . " " ' ' ' . ' ' ' ' . l -20

0

20

40

60

80

Relative Energy (ev)

Figure 11. Mixtures of U(IV) and U(V1) standards demonstrating the regular shift in the edge position with valence and additional diagnostic features of U(W) in the multiple scattering region of the spectra (modified from Bertsch er al., 1994b).

U-bonding environments can also be extracted from the multiple scattering resonances on the high-energy side of the main absorption feature, which are related to the local geometry around the central absorbing atom. The multiple scattering resonance to the immediate high-energy side of the main absorption feature is distinct for U(V1)-containing phases, representing multiple scattering associated with the axial U-0 bonds, whereas the second multiple scattering resonance at still higher energy is related to multiple scattering associated with the equatorial U-0 bonds. The central position of the multiple scattering resonances is inversely related to the U-0 bond distance, allowing qualitative information to be extracted by comparing the "fingerprint" regions of XANES spectra taken for various samples and reference phases (Farges er al., 1992). Another application of the micro-XANES technique involved the elucidation of Mn oxidation states in the vicinity of wheat (Triricum aestivum L.) roots infected with the fungus Gaeumunnomyces graminis (Sacc.) var. rririci (Ggt), which is responsible for the take-all disease. Plants infected with Ggt are often Mn deficient, which has been hypothesized to be related to fungus-induced oxidation of Mn from the soluble Mn*+ species to the Mn(IV) species, which forms insoluble oxyhydroxide mineral phases. By using micro-XANES at the

SYNCHROTRON X-RAY TECHNIQUES

31

MnKa edge, Schulze et al. (1995a) tested this hypothesis by examining the rhizosphere of wheat plants grown on agar amended with 50 pg-1 Mn2+ and infected with Ggt. Manganese in the clear agar was found to remain in the Mn2+ form, while Mn in darkened regions of roots infected with Ggt was predominantly in the Mn(IV) oxidation state, present as discrete oxyhydmxide precipitates. This study represents the first unambiguous test of the fungus-induced oxidation hypothesis and further demonstratesthe utility of the micro-XAS technique to examine plant, soil, or other environmental samples in siru or with little manipulation or sample pretreatment. Investigations in progress are the examination of Fe and Mn oxidation states in the rhizosphere of plants growing in soil media and in thin sections of soils around zones of active oxidation-reduction activity. In addition to oxidation state determinations, other qualitative information on the binding environments of these and other metals is also available by the analysis of additional features in the XANES spectra, as discussed previously for U. The microprobe capabilities on X-26A at NSLS also allow the determination of complete elemental distributions in the regions examined by micro-XAS, thus providing additional information for inferring the chemical speciation of metals and metalloids in soils or other environmental samples on the basis of coassociated elements (Bertsch et al., 1994a,b; Hunter et al., 1994; see the following).

i. Oxidation State Mapping In addition to the determination of oxidation states of metals and metalloids on 10-pm regions within a sample, it is also possible to generate oxidation state maps that are analogous and comparable to the elemental maps generated with the X-ray microprobe by using white light (see the following). The method involves scanning a sample in two dimensions with the monochromator set to the appropriate energy for the edge of one oxidation state, then scanning the identical points with the monochromator position set at the energy where the intensities of both oxidation states are observed, and then generating an appropriate ratio of the intensities at both monochromator positions. The method is calibrated with pure end member and mixed oxidation state standards and can be verified on actual samples by taking full XANES spectra on regions demonstrated to be composed of predominantly one or another oxidation state (Sutton et al., 1995a). This method was employed to map Mn oxidation states in the rhizosphere of the wheat plants previously described in the take-all experiment (Schulze et al., 1995a)and has been used to map U oxidation states on identical 50-pm regions in contaminated soils both before and after implementation of a chemical intervention remediation strategy (Bertsch et al., 1994a). Combination of both elemental and oxidation state maps of the hyperaccumulating aquatic plant Salvinia rotundiJlora following exposure to chromate and selenate demonstrated the heterogeneous distribution of Se and Cr in the plant tissue, with Se being concentrated within

32

D. G. SCHULZE AND P. M. BERTSCH

the nuclei of cells and Cr in the cytoplasm, probably associated with vacuoles. Micro-XANES measurements taken on these same regions demonstrated the total loss of the Cr(V1) diagnostic preedge feature in the XANES spectrum, indicating the complete reduction of the Cr(V1) to Cr(1II) after uptake by the plants. A shift of the Se absorption edge to lower energies was observed for the Se in tissues of Safviniurotundflora, suggesting partial reduction of the accumulated Se(V1) to Se(-11) and incorporation into proteins as cysteine type groups following uptake (Bertsch et al., 1994a; Hunter et al., 1994). j. XAS of Low-Z-Elements Using the Scanning Transmission X-Ray Microscope Another novel and exciting application of micro-XANES involves the use of the scanning transmission X-ray microscope (STXM), which utilizes a soft X-ray undulator at NSLS beam line X-1A operating in the 250-750-eV energy range. In the imaging mode, the X-ray energy is fixed and the sample is scanned in the focal plane of the focusing device, a Fresnel zone plate, to generate a 2D image with spatial resolution on the order of 55 nm (Ade et a f . , 1992; Ade and Hsiao, 1993; Zhang et af., 1994). The device has been configured so that a XANES spectrum can be collected in a region as small as 0.2 X 0.2 p m by scanning the monochromator over an energy range while simultaneously adjusting the distance of the zone plate from the sample to maintain the focus. The technique has been used to collect C XANES on a number of polymers and biological samples, including proteins and DNA (Zhang et a f . , 1994). The results have demonstrated the ability of the micro-XANES technique to distinguish various functional groups such as C - C , C 4 , m,C=O, COOH, and C=N. On the basis of the relative proportions of C = C , C=N, and C = O functional groups within a biological molecule, differences in resonance intensities were observed, as was the energy of the resonance for a given chemical bond as a result of the different chemical environment within a specific molecule (Ade and Hsiao, 1993). The chemical shift in the C resonances between protein and DNA, for example, has been exploited to provide chemical contrast images of chromosomes and is currently being used to develop quantitative mapping of DNA and protein within biological samples (Ade and Hsiao, 1993;Zhang et a f . , 1994). In principle, other elements with K or L edges in the operating range of this beam line (250-750 eV), such as 0, N, K, Ca, Cr, Mn, and Fe, should be accessible with this device, and it is anticipated that applications of chemical or structural imaging will become powerful tools in the future. Furthermore, the polarization dependence of XANES and the resulting X-ray dichism can be exploited to image bond orientations in ordered and partially ordered systems.

k. Future Developments in XAS There are a number of other developmentsin XAS and spatially resolved XAS that will greatly facilitate future applications in the soil, plant, and environmental

SYNCHROTRON X-RAY TECHNIQUES

33

sciences. The quick EXAFS technique, or QEXAFS, relies on a constant monochromator scan rate and fast array detectors to collect a full EXAFS spectrum in a few seconds versus tens of minutes in a conventional experiment. In principle, millisecond or microsecond time scales are achievable (Lytle and Greegor, 1991; Frahm, 1991; Dobson, 1994). This technique will greatly facilitate the collection of data for elements that are present in low concentrations and will allow the design of experiments in which dynamic processes can be examined in siru. Clausen and Topsm (1994) describe an experimental arrangement in which a QEXAFS spectrum and an X-ray diffraction pattern can be obtained simultaneously from the same sample, allowing one to obtain both short range (EXAFS) and longer range (XRD) data on the sample. Another technique utilized for time-resolved XAS is energy-dispersive EXAFS or DEXAFS, which employs a bent triangular monochromator crystal and a large area photodiode array detector, which allows parallel acquisition of a full EXAFS spectrum in a small fraction of a second (Baker et al., 1991). The disadvantage of this approach compared to QEXAFS is that detection is limited only to the transmission mode. Another surface-sensitive EXAFS technique, which does not require the ultra-high vacuum environment of surface EXAFS, is glancing angle EXAFS or REFLEXAFS. In this technique, the sample is oriented in the incident beam at an angle that is smaller than the critical angle for total reflection. By varying the incidence angle, EXAFS spectra can be obtained from a few nanometers in depth to a total depth approaching the reciprocal of the absorption coefficient, which can be on the order of several millimeters (Greaves er af., 1991). This technique should have a number of applications in the soil and environmental sciences where information on surface-sorbed species is desired. Another technique that may hold promise for applications in the soil and environmental sciences is diffraction anomalous fine structure (DAFS) (Stragier er al., 1992; Finkelstein and Sutton, 1994; Lee et af., 1994). In this technique, XANES and EXAFS are collected at the Bragg diffraction angle of a given phase, thus providing XAS data specific to the diffracting phase. This measurement scheme may be particularly useful when operating in the microprobe mode especially if the heterogeneous sample contains discrete, yet multiple phases of a given element. Perhaps the most powerful technique for examining heterogeneous plant, soil, and other environmental samples is the spatially resolved XAS described earlier. One of the requirements for the successful application of spatially resolved or micro-XAS is that the X-ray beam size is on the order of, or smaller than, the heterogeneous mosaic of the element of interest within the sample. Current microbeam optics for hard X-ray sources include a simple pinhole collimator, which wastes photons and, therefore, reduces the potential flux at the sample. As described previously, Fresnal zone plates are used to focus soft X-rays and have effectively delivered submicrometer spot diameters for use in XANES spectroscopy. Other focusing devices include ellipsoidal mirrors, double elliptical Kirkpatrick-Baez mirrors, and tapered capillary concentrators. The double ellip-

34

D. G . SCHULZE AND P. M. BERTSCH

tical Kirkpatrick-Baez mirrors have achieved spot diameters as small as 2 Km, while tapered capillary concentrators have achieved submicrometer spot diameters. Both devices are under active research and development (Thiel et al., 1993; Bilderback et al., 1994; Hoffman et al., 1994; Ullrich et al., 1994; Sutton et al., 1995b; Yang ef al., 1995). In addition to producing micrometer or submicrometer X-ray beams, these devices increase the flux per unit area by a factor of lo2103 at existing second generation synchrotrons. This translates into an enhancement factor of 105-106 at the third generation sources, such as the APS, greatly facilitating applications to the complex samples typically encountered in the plant, soil, and environmental sciences. Active research and development efforts in the area of spatially resolved EXAFS on the micrometer scale are currently under way, but will require additional advances in a number of areas, including stabilization of the beam position during the scan and implementation of active vibration control of the experimental setup. Continuing advances in monochromators, X-ray-focusing devices, and detectors, coupled with the availability of the increased brilliance of third generation sources, will undoubtedly result in novel applications of XAS in the plant, soil, and environmental sciences that will ultimately define new standards in the determination of chemical speciation in complex environmental samples.

B. SYNCHROTRON X-RAYFLUORFSCENCE SPECTROSCOPY Synchrotron X-ray fluorescence spectroscopy (SXRF)is a rapid, nondestructive technique for the quantitative determination of elements (Z > 20) in a wide variety of samples in the parts per million to parts per billion concentration range under ambient conditions. The high intensity, linear polarization, and natural collimation of synchrotron radiation (see Section ILB) contribute to the high sensitivity and achievable spatial resolution (-1 pm) of SXRF. The physical processes behind SXRF are identical to those discussed previously for XAS, i.e., there is a high probability that an X-ray photon interacting with an atom will eject a core level electron when the energy of the impinging X rays is approximately equivalent to or slightly greater than the binding energy of the core level. The deexcitation of the atom via fluorescence X-ray production and the measurement of the integrated intensity of the X-ray fluorescence spectrum are directly related to the elemental concentration. Each element produces fluorescent X rays with a different but characteristic energy, providing the basis for the elemental specificity of the technique. There are a number of advantages of SXRF over other commonly employed microprobe techniques. Electron microprobe techniques offer outstanding spatial resolution, but they suffer from relatively poor sensitivity (especially for transition and heavier elements), require a high-vacuum environment, and can result in appreciably greater beam damage to the sample. Like-

SYNCHROTRON X-RAYTECHNIQUES

35

wise, proton-induced X-ray emission (PIXE),although it achieves spatial resolution comparable to that of SXRF, is not as sensitive for heavier elements and can impart beam damage to the sample (Sutton er af., 1994). Furthermore, advances in X-ray-focusing devices and the increased brilliance of the third generation sources will result in spatial resolution and sensitivity enhancements of an order of magnitude or more. A very comprehensive discussion and comparison of SXRF and PIXE techniques as applied to bulk elemental analysis of soils has been presented by Amonette and Sanders (1994), and interested readers are referred to this reference and the references contained therein for a complete discussion of the relative advantages and disadvantages of the various microprobe techniques. The X-26A beam line at the National Synchrotron Light Source at Brookhaven National Laboratory is the only dedicated hard X-ray microprobe beam line currently operating at any synchrotronfacility. The experimental setup includes a primary aperture, a beam-defining aperture composed of computer-controlled tantalum shutters, an optional monochromator for operating in the XAS mode, a pinhole or elliptical focusing mirror to produce small X-ray beams, and a computer-controlled x, y, z, micrometer stepping motor sample stage that can move the samples to appropriate locations within the beam (Fig. 12). The system includes a petrographic microscope with a TV camera attachment so that the location of the sample stage can be properly adjusted from outside the expenmental hutch via computer control. The system is equipped with a number of different detectors, including a solid state Si(Li) energy-dispersivespectrometer, a wavelength-dispersivespectrometer, and a 13-elementsolid state detector. The sample is typically oriented at 45" to the incident beam so that the detector

Figure 12. Configurationof the X-26A dedicated microprobe beam line at the National Synchrotron Light Source, Brookhaven National Laboratory.

36

D. G. SCHULZE AND P. M. BERTSCH

resides at 90"to the incident beam and in the plane of the storage ring. This geometry is required to minimize the detection of scattered X rays, which are the primary source of background intensity in XRF analysis (Sutton et al., 1994). Particulars related to setting up a SXRFexperiment and operating the microprobe on this dedicated beam line have been described in detail elsewhere (Jones and Gordon, 1989; Smith and Rivers, 1994). The SXRF technique has been applied to a very wide range of problems in the earth sciences, including cosmic dust particles, mineral inclusions, and noble gas diffusion in silicates. These and many other applications of SXRF have been comprehensivelyreviewed (Sutton et al., 1993, 1994; Smith and Rivers, 1994). SXRF is particularly useful for trace element analysis in biological samples and in soils and geological materials, particularly where spatial resolution of intact samples or nondestructive analyses of minute samples are required. There are numerous applications in the plant, soil, and environmental sciences, including micronutrient distribution in the rhizosphere, spatially heterogeneous elemental distributions within micromoxphological thin sections, and trace metal concentrations in Mn nodules, secondary carbonate accumulations, various waste forms, and phyllosilicates. The first applications of SXRF to problems in the plant, soil, and environmental sciences include applications dealing with Cr and Se distributions in hyperaccumulator plants employed in phytoremediation (Bertsch et al., 1994a;Hunter et al., 1994), nutrient and transition metal distributions in tree rings (Hunter et al., 1994), contaminant (Ni, As, and Se) distributions in annuli of bone fragments and protein coatings from turtle shells (Bertsch et al., 1994a), trace element distributions in atmospheric particulates (Grant et al., 1992), Se distributions surrounding anaerobic microsites in soil microaggregatesand ponded sediments (Tokunaga et al., 1994b), contarninant metal concentrations associated with ground water colloids (Kaplan et al., 1994), and contaminant metal and elemental distributions and associations in soil and in coal and coal combustion waste products (Bertsch et al., 1994a; Torok et al., 1994). These applications involve nondestructive, in situ determinationof elements in small samples or elemental distributions in heterogeneous samples. The applications have been particularly useful for transition metal analysis of samples for which both energy-dispersiveX-ray analysis (EDX) with an electron microscope and PIXE are not sufficiently sensitive and/or when beam damage to small samples is a concern. Such applications include the determination of Cu and other transition metals associated with atmosphericparticulates that were used as source term indicators (Grant et al., 1992),the determinationof metals in regions associated with the growth annuli of trees and turtle shells as chronicled indicators of contaminant exposure (Bertsch et al., 1994a), and the determination of contaminant metals associated with ground water colloids, which provided

SYNCHROTRON X-RAYTECHNIQUES

37

strong evidence for the role of colloids in the facilitated transport of the contaminants (Kaplan et al., 1994). Other applications have employed 2D scans of samples or regions within a sample to provide complete elemental maps. The spatial resolution of the images of one-half the beam size (-4 Fm) is easily achieved by acquiring data at points separated by a commensurate step size. Examples of applications of this approach are the mapping of Se distributions in anaerotic microsites of soil aggregates, U and coassociated element distribution maps of contaminated soil particles, As distribution maps of fly ash particles, and Se and Cr distribution maps of hyperaccumulating aquatic plants (Bertsch et af., 1994a,b; Tokunaga et al., 1994b; Toriik et al., 1994). Elemental distribution maps generated by 2D scans for U-and Ni-contaminated soil particles and Se localized in the hyperaccumulator species, Lemna minor, are illustrated in Figs. 13 and 14. The image of the plant illustrates the accumulation of Se primarily in the nuclei of individual cells and provides a detailed and accurate representation of morphology when compared to images generated by light microscopy (Hunter et al., 1994). For the Uand Ni-contaminated soil particles, comparison of elemental associations in the images is useful for infemng additional information concerning the chemical speciation of the contaminants, particularly when comparisons are made prior to and following sequential extraction techniques (Bertsch et af., 1994a,b). Perhaps the most exciting and powerful application of SXRF to heterogeneous samples commonly encountered in the plant, soil, and environmental sciences is when it is combined with micro-XAS. Combination of the techniques allows an investigator to probe the chemical environment of an element that has been determined by SXRF to have a spatially variable distribution pattern within a sample or one that is localized in regions having different elemental coassociation patterns. These applications have been discussed in Section III.A.2.g. It is also possible to generate 2D images on the basis of chemical contrast, such as oxidation state, and to superimpose these on the elemental distribution images to provide additional information concerning the chemical speciation of elements of interest. A newly commissioned beam line at the Advanced Light Source (ALS) may provide new opportunities for applying synchrotron microanalytical techniques with emphasis on lighter elements that are currently not accessible at X-26A, but are quite relevant to a number of applications in the plant and soil sciences. The future developments discussed previously for X-ray microfocusing and fast response array detectors, coupled with the increased brilliance of the third generation sources, such as the A P S , will greatly enhance the sensitivity and spatial resolution capabilities of synchrotron-based microanalytical techniques, which will undoubtedly lead to a number of novel applications in the plant, soil, and environmental sciences in the near future.

38 D. G. SCHULZE AND P. M. BERTSCH

Flgwe 13. 2D elemental maps of a fine sand particle from a U-contaminated soil sample. The white tight images were generated on

8000.0--J

0.0

1.o

Fsgure 14. 2D elemental map generated at the CrKa edge of a rootlet from a plant exposed to chromate, showing significant accumulation in the menstem (A), elevated Cr concentrations throughout the root cap region (B), and very low Cr concentrations in the mature root (C). Images were generated on beam line X-26A at the NSLS.

40

D. G. SCHULZE AND P. M. BERTSCH

C. STANDING WAVEAND FLUORESCENT X-RAY INTERFERENCETECHNIQUES X-rays are ideal for probing the liquid-solid interface because they can easily penetrate low-Z materials, but the angstrom wavelength scale of X rays is too fine for probing the ionic distribution in an electrolyte in contact with a charged surface. Long period X-ray standing waves generated by strong Bragg diffraction or by total external reflection of X rays by a surface (Bedzyk, 1990) and fluorescent X-ray interference (Sasaki et al., 1994) are two different approaches for obtaining information about atoms sorbed onto a surface at length scales larger than can be probed using XAS. Standing waves are generated within and above a single crystal by interference of the incident X-ray wave with the coherently diffracted wave during Bragg diffraction (Bedzyk, 1990). This standing wave has antinodal planes that are parallel to the diffracting planes, a period equal to the diffraction plane spacing, and a phase that depends on the incident angle 8. The nodes of the standing wave shift as 8 increases or decreases through the angular region of total reflection. Since the production of fluorescent X rays by an atom is proportional to the electric field at the center of the atom, the production of fluorescent X rays varies with the precise position of the standing wave within and immediately above the crystal. Since the standing wave also exists above the single crystal surface, it can be used to probe the distribution of atoms in an adsorbed layer with respect to the underlying bulk diffraction planes. Qian et al. (1994) used X-ray standing waves generated by dynamical Bragg diffraction to measure the location of naturally occurring trace manganese relative to the (1014) planes of a single crystal of Iceland spar calcite. They concluded that the Mn was located on the same plane as the Ca ions, consistent with its isomorphous substitution for Ca within the crystal. After the calcite crystal was reacted with a solution containing 10 p o l of Pb, sorbed Pb was found to be highly ordered and to occur mostly within the (1014) planes. Standing waves are also generated above a surface when X rays strike the surface at an angle smaller than the critical angle (Fig. 15a). The interference between the incident and specular reflected X-ray plane waves during total external reflection produces an X-ray standing wave above the mirror surface, whose antinodal planes are parallel to the mirror surface (Bedzyk, 1990). As the incident angle 8 is increased from 0 to the critical angle, the nodes and antinodes of the standing wave progressively approach the mirror surface. These standing waves have a period ranging from about 70 to lo00 A. By measuring the fluorescent X-ray intensity from the atoms of interest as a function of incident angle, information can be obtained about the distribution of the atoms above the surface. Bedzyk et af. (1990) used this approach to study the distribution of Zn

SYNCHROTRON X-RAYTECHNIQUES

41

F l p r e 15. (a) Illustration of the X-ray standing wave field formed by the interference between incident and specula-reflected plane waves when the incident X rays strike the sample at an angle less than the critical angle, i.e., when the surface acts as a mirror. The distance of the nodes above the surface varies as the angle of incidence, 0, is changed. (b) Illustration of the principle of fluorescent X-ray interference. The directly emitted and reflected fluorescent X rays from an excited source atom located at a distance z above the surface interfere constructively and destructively as a function of the take-off angle, 8,. of the fluorescent X rays.

ions in the diffusedouble layer in contact with a charged phospholipid membrane at different pH values. Whereas standing wave techniques rely on the interference of incident and diffracted or reflected X rays to probe the interfacial region at a solid surface, Sasaki ef al. (1994) describe an approach that relies only on fluorescent X rays to obtain similar information. A fluorescent X-ray interference pattern is obtained for source atoms above a substrate surface when the take-off angle for fluorescent X-ray detection is less than the critical angle for the total reflection of fluorescent X rays by the surface, i.e., the detector views the surface at angles less than the critical angle. Under these conditions, the fluorescent X rays from the source atom can take one of two paths: either directly from the source atom to the detector, or they are first reflected off of the substrate before traveling to the detector. The path the reflected X ray takes is longer than the path the direct X ray takes by an amount determined by the distance of the source atom above the surface and by the take-off angle of the detector (Fig. 15b). The direct and reflected X rays interfere constructively and destructively with one another to a greater or lesser extent as a function of the take-off angle. By analyzing this interference pattern, one can obtain information on the mean and standard deviation of the distance between the source atom and the substrate. Sasaki et al. (1994) used this approach to measure the size of the Fe-containingcore of femtin and the orientation of Zn-labeled bovine serum albumin. The size of the Fe core of ferritin obtained from the fluorescent X-ray interference measurementsagreed well with the size measured by scanning electron microscopy. The bovine serum albumin molecule is a prolate ellipsoid, and the fluorescent X-ray interference measurements were consistent with the short axis of the molecule being perpendicular to the surface of the substrate. Both the standing wave and the fluorescent X-ray interference techniques

42

D. G . SCHULZE AND P. M.BERTSCH

require the presence of large, perfect surfaces, such as large single crystals, layered synthetic microstructures, or organic-coated gold substrates. Given the relative ease with which layer silicate clays can be deposited on flat surfaces, one or both techniques might be applicable to studying the sorption of various inorganic cations or metal-labeled organic molecules onto clay surfaces.

D. INFRARED MICROSPE~TROSCOPY Infrared radiation from a synchrotron source can provide the source for infrared spectroscopy. Reffner et al. (1994) describe the installation of a standard

Spectra-Tech IRps scanning infrared microprobe onto a port on the vacuum ultraviolet (VUV) ring at the National Synchrotron Light Source. They have found the synchrotron source to be 100- lo00 times brighter than the conventional black-body infrared source originally provided with the instrument. By using the synchrotron source, they were able to obtain roughly the same signal from a 10-p.mZ spot size as could be obtained from a 100-pm2spot by using the original black-body radiator. Overall, they found an enhancement in sensitivity of about 100, although the improved stability of the synchrotron source produced spectra with lower noise for a given measuring time (Reffner et al., 1994). Initial experiments with this instrument have included studies focused on identifying the mineral composition of 10-p,m interplanetary dust particles (Reffner et al., 1994) and hydrogen in defect sites in microscopic single crystals of perovskite (MgSiO,) formed under conditions that simulate the pressures of the earth's interior (Meade et al., 1994). Potential applications in the soil and environmental sciences include spatially resolved identification of mineral phases in thin SIXtions of soils or waste forms to complement the elemental and chemical information obtained by X-ray fluorescence or X-ray absorption spectroscopy.

E. MOSSBAUER SPECTROSCOPY The past decade has seen considerable research aimed at using synchrotron radiation as an alternative to radioactive sources for Mossbauer spectroscopy. Third generation synchrotron sources promise to reduce Mossbauer data acquisition times to the order of seconds from the hours or days required with current radioactive sources (Alp et al., 1993a). Currently, 57Fe, I6QTm,and "9Sn have been measured, but it should be possible to measure additional isotopes in the next 5 years (Alp et al., 1993a,b). The challenge, at the present time, is to develop appropriate beam line optics to obtain the very narrow energy resolution needed for Mossbauer spectroscopy. Alp et al. (1992, 1993a) describe the various strategies currently under consideration. There are plans for Mossbauer

SYNCHROTRON X-RAYTECHNIQUES

43

spectroscopy instrumentation on undulator beam lines at A P S , ESRF, and SPring-8 (Alp er al., 1993a). When these beam lines become available to general users, they will open up the use of Mossbauer spectroscopy to the study of the dynamics of processes such as the oxidation or reduction of iron in the structure of clay minerals.

F. X-RAY DIFFRACTION The principles of X-ray diffraction utilizing a synchrotronX-ray source are not fundamentally different from those of diffraction utilizing a laboratory X-ray source. Synchrotron X-ray sources, however, provide a number of advantages for both single crystal and powder diffraction (Prewitt et al., 1987). We focus here on powder diffraction techniques, because they are most likely to be utilized by soil and environmental scientists, and on some applications of single crystal diffraction to structural biology relevant to the plant sciences. 1. Powder Diffraction of Minerals

There are at least four advantages of powder X-ray diffraction using a synchrotron source compared to conventional X-ray sources: (1) there is roughly an order of magnitude better peak resolution; (2) anomalous scattering effects can be exploited; (3) diffraction patterns can be obtained from small area of samples; and (4) time-resolved diffractioncan be used to follow chemical reactions. Finger (1989) provides a brief introduction to synchrotron powder diffraction, while Cox (1991) provides a more comprehensive review. a. Peak Resolution Only a narrow energy range is selected from a broad continuum of radiation by the crystal monochromators used for synchrotron powder diffraction, eliminating the problem of Ka,-Kct2 doublet broadening inherent with the use of conventional X-ray tubes. The well-defined incident wavelength, along with the highly collimated nature of the synchrotron X rays, allow diffractometer designs that can produce diffraction lines roughly an order of magnitude narrower, and peak to background ratios an order of magnitude better, than those obtained from a conventional laboratory diffractometer (Cox, 1991). This facilitates the identification of trace mineral phases, particularly if they are well crystallized and do not exhibit significant particle size broadening. It should be emphasized that the increase in peak resolution will be significant only for phases for which peak width is controlled by the diffractometeroptics, not by particle size effects. Thus, the peak widths of most soil clay minerals are not significantlyreduced by using a synchrotron source. Nevertheless, there are classes of problems for which this

44

D. G. SCHULZE AND P. M. BERTSCH

increased resolution will be advantageous. Jones and Malik (1993), for example, compared the powder X-ray diffractionpattern of a Hawaiian bauxite obtained by using synchrotron radiation with a pattern of the same sample collected by using a conventional difhctometer. They found a dramatic improvement in the sharpness of gibbsite diffraction lines in the synchrotron powder pattern because the gibbsite line width was controlled by instrumental line-broadening effects. The goethite diffraction lines, which were broadened by particle size effects, had almost the same width-at-half-height in both patterns. The diffraction peaks in synchrotronpowder diffraction patterns are also much more symmetric than peaks obtained from conventional dihctometers. The narrow, symmetric diffraction peaks allow easier use of deconvolution programs and facilitate the application of whole pattern fitting, or Rietveld refinement. In a Rietveld refinement, the parameters of a previously known crystal structure are refined by fitting a calculated pattern to the measured diffraction data by using a least-squares, nonlinear fitting procedure (Rietveld, 1969; Post and Bish, 1989). Thus, accurate structural information such as unit cell parameters, atomic positions, and bond angles can be obtained from materials that occur only as powders and that cannot normally be analyzed by single crystal diffraction techniques. Examples of RieWeld analysis applied to synchrotron powder diffraction data of minerals that could occur in soils or environmental samples include a-CrPO, (Attfield et al., 1988), synthetic leucites (Bell and Henderson, 1994), chrysotile asbestos (Cressey et al., 1994), and palygorskite (Artioli et al., 1994). Artioli et al. (1994) were able to distinguish the orthorhombic form of palygorskite from the monoclinic form in a sample containing both varieties. The swelling behavior of clays is another application that can take advantage of synchrotron diffraction techniques. Huang et al. (1994) studied the dehydration and hydration of montmorilloniteat temperatures up to 777OC and pressures up to 5.3 kbar in a diamond anvil cell. Diffraction patterns were collected in an energy-dispersive mode from microgram quantities of Na-saturated montmorillonite loaded into the diamond anvil cell. b. Anomalous Scattering X-ray scattering by elements changes as the incident wavelength is varied. The change in scattering is particularly large near an adsorption edge of an element and is the result of resonance between the incident X-ray waves and electronic transitions from bound atomic orbitals. This anomalous scattering effect is used increasingly in the determinationof macromolecularstructures using single crystal techniques (Hendrickson, 1991) and can be used to obtain additional phase and structural information for powder samples (Nichols e l af., 1985;Wood er al., 1986). In a differential anomalous powder diffraction experiment, two diffraction patterns are obtained, one using an incident wavelength below the absorption

SYNCHROTRON X-RAY TECHNIQUES

45

edge of the element of interest and the other using a wavelength at the absorption edge. The change in scattering factor of the element whose edge was crossed results in a change in the peak intensities of minerals containing that element. Subtraction of the two patterns results in the diffraction pattern of the phases containing the element of interest. This is similar to differential X-ray diffraction procedures which rely on the selective dissolution of a particular mineral phase (Schulze, 1981, 1994) but with the advantage that exactly the same sample is used for both patterns, allowing much better matching of peaks common to both patterns. The spectral range accessible for anomalous diffraction is for incident X-ray wavelengths of -0.3 to -3.0 (-4-40 keV) (Hendrickson, 1991). This includes the K edges from atomic number Z = 20 (Ca) to Z = 58 (Ce) and L(III) edges from Z = 50 (Sb) to Z = 92 (U). Thus, anomalous diffraction should be applicable to the study of Fe- and Mn-containing minerals in soil clays and to the identification of contaminant phases in environmental samples. Anomalous diffraction effects and computer programs such as NEWMOD (Reynolds, 1985) could be exploited to obtain more precise structural information on mixed layered clays, because layer silicate clays often contain Fe and other heavier elements in their crystal structures, and because a wide variety of elements can be introduced into the interlayer.

a

c. Microdiffraction Microdiffraction techniques are still in the early stages of development. Skelton et al. (1991) obtained diffraction patterns of cylindrical single crystals of bismuth 0.22 pm in diameter by using a wiggler beam line, but in this case the small crystals were bathed in a much larger X-ray beam. To obtain extremely small hard X-ray beams for X-ray microtechniques is a challenge. Various types of X-ray mirrors can be used to focus synchrotron X-ray beams down to 1 pm or slightly less (Bilderback et al., 1994). Even smaller X-ray beams have been produced by condensing hard X-ray beams utilizing the total external reflection of X rays from the inside wall of tapered glass capillaries. Bilderback et al. (1994) obtained X-ray beams as small as 95 nm by utilizing this technique. They also obtained Laue diffraction patterns of a 50-nm-thick single crystal gold foil with a 360-nm-diameter X-ray beam, the smallest single crystal volume ever studied by X-ray diffraction methods. With an appropriately designed beam line, it should be possible to obtain diffraction patterns from pm2 areas of soil thin sections or clay or iron oxide coatings on ped faces and root channels. Thus, in principle, one should be able to obtain information on the spatial distribution of clay-sized minerals in soils or environmental samples to complement the bulk information obtained by conventional clay mineralogy techniques. If glass capillaries come into routine use, it may be possible to study individual crystals of clay-sized minerals by using single crystal diffraction techniques.

46

D. G . SCHULZE AND P. M. BERTSCH

d. Time-Resolved Diffraction The high X-ray flux available from synchrotron sources has allowed the development of time-resolved powder diffraction techniques at time scales inaccessible using laboratory X-ray sources. A particular challenge for time-resolved diffraction is the development of detectors capable of dealing with high count rates and rapid cycling times. In most cases the detectors are the limiting factor in determining how rapidly successive diffraction patterns can be collected. Pennartz et al. (1992) were able to obtain a complete powder diffraction pattern of CaF, in only 200 ms. They also demonstrated that they could follow the phase transitions of Na2S0, at 1-s intervals as the sample was heated at a rate of 1.5" s-1 or cooled at a rate of 3" s-1 between 30 and 320°C. Muhamad et al. (1993) followed the formation of ettringite during the hydration of a rapid-hardening cement used in coal mining. They used an energy-dispersivediffraction geometry, which allows the collection of a complete diffraction pattern for a sample within a 10-mm-diameter sample cell as rapidly as once per second. They concluded that an increase in the ettringite a dimension as the reaction progressed was caused by the initial ettringite being sulfate deficient, but as the reaction progressed, ettringite with increasing sulfate content formed. Rashid et al. (1994) used a similar energy-dispersivedifiaction approach to follow the formation of different calcium aluminate cement hydrate phases over a 200-min time period following mixing. StAhl and Hanson (1994) used a curved positionsensitive detector to obtain complete powder difiaction patterns suitable for Rietveld analysis, with data collection times of only a few minutes per pattern. They obtained detailed structural information on the dehydration behavior of two zeolites, scolecite and mesolite, from consecutive diffraction patterns collected as the samples were heated at a rate of 5" min-1. Synchrotron powder diffraction techniques have not been utilized to any appreciable extent in soil science research, but the examples cited earlier suggest a wide variety of future applications. These might include the identification of trace and poorly crystallized mineral phases in soil and environmental samples and studies of the dynamic nature of shrinking and swelling of soil clays in response to wetting and drying. 2. Single Crystal Dihction of Biological Molecules

Single crystal X-ray dif€raction techniques are one of the major tools used to determine the molecular structures of biological molecules, and synchrotron X-ray sources have allowed crystallographers to determine the structures of molecules of ever increasing complexity. In contrast to inorganic crystals, which typically contain a few tens of atoms in the unit cell, crystals of large biological molecules, such as proteins and viruses, may have unit cells containing thousands of atoms. Standard, rotating anode X-ray tubes can and are used to collect

SYNCHROTRON X-RAYTECHNIQUES

47

the data sets needed for crystal structure determinations, but synchrotron X-ray sources have proven to be so superior to laboratory X-ray sources for large molecule crystallography that facilities for biological crystallography have been developed at synchrotron facilities around the world. A data set that takes several weeks or months to collect using a rotating anode X-ray generator in the laboratory may take only a few hours to collect using a synchrotron source, with the added advantage that the data are of much higher quality. There is voluminous literature on biological crystallography, and it is possible to cite only a few examples to illustrate the type of problems in the plant sciences that have been approached using this technique. Johnson and colleagues have relied heavily on synchrotron diffraction techniques to test a theory that plant viruses can be used as vaccines to deliver harmless proteins that stimulate the production of large quantities of antibodies for resistance against a number of serious diseases. They have focused on cowpea mosaic virus (CPMV) because the virus is thermally stable, easy to purify, and grows extremely well in host plants, with yields of 1-2 g of virus kg- * plant tissue (Usha et al., 1993; Anonymous, 1994). Single crystal X-ray diffraction techniques utilizing a synchrotron X-ray source were used to determine the crystal structure of the native virus, which is a large, icosahedral molecule that crystallized in a cubic space group with a unit cell dimension of a = 317 8, (Chen et al., 1989). Once the structure was determined, they identified points on the virus surface where foreign protein sequencescould be inserted by using molecular genetics techniques. They have successfully inserted protein sequences capable of stimulating the production of antibodies against foot-and-mouth disease virus (a devastating animal disease) (Usha et al., 1993), human rhinovirus 14 (common cold), and human immunodeficiency virus type 1 (HIV or AIDS) (Porta et al., 1994). The human rhinovirus 14 construct has been shown to be immunogenic in rabbits (Porta et al., 1994). If this research is ultimately successful, it will allow cowpea plants (Virgnu unguiculafu) to be used to produce vaccines for a number of serious diseases. The molecular basis of biological nitrogen fixation is another area of intense research that relies heavily on synchrotron-based diffraction and spectroscopy techniques. Nitrogenase, the metalloenzyme that catalyzes the reduction of N, to NH,, consists of two component proteins, the Fe protein with a molecular weight of -60,000 and the MoFe protein with a molecular weight of -250,000 (Dean et al., 1993).The crystallographic structures of these large proteins and their metalloclusters are now worked out (see Dean et al., 1993, for references). Since the molecules contain Mo and Fe, anomalous scattering techniques were used in collecting diffraction data (Bolin et al., 1990), while EXAFS spectroscopy provided additional information on bond lengths and nearest neighbor atoms (Chen et al., 1993a,b; Section III.A.2.d). At the frontiers of structural biology research are efforts to obtain real-time

48

D.G. SCHULZE AND P. M. BERTSCH

moving pictures of molecules and enzymes changing conformation during chemical reactions (Taubes, 1994). The goal is to initiate a chemical reaction in a single crystal in response to a chemical trigger or flash of light and then follow the progress of the reaction by using single crystal h u e diffraction. Current work has focused on crystals cooled with liquid nitrogen to slow the reaction, but the ultimate goal is to allow the reactions to proceed at physiological temperatures.

G. SMALL ANGLESCATTERING Small angle X-ray scattering (SAXS) is an established technique for providing information on the microstructure of disordered solids, biological and inorganic colloids, and soluble polynuclear species (Glatter and Kraty, 1982). The scattering intensity is recorded over a scattering vector amplitude (Q = IT sin WX) range. The information that can be extracted from an SAXS experiment for polymers includes values for the radius of gyration, molecular weight, and other properties concerning tertiary structure. Additionally, significant information concerning the aggregation of polymers can be extracted, including information on individual subunits including size, coordination number of individual particles within an aggregate, intersubunit arrangement within an aggregate, characteristic length and shape of an aggregate, and fractal dimension. Applications of this technique to problems in the plant, soil, and environmental sciences, which are typically represented by weakly scattering systems, have been limited in the past by the low intensity and lack of beam stability of conventional X-ray sources. The extremely high intensity and beam quality from synchrotron X-ray sources, especially the third generation sources, have greatly expanded the potential use of SAXS to a variety of problems of relevance to plant, soil, and environmental scientists. Furthermore, these capabilities have enabled the collection of usable data on weakly scattering systems in milliseconds, thus allowing time-dependent phenomena to be followed. Applications of SAXS to systems of relevance to the soil and environmental sciences have included investigations of the microstructure of phyllosilicate suspensions and the hydrolysis products of A1 and Fe in the presence and absence of various complexing ligands (Bottero el al., 1982, 1987, 1991, 1993; Ben Rhaiem et al., 1987; Tchoubar et al., 1991; Masion et al., 1994a-c). The properties of A1 and Fe hydrolysis products are important in understanding oxyhydroxide mineral phase formation and transformations, as well as having a number of environmental applications related to the role of poorly ordered A1 and Fe oxyhydroxides as critical sorbent phases in soils and sediments and in water and wastewater treatment. Microstructural properties of clays control such factors as chemical reactivity of clay mineral surfaces and macroscopic soil properties such as water-holding capacity, swelling, and permeability. SAXS investiga-

49

SYNCHROTRON X-RAY TECHNIQUES

tions of the microstructure of smectite gels have indicated a structure composed of tactoids represented by 4-5 layers, well ordered in the crystallographic c-axis direction but highly disordered in the lateral orientations (Pons et al., 1982a,b). Other investigations have examined the intercalation of smectite clays by polynuclear hydroxometal cations by SAXS, which demonstrated that the pillaring process is influenced by the relative distribution of individual crystallites and tactoids in the starting clay suspension (Henderson et al., 1989). With the capabilities of the third generation synchrotron sources, it is anticipated that the dynamics of microstructural changes in clay suspensions as a result of the nature of the saturating cation, the oxidation-reduction status of structural Fe, or the intercalation of the clays with hydroxo polynuclear cationic centers will be monitored. The formation of Al(OH), solid phases from solutions containing A1 polynuclear species has been followed with SAXS, and these studies have been instrumental in providing detailed information concerning the rearrangement of the soluble polynuclear structural component characterized by the A11304(OH)24(H20)127+ species (All,) (Bottero et af., 1982, 1987). At OH/Al molar ratios (ii) between 2.3 and 2.6, the All, polynuclear units were observed to gradually form tenuous linear aggregates with some polymerization between octahedral units, composing clusters of -400 8, outside diameter and having a fractal dimension of 1.4. On aging, these aggregates were found to transform to elongated platelets on the order of 550 8, long and 25 A thick. At higher values of ii, more dense, less opened clusters of -1000 A diameter were observed before the initial 100-s acquisition time. At still higher ii, the formation of highly polymerized octahedral layers having a fractal dimension of 1.85 A was observed just following hydrolysis (Bottero et af., 1987). Other SAXS studies have examined the role of organic ligands in modifying the A113 condensation reaction and the structure of the resulting aggregates (Masion et al., 1994a-c). In general, the results of these investigations revealed that organic ligands inhibit the formation of the Al,, species, with the resultant solid phase primarily composed of subunits dominated by uncondensed monomeric species and dense Al(0H)AI colloids, with only small amounts of All, and oligomers that form locally small linear aggregates. Similar detailed information has been derived from SAXS studies of hydrolyzed Fe3+ solutions in the ii range of 0-3 (Tchoubar et al., 1991; Bottero et al., 1991). For hydrolysis products synthesized from the C1- salt, the basic subunit of the colloidal aggregates was characterized by a 16-A-diameter entity, regardless of the ii or the aging time. The local structure of the clusters was found to vary with aging time and ii, characterized by linear aggregates at low ii (1 .O) that exhibited a slightly more branched structure on aging. At higher ri (2.0-3.0), the aggregates were characterized by structures having fractal dimensions of between 1.7 and 2. At the higher ii the local organization was not changed on

-

-

50

D. G. SCHULZE AND P. M.BERTSCH

aging, although photon correlation spectroscopy indicated a size distribution decreasing from 700 to approximately 10 nm, suggesting rearrangement into denser aggregates (Tchoubar et al., 1991). More recent Fe EXAFS investigations of similar systems revealed the presence of edge-sharing dimers and edge- and comer-sharing trimers, which coalesce into a polynuclear unit of approximately 24 Fe atoms having local structure resembling the P-FeOOH structure (Bottero el al., 1994). For hydrolysis products synthesized from the NO3 salt, the size of the subunits was found to vary with A, ranging from 7 8, at A = 1.5 to 13.5 8, at A = 2.5 (Bottero, 1991). The colloidal aggregates formed from the subunits varied in a manner similar to that observed for the C1- salts, i.e., with increasing A the aggregates are characterized by more branched and dense structures. The mechanism proposed to explain the invariant subunit size within colloids formed from the hydrolysis of Fe chloride solutions compared to the increasing subunit size calculated for the hydrolyzed Fe nitrate solutions was the tendency for C1- to form inner-sphere complexes with Fe nucleation centers, whereas the NO3ligand forms weaker outer-sphere complexes that interact with larger polynuclear units once formed (Bottero et al., 1991). These investigations have provided information on these very important poorly ordered phases in soils that has not been obtainable by alternative methods. Future studies on such systems examining a wider range of synthesis conditions and time-dependent processes should greatly enhance the understanding of nucleation mechanisms leading to the formation of both poorly ordered and crystalline Fe oxyhydroxide phases and of those factors that modify the surface chemistry of these phases. Other applications of SAXS relevant to problems in the plant, soil, and environmental sciences include the probing of conformational behavior of organic polymers, such as proteins and humic macromolecules as influenced by a variety of solution parameters including countercation identity, ionic strength, and pH (Russell, 1988; Oades, 1989; Phillips, 1989). It is well recognized from macroscopic investigations that variations in conformation influence the activity of important biopolymers and the sorption behavior of metals and organic contaminants to natural humic macromolecules. Little microscopicdata are available on these systems, although such information is critical to understanding and modeling the reactivity of these important components (Oades, 1989). SAXS has the potential to directly probe these interactions and provide time-resolved information related to the conformational changes that occur with changing solution chemistries or polymer concentrations. Another area having potential applications in the soil, plant, and environmental sciences is anomalous small angle X-ray scattering (ASAXS). The technique is analogous to differential anomalous X-ray scattering described previously in Section 1II.F. 1.b, but is conducted at scattering angles usually <1" 28, thus providing element-specificinformation on long range interactions related to molecular size and configuration.

SYNCHROTRON X-RAYTECHNIQUES

51

H.X-RAYMICROSCOPY X rays are much more penetrating than electrons, and microscopes that use X rays rather than electrons to generate an image allow the sample to remain in air or water during observation. Focusing X rays, however, is much more difficult than focusing electrons; thus, X-ray microscopes are less common. Nevertheless, a number of soft X-ray microscopes have been constructed by using Fresnel zone plates as lenses for focusing X rays. Zone plates are circular diffraction gratings fabricated by using techniques commonly used for the manufacture of integrated circuits in the electronics industry. Zone plates consist of concentric rings of X-ray-opaque and X-ray-transparent material and, when placed in an X-ray beam, behave similarly to a convex refractive lens (Buckley and Rarback, 1990). A phase contrast X-ray microscope has been built at the BESSY electron storage ring in Berlin (Schmahl et al., 1994a,b). It relies on a condenser zone plate to focus the incident X rays onto the sample and a micro zone plate to form the image and can utilize either phase contrast or amplitude contrast to form an image. The image is formed directly like a transmission electron microscope or a light microscope. Schmahl et al. (1994a,b) illustrate the use of this instrument to image artificial lipid membranes, macrophage-specific antigens on the surface of Kupffer cells of a rat liver, and hematite particles. In all cases, the images were obtained while the samples remained fully hydrated. Features as small as about 25 nm have been resolved (Schmahl et al., 1994a). Potential applications include direct imaging of hydrated clay aggregates and particle associations. A different imaging approach is used by the scanning transmission X-ray microscope at the National Synchrotron Light Source (Kirz et al., 1992). This instrument utilizes a zone plate to focus the coherent X-ray flux from an undulator to a submicrometer spot. The sample, which can remain hydrated, is scanned under the incident beam and the image is formed indirectly, like a scanning electron microscope. Contrast is generated by utilizing the rapid change in mass absorption coefficient with a change in X-ray energy above and below an absorption edge. Detail down to 36 nm has been observed by using a gold test pattern. Biological applications have included the study of secretion granules, hydrated chromosomes, and cell cultures (Kinet al., 1992; Zhang et al., 1994). Ade and co-workers (Ade et al., 1992; Ade, 1994; Zhang et al., 1994) have demonstrated that chemical contrast can be obtained by utilizing differences in the X-ray absorption near edge structure near the carbon K edge. For example, the distribution of DNA in chromosomes embedded in bovine serum albumin was mapped with a resolution of 55 nm (Me et al., 1992; Zhang et al., 1994), and Botto et al. (1994) mapped the spatial distribution of aromatic and aliphatic carbons in ultramicrotomed coal thin secretions. As described in Section III.A.2.j, the absorption spectra of particular features can be collected by holding the sample stage

52

D. G. SCHULZE AND P.M. BERTSCH

still and scanning the monochromator. The carbon K edge has been predominately used to obtain images and spectra, but the nitrogen, oxygen, chlorine, fluorine, potassium, and calcium edges, as well as L edges of other elements in the range 220-750 eV, could be used as well (Ade, 1994). For example, Buckley et af. (1992) used the calcium L edge to map the distribution of calcium in sections of tendons and cartilage. Haddad et al. (1994) used the scanning transmission X-ray microscope to obtain ultrahigh resolution, three dimensional X-ray tomographic images of a gold test object, demonstrating a transverse resolution of 0.1 pm and a depth resolution of -0.6 pm. The resolution is limited by the quality of the zone plates, but resolutions on the order of 0.03 Fm should be possible using today’s best available zone plates. This tomographic approach could be used to obtain three-dimensional images of a wide array of biological and inorganic materials that are not easily studied with electron or other probes. X-ray microscopes are also operational at the SRS in Daresbury, United Kingdom, and are planned for the ALS in Berkeley, CA. There are numerous potential applications of the scanning transmission X-ray microscope to the soil, plant, or environmental sciences. For example, the diffusion of potassium in the interlayer of hydrated clay could be studied using an L edge of potassium, or the distribution of calcium, potassium, or other elements within plant cells could be mapped using hydrated samples.

I. X-RAY COMPUTED MICROTOMOGRAPHY X-ray computed tomography is commonly used for medical imaging, where it is referred to as CAT (computer-assisted tomography) or CT (computed tomography) scanning. Computed tomography allows one to observe the interior structure of a solid without physically sectioning the object (NueBhardt et al., 1991; Anderson and Hopmans, 1994). Medical CT instruments, which utilize conventional X-ray tubes, have been used to image large pores and physical features in soils, but features smaller than about 1 mm2 cannot be observed (Aylmore, 1994; Hopmans et al., 1994). Industrial CT scanners provide resolution down to 10010 pm2 (Steude et af., 1994). Synchrotron X-ray sources extend computed tomography to spatial resolutions as small as 1 pm* (Kinney et af., 1994; Spanne et al., 1994a,b; Spanne and Rivers, 1987). An X-ray tomographic image is obtained by measuring the linear X-ray attenuation coefficient for a large number of rays that intersect the object of interest at various angles. In the simplest case, in which a pencil beam of radiation is used, the object is translated through the beam and the transmitted X-ray intensity is measured at each translation step. The object is then rotated by a small amount, and the translation is repeated. The process of rotation and translation is repeated

SYNCHROTRON X-RAYTECHNIQUES

53

EIgure 16. Schematic illustration of a simple computed tomography experiment. The object of interest is rotated and translated in the X-ray beam to measure the X-ray attenuation along a large number of rays through the object. Computer software is then used to reconstruct an image of a “slice” through the interior of the object.

until the object has been rotated 180”. Computer software is then used to reconstruct an image of the density distribution on a plane through the object (Fig. 16). If images of successive “slices” are made at closely spaced intervals, a threedimensional image of the interior of the object can be constructed. By using different incident X-ray wavelengths, the spatial distribution of specific elements can also be obtained in favorable cases. To obtain an image of the spatial distribution of iron in a soil sample, for example, two CMT images are obtained, one using an incident X-ray wavelength just below and the other using a wavelength just above the absorption edge of the element of interest. The difference between the two images is then the distribution of iron within the sample. There have been only a few demonstration projects utilizing computed microtomography to study soils or rocks. Spanne et al. (1994b) determined the geometrical structure of pores in a piece of Fontainebleau sandstone with a voxel size of 10 X 10 X 10 pm. The permeability and conductivity computed from the CT data were in good agreement with those of the experimentally measured data. Spanne et al. (1994a) imaged the distribution of water- and air-filled pores in samples of glass beads or sand grains. An example of a tomogram through sand grains with a pixel size of 10 X 10 pm is illustrated in Fig. 17. A feasibility study has shown that computed microtomography can be used to compare differences between the internal structure of soil aggregates that slake easily when wetted and that of aggregates that do not slake when wetted (D. E. Stott, USDANational Soil Erosion Research Laboratory, West Lafayette, IN, unpublished data). Samples can be imaged moist or dry, thus providing capabilities for examining moist soil microaggregates and soil crusts important in understanding the fundamental processes of soil erosion. Ultimately, it may be possible to image the distribution of water, air, and immiscible organic liquids, such as oil and

54

D.G. SCHULZE AND P.M. BERTSCH

Flpre 17. Computed microtomography image through grains of sand: pixel size, 10 X 10 pnZ; slice thickness, 2 Fm; 407 by 407 element image matrix. (From Spanne er al., IWa, reprinted by permission of the Soil Science Society of America.)

gasoline, in soil cores, providing information important for better understanding the environmental fate of organic pollutants in soils. With fast array detectors and massively parallel computing systems to handle the large data volumes, it may be possible to approach real-time tomography by collecting CT data as rapidly as 10 times a second, allowing one to study the dynamics of processes such as the flow of fluids through soil pores.

IV. ACCESSING SYNCHROTRON FACILITIES Obtaining initial access to a synchrotron research facility may appear to be a formidable task. It is helpful to remember, however, that synchrotron X-ray facilities that provide access to the scientific community at-large must justify their existence by the amount and quality of science done at the facility. In our own experience, synchrotron research facilities generally go out of their way to help new users access the facility. Whether or not a facility is available for the type of research one wishes to

SYNCHROTRON X-RAY TECHNIQUES

55

conduct is determined by the user access policy of the facility and by the type of instrumentation developed on the individual beam lines. There are several approaches to obtaining access for the first time. One approach is to contact a colleague who regularly does synchrotron-based research to see if he or she can run initial samples. This is the most direct route, and the one most likely to get a new user started in the shortest amount of time. Even if the colleague’s interests are different, he or she can be a helpful source of information for getting to the synchrotron facility, finding housing, and providing suggestions of who to talk to for more information. Another approach is to call the user administrationoffice at the synchrotron facility. The user administration office should be able to provide names of beamline scientists to contact to discuss the technical details of one’s proposed experiment. Yet another way to make initial contacts is to attend the synchrotron facility’s annual user meeting. These meetings usually consist of a day or two of technical presentations held at or near the synchrotron facility. The user administration office can provide the date and location of the next user meeting. Finally, some groups occasionally hold hands-on workshops to introduce new users to synchrotron-based techniques. There is no substitute for an initial visit to the synchrotron facility prior to planning more extensive experiments. One should take along the samples to be studied and any special sample chambers so that the beamline scientists will better understand the proposed experiments. Sometimes an experiment which seemed possible on a particular beamline when it was described over the phone, will turn out to be unfeasible when the beamline scientist sees the samples or equipment. The beamline may not be set up to handle a particular type of sample, or the sample chamber may not physically fit the equipment available at the beamline. In this case, the beamline scientist should be able to suggest other beamlines to contact concerning the experiment. If the experiment is feasible, the beamline scientist may be willing to run one or two samples to obtain initial data. This initial data is often essential for obtaining additional beam time later. There are generally two groups of users at synchrotron facilities. One consists of groups of scientists who build the instrumentation on particular beamlines and who maintain and operate the beam lines on a day-to-day basis. At U.S.synchrotron facilities, these groups are called the participating research teams (PRT) or collaborative access teams (CAT). At the NSLS and APS, the PRT or CAT members also are required to raise the money needed to construct, maintain, and operate the instrumentation. The beamline scientists who work at a synchrotron facility full time are members of, or work for, a PRT or CAT. The PRT or CAT members receive a large percentage of the available beam time for their own research in exchange for building, maintaining, and operating the beamline instrumentation. The second group is the general users. General users travel to the synchrotron for a few days at a time to conduct their own experiments at the synchrotron and then return to their home institutions. Anyone may apply for

56

D. G. SCHULZE AND P. M. BERTSCH

general user beam time by submitting a proposal to the synchrotron facility. The mechanics of the process vary from facility to facility, but generally the proposal is 3-4 pages in length. A typical experimental run is 2 - 4 days in length or perhaps more, depending on the policies of the facility and the type of experiment. The synchrotron operates 24 h per day. To keep from wasting valuable beam time, it is important to have the assistance of one or two graduate students, technicians, or colleagues for data collection. It is also a good idea to plan to stay a few days after the experimental run to work on data analysis. Data analysis usually requires the assistance of the beamline scientist and it is easier to begin the data analysis before returning home. In the U.S.,soil and environmental scientists from state land-grant universities and other research agencies within the U.S. Department of Agriculture (USDA) have organized a national committee within the Cooperative States Research, Education, and Extension Service of USDA. The objective of this committee, designated NCR- 174 and entitled Synchrotron X-ray Sources in Soil Science Research, is to facilitate the usage of synchrotron X-ray techniques by soil and environmental scientists within the U.S. agricultural research system. This national group of soil and environmental scientists has aligned itself with a national group of geoscientists to plan, build, and operate two beamlines at the Advanced Photon Source. The group is a member of the Consortium for Advanced Radiation Sources (CARS), a large collaborative access team planning instrumentation at the APS for a number of scientific disciplines. Once completed, the GeoSoilEnviroCARS beamlines will serve as a national user facility for synchrotronbased research in the geological, soil, and environmental sciences at the APS. Beam time for members of the general geological, soil, and environmental sciences research communities will become available on a peer-reviewed basis beginning in 1996. Contact the authors of this chapter for additional information on NCR-174, CARS, or the GeoSoilEnviroCARSbeamlines. Synchrotron X-ray sources provide a wide variety of analytical techniques with applications in the soil, plant, and environmental sciences. Although synchrotron-basedresearch requires a large, centrally funded and staffed facility, much of the research done at synchrotron facilities is carried out by a large number of individual researchers, each working with only a few students and support scientists. Third generation synchrotron sources with exceedingly brilliant X-ray beams promise orders of magnitude greater capability over current first and second generation sources and will provide exceptional research capabilities for soil, plant, and environmental research well into the 21st century.

ACKNOWLEDGMENTS We thank Mark Rivers for providing the figures of X-raybrightness for selected synchrotron X-ray sources, John Johnson and Jeff Bolin for providing reprints of their work, Susan Barr for providing

SYNCHROTRON X-RAYTECHNIQUES

57

the drawing of the APS, Keith Jones and Lyle Runty for providing a selection of computed tomography figures, and Doug Hunter for providing the SXRF images. We also thank Mark Rivers, Steve Sutton, Diane Stott, Don Franzmeier, Sara Gruenhagen, Chris Guest, and Doug Hunter for reviewing all or portions of the manuscript. Paul Bertsch was partially supported by contract D E - A m 76SROO-819 between the U.S. Department of Energy and the University of Georgia during the preparation of this manuscript. This is article number 14,639 of the Purdue Office of Agricultural Research FVograms.

REFERENCES Me, H. 1994. NEXAFS microscopy of polymeric samples. Synchrotron Radiat. News 7(2), 11-15. Ade, H., and Hsiao, B. 1993. X-ray linear dichroism microscopy. Science 262, 1427-1429. M e , H., Zhang, X., Cameron, S., Costello, C., Kin, J., and Williams, S. 1992. Chemical contrast in x-ray microscopy and spatially resolved XANES spectroscopy of organic specimens. Science 2589972-975. Alp, E. E., Mooney, T. M., and Toellner, T. 1992. Coherent nuclear resonant optics for third generation synchrotron radiation sources. SPIE Proc. 1740, 95- 107. Alp, E. E., Mooney, T. M., Toellner, T.. and Sturhahn, W. 1993a. Mossbauer spectroscopy with synchrotron radiation. h “Proc. 2nd Nassau Mossbauer Conf.” (W. Wynter and E. E. Alp, eds.), in press. W. C. Brown Publ. Alp, E. E., Mooney, T. M., Toellner, T., Sturhahn, W., Witthoff, E., Rohlsberger, R., Gerdau, E., Homma, H., and Kentjana, M. 1993b. Time resolved nuclear resonant scattering from “9Sn nuclei using synchrotron radiation. Phys. Rev. Len. 70, 3351-3354. Amonette, J. E., and Sanders, R. W. 1994. Nondestructivetechniques for bulk elemental analysis. I n “Quantitative methods in soil mineralogy” (J. E. Amonette and L. W. Zelazny, eds.), Soil Sci. Soc. Amer. Misc. Publ., pp. 1-48. Soil Science Society of America, Madison, WI. Anderson, S. H., and Hopmans, J. W. 1994. “Tomography of soil-water-root processes,” Soil Sci. Soc. Amer. Spec. Publ. 36. Soil Sci. Soc. Amer., Madison, WI. Anonymous. 1994. Reaping human vaccines from plants. Science 263, 1211. Artioli, G., Galli, E., Burattini, E.. Cappuccio, G., and Simeoni, S. 1994. Palygorskite from Bolca, Italy: A characterization by high-resolution synchrotron radiation powder diffraction and computer modeling. N. Jb. Miner. Mh.. 217-229. Attfield, J. P., Cheetham, A. K., Cox, D. E., and Sleight, A. W. 1988. Synchrotron x-ray and neutron powder diffraction studies of the structure of a-CrPO,. J. Appl. Crysrallogr. 21,452457. Aylmore, L. A. G. 1994. Application of computer assisted tomography to soil-plant-water studies: An overview. I n ‘Tomography of soil-water-mt processes” (S. H. Anderson and J. W. Hopmans, eds.), SSSA Spec. Publ., pp. 7-15. Soil Science Society of America, Madison, WI. Bajt, S., Clark, S. B., Sutton, S. R., Rivers, M. L., and Smith, J. V. 1993. Synchrotron x-ray microprobe determination of chromate content using x-ray absorption near-edge structure. Anal. Chem. 65, 1800- 1804. Bajt, S., Sutton, S.R., and Delaney, J. S. 1994. X-ray microprobe analysis of iron oxidation states in silicates and oxides using x-ray absorption near edge structure (XANES). Geochim. Cosmochim. Acra 58 5209-5214. Baker, G., Dent, A. J., Derbyshire, G., Greaves, G. N., Catlow, C. R. A., Couves, J. W., and Thomas, J. M. 1991. Time resolved structural studies of nickel exchanged zeolite Y and nickel oxide using energy dispersive EXAFS. I n “X-ray absorption fine structure” (S. S.Hasnain, ed.), pp. 738-741. Ellis Horwood, New York.

D.G. SCHULZE AND P. M. BERTSCH Bedzyk, M. J. 1990. Measuring the diffuse-double layer at an electrical interface with long period X-ray standing waves. Synchmtmn Radiat. News 3(5), 25-29. Bedzyk, M. J., Bommarito. G. M.,Catky, M., and Penner, T. L. 1990. Diffuse-double layer at a membrane-aqueous interface measured with x-ray standing waves. Science 248, 52-56. Bell, A. M. T.,and Henderson, C. M. B. 1994. Rietveld refinement of the structures of drysynthesized MFe”*Si,O, leucites (M = K, Rb, Cs) by synchrotron x-ray powder diffraction. Acta Crystallogr. C50, 1531-1536. Ben Rhaiem, H., Pons, C. H., and Tessier, D. 1987. Factors affecting the microstructure of smectites: Role of cation and history of applied stresses. In “Proceedings of the International Clay Conference, Denver, 1985” (L. G. Schultz, H. van Olphen, and F. A. Mumpton, eds.), pp. 292-297. Clay Minerals Society, Bloomington, IN. Bertsch. P. M., Hunter, D. B., and Clark, S. B. 1994a. In situ speciation of metals and radionuclides in soils, sediments, waste forms, and biota by micro X-ray absorption spectroscopy. Am. Chem. Soc.,I&EC Special Symposium, Atlanta, GA. pp. 1356-1359. Bertsch, P. M., Hunter, D. B., Sutton, S. R., Bajt, S., and Rivers, M. L. 1994b. In situ chemical speciation of uranium in soils and sediments by micro X-ray absorption spectroscopy.Environ. Sci. Technol. 28, 980-984. Bianconi, A. 1988. XANES spectroscopy.In “X-ray absorption: Principles, applications, techniques of EXAFS, SEXAFS and XANES” (D. C. Koningsberger and R. Prins, eds.), pp. 573-662. John Wiley, New Yo&. Bidoglio, G., Gibbson, P. N., Haltier, E., Omenetto, N.,and Lipponen, M. 1992. XANES and laser fluorescencespectroscopy for rare earth speciation at mineral water interfaces. Radiochim. Acta 58-59, 191-197. Bidoglio, G.. Gibson, P. N., O’Gorman, M. 0.. and Roberts, K. J. 1993. X-ray absorption spectroscopy investigation of surface redox transformation of thallium and chromium on colloidal mineral oxides. Geochirn. Cosmochim. Acta 57, 2389-2394. Bilderback, D. H., Hoffman, S. A., and Thiel, D. J. 1994. Nanometer spatial resolution achieved in hard x-ray imaging and h u e diffraction experiments. Science 263, 201-203. Bolin, J. T., Ronco, A. E., Mortenson, L. E., Morgan, T.V., Williamson, M., and Xuong, N.-H. 1990. The structure of nitrogenase MoFe protein: Spatial distribution of the intrinsic metal atoms determined by x-ray anomalous scattering. In “Nitrogen fixation: Achievements and objectives” (P. M. Gresshoff, L. E. Roth, G.Stacey, and W. E. Newton, eds.), pp. 117-122. Chapman and Hall, New York. Bottero, J. Y., Tchoubar, D., Cases, J. M., and Flessinger, F. 1982. Investigation of the hydrolysis of aqueous solutions of aluminum chloride. 2. Nature and structureby small-angle x-ray scattering. J. Phys. Chem. 86, 3667-3673. Bottero, J. Y..Axelos, M., Tchoubar, D.. Cases, J. M., Fripiat, J. J., and Flessinger, F. 1987. Mechanism of formation of aluminum trihydroxide from Keggin Al,, polymers. J. Colloid Interface Sci. 117,4747. Bottero, J. Y.,Tchoubar, D., Amaud, M.. and Quienne, P. 1991. Partial hydrolysis of femc nitrate salt. Structural investigation by dynamic light scattering and small-angle x-ray scattering. Lungmuir 7 , 1365-1369. Bottero, J. Y.,Masion, A., Lartiges, B. S., Thomas, F., Tchoubar, D., and Axelos, M. A. V. 1993. Hydrolysis and flocculation: A structural approach through small-angle x-ray scattering. J. PhyS. 3, 211-218. Bottero, J. Y., Manceau, A., Villiiras, F., and Tchoubar, D. 1994. Structure and mechanisms of formation of FeOOH(C1) polymers. Lungmuir 10, 316-319. Botto, R. E., Cody, G. D., Km,J., Ade, H., Behal, S., and Disko, M. 1994. Selective chemical mapping of coal microheterogeneity by scanning transmission x-ray microscopy. Energy Fuels 8, 151-154.

SYNCHROTRON X-RAYTECHNIQUES

59

Brown, G. E., Jr.. Calas, G.. Waychunas, G. A., and Petiau, J. 1988. X-ray absorption spectroscopy and its applications in mineralogy and geochemistry. In “Spectroscopic methods in mineralogy and geology” (F. C. Hawthorne, ed.), Reviews in Mineralogy 18, pp. 431-512. Miner. Soc. Amer., Washington, DC. Brown, G. E., Jr., Parks, G. A., and Chisholm-Brause, C. 1. 1989. In situ x-ray absorption spectroscopic studies of ions at oxide-water interfaces. Chimia 43, 248-256. Brown, G. E., Jr., Parkhurst, D. A., and Parks, G. A. 1995. Zinc complexes in aqueous chloride solutions: Structure and thermodynamic modelling. Geochim. Cosmochim. Acra, in press. Brown, G . S . , and Doniach, S. 1980. The principles of x-ray absorption spectroscopy. In “Synchrotron radiation research” (H. Winick and S. Doniach, eds.), pp. 353-385. Plenum Press, New York. Buckley, C. J., and Rarback, H. 1990. Scanning x-ray microscopy. In “Modern microscopies: Techniques and applications” (P. J. Duke and A. G. Michette, eds.), pp. 69-85. Plenum Ress, New York. Buckley, C. J., Foster, G. F., Burge, R. E., Ali, S. Y.,Scotchford, C. A., Kin, J., and Rivers, M. L. 1992. Elemental imaging of cartilage by scanning x-ray microscopy. Rev. Sci. Insrrum. 63,588-590. Calas, C., Brown, G. E., Waychunas, G. A., and Petiau, J. 1987. X-ray absorption spectroscopic studies of silicate glasses and minerals. Phys. Chem. Minerals 15, 19-29. Charlet, L., and Manceau, A. A. 1992a. X-ray absorption spectroscopic study of the sorption of Cr(1ll) at the oxide-water interface. 1. Molecular mechanism of Cr(II1) oxidation on Mn oxides. J. Colloid Interface Sci. 148, 425-442. Charlet, L., and Manceau, A. A. 1992b. X-ray absorption spectroscopic study of the sorption of Cr(II1) at the oxide-water interface. 11. Adsorption, coprecipitation,and surface precipitation on hydrous femc oxide. J. Colloid Inrerface Sci. 148, 443-458. Charlet, L., and Manceau, A. 1993. Structure, formation, and reactivity of hydrous oxide particles: Insights from x-ray absorption spectroscopy. In “Environmental particles” (J. Buffle and H. P. van k u w e n , eds.), Vol. 2, pp. 117-164. Lewis Publ., Ann Arbor, MI. Chen, Z., Stauffacher, C., Li, Y.,Schmidt, T., Bomu, W., Kamer, G., Shanks, M., Lomonossoff, G., and Johnson, J. E. 1989. Protein-RNA interactions in an icosahedral virus at 3.0 8, resolution. Science 245, 154-159. Chen, J., Christiansen, J., Campobasso, N., Bolin, J. T., Tittsworth, R. C., Hales, B. J., Rehr, J. J., and Cramer, S. P. 1993a. Refinement of a model for the nitrogenase Mo-Fe cluster using singlecrystal Mo and Fe EXAFS. Angew. Chem., Int. Ed. Engl. 32, 1592-1594. Chen, J., Christiansen, J., George, S. J., van Elp, J., Tittsworth, R., Hales, B. J., Al-Ahmad, S., Coucouvanis, D., Campobasso, N., Bolin, J. T., and Cramer, S. P. 1593b. Extended x-ray absorption fine structure and Ledge spectroscopy of nitrogenase molydenum-iron protein. In “Molydenum enzymes, cofactors, and model systems” (E. I. Stiefel, D.Coucouvanis, and W. E. Newton, eds.), pp. 231-242. American Chemical Society, Washington, D.C. Chisholm-Brause, C. J.. Brown, G. E., Jr., and Parks, G. A. 1989a. EXAFS investigation of aqueous Co(I1) adsorbed on oxide surfaces in-situ. Physica B+C 158, 6 4 6 6 4 8 . Chisholm-Brause, C. J., Row, A. L., Hayes, K.F., Brown, G. E., Jr., Parks, 0 . A.. and Leckie, J. 0. 1989b. XANES and EXAFS study of aqueous Pb(I1) absorbed on oxide surfaces. Physica B, 158, 674-675. Chisholm-Brause, C. J., Hayes, K. F., Roe. A. L., Brown, G. E., Jr., Parks, G. A., and Leckie, J. 0. 1990a. Spectroscopic investigation of Pb(I1) complexes at the y-A1,OJwater interface. Geochim. Cosmochim. Acra 54, 1897-1909. Chisholm-Brause, C. I., O’Day, P. A., Brown, G. E., Jr., and Parks, G. A. 1990b. Evidence for multinuclear metal-ion complexes at solid-solution interfaces from x-ray absorption spectroscopy. Nature 348, 528-530.

D. G. SCHULZE AND P. M. BERTSCH

60

Clausen, B. S., and Topsge,H. 1994. Combined QEXAFS/XRD A new experimental approach in materials science. S y n c h m n Radiar. News 7(1), 32-36. Combes, J. M., Manceau, A., and Calas, G. 1986. Study of the local structure in poorly-ordered precursors of iron oxi-hydroxides. J. Phys. C8,697-701 Combes, J. M.. Manceau, A., Calas, G.,and Bottero, J. Y. 1989. Formation of ferric oxides from aqueous solutions: A polyhedral approach by X-ray absorption spectroscopy. 1. Hydrolysis and formation of ferric gels. Geochim. Cosmochim. Acra 53, 583-594. Combes. J. M., Manceau, A., and Calas, G. 1990. Formation of ferric oxides from aqueous solutions: A polyhedral approach by X-ray absorption spectroscopy. 11. Hematite formation from ferric gels. Geochim. Cosmochim. Acra 54, 1084-1094. Combes, J. M., Chisholm-Brause, C. J., Brown, G.E., If.. Parks, G. A., Conradson, S. D., Eller. P. G . . Ray, I. R., Hobart, D. E., and Meijer, A. 1992. EXAFS specroscopic study of neptunium(V) sorption at the a-Fe00Hlwater interface. Environ. Sci. Technol. 26, 376-382. Come, A. R., Evans, M. C. W., Hubbard. J. A. M., Strange, R. W.. and Hasnain, S. S. 1991. An EXAFS study of the oxygen-evolving complex of photosystem 11. I n “X-ray absorption fine structure” (S. S. Hasnain, ed.), pp. 178-180. Ellis Horwood. New York. Cotter-Howells, J. D., Champness, P. E., Charnock, J. M., and Pattrick. R. A. D. 1994. Identification of pyromorphite in mine-waste contaminated soil by ATEM and EXAFS. Eur. J. Soil Sci. 45,393-402.

Cox, D. E. 1991. Powder diffrection. I n “Handbook on synchrotron radiation” (G. S. Brown and D. E. Moncton, eds.), Vol. 3, pp. 155-200. Elsevier, New Yo&. Cramer, S. P., Hodgson, K. O., Stiefel, E. I., and Newton, W. E. 1978. A systematic x-ray absorption study of molybdenum complexes. The accuracy of structural information from extended x-ray absorption fine structure. J. Am. Chem. SOC. 100, 2748-2761. Cressey, B. A., Cressey, G., and Cemik, R. J. 1994. Structural variations in chrysotile asbestos fibers revealed by synchrotron x-ray diffraction and high-resolution transmission electron microscopy. Can. Mineral. 32, 257-270. Dean, D. R., Bolin, J. T., and Zheng, L. 1993. Nitrogenase metalloclusters:structures, organization, and synthesis. J. Bucreriol. 175, 6737-6744. de G m t , F. M. F. 1991. X-ray absorption of transition metal oxides. Chapter 2, pp. 15-25. Krips Repro, Meppel. Dent, A. J., Ramsay, J. D. F., and Swanton, W. 1992. An EXAFS study of uranyl ion in solution and sorbed onto silica and montmorillonite clay colloids. J. Colloid Interface Sci. 150, 45-60. DeWitt, J. G., Hodgson, K. O., Bentsen, J. B.. Rosenzweig, A. C., Lippard, S. J., Hedman, B., Green, J., Pilkington. S., and Dalton, H. 1991. X-ray absorption spectroscopy of difemus and differicprotein A of soluble methane monooxygenasefrom Methylococcuscapsulatus (Bath). I n ‘‘X-ray absorption fine structure”(S. S. Hasnain, ed.). pp. 128-130. Ellis Horwood, New York. Dobson, B. R. 1994. Quick scanning EXAFS facilities at Dambury SRS.Synchmron Radiar. News 7(1), 21-24.

Drits, V. A., Sakharov, B. A., Salyn, A. L., and Manceau, A. 1994. Structural model for ferrihydrite. Clay Miner. 28, 185-207. Farges, F., Ponader, C. W., Calas, G., and Brown, G. E., Jr. 1992. Structural environments in silicate glasslmelt systems: UIV,Uv, and Uvl. Geochim. Cosmochim. Actu. 56, 4205-4220. Farges, F., Peck, J. A., and Brown, G. E., Jr. 1994. Local environment around gold(II1) in aqueous chloride solutions: An EXAFS spectroscopic study. Geochim. Cosmochim. Acra. 57, 12431252.

Fendorf, S. E., Lamble, G . M., Stapleton, M. G., Kelley, M. J., and Sparks, D. L. 1994a. Mechanismsof chomium(II1) sorption on silica. I . Cr(I11) surface structure derived by extended x-ray absorption fine structure spectroscopy. Envimn. Sci. Technol. 28, 284-289. Fendorf, S. E., Sparks, D. L., h b l e , G. M., and Kelley, M. I. 1994b. Applications of x-ray absorption fine structure spectroscopy to soils. Soil Sci. Soc. Am. J. 58, 1583-1595.

SYNCHROTRON X-RAYTECHNIQUES

61

Finger, L. W. 1989. Synchrotron powder diffraction. In “Modern powder diffraction” (D. L. Bish and J. E. Post, eds.), Reviews in Mineralogy Vol. 20. pp. 307-331. Miner. Soc. Amer., Washington, D.C. Finkelstein, K. D., and Sutton. M. 1994. On the collection of diffraction data simultaneously over a range of energy. Nucl. Insirurn. Methods A347, 495-498. Frahm, R. 1991. Quick XAFS: Potentials and practical applications in materials science. In “X-ray absorption fine structure” (S. S. Hasnain, ed.), pp. 731-737. Ellis H o r w d , New York. Glatter, 0..and Kraty, 0. 1982. “Small angle x-ray scattering.” Academic Press, New York. Grant, R. H., Schulze, D. G.,and Sutton, S. R. 1992. Trace element composition of background mid-continental atmospheric aerosols. Agmn. Abstr., 371. Greaves, G. N., Pizzini, S., Roberts, K. J., Barrett. N. T., and Kalbitzer, S. 1991. Glancing angle XAFS for the study of real surfaces. In “X-ray absorption fine structure’’ (S. S. Hasnain, ed.), pp. 232-237. Ellis Horwood, New York. Haddad, W. S., McNulty. 1.. Treks, J. E., Anderson, E. H., Levesque, R. A., Yang, L. 1994. Ultrahigh-resolution x-ray tomography. Science 266, 12 13- 12 15. Harris, J. E., Hukins, D. W. L., and Hasnain, S. S. 1988. Calcium environment in bone mineral determined by EXAFS spectroscopy. Calcif. Tissue Inr. 43, 250-253. Hasnain, S. S. 1991. “X-Ray Absorption Fine Structure.” Ellis Horwood, New York. Hayes, K. F., Roe, A. L., Brown, G.E., Jr., Hodgson, K. 0.. Leckie. J. 0.. and Parks, G. A. 1987. In situ X-ray absorption study of surface complexes at oxide/water interfaces: Selenium oxyanions on a-FeOOH. Science 238, 783-786. Henderson, S. I., Dai. L., and White, J. W. 1989. Some applications of small angle scattering in chemistry. In “Chemical Applications of Synchrotron Radiation: Workshop report’’ (M. Beno and S. Rice, eds.), ANLIAPS-TM-4, pp. 57-96. Argonne National Laboratory, Argonne, IL. Hendrickson, W. A. 1991. Determination of macromolecular structures from anomalous diffraction of synchrotron radiation. Science 254, 51-58. Hoffman, S. A., Thiel, D. J., and Bilderback, D. H. 1994. Developments in tapered monocapillary and polycapillary glass X-ray concentrators. Nucl. Instrum. Methods Phys. Res. A347, 384389.

Hopmans, J. W.. Cislerova, M., and Vogel, T. 1994. X-ray tomography of soil properties. In ‘Tomography of soil-water-root processes” (S.H. Anderson and J. W. Hopmans, eds.), SSSA Spec. Publ., pp. 17-28. Soil Sci. Soc. Amer., Madison, WI. Huang, W.-L., Bassett, W. A., and Wu, T.-C. 1994. Dehydration and hydration of montmorillonite at elevated temperatures and pressures monitored using synchrotron radiation. Am. Miner. 79, 683-691.

Huffman, G. P., Mitra, S., Huggins, F. E., Shah, N., Vaidya, S., and Lu, F. 1991. Quantitative analysis of all major forms of sulfur in coal by x-ray absorption fine structure spectroscopy. Energy Fuels 5 , 574-581. Huggins, F. E., Huffman, G.P., Mitra, S., and Shah, N. 1991. Determination of sulphur forms in coal by XAFS spectroscopy. In “X-ray absorption fine structure” (S. S. Hasnain, ed.), pp. 610612. Ellis H o r w d , New York. Hunter, D. B., Clark, S. B., and Bertsch, P. M. 1994. Uptake and metabolism of metals by aquatic plants determined by synchrutron X-ray absorption micro-analysis. Agmn. Abstr., 256. Ildefonse, P., Kirkpatrick, R. J., Montez, B., Calas, G.,Flank,A. M., and Lagarde, P. 1994. *’A1 MAS NMR and aluminum x-ray absorption near edge structure study of imogolite and allophanes. Clays Clay Miner. 42, 276-287. Irlam, J. C., Holt. C., Hasnain, S. S., and Hukins, D. W. L. 1985. Comparison of the structure of micellar calcium phosphate in milk from six species by extended x-ray absorption fine structure spectroscopy. J. Dairy Res. 52, 267-273. Johnston, C. T., Sposito, G.,and Earl, W. L. 1993. Surface spectroscopy of environmental particles by fourier-transform infrared and nuclear magnetic resonance spectroscopy. In “Environmental

62

D. G. SCHULZE AND P.M. BERTSCH

particles” (1. Bume and H. P. van Leeuwen, eds.), Vol. 2, pp. 1-36. Lewis Publ., Ann Arbor, MI. Jones, K. W.. and Gordon, B. M. 1989. Trace element determinations with synchrotron-induced x-ray emission. Anal. Chem. 61, 341A-358A. Jones, R. C.. and Malik, H. U. 1993. A computer technique for rapid decomposition of x-ray diffraction instrumental aberrations from mineral line profiles. I n “Computer applications to x-ray powder difFraction analysis of clay minerals” (R. C. Reynolds. Jr. and J. R. Walker, eds.), pp. 156-171. The Clay Minerals Society, Boulder, CO. Kaplan, D. I., Hunter, D. B., Bertsch, P. M.. Bajt, S.,and Adriano, D. C. 1994. Application of synchrotron x-ray fluorescence spectroscopy and energy dispersive x-ray analysis to identify contaminant metals on groundwater colloids. Envimn. Sci. Technol. 28, 1186-1 189. Kemner, K. M., Elam, W. T., Hunter, D. B., and Bertsch, P. M. 1995. EXAFS studies of CsBrdibenzo-IS-crown-6ether solutions. XAFS Conference VIII. Phys. B , in press. Kim, K.-J. 1986. Characteristics of synchrotron radiation. In “X-ray data booklet” (D. Vaughan, ed.). PUB490 rev., pp. 4-1 to 4-16. Lawrence Berkeley Laboratory, Berkeley, CA. Kinney, 1. H., Haupt, D. L., Nichols, M. C., Breunig, T. M., Marshall, G. W., Jr., and Marshall, S. J. 1994. The x-ray tomographic microscope: Three-dimensional perspectives of evolving microstructures. Nucl. Instrum. Methods A347,480-486. Kin, J., Ade. H., Jacobsen, C.. Ko, C.-H., Lindaas, S., McNulty. I., Sayre, D., Williams, S., Zhang, X., and Howells, M. 1992. Soft x-ray microscopy with coherent x-rays. Rev. Sci. Instrum. 63,557-563. Koningsberger,D. C., and Prins, R. 1988. “X-ray absorption: Principles, applications, techniques of EXAFS, SEXAFS, and XANES.” John Wiley & Sons, New York. Korystova, A. F., Shelestov, V. M., Vazina, A. A., and Kochubei, D. 1. 1991. EXAFS studies of differentcenters of Ca2+-bindingproteins. In “X-ray absorption fine structure” (S.S. Hasnain, ed.), pp. 187-190. Ellis Horwood, New York. Kusunoki, M., Ono, T.. Inoue, Y., Suzuki, M., Uehara, A., Matsushita, T., and Oyanagi, H. 1991. Mn K-edge XANES spectroscopy for water-splittingMnenzyme in photosynthesis. High quality pre-edge features in the SI and S2 states. I n “X-ray absorption fine structure” (S.S.Hasnain, ed.),pp. 174-177. Ellis Horwood, New York. Lee. P. L.. Beno, M. A., Knapp, G. S., and Jennings, G. 1994. Continuous energy diffraction spectroscopy: A new d-space matching technique for energy dispersive synchrotron radiation diffraction. Rev. Sci. Instrum. 65, 2206-2209. Li, D., Bancroft, G. M., Kasrai, M.. Fleet, M. E., Secco, R. A., Feng, X. H.. Tan, K. H.. and Yang, B. X. 1994. X-ray absorption spectroscopy of silicon dioxide (SiO,) polymorphs: The structural characterization of opal. Am. Miner. 79, 622-632. Lindley, P. F. 1991. EXAFS and crystallographic studies of metallo-proteins containing iron. I n “X-ray absorption fine structure” (S.S. Hasnain, ed.), pp. 115-121. Ellis Horwood, New York. Lytle, F. W.,and Greegor. R. B. 1991. New developments in XAS experiments. I n “X-ray absorption fine structure” (S.S. Hasnain, ed.), pp. 625-633. Ellis Honvood, New York. Manceau, A., and Charlet, L. 1992. X-ray absorption spectroscopic study of the sorption of Cr(I1) at the oxide-water interface. 1. Molecular mechanism of Cr(1II) oxidation on Mn oxides. 1.Colloid Intetj4ace Sci. 148, 425-442. Manceau, A., and Drits, V. A. 1993. Local structure of fenihydrite and feroxyhite by EXAFS spectroscopy. Clay Miner. 28, 165-184. Manceau, A., Combes. J. A., and Calas, G. 1990. New data and a revised model for fenihydrite: A comment on a paper by R. A. Eggleton and R. W. Fitzpatrick. Clays Clay Miner. 38,331-334. Manceau, A., Charlet, L., Boisset, M. C., Didier. B., and Spadini, L. 1992a. Sorption and speciation of heavy metals on hydrous Fe and Mn oxides. From microscopic to macroscopic. Appl. Clay Sci. 7, 201-223.

SYNCHROTRON X - R w TECHNIQUES

63

Manceau, A., Gorshkov, A. I., and Drits,V. 1992b. Structural chemistry of Mn, Fe, Co and Ni in Mn hydrous oxides. I. Information from XANES spectroscopy. Am. Miner. 77, 1133-1143. Manceau, A., Gorshkov, A. I., and Drits, V. 1992c. Structural chemistry of Mn, Fe, Co and Ni in Mn hydrous oxides. 11. Information from EXAFS spectroscopy, electron and x-ray diffraction. Am. Miner. 77, 1144-1157. Masion, A., Bottero, J. Y., Thomas, F., and Tchoubar, D. 1994a. Chemistry and structure of Al(0H)lorganics precipitates. A small angle x-ray scattering study. Part 11: Speciation and structure of the precipitates. lmngmuir 10,4349-4352. Masion. A., Tchoubar, D., Bottero, J. Y., Thomas, F., and Villiiras, F. 1994b. Chemistry and structure of AI(0H)lorganic.xprecipitates. A small angle x-ray scattering study. Part I: Numerical procedure for speciation from scattering curves. Langrnuir 10,4344-4348. Masion, A., Thomas, F., Tchoubar, D., Bottero, J. Y., and Tekely, P. 1994c. Chemistry and structure of AI(0H)lorganic.xprecipitates. A small angle x-ray scattering study. Part 111: Depolymerization of the All, polycation by organic ligands. Langrnuir 10, 4353-4356. Meade, C., Reffner, J. A., and Ito, E. 1994. Synchrotron infrared absorbance measurements of hydrogen in MgSiO, pervoskite. Science 264, 1558-1560. Miller, F., Jr. 1972. “College Physics.” 3rd Ed. Hancourt Brace Jovanovich, Inc., New York. Moggridge, G. D., Parent, P., and Tourillon, G. 1994. A NEXAFS study of the orientation of benzoate intercalated into a layer double hydroxide. Clays Clay Miner. 42, 462-472. Motschi, H. 1987. Aspects of the molecular structure in surface complexes: spectroscopic investigations. In “Aquatic Surface Chemistry” (W. Stumm, ed.). pp. I 11-125. Wiley-Interscience, New York. Muhamad, M. N., Barnes, P., Fentiman, C. H., Hausermann, D., Pollman, H., and Rashid, S. 1993. A time-resolved synchrotron energy dispersive diffraction study of the dynamic aspects of the synthesis of ettringite during minepacking. Cem. Concr. Res. 23,267-272. Murphy, L. M., Hasnain, S.S.,Strange, R. W., Harvey, I., and Ingledew, W. J. 1991. XAFS studies on blue copper proteins: The effect of pH and oxidation state changes on the copper site. I n “X-ray absorption fine structure” (S. S. Hasnain, ed.). pp. 152-156. Ellis Horwood, New York. Nichols, M. C., Smith, D. K., and Johnson, Q . 1985. Differential x-ray diffraction: A theoretical basis for a technique based on wavelength variation. J. Appl. Crystallogr. 18, 8-15. NuShardt, R., Bonse, U., Busch, F., Kinney, J. H., Saroyan, R. A., and Nichols, M. C. 1991. Microtomography: A tool for nondestructive study of materials. SynchrorronRadiat. News 4(3), 21-23. Oades, J. M. 1989. An introduction to organic matter in mineral soils. I n “Minerals in soil environments” (J. B. Dixon and S. B. Weed, eds.), pp. 89-159. Soil Science Society of America, Madison, WI. O’Day. P. A., Brown, G. E., Jr., and Parks, G. A. 1994a. X-ray absorption spectroscopy of cobalt(l1) multinuclear surface complexes and surface precipitates on kaolinite. J. Colloid Inrerface Sci. 165, 269-289. O’Day, P. A., Parks, G. A., and Brown, G. E.,Jr. 1994b. Molecular structure and binding sites of cobalt(I1) surface complexes on kaolinite from x-ray absorption spectroscopy.Clays C k y Miner. 42, 337-355. Pennartz, P. U., Lachner, U., and Fuess, H.1992. Powder diffraction in the range of milliseconds. J. Appl. Crysiallogr. 25, 571-577. Phillips, J. C. 1989. Macromolecular structure changes in solution observed by time-resolved synchrotron x-ray scattering. I n “Chemical applications of synchrotron radiation: Workshop report’’ (M. Beno and S. Rice, eds.), ANLIAPS-TM-4, pp. 101-114. Argonne National Laboratory, Argonne, IL. Pons, C. H., Rousseaux, F., and Tchoubar, D. 1982a. Utlisation du rayonnement synchrotron en

64

D. G. SCKULZE AND P.M.BERTSCH

diffusion aux petits angles pour I’itude du godement des smectites. I. ltude du systhme eaumontmorillonite-Na en fonction des smectites. CIay Miner. 17,327-338. Pons, C. H.. Tessier, D., Ben Rhaiem, H.. and Tchoubar, D. 1982b. A comparison between x-ray studies and electron microscopy observations of smectite fabric. I n “Proceedings of the International Clay Conference, Bologna and Pavia” (H. van Olphen and F. Veniale, eds.), pp. 177185. Elsevier, Amsterdam. Porta, C., Spall. V. E.. Loveland. J., Johnson. J. E.,Barker, P. I.. and Lomonossoff, G. P. 1994. Development of cowpea mosaic virus as a high-yielding system for the presentation of foreign peptides. Vimlogy 2@2,949-955. Post, J. E.. and Bish. D. L. 1989. Rietveld refinement of crystal structures using powder x-ray dif€raction data. In “Modern powder dihction” (D.L. Bish and J. E. Post, eds.). Reviews in Mineralogy Vol. 20. pp. 277-308. Miner. Soc. Amer.. Washington, D.C. Povey, J. F., Diakun, G. P., Gamer, C. D., Wilson, S. P., and h u e , E. D. 1991. EXAFS studies of zinc model compounds: Metal ion coordination in the DNA binding domain of the yeast transcriptional activator GAL4. I n “X-ray absorption fine structure” (S. S. Hasnain, ed.), pp. 181-183. Ellis Horwood, New York. Prewitt, C. T., Coppens, P., Phillips, I. C., and Finger, L. W. 1987. New opportunities in synchrotron x-ray crystallography. Science 238, 312-3 19. Priggemeyer, S., Eggers-Borkenstein. P.. Rompel, A., Krebs, B.. Henkel, G.,Witzel, H., Kvrner, M., Nolting, H.-F.. and Hermes, C. 1991. XAS investigations on the Fe(lI1)-Zn(II)center of purple acid phosphatase from red kidney beans. I n “X-ray absorption fine structure’’ (S. S. Hasnain. ed.). pp. 131-133. Ellis Horwood, New York. Qian, Y.,Sturchio, N. C.,, Chiarello. R. P.,Lyman, P. F., Lee, T.-L., and Bedzyk, M. J. 1994. Lattice location of trace elements within minerals and at their surfaces with x-ray standing waves. Science 265, 1555-1557. Rashid, S., Bames, P., Bensted, J., and ’hnillas, X. 1994. Conversion of calcium aluminate cement hydrates re-examined with synchrotron energydispensive diffraction. J. Muter. Sci. Len. 13, 1232-1234. Reher, J., Cam, G. L., Sutton, S., Hemley, R. J., and Williams, G. P. 1994. Infrared microspectroscopy at the NSLS. Synchrotron Radiar. News 7(2), 30-37. Reynolds, R. C. 1985. NEWMOD, a computer program for the calculation of basal x-ray diffraction intensities of mixed-layered clays. R. C. Reynolds, Hanover, NH. Rietveld. H. M. 1%9. A profile refinement method for nuclear and magnetic structures. J . Appl. Crysrallogr. 2, 65-7 I. Rivers, M.L. 1990. Characteristics of the Advanced Photon Source and comparison with existing synchrotron facilities. In “Synchrotron x-ray sources and new opportunities in the soil and environmental sciences: workshop report” (D. G. Schulze and J. V. Smith, eds.), ANLIAPSITM-7, pp. 5-23. Argonne National Laboratory, Argonne, IL. Roe. A. L.. Hayes, K. F., Chisholm-Brause, C. J., Brown, G. E., Jr.. Parks, G. A,, and Leckie, J. 0. 1991. X-ray absorption study of lead complexes at a-FeOOHIwater interfaces. Langmuir 7, 361-373. Russell, T. P. 1988. Time-resolved SAXS studies on polymers. In ‘Time-resolved studies and ultrafast detectors: Workshop report” (R. Clarke, P. Sigler, and D. Mills, eds.), ANLIAPSTM-2, pp. 29. Argonne National Laboratory. Argonne, IL. Sandstrom, D. R. 1984. EXAFS studies of electrolyte solutions. In “EXAFS and near-edge structure 111” (K. 0. Hodgson era/., eds.), pp. 409-412. Springer Verlag. New York. Sasaki. Y. C.. Suzuki, Y., and Ishibashi, T. 1994. Fluorescent x-ray interference from a protein monolayer. Science 263, 62-64. Sayers, D. E., Stern, E. A., and Lytle, F. W. 1971. New technique for investigating noncrystalline structures: Fourier analysis of the extended x-ray absorption fine structure. Phys. Rev. Lerr. 27, 1204- 1207.

SYNCHROTRON X-RAYTECHNIQUES

65

Schmahl, G., Rudolph, D., Guttmann, P., Schneider, G., Thieme, J., Niemann. B., and Wilhein, T. 1994a. Phase contrast x-ray microscopy. Synchrotron Radiat. News 7(4), 19-22. Schmahl, G.. Rudolph, D.. Schneider, G., Guttmann, P., and Niemann, B. 1994b. Phase contrast x-ray microscopy studies. Oprik 97, 181-182. Schulze, D. G. 1981. Identificationof soil iron oxide minerals by differential X-ray diffraction. Soil Sci. Soc. Am. J. 45, 437-440. Schulze, D. G. 1994. Differential x-ray diffraction analysis of soil minerals. I n “Quantitative methods in soil mineralogy” (J. E. Amonette and L. W. Zelazny, eds.), Soil Sci. Soc. Amer. Mix. h b l . . pp. 412-428. Soil Science Society of America, Madison, WI. Schulze, D. G., McCay-Buis, T.,Sutton, S. R., and Huber, D. M. 1995a. Manganese oxidation states in Gaeumannomyces infested wheat rhizospheres probed by micro-XANES spectroscopy. Phytopathology, in press. Schulze, D. G., Sutton, S. R., and Bajt, S. 1995b. Deknnhation of manganese oxidation state in soils Wing x-ray absorption near-edge structure (XANES) spectmsapy. Soil Sci. SOC.Am. J.. in press. Skelton, E. F.,Ayers, J. D., Qadri, S. B., Moulton, N. E., Cooper, K. P., Finger, L. W., Mao, H. K.,and Hu. 2. 1991. Synchrotron x-ray diffraction from a microscopic single crystal under pressure. Science 253, 1123-1125. Smith, J. V., and Rivers, M.L. 1994. Synchrotron x-ray microanalysis. I n “Microprobe techniques in the earth sciences” (P.J. Potts, J. F. W. Bowles. S. 1. B. Reed, and M. R. Cave, eds.), in press. Chapman and Hall, London. Spanne. P., and Rivers, M. L. 1987. Computerized microtomography using synchrotron radiation from the NSLS. Nucl. Instrum. Methods B24I25, 1063-1067. Spanne, P.. Jones, K. W., Prunty, L. D., and Anderson, S. H. 1994a. Potential applications of synchrotron computed microtomography to soil science. I n ‘“Tomography of soil-water-root processes” (S. H. Anderson and J. W. Hopmans, eds.), SSSA Spec. h b l . , pp. 43-57. Soil Science Society of America, Madison, WI. Spanne, P., Thovert, J. F.,Jacquin, C. J., Lindquist, W. B., Jones, K.W., and Adler, P. M. 1994b. Synchrotron computed microtomographyof porous media: Topology and Transports. Phys. Rev. Lett. 73,2001-2004. Sposito, G. 1986. Distinguishing adsorption from surface precipitation. I n “Geochemical processes at mineral surfaces” (J. A. Davis and K.F. Hayes, eds.), Symp. Ser. Vol. 323, pp. 217-228. American Chemical Society, Washington, DC. Stihl, K., and Hanson. J. 1994. Real-time x-ray synchrotron powder diffraction studies of the dehydration processes in scolecite and mesolite. J. Appl. Crystallogr. 27, 543-550. Stem, E. A. 1974. Theory of extended x-ray absorption fine structure. Phys. Rev. B10,3027-3037. Steude. J. S., Hopkins, F., and Anders, J. E. 1994. Industrial x-ray computed tomography applied to soil research. I n ‘“Tomography of soil-water-mt processes” (S. H. Anderson and J. W. Hopmans, ed~.),SSSA Spec. Publ., pp. 29-41. Soil Science Society of America, Madison, WI. Stragier, H., Cross, J. 0.. Rehr. J. J.. Sorensen, L. B., Bouldin, C. E., and Woicik, J. C. 1992. Diffraction anomalous fine structure: A new x-ray structural technique. Phys. Rev. Lett. 69, 3064-3067. Sutton, S. R., Rivers. M. L., Bajt, S., and Jones, K. W. 1993. Synchrotron X-ray fluorescence microprobe analysis with bending magnets and insertion devices. Nucl. Insfrum. Methods Phys. Res. B75, 553-558. Sutton, S. R., Rivers, M. L., Bajt, S., Jones, K.,and Smith, J. V. 1994. Synchrotron X-ray fluorescence microprobe: A microanalytical instrument for trace element studies in geochemistry, cosmochemistry, and the soil and environmental sciences. Nucl. Instrum. Methods Phys. Res. A347, 412-416. Sutton, S. R.. Bajt, S., Delaney, J., Schulze, D., and Tokunaga, T. 1995a. Synchrotron x-ray fluorescence microprobe: Quantification and mapping of mixed valence state samples using micro-XANES. Rev. Sci. Insmun. 66, 1464-1467.

66

D. G. SCHULZE AND P. M.BERTSCH

Sutton, S. R.. Hoffman, S., Bajt, S., and Jones, K. 1995b. Synchrotron x-ray fluorescence microscopy with a capillary x-ray concentrator. Chem Geol., submitted for publication. Taubes, G. 1994. X-ray movies start to capture enzyme molecules in action. Science 266,364-365. Tchoubar. D., Bottero. J.-Y., Quienne, P., and Amaud. M. 1991. Partial hydrolysis of ferric chloride salt. Structural investigation by photon-correlation spectroscopy and small-angle x-ray scattering. Langmuir 7 , 398-402. Teo, B. 1. 1986. “EXAFS: Basic Principles and Data Analysis,” Inorganic Chemistry Concepts 9. Springer-Verlag. Berlin. Thiel, D. J., Bilderback, D. H., and Lewis, A. 1993. Production of intense micrometer-sized x-ray beams with tapered glass monocapillaries. Rev. Sci. Instrum. 64, 2872-2875. Tokunaga, T. K., F’ickering, I. J., and Brown, G. E., Jr. 1994a. X-ray absorption spectroscopy studies of selenium transformations in ponded sediments. (in preparation). Tokunaga, T. K., Sutton, S. R., and Bajt, S. 1994b. Mapping of selenium concentrations in soil aggregates with synchrotron x-ray fluorescence microprobe. Soil Sci. 158,421-434. Tomk, S . , Faigel, G.. Jones, K. W., Rivers, M. L., Sutton. S. R., and Bajt, S. 1994. Chemical characterization of environmental particulate matter using synchrotron radiation. X-Ray Spectmtn. 23,3-6. Ullrich, J. B., Gibson, W. M., Gubarev, M. V., MacDonald, C. A., and Xiao, Q.F. 1994. Potential for concentration of synchrotron beams with capillary optics. Nucl. Instrum. Methods Phys. Res. A347.401-406. Usha, R., Rohll, J. B., Spall, V. E., Shanks, M., Maule, A. J., Johnson, J. E., and Lomonossoff, G. P. 1993. Expression of and animal virus antigenic site on the surface of a plant virus particle. Virology 197, 366-374. Van der Heijdt, L. M., Veldink, G. A., Vliegenthart, I. F. G.. Feiters, M. C., Navaratnam, S., Nolting, H.-F., and Hermes, C. 1991. EXAFS of soybean lipoxygenase-I: Influence of lipid hydroperoxide activation and lyophilization on the structure of the soybean lipoxygenase nonheme iron active site. I n “X-ray absorption fine structure” (S. S. Hasnain, ed.), pp. 125-127. Ellis Horwood, New York. Waldo, G. S., Carlson, R. M. K., Modowan, 1. M., Peters, K. E., and Penner-Hahn, J. E. 1991. Sulfur speciation in heavy petroleums: Information from x-ray absorption near-edge structure. Geochim. Cosmochim. Acra 55, 801-814. Waychunas, G. A., Rea, B. A., Fuller, C. C., and Davis, J. A. 1993. Surface chemistry of ferrihydrite: Part 1: EXAFS studies of the geometry of coprecipitated and adsorbed arsenate. Geochim. Cosmochim. Acra 57, 2251-2269. Winick, H. 1987. Synchrotron radiation. Sci. Am. 257(5), 88-99. Winick, H., and Williams, G. P. 1991. Overview of synchrotron radiation sources world-wide. Synchmrron Radiar. News 4(5), 23-26. Wood, I. G., Nicholls, L., and Brown, 0.1986. X-ray anomalous scattering difference patterns in qualitative and quantitative powder diffraction analysis. J. Appl. Crysrallogr. 19, 364-371. Yang, B. X., Rivers, M. L., Schildkamp, W., and Eng, P. 1995. GeoCARS microfocusing Kirkpatrick-Baez mirror bender development. Rev. Sci. Instrum., in press. Zhang, X., Ade, H., Jacobsen, C.. Kin, J.. Lindaas, S.,Williams, S., and Wirick, S. 1994. MicroXANES: chemical contrast in the scanning transmission x-ray microscope. Nucl. In.wum. 431-435. Methods Phys. Res. -7,

GEOGRAPHIC INFORMATION SYSTEMS IN AGRONOMY G. W. Petersen', J. C. Bell*, K. McSweeney3, G. A. Nielsen4, and P. C. Robert2 'Department of Agronomy and Environmental Resources Research Institute, The Pennsylvania State University, University Park, Pennsylvania 16802 %oil Science Department, University of Minnesota, St. Paul, Minnesota 55108 'Department of Soil Science, University of Wisconsin, Madison, Wisconsin 5 3706 4Department of Plant & Soil Science, Montana State University, Bozeman, Montana 59717

I. Introduction 11. Overview of GIs Technology A. Spatial Data Capture B. Typical GIS Operations C. Advantages and Disadvantages of GISs 111. Remote Sensing A. Remote Sensing as a Data Source B. Remote Sensing-Supported GIS Operations C. Trends in Remote Sensing Iv. Terrain Analysis and Soil-Landscape Modeling V Site-Specific Fanning A. Current Applications of Site-Specific Farming B. Emerging Role of GIS and GPS C. Components of GIS/GPS Site-Specific Fanning VI. Environmental Applications A. Regional-Scale Pollution Assessment B. Watershed-Based Nonpoint Source Assessment C. Spatial Decision Support Systems VII. Conclusions References

67 Mwnrrr b Agrmnnrj( Vdvnr 55 C o m g h t 0 I995 by Aerdemic Pms, Inc. All rights of reproduction in any form rcservcd.

68

G. W. PETERSEN ET AL.

I. INTRODUCTION Interest in and use of geographic information systems (GIS) have accelerated in recent years. The management and manipulation of spatially related data using computer-aided techniques began in the 1960s and has grown rapidly in the 1990s.

An increasing worldwide population, coupled with misuse of land resources, requires the application of new technologies like GIS to help maintain a sustainable food and water supply without degrading the environment. We are entering an era in which the use of GIS can better organize and integrate scientific data, address spatial and temporal variability, model soil-landscapes using quantitative and statistical methods to define relationships, advance knowledge, and manage landscapes as ecosystems. With the development of GIS technology, decision makers have at their disposal information systems in which data are readily accessible and easily combined and modified. This technology allows the examination of a wider range of variables than is usually considered in land management decisions and will, therefore, lead to a better understanding of how landscape systems function and interact. Scientists, managers, and politicians can use this technology to make informed decisions on the basis of more complete exploration of land use options, greater sensitivity to environmentalquality, and increased public participation. It is not the intent of this chapter to provide an extensive literature review of geographic information systems. To review such an extensive body of literature is well beyond the scope of the current work. This chapter is intended to provide an introduction to GIS and associated landscape tools and to illustrate how they are being used in various aspects of agronomy.

II. OVERVIEW OF GIs TECHNOLOGY As scientists concerned with land management, agronomists must work with spatial information involving patterns of soil properties, cropping practices, pest infestations, weather conditions, and topography. Historically, the analog map was the most common method of depicting patterns of spatial information, with early versions apparently predating the first alphabet (Marble, 1990). The availability of high speed computers led to the developmentof digital tools to capture, store, analyze, and report spatial information that are commonly referred to as geographic information systems (GIS). GIS have existed for over three decades, but only recently have they been widely used for natural resource planning (Burrough, 1986).

GEOGRAPHIC INFORMATION SYSTEMS

69

The information stored within a GIS can be divided into two distinct categories. The first category includes spatially referenced information that can be represented by points (i.e., well locations or sampling points), lines (i.e., river and road networks), and polygons (i.e., fields, soil delineations, or land use classes). This information is referenced to a geographic coordinate system and is usually stored in either a raster (grid-cell) or vector (arc-node) digital format (Petersen ef af., 1990). For raster formats, an area is divided into a series of equidimensional rasters, and each raster is assigned a value representing the status of the geographic feature for that location (Fig. la). Rasters of various sizes are used, depending on the scale of the map being digitized. Vector formats are based on the explicit definition of coordinates that define the location of a point, line, or polygon feature (Fig. lb). For example, a soil map unit is defined by a series of points defining vectors that outline the map unit boundaries. Each vector is assigned two identifiers, one for each side of the vector; each identifier defines the attribute for the polygon. In general, more sophisticated software systems are needed for geographic data stored in vector formats, compared with raster formats. However, vector formats provide a more precise representation of the spatial feature; this may be required for certain applications, such as delineating property boundaries. The second type of information stored in a GIS is attribute data or information describing the characteristics of the spatial feature. For example, attribute data for a soil map unit could include the predominant soil series, soil drainage class, and texture of the surface soil horizon. Relational database programs are frequently used to store and manipulate attribute data. A defining feature of a GIS is the ability to link this attribute database to the spatial feature database for geographic analysis. Geographic information systems range from relatively simple to highly sophisticated and are adapted for use on a variety of computer platforms. Each of these systems includes the following capabilities: 1. The ability to enter or capture spatially referenced information from existing analog maps or remote sensors. 2. The ability to store, retrieve, and edit spatial data and attribute information. 3. The ability to perform spatial analysis such as map overlay, spatial buffer analysis, attribute reclassification, and spatial summaries. Many systems also have the ability to link models with spatial databases. 4. The ability to display maps and generate tabular reports of attribute data.

A GIS consists of three major components: (1) computer hardware, (2) computer software, and (3) digital geographic data. Computer hardware configurations for GIS applications vary, depending on the complexity of software and size of the digital database. Basic components usually include a central processing unit, such as a personal computer or workstation; mass storage devices using magnetic disks, optical disks, or magnetic tapes; data capture devices, such as a

70

G. W. PETERSEN ET AL. Orlginai Map

Y

RasterMap

b

Original Map

Dlgltlzed Polnts

Vector Map

:

ngure 1. (a) Conversion of a map to a raster (grid-cell) digital format. (b) Conversion of a map to a vector (arc-node) digital format.

manual digitizer or scanner; and map production devices, such as pen or electrostatic plotters. GIS software can be purchased from commercial vendors, or public domain software usually may be obtained for a minimal cost. The availability and quality of technical support after purchasing the software is also an important consideration. Land resources data for GIS applications can be obtained from the Natural Resource Conservation Service, the United States Geological Survey, and state agencies. .

A. SPATIALDATACAPTURE The capture of existing information on analog maps into a digital format requires careful consideration of the quality of the map delineations and the mapping base characteristics. Because many GIS applications rely on the ability

GEOGRAPHIC INFORMATION SYSTEMS

71

to make accurate spatial measurements, and because multiple maps of geographic features are layered, these maps must be spatially registered to a common, controlled geographic coordinate system and must be of similar scales. This may be problematic for maps prepared from aerial photography such as soil surveys, which may contain numerous spatial distortions that would be inherited. Day er al. (1988) provide a good discussion of base map problems associated with soil surveys. The coordinates of geographic features can be captured from maps by either manually digitizing or scanning or in the field by using a survey instrument such as a global positioning system. Remotely sensed data, such as those collected from satellite and aircraft scanners, can be used for GIS applications after image classification.

B. TYPICAL GIs OPERATIONS Although the capabilities of individual geographic information systems will vary, most systems are able to perform the following basic operations: 1. Data retrieval. After data are stored in a digital format, the maps are displayed on a computer monitor. Most systems allow the user to browse the map and query the attributes of geographic features. For example, a user may move a cursor to select a polygon and query the attribute data for that polygon. 2. Map reclassification and common boundary dissolution (Fig. 2a). This procedure is used to create a series of thematic maps from the original map. In Fig. 2a, soil mapping units are reclassified to create a map of the predominant soil drainage class for each polygon. The boundaries between polygons that have a common value are then dissolved to create the final soil drainage class map. 3. Map manipulation (Figs. 2b-d). vpical operations include edge matching to seam together several adjacent maps, creating a subset, transformations between different map projections or coordinate systems, and conversion between raster and vector data storage formats. 4. Buffer generation (Fig. 2e). All of the area falling within a specified distance of a point, line, or polygon feature is identified. In Fig. 2e, all of the area within 100 m of a stream is identified by proximity analysis. 5 . Map overlay (Fig. 2f). The combination of features found on two or more maps that are referenced to the same map base is identified. In Fig. 2f, soil type and management area maps are overlaid to identify the soil types found within each management area. 6. Spatial measurement. The linear and curvilinear distances of lines, area, perimeter, and centroids of polygons, as well as volumes of spatial features, are calculated. 7. Data output. Analog maps and tabular reports summarizing the results of spatial analysis and modeling are generated.

-

ipping Units a Soil Mapping

p Sol1 DrainageClass

I T I

Reclassify

Common Dissolve Boundary

Npre 2. (a) Map reclassification and common boundary dissolve. (b) Edge matching to seam together several adjacent maps. (c) Creation of subset h m a mapped area. (d) Conversion between raster and vector data storage formats. (e) Buffer generation. (f) Map overlay of soils and management areas. 72

d

Vector

- Conversion

e

z Streams

Stream Bufler

Proxlmtty Analysts

Soils and Management Areas

73

74

G. W. PETERSEN ET AL.

C. ADVANTAGES AND DISADVANTAGES OF GISs Although a GIS is a powerful tool for certain types of applications, it is not a universal solution for all problems involving spatial data. Like many new technologies, the capabilities of GISs are exaggerated at times, and it is important that a thoughtful assessment of the applicability of GIS technology is made for each proposed application. A general list of some of the advantages and disadvantages of GIS technology (adapted from Dangermond, 1990) follows. Advantages 1. Data are stored in a physically compact format and can be retrieved quick-

ly. 2. Spatial analysis is conducted by computer algorithms that, from a practical perspective, are not performed on analog map data, such as multiparameter spatial modeling and change analysis. 3. Spatial and attribute data are integrated into a single system. 4. It is cost effective for certain complex spatial modeling tasks. 5 . Data collection, spatial analysis, and decision making are integrated into a single system. Disadvantages 1. The cost can be prohibitively high to convert existing maps and attribute data into a digital format suitable for GIS applications. 2. Purchase and maintenance costs of computer software and hardware are high for complex modeling tasks or sustaining large databases. 3. A relatively high level of technical expertise is required for successful GIS ventures. The primary cost when establishing a working GIS involves database development; this accounts for over 90% of the total system cost in some cases. After a digital database is established, however, it can be easily updated and used for numerous applications. GIS technology enables the exploration of new and innovative approaches to the scientific management of the land to help ensure a sustainable agricultural resource base and to protect our environment.

III. REMOTESENSING Remote sensing is the noninvasive and nondestructive gathering of information about a feature removed from the sensing instrument. This technology has

GEOGRAPHIC INFORMATION SYSTEMS

75

become a valuable source of input for GIS databases because the instruments are capable of quickly collecting data (often in a digital format) over large areas.

A. REMOTE SENSINGAS A DATASOURCE Remote sensing systems vary in the nature of the observation platform, meth-

od of sensing, and instrument configuration. Instrument platforms are designed to operate as ground-based, (e.g., ground-imaging radar either as hand-held or vehicle-mounted instruments), airborne (e.g., airplane- and helicopter-mounted), and orbital (satellites and manned spacecraft) systems. The variety of platforms is matched by the diversity in instrument characteristics, as defined by type of sensing and instrument sensitivity. Qpe of sensing has two components: the way sensed information is recorded and the use of generated or natural illumination. Information is recorded primarily through either photographic or imaging methods. Imaging is more convenient for GIS applications, as it is collected and stored in digital form. Photographs and maps must be scanned or digitized (i.e., converted into digital form) for GIS use. The use of natural versus generated illumination is an important component in remote sensing. Passive systems rely on natural illumination to provide the energy needed to produce the photographic or imaged record of the area of interest. Detected energy is reflected, transmitted, or emitted prior to passive sensing. By contrast, active systems, such as synthetic aperture radar (SAR) or ground-penetrating radar, generate predetermined energy to illuminate targets. Instrument sensitivity to the electromagnetic spectrum is the second way to differentiate between remote sensing systems. Some systems have instruments that measure the intensity of detected energy in specific and discrete ranges. One example is the French satellite SPOT, which when imaging in XS mode has three sensors, each sensitive to different parts of the spectrum, including blue (0.5000.590 pm), red (0.610-0.680 pm), and near-infrared (0.790-0.890pm). Other instruments measure the intensity of detected energy over a continuous spectral range. Aerial photography or the SPOT panchromatic instrument, which is sensitive in the 0.510-0.730-pm range, are examples of instruments with this kind of sensitivity. These instruments sense the part of the electromagnetic spectrum referred to as the optical and reflective infrared. Other passive instruments measure thermal infrared and microwave sections of the spectrum. Active sensors detect the intensity of the response to reflected, instrumentgenerated wave energy. Ulaby et al. (1981, 1982, 1986) describe active and passive microwave sensing theory, applications, systems, and output. Active systems are attractive for acquiring data from areas frequently covered by clouds. Wavelengths commonly used by these instruments have much greater atmo-

76

G. W. PETERSEN ET AL.

spheric penetrability when compared with instruments sensitive to the optical and reflective infrared parts of the spectrum.

B. REMOTESENSING-SUPPORTED GIs OPERATIONS As one possible source of GIS input data, remote sensing is useful in four capacities: as a display backdrop; for the generation of thematic maps; for the derivation of input variables for models; and as a real-time link. As a display backdrop, the primary function of a remote sensing image is as a visual aid or for the interactive creation and updating of maps. If the desired base is a photograph, then scanning is needed to convert it to digital form. This highlights an advantage of imaged data generated in digital form. Remotely sensed data used for displays can be processed to correct for distorting influences. For example, digital elevation models are used to remove topographic influences on feature location when generating digital orthophotos. Atmospheric and terrain corrections are sometimes undertaken with multispectral imagery. Corrective procedures enable the production of displays that are more accurately geocoded to a selected coordinate system than uncorrected forms of the data. Geocoding is central to thematic map creation, which is the second use of remote sensing data for GIS functions. Remotely sensed data are often used to generate thematic maps that can subsequentlybe utilized as a data layer for GIS functions. As before, if photography is used for map creation, then the extra step of either digitizing or scanning a thematic map is needed. The photography can also be scanned, with the mapping done onscreen to produce a digital form of the map. The current trend is toward the use of softcopy photogrammetry to generate digital orthophotos, elevation data, and feature extraction from scanned aerial photography. Softcopy systems provide a significant time savings compared to the use of conventional stereoplotters. Digital orthophotography also provides a mapping base that can be directly used for a variety of GIs-based applications. Softcopy photogrammetry requires the use of special hardware and software, and good quality aerial photography is necessary to offset any losses of resolution due to the scanning process. Corbley (1995) describes soficopy photogrammetry and its application in the generation of a countywide database. Accuracy is an issue when remote sensing data is used for mapping purposes. Mapping accuracy is of concern because the integrity of decisions made using a GIS database depends on the quality of the input data. The process of creating thematic maps has two major sources of error. The first is the content of the classified units of the map (i.e., is the area what the map says it is?). The second source of error is the accuracy of the location of mapped features. Classification involves mapping categories that are determined for the area covered by images or photographs. The ideal classification process uses guide-

GEOGWHIC INFORMATION SYSTEMS

77

lines and, where possible, quantitative methods to remove subjectivity introduced by the mapper. The use of a classification system aids this mapping process. Congalton (1991) examines accuracy assessment for the classification of remotely sensed data. Descriptions and a comparison of sampling schemes for conducting accuracy assessments can be found in Congalton (1988). A frequently employed system, used either out right or modified, is that of Anderson el al. (1976). The system is hierarchical in structure, where land usekovers can be categorized into subgroups according to increasingly specific criteria. The approach is similar to taxonomic systems used in soil science and other natural sciences. Although commonly used for aerial photography interpretation, the system is applicable to the classification of other remotely sensed data, when the goal is to delineate mapping units of defined and unique characteristics. The practice of classification itself varies with the source of the remote sensing data. The use of hardcopy aerial photography requires visual interpretation by an analyst to generate a thematic map. If imaged data are used, then computer-aided interpretation can be employed in the mapping process. A variety of statistical and numerical algorithms aid the classification process. Traditional approaches include the use of classification techniques such as supervised and unsupervised procedures. Supervised procedures use areas located by the analyst to identify other areas, while unsupervised procedures uses groupings of pixels in the image data when plotted in feature space. These options are available in most image processing software and some GISs. Other interpretive algorithms are available either as selectable options or can be constructed with algebraic operations internal to the system. Ratioing is a procedure that uses visible and near-infrared reflectance in vegetation indices. Ratios of visible and near-infrared reflectance have traditionally been used for vegetation indices such as the normalized difference vegetation index (NDVI). Other ratio indices have been proposed or developed, and understanding of the physical processes represented by the ratios has expanded, such as with the development of the soil-adjusted vegetation index (Huete, 1988). Improvements in hardware have also provided the basis for changes in ratiobased indices. Landsat Thematic Mapper (TM) data has two reflective infrared bands that are used in ratios to distinguish between soils containing hematite and geothite (Fraser, 1991) and between changes in the organic carbon/iron oxide ratio for erosion mapping (Frazier and Cheng, 1989). Analytic techniques and spatial statistics have also been used for interpreting landscapes. Woodcock et al. (1988a,b) discuss and apply geostatistics to relate TM reflectance to variations in soil properties. Despite the efforts of software developers to keep up with changes in remote sensing, the pace of developments sometimes outstrips the ability to incorporate new ideas. Some advances involve the development of new imaging hardware. One example is the development of the hyperspectral airborne visible/infrared imaging

78

G. W. PETERSEN ET AL.

spectrometer (AVWS), which permits fine distinctions to be made between imaged scene components. The spatial imaging processing system (SIPS) (Kruse et al., 1993) was developed to permit investigators to use techniques not available or not conveniently undertaken with existing image processing systems or GISs. Some of these techniques are described in a special issue of Remote Sensing of Environment (1993), dedicated to AVIRIS. Notable among the presented works for those interested in soils and vegetation is the article by Roberts et al. (1993). The use of nonlinear mixing in spectral mixture analysis enables the separation of soil and plant contributions to scene reflectance and improves estimates of vegetation and shade fractions. The combination of estimates of shade and nonphotosynthetic vegetation improves the discrimination of vegetation communities. Other processing techniques used with AVIRIS data enable finer distinctions of mineral composition than was previously possible with older instruments. In the past, image processing results were imported into the GIs. Today, some GIS software packages have image processing capabilities, making the map generation process an internal operation. The reverse is also true, with image processing systems taking on some GIS operations. True do-it-all systems do not exist, but increasing integration of image processing and GIS software is the trend. The trend toward integration makes remote sensing data processing and manipulation more convenient for users, but an understanding of the error in the final product is important. De Gloria (1991). The second aspect of accuracy in mapping involves locational precision. Photogrammetric techniques are used to determine coordinates for products generated from aerial photographs (Moffitt and Mikhail, 1980). A general approach is to use stereoscopic photography on which georeferenced control points are visible and can be used to produce geocoded planar and topographic maps. Errors are usually assessed in terms of the root mean square error of a mapped feature from its location determined from ground survey. Procedures similar to those used for aerial photography can be adopted for satellite imagery. Rodriguez er al. (1988) and Gugan and Dowman (1988) found that the stereoscopic location of check points from SPOT imagery yielded planimetric root mean square errors of 6.0 and 17.7 m, depending on the level of preprocessing and number of control points. The vertical root mean square varied from 3.5 to 6 m, depending on the number of control points and base to height ratio. Welch el al. (1985) evaluated Landsat-4 and Landsat-5 Thematic Mapper images for terrain mapping. Errors of 20.3 pixels in planimetric correlation were found, allowing the generation of topographic maps with 100-mcontour intervals with a a33-m error. Error reduction or definition permits the use of remotely sensed data as input into GIs-based models. Vieux et al. (1988) outline an example utilizing a GIS as part of a modeling effort that employs finite element methodology. The GIS incorporates land

GEOGRAPHIC INFORMATION SYSTEMS

79

use/land cover (mapped from aerial photography) and soils and topographic information. Areas of uniform hydrologic characteristics within a watershed are then determined. These areas become the finite elements used in the calculations of overland flow. Because the elements are geographically related within the GIs, the potential exists for extending flow to predict contaminant movement within the watershed. Several types of agronomically related models use remotely sensed data. Petersen e? al. (1991b) employs land uselland cover data derived from satellite imagery to assess potential nonpoint pollution water quality impacts on a statewide basis. Jackson (1984) provides an overview of remotely sensed inputs to plant stress, growth, and evapotranspiration models. Johnston et af. (1988) utilizes multidate aerial photography to map wetlands in a cumulative impact assessment. Losses due to conversion to other land uses, both urban and agricultural, are associated with stream water quality. In most applications there is a delay in delivery of data to users. Farm management and crop forecasting and monitoring, as well as other agricultural applicacations, could benefit from real-time data delivery. The only orbiting remote sensing system that approximates real-time data delivery is the NOAA Geostationary Operational Environmental Satellites (GOES) used to track and predict weather on a global scale. Details about the system are provided to illustrate the steps needed to provide the data in a timely manner. The GOES satellites are part of an international network of geosynchronous satellites. Some were launched and are maintained by other governments, including Japan’s Global Meteorological Satellite (GMS), the European Space Agency’s METEOSAT, and Russia’s Geosynchronous Meteorological Satellite (GOMS). An international nongovernmental body, Coordination of Geostationary Meteorological Satellites (CGMS), establishes formats for data relay to ensure access by member nations. Nonmember nations, including India, Brazil, and China, either have or anticipate having their own satellites in orbit in the near future. The GOES series of instruments is gradually being supplemented by a nationwide system for the U.S.of new ground-based Doppler radar sites called NEXRAD, which are being constructed to replace the current weather radar system installed in 1957. A total of 137 NEXRAD radars will be sited at military bases, airports, and topographically advantageous locations. Each radar will provide effective coverage of two radii, 230 and 460 km. Data updates are provided every 6 min, including preprocessing. Access to data for private users is provided at a cost by four private commercial companies; the National Weather Service uses the data at no cost. The American Meteorological Society (1990b) provides insight into the range of weather phenomena that can be monitored by using Doppler radar. Products of forecasting, either NEXRAD- or GOES-derived, are registered to geocoded

80

G. W. PETERSEN ET AL.

displays that plate forecasted events to ground areas. This is the most common function among the GIS-based uses of meteorological data. Other functions, such as the management of data for historical trend analyses and overlay of measurements of differing atmospheric layers, are also used in GIS. The sole terrestrial monitoring satellite that approaches the GOES revisit frequency is another NOAA instrument, the advanced very high resolution radiometer (AVHRR). AVHRR has a revisit capability of 1 day, but its spatial resolution of 1.1 km limits data use to large-area studies. Harris (1985) describes and outlines applications of AVHRR data, including changes in vegetation cover in Saharan Africa and the impact of a volcanic eruption on plant communities.

C. TRENDS IN REMOTESENSING Remote sensing has changed a great deal in the past 20 years and the next 10 years promise to be just as dynamic. Changes include modifications in scanner configurations, extension of known technologies, and new areas of application. Imaging systems changed rapidly in the past two decades. Aerial photography remains important, as witnessed by new guides on interpretation (e.g., Rasher and Weaver, 1990). In fact, some of the changes in imaging systems involve the desire to capture a share of the $2 billion dollars spent annually on aerial photography acquisition (Leonard, 1993). David (1995), Henderson (1993, and Budge and Morain (1995) present reviews on specific systems. The major changes in instrument systems include increasing spatial and spectral resolution. For imaging systems, spatial resolution is determined by pixel size, which is the unit of ground area imaged by the sensor's instantaneous field of view (IFOV). Table I displays the spatial and spectral characteristics of various multispectral imaging systems. Note the reduction in pixel size, increasing spatial resolution, since the 1970s. Planned systems for the multi-instrument EOS-A platform to be launched around 1998 approximate current pixel sizes. Pixel size affects interpretation of imagery. As for many applications, smaller pixel size aids interpretation. As pixel size is reduced, however, the amount of data and, therefore, computer time and storage required for processing also increase. Faster computer speed and expanding memory make data storage and access less of a problem than in the past, but future imaging systems will provide ever increasing amounts of data. Another factor in the interpretability of imagery is spectral resolution. The number and width of the spectral bands of the sensor influence the discrimination of features. Buttner and Csillag (1989) found that the extra infrared bands of the Thematic Mapper ('I'M) sensor, as compared with higher spatial resolution SPOT imagery, improved the accuracy of land uselland cover maps. Future changes in spectral discriminationwill be dramatic. The high resolution imaging spectrometer (HIRIS) instrument due to be launched as part of the

GEOGRAPHIC INFORMATION SYSTEMS

81

Table I coaeguration of selected sensors Band

Spectral resolution (prn)

Spatial resolution (rn)

Multispectral scanner (MSS), U.S.A., first launched in 1972 4 Green 0.500-0.600 80 5 Red 0.600-0.700 80 6 Near-IR 0.700-0.800 80 7 SWIRa 0.800- I . 10 80 Thematic mapper (TM), U.S.A., first launched in 1982 1 Blue-green 0.450-0.520 30 2 Green 0.520-0.600 30 3 Red 0.630-0.690 30 4 Near-IR 0.760-0.900 30 5 SWIR 1.55-1.75 30 6 Thermal IR 10.4-12.5 120 7 SWIR 2.08-2.35 30 Systeme probatoire d’observation de la terre (SPOT), France, first launched in 1986 1 Blue 0.500-0.590 20 2 Red 0.6 10-0.680 20 3 Near-IR 0.790-0.890 20 4 Panchrornaticb 0.5 10-0.730 10 Intermediate thermal infrared radiometer (ITRI), U.S.A., launched late 1990s 1 Blueb 0.520-0.600 15 2 Redb 0.630-0.690 I5 3 Near-IRb 0.760-0.860 15 4 SWIR 1.60- 1.71 30 5 SWIR 2.02-2.12 30 6 SWIR 2.12-2.19 30 2.19-2.26 30 7 SWIR 8 SWIR 2.29-2.36 30 9 SWIR 2.36-2.43 30 10 Thermal IR 8.025-8.375 90 11 Thermal IR 8.375-8.725 90 12 Thermal IR 8.725-9.075 90 13 Thermal IR 10.25-10.95 90 10.95-11.65 90 14 Thermal IR SW, short wavelength. IR, infrared. Indicates stereoscopic viewing capability.

multiinstrument EOS-A platform in the late 1990s will sense the range of 0.400 to 2.45 pm with 192 bands, dramatically increasing the capability to discriminate features on the basis of spectral properties. The previously noted airborne instrument, AVIRIS, has a similar configuration.

82

G. W. PETERSEN ET AL.

In addition to changes in the configuration of multispectral scanners, extensions of old technology provide new ways to gather information. These new applications include airborne video, ground-based systems, and synthetic aperture radar. Airborne video is the use of super VHS technology mounted in an airplane. The system records images with cameras fitted with special filters to permit imaging in specific spectral ranges. Although a variety of applications are possible (Agricultural Research Service, 1991; Petersen et al., 1991a), agriculture is a primary focus of the systems currently being examined. Similarities with orbiting multispectral instruments make the interpretation of VHS images a familiar exercise. Applications include determining water and pest stress on crops, plant nutrient deficiencies, natural disaster impacts, and crop inventory. Advantages of the super VHS technology are the quick turnaround time from image acquisition to delivery and the relatively low cost. The Agricultural Research Service (1991) notes that aerial photography film processing may cost $200-1000 per roll, compared to $15 for a tape. After imaging is complete, it can be immediately viewed and visually interpreted. Aerial photography and satellite imagery have delays of days or weeks. Although the current platform is an airplane, other platforms under consideration include the space shuttle and space station. Like VHS, synthetic aperture radar (SAR) was originally mounted on an airplane. It is now providing data from an orbiting platform. The European Space Agency’s Earth Resources Satellite 1 (ERS-1) was launched in 1991 with a C-band (1.5 GHz) HH polarized instrument. The U.S. put SAR instruments in space in 1978 and 1981. Unfortunately, both had technical problems that limited data collection. More SAR instruments will be put into space during the 1990s according to Leberl (1990). The advantage of using an active instrument like SAR, rather than passive systems like TM and SPOT, is its ability to acquire data through clouds and in darkness. The disadvantage is the amount of processing required. The Canadian Center for Remote Sensing is undertaking the processing and distribution of ERS-1 data, in preparation for the launch of Canada’s RADARSAT. The utility of SAR data for agricultural applications is still under investigation. Major areas of investigation include soil moisture measurements and land coverlland use mapping. Progress in these areas has been made but problems still exist, making routine quantitative interpretation difficult. Remote sensing is a valuable method for collecting elevation data that can be used for terrain investigations. The acquisition of elevation data was traditionally performed by using aerial photography. The differences in the position of a feature on the photograph is related to the feature’s height. Elevation data is now acquired on an orbiting basis by using SPOT 10-m panchromatic data acquired at different angles of incidence on sequential passes over an area. The soon to be launched RADARSAT will have topographic mapping capa-

GEOGRAPHIC INFORMATION SYSTEMS

83

bilities. The use of radar for topographic mapping is advantageous, compared with visible and near-infrared sensors, as it acquires data from areas frequently covered by clouds. Leberl (1990) describes the process of topographic mapping with synthetic aperture radar. An alternative to imaging-based topographic mapping is the use of laser-based altimeters. NASA has a range of airborne altimeters under evaluation as candidate instruments for orbiting platforms. Remotely sensed elevation data enables investigators to conduct hydrologic studies of currently unmapped areas and topographic investigations of dynamic land surfaces such as volcanoes and slide-prone areas. Not all remote sensing systems involve airborne or orbiting instruments. Ground-based systems, including ground-penetrating radar (GPR) and conductivity meters, are useful for the collection of subsurface soil information. Ground-penetrating radar is applied for a variety of purposes. Doolittle and Asmussen (1992) review 10 years of GPR use by the Soil Conservation Service and the Agricultural Research Service. General application areas include quality control for soil surveys, location and description of specific soil properties and features, and improved soil-landscape modeling. GPR uses antennas operating with center bandwidth frequencies of 80, 100, 120,300,400,500, and 900 MHz, with the most common being 120 MHz. The units are primarily towed behind a vehicle at slow speeds. In rugged or forested terrain, GPR equipment is hand towed. Doolittle (1987) found that the use of GPR was optimal in areas of highly weathered soils with coarse or moderately coarse texture, low proportions of 2:l type clays and concentrations of soluble salts, and depth-to-bedrock distances of 0 to 4 m. A second group of ground-based instruments, conductivity meters based on electromagnetic (EM) induction, is valuable in the detection of subsurface soil properties and features. These instruments measure average electrical conductivity for a given soil depth. Conductivity varies with water content, amount and type of salts, and amount and type of clays (Bmne and Doolittle, 1990). The depth-weighted values depend on the spacing between transmitter and receiver coils and the electrical frequency of the instrument. Rhoades et al. (1989) describe calibration equations used for deriving electrical conductivity, based on differential orientation of the instrument parallel and perpendicular to the soil surface. The application of electromagnetic induction meters includes a range of soil, environmental, and geological investigations. Williams and Baker (1982) utilized an EM unit to map areas of potential salinity hazard. It was hoped that conductivity could be related to the salinity of individual soil horizons, but it was not possible without supplemental deep soil core data. Brune and Doolittle (1990) found that EM provided a valuable subsurface mapping tool for the detection of plumes of animal wastes from storage lagoons. Zalasiewicz et al. (1985) utilized an EM unit to map subsurface geology in East Anglia, UK.

84

G. W. PETERSEN ET AL.

Results established boundaries between different lithologies under variable depth till, which masked bedrock transitions. As expected, the EM unit worked best in areas of bedrock and till with high conductivity contrasts. For both applications, it was necessary to obtain ancillary information to correlate conductivity readings with the subsurface feature under study. The GIS.use of such data is aided by the collection of coordinate information with the instrument readings. Sampling in the previously mentioned studies was done according to grids. Readings were taken at known grid locations and stored so that the trends in the data could be analyzed. Maps generated from such analyses can be entered into a GIS and correlated with other data. Not all trends in remote sensing involve hardware. The increasing variety of data sources provides researchers with an opportunity to combine them (Chavez et a f . , 1991; Morain, 1990). This approach emphasizes the ability of GIS to integrate different types of data. A number of works were undertaken using satellite imagery and digital elevation models (DEMs) for soil survey and detection of soil properties. Su ef af. (1989, 1990)and Lee er al. (1988) found that the use of DEMs with SPOT and TM imagery, respectively, improved the classification of soil mapping units. Although the analysis was conducted with image interpretation software, the trend in merging the software with GIS may lead to interpretation within a GIS or the display of results by a GIs. The modeling of global and regional energy and water balances requires even more extensive combinations of remotely sensed and other instrument-collected data. The Monsoon ‘90project (WaferResources Research, 1994) is an example of such a multi-instrument, multitemporal investigation in which remote sensing plays a vital role in acquiring synoptic information about the dynamics of the land surface. GIS provides the method by which the collected data is managed for subsequent operations. By combining the instrument advances with the progress in processing and analytical techniques, remote sensing should be a growing source of useful input to GIS databases.

lV.TERRAIN ANALYSIS AND SOIL-LANDSCAPE MODELING The demand for reliable information on patterns of soil variability has expanded to a wide range of human activities and will continue to expand in the future. The ultimate success of many intensive land uses, such as precision agriculture, is dependent on the ability to precisely define the variability of certain soil properties at more detailed scales than are employed by conventional soil surveys. At the other end of the spectrum, ecological questions arising from interest in global change are driving the examination of soil as part of the global

GEOGRAPHIC INFORMATION SYSTEMS

85

terrestrial ecosystem. Quantification of spatial relationships of observable landscape attributes to soil spatial variability is a scientifically based and potentially cost-effective means of both mapping soils and understanding pedology at multiple scales. The quest to refine the understanding of soil-landscape relationships is a central theme in pedology since its emergence as a distinct field of scientific inquiry in the late 19th-century (Tandarich and Sprecher, 1994). The formalization of the factors of soil formation by Jenny (1941) highlights the importance of topography as a key influence on soil morphological diversity and soil genesis. Milne’s (1935, 1936) pioneering work in East Africa crystallized the catena concept through his demonstration of the close relationships that may occur among landform, hydrology, and soil. Systematic frameworks for investigating soil-landscape relationships emerged in the 1950s (Ruhe, 1956; Butler, 1959). The Ruhe-Butler approach is based on geomorphic and stratigraphic principles and explicitly recognizes the periodic nature of soil-landscape relationships. This approach strongly influenced many subsequent soil-landscape studies and provided the backbone for modem soil survey (Olsen, 1989). Hewitt (1993) emphasizes the importance of soil-landscape models for predicting soil classes and their spatial arrangement, or soil properties and their trends, from soil-landscape features. Opportunities to describe the threedimensional (3D) complexity of the soil-landscape continuum and its temporal variability are limited by a lack of tools rather than concepts. GIS and its attendant spatial analytical and display capabilities provide a platform for supporting a new era of soil-landscape research. McSweeney et al. (1994) and Slater et al. (1994) provide conceptual and methodological frameworks for building 3D soil-landscape models that involve the use of GIs, scalevariable experimental design, and a selection of spatial analytical techniques. Ventura et al. (1995) provide a review of data structures that show promise for linking 3D soil horizon reconstruction with terrain model surfaces. A key component of these approaches is the development of landscape attributes from analyses of digital terrain models (Dikau, 1989; Moore et al., 1991). If reliable and co,,sistent relationships can be established between landscape attributes and underlying soil, then these relationships may be used more broadly within surrounding areas that have similar geological, pedological, and ecological histories. In this context, terrain models serve as the key for broadening the utility of the soil-landscape model. Two general methods used to quantitatively describe topographic surfaces in GIS applications include triangular irregular networks (TINS)and digital elevation models (DEMs). A TIN (Mark, 1975) divides the surface into a series of irregularly shaped triangles conforming to the shape of the land surface. Although this approach provides an accurate portrayal of the topographic surface,

86

G. W. PETERSEN ET AL.

the calculation of certain terrain attributes is complicated by the varying size of the unit elements, and concurrent spatial modeling with other data layers is confounded. Alternatively, DEMs, sometimes called digital terrain models (DTMs), are regular arrays of elevation observations depicting the shape of the topographic surface. Several researchers (Moore et al., 1993a; Bell er al., 1992; Dikau, 1989; Jenson and Domingue, 1988) developed algorithms to calculate primary terrain attributes (i.e., slope gradient, aspect direction, plan and profile slope curvature, specific catchment area, flow paths, watershed divides, and proximity variables) and secondary attributes (i.e., compound terrain index, stream power index, and depression proximity index). These terrain variables describe specific characteristics of the topographic surface and can be displayed three-dimensionally (Fig. 3). DEMs are available through the U.S. Geological Survey at 1:24,000 scale (30-m grid spacing) for portions of the United States and at 1:25O,OOO scale (approximately 100-m grid spacing) for the entire country (Elassal and Caruso, 1983). Higher resolution DEMs are required to model certain hydrologic processes associated with soil genesis and are created by field surveys (Bell et al., 1994a) or photogrammetric techniques. Differential global positioning system (DGPS) technology may provide the capability to develop high resolution DEMs in the future, particularly for agricultural lands. Bell et al. (1992, 1994b) developed a statistical model that relates soil drainage class to eight landscape parameters describing slope morphology, proximity to surface drainage features, and soil parent material. The model drew upon an extensive GIS database for an area of the ridge and valley physiographic province of south-central Pennsylvania. The model was calibrated and validated and found to be in slightly better agreement with field observations than the conventional soil map provided. Soil drainage class probability maps were generated to document the uncertainty of model estimations. The research integrated pedology, statistical modeling, and GIS for soil resource inventory. The authors pointed out that GIS technology and statistical modeling are not intended to replace the soil surveyor’sjudgment, but rather to enhance decision making and to complement experience and judgment in soil resource inventories. Moore et al. (1993a) and Bell et al. (1994a) provide examples of similar GISbased approaches that strongly rely on terrain model analysis to predict spatial patterns of A-horizon thickness, organic carbon content, and depth to carbonates. Pennock et al. (1994) employ a quantitative, 3D classification of landscape derived from terrain model analysis to address scale-variable changes in indicators of soil quality due to cultivation in a gently rolling, hummocky till landscape of Saskatchewan. The research draws upon early work (Pennock er al., 1987) demonstrating the strong relationship that exists between small (5 X 5 m) slope segments and soil distribution in order to define larger landform element complexes characterized by distinctive hydrological/pedogenic regimes. Landform element complexes were used as the basic spatial unit of analysis for assessing

GEOGRAPHIC INFORMATION SYSTEMS

87

changes in soil quality within the complexes. Soil redistribution within the landscape was assessed by using spatial patterns of cesium 137. GIS-based techniques promise to assist many facets of basic and applied aspects of soil-landscape research. The array of data types that GISs can support for spatiotemporal analysis of landscape phenomena and attributes provides a framework for much more rigorous and integrated analysis of patterns and processes operating in terrestrial ecosystems than was previously possible. The technology provides a flexible approach for querying and displaying soil information. The challenge is to formulate sound research questions that can utilize the technology. Burrough (1993, p. 550), pointed out the following. Our knowledge of the geological, geomorphological, soil-forming processes in a landscape gives us good reason to think that certain processes do operate only at certain discrete scale (e.g., formation of volcanoes) while others (e.g., wind and water erosion) operate at a wide range of scales. Because many landscapes are the result of many processes operating simultaneously and in historical sequence, it is no wonder that patterns of variation on the earth’s surface are so complex. It is easy to criticize the rigid approaches to soil classification and survey. However, alternative approaches must address the scale and variability issues that Burrough highlights if the quality and utility of soil information are to improve to meet the increasingly diverse array of applications for which it is being used (e.g., global change research, contaminant transport, and sustainable land management).

V. SITE-SPECIFIC FARMING Site-specific farming (SSF) is farm management based upon variable soil and microclimate conditions that occur within most fields. Although fields have variable conditions, conventional management is usually uniform. In SSF,fertilizers, pesticides, crop varieties, and management practices are precisely matched with land and climate attributes, providing economic and environmentalbenefits. Site-specific farming reduces waste, because fertilizer and herbicide, for example, are applied only where needed. The conventional application of fertilizer often supplies either too much or too little fertilizer on specific soils. New opportunities are now available for replacing uniform management with variable management. These derive from 1.) microcomputer capabilities 2.) soil information sources

a

Fpgure 3. (a) Three-dimensional view of 10-m (horizontal resolution) digital elevation model for a 3-ha hill slope near Lancaster, WI (3X vertical exaggeration). (b) Slope gradient draped over the three-dimensional surface. (c) Aspect direction draped over the three-dimensional surface. (d) Profile slope curvature (in direction of maximum slope gradient) draped over the three-dimensional surface.

88

89

90

G. W. PETERSEN ET AL.

3.) enhanced imagery for soil mapping (Long et af., 1989; Smith et af., 1989) 4.) comprehensive and readily available climatic databases 5.) GIS and GIS-based models for terrain analysis and landscape management (Moore et af., 1991, 1993b; Nielsen et al., 1990; Smith et al., 1989) 6.) decision support systems (DSS)for crop management (Petersen et af., 1993) and environmental hazard assessment 7.) variable-rate fertilizer and pesticide applicators (Larson and Robert, 1991) 8.) precise positioning systems, such as global positioning systems (GPS) for field navigation (Larsen et af., 1989; Petersen, 1991). Spatially variable management practices have been called farming soils not fields, precision farming, farming by soil, soil-specific crop management, grid fanning, farming by the foot, spatially prescriptive farming, computer-aided farming, and variable-rate technology (Abelson, 1990; Carr er af., 1991; Larson and Robert, 1991; Luellen, 1985; Mann, 1993; Peck, 1990). Together, these practices have received extensive coverage by the media (Peck, 1990; Petersen, 1991; Reichenbergerand Russnogle, 1989; Richter, 1991; Wilson, 1990). Public reaction was usually favorable, because the practices are intuitively sound, both economically and environmentally.

A. CURRENTAPPLICATIONSOF SITE-SPECIFIC FARMING On-farm use of individual SSF components, such as variable-rate fertilizer applicators, is growing rapidly. Integration of many SSF components into a comprehensivefarming system, however, is just beginning. Nearly 200 variablerate fertilizer applicators are in service across the United States. They are usually operated by farm service companies under contract to fanners. One service company in Illinois contracted 28,000 ha in 1992 for variablerate fertilizer application and information mapping. New customers accounted for 30% of this land (Mann, 1993). A grain fann in Indiana uses spatially variable control mechanisms for the application of all materials. Estimated savings from reduced use of starter fertilizer, dry fertilizer, ammonia, other chemicals, and seed was $33.78/ha in 1991. Chemical inputs were reduced, especially where groundwater contamination might occur (Macy, 1993). An area in Missouri, threatened by nutrient and pesticide leachate from intensive cropping practices, is the site of comprehensive SSF applications. Some 3600 ha were grid sampled and 2500 ha fully processed for variablerate fertilizer application. Phosphorus levels from C20 to >200 kg/ha and potassium levels from C 100 to >400 kg/ha were not uncommon. Plans for this prescription farming project included expansion to 10,OOO ha, use of

GEOGRAPHIC INFORMATION SYSTEMS

91

GPS for locating soil sample sites and field boundaries, and bar-code readers for sample identification (Holmes, 1993). Variable fertilizer application on eight potato fields in Washington increased costs over conventional application by $5.98/ha, for an average 3% increase in fertilizer applied and $23.45/ha for additional management services. These increases were small compared with average costs of production, which totaled about $5000/ha. With minor increases in yield and improved crop quality, the returns for variable management could be up to $7001000/ha, which are far more than the costs (Hammond, 1993). Variable-rate fertilizer application on eight projects in Montana, Minnesota,

North Dakota, and Missouri indicated that variable-rate fertilizer management cost was $lO-l7/ha higher than a conventional, single rate per field program. The higher cost included additional sampling and soil analysis, data management, map making, and use of special equipment (Wollenhaupt and Buchholz, 1993). The value of managing within field variability depends upon the magnitude of spatial variability, the value of materials applied, the cost of differential application, and the value of environmental benefits (Forcella, 1993). The greatest societal benefits of site-specific application of agricultural chemicals might be in the maintenance or improvement of soil and water quality, but the practice will not be widely accepted unless it is also profitable. Demonstrations of present SSF practices do not consistently show increased profitability. Low profitability can often be attributed to inability to recommend optimum practices (e.g., nitrogen rates) for different soil management units within fields. In their review of this subject, Wollenhaupt and Buchholz (1993, p. 210) discuss limitations associated with SSF,but said that they “strongly believe variable rate fertilization shows promise as a practice that will be profitable.” They identify crop management, based on a soil yield potential map, combined with N, P, K, and lime applications, based on soil nutrient grid sampling, as offering the greatest potential for efficient fertilizer use.

B. EMERGING ROLEOF GIs AND GPS GIS and GPS technology is advancing rapidly and is driven by major nonagricultural markets. This reduces development costs to agriculture. GIS/GPS, combined with real-time sensing of soil variability within fields, may reduce the cost of acquiring precise maps of soil organic matter and plant available nutrients. Research on devices that sense soil, crop, and weed conditions is a major thrust in agricultural science and engineering. GPS and GIS technology may soon allow more efficient weed control, for example, by treating only sites mapped as weed infested rather than treating whole fields, by spraying weeds at

92

G. W. PETERSEN ET AL.

night, and by returning to sites when weeds are vulnerable to treatment but difficult to find by conventional methods. Precise GPS/GIS automated navigation systems, foreseen in a few years, will reduce skips and overlaps in tillage, materials added, and seeding (Borgelt, 1993; Larsen er al.. 1994). Traffic lanes could be precisely designated, thereby reducing compaction where plants are grown. Precise GPS/GIS navigation systems will also allow night operations that increase the safety (reduced drift and lower rates) and effectiveness (temperatures) of chemical applications. Night harvesting improves the quality of some harvested crops. Ascheman (1993, p. 86), an agricultural consultant, concludes that “site-specific crop management is now at the cutting edge of applied science in food and fiber production. We are approaching or are now at the critical mass stage where logarithmic growth of this technology is about to start.”

C. COMPONENTS OF GIS/GPS SITE-SPECIFIC FARMING The first SSF component, GIS data layers (Fig. 4), requires that someone (consultant, service agent, farmer) inventory, monitor, and map spatial data. Data from the first component are applied to the second component (models) to make interpretations and predictions. In the third component, the farmer or advisor uses a decision support system (DSS) to integrate results from modeling and to plan for different management zones within fields. The management plan is applied in the field, for example, by taking soil samples, tilling, planting, fertilizing, controlling weeds, and even disposing of waste, according to the plan. The first cycle of SSF components and actions is complete when crop yields, resource conditions (e.g., nutrient status), and pest distributions are monitored and recorded for use in the next cycle. All SSF components from the inventory and processing of site-specific data through modeling, DSS, and field applications rely upon the generation of digitized maps (i.e., geographic data files). The locations of all attributes, conditions, predictions, constraints, and management plans are recorded according to an accurate designation of latitude, longitude, and elevation. GPS and GIS technologies offer a means of georeferencing points where data are acquired and where treatments are applied in the field. “Determining a position, to within a few centimeters if necessary, is going to be very straight forward and easy in the future” according to Qler (1993, p. 164). The following is a description of current and developing SSF components.

1. GIS Data Layers. a. Crop yield maps are accurate and inexpensive mapping methods that use monitors on crop harvesters, facilitate selection of fields for SSF application, and

GEOGRAPHIC INFORMATION SYSTEMS

93

FLgure 4. Site-specific farming represented as a repeating cycle of actions (inner ring) and components (outer ring).

provide feedback loops to evaluate SSF treatments (Schnug el al., 1993). Several yield monitor/mapping devices for crop harvesters are on the market and others are anticipated. b. Soil arrribure maps are essential to SSF. These maps express the variability of specific attributes that are biologically or environmentally significant and useful for the development of simulation models and management plans. Soil survey maps and databases are readily available sources of estimated soil attribute data, but the map resolution is often too coarse for SSF (Robert, 1993). However, data from these sources might be enhanced with terrain attribute data computed from fine resolution digital elevation models (DEMs). In the future, DEMs will be easily developed from GPS surveys. Figure 5 shows the latitude and longitude (x and y in UTM coordinates) of the x , y, z triplets measured during a kinematic GPS survey of a 20-ha field in north-central Montana. The survey was conducted in 1989 by using GPS receivers in differential mode. A stationary receiver was located at a benchmark elevation site and a roving receiver was mounted on a vehicle. The goal was to generate a DEM for use in the analysis of autocorrelated terrain, soil, and crop variables. The data presented were obtained

G. W. PETERSEN ET AL.

94

1 . , . , . 1

1000 7700

I100

1900

(a)

1300

(b)

Ffgure 5. (a) GPS sampling points. (b) Isometric projection of the GPS-derived DEM for a field in north-central Montana.

after less than 1 h of field work. GPS/GIS methods offer a promising, costeffectivemeans of creating the fine resolution (about 1:6OOO scale) maps needed for SSF (Moore et a f . , 1993b). The conventional method of obtaining field information by discrete sampling within a field, based on a grid, is expensive. Stein et a f . (1988a) reported that kriging reduced the number of samplingpoints by 33% without a reduction in the accuracy of predictions. The original survey used a boring density that was a function of the scale of the soil map. Methods are needed to select appropriategrid sampling sizes and to interpolate values between sampling points. Mulla (1993) effectively applied kriging methods to SSF. The development of semivariograms for kriging is computationally intensive, and the method was not established as more appropriate for crop management than other simpler techniques such as block averaging (Searcy and Motz, 1992). Use of soil maps, field records, and modeling would reduce the number of samples required (Stein et al., 1988b). Research continues on devices that continually sense soil and plant conditions as implements move through fields (Borgelt, 1993). These devices, coupled with GPS-based navigation and GIS technology, could produce fine resolution soil attribute maps. c. Microclimate attribute maps show within field variations in net solar radia-

GEOGRAPHIC INFORMATION SYSTEMS

95

tion, temperature, and water supply as influenced largely by terrain position. These could be considered characteristics of specific soil individuals or sites. Fine resolution microclimate attribute maps could be generated by modifying GIS climate database values (Nielsen et al., 1990) on the basis of local digital terrain models (Moore et al., 1993b). d. Crop condition mups represent crop development and health at specified times using remote sensing methods. These maps supplement data often acquired by on-site field scouting procedures. Crop condition can be documented with cameras, video recorders, or multispectral scanning devices. Near-infrared aerial photography is especially effective. However, the phenological stage of crop development when the image is acquired may be more important than the kind of film or remote sensing device used. For small grain crops, the best time is when about one-third of the field has started to ripen, changing from green to yellow (Long et al., 1989). Unfortunately. timing is often constrained by weather conditions (cloud cover). Acquisition from satellites is further constrained by the timing of orbits and the wide range of crop maturity that occurs within a single remote sensing frame or record. GIS techniques are needed to extract information from images. Estimation of crop conditions requires both ground truth acquisition and human expert interpretation. Readily available imagery from space satellites (Landsat series, TM, or SPOT) lacks the precision needed for most SSF applications. As with aerial photography, ground-truthed observations are needed to determine what is actually represented in the image. Hand-held imaging spectroscopingmeters, such as CROPSCAN, are quite useful for field application (Finke, 1992). Remote sensing is ideal for monitoring seasonal changes in crop conditions and long-term changes in field variability. Leaf area index data could be used to update crop growth estimates from models (Finke, 1992). In short, the technology for acquiring images from remote sensing is well developed and commonly available, but better methods are needed to convert the images into high resolution records of crop conditions for planning and evaluating SSF treatments. Furthermore, little expertise has been available for image analysis in rural areas. SSF high technology management systems will require skilled consultants capable of using remote sensing techniques efficiently. e. Environmental condition mups represent field variations in soil and water quality and provide a baseline record against which changes in quality can be measured. Accurate GPS-based georeferencing allows successive sampling at selected locations. Appropriate methods of monitoring environmentalchanges in farm fields have not been developed. But periodic measurements of water quality in the upper vadose zone would document the impact of soil-specific management. f. Pest distribution maps will most likely be generated by a combination of remote sensing techniques and recognition of infestations by people operating

96

G. W. PETERSEN ET AL.

field implements equipped with GPS/GIS navigation and mapping devices. Automated recognition of weeds within crops remains a challenging problem, but entering weed names and cover information with GPS/GIS and a computer keyboard can be accomplished (Mangold, 1990). g. Management constraint maps show, for example, herbicide use areas, compacted areas, and areas with high potential for erosion or leaching. These maps can be developed from farm records, soil attribute maps, and accurate surveys, assuming that all locations are recorded in a common geodetic referencing system. 2. Models

Simulation models are available to predict water and solute movement in combination with crop growth (Finke, 1992; Wagenet and Hutson, 1989). An integrated analysis of physical, chemical, and biological processes is preferred for SSF, because the use of separate models presents problems in integration. For SSF, a deterministic model is needed that reflects different soil conditions within fields, as opposed to generalized capacity models that are more suitable for regional applications. The major driving variables for deterministic models are weather conditions, groundwater fluctuations, and soil parameters, such as hydraulic conductivity and soil water retention. They are measured or, more realistically, estimated from site-specific GIS data layers (Wagenet et al., 1991). Simulation models need to accommodate spatial dimensions and areas to a field GIs.

3. Decision Support Systems DSSs would employ the power and memory of computers to help farmers and consultants integrate information from many sources, synthesize management plans, and evaluate them. An SSF decision support system could link sitespecific attribute data with models and GIS technology (Petersen et al., 1993). AEGIS, an agricultural and environmentalgeographic information system envisaged by the International Benchmark Soil Network for Agrotechnology Transfer (IBSNAT, 1992), is an example, but it has not been adapted for SSF applications. The DSS management plan must also incorporate water and soil quality goals that are driven by farmers’ personal concerns about the quality of the land that supports them and the water they drink, as well as by laws designed to protect public interests in soil and water quality. A GISIGPS-based DSS that includes models and a record keeping system would help farmers develop plans that meet their goals and maintain records of their actions. The primary output of a DSS is a field management plan map composed of management zones, each with a specified rate of material application or treat-

GEOGRAPHIC INFORMATION SYSTEMS

97

ment. The smallest management unit or management cell size is determined by equipment size and the rate at which treatments can be adjusted. Commercially available GIS software merges data stored at a variety of cell sizes and produces new records with cell widths that can be adjusted to conform with equipment widths. Site-specific data are stored to represent the precise points at which they are acquired. Data are not averaged before they are stored.

4. Field Applications Management plan maps provide direction to field applications such as the following: a. Tilling soils at variable depths and intensities would, for example, save energy, reduce equipment wear, and leave variable amounts of crop residue near the surface. More residue would reduce erosion on steep sites, and less residue would favor seed germination on cold, wet sites. b. Sowing seed varieties adapted to wet or dry parts of fields, changing sowing depth according to soil texture, and adjusting seeding rate in response to yield goals illustrate other potential applications of SSF technology. c. Fertilization with equipment that varies fertilizer and application rates across fields according to yield goals, soil fertility levels, and environmental constraints is already a commercially established practice. Manure and lime can also be applied at variable rates. d. Crop protection (i.e., control of weeds, insects, and diseases) with variable treatmenthate applicators offers many potential economic and environmental benefits. Applications occur only where pests are located or expected on the basis of site conditions. Environmentally sensitive areas are avoided or receive safe alternative treatments. Rates are adjusted according to maps of soil pH, organic matter, texture, leaching potential, and other attributes. e. Waste disposal is accomplished with variable-rate applicators and SSF technology in order to make applications according to variable soil and microclimate conditions within fields. Furthermore, the load distribution of waste materials in the field is recorded. This record constitutes site-specific documentation showing that regulations were met. It also provides input for models that predict nutrient flux, leaching potential, and productivity. SSF will be the agricultural system of the 21st century. It is made possible by the development of technologies such as microcomputers, positioning systems (e.g., GPS), and GISs. A major challenge in optimizing SSF is the development of information systems and decision support systems based on GISs to efficiently merge field information databases and provide sound management recommendations to producers.

98

G. W. PETERSEN ET AL.

VI. E"MENTAL APPLICATIONS Geographic information systems are used for a variety of agronomic environmental applications. GIS, combined with modeling capabilities, offer an efficient and effective means of identifying and understanding environmental problems (Petersen e?al., 1991a). One of the major areas of environmental concern in the agricultural community is the nonpoint pollution problem. The Federal Clean Water Act, administered by the EPA, requires states to develop assessments of nonpoint source water quality problems and management plans to remediate nonpoint pollution problems. Because resources for combating nonpoint source pollution are limited, states are required by the EPA to develop a list of high priority, or critical, water bodies on a watershed basis (Hession et al., 1992). High priority regions receive preference for funds to help alleviate nonpoint sources of pollution. Recent emphasis is directed to groundwater pollution. Contamination of groundwater results from the generation of potential pollutants and the transfer of these pollutants from the source to the receiving aquifer. GIS, combined with modeling capabilities, offer an efficient means of identifying and ranking nonpoint pollution potential for both surface water and groundwater. Various agencies are starting to rely on these models to help them direct programs and limited monies available for reducing nonpoint source pollution. It is a well established fact that the inherent soil composition of a given geographic region greatly influences the amount and type of nonpoint source pollution emanating from that region via surface water runoff. Similarly, the inherent properties of the soil types comprising the land surface also determine the ease with which potential contaminants introduced to the surface (either intentionally or unintentionally)can migrate to underlying groundwater. Because the spatial distribution of properties, such as soil erodibility (K factor) and soil permeability, can vary substantiallyover relatively short distances, it is necessary to account for spatial variability in evaluating the pollution potential of a given geographic region. GIS technology is ideally suited for this type of spatial analysis. GIS techniques are available that allow for the incorporation of soil factors in watershed-based analyses of nonpoint source pollution and in regionalscale empirical modeling of surface and groundwater pollution potential.

A. REGIONAL-SCALE POLLUTIONASSESSMENT The assessment of surface and groundwater pollution potential using GISbased empirical modeling techniques is quite widespread due to the ease with which this type of analysis can be accomplished. This type of spatial modeling is

GEOGRAPHIC INFORMATION SYSTEMS

99

not based upon a rigorous simulation of physical, chemical or biological processes, but rather uses weighted indexing schemes to quantify the relative influence of various factors that contribute to pollution problems. One example of this type of analysis is the use of GIS for evaluating potential groundwater contamination from agricultural pesticides (Hamlett et al., 1994). Another example is the nonpoint source assessment technique described by Petersen et al. (1991b) and Hamlett et al. (1992). In Hamlett et al. (1994), GIS software was used to assess the susceptibility of public groundwater supply systems throughout the state of Pennsylvania to contamination from agricultural pesticides. A groundwater pollution potential map was created for the entire state by using the DRASTIC ranking methodology developed by the U.S. EPA (1985). DRASTIC rates the inherent pollution potential of an area on the basis of the hydrogeological, topographical, and soil characteristics that are mapped for that area. For each of over 60oO groundwater supply wells, the likelihood of contamination from pesticides was evaluated by using a rather complex sequence of GIS operations that involved generating buffer zones around wells, overlaying multiple GIS files, and cross-referencing GIS data with external tabular files (or “look-up” tables) that contained information on pesticide usage by crop type and crop type distribution by county. The leachability of various pesticides was also considered via use of the groundwater ubiquity score (GUS) rating scheme developed by Gustafson (1989). With GUS,

3 (0

3

8

5>

Depth to water Net recharge Aquifer media Soiltype Topography Impact of the vadose zone Hydraulic conductivity

x

3 Pesticide use by crop type Q)

.U c,

M

Crop type distribution Cropland acreage

Q

x

CI

rr.v = .-Q Pesticide soil half life -r U

L

Pesticide linear adsorption coefficient for organic carbon

Figure 6. Procedure to determine the relative susceptibility of groundwater subbasins to pesticide contamination.

100

G. W. PETERSEN ET AL.

the inherent leaching potential of pesticides is rated on the basis of persistence (soil half-life) and mobility (adsorption coefficient for organic carbon). GUS is linked with pesticide use and the DRASTIC rankings to provide an overall assessment of groundwater subbasin pesticide pollution susceptibility (Fig. 6). In Petersen et af. (1991b), a GIs-based methodology for ranking the relative contribution of sediment, nutrients, and pesticides from 104 watersheds in Pennsylvania was developed. The transport of these pollutants was considered by estimating the runoff and sediment delivery from the 100 X 100-m cells used in the digital map database. This database included information on watershed boundaries, land use/cover, animal densities, topography, soils, precipitation, and rainfall runoff factors. An agricultural pollution potential index (APPI) was derived by overlaying and analyzing these data layers, and each of the 104 major watersheds in the state was ranked according to its relative nonpoint pollution potential (see Fig. 7). Each of the preceding projects addressed nonpoint source pollution potential solely from a surface water or groundwater perspective. It is possible, however, to integrate the results from both to more comprehensivelydepict statewide NPS pollution potential on a watershed basis. For example, as shown in Fig. 7, GISs could be used to derive a single index value incorporatingboth surface water and groundwater pollution from nonpoint sources.

Runoff Index

D f ~ d

Sediment Production Index

-

f \ Surface Water A ricultural Lution Potential

Spatial Analysis

L

E:mical Index

Watershed Index of

k Pesticide Susceptibili~ Index Nitrate

Suxeotibilitv

Groundwater A ricultural gollution Potential

Numerical Analysis

2A 28

....

101

....

Indez--

F l p r e 7.

Schematic diagram of agricultural pollution potential screening in Pennsylvania.

GEOGRAPHIC INFORMATION SYSTEMS

101

B. WATERSHED-BASED NONFOINT SOURCEASSESSMENT Several approaches to using GIS for evaluating the generation and movement of pollutants such as nutrients, sediment, and pesticides within and through a watershed are possible. One of the most common approaches involves parameter estimation. In this case, the objective is to determine and quantify parameters that can be used as input to watershed models via the manipulation and analysis of various temin-related data sets. With this approach, GIS software typically is used to overlay preprepared digital maps depicting soils, land use/cover, and topography to produce area-weighted estimates (on either a watershed or subarea basis) for a number of parameters commonly used in watershed-based simulation models. Example parameters for which estimates are derived in this fashion include curve number, percent impervious surface, evapotranspiration, soil moisture storage, roughness coefficient, and average land slope. The resulting estimates normally need to be reformatted for input into the watershed model although, as described in the following, other more automated data transferal mechanisms are possible. Figure 8 illustrates conceptually how a raster GIS package might be used to compute the area-weighted value for a given input variable for a watershed model such as AGNPS (agricultural nonpoint source) that uses rectangular subareas. In this case, various GIS routines are used to overlay a digital AGNPS subarea layer with another layer depicting a particular parameter, sum the parameter cell values within the area bounded by each AGNPS cell (i.e., subarea), and calculate the average value (Evans and Miller, 1988). It is not necessary that the subarea cells and GIS cells be the same size. In fact, the spatial resolution of GIS data cells is almost always an order of magnitude finer than that of the AGNPS subarea cells. Another way in which a GIS can be used to derive hydrologic parameters is via linkage to a library of georeferenced parameter values. For example, the SWRRBWQ (Simulation for Water Resources in Rural Basins-Water Quality) model for simulation of the water resources of rural basins has a library of weather parameters defined for about 100 weather stations in the U.S., so that estimates of required climatic variables can be automatically extracted for modeling purposes (Arnold et al., 1990). Likewise, for soils information, SWRRBWQ detailed data on soil properties for hundreds of soil types as depicted on countylevel SCS soil maps. Descriptions of how GIS was used to automate the parametrization process for this particular model are provided by Evans et al. (1992) and Rosenthal et al. (1993). It is also possible to perform some hydrologic modeling directly within a GIs, so long as time variability is not an issue. This is the case when considering annual averages of variables, such as annual average flow or pollutant loadings

G. W. PETERSEN ET AL.

102

I I I

I

4 I I

Conceptual Grid Overlay I

I I

Columns

I

Gridded CIS Cell Values

Averaged ACNPS Cell Values

Flgure 8. Conceptual representation of creating grid cell data within a GIS and deriving averaged parameter values for AGNPS cells.

from a watershed. For example, one could implement spreadsheet-type models in which flows or loadings are computed as flow or load per unit area multiplied by the area (Evans et al., 1994). This type of analysis is accomplished by a database management module of the type found with most of the more popular GIS packages. One could also implement more complex equations, such as those for pollutant loadings derived from a regression, where the independent variables in the regression equations are mapped in coverages and then the loadings are worked out on the basis of mathematical combinations of coverage data. For example, the amount of soil loss occumng in a given watershed is to a large degree a function of soil texture and amount of precipitation. For a given analysis, a GIS macroprogram could be written that, upon execution, overlays a soil layer with a precipitation layer and applies different preprogrammed equations to different aerial combinations of the two data sets to predict soil erosion and removal rates. As described earlier, GIs-based watershed modeling is accomplished in sev-

GEOGRAPHIC INFORMATION SYSTEMS

103

era1 ways, including direct modeling within a GIS. Hydrologic modeling capabilities and/or functions are incorporated into very few widely used GIS packages. l b o notable exceptions, however, are GRASS and ARC/INFO. GRASS (Geographic Resource Analysis and Support System) is public domain software! that was developed by the U.S. Army Corps of Engineers and is used by many government agencies, institutes, and universities within the United States. ARCIINFO is a commercial package developed by ESRI, Inc., that is probably the most widely recognized GIS software in the world. GRASS is primarily a grid cell-based system with some vector handling capability. ARC/INFO was initially created as a vector-based GIS package, but was upgraded to include extensive grid cell processing capabilities. Over the last few years, both packages were enhanced with fairly powerful watershed modeling capabilities, including such functions as automated drainage area and flow path delineation, stream ordering, flow accumulation within subareas, and network-based flow modeling, to name a few. In the case of GRASS, such functions were primarily developed by end users and distributed in shareware fashion within the GRASS user community. In the case of ARC/INFO, however, ESRI has made a corporate commitment toward the continuing development of specialized hydrologic analysis and modeling tools. Concomitant with ESRI's effort, several third-party software developers began to independently enhance the already considerable hydrologic modeling capabilities of ARC/INFO. For example, Innovative Software Developers, Inc., of suburban Baltimore, MD, developed a product called GeoStorm that has the ability to directly run several SCS watershed models, such as TR-20 and TR-55, within a completely menu-driven ARC/INFO environment. Future plans for this product call for the incorporation of the Army Corps of Engineers' HEC-1 and HEC-2 models as well. A similar package that allows execution of the U.S. EPA's SWMM model within ARC/INFO was developed by the University of Oregon.

C. SPATIAL DECISION SUPPORTSYSTEMS Most GIS packages are quite sophisticated and powerful in terms of their analytical capabilities, and many provide literally hundreds of software operations and functions. Although this almost overwhelming availability of options may be very appealing to GIS experts, it can be a source of frustration for GIS novices. Similarly, many decision makers would like to take advantage of the problem-solving capabilities of GIS, but lack the time needed to keep up with changes in GIS software and related technologies. To satisfy the needs of these types of users, GIS designers and consultants are developing customized applications of GIS called spatial decision support systems (SDSS).

104

G. W. PETERSEN ET AL.

In a typical SDSS,the user is guided through a series of data manipulation and analysis steps via the use of a customized user interface. For many more straightforward applications, such an interface can often be constructed by using macroprogramming facilities provided within various CIS packages. In more complex cases, an SDSS may combine the technologies of GIs, decision support systems (DSS),and mathematical modeling. The deductive power of logic inference systems is typically used in artificial intelligence (AI) systems, and many researchers in the GIS arena believe that similar methods are now needed to help integrate and analyze data from different sources within a CIS environment. However, very limited use of AI, spatial query language, or expert systems capabilities has been made so far in designing GIS software. As a rule, an SDSS should be fairly focused in terms of its intended user community, but it should allow for relatively easy completion of a number of both spatial and nonspatial data operations. For example, an SDSS oriented toward the evaluation of nonpoint source pollution problems within a state might provide for the following functions: Quickly display a multitude of NPS-related data sets such as soils, land use/cover, topography, geology, streams, etc. Track the status of NPS mitigation projects by watershed Evaluate NPS problems at the watershed level by using user-defined empirical or physically based modeling techniques Extract water quality data from another database (e.g., STORET) and display it by watershed or stream reach Compare water quality data between areas by using statistical graphing and data visualization techniques Produce standardized data summaries and graphics from a predefined menu It is possible, of course, to build an SDSS for a very wide range of CIS applications. The more successful ones tend to be those that provide the most flexibility in terms of analyses and output, while at the same time requiring the least amount of prior CIS experience from the user.

VII. CONCLUSIONS CIS technology is bringing about rapid changes in the way that agronomic analysis and management are being conducted. CIS coupled with remote sensing, GPS, electronic sensors, and computer technologies is providing new methods for data acquisition, storage, processing, analysis, and modeling. These new tools allow us to quantitatively describe landscapes and processes. However, new and/or improved models need to be developed to fully take

GEOGRAPHIC INFORMATION SYSTEMS

105

advantage of the spatial nature of the data provided by these tools. The development of these models will rely on spatial statistical analysis techniques to quantify the accuracy of input parameters and model output. Many new tools are being used in this rapidly evolving field of GIS. Threedimensional scene simulation, visualization, and animation linked with remote sensing and image processing technologies and real time data collection will be needed in the study of agronomic systems. The development and use of threedimensional GIS and spatiotemporal GIS will be an increasingly important area of research. The agronomic community, including farmers, land managers, fellow scientists, policymakers, and the general public should benefit from this evolving and expanding field.

ACKNOWLEDGMENTS The authors acknowledge Eric D. Warner and Barry M. Evans for their contributions to this chapter. We are also grateful to Tawna Mertz and Joy Robert for their editorial assistance and to Michelle Barnyak for the preparation of this manuscript.

REFERENCES Abelson, P. H. (1990). Dialog on the future of agriculture. Science 249(4968), 457. Agricultural Research Service. (1991). Airborne video-A Ride on ARS’ own “Air Force 1”. Agric. Res. 39(2), 4-9. American Meteorological Society. (1990a). “Weather satellites: Systems, data, and environmental applications” (P. K. Rao, S . J. Holmes, R. K. Anderson, J. S . Winston, and P. E. Lehr, eds.). American Meteorological Society, Boston, MA. American Meteorological Society. ( 1990b). “Radar in meteorology” (D. Atlas, ed.). American Meteorological Society, Boston, MA. Anderson, 1. R., Hardy, E. E., Roach, 1. T.,and Witmer, R. E. (1976). A land use and land cover classification system for use with remote sensor data. Geological Survey Professional Paper 964. U.S. Geological Survey, U.S. Department of the Interior, Washington, D.C. Arnold, J. G.,Williams, J. R., Griggs. R. H., and Sammons, N. B. 1990. SWRRBWQ-A basin scale model for assessing management impacts on water quality. USDA Agricultural Research Service. Ascheman, R. E. (1993). Some practical field applications. In “Soil Specific Crop Management” (P. C. Robert, R. H.Rust, and W. E. Larson, eds.), pp. 79-86. ASA. CSSA, and SSSA, Madison, WI. Bell, J. C., Cunningham, R. L., and Havens, M. W. (1992). Calibration and validation of soillandscape model for predicting soil drainage class. Soil Sci. SOC. Am. J . 56, 1860-1866. Bell, J. C., Thompson, J. A., Butler, C. A., and McSweeney, K. (1994a). Modeling soil genesis from a soil-landscape perspective. In “Transactions 15th World Congress of Soil Science” (J. D. Etchevers, A. Aguilar, R. NiiBez, G. Alchtar, and P. Shchez. eds.), Vol. 6a, pp. 179-195. Acapulco, Mexico. Bell, J. C., Cunningham, R. L., and Havens, M. W. (1994b). Soil drainage class probability mapping using a soil-landscape model. Soil Sci. SOC.Am. J . 58, 464-470.

106

G. W. PETERSEN ET AL.

Borgelt, S. C. (1993). Sensing and measurement technologies for site specific management. I n “Soil Specific Crop Management” (P. C. Robert, R. H. Rust, and W. E. Larson, eds.). pp. 141-157. ASA, CSSA, and SSSA, Madison, WI. Brune, D. E., and Doolittle, J. (1990). Locating lagoon seepage with radar and electromagnetic survey. Environ. Geol. Water Sci. 16(3), 195-207. Budge, A. M., and Morain, S. A. (1995). Access remote sensing data for CIS. CIS World February, 45-49. Burrough, P. A. (1986). “Principles of geographical information systems for land resource assessment.” Oxford University Press, New York. Burmugh, P. A. (1993). Soil variability: A late 20th century review. Soils Fertilizers 56, 529562. Butler, B. E. (1959). Periodic phenomena in landscapes as a basis for soil studies. Soil Publ. 14. CSIRO, Div. of Soils, Canberra, Australia. Buttner, G . , and Csillag, F. (1989). Comparative study of crop and soil mapping using multitemporal and multispectral SPOT and Landsat Thematic Mapper data. Remote Sensing Environ. 29,241249. Cam, P. M., Carlson. G.R., Jacobsen, J. S.,Nielsen, G.A., and Skogley. E. 0. (1991). Farming soils, not fields: A strategy for increasing fertilizer profitability. J . Prod. Agric. 4, 57-61. Chavez, P. S., Jr., Sides, S.,and Anderson, J. A. (1991). Comparison of three different methods to merge multiresolution and multispectral data: Landsat TM and SPOT panchromatic. Phorogram. Eng. Remote Sensing 57(3), 295-303. Congalton, R. G.(1988). A comparison of sampling schemes used in generating ermr matrices for assessing the accuracy of maps generated from remotely sensed data. Photogram. Eng. Remote Sensing 54(5), 593-600. Congalton. R. G. (1991). A review of assessing the accuracy of classification of remotely sensed data. Remote Sensing Environ. 37, 35-46. Corbley, K.P. (1995). Iowa mapping firm meets growing customer needs with softcopy photogrammetry. Earth Observation Magazine, February. Dangermond, J. (1990). A classification of software components commonly used in geographic information systems. I n “Introductory readings in geographic information systems” (D. 1. Peuquet and D. F. Marble. eds.), pp. 30-51. Taylor and Francis, New York. David, L. (1995). Private firms unveil high-resolution satellite ventures. Earth Observation Mugazine NovemberDecember. 26-28. Day, R. L., Petersen, G.W., and Miller, D. A. (1988). The base map dilemma: Solution through digital cartography and remote sensing. “Proc. International Interactive Workshop on Soil Resources: Their Inventory, Analysis and Interpretation for Use in 1990’s.” pp. 40-50. Minnesota Extension Service, University of Minnesota, St. Paul, MN. De Gloria, S. D. (1991). Elements of geographic information systems for resources conservation.I n “Challenges in Conservation of Biological Resources” (D. J. Decker, M. E. Krasny, G.R. Goff, C. R. Smith, and D. W. Gross,eds.). Westview Press, Boulder, San Francisco, Oxford. Dikau, R. (1989). The application of a digital relief model to landform analysis in geomorphology.I n “Three Dimensional Applications in Geographic Information Systems” (J. Raper, ed.), pp. 5177. Taylor and Francis, New York. Doolittle, J. A. (1987). Using ground-penetrating radar to increase the quality and efficiency of soil surveys. I n “Soil Survey Techniques” (W. U. Reybold and G.W.Petersen, eds.), Soil Science Society of America Special Publication 20, pp. 11-32. Doolittle, J. A., and Asmussen, L. E. (1992). Ten years of ground penetrating radar by the United States Department of Agriculture. Geological Survey of Finland, Special Paper 16 (P. Hanninen and S. Autio, eds.), pp. 139-147. Fourth International Conference on Ground Penetrating Radar, June 8-13, 1992.

GEOGRAPHIC INFORMATION SYSTEMS

107

Elassal, A. A., and Caruso, V. M. (1983). Digital elevation models. U.S. Geological Survey Circular 895-B, USGS Digital Cartographic Data Standards. U.S. Dept. of the Interior, Geological Survey, Reston, VA. Evans, B. M., and Miller, D. M. (1988). Modeling nonpoint pollution at the watershed level with the aid of a geographic information system. In “Roc. Symp. on Nonpoint Pollution: 1988-Policy, Economy, Management and Appropriate Technology” (V.Novohy, ed.), pp. 283-291. American Water Resources Association. Evans, B. M., Grimm, J., Thomton. L.. and Sanders, P. (1992). Linking CIS with hydrologic modeling. In “Pmc. Water Forum ’92,” pp. 499-504. Baltimore, MD. Evans, B. M., White, R. A., Petersen, G. W., Hamlett, J. W., Baumer, G. M., and McDonnell, A. I. (1994). Land use and nonpoint pollution study of the Delaware River Basin. Environ. Resources Res. Inst. Penn State University, ER94-6, 88 pp. Finke, P. A. (1992). Integration of remote sensing data in the simulation of spatially variable yield of potatoes. Soil Technol. 5, 257-270. Forcella, F. (1993). Value of managing within-field variability. In “Soil Specific Crop Management” (P. C. Robert, R. H. Rust, and W. E. Larson, eds.), pp. 125-132. ASA, CSSA, and SSSA, Madison, WI. Fraser, S.J. (1991). Discrimination and identificationof ferric oxides using satellite thematic mapper data: A Newman Case Study. Inr. J . Remore Sensing 12(3). 635-641. Frazier, B. E., and Cheng, Y. (1989). Remote sensing of soils in the Eastern Palouse Region with Landsat Thematic Mapper. Remote Sensing Environ. 28. 317-325. Gugan. A. J., and Dowman, I. J. (1988). Topographic mapping from SPOT imagery. Photogram. Eng. Remore Sensing 54(10), 1409-1414. Gustafson, D. I. (1989). Groundwater ubiquity score: A simple method for assessing pesticide leachability. Envimn. Toxicol.Chem. 8 , 339-357. Hamlett, I. M., Miller, D. A., Day, R. L., Petersen, G. W., Baumer, G. M., and Russo, J. (1992). Statewide CIS-based ranking of watersheds for agricultural pollution prevention. J. Soil Water Conserv. 47(5). 399-404. Hamlett, J. M., Petersen, G. W., Harrison, S.,Baumer, G. M., Deichert, L. A., Messier, S. R., and Anderson, M.C. (1994). Assessment of ground water pollution using a geographic information system. In “Transactions 15th World Congress of Soil Science” (J. D. Etchevers, A. Aguilar, R. Ndiiez, G. Alcbtar, and P. Shchez, eds.), Vol. 6b, pp. 335-336. Acapulco, Mexico. Hammond, M. W. (1993). Cost analysis of variable fertility management of phosphorus and potassium for potato production in central Washington. In “Soil SpecificCrop Management” (P. C. Robert, R. H. Rust, and W. E. W o n , eds.), pp. 213-228. ASA, CSSA, and SSSA, Madison, WI. Harris, R. (1985). Satellite remote sensing: Low spatial resolution. Prog. Phys. Geogr. 9(4), 600606. Henderson, F. B., I11 (1995). Remote sensing for CIS. CIS World February, 42-44. Hession, W. C., Flagg, J. M., Wilson, S. D., Biddix, R. W., and Shanholtz, V.0. (1992). Targeting Virginia’s nonpoint source programs, ASAE Paper No. 92-2092. American Society of Agricultural Engineers, St. Joseph, MI. Hewitt, A. E. (1993). Predictive modeling in soil survey. Soils Feriilizers 56, 305-314. Holmes, B. (1993). Prescription farming. In “Soil Specific Crop Management” (P. C. Robert, R. H. Rust, and W. E. Larson, ed~.),pp. 311-315. ASA, CSSA, and SSSA, Madison, WI. Huete, A. (1988). A soil adjusted vegetation index (SAW). Remore Sensing Envimn. 25,295-309. IBSNAT. (1992). Linking DSSAT to a geographic information system. International Benchmark Sites Network for agrotechnology transfer. Agmtechnol. Transfer IS, 1-6. Jackson, R. D. (1984). Remote sensing of vegetation characteristics for farm management. Soc. Phoro-Optical Instrum. Eng. 475. 81-96.

108

G. W. PETERSEN ET AL.

Jenny, H. (1941). “Factors of soil formation: A system of quantitative pedology.” McGraw-Hill, New

Yo*. Jenson. S. K.,and Domingue, J. 0. (1988). Extracting topographic structue from digital elevation model data for geographic information system analysis. Phorogram. Eng. Remote Sensing 54, 1593- 1600. Johnston, C. A., Detenbesk, N. E.,Boule, J. R., and Niemi, G . J. (1988). Geographic information systems for cumulative impact assessment. Phorogram. Eng. Remote Sensing 54(1 I), 16091615.

Kruse, F. A., Lefkoff, A. B., Boardman, J. W., Heidebrecht, K. B., Shapiro, A. T., Barloon, P. J., and Goetz, A. F. H. (1993). The Spectral Image Processing System (SIRS)-interactive visualization and analysis of imaging spectrometer data. Remote Sensing Envimn. 44, 145-163. Larsen, W. E.,Abousabe, A., nler, D., and Nielsen, G.A. (1989). Tests with the field navigation system. Paper PNR-89-10s. American Society of Agricultural Engineers, St. Joseph, MI. Larsen, W. E., Qler, D. A.. and Nielsen, G. A. (1994). Precision navigation with GPS. In “Computers and Electronics in Agriculture” Vol. 1 I , pp. 85-95. Elsevier Science Publishers, Amsterdam. Larson, W.E.. and Robert, P. C. (1991). Farming by soil. I n “Soil Management for Sustainability” (R. Lai and F. J. Pierce, eds.), pp. 103-1 I 1. Soil Water Conserv. SOC., Ankeny, IA. Leberl, F. W. (1990). “Radargrammetric image processing.” Artech House, Norwood, MA. Lee, K.-S.,Lee, 0 . B., and vier, E. J. (1988). Thematic mapper and digital elevation modeling of soil characteristics in hilly terrain. Soil Sci. Soc. Am. J . 52, 1104-1 107. Leonard, D. (1993). Private firms unveil high-resolution satellite ventures. Earrh Observation Magazine, November/December. Long, D. S., Nielsen. G.A., and Carlson, G. R. (1989). Use of aerial photographs for improving layout of field research plots. Appl. Agric. Res. 4(2), 96-100. Luellen, W. R. (1985). Fine-tuned fertility: Tomorrow’s technology here today. Cmps Soils 38(2), 18-22. Macy, T. S. (1993). Macy farms-site-specific experiences. I n “Soil Specific Crop Management” (P. C. Robert, R. H. Rust, and W. E. Larson. eds.), pp. 229-244. ASA, CSSA, and SSSA, Madison, WI. Mangold. G . (1990). Bytes and beans. Soybean Digest February, 24. Mann, J. (1993). Illinois FS variable rate technology: Technology transfer needs from a dealer’s viewpoint. I n “Soil Specific Crop Management” (P. C. Robert, R. H.Rust, and W.E. Larson, eds.), pp. 317-323. ASA, CSSA, and SSSA, Madison, WI. Marble, D. F. (1990). Geographic information systems: an overview. I n “Introductory Readings in Geographic Information Systems” (D. J. Peuquet and D. F. Marble, eds.), pp. 8-17. Taylor and Francis, New York. Mark, D. M. (1975). Computer analysis of topography: A comparison of terrain storage methods. Geogr. Ann. 57A, 179-188. McSweeney. K.,Gessler, P. E., Slater, B., Hammer, R. D., Bell, I., and Petersen, G. W. (1994). Towards a new framework for modeling the soil-landscape continuum. In “Factors of Soil Formation: A Fiftieth AMiVersary Perspective” (R. G. Amundsen, M. J. Singer, and J. W. Hardin, eds.), Soil Sci. SOC. Am. Spec. Publ. No. 33, pp. 127-145. SSSA, Madison, WI. Milne, G.(1935). Some suggested units of classification and mapping particularly for East Africa. Soil Res. 4, 3. Milne, G.(1936). Normal erosion as a factor in soil profile development. Nature (London) 138, 548549. Moffitt, F. H., and Mikhail, E. M. (1980). “Photogrammetry.” Harper and Row, Publishers Inc., New York. Moore, 1. D., Grayson, R. B., and Ladson, A. R. (1991). Digital terrain modeling: A review of hydrological, geomorphological, and biological applications. Hydml. Proc. 5. 3-30.

GEOGRAPHIC INFORMATION SYSTEMS

109

Moore. 1. D., Gessler. P. E.. Nielsen, G. A., and Peterson, G. A. (1993a). Soil attribute prediction using terrain analysis. Soil Sci. Soc.Am. J. 57, 443-452. Moore, I. D., Gessler. P. E., Nielsen, G. A., and Peterson, G. A. (1993b). Terrain analysis for soilspecific management. In “Soil Specific Crop Management” (P. C. Robert, R. H. Rust, and W.E. Larson, 4 s . ) . pp. 27-55. ASA. CSSA, and SSSA. Madison, WI. Morain, M. S. (1990). A window-based technique for combining Landsat Thematic Mapper thermal data with higher-resolutionLandsat Thematic Mapper multispectral data over agricultural lands. Photogrum. Eng. Remote Sensing 56(3), 337-342. Mulla, D. J. (1993). Mapping and managing spatial patterns in soil fertility and crop yield. In “Soil Specific Crop Management. Research and Development Issues. Proc. Workshop” (P. C. Robert, R. H. Rust, and W. E. Larson, eds.), pp. 15-26. ASA, CSSA, and SSSA, Madison, WI. Nielsen, G. A., Caprio, J. M., McDaniel, P. A., Snyder, R. D., and Montagne. C. (1990). MAPS: A CIS for land resource management in Montana. J. Soil Water Conserv. 45(4), 450-453. Olsen, C. G. (1989). Soil geomorphic research and the importance of paleosol stratigraphy to Quaternary investigations, Midwestern USA. In “Paleopedology: Nature and Application of Paleosols. Catena Suppl. 1 6 (A. Bronger and J. Can, eds.). pp. 129-142. Catena Verlag h b l . , Cremlingen-Destedt, The Netherlands. Peck, C. (1990). Recision fanning coming in the ‘90s. Montana Farmer-Srockman November, 6-9. Pennock, D. J., Zebarth, B. J., and de Jong, E. (1987). Landform classification and soil distribution in hummocky terrain. Geoderma 40,297-3 15. Pennock, D. J., Anderson, D. W., and de Jong, E. (1994). Landscape-scale indicators of soil quality due to cultivation in Saskatchewan, Canada. Geoderma 64, 1-19. Petersen, C. (1991). Precision GPS navigation for improving agricultural productivity. GPS World 2(1), 38-44.

Petersen, G. W., Miller, D. A., Day, R. L., Sasowsky, K. C., and Evans, B. M. (1990). An introduction to geographic information systems and their role in soil and hydrologic studies. In “Proceedings of Soil Erosion and Productivity Workshop” (W. E. Larson, G. R. Foster, R. R. Allmaras, and C. M. Smith, eds.), pp. 105-116. University of Minnesota, Saint Paul, MN. Petersen, G. W.. Nielsen, G. A., and Wilding, L. P. (1991a). Geographic information systems and remote sensing in land resources analysis and management. Suelo Planr. 1, 531-543. Petersen, G. W., Hamlett, J. M., Baumer. G. M., Miller, D. A., Day, R. L., and Russo, J. M. (1991b). Evaluation of agricultural nonpoint pollution potential in Pennsylvania using a geographic information system. Environmental Resources Research Institute, University Park, PA. Petersen. G. W., Russo, J. M., Day, R. L., Anthony, C. T., and Pollack, J. (1993). Importance of spatial variability in agricultural decision support systems. In “Soil Specific Crop’’ (P. C. Robert, R. H. Rust, and W. E. Larson, eds.), pp. 167-179. ASA, CSSA, and SSSA, Madison, WI . Rasher, M. E., and Weaver, W. (1990). Basic photo interpretation: A comprehensive approach to interpretation of vertical aerial photographs for natural resource applications. National Cartographic Center, Soil Conservation Service, US. Department of Agriculture, Fort Worth, TX. Reichenberger, L., and Russnogle, 1. (1989). Farm by the foot. FannJ. March, 11-15. Remore Sensing of Environment. (1993). Special Issue: Airborne Imaging Spectrometry 44 (May/June), 107-356. Rhoades, J. D., Lesch, S. M., Shouse, P. J., and Alves, W. J. (1989). New calibrations for determining soil electrical conductivitydepth relations from electromagnetic measurements. Soil Sci. Soc. Am. J . 53, 74-79. Richter, S. (1991). Fanning soils: applying plant food by bits and by bytes, pp. 1-4. Cooperative Partners, Cenex/Land O’Lakes Ag. Service, St. Paul, MN. Robert, P. C. (1993). Characterization of soil conditions at the field level for soil-specific management. Geoderma 60. 57-72.

110

G . W. PETERSEN ET AL.

Roberts, D. A.. Smith, M. O., and Adams, J. 8. (1993). Green vegetation, nonphotosynthetic vegetation, and soils in AVIRIS data. Remore Sensing Environ. 44, 255-269. Rodriguez, V., Gigod, P., de Ganjac, A. C., Munier, P.. and Begni, G. (1988). Evaluation of the stereoscopic accuracy of the SPOT Satellite. Photogram. Eng. Remore Sensing 54(2), 217-221. Rosenthal, W. D., Srinivasan, R., and Arnold, I. (1993). A GIS-based watershed hydrology model link to evaluate water resources of the Lower Colorado River in Texas. I n “Roc.Application of Advanced Information Technologies.” Spokane, Wash. Ruhe, R. V. (1956). Geomorphic surfaces and the nature of soils. Soil Sci. 82, 441-445. Schnug, E., Murphy, D., Evans, E., Haneklaus, S., and Lamp, I. (1993). Yield mapping and application of yield maps to computer aided local resource management. I n “Soil Specific Crop Management”(P. C. Robert, R. H. Rust, and W. E. Larson, eds.), pp. 87-93. ASA, CSSA. and SSSA, Madison, WI. Searcy, S. W., and Motz, D. S. (1992). Engineering technology for soil specific crop management. In “Soil Specific Crop Management. Research and Development Issues. Proc. Workshop” (P.C. Robert, R. H. Rust, and W.E. Larson, eds.), pp. 181-196. ASA, CSSA, and SSSA, Madison, WI. Slater, B. K., McSweeney, K., McBramey, A. B., Ventura, S. J., and Iwin, B. (1994). A spatial framework for integrating soil-landscape and pedogenic models. I n “Quantitative Modeling of Soil Processes” (R. B. Bryant and R. W. Arnold. eds.), Soil Sci. Soc. Am. Spec. Publ. No. 39, pp. 169-185. SSSA. Madison, WI. Smith, S. M.. Schreier, H. E., and Brown, S. (1989). Analysis of forage crops using geographic information system and image analysis techniques. Inaugural Colloquium of the Spatial Information Research Centre, Dunedin, New Zealand. Stein, A., van Dooremolen, W.,Bouma, J., and Bregt, A. K. (1988a). Co-kriging point data on moisture deficit. Soil Sci. SOC.Am. J . 52, 1418-1423. Stein, A., Hoogemerf, R., and Bouma, I. (1988b). Use of soil map delineations to improve cokriging of point data on moisture deficit. Geoderma 43, 163-177. Su. H., Ransom, M. D., and Kanemasu, E. T. (1989). Detecting soil information on a native prairie using Landsat TM and SPOT satellite data. Soil Sci. SOC. Am. J . 53, 1479-1483. Su, H., Kanemasu, E.T., Ransom, M. D., and Yang, S.-s. (19%). Separabilityof soils in a tallgrass prairie using SPOT and DEM data. Remote Sensing Environ. 33, 157-163. Tandarich, J. P., and Sprecher, S. W. (1994). The intellectual background for the factors of soil formation. In “Factors of Soil Formation: A Fiftieth Anniversary Perspective” (R. G. Amundsen. M. J. Singer, and J. W. Hardin. eds.), Soil Sci. Soc. Am. Spec. Publ. No. 33, pp. I 13. SSSA, Madison, WI. bier, D. (1993). Positioning technology (GPS). I n “Soil Specific Crop Management” (P. C. Robert, R. H. Rust, and W. E. Larson, eds.), pp. 159-165. ASA, CSSA, and SSSA, Madison, W1. U.S.EPA. (1985). DRASTIC: A standardized system for evaluating ground water pollution potential using hydrogeologic settings. EPA/600/2-85/018, 163 pp. Ulaby, F. T., Moore. R. K.. and Fung, A. K.(1981). “Microwave remote sensing active and passive. Volume I: MicrowaveRemote Sensing Fundamentalsand Radiometry.” Artech House, Dedham, MA. Ulaby, F. T., Moore, R. K., and Fung, A. K. (1982). “Microwaveremote sensing active and passive. Volume 11: Radar Remote Sensing and Surface Scattering and Emission Theory.” Artech House, Dedham, MA. Ulaby, F. T., Moore. R. K.. and Fung, A. K.(1986). “Microwave remote sensing active and passive. Volume 111: From Theory to Applications.” Artech House, Dedham. MA. Ventura, S. J., Irvin, B. J., Slater, B. K., and McSweeney. K. (1995). Data structures for representation of soil stratigraphy. I n “GIs and Environmental Modeling: Progress and Research Issues” (M. Goodchild, er al., eds.), Natl. Ctr. for Geographic Information and Analysis (NCIGA) Spec. Publ., in press. GIS World, Boulder, CO.

GEOGRAPHIC INFORMATION SYSTEMS

111

Vieux, B. E., Bralts, V. F., and Segerlind, L. J. (1988). Hydrologic modeling using the finite element method and geographic information systems. American Society of Agricultural Engineers. Wagenet, R. L., and Hutson. J. L. (1989). LEACHM: Leaching estimation and chemistry model: A process-based model of water and solute movement, transformations, plant uptake and chemical reactions in the unsaturated zone. Version 2.0, Continuum Vol. 2. Water Resources Institute, Cornell Univ., Ithaca, NY. Wagenet, R. J.. Bouma. J.. and Grossman, R. B. (1991). Minimum data sets for use of soil survey information in soil interpretive models. In “Spatial Variabilitiesof Soils and Landforms’’ (M. J. Mausbach and L. P. Wilding, eds.),Soil Sci. Soc. Amer. Spec. Publ. 28, pp. 161-182. SSSA, Madison. WI. Warer Resources Research. (1994). Special Issue. Monsoon ‘90Experiment 30(5), 1209-1393. Welch, R., Jordan, T. R., and Ehlers, M. (1985). Comparative evaluations of the geodetic accuracy and cartographic potential of Landsat-4 and Landsat-5 Thematic Mapper image data. Photogram. Eng. Remote Sensing 51(11), 1799-1812. Williams, B. G., and Baker, G. C. (1982). An electromagnetic induction technique for reconnaissance surveys of soil salinity hazards. Austral. J. Soil Res. 20, 107-118. Wilson, W. (1990). Precision farming. Prairie Farm Report Video. Ag-Com Productions, Great Plains Building, Regina, Saskatchewan, Canada. Wollenhaupt, N. C., and Buchholz, D.D.(1993). Profitability of farming by soils. In “Soil Specific Crop Management” (P. C. Robert, R. H. Rust, and W. E. Larson, eds.), pp. 199-211. ASA, CSSA. and SSSA, Madison, WI. Woodcock, C. E., Strahler, A. H.,and Jupp, D. B. (1988a). The use of variograms in remote sensing: I. Scene models and simulated images. Remore Sensing Environ. 25, 323-348. Woodcock, C. E., Strahler, A. H., and Jupp, D. B. (1988b). The use of variograms in remote sensing: 11. Real digital images. Remote Sensing Environ. 25, 349-379. Zalasiewicz, J. A., Mathers, S. J., and Cornwell, J. D. (1985). The application of ground conductivity measurements to geological mapping. Q . J . Eng. Geol. 18, 139-148.

This Page Intentionally Left Blank

USDA PLANTGENOME RESEARCH PROGRAM USDA Plant Genome Research Program Participants' USDA, Agricultural Research Service, BARC-West, Building 005,Room 331C, Beltsville, Maryland 20705

I. Introduction A. History and Program Establishment B. Program Goals, Structure, and Operation 11. Progress A. Summary of Four Years: 1991- 1994 B. Graminae C. Leguminosae D. Cruciferae E. Malvaceae: Cotton (Gossypium spp.) F. Solanaceae G. Woody Species 111. Plant Genome Database A. Introduction B. History C. Discussion of the PGD Information Resources D. Accessing the PGD Information Resources E. ACEDB Iv.Future Projections References

I. INTRODUCTION A. HISTORY AND PROGRAM ESTABLISHMENT The U.S. Congress appropriated funds in 1991 for the USDA Plant Genome Research Program, 4 years after its initial conception in 1987. Early in 1988, a formal proposal for the program was presented to the then Assistant Secretary for Science and Education Orville Bentley, and later in the same year the leadership role of the program was assigned to the Agricultural Research Service (ARS). A 'The names of the participants are listed under Acknowledgments. 113 in A p m y , filumr 55 Copyright 0 1995 by Academic Press, Inc. All rights of reproduction in any form reserved. &mu

114

S. ALTENBACH ET AL.

conference held in December 1988 in Washington, DC, with plant science researchers from public, private, and government institutions addressed the need for and the goals of the program. The need for a plant genome mapping effort was uniformly recognized as important to U.S. agriculture. The director of the program was appointed April 11, 1989, after which an Interagency Plant Genome Coordinating Committee was formed, and the committee met twice that same year. The committee’s consensus states that the program would not specify one agricultural species but the effort would address genes of agricultural importance. ARS was given $99,000 in “seed money” in 1990 for planning activities and formulating a proposal for presentation to USDA administration and legislative bodies. An Informatics Project Leader was hired during the same year to begin addressing database information handling of the Plant Genome Research Program, and the database effort was housed at the National Agricultural Library.

B. PROGRAM GOALS,STRUCTURE, AND OPERATION Discussions of the Plant Genome Science and Technology Coordinating Committee concluded that the project is more than mapping alone. It includes additional molecular biology techniques to pull out the gene system, characterize, and develop methods for transfer and gene expression. Because of the broad scope of the effort, the committee named the activity the USDA Plant Genome Research Program. Overriding the committee’s discourse is the paramount importance of placing the genes or map marker locations in the hands of breeders for full application of the program. The goal, therefore, of the USDA Plant Genome Research Program (PGRP)is to improve plants (agronomic, horticultural, and forest tree species) by locating marker DNA or genes on chromosomes, determining gene structure, and transferring genes to improve plant performance with accompanying reduced environmental impact to meet marketplace needs and niches (Miksche, 1991). New cultivars will offer pest and disease resistance, which reduces chemical applications, and tolerance to abiotic stresses such as heat, cold, and drought conditions. The Plant Genome Research Program is one program with two parts: (1) National Research Initiative and (2) Plant Genome Database. 1. National Research Initiative and Agricultural Research Service Cooperative Components

The grant proposals address the goal of improving plants through genome research. Although this is an applied goal, the research efforts proposed by scientists that lean toward basic research are also considered for funding by the National Research Initiative (NRI) evaluation panel. The line of demarcation

USDA PLANT GENOME RESEARCH PROGRAM

115

between basic and applied efforts in genomic research is not sharp and one cannot advance the science of one aspect without the other’s information. The Request For Proposals (RFP) directed the proposals to address three categories: (1) broad genome maps; (2) fine maps, including physical mapping; and (3) new technology development to increase the efficiency of mapping, gene localization and characterization, and sequencing of desirable genes.

II. PROGRESS A. SUMMARY OF FOURYEARS:1991-1994 The total appropriation from Congress from 1991 through 1994 was $58.79 million for the program. The National Research Initiative and the Agricultural Research Service received $46.55 and $12.24 million, respectively, for the grants and database efforts. The NRI plant genome competitive component awarded 381 grants to scientists from 84 public, private, and government research institutions. Awards covered research on 5 1 agronomic, horticultural, and forest tree species and 4 nonagricultural taxa. Within the 55 taxa, 84% of the research award dollars went to 5 plant groups: (1) tree species, $1.8 million; (2) crucifers, $4.3 million; (3) legumes, $5.9 million; (4) solanaceae, $9.8 million; and (5) grasses, $16.6 million. Over 80 gene/trait/genetic phenomena are at various stages of progress, as listed in Table I. Some important accomplishments made by the grant awardees are as follows: For the first time ever, a disease-resistant gene was located and removed from the genome by map-based cloning technology. The bacterial speck resistance trait was transferred to a susceptible variety, resulting in resistance. One researcher is part of a team that discovered a new class of genes that allows plants to recognize a diverse group of pathogens. Quantitative trait loci (QTL)methods have been used to develop a barley line resistant to barley stripe rust, and another researcher demonstrated an increase of 15% in corn yield. Forest researchers have analyzed the loblolly pine genome and mapped over 200 genetic markers as part of a tree improvement program in the southeastern United States. Tree breeders can now expedite the improvement of loblolly pines by time compression and use of genetic resources with meaningful parents and desirable offspring. The Plant Genome Database is now a real and functioning information and data resource for agricultural and other plant science genome researchers, and it is in the public domain. The preceding progress represents only a small summary

116

S . ALTENBACH ET AL. Table I Gene System or Traits, NRI Plant Genome Awards, 1991-1994 Abscission AC/DC transposons AC/DC muatgenesis and integration Alcohol dehydrogenase Amylase activation Apomixis Bacterial blight resistance Ca + modulated leaf receptor protein Chromatin structure Cytolcinin response Cytoplasmic male sterility Disease resistance Drought tolerance Floral homeotic genes and sterility Flowering Gametophytic lethals Gene targeting for excision of foreign DNA Gibberellin synthesis I n situ hybridization Insect resistance Lipid &saturation Microsatellite sequences mRNA stability Nodulin N fixation Phytochrome A mRNA degradation Plasmid-directed conjugation Polyadenylation Polyamines and stress tolerance QTLs for wood quality QTLs for yield Ribosomal protein synthesis Ripening Rust resistance Scaffold attachment Seed maturation Stable transformation Starch synthesis Targeted DNA integration Trichomes and insect resistance DNA transport and integration Winter hardiness YAC size DNA

Acylsugar biosynthesis Agrobacterium virulence Anthocyanin biosynthesis Anthocyanin biosynthesis Aphid resistance Blight resistance Centromere organization Endodormancy chilling requirement Ethylene biosynthesis Fatty acid biosynthesis Fertility Fiber quality Fiber yield Fruit quality Hessian fly resistance Influorescence development Kernel starch Kernel sucrose metabolism Leaf epidermal gmwth Leaf morphology Mildew resistance Mitochondria protein synthesis Nematode resistance Nodulation and N fixation Organogenesis Phaseolin and seed protein Photorespiration Photosynthesis Phytoalexin enzyme Plastid light response Rotein synthesis Quantitative trait mapping Rust resistance Seed oil synthesis Self incompatibility Signal transduction Transcription Ubiquitin ligation Vigor and plant morphology Virulence genes Virus resistance Wood specific gravity

USDA PLANT GENOME RESEARCH PROGRAM

117

of the findings, but illustrates the subject areas of the program. Additional progress is given in the following according to major plant groups.

B. GRAMINAE 1. Genomes of Maize and Sorghum Maize has been called “the human of the plant world,” considering that the corn genome is of similar bulk (approximately 3 billion base pairs), similar complexity (estimated 20,000-60,000 genes interspersed with much repetitive DNA), and similarly high polymorphism per locus. The haploid number of chromosomes in maize is 10, and there is compelling evidence for extensive duplications in the genome (Helentjaris et al., 1988), reflecting an ancient hybridization between two species so diverged and rearranged that the genomes in the hybrid formed an effective, allopolyploidproduct that became modem maize. The state of genome analysis in this crop is reflected in the following: Genes. defined or prospective Genes, defined to unique location or function Genes, located to chromosome Genes, sequences in GenBank Genes, mitochondrion Genes, plastid Transposable elements Mutants RFLP probles Break points Maps Quantitative trait loci

4980 1021 914 11%

38 79 61 6205 2221 235 1 152 169

Maize and sorghum are closely related to the other cereals and grasses (Bennetzen and Freeling, 1993), an insight made evident by mapping with molecular markers across species. The amount of genetic knowledge, plant breeding techniques, and genetic technology for maize is exceptional. By combining this information with other crops, prompt and efficient applications of new knowledge and new concepts among maize and the other cereals are the promise of coordinated research. Sorghum has the same number of chromosomes as maize and a genome of less than approximately 1 billion base pairs. Genetic studies have identified over 200 morphological and other variants (Melake-Berhan et al., 1993), but until the advent of molecular markers, linkage mapping has been limited by a shortage of suitable tools.

118

S. ALTENBACH ET AL.

a. Markers for Analysis and Manipulation of the Genome Dramatic advances have been produced in a short time with tools that associate traits with efficient markers for the genes controlling them. These DNA probe tools can detect polymorphisms (RFLPs) at specific places in the genome. In the public sector, random genomic probes from maize developed at Brookhaven National Laboratory (BNL) and the University of Missouri are designated with prefixes bnl and umc, respectively. in the private sector, prefixes include Native Plants Incorporated (NPI) with npi, Pioneer HiBred International with pi0 (now php), and Agrigenetics (Mycogen) with agr, among others. By marking the genome, simple and complex traits now can be dissected, analyzed, and manipulated with predictability and efficiency, multiplying the power of the breeder and the biotechnologist. Map development for maize is moving very rapidly in several laboratories, and there are several parallel, mostly equivalent maps [table in Coe and Gardiner (1994)l. The limits of computational tools needed for harmonizing and merging the data have so far delayed combined representation of all the information. Core markers (Gardiner et al., 1993), chosen for their clarity, reliability, spacing, and high rate of polymorphism, will aid harmonizationamong maps, when use is made of these markers as pegs in common among the different maps and among maps in different species. A high-precision, statisticallyqualified core map for maize, with over 700 markers, has just been generated as a standard for use in comparative mapping of genes and gene candidates probed by cDNAs, with visibly defined genes (Yerk-Davis, G. L., Grant, D., McMullen, M. D., Musket, T., Xu, G., Chao, S., and Coe, E. H. 1995, in preparation). The first published RFLP maps (Helentjaris et al., 1986a,b; Burr et al., 1988) have been subsumed in or superseded by more current maps (Gardiner et al., 1993). Additional maps are being generated in numerous populations prepared for QTL studies. Map distances among populations differ as expected because of widely differing parentages, progeny types, progeny sizes, probe numbers and probed loci, numbers of trait loci, marker coverage, and estimated genome size. The proximate order of markers varies in relatively few instances; within chance variations among samples and are open to reinterpretations or reevaluations. Maps of hybrids between maize and Zea diploperennis and between maize and Zea fuurians show some differences (Doebley et al., 1990). Maps in Sorghum bicolor, developed with probes from maize (Hulbert et al., 1990;Melake-Berhan et al., 1993; Whitkus et al., 1992;Pereira et al., 1994),and one in S . bicolor X Sorghumpmpinquum, developed largely with probes from sorghum (Chittenden e? al., 1994), have been advanced. These serve as the first genetic maps for sorghumand now have several morphological and quantitativetraits on them. Less duplication is found within the sorghum genome than with maize. b. Map Locations of Genes Encoding Key Cell Functions Several probes used in the public sector mapping projects with molecular

USDA PLANT GENOME RESEARCH PROGRAM

119

markers have been derived by techniques that genetically define them as clones for specific cell functions. Examples include genes isolated by transposon tagging or by messenger selection and reverse transcription to produce a defined cDNA. In contrast, the agr probes developed and mapped by Murray et al. (1989) are random cDNAs from tissue-specific libraries, which have been mapped with bnl and umc RFLP markers. Substantial numbers of random cDNA clones have been sequenced (Keith er al., 1993) and loci for them are being mapped (Chao et al., 1994; Personal Communication, Maize Genet. Coop. Newsl. 68, 101-104). Because these sequences can be compared against existing databanks, potential functions often can be attributed to the loci that are mapped: Among 130 clones studied by Keith et al. (l993), 18 were found to show strong similarity to genes known in maize or other species. The implication is that expansion of numbers of sequences in various species will synergisticallyexpand the knowledge of functions and candidate functions of the others, including knowledge not only to maize but from maize as well. c. Genome Structure and Synteny Understanding of the relationships among species is now greatly enhanced by finding extensive homology seen in probe hybridization across species boundaries. Second, the common order for extended segments of the genome between species that have been generally regarded as only distantly related, at best, also increases understanding of connection. Probes applied to sugarcane, foxtail millet, or sorghum display high frequencies of strong hybridization (Hulbert et al., 1990). All but 4 out of 250 maize probes showed strong hybridization with sorghum DNA (Bennetzen and Melake-Berhan, 1994). Up to 30% of the two genomes have been estimated to have common order (Whitkus et al., 1992), and the percentage is likely much higher. Such a considerable degree of synteny suggests that the genomes are moderately diverged. They may even be subject to the transfer of genetic properties from one to the other, if technologies permit some method of hybrid formation, among either whole genomes or genome segments (extensive attempts to cross maize with sorghum have been made without success; note, however, the instructive experiences with very wide crosses, in the next section). A species more closely related to maize, and crossable with it, Tripsacum dactyloides, shows considerable divergence in genome order (Blakey, 1993). Rice probes on maize and maize probes on rice display extensive regions of synteny (Ahn and Tanksley, 1993), as also found for wheat (Ahn et al., 1993). d. Very Wide Crosses Wheat has been crossed with maize pollen, after which the resulting zygotes lose the maize chromosomes in early divisions (Laurie and Bennett, 1986). This result is an effective method by which to generate haploids in wheat (Laurie and

120

S. ALTENBACH ET AL.

Bennett, 1988; Laurie ef al., 1990; Suenaga and Nakajima, 1989). Haploids are generated (Rines and Dahleen, 1990) from crosses with maize pollen in oats also, and occasionally plants are obtained that carry one or more maize chromosomes. This work promises to provide needed cytogenetic tools for the analysis and manipulation of the oat genomes, and for maize as well. Crosses between maize and sorghum, which are much more closely related, have been attempted many times but have been unsuccessful, perhaps in part because of the inhibition of pollen tube growth (Laurie and Bennett, 1989). e. Quantitative Trait Loci and Marker-Assisted Selection The inheritance of quantitative traits in maize has been studied extensively by measurements and statistical analyses, especially by analysis of variance. The design of quantitative experiments, and their interpretation, is grounded in the Mendelian behavior of large numbers of genes. The very substantial advances in analysis that have become possible because of molecular markers are due to their clarity in most applications and the irrelevance of variations in molecular markers to the genetics of the traits themselves. By the use of RFLP markers, distributed at intervals of 20 centimorgan (cM) or so, the definition of intervals carrying a gene or genes affecting a measured trait has begun to advance at a rapid pace. Experiments to map quantitative trait loci (QTL) began as soon as the technology could be applied (Beavis et al., 1991; Edwards e l al., 1992; Romero-Severson et al., 1989) and are expanding rapidly. Notable in particular is the fact that a substantial part of the variation for many traits is attributable to variations in only a few segments of the genome. These findings are a prediction of the proposal of Robertson (1989) that subliminal variations at loci known from drastic mutants (e.g., dwarfs) are a source of quantitative trait variation (e.g., plant height). Marking of genetic regions for inclusion or exclusion, to save repeated testing and to reduce population sizes that must be advanced during breeding and selection, has been one of the most sought-after consequences of map development. Now that maps are available whose coverage is adequate for the purpose, selection of targeted segments is proving effective: Demonstrated increases in yield substantially exceed those of the parent hybrid and a quality commercial hybrid (Stuber and Sisco, 1991).

2. Genome Analysis in Small Grains and Sugarcane Wheat, barley, rice, and oats are the major small grain crops, with wheat accounting for more acreage in the world than any other cereal and wheat and rice vying yearly for the most tonnage worldwide. In contrast to rice, the other small grain genomes are large and complex. Bread wheat contains one of the largest genomes of the major crop plants, about 1.6 X 10'0 bp of DNA per haploid nucleus, or approximately40 times that of rice. The genome sizes of oats

USDA PLANT GENOME RESEARCH PROGRAM

121

and barley are also large, 1.1 X 1O’O bp for oats and 5.0 X lo9 bp per haploid nucleus for barley. The large genome size of these plants coupled with the polyploid natures of both wheat and oats makes genome analysis particularly challenging in these species. a. Barley Doubled haploid populations developed from principal North American germ plasm groups have been used to generate an extensive molecular map for barley (Hordeurn vulgare L.). A cross between Steptoe, a six-rowed high-yielding feed barley, and Morex, a six-rowed barley that is used as the standard of the American malting industry, has been used to map 450 loci with an average distance of 3 (cM) between markers (Kleinhofs et al., 1993, 1994). Several additional RFLP maps have also been developed using other populations (Heun et a f . , 1991; Graner et af., 1991; Kleinhofs et al., 1994). The development of linkage maps is important for studies aimed at locating quantitative trait loci (QTL) that may be targets for map-based cloning or molecular marker-assisted selection (MMAS). On the basis of agronomic and malting quality phenotypes, data generated in 5 environments in 1991, and a 123point skeleton linkage map, Hayes et a f . (1993) located QTLs for grain yield, lodging percentage, plant height, heading date, grain protein, a-amylase, diastatic power, and malt extract in the Steptoe X Morex population. In 1992, agronomic phenotypes were assessed in 11 additional environments and malting quality traits in 5 environments (Hayes et al., 1994). b. Oats Since oats (Avena sativa L.), similar to wheat, is an allohexaploid species (genome designation AACCDD), initial efforts at genome analysis have centered around the development of RFLP linkage maps for the A genome of Avena using F3 families from a cross between two diploid species, A . atlantica and A. hiratula (O’Donoughue et al., 1992). A total of 192 RFLP markers, most derived from either oat or barley cDNA libraries, was mapped or assigned to 7 linkage groups. c. Rice Rice (Oryza sativa L.) is one of the most important food crops in the world and is a staple food for much of the world’s population. Rice is a diploid species with 12 chromosomes and has the smallest genome of any monocot known, about 4 X 108 bp per haploid nucleus. Rice has also become a model plant among the cereals for molecular genetics studies since plants can be regenerated from protoplasts and transformed at relatively high efficiencies [for reviews, see Lynch et al. (1991), Hodges et al. (1991), and Kothari et al. (1993)l. McCouch er a f . ( 1988) described the construction of the first RFLP map in

122

S. ALTENBACH ET AL.

rice. This map was constructed from an F2 population derived from a cross between varieties representing the two major subspecies (indica and japonica) of cultivated rice. Primary trisomic stacks (Khush et al., 1984) were used to assign the 12 linkage groups to their respective chromosomes. The RFLP map generated has provided the basis for tagging a number of agronomically important genes with RFLP markers, including both single genes and quantitative trait loci (QTL)linked to blast resistance (Yu et al., 1991; Wang et al., 1994), insect resistance (McCouch and Tanksley, 1991; Mohan et al., 1993), bacterial blight resistance (McCouch et al., 1991; Ronald et al., 1992), photoperiod sensitivity (Mackill et al., 1993), grain aroma (Ahn et al., 1992), wide compatibility (Liu et al., 1992; Zheng et al., 1992), and semidwarf character, sd-1 (Cho et al., 1994). A second RFLP map of rice based on a different indica/japonica cross was reported by Saito et al. (1991). A third map based on an indica/japonica cross is under development in Japan (Nagamura et af., 1993). That map is composed of a combination of genomic and cDNA markers and consists of over 1400 markers. Efforts to integrate the rice maps are underway (Xiao et al., 1992). d. Wheat The use of complementary approaches has been critical to the advancement of genome mapping efforts in wheat (Triticum aestivum L.). Hexaploid bread wheat exhibits relatively low levels of polymorphism, making RFLP linkage analysis somewhat difficult. However, wheat has the distinct advantage of having excellent cytogenetics, and the availability of extensive sets of aneuploid stocks has proven to be invaluable in genome research. Aneuploid stocks with either whole chromosomes or segments of chromosomes added or subtracted from the genome have been used successfully to develop chromosome arm maps. Anderson er al. (1992) determined the locations of 800 restriction fragments in Chinese spring that were homeologous to 210 barley cDNA, oat cDNA, and wheat genomic clones using ditelosomic and nullitetrasomic stocks. The construction of cytogenetically based physical maps of wheat chromosomes has also been possible by using deletion stocks and has facilitated analyses of recombination in defined regions of wheat chromosomes (Werner et al., 1992). RFLP linkage mapping has been performed in polymorphic diploid species such as Triticum tauschii, the D genome progenitor of hexaploid wheat (Gill et al., 1991). Partial RFLP linkage maps of hexaploid wheat have also been generated for homeologous group 2, 3, 5, and 7 chromosomes by using populations derived from wide crosses (Devos et al., 1992, 1993; Xie et af., 1993; Chao et al., 1989). Along with the development of RFLP maps for hexaploid wheat have come a number of important observations about the structure of the wheat genome. When Werner et al. (1992) attempted to integrate existing linkage maps with

USDA PLANT GENOME RESEARCH PROGRAM

123

cytogenetically based physical maps constructed by using a collection of 4 1 aneuploids containing partial arm deletions, they found that loci that were close to the centromere from genetic analysis were physically located on more distal regions of the chromosomes. This analysis indicated that recombination was suppressed in the proximal 70% of the chromosome and was high in the distal ends, resulting in a compression of the genetic map in the middle and an expansion at the ends of the chromosomes. Thus, it appears that genes tend to be clustered on the distal portions of the wheat chromosomes. Such results have important implications when map-based cloning approaches are pursued to isolate genes of agronomic importance. Perhaps the most exciting observation comes from comparative genetic mapping studies between wheat and other members of the Gramineae. Ahn et al. (1993) mapped 66 cDNA markers that had previously been assigned to wheat chromosome arms onto the 12 linkage groups of rice and found that the synteny of many loci appears to be conserved between the 2 species. A number of linkage group rearrangements could also be inferred from comparisons between the rice and wheat genomes. For example, 9 contiguous loci on rice chromosome 3 were located on wheat chromosome 4, while an additional 2 loci from the end of the rice chromosome were localized on wheat chromosome 5 , suggesting that a translocation had occurred since the divergence of rice and wheat. By comparing the results of this study with those of a previous study (Ahn and Tanksley, 1993) in which comparative maps were generated for rice and maize, the relationships between the wheat and maize genomes have been established as well. An example of these comparative mapping results is shown in Fig. 1. The observed synteny between the various species in the Gramineae should help to accelerate genome mapping in wheat, since molecular maps for both maize and rice are well developed and a greater number of isozyme and morphological loci have been identified and mapped in these species. Sets of “anchor probes” might also be developed to serve as a framework for molecular mapping studies in the various cereal species. Since the rice genome is 40-fold smaller than that of wheat, it may also be possible to exploit the rice genome when considering map-based cloning approaches for the isolation of agronomically important genes. e. Genome Analysis in Sugarcane Sugarcane presents even greater challenges than the other cereals in terms of genome analysis. Cultivated sugarcane is a genetically complex, multispecies hybrid that is generally considered to be an aneuploid of a basic octaploid. Sugarcane has a chromosome number of approximately 120, and the genome size is estimated to be greater than 3 X lo9 bp per haploid nucleus. Modem cultivars are derived from interspecific crosses between Saccharum oflcinarum

S. LTENBACH ET AL. Rice 5

Triticeae 1

Oat A

0-

20-

4-

60-

8-

loo-

120

14

-

160-

Figure 1. Comparative maps of cereal chromosomes. The synteny of molecular markers is shown for rice chromosome 5, a composite Triticeae group 1 chromosome, and chromosome A of oat. This figure is redrawn from the data of Van Deynze et al. (1995). Data for constructing the maps are from Van Deynze et al. (1995). Causse er al. (1994) (rice 5), and O’Donoughue et al. (1992) (oat). Markers in parentheses indicate low lod scores.

and Saccharum spontaneum, Saccharurn barberi, Saccharum sinense, or Saccharum robusturn, with subsequent recurrent back-crossing to the female parent. Single-dose markers that are present in one parent, absent in the other, and segregate 1:l in the progeny are useful in the genetic mapping of polyploid species where there are no known diploid relatives. By using such an approach, an RFLP linkage map of the wild sugarcane species S. spontaneum L. has been constructed. A total of 216 loci defined by 116 DNA probes from sugarcane, oat, rice, and barley cDNA libraries and sugarcane and maize genomic libraries was mapped on 44 linkage groups (da Silva et al., 1993). The 44 linkage groups were shown to comprise 8 sets of homologous chromosomes. An additional 208 loci were placed on this map by using RAPD technology (Al-Janabi et al., 1993).

USDA PLANT GENOME RESEARCH PROGRAM

125

C. LEGUMINOSAE The Leguminosae is the third largest family of flowering plants and contains approximately 650 genera and over 18,000 species (Polhill et af., 1981). The subfamily Papilionoideae is the largest of the three subfamilies of the Leguminosae, whose members include the most diverse and economically important legumes. Accurate phylogenetic relationships among these complex and diverse plants, for the most part, probably have not yet been fully resolved. The great diversity within the family and their potential for food and forage contribute to their increasing economic importance. Perhaps the most striking characteristic of the legumes is their ability to fix atmospheric nitrogen in a symbiotic relationship with Rhizobiurn. In a survey of approximately 3000 legume species, it was found that about 90% formed N-fixing nodules, thus making this family crucial in environmentally friendly, sustainable agriculture.

1. Genome Mapping among the Legumes The vast amount of mapping progress has been accomplished by using molecular genetic markers (Table 11). Not surprisingly, most studies have involved crops of major importance as sources of seed protein or oil or for high-quality forage. These mapping studies employed a wide range of intra- and interspecific crosses and F2, as well as recombinant inbred populations. In some instances maps were derived from the integration of two or more populations (Echt et af., 1994; Shoemaker and Specht, 1995; Ellis et al., 1992). Molecular probes detecting two or more loci are commonly reported in mapping studies involving legumes. It has been estimated that approximately 4752% of all flowering plants are polypoids. It is generally thought that plants with approximately n = 13 or greater should be considered polyploid. Consequently, the greatest number of duplicate loci within legume genomes have been reported within those genera possessing the greatest chromosome number, e.g., peanut and soybean.

2. Nontraditional Genome Map Applications A molecular genetic map alone is of limited value unless it can be applied to crop improvement breeding programs or to enhance our basic understanding of gene expression or our understanding and determination of genome organization and evolution. One of the first efforts to demonstrate the application of mapbased genome analysis of soybean cultivar pedigrees was reported by Shoemaker et al. (1992). This study demonstrated that segments of linkage groups could be followed through two generations, from parental cultivars to cultivars derived

Table Il

Genome Mapping Studies Reported among the Leguminmeae

Population information

Alfalfa (Meakugo) (2n = 16) M. sativa (CADL) X M. sativa (CADL) M. sativa ssp. quasi falcata X M. sativa ssp. coerula M. sativa ssp. sativa X M. sativa ssp. coerulea Soybean (Glycine) (2n = 40) G . mar x G. soja G. mar x G . max G. max x G. soja G. mar x G. soja G. max X G. mar G. mar X G. soja Mmgbean ( V W )(2n = 22) V. radiata CV. vc3890 x V. radiara ssp. subloha Cowpea (Vignu) (2n = 22) V. unguiculara cv. IT22464 x V. unguiculata spp. dekindtiana Peaont ( A m h k ) (2n = 40) A. stemspenna X A. cardenasii Lentil (Lms)(2n = 14) L. culiananis X L. orientalis L. ervoides X L. culinaris Common bean (Phaseolus)(2n = 22) P. vulgaris (Andean) X P. vulgaris (mesoamerican) P. vulgaris (Andean) X P. vulgaris (mesoamerican) Pea ( B u m ) (2n = 14) P. sarivum X P. sah'vwn

Linkage gmw

Loci

Map size (cM)

8 8 10

130 89 108

603 659 468

23 31 21 26 26 25

371 132 600 130

3371

Reference Echt et al., 1994 Kiss et al., 1993 Brummer er al., 1993

373

267 1200 1055 2461

Shoemaker and Olson, 1992 Lark et at., 1993 Fbfalski and Tmgey, 1992 Keim et al., 1990 Shoemaker and Specht, 1995 Shoemaker and Specht. 1995

14

171

1570

Menancic-Hautea et al., 1993

10

97

684

Menancio-Hautea et al., 1993

11

117

1063

9 11

34 64

333 560

Harvey and Muehlbauer, 1989 Weeden et al., 1992

11 15

224 143

960 827

Vallejos et al., 1992 Nodari et al., 1993a,b

7

I51

1700

11

66

350

100

1550

Halward et al., 1993

Ellis et al., 1992 Weeden and Wolko, 1990

F a b bean (Vuia)

v.foba

x

v.foba

Torres et al., 1993

USDA PLANT GENOME RESEARCH PROGRAM

127

from them, thus correlating changes in genome allelic structure with cultivar development. Lorenzen (1994) applied this concept to the pedigree analysis of 43 soybean cultivars spanning nearly 70 years and 6 generations of cultivar development. This approach facilitates the identification of those regions of a genome that have been selected for or against in breeding programs. This retrospective analysis of genome manipulation during cultivar development has the potential to allow breeders to identify genome regions important in their breeding programs and to make predictions of the success or failure of particular allele combinations. Extensions of this type of analysis have shown that, by combining pedigrees with graphical map-based analyses, it is possible to identify regions of the genome that are undergoing high or low rates of recombination during breeding programs. Alleles that are selected for or against within northern and southern germ plasms (different maturity groups) can be used to estimate losses of genetic diversity during consecutive generations of cultivar development and where losses in the genome have occurred (Lmenzen, 1994). In cultivated alfalfa, an out-crossing autotetraploid (2n = 4x = 32), most mapping studies involved diploids. In this crop, where maximum heterozygosity is desirable, a high degree of segregation distortion is evident (Brummer et al., 1993; Kiss et al., 1993; Echt et al., 1994). The amount of segregation distortion is higher than that reported for most other plants. In the majority of instances segregation was skewed in the direction of the heterozygote. The maximum heterozygosity theory maintains that multiple alleles at orthologous loci are a prerequisite for the successful expression of many of the traits associated with quality forage characteristics. Brummer et al. (1993) proposed that the genome regions associated with distorted segregation may be important targets for manipulation during forage breeding programs.

3. Integration of Genes into Legume Maps Many of the mapping population studies cited in Table I1 segregated for isoenzyme, morphological, or developmental qualitative traits. Consequently, many of these projects were able to incorporate agronomically important genes directly into the species map. Additionally, several groups attempted to integrate specific genes into molecular maps by using populations unique to that purpose. Meuhlbauer et al. (1991) used a combination of near-isogenic lines and segregating populations to map genes for morphological characters in soybean (PZ,r, and LfZ). Other groups mapped agronomically important traits such as Fap2, a gene for fatty acid content in soybean seed oil (Nickell et al., 1994), and nrs, a gene controlling nodulation in soybean (Landau-Ellis et al., 1991). Shoemaker and Specht (1995), using a specially constructed population segregating for

128

S . ALTENBACH ET AL.

nearly 20 genes, integrated nearly half of the classical soybean linkage groups into the molecular map in a single study. Several mapping studies focused on disease resistance genes in legumes. Diers et al. (1992a.b) mapped the location of five of seven known genes conferring resistance to Phytophthora root rot in soybean. Several groups have concentrated efforts on mapping the genes conferring oligogenic resistance to soybean cyst nematode (Webb et al., 1994; Weisemann et al., 1992; Concibido et al., 1994). A gene for the insect pests bruchids was mapped in mungbean (Young et af., 1992), as well as oligogenic resistance loci for powdery mildew (Young et af., 1993). Resistance to necrosis-inducing strains of bean common mosaic potyvirus (Nodari et af., 1993a) and common bacterial blight (Nodari et af., 1993b) were mapped in common bean. A resistance gene for soybean mosaic virus was mapped in soybean by Yu et af. (1994). The incorporation of many more disease resistance genes into molecular maps will likely occur as the integration of classical and molecular maps proceeds. Other projects incorporated agronomically important genomic regions into maps through QTL mapping. QTL mapping results to date have been carried out predominantly in soybean. QTL for seed protein and oil have been reported (Diers er af., 1992a; Mansur et af., 1993; Lark et al., 1993). Other studies have been reported on fatty acid composition (Diers and Shoemaker, 1992), seed coat hardness (Keim er af., 1990), seed weight (Mansur et af., 1993), nutrient efficiency (Diers et af., 1992b), and reproductive and morphological traits (Keim et af., 1990; Mansur et af., 1993).

4. Microsatellites or Simple Sequence Repeats Hypervariable “minisatellites,” or tandemly repeated short nucleotide sequences of variable length, were first reported in the human genome. “Microsatellites” have since proven to be invaluable, when coupled with polymerase chain reaction, in uncovering high levels of polymorphism in the human genome, as well as in plant genomes. One of the first, and most detailed, studies of microsatellites in higher plants was conducted by M a y a et al. (1992) within the soybean (Glycine m a (L.) Merr.). They identified a number of di- and trinucleotide minisatellites (simple sequence repeats or SSRs), demonstrated their Mendelian inheritance, and established their multiallelic properties. The same laboratory demonstrated in soybean, which has a paucity of polymorphism among elite breeding lines, that SSRs are extremely valuable in genotype identification. SSRs have since been used in a study to locate and map the location of Rsv, a gene conferring resistance to soybean mosaic virus (Yu et af., 1994). A concerted effort is in progress to fully integrate a high number of SSRs into a combined molecular and classical genetic map of soybean (Cregan, P. B., Specht, J., and Shoemaker, R. C., unpublished data), and the development of

USDA PLANT GENOME RESEARCH PROGRAM

129

microsatellites is progressing in peanuts (Aruchis sp.) (S. Kresovich, personal communication; G. Kochert, personal communication).

5. Comparative Mapping among Legumes A comparison of genetic maps established from an interspecific cross in lentil (Lens) and from garden pea (Pisum) suggested that extensive conservation of linkage relationships exists between these members of the legume tribe Viceae (Weeden er a f . , 1992). These authors showed that approximately 40% of the linkage map for Lens remained conserved in Pisum. They suggested that all members of this tribe may possess linkage groups similar in structure to those of Lens and Pisum. It was noted that linkages conserved between isoenzyme loci in lentil and pea could also be observed in faba bean (Viciu) (Torres et uf., 1993). However, due to the lack of high-resolution maps and the presence of many markers “bridging” the three genera, very few conserved linkages were observed. Comparative mapping between mungbean (V. rudiuru) and cowpea (V. unguiculuru), both members of the tribe Phaseoleae, also demonstrated a relatively high degree of linkage conservation between contiguous probes (MenancioHautea et a f . , 1993). The authors showed that 49 out of 53 loci retained linkage association. Although most regions of conservation were relatively small, a few large linkage blocks were retained between species. However, the linear order of loci within conserved linkage blocks occasionally was substantially rearranged. The soybean is considered to be a “diploidized” tetraploid, and many examples of duplicated genetic factors and duplicate loci are known. This conclusion is supported by the work of Funke er uf. (1993), who also noted a high degree of sequence conservation between duplicated regions in soybean. This type of information has the potential to enrich our understanding of gene expression of duplicated genes, as well as our understanding of the genetics of multigenic agronomic traits.

D. CRUCIFERAE The most economically important group of plants in the Cruciferae are in the genus Brussica (tribe Brassiceae) and include six cultivated species that are grown worldwide for several different uses. Three of the species are diploid (Brussica rupa, A genome, n = 10; Brussica nigru, B genome, n = 8; and Brassica oleraceu, C genome, n = 9) and three are amphidiploid (Brussica junceu, AB, n = 18; Brussica napus, AC, n = 19; and Brussica carinata, BC, n = 17), which are believed to have arisen by interspecific hybridization of the diploid species. One of the most distinctive features of this genus is its wide

130

S . ALTENBACH ET AL.

range of morphological variation and end uses, including vegetables, fodder, oil, and condiments. Even within species there is tremendous variation in form and use. For example, B . rupu includes turnip, oilseed rape, Chinese cabbage, and many other vegetable types. One might expect this wide range of morphological variations to reflect a high level of DNA polymorphism. Indeed, Brussicu species are among the most polymorphic of the crop species surveyed. Not only are the different morphological types within a species highly polymorphic for both RFLP and RAPD markers, but cultivars within morphotypes also show extensive polymorphism (Diers and Osborn,1994; Figdore et ul., 1988; Hu and Quiros, 1991; Thormann et al., 1994). DNA markers have proved to be very useful for phylogenetic studies of Brussicu and related genera. Results from RFLP analyses provide evidence for the evolutionary pathways of cultivated diploid forms from wild relatives (Song et ul., 1990) and for the hybridization of specific diploid accessions to create amphidiploids (Song and Osborn, 1992). 1. Genetic Maps

This high level of DNA polymorphism accelerated the development of genetic linkage maps, and several maps are available for four of the Brussicu species (Table 111). These maps consist mostly of RFLP markers, although some also include RAPD and isozyme markers. The number of marker loci for each map ranges from 49 to 360. Some of these maps have been developed independently by different researchers using different sets of marker loci. Therefore, the total number of marker loci mapped in Brussicu species probably exceeds 1OOO. An important future task is the integration of different maps so that researchers can have linkage information for a larger set of markers, and efforts toward this goal are currently underway (C. Quiros, personal communication). The use of alien chromosome addition lines, which have been developed for some of the Brussicu genomes [reviewed by Quiros et ul. (1994)], may help in this endeavor. Several important themes emerged from the inspection and comparison of Brussicu genetic maps. One is that even diploid genomes are highly duplicated. This was anticipated on the basis of previous cytogenetic studies, but RFLP maps have added a large degree of precision to the analysis of genome duplication. All three diploid species contain duplicated RFLP loci, and some loci are present in at least four copies. The most detailed analyses of the arrangementsof duplicated loci were reported for mapping populations of B . oleruceu (Slocum et ul., 1990) and B . rupu (Song et ul., 1991). For these populations, over one-third of the DNA clones used detected, replicated, segregating loci. Many of these loci were scattered throughout the genomes with no apparent conservation in linkage arrangement; however, a large portion were duplicated, or even triplicated, as conserved linkage blocks ranging from 2 to 10 loci. There was no evidence from

Table 3 Summary of Published Mdecular Marker M a p and Quantitative Trait h

Podation

Marker

No. of

types

loci

i Identified in Segregating Populations of Four Bmsska Species

Map lengtha

Trait loci

References

B. oleracea

WGA (cabbage) X Packman (broccoli)b EW (cauliflower) X CR7 (broccoli)b Four crosses (composite map)' 86-15-5 (cabbage) X CrGC 85 (rapid cycling) B116 (cabbage) X CY7 (broccoli)d

RFLP

258

RFLP

58

820 N

22 morph. traits

N

clubroot resistance, four morph. traits annual habit, glossy foliage clubroot resistance, fern leaf

Slocum er al., 1990 Kennard er a/., 19% Figdore et al., 1993 Kianian and Quiros, I992 land^^ et al., 1992

RFLP, isozyme RFLP

I08 198

747 cM 1112 cM

RFLP

114

980 cM

vernalization. blackmt resistance

camargo, 1994

RFLP

280

1850 cM

28 morph. traits

49

286 cM

N

Song er al., 1991 Song et al., 1994 McGrath and Quiros, 1991

RFLP

360

I876 N

N

Chi et al., 1992

RFLP

139

1785 cM

yellow seeds, erucic acid, pubesense

Teutonic0 and Osborn, 1994

RFLP, RAPD,

124

677 cM

N

TNCOand Quiros, 1994

120 130 138

1413 N 1350 N 1016 cM

N N

Landry et al., 1991 Honecke and Chyi, 1991 Ferreria et al., 1994a-d

B. r a y

Michihili (Chinese cabbage) X spring broccolib Yorii spring (turnip) x KwanHoo Choi (pak-choi)' Horizon (canola) x R500 (yellow sarson)= Per (rapeseed) X R500 (yellow sarsonld B. nigra B1164 x B1157

RFLP, isozyme

isozyme B . napus Westar (canola) x Topas (canola) BN0011 X BN0019= Major (rapeseed) x Stellar (cano1a)d

recombination units; cM, centimorgan (Kosambi map function); nr, not reported. Populations with the same letter have maps with marker loci in common.

a N,

b-e

RFLP RFLP RFLP

vernalization, white rust resist., blackleg resist.

132

S. ALTENBACH ET AL.

these, or any other, studies for the duplication of entire linkage groups. Thus, if the diploid Brussicu species have evolved as an ascending aneuploid series from a protype species of x = 6, as hypothesized from cytological evidence (Prakash and Hinata, 1980), there appears to have been extensive rearrangement of the replicated chromosomes. A second theme emerging from mapping studies is that the overall length of the genetic maps in different Brussicu species does not necessarily correspond to their nuclear genome size. This is most clearly illustrated for B. rupu and B. oleruceu, which have very similar DNA contents but very different map lengths. Several factors can influence the overall length of a genetic map; however, the reported maps, which were developed by different researchers using different probes and different populations, are consistently larger for B. rupu than for B. oleruceu. The three B. rupu maps with more than 100 marker loci range in length from 1785 recombination units to 1876 cM and are about twice as long as the four B. oleruceu maps, which range in length from 747 to 1127 cM (Table 111). These differences may be due to overall species differences in the frequency of recombination. On the basis of only one map, the diploid B. n i p appears to have a recombination frequency similar to that of B. oleruceu. The amphidiploid B. nupus, which consists of the A and C genomes, appears to have much less recombination than the sum of the diploid genomes; however, this comparison is based on fewer marker loci than are present in the combined diploid maps. A third theme that also has emerged from comparison of genetic maps is that Brussicu chromosomes appear to have been extensively repatterned during evolution of the species. Linkage arrangement of duplicated loci within species has provided some evidence for this, but comparison between species of maps containing RFLP loci detected by the same set of clones has provided additional evidence. B . oleruceu and B. rupu are closely related, produce viable but sterile hybrid progeny, and differ in chromosome number by only 1. Comparisons of linkage maps with common marker loci reveal extensive regions of conserved linkage arrangements (Osbornet ul., 1991; McGrath and Quiros, 1991; Slocum er ul., 1990; Song et u f . , 1991; Teutonico and Osborn,1994). However, these comparisons also provide evidence for chromosomal rearrangements, such as translocations or inversions, after divergence of the species. Conservation of the chromosomal integrity of these Brussicu species appears to be much less than that for other closely related species, such as tomato and potato (Bonierbale et ul., 1988). Maps with common marker loci also have allowed the comparison of the amphidiploid B. nupus with its hypothesized progenitor species B. rupu and B. oferuceu (Honecke and Chyi, 1991; Teutonico and Osborn, 1994; Camargo, 1994). As one might expect, there is evidence for extensive regions of conserved linkages; however, there is also evidence for rearrangements. These comparisons have not allowed the identification of the A and C genome chromosomes in B. nupus, and in fact, they suggest that these chromosomes, as they exist inB. rupu

USDA PLANT GENOME RESEARCH PROGRAM

133

and B. oleracea, may not be intact in B. nupus. By using a different strategy for genome comparison, Lydiate er al. (1993) developed an RFLP map for a population from a natural B. nupus crossed to a synthetic B. napus. This analysis allowed the assignment of A and C genome linkage groups in natural B. nupus, and the results suggest that these genomes have remained relatively intact in B. nupus.

2. Mapping Trait Loci There are many useful and interesting traits in Brassica that could be better understood genetically, and perhaps manipulated in breeding programs, by using molecular markers. Mapping information for trait loci is only just beginning to accumulate, but significant progress has been made in some areas (Table 111). An obvious target for these studies is morphological variation, and several genes controlling the qualitative variation for morphology have been mapped in B. oferacea (Kianian and Quiros, 1992; Landry et a f . , 1991, 1992) and B. rapa (Teutonic0 and Osborn, 1994). However, much of the variation for morphology in Brassica is under polygenic control. Some of these quantitative trait loci (QTL) were identified by analyzing populations from crosses of very different morphological forms in B. oferacea (Kennard et a f . , 1994) and B. rapa (Song ef af., 1994). Although analyzed as QTL, alleles at many of the loci identified in these studies had very large effects, suggesting that major genes have played an important role in the evolution of morphological variations in Brassica. However, loci with small effects also were identified, and it is alleles at these types of loci that breeders have probably manipulated to fine tune the current forms of our cultivars. The vernalization requirement is an important component of morphological variation, and loci controlling this have been mapped in several studies (Camargo, 1994; Ferreira et al., 1994b; Kennard et a f . , 1994; Kianian and Quiros, 1992). Another obvious target for mapping studies is disease resistance. In B. oleracea, QTL have been identified for clubroot resistance (Figdore et af., 1993; Landry et af., 1991, 1992) and blackrot resistance (Camargo, 1994). These diseases are sometimes difficult to screen in a breeding program, and linked markers could prove useful for manipulating alleles from the resistance sources used in these studies. In B. nupus, single major genes for cotyledon resistance to white rust and blackleg were mapped, along with field resistance to blackleg (Ferreira et al., 1994c,d).

3. Comparative Chromosome Organization of Brasska and Arabidopsis Arabidopsis thaliana, n = 5 (tribe Sisymbrieae), an extensively utilized model system in plant biochemistry, physiology, and classical and molecular genetics,

134

S . ALTENBACH ET AL.

is often referred to as a close relative of plants within the genus Brussicu. This relationship is further suggested by extensive conservation of coding sequences between Brassicu and Arubidopsis (Lydiate et ul., 1993). Comparative mapping of these genera is important because it will permit the cross-utilization of tools and resources that have been developed for each, and it will help us to understand the processes of evolution in greater detail. Moreover, use of the many cloned genes from Arubidopsis as RFLP probes in Brussicu may provide insight into the possible function of these genes in regulating traits of interest in Brussicu (Teutonico and Osbom, 1994). A comparative linkage map of the chromosomes of A . thaliana and B . oleruceu has been constructed by applying previously mapped Brussicu genomic DNA clones (Slocum er ul., 1990) to two segregating populations of A . thaliunu (Kowalski et ul., 1994). Although extensive chromosomal rearrangements have occurred since the divergence of B. oleruceu and A . rhuliunu, islands of conserved organization are discernible. At least one conserved region was detected on each of the five chromosomes. In total, 11 regions spanning 24.6% of the A. thaliunu genetic map were closely conserved with 29.9% of the B . oleruceu genetic map. Chromosomal segments with an average length of 21.3 cM in A . rhuliunu were estimated to be uninterrupted by rearrangements distinguishing them from their order in B. oleruceu. This calculation predicts that approximately 25 chromosomal rearrangements have occurred since the divergence of these two species, at a rate of 2.5 rearrangements per million years, since appearance of this plant family about 10 million years ago (paleopalynological evidence indicates that the plant order Capparales, including the families Capparaceae, Resedaceae, and Cruciferae, first appeared during the upper Miocene, approximately 10 million years ago; Muller, 1981, 1984). Relative to other plant species for which equivalent comparisons can be made, the chromosomes of B. oleruceu and A . rhuliunu appear to have diverged relatively rapidly [see Kowalski et ul. (1994)l. Chromosomal inversions appear to account for the synteny of unlinked markers. Several DNA markers that are closely linked in Brussicu were found to be syntenic in Arubidopsis, but separated by intervening markers from other B. oleruceu chromosomes or linkage groups. Inversion is the most likely means by which such markers have become separated (or joined). Syntenic markers at disparate sites on A. thuliunu chromosomes usually were closely linked in B. oleruceu (Kowalski et ul., 1994). This further supports the inferences that such markers reflect localized regions of conservation between A . rhuliunu and B . oleruceu and that these regions are distinguished by inversions. Two independent experiments suggest that ancient duplications have contributed to the present organization of the Arubidopsis genome. Kowalski et ul. (1994) identified one region of chromosome 1 that may be homeologous with a region of chromosome 5 . McGrath et ul. (1993) reported that three DNA probes

USDA PLANT GENOME RESEARCH PROGRAM

135

detect RFLP loci duplicated on Arubidopsis chromosomes 1 and 5 . Each of the duplicated regions is close to the respective homeologous regions reported by Kowalski et ul. (1994) on the basis of anchor loci common to chromosome 1 and tightly linked reference loci on chromosome 5 . In the McGrath et ul. (1993) map, putative homeologous regions of the genome may also be present between regions on chromosomes2 and 3 and between regions on chromosomes 3 and 4. It must be noted that duplicate loci that contradict evidence of Kowalski et ul. (1994) and McGrath et ul. (1993) have also been reported, with duplications between chromosomes 1 and 3 (Hauge et ul., 1993). On the basis of the analysis of organization of duplicated loci, it is clear that Arubidopsis and Brussicu diverged from a common ancestor with less chromosomal duplication than B. oleruceu. The relative orders of DNA markers along homeologous chromosomal regions permit inference of whether specific chromosomal rearrangements predate, or postdate, the duplication of Brussicu chromosomes. A segment of chromosome 3 of A . fhuliuna spanning 7 marker loci displays nearly complete linkage conservation with homeologous regions on C8 and C3 of B. oleruceu, except for 2 markers. Although these markers cosegregate on both C8 and C3 of B . oleruceu, they are separated by a distance of 5.7 cM, and two other markers, in A. thafiunu.The simplest explanation for this would be that the prototypical B. oleruceu and A . thuliunu chromosomes differ by a rearrangement in this region and that chromosomal duplications then propagate this region in B. oleruceu, i.e., the rearrangement predates duplication of the Brussicu chromosomes. Sadowski et ul. (1994) reported a complex of three tightly linked genes in A . thuliunu mapping to a single locus. Each of these probes maps to duplicated loci in B. oleruceu, cosegregating on one homolog, but with duplicated sites dispersed over three chromosomes. The simplest explanation for this would be that the prototypical B . oleruceu and A . thuliunu chromosomes show close linkage of these markers and that one B . oleruceu homolog is rearranged subsequent to duplication.

E. &VACEAE:

COTTON(GOSSYPIUM SPP.)

Cotton is cultivated for the production of spinnable fiber. The U.S.cotton crop of ca. 18 million bales (218 kg/bale) has a value of ca. $4-6 billiodyear. Cotton was among the first species to which the Mendelian principles were applied (Balls, 1906) and has a long history of improvement through breeding, with sustained long-term yield gains of 7- 10 kg of lint/ha/year (Meredith and Bridge, 1984). Cultivated cottons derived from four species, Gossypium hirsutum L., Gossypium burbudense L., Gossypium urboreum L., and Gossypium herbuceum L., provide the world’s leading natural fiber. Other wild relatives of cotton produce

136

S. ALTENBACH ET AL.

little or no fiber. G . hirsutum accounts for about 90% of world production. G . barbadense (Pima, Egyptian, and Sea Island cottons) fills a niche with highquality fibers that exceed those found in G . hirsutum germplasm. G . arboreum and G . herbaceum were the first cottons of world commerce, but commercial production of diploid cotton now is confined to India and Asia and is primarily for domestic use. Cotton is also the world’s second most important oilseed. The use of cottonseed, or limitations of its use, is determined by its composition (Kohel, 1989). 1. Germ Plasm

The genus Gossypium L. comprises about 50 diploid and tetraploid species indigenous to Africa, Central and South America, Asia, Australia, the Galapagos, and Hawaii (Fryxell, 1979, 1992). Diploid species of the genus Gossypium all have 13 gametic chromosomes (n = 13) and fall into 7 different genome types, designated A-G on the basis of chromosome pairing relationships (Beasley, 1942; Endrizzi et al., 1984). A total of five tetraploid (n = 2x = 26) species is recognized, all of which exhibit disomic chromosome pairing (Kimber, 1961). Tetraploid cottons contain two distinct genomes, which resemble the extant A genome of G . herbaceum (n = 13) and the D genome of G . raimondii Ulbrich (n = 13), respectively. The A and D genome species diverged from a common ancestor about 6-1 1 million years ago (Wendel, 1989). The A X D polyploidization occurred in the New World about 1.1-1.9 million years ago and required transoceanic migration of the maternal A genome ancestor (Wendel, 1989; Wendel and Albert, 1992). An extensive body of germ plasm is maintained in the USDA Cotton Germplasm Collection (Percival, 1987). The 5000 accessions of G . hirsurum include wild and/or feral types collected in their native habitat, obsolete cultivars, and improved cottons. G . burbadense includes accessions collected in their native habitat, and improved cottons, for a total of over 2000 accessions. The collection of cultivated diploids includes from obsolete landraces to contemporary improved types; recent germ plasm exchanges should increase this collection to about 2000 accessions. Many wild diploids are not productive enough for routine maintenance and have limited seed reserves or are maintained as live specimens (Percival, A. E., personal communication). Information regarding accessions and their availability is included in the Germplasm Resources Infon-nation Network (GRIN). Evaluation of the cotton germ plasm has included only a limited number of accessions or traits that were readily measured. The wild diploids are difficult to grow and do not have fiber, so that many desirable traits have no direct means of measurement. The tetraploids represent less of a problem, but perennial growth habits, late maturity, and photoperiodism make screening difficult. Conversion

USDA PLANT GENOME RESEARCH PROGRAM

137

programs for G. hirsutum and G . barbadense have begun (McCarthy et al., 1979; Percy, R. G., personal communication).

2. Molecular Map of the Cotton Chromosomes A detailed molecular map of cotton has been established (Reinisch et al., 1994), including the identity and ancestry of most cotton chromosomes, the relationships among the chromosomes, and most of the rearrangements that distinguish corresponding (homeologous) chromosomes. The cotton map spans 4675 (cM), representing 92-97% of the cotton genome. Among the more than 1200 DNA probes examined with 4-6 restriction enzymes, 563 DNA probes revealed RFLR at 705 loci, distributed at average intervals of 7.1 cM along the chromosomes. Among these, 683 (96.8%)have been assembled into 41 linkage groups of two or more loci, with the remaining 22 not yet linked to the map. Mapped DNA probes included cloned genes and genomic DNA fragments from several diploid and tetraploid cotton species. Low-copy DNA sequences from the A, D, and AD genomes have not diverged extensively, as genomic probes from each source readily detect genomic fragments across all genomes. Other work has added sequence-tagged microsatellites (Zhao et al., 1994) and tandemly repeated DNA elements to the map (Zhao, X., Dong, J., and Paterson, A. H., unpublished data). The map is based on 57 F2 progeny of a cross between single individuals of G. hirsutum race “palmeri” and G. barbadense acc. “K 101 .” These accessions were selected because they are relatively free from the interspecific introgression that characterizes cultivated types (Percy and Wendel, 1990; Wang, Dong, Paterson, submitted), as well as some wild populations of these two species from overlapping portions of their indigenous ranges (Brubaker eral., 1993; Percy and Wendel, 1990).

3. Cytogenetic Stocks and Assignment of Linkage Groups to Chromosomes Tetraploid cotton has the ability to tolerate haplodeficiencies, but no nullisomics have been identified. Therefore, cytogenetic aneuploids are represented by primary monosomes, which have one chromosome in the pair missing, or telosomes (monotelodisomes), which have one arm of a chromosome pair missing. Monosomes have been identified for 15 of the 26 chromosomes, and 29 telosomes have been identified that include one arm of 4 additional chromosomes (Endrizzi et al., 1984). Translocations are available that involve 25 of the 26 chromosomes of cotton. Through the use of cytogenetic markers, all 26 chromosomes are marked in at least one arm (Endrizzi et al., 1984).

138

S . ALTENBACH ET AL.

To integrate the molecular and cytogenetic maps, a subset of the mapped DNA probes was applied to monosomic and monotelodisomic substitution stocks, each with a single G.burbadense chromosome substituted for one pair of G.hirsutum chromosomes. On the basis of multiple genetically linked loci, the linkage groups that correspond to chromosomes (chrs.) 1 , 2 , 4 , 6 , 9 . 10, 17,22, and 25 have been determined (Reinisch et al., 1994).The identity of chrs. 5 , 14, 15, 18, and 20 is suggested by single loci, which are neither corroborated nor contradicted by any other locus on the linkage group. Three pairs of (tentatively) identified chromosomes showed homology: chrs. 1 and 15, chrs. 5 and 20, and chrs. 6 and 25. In each case, these homologies are corroborated by classical genetic analysis of mutant phenotypes (Endrizzi and Ramsay, 1979; Endrizzi et al., 1984). Additional aneuploid stocks are being characterized to verify results and to determine the chromosomal identity of additional linkage groups.

4. Deducing the Ancestry of Linkage Groups in Allotetraploid Cotton Many valuable traits might be transferred to cultivated cottons from wild relatives, especially wild diploids. Consequently, it was important to determine the diploid genomic origin of linkage groups (chromosomes) in the cotton map. Some DNA probes detected genomic fragments in tetraploid cottons that were shared with either A or D genome ancestors, but not both. On the basis of these “alloallelic” loci, the genomic origin of 33 of the 41 linkage groups, including all of the identified chromosomes, was determined (Reinisch et al., 1994). In 100% of the cases, the majority of the alloallelic information for a chromosome coincided with the prior classical assignment of chromosomes to genomes (based on pairing in diploid X tetraploid hybrids; A, chrs. 1-13; D, chrs. 14-26).

5. Unraveling the History of Cotton Evolution Based upon Duplicated Loci To better focus future efforts in cotton improvement, it was important to gain a better understanding of the history of cotton evolution. The distribution of linked, duplicated loci across the map reveals strong evidence of a recent (ca. 12 million years ago) chromosomal duplication event in cotton and tenuous evidence of a second, earlier event (Reinisch et al., 1994). Evidence for the duplication of at least 23 of our 41 linkage groups, covering 1668 cM (36% of the genome) and including 11 of the expected 13 pairs of homologs, has been described (Reinisch et af., 1994). Further mapping is likely to show that the entire genome was duplicated in this event. All except two homologous chromosome pairs in n = 26 cottons show one or more rearrangements of gene order as based on the present data.

USDA PLANT GENOME RESEARCH PROGRAM

139

While most duplicated loci mapped in “tetraploid (n = 26) cotton can be accounted for by this relatively recent event (1. I- 1.9 million years ago), new results support a classical theory that “diploid (n = 13) cottons may be “paleopolyploids” derived from an earlier event involving ancestors with fewer chromosomes (Reinisch et af., 1994). This putative n = 6-7 to n = 13 transition must have antedated the evolution of the genus Gossypiurn (estimated to be at least 25 million years old; Wendel and Albert, 1992) and, indeed, must have antedated the entire tribe Gossypieae and the closely related tribe Hibisceae, wherein all genera have high gametic chromosome numbers (Fryxell, 1979).

6. Prospects for Map-Based Cloning of Agriculturally Important Genes in Cotton and Other Polyploids Several complications associated with map-based cloning in disomic polyploids are partly compensated for by the unique advantages of polyploid genomes such as cotton, soybean, wheat, oat, canola, tobacco, peanut, and others. The physical amount of DNA in cotton is not prohibitive to map-based cloning; however, the lengthy genetic map will require a large number of markers in order to be sufficiently close to most genes for “chromosome walking.” The average physical size of a centimorgan in cotton is about 400 kb (Reinisch et al., 1994), which is only moderately larger than that of Arabidopsis (ca. 290 kb) and smaller than that of tomato (ca. 600 kb), both of which are species in which mapbased gene cloning has been accomplished. However, even with the advantage afforded by homologous information, the cotton map of 5000 cM (Reinisch ef al., 1994) will require ca. 3000 DNA markers to map at an average 1 cM density, and the physical genome of 2246 Mb will require ca. 75,000 YACs/BACs of average size 150 kb for 5x coverage. Map-based cloning in polyploids such as cotton introduces a new technical challenge not encountered in diploids, e.g., virtually all “single-copy” DNA probes occur at two or more unlinked loci. This makes it difficult to assign YACs (or other large DNA vectors) to their site of origin. However, interspersed repetitive DNA elements, which differ between the two genomes of tetraploid cotton, may provide a means of determining the genomic identity of individual YACs from tetraploid cottons (Zhao, X.,Wing, R., and Paterson, A. H., submitted). Such an approach may prove generally applicable to map-based cloning in other major crops, many of which are disomic polyploids (e.g., soybean, wheat, oat, canola, tobacco, peanut, and many others).

7. Application of DNA Markers to Cotton Improvement The cotton RFLP map is a starting point for the use of DNA markers to identify and manipulate determinants of agricultural productivity and quality.

140

S. ALTENBACH ET AL.

Historically, introgression has been practiced to transfer desirable traits into cultivated cottons. Such introgression events may be detectable using existing DNA markers, including the transfer of numerous genomic regions from G. hirsutum into cultivated G . barbdense (Wang, G . , Dong, J., and Paterson, A. H., submitted), as well as introgression of specific traits such as Verticifliurn wilt resistance (Staten, 1971), bacterial blight resistance (Staten, 1971), nectarless leaves (Meyer and Meyer, 1961; Vler, 1908), restoration of cytoplasmic male sterity (Meyer, 1975; Weaver and Weaver, 1977), and improved fiber quality (Culp and Harrell, 1974; Culp et a f . , 1979). Rust resistance was transferred (Blank and Leathers, 1963), but has limited importance and is not widely used. Other traits such as nematode resistance (Yik and Birchfield, 1984) and unique natural products have been identified (Balls, 1906) in wild diploids. Efforts to introgress quantitative traits, such as fiber properties, have had limited success (Meredith, 1984). The availability of a detailed molecular map of cotton affords opportunities to pursue comprehensive mapping of both simple and complex traits by welldeveloped strategies (cf. Paterson et al., 1988). As in many predominantly self-pollinated crops, the gene pools of each of the cultivated cotton species show only modest levels of DNA polymorphism. While the large number of DNA markers now mapped in cotton partly compensate for this limitation, routine application of DNA markers to cotton breeding may benefit from new technologies such as microsatellite-based DNA markers, which are being superimposed on the existing RFLP map (Zhao et a f . , 1994) to create a unified body of information on cotton genetics and evolution. A detailed map of DNA markers offers the opportunities to utilize the polymorphic genus Gossypium and its diverse relatives such as Hibiscus (kenaf, roselle) and Abefmoschus (okra), to investigate plant chromosome evolution in exquisite detail, and to have a major impact on the improvement of one of the world’s oldest and most important crops.

F. SOLANACEAE 1. Comparative Mapping in the Family Solanaceae The family Solanaceae (nightshade family) contains a number of economically important plant species, including tomato, pepper, potato, eggplant, petunia, and tobacco. The majority of nightshade species have a basic chromosome number of x = 12. Comparative linkage maps have been constructed for three of these species, tomato, potato, and pepper (Bonierbale et al., 1988; Tanksley, 1992; Prince et al., 1993). The results from these comparative maps reveal several interesting aspects of plant chromosome evolution.

USDA PLANT GENOME RESEARCH PROGRAM

141

2. Speciation and Morphological Differentiation Are Not Always Accompanied by Major Genome Rearrangements Tomato and potato are well-differentiated species and are sexually incompatible. Nonetheless, the genetic content and chromosomal organization of tomato and potato are nearly identical. Comparative maps reveal no interchromosomal rearrangements (e.g., reciprocal translocations). The only apparent gross differences between the genomes are five paracentric inversions (chromosomes 5 , 9, 10, I I , and 12). Four of these involve short arm inversions (chromosomes 5 , 9, 11, 12), while one (chromosome 10) involves a long arm inversion. The high level of conservation in the linkage of tomato and potato permits easy cross-use of probes for genome mapping between these two species.

3. Break Points for Chromosomal Rearrangements Often Occur at or near Centromeres All of the inversions that differentiate the tomato and potato genomes appear to involve a break point at that chromosomal location of the affected chromosome, resulting in an inversion of the entire chromosome arm. In no instances could a second break point for an inversion be seen in the distal part of the arm,indicating that the entire arm had been inverted (Fig. 2). Genome studies in tomato show that telomeres and centromeres share some repetitive DNA sequences, which may involve occasional recombination events that result in the inversions of entire chromosome arms as found in tomato and potato (Tanksley, 1992). Pepper is more distantly related to tomato and potato and has a higher nuclear DNA content. While there are many more chromosomal rearrangements differentiating the pepper genome from the tomato and potato genomes, conserved linkage blocks can still be observed, and these often correspond to segments of chromosomes with putative break points at or near centromeres (Prince e? al., 1993; Fig. 2).

4. Gene Repertoire is More Conserved than Gene Order The most conserved feature of the tomato, potato, and pepper genomes is the gene content. Cloned genes from any one of these species usually crosshybridize with orthologous gene copies in each of the other species. The level of gene conservation (based on cross-hybridization with cDNA clones) is greater than 99% for tomato and potato and greater than 90% for most solanaceous species (e.g., tomato, pepper, potato, petunia, tobacco) (Zamir and Tanksley, 1988). The high degree of gene conservation makes cDNA clones ideal for comparative genome mapping in these species.

142

S . ALTENBACH ET AL. 0

0 c Q

c

0 P

c

4-

(A

s

Hgure 2. Synteny among pepper, potato, and tomato.

5. Application of Comparative Maps in the Family Solanaceae Molecular mapping of nightshade species has advanced dramatically. These maps and their associated technology have been used successfully for a number of applications in plant breeding and genetics: (1) characterization of genetic variation in germ plasm collections; (2) gene tagging (i.e., identification of markers tightly linked to major genes); (3) map-based gene cloning; and (4) analysis of quantitative traits. The ability to cross-use probes has accelerated genome applications in nightshade species-especially in potato and pepper, for which the tomato genetic map has offered a source of probes whose homologous chromosome positions can often be identified through the use of comparative

USDA PLANT GENOME RESEARCH PROGRAM

143

genetic maps. It is likely that a more synergistic relationship will develop between breeders and geneticists of these crop species since they are tied together through common mapping information. We can also expect that comparative genetic maps will also be available for other nightshade species, including petunia, tobacco, and eggplant.

1. Genetic Mapping in Forest Trees

a. Introduction There are many aspects of genome mapping research in forest trees that parallel genome mapping in agronomic crops, for example, types of markers used, methods of linkage analysis, and some of the applications of genetic marker technology. There are, however, many unique aspects to genome mapping in forest trees that will be the focus of this section. Several more comprehensive reviews of genome research in forest trees have been written (Neale and Williams, 1991; Grattapaglia et al., 1993; Neale and Harry, 1994). Trees, unlike most crops, are long-lived perennial plants. Generation times are long, and multigeneration pedigrees are rarely available. Most species are difficult to self-pollinate and inbreeding depression is common; thus, typical mapping pedigrees such as F2’s or back-crosses generally are not used. Trees are also highly genetically variable, so that mapping populations are constructed from matings among highly heterozygous parents. Forest tree species are found in both the angiosperms and gymnosperms. Taxa of great interest in the angiosperms include Populus and Eucalyptus. Within the gymnosperms the pines (Pinus) predominate, but spruces (Picea), firs (Abies), larches (Larix), and Douglas firs (Pseudotsuga) are also important. An interesting but challenging aspect of conifer genome research is the size of the genomes. C-value estimates range from 20 to 30 pg among pines (Wakamiya et al. , 1993). One theory for the large genomes of pines is that they appear to have exceptionally large gene family sizes on the basis of Southern blot analysis (Kinlaw and Gerttula, 1993; Devey et al., 1994a,b). Reasons similar to those for crop species exist for constructing genetic maps in forest trees, e.g., understanding genome organization and evolution, gene and quantitative trait mapping, and marker-aided breeding. In contrast to most crops, however, trees are relatively undomesticated and are grown in wildland environments. There is an urgent need to monitor changes in genetic diversity in these populations as they are effected by human intervention and global climate change. Tree genome research will play an important role in solving these practical forest management problems.

144

S. ALTENBACH ET AL.

b. Genetic Mapping Genetic maps have been constructed for a number of forest tree species, and a variety of approaches have been employed. The first mapping projects in forest trees were based on RFLP markers and mutigeneration pedigrees. Devey er al. (1991, 1994a,b) mapped 73 RFLP loci to 20 linkage groups (N = 12) using 65 cDNA and 3 genomic DNA probes in a three-generation outbred pedigree of loblolly pine (Pinus raedu L.). Groover er al. (1994) subsequently constructed a second RFLP map for loblolly pine by using many of the same DNA probes used in the earlier study. Genetic maps were constructed for both the female and male parents of the mapping population and were used to demonstrate that the rate of meiotic recombination is slightly greater in the male than in the female in loblolly pine (Groover et al., 1995). The two loblolly pine RFLP maps are being integrated to form a consensus map. Bradshaw er al. (1994) constructed a detailed map for Populus by using RFLP, STS, and RAPD markers in an F2 population derived from a hybrid cross between Populus rrichocarpa and Populus delroides. Populus is dioecious; thus, the F2 resulted from a full-sib mating of two F1 plants. A total of 343 markers was mapped to 25 linkage groups (N = 19). The Populus mapping project has led to some interesting basic biology. Bradshaw and Stettler (1994) demonstrated that segregation distortion at one linkage group was due to the presence of a linked recessive allele in one of the progeny homozygous classes. An RFLP map has also been constructed for a second species of Populus, trembling aspen ( P . rremuloides Michx.). Liu and Fumier (1993) mapped 54 RFLP and 3 arozyme loci to 14 linkage groups (N = 19) by using genomic DNA probes from trembling aspen. Although RFLP markers are highly informative for most applications in forest trees, very few forest genetics labs have used this technology because of the technical difficulty of RFLPs with large genomes and the slow rate of data acquisition. The development of the RAPD marker system (Williams er al., 1990) profoundly changed the view of many researchers toward the feasibility of genetic mapping in trees. The speed and ease of RAPDs were obvious; but more important for tree geneticists was the ability to perform RAPD assays and construct genetic maps using the haploid megagametophyte tissue of conifer seeds, Tree geneticists had long taken advantage of this unique genetic system in conifers to establish inheritance and linkage relationships among allozyme loci and to study the population genetics of conifers. Carlson er al. (1991) first reported on the inheritance of RAPD markers in a conifer; however, this study was based on segregations in diploid Fl’s of a controlled mating of Douglas fir (Pseudorsuga menziessii). In a subsequent paper, this group reported on the construction of a RAPD map in white spruce (Picea glauca) based on megagametophyte segregations from a single mother tree (lhlsieram ef al., 1992). Since then single tree RAPD maps have been reported for a number of conifers (Nelson ef al.,

USDA PLANT GENOME RESEARCH PROGRAM

145

1993a,b; Binelli and Bucci, 1994). These RAPD maps will generally be used for QTL mapping and marker-aided breeding. RAPD maps have also been developed in angiosperm forest tree species. Grattapaglia and Sederoff ( 1994) used a “two-way pseudotestcross” mapping strategy to construct maps for both the female and male parents of a hybrid cross of E. grandis X E. urophyffa.These maps had 240 markers in 14 linkage groups and 251 markers in 11 linkage groups, respectively.

2. Quantitative Trait Mapping in Forest Trees Most traits of economic interest in forest trees are quantitatively inherited, for example, height growth, volume, wood density, and bud phenology. Unlike many crop species, there are almost no traits of interest that are known to be caused by single genes (see the discussion of disease resistance mapping to follow for an exception), and mutant stocks have never been developed for forest trees. Thus, the demonstration that QTLs could be mapped in plants (Paterson et a f . , 1988) was met with great enthusiasm by forest geneticists. Groover et a f . (1994) mapped five major QTLs for wood specific gravity in loblolly pine by using a large F2 family from a three-generation outbred pedigree. They used RFLP markers so that in cases where the linked RFLP marker segregated for more than three alleles (fully informative marker) it was possible to estimate the number and relative effects of the QTL alleles segregating in each of the two parents. Information of this type will be valuable if the markers are to be used in a marker breeding application in a full-sib but outbred mating design. QTLs have also been mapped in interspecific crosses of two angiosperm species. Bradshaw and Stettler (1995) mapped QTLs for a number of growth and adaptative traits in an F2 population of a P. trichocarpa X P . deftoides hybrid. In several cases, a small number (1-5) of QTLs explained a very large percentage of the total genetic variance for a quantitative trait. These results demonstrate that the purely polygenic and additive model for inheritance of growth and adaptative traits in trees may not be appropriate. The high proportion of genotypic variance attributed to a few QTLs should not be generalized to intraspecific crosses because these data are from interspecific crosses where linkage disequilibrium is expected to be high. In Eucalyptus, Grattapaglia (1994) mapped QTLs for traits related to vegetative propagation, growth, and wood properties. QTLs for vegetative propagation traits were mapped by using RAPD markers and the “pseudotestcross”strategy in a E . grandis X E. urophyffahybrid cross. It was shown that shooting responses are inherited from E. grandis, whereas rooting responses are inherited from E. urophyfla. The growth and wood property QTLs were mapped in half-sib families from E. grandis.

146

S . ALTENBACH ET AL.

3. Mapping of Disease Resistance Genes Protection from disease and insects is an objective of many forest tree breeding programs. A quantitative genetic approach to breeding for resistance has almost always been applied and progress has been made (Carson and Carson, 1989). However, gene-for-gene relationships have also been identified and this theory has been applied to breeding for resistance. One of the most destructive pathogens of forest trees are the pine stem rusts (Cromrtium). Fusiform rust(C. quercuum)attacks pines of the subsectionAustrales in the southeasternUnited States. Research has been initiated to identify and map fusiform rust resistance genes in lobolly pine (Wilcox er al., 1993) and slash pine ( P . ellioffiiEngelm.) (Nance ef al., 1992; Nelson et al., 1993a,b). Gene-for-generelationships have been established in both species, and significant progress toward mapping resistance genes has been made. A second important pine rust is white pine blister rust (C. ribicola Fisch.), which attacks members of the section Strobus (white pines). In a pioneering study, Kinloch er al. (1970) discovered a single dominant gene for resistance in sugar pine ( P . lamberriana Dougl.). Later, a gene-for-gene relationship with the rust was demonstrated (Kinloch and Comstock, 1981). On the basis of the wellcharacterized genetics of resistance in sugar pine, Devey et al. (1994a) mapped the dominant resistance gene, R. Their strategy employed haploid genetics, RAPD markers, and bulked segregant analysis (Michelmore et al., 1991). Ten linked RAPD markers were identified, the closest being 0.9 cM from R. These markers ultimately may assist in the cloning of this gene.

4. Gene Sequencing Tree molecular geneticists have cloned, sequenced, and studied the expression of a wide array of genes [see the review by Davis (1995)l. These efforts have largely been one gene at a time using standard “cloning by homology” or “reverse genetic” approaches. The Human Genome Project has led to an alternative approach to identifying genes on the basis of automated sequencing of anonymous cDNAs (Adams et al., 1991, 1992). In plants, this approach has been applied to rice (Uchimiya et al., 1992), Arabidopsis (Hofte et al., 1993); corn (Keith er al., 1993), and Brassica (Park et al., 1993). Tens of thousands of cDNAs have already been sequenced and several thousand have been identified on the basis of homology searches of databases. A gene sequencing project has been initiated at the Institute of Forest Genetics. Approximately 200 cDNAs, which had previously been used as RFLP mapping probes (Devey et al., 1994b; Groover et al., 1994), have been sequenced. Identities were determined for approximately 30% of the cDNAs (Kinlaw and Gerttula, 1993). In the future, random cDNAs will be sequenced from tissue and developmentally specific libraries.

USDA PLANT GENOME RESEARCH PROGRAM

147

5. Marker-Aided Breeding and Selection The potential of marker-aided breeding (MAB) and/or marker-aided selection (MAS) in forestry has been hotly debated. The potential for MAB/MAS with inbred agricultural crops and back-cross-breeding has been firmly established; however, it is not certain that MAS/MAB could be applied to highly heterozygous and outbred plants such as forest trees. In their review, Neale and Williams (1991) identified three potential limitations: (1) polygenic inheritance of economically important traits, (2) genotypic x environment interactions, and (3) linkage equilibrium conditions between marker alleles and QTL alleles. In June, 1991, a symposium was organized to debate the pros and cons of MAB/MAS ['hskan (1992) and papers cited therein]. In a thorough review on this topic, Strauss et al. (1992) identified additional limitations to MAB/MAS, such as (1) high cost, (2) QTL x genetic background interactions, and (3) changes in QTL gene frequencies over generations. Opinions among the symposium participants ranged from highly optimistic to completely pessimistic; the only consensus reached was that empirical data was needed before MAB/MAS could be fully evaluated for application in forestry. In the past 2 years, several developments have occurred that provide optimism for MAB/MAS in forestry. First, it has been demonstrated that QTLs of major effect can be identified for economic traits (Groover et al., 1994; Grattapaglia, 1994; Bradshaw and Stettler, 1995) and that inheritance of such traits is not strictly polygenic and/or additive. Second, the problem of linkage equilibrium can be overcome if marker and QTL allele phase relationships can be determined for all members of a breeding population. These relationships, however, are best determined in experimental populations where genetic segregation of the quantitative trait has been maximized. Once marker-QTL phase relationships are determined for members of a breeding program, this information can be used to guide breedings in an applied breeding program. This task is possible with highly informative RFLP markers (Groover et al., 1994), but is not likely to be cost effective. RAPD markers are less informative but much more cost effective for genotyping large numbers of individuals (Grattapaglia et al., 1993). Clearly, more empirical testing is needed to identify the best strategies and opportunities for MAB/MAS in forestry.

m. PLANT GENOME DATABASE A. INTRODUCTION The Plant Genome Database (PGD) is actually a suite of several information products produced at the National Agricultural Library (NAL) in collaboration

148

S. ALTENBACH ET AL.

with the Agricultural Research Service and Forest Service species coordinators. The species groups initially consisted of wheat (which soon expanded to the Triticeae), soybean, maize, and lobolly pine (which has expanded to include other conifers). Arabihpsis was added soon after. At present, the Solanaceae, rice, and Chlamydomonusare included, and cotton and sorghum will be added in the near future. Expansion to include other species will continue. The products include a Gopher server (PGD/Gopher), a World Wide Web server (PGDI WWW), a CD-ROM (FGDICD), an electronic mail query server (PGD/Email), and an anonymous ftp server (PGDIFtp). These information products, along with their evolution, will be discussed. The species coordinators are responsible for collecting, organizing, and evaluating data for their respective species into separate databases. These databases are discussed elsewhere in this review; however, a description of the software used by the majority of the coordinators, ACEDB, is presented in Section 1II.E.

B. HISTORY During the initial phases of the database project, a number of site visits were made to locations that were already producing similar databases. Several were involved in the Human Genome program, including the GDB at Johns Hopkins, Lawrence Berkeley Labs, Lawrence Livermore Labs, and the Los Alamos National Lab. Site visits were also made to the National Center for Biotechnology Information at the National Library of Medicine and to two private companies, Agrigenetics and DuPont. These visits were useful with respect to database design and hardware and software selection. The primary goal of the first 2.5 years of the project was to develop a working prototype of an integrated genome database for plants. The focus would then change to releasing the database to the users, reevaluating the initial design, redesigning portions of the database where necessary, and starting to forge tighter links to external data sources. The goals and timetable have been followed closely, although the number of information products has increased significantly over what was anticipated.

C. DISCUSSION OF THE PGD INFORMATION RESOURCES The initial phase of design development primarily consisted of determining the scope of the database. It was decided that the core of the database would consist of maps and loci, along with alleles, probes, phenotypic traits, gene products, and metabolic pathways. It was also deemed essential that this information be linked with existing information, such as germ plasm (via the USDA Germplasm

USDA PLANT GENOME RESEARCH PROGRAM

149

Resources Information Network) and sequence data (GenBank/EMBL/DDBJ and SwissProt/PIR), as well as NAL‘s AGRICOLA bibliographic database. As database design was taking place not only at NAL but also at the species group level, it was decided to hold review meetings approximately every 3 months in order to exchange ideas about how to achieve an optimal design and to ensure that all collaborators were maintaining design compatibility. A decision was made early on not to enforce a single design for two reasons: (1) one design was not necessarily optimal for all groups; and (2) as the goal was to eventually include data from many additional sources, it was necessary to develop a paradigm for integrating data that would likely be formatted in many different ways. Therefore, a major design goal was to develop a generic mechanism that would allow data formatted in multiple ways to be accessed in a consistent manner. Initially, it was assumed that this would require a centralized relational database, but with the advent of new data delivery mechanisms on the Internet, particularly the World Wide Web (WWW), Gopher, and WAIS, this approach was replaced by one in which databases are “federated,” that is, they remain separate entities, but are accessed as if they are a single entity. This paradigm allows such extreme flexibility that a new database with a unique data model can be integrated into this system in less than an hour. Also, species database curators can alter their data models at will without worrying about the impact to PGD. Gopher and WAIS caused a virtual information explosion in 1992. A complete description of Gopher, WAIS, Mosaic, and WWW is beyond the scope of this document. Further information, including how to obtain and install Gopher and WWW clients, should be obtained from your local computer center or network “guru.” Gopher uses a menu-oriented paradigm and a simple protocol to allow the user to access text documents, files, and images. When coupled with WAIS, it provides full-text searching of data files. PGDIGopher provides access to all of the plant genome data, as well as the retrieval of AGRICOLA records. Although Gopher is simple to use, it is limited in the ways that data can be retrieved and presented. Figure 3 shows a menu available through PGD/Gopher. A more important addition to data retrieval via the Internet has been WWW. It is similar to Gopher in that it uses a simple protocol to allow users to navigate through the “web” of information available on the Internet. It is, however, more sophisticated and flexible than Gopher, primarily due to its use of hypertext links and, with the appropriate WWW viewer, in-line graphics (e.g., genetic maps) and a point-and-click method of navigation. Most viewers, including the most popular one, Mosaic, also “understand” the Gopher, WAIS, and ftp protocols. Versions of Mosaic exist for Unix (several flavors), MS-Windows, and Macintosh. See Fig. 4 for an example of a screen from PGDIWWW. PGD/WWW, in addition to providing for simple database navigation, also gives the user several options for data retrieval, including fuzzy and WAIS searching. When one of these options is selected, the user is presented with a

S . ALTENBACH ET AL.

150

a

About, Genone Oatabases at Help with Searching About, I n d i v i d u a l Databases

0 Scanned Inanes -I

- * * *.

LU.

* * * - *.

--

fibout the Plant Genone UUU Inherface

-m Plant

Genone UUU I n t e r f a c e

-

no Passuord

W.'.**UL

H Search A l l Plant Genone Databases p J

LLYUYLUL

Search AAtDB (Arabidopsis thaliana) Search SoltBase (Sorlbeans)

w

Search RiceGenes (Rice) Search SolGenes (Solanaceae) Search HaizeOB (naize) Search GrainGenes (uheat and r e l a t i v e s ) Search TreeGenes ( f o r e s t trees) Search ChlandlB (Chlanwhwmas)

Flgure 3. Example of a PGD/Gopher menu.

form into which the search criteria are entered. Simple searches, as well as boolean and wildcard searches, can be carried out. Fuzzy searching, in which mismatches (insertions, deletions, or substitutions) are allowed, is particularly useful in cases where the exact spelling of a search term is not known (e.g., a person's name) or to accommodate slight differences in nomenclature among species databases. An example of the latter is the symbol for alcohol dehydrogenase 1, which might be either adhl or adh- 1 depending on the database of origin. Either symbol could be used as a search term for fuzzy searching and all records containing either of these symbols would be retrieved. WAIS searching is provided as an alternative for those who are more familiar with its capabilities or for those who do not require fuzzy searching. In either case, the user can select which database, or combination of databases, to search. Also, the user can choose to search all plants or all grasses. One major advantage of the WWW that is used extensively in PGD is the ability to link to external data sources on the WWW. A simple example: If a

USDA PLANT GENOME RESEARCH PROGRAM

151

Database: RiceGenes (rice) uddhwX-----

(hShCUSAdh-1 Type Isozyme Name Pull-name Alcohol dehydrogenase-1 Reference Linkaae analvsis for four isozvme loci, Mh-1, ACD-1, Pox-: pad-l Chromosomal analvsis of isozvme loci and the allelic emre.: at cellular level in rice Chromosomal location of four isozvme loci by trisomic anal\ w z a sativa L.). Location Hap Rice-Momh-11 Position 95 pice-Cu-11 Position 65.4 Error 2.25 Map-data Rice-lk-mholoaicak Remarks On Rice-B8125/2/BB125/WL02 map: Knovn gene, position approximatad based on mapping in a different population.

Figure 4. ing.

Sample screen from PGDIWWW. Note that hypertext links are indicated by underlin-

database curator provides an accession number for a SwissProt protein sequence anywhere in a database object, a link is made directly from that object to the ExPASy WWW server in Geneva where SwissProt is maintained. By clicking on that link, the user retrieves the full SwissProt sequence record. In addition to SwissProt, links are provided in PGD/WWW to Genbank/EMBL/DDBJ, GRIN, AGRICOLA, dbEST, Enzyme, Plant Variety Protection data, and Mendel (Commission on Plant Gene Nomenclature database). It is anticipated that many additional links, such as to metabolic pathways data, will be made in the future. The PGD/CD (the initial prototype was pressed in April 1994 and the first release was in January 1995) contains a version of the database that is similar to PGD/WWW, that is, it is browsable with a WWW client. Navigation through the data is also accomplished via hypertext links. Three versions of the Mosaic viewer are included on the CD. Although lacking in the prototype, subsequent CDs will contain full-text searching. It should be noted that PGD/CD is, of necessity, not updated nearly as frequently as other versions of PGD and should be used only if Internet connectivity is not available to the user. PGD/Email is an electronic mail server that allows users to query the plant genome database by sending electronic mail to the server. This allows searching for people who may have an electronic mail gateway to the Internet, but who are not directly connected to it. Also, those who are directly connected but may have slow or otherwise poor Internet connections may prefer to use this method. The results are returned to the user by electronic mail. PGD/Ftp allows users to use anonymous ftp to retrieve certain data. All of the species databases are available in native ACEDB format. Many graphical images, mainly in gif and jpeg formats, are also available. Additional data and tools will be deposited in the anonymous ftp directories in the future.

152

S. ALTENBACH ET AL.

All of these resources are changing rapidly, and more up to date information can be found by accessing the particular resource. This information is given in the next section.

D. ACCESSINGTHE PGD INFORMATION RESOURCES Regardless of which of the various PGD resources is being accessed, the user should always consult the README, ABOUT, or HELP files to learn about that particular resource. Also, WHAT’S NEW files are included when applicable. Technical help for any of the resources can be obtained by sending an email message to [email protected] or by calling 301-504-6813. Database content questions or comments should be addressed to the curators for a particular species database. PGD/Gopher: Host, probe.nalusda.gov; port, 70 E D / W W W URL: http:/ /probe. nalusda.gov: 8OOO/ PGD/Email: Send an electronic mail message with only the word help in the body to [email protected] PGD/Ftp: Ftp to probe.nalusda.gov [login: anonymous; password: (your electronic mail address)] PGD/CD: Send electronic mail to [email protected] or call 30 1-5046613 for availability

E. ACEDB A major challenge facing the research community is how to present and integrate complex and rapidly accumulating data. Unfortunately, the solutionthe development of specialized “genome” databases-has come about more slowly than anticipated. One problem is that resources are limited; a full-fledged database design and implementation team using a commercial database management system can easily consume over $1 million per year for software, personnel, and hardware. Second, and more critical, a new group desiring to “clone” an existing database is faced with software licensing fees and the need to hire expensive experts for maintenance and to make modifications. These factors discourage the reuse and proliferation of database technology, even though user demand for it is greater than ever. A product that largely overcomes these problems has become available to biologists. ACEDB, “A C. elegans Data Base”, was created in 1991 by Jean Thierry-Mieg and Richard Durbin (1991) to represent information for the C. elegans genome project. In the short time since then, it has been adopted by a

USDA PLANT GENOME RESEARCH PROGRAM

153

dozen different genome projects and is in use at over 100 research sites worldwide. The rapid acceptance of ACEDB-based genomic databases is due to several attractive features: the software is free and includes a sophisticated graphical interface immediately capable of displaying genetic maps (see Fig. 3,physical maps, sequences, and any text-based biological information. The software can also present scanned images such as autoradiograms and photographs. Finally, reconfiguration for a new species is straightforward and requires no computer expertise. Thus, biologists can initiate and control the entire database development effort from prototyping to public release. This has vastly accelerated the rate at which new databases appear. Users favor the software because data is accessible for casual browsing, as well as by formal query. Browsing is supported by a hypertext interface, similar in spirit to Hypercard applications on the Macintosh. Most objects that appear on

H p r e 5. Genetic map display from ACEDB.

I54

S. ALTENBACH ET AL.

the screen-locus symbols, lines representing cloned DNAs, names of stocks, people, sequences, publications, and so forth-are sensitive to the mouse. Users traverse links between information by clicking with the mouse. The related information appears in a new window that itself will have hypertext links. ACEDB is available via anonymous ftp from several sites. The major site for the software in the United States is ncbi.nlm.nih.gov (130.14.25.1) in /repository/acedb. Fully loaded databases are available from probe.nalusda.gov (192.54.138.44) in /pub/ACEDB.databases. ACEDB runs on the Macintosh and many Unix systems: under X l l , any machine running SunOS 4.x, e.g., Sun SPARCstation 1, 1+, 2, IPC, IPX; any machine running Solaris; DEC station 3100, 5100, etc.; DEC AlphalOSF-1; Silicon Graphics Iris series; PC 386/486 with Linux (public domain Unix). Information about ACEDB can be obtained via anonymous ftp from several on-line sources (Cherry and Cartinhour, 1993; Dunham et al., 1993). An introductory manual (Cartinhour el a!., 1992) is available for free. In addition, a Usenet/Biosci conference (bionetsoftware. acedb) has been established. Extensive documentation on ACEDB can be found on PGD/WWW.

IV. FUTURE PROJECTIONS Plant genome research is here to stay. The future looks excellent. The advancements made during the past decade in gene mapping and associated research activities permits optimism. This tendency toward a hopeful outlook is justifiable on the basis of results to date. Finally, it is becoming evident from genome research that geneticists can increase their understanding of genotypes and related gene expression processes. They have the tools to learn how genotype x environment interactions work and how to use that information to productively manipulate quantitative characters in breeding schemes. The capabilities of handling single gene traits by plant breeders are more effective than ever before with the new molecular techniques. Molecular map markers or marker-assisted selection gives the breeder increased research power and precision. This new strength hastens finding and manipulating genes that code for desirable traits over the use of traditional phenotypic selection. In addition to precision, the new methods speed up the breeding process by reducing the time it takes to develop a new cultivar with desired traits. Concurrent with time, reductions in expenditures are evident. This occurs even when considering the initial high cost of startup for equipment and chemicals for a molecular biology laboratory. As new tools develop, the breeders will enhance and expand their capacities to address breeding problems and develop solutions. The challenges facing the breeder are still the same, e.g., disease and pest

USDA PLANT GENOME RESEARCH PROGRAM

155

resistance, yield enhancement, and increased tolerance to abiotic stresses such as heat, drought, and salt. This battle has been conducted for hundreds of years, The difference now is the speed of the counterattack against those ever present obstacles. Further, rzpid instrument throughput of experimental data and the incorporation of digital information into a readily accessible open database hasten development of research results. It is also becoming evident that future laboratory and field experiments will require the use of a computer; i.e., the recording results lab trend is over. Genome research is bridging the gap by serving as a catalyst between basic and applied genetics. That gap also contains plant physiologists/biochemists,as they also use molecular tools to address geneslgene systems/domains that regulate primary and secondary metabolism and that form wanted products and increased understanding of processes and functions in plants. The future will yield a continuous spectrum of plant science research from the most fundamental molecular to applied efforts. Therc will be system development that will encompass vertical integration of the agricultural industry from before and after the farm gate to processing, new farm and nonfarm products, retailing, and exports. This kind of approach is necessary to keep U.S. agriculture competitive in domestic and foreign markets. Plant genome research will play a critical role in advancing agriculture.

ACKNOWLEDGMENTS The following authors are participants of the USDA Plant Genome Research Group and summarized genome research in various plant groups and database parts of the document: S. Altenbach (small grains), USDA, ARS, WRRC, 800 Buchanan St., Albany, CA 94710; 0. Anderson (small grains), USDA, ARS, WRRC, 800 Buchanan St., Albany, CA 94710; D. Bigwood (Plant Genome Database) USDA, NAL, BARC-E, Beltsville, MD 20705; S. Cantinhour (Plant Genome Database), USDA, NAL, BARC-E, Beltsville, MD 20705; E. Coe (maize and sorghum), USDA, ARS, Plant Genetics Research, Curtis Hall, Rm. 210, Columbia, MO 6521 I; A. Datko (competitive grants), USDA, CRSEES, NRICG, 901 D St. SW, Rm. 323, Washington, DC 20250; S. Heller (Plant Genome Database), USDA, ARS, BARC-W, Bldg. 005, Rm 337, Beltsville, MD 20705; E. Kaleikau (competitive grants), USDA, CSREES, NRICG, 901 D St. SW, Rm. 323, Washington, DC 20250; S. McCouch (rice), Plant Breeding Dept., Bradford Hall 418, Cornell University, Ithaca, NY 14853-1901; R. Kohel (cotton), USDA, ARS, Crop Germplasm Research, Route 5, Box 805, College Station, TX 77845; J. Miksche (plant genome), USDA, ARS, BARC-W, Bldg. 005, Rm 331C. Beltsville, MD 20705; P. Moore (sugarcane), USDA, ARS, HI Sugar Planters Association, P.O. Box 1057, Aiea, HI 97601; D. Neale (woody species), USDA, FS, SWFES, 800 Buchanan St., Albany, CA 94710; T. Osborne (crucifers) Dept. of Agronomy, University of Wisconsin, Madison, WI 53706; A. Paterson (cotton and crucifers), Dept. of Soil and Crop Science, Texas A&M University, College Station, TX 77843-2474; R. Shoemaker (legumes), USDA, ARS, 1575 Agronomy Bldg., Rm G401, Iowa State University, Ames, IA 50011; G. Smith (plant genome data), USDA, ARS, NCRL, 1307 N. 18th St., P.O. Box 5677, Univ. Station, Fargo, ND 58105; S . Tanksley (Solanaceae), Plant Breeding Dept., Emerson Hall 248, Cornell University, Ithaca, NY 14853-1901; J. Wendel (cotton), Dept. of Botany, Iowa State University, Ames, IA 5001 1.

S. ALTENBACH ET AL.

156

REFERENCES

Adams, M. D., Kelley, J. M., Gocayne, J. D., Dubnick, M., Polymeropoulos, M. H., Xiao, H., Merril, C. R., Wu. A,, Olde, B., Moreno, R. F.. Kerlavage, A. R.. McCombie, W. R., and Venter, J. C. (1991). Complementary DNA sequencing: Expressed sequence tags and human genome project. Science 242, 1651-1656. Adams, M. A.. Dubnick, M.. Kerlavage, A. R., Moreno, R., Kelley, J. M., Utterback, T. R., Nagle, J. W., Fields. C., and Venter, J. C. (1992). Sequence identification of 2,375 human brain genes. Nature 355, 632-634. Ahn, S. N.,Bollich, C. N., and Tanksley. S. D. (1992). RFLP tagging of a gene for aroma in rice. Theorer. Appl. Gener. 84, 825-828. A h , S. N., Anderson, 1. A., Sorrells, M. E., and Tanksley, S. D. (1993). Homoeologous relationships of rice, wheat and maize chromosomes. Mol. Gen. Genet. 241. 483-490. Ahn, S. N., and Tanksley, S. D. (1993). Comparative linkage maps of the rice and maize genomes. h c . Narl. Acad. Sci. USA 90,1980-7984. Akkaya, M. S., Bhagwat, A. A.. and Cregan, P. B. (1992). Length polymorphisms of simple sequence repeat DNA in soybean. Generics 132, 1131-1139. Al-Janabi. S. M., Honeycutt, R. J., McClelland, M., and Sobral, B. W. (1993). A genetic linkage map of Saccharwn sponnirieuin L. ‘SES 208.’ Genetics 134, 1249-1260. Anderson, J. A., Ogihara, Y., Sorrells, M. E., and Tanksley. S. D. (1992). Development of a chromosomal arm map for wheat based on RFLP markers. Theorer. Appl. Genet. 83, 10351043.

Balls, W. L. (1906). Studies in Egyptian cotton. I n “Yearbook Khediv. Agric. Soc. for 1906,” pp. 29-89. Cairo, Egypt. Beasley, J. 0. (1942). Meiotic chromosome behavior in species hybrids, haploids, and induced polyploids of Gossypium. Generics 27, 25-54. Beavis, W. D., Grant, D., Albertsen, M. C., and Fincher, R. R. (1991). Quantitative trait loci for plant height in four maize populations and their association with qualitative genetic loci. Theore?. Appl. Gene?. 83, 141-145. Bennetzen, J. L.. and Freeling, M. (1993). Trends Genet. 9, 259-261. Bennetzen, J. L., and Melake-Berhan, A. (1994). I n “DNA-Based Markers in Plants” (R. L. Phillips and 1. K. Vasil. Eds.), pp. 291-298. Kluwer Academic, Norwell, MA. Binelli, G., and Bucci, B. (1994). A genetic linkage map of Picea abies Karst., based on RAPD markers, as a tool in population genetics. Theore?. Appl. Gene?. 88, 283-288. Blakey, C. A. (1993). Ph.D. Dissertation, University of Missouri, Columbia, MO. Blank, L. M., and Leathers, C. R. (1963). Environmental and other factors influencing development of southwestern cotton rust. Phyroparhology 53, 921-928. Bonierbale, M. W., Plaisted, R. L., and Tanksley, S. D. (1988). RFLP maps based on a common set of clones reveal modes of chromosomal evolution in potato and tomato. Genetics 120, 10951103.

Bradshaw, H. D., and Stealer, R. F. (1994). Molecular genetics of growth and development in Populus H. Segregation distortion due to genetic load. Theorer. Appl. Gener., (in press). Bradshaw, H. D., and Stettler, R. F. (1995). Molecular genetics of growth and development in fopulus. IV. Mapping QTLs with large effects on growth, fonn, and phenology traits in a forest tree. Generics, (in press). Bradshaw, H. D., Villar, M., Watson, B. D., Otto, K. G., Stewart, S., and Stealer, R. F. (1994). Molecular genetics of growth and development in fopulus. 111. A genetic linkage map of a hybrid poplar composed of RFLP, STS,and RAPD markers. Theorer. Appl. Genet., (in press). Brubaker, C. L., Koontz, J. A., and Wendel, J. F. (1993). Bidirectional cytoplasmic and nuclear

USDA PLANT GENOME RESEARCH PROGRAM

157

introgression in the New World cottons, Gossypium barbudense and G. hirsutum. Am. J. Bor. 80, 222-227. hummer. E. C., Bouton, J. H., and Kochert, G. (1993). Development of an RFLP map in diploid alfalfa. Theorer. Appl. Genet. 86, 329-332. Burr, B., Burr, F. A., Thompson, K. H., Albertsen, M. C., and Stuber, C. W. (1988). Genetics 118, 5 19-526. Camargo, L. E. (1994). “Mapping RFLP and Quantitative Trait Loci in Erussicu oleruceu.” Ph.D. Thesis, University of Wisconsin, Madison, WI. Carlson, J. E., ’hlsieram, L. K., Glaubitz, J. C., Luk, V. W., Kauffeldt, C., and Rutledge, R. (1991). Segregation of random amplified DNA markers in FI progeny of conifers. Theoret. Appl. Gener. 83, 194-200. Carson, S . D., and Carson, M. J. (1989). Breeding for resistance in forest trees: A quantitative genetic approach. Annu. Rev. Phyropurhol. 27, 373-395. Cartinhour, S . , Cherry, J. M., and Goodman, H. M. (1992). “An Introduction to ACEDB: For AAtDB, An Arubidopsis thulium Database,” Massachusetts General Hospital (available upon request in printed form from the AAtDB curator: ([email protected]). Chao, S . , Sharp, P. J., Worland, A. J., Warham, E. J., Koebner, R. M., and Gale, M. D. (1989). RFLP-based genetic maps of wheat homoeologous group 7 chromosomes. Theorer. Appl. Gener. 78, 495-504. Chao, S . , Baysdorfer, C., Heredia-Diaz, O., Musket, T., Xu, G., and Coe, E. H. (1994).Theorer. Appl. Gener. 88, 717-721. Cherry, J. M., and Cartinhour, S . W. (1993). ACEDB, A tool for biological information. I n “Automated DNA Sequencing and Analysis” (M. Adams, C. Fields, and C . Venter, Eds.), in press. Academic Press [text is available through ftp or gopher from weeds.mgh.harvard.edu (132.183.190.21)]. Chittenden, L. M., Schertz, K. F., Lin, Y. R., Wing, R. A., and Paterson, A. H. (1994). Theoret. Appl. Gener. 87. 925-933. Cho, Y. G., Eun, M. Y.,McCouch, S . R., and Chae, Y. A. (1994). Molecular mapping and genotypic selection for the semi-dwarf gene, sd-I, in rice (Oiyza sririvu L.).Theorer. Appl. Genet., (in press).

Coe, E., and Gardiner, J. (1994). In “DNA-Based Markers in Plants” (R. L. Phillips and 1. K. Vasil. Eds.), pp. 240-245. Kluwer Academic, Norwell, MA. Concibido, V. C., Denny, R. L., Boutin, S . R., Hautea, R., Orf, J. H.. and Young, N. D. (1994). DNA marker analysis of loci underlying resistance to soybean cyst nematode (Hererodera glycines Ichinohe). Crop Sci. 34, 240-246. Culp, T. W., and Harrell, D. C. (1974). “Breeding Quality Cotton at the Pee Dee Experiment Station, Florence, SC.” USDA Publ. ARS-5-30. U.S. Government Printing Office, Washington, DC. Culp, T. W., Harrell, D. C., and Kerr, T. (1979). Some genetic implications in the transfer of highfiber strength genes to upland cotton. Crop Sci. 19, 481-484. da Silva, J. A,, Sorrells, M. E., Nurnquist, W. L., and Tanksley, S. D. (1993). RFLP linkage map and genome analysis of Succhurum sponruneum. Genome 36, 782-791. Davis, J. M. (1995). Reverse genetics: An approach to study gene function in forest trees. Forest Genet., (in press).

Devey, M. E., Delfino-Mix, A,, Kinloch, B. B.. Jr., and Neale. D. B. (1994a). Efficient mapping of a gene for resistance to white pine blister rust in sugar pine. Proc. Nurl. Acud. Sci. USA., (in press).

Devey, M. E., Fiddler, T. A,, Liu, B. H.. Knapp, S. J., and Neale, D. B. (IW4b). An RFLP linkage map for loblolly pine based on a three generation outbred pedigree. Theorer. Appl. Genet. 88. 273-278. Devos. K. M., Atkinson, M. D., Chinoy. C. N., Liu, C. J., and Gale, M. D. (1992). RFLP-based

158

S . ALTENBACH ET AL.

genetic map of the homoeologous group 3 chromosomes of wheat and rye. Theorer. Appf. Genet. 83, 931-939. Devos, K. M., Millan, T., and Gale, M. D. (1993). Comparative RFLP maps of the homoeologous group 2 chromosomes of wheat, rye and barley. Theorer. Appl. Gener. 85. 784-792. Diers, B. W., and Osbom, T. C. (1994). Genetic diversity of oilseed Brussicu M ~ U Sgermplasm based on restriction fragment length polymorphisms. Theorer. Appl. Genet. 88. 662-668. Diers, B. W., and Shoemaker, R. C. (1992). Restriction fragment length polymorphism analysis of soybean fatty acid content. JAOCS 69, 1242-1244. Diers, B. W.,Keim, P., Fehr, W. R., and Shoemaker, R. C. (1992a). RFLP analysis of soybean seed protein and oil content. Theorer. Appl. Genet. 83, 608-612. Diers, B. W., Cianzio, S. R., and Shoemaker, R. C. (1992b). Possible identificationof quantitative trait loci affecting iron efficiency in soybean. J . Plant Nutr. 15, 2127-2136. Doebley, J., Stec, A.. Wendel, J., and Edwards, M. (1990). Pmc. Nurl. Acud. Sci. USA 87, 98889892. Dunham, I., Durbin, R., Mieg, J. T., and Bentley, D. R. (1993). Physical mapping projects and ACEDB. I n “Guide to Human Genome Computing” (M. J. Bishop, Ed.), in press. Academic Press (review) [text is available through ftp or gopher from weeds.mgh.harvard.edu ( 132.183.190.2 I)]. Echt, C. S., Kidwell, K. K., Knapp, S. J,, Osbom, T. C., and McCoy, T. J. (1994). Linkage mapping in diploid alfalfa (Medicugo saliva). Genome 37, 61-71. Edwards, M., Helentjaris, T., Wright, S., and Stuber, C. W. (1992). Molecular-marker-facilitated investigations of quantitative trait loci in maize. 4. Analysis based on genome saturation with isozyme and restriction fragment length polymorphism markers. Theorer. Appl. Genet. 83,765714. Ellis, T. H., ’Ibmer, L., Hellens, R. P., Lee, D., Harker, C. L., E n d , C., Domoney, C., and Davies, D. R. (1992). Linkage maps in pea. Generics 130, 649-663. Endrizzi, J. E., and Ramsay, G. (1979). Monosomes and telosomes for 18 of the 26 chromosomes of Gossypiurn hirsutum. Can. 1. Genet. Cyrol. 21, 531-536. Endrizzi, J. R., lbrcotte, E. L., and Kohel, R. J. (1984). Qualitative genetics, cytology, and cytogenetics. In “Cotton: Monograph Series Agronomy No. 24” (R. Kohel and C. Lewis, Eds.), pp. 81-129. American Society of Agronomy, Madison, WI. Ferreira, M. E., Satagopan, J., Yandell, B. S., Williams, P. H., and Osbom, T. C. (1994a). Mapping loci controlling vernalization requirement and flowering time in Brussicu nupus. Theorer. Appl. Genet., (in press).

Ferreira, M. E., Williams, P. H., and Osbom, T. C. (1994b). Mapping of a locus controlling resistance to AIbugo cundidu in Brussicu nupus using molecular markers. Phyroparhology, (in press).

Ferreira, M. E., Rimmer, S. R., Williams, P. H., and Osborn, T. C. (1994~).Mapping loci controlling Brassicu nupus resistance to Leptosphueria mucufuns under different screening conditions. Phyropurhofogy, (in press). Figdore, S. S.. Ferreira, M. E., Slocum, M. K.. and Williams, P. H. (1993). Association of RFLP markers with trait loci affecting clubroot resistance and morphological characters in B . oferuceu L. Euphyrica 69, 33-44. Figdore, S. S., Kennard, W., Song, K. M., Slocum, M. K.,and Osbom, T. C. (1988). Assessment of the degree of restriction fragment length polymorphism in Erussicu. Theorer. Appl. Genet. 75. 833-840. Fryxell, P. A. (1979). I n “The Natural History of the Cotton Tribe.” Texas A&M Univ. Press, College Station, TX. Fryxell, P. A. (1992). A revised taxonomic interpretation of Gossypium L. (Malvaceae). Rheedeu 2, 108- 165.

USDA PLANT GENOME RESEARCH PROGRAM

159

Funke, R. P., Kolchinsky. A., and Gresshoff, P. M. (1993). Physical mapping of a region in the soybean (Glycinemar) genome containing duplicated sequences. Plant Mol. B i d . 22,437-446. Gardiner, J., Coe, E. H., Melia-Hancock, S., Hoisington, D. A,, and Chao, S . (1993). Generics 134, 9 17-930. Gill, K. S., Lubbers, E. L., Gill, B. S., Raupp, W. J., and Cox, T. S. (1991). A genetic linkage map of Triricum tauschii (DD) and its relationship to the D genome of bread wheat (AABBDD). Genome 34, 362-374. Graner, A., Jahoor, A., Schondelmaier, J., Siedler, H., Pollen, K., Fischbeck, G., Wenzel, G., and Hemnann, R. G. (1991). Construction of an RFLP map of barley. Theoret. Appl. Genet. 83. 250-256. Grattapaglia, D. (1994). “Genetic Mapping of Quantitatively Inherited Economically Important Traits in Eucalyptus.” Ph.D. Dissertation, North Carolina State University. Grattapaglia, D., and Sederoff, R. (1994). Genetic linkage maps of Eucalyptus grandis and Eucalyptus urophylla using a pseudo-testcross: Mapping strategy and RAPD markers. Genetics 137, 1121-1137. Grattapaglia, D., Chapparro, J., Wilcox, P., McCord, S . , Werner, D., Amerson, H., McKeand, S., Bridgwater, F., Whetten, R., O’Malley, D., and Sederoff, R. (1992). Mapping in woody plants with RAPD markers: application to breeding in forestry and horticulture. In “Applications of RAPD Technology to Plant Breeding,” pp. 37-40. Crop Science Society of America/American Society for Horticultural Science/American Genetic Association, Minneapolis, MN. Grattapaglia, D., Chapparro, J., Wilcox, P., McCord, S., Crane, B., Amerson, H., Wemer, D., Liu. B., O’Malley, D., Whetten, R., McKeand, S., Goldfarb, B., Greenwood, M., Kuhlman, G.. Bridgwater, F., and Sederoff, R. (1993). Application of genetic markers to tree breeding. In “Proceedings of the 22nd Southern Forest Tree Improvement Conference, Atlanta, GA, June 14- 17, 1993.” pp. 452-463. National Technical Information Service, Springfield, VA. Groover, A,, Devey, M., Fiddler, T., Lee, J., Megraw, R., Mitchell-Olds, T., Shennan, B., Vujcic, S., Williams, C., and Neale, D. (1994). Identification of quantitative trait loci influencing wood specific gravity in an outbred pedigree of loblolly pine. Generics, (in press). Groover, A. T., Williams, C. G., Devey, M.E.,Lee, J. M., and Neale, D. B. (1995). Sex-related differences in meiotic recombination frequency in Pinus taeda. J . Hered., (in press). Hauge. B. M., Hanley, S. M., Cartinhour, S., Cherry, J. M., Goodman, H. M., Koornneef, M., Stam, P., Chang, C., Kempin, S., Medrano, L., and Meyerowitz, E. (1993). An integrated genetic/RFLP map of the Arabidopsis thaliana genome. Plant J . 3, 745-754. Hayes, P. M., Liu, B. H., Knapp, S. J., Chen, F., Jones, B., Blake, T., Franckowiak, J., Rasmussion, D., Sorrells, M., Ullrich, S. E.,Wesenberg, D., and Kleinhofs, A. (1993). Quantitative trait locus effects and environment interaction in a sample of North American barley gennplasm. Theorer. Appl. Genet. 87, 392-401. Hayes, P. M.. Matthews, D. E.,and the North American Barley Genome Mapping Project. (1994). Online dataset for the barley Steptoe X Morex mapping population. Files available via Internet Gopher, host greengenes.cit.cornelI.edu,poit 70, menu “NABGMP Steptoe X Morex dataset.” Helentjaris, T., Slocum, M., Wright, S., Schaefer, A., and Nienhuis, J. (1986a). Theorer. Appl. Genet. 72, 761-769. Helentjaris, T., Weber, D. F., and Wright, S . (1986b). Proc. Narl. Acad. Sci. USA 83, 6035-6039. Helentjaris, T., Weber, D., and Wright, S . (1988). Genetics 118, 353-363. Heun, M.,Kennedy, A. E., Anderson, J. A., Lapitan, N. L., Sorrells, M. E..andTanksley, S. D. (1991). Construction of a restriction fragment length polymorphism map for barley (Hordeurn vulgare). Genome 34, 437-447. Hodges, T. K., Peng, J., Lyznik, L. A., and Koetje, D. S. (1991). Transformation and regeneration of rice protoplasts. In “Rice Biotechnology” (G. Toenniessen and G. Khush, Eds.), pp. 157174. CAB International in association with International Rice Research Institute, Oxon, UK.

160

S . ALTENBACH ET AL.

Hofte, H., Desprez, T., Amselem, J., Chiapello, H., Caboche, M., Moisan, A., Jourjon, M., Charpenteau, J., Berthomieu, P., Guenier, D., Giraudat, J., Quigley, F., Thomas, F., Yu, D., Mache, R., Raynal, M., Cooke, R., Grellet, F., Delseny, M., Parmentier, Y., Marcillac, G., Gigot, C.. Fleck, J., Phillips, G., Axelos, M., Bardet, C., Tremousaygue, D., and Lescure, B. (1993). An inventory of 1152 expressed sequence tags obtained by partial sequencing of cDNAs from Arabidopsis rhaliana. Phnr J . 4, 1051-1061. Honecke, M., and Chyi, Y. S. (1991). Comparison of Erassica M P U S and E . rapa genomes based on RFLP mapping. I n “GCIRC Eight Intl. Rapeseed Congress,” Vol. 4, pp. 1102-1 107. Organizing Committee, Saskatoon, Saskatchewan, Canada. Hu, J., and Quiros, C. F. (1991). Identification of broccoli and cauliflower cultivars with RAPD markers. Phnr Cell Rep. 10, 505-51 I . Hulbert, S. H., Richter, T. E., Axtell, J. D., and Bennetzen, J. L. (1990). Pmc. Nafl.Acad. Sci. USA 87,4251-4255. Keim, P.. Diers, B. W., Olson, T. C., and Shoemaker, R. C. (1990). RFLP mapping in soybean: Association between marker loci and variation in quantitative traits. Generics 126, 735-742. Keith, C. S., Hoang, D. O., Barrett. B. M., Feigelman, B., Nelson, M. C., Thai, H., and Baysdorfer, C. (1993). Partial sequence analysis of 130 randomly selected maize CDNA clones. Planr Physiol, 101, 329-332. Kennard, W. C., Slocum, K. M., Figdore, S. S., and Osborn, T. C. (1994). Genetic analysis of morphological variation in Erassica oleracea using molecular markers. Theorer. Appl. Genet. 87, 721-732. Khush, G . S . , Singh, R. I., Sur, S. C., and Librojo, A. L. (1984). Primary trisomic of rice: origin, morphology, cytology and use in linkage mapping. Generics 107, 141-163. Kianian. S. F., and Quiros, C. F. (1992). Generation of a Erassica oleracea composite RFLP map: linkage arrangements among various populations and evolutionary implications. Theorer. Appl. Genet. 84, 544-554. Kimber, G. (l%l). Basis of the diploid-like meiotic behavior of polyploid cotton. Narure 191, 98-99. Kinlaw, C. S., and Gerttula, S. M. (1993). Complex gene families of pines. In “Proceedings of the 22nd Southern Forest Tree Improvement Conference, Atlanta, GA, June 14-17, 1993,” pp. 275-283. National Technical Information Service, Springfield, VA. Kinloch, B. B., Jr., and Comstock, M. (1981). Race of Cronarrium ribicola virulent to major gene resistance in sugar pine. Plant Dis. 65, 604-605. Kinloch, B. B., Jr., Parks, G. K., and Fowler, C. W. (1970). White pine blisterrust: simply inherited resistance in sugar pine. Science 167, 193-195. Kiss, G. C., Kalman, K., Kalo, P., and Okresz, L. (1993). Constmction of a basic genetic map for alfalfa using RFLP, RAPD. isozyme and morphological markers. Mol. Gen. Gener. 238,129137. Kleinhofs, A., Kilian, A., Saghai Maroof, M. A,, Biyashev. R. M., Hayes, P., Chen, F. Q . , Lapitan, N., Fenwick, A., Blake, T. K., Kanazin, V., Ananiev, E., Dahleen, L., Kudrna, D., Bollinger, J., Knapp, S. J., Liu, B., Sorrells, M., Heun, M., Franckowiak, J. D.. Hoffman, D., Skasden, R.,and Steffenson, B. J. (1993). A molecular, isozyme and morphological map of the barley (Hordeurn vulgare) genome. Theorer. Appl. Genet. 86, 705-712. Kleinhofs, A.. Kilian, D., Kudma, D., Saghai Maroof, M. A., and Steffenson, B. J. (1994). The barley genome map based on the Steptoe x Morex cross. Planr Genome 11: The Second Infernational Conference on rhe Plant Genome, 42. Kohel, R. J. (1989). Cotton. I n “Oil Crops of the World (G. Robbelen, R. Downey, and A. Ashir, Eds.), pp. 404-415. McGraw-Hill, New York. Kothari, S. L., Davey, M. R., Lynch, P. T., Finch, R. P., and Cocking, E. C. (1993). Transgenic rice. I n ‘Transgenic Plants” (S. Kung and R. Wu, Eds.), Vol. 2, pp. 3-20. Academic Press, New York.

USDA PLANT GENOME RESEARCH PROGRAM

161

Kowalski, S. P.. Lan, T. H., Feldman, K. A., and Paterson, A. H. (1994). Comparative mapping of Arubidopsis rhulium and Brussicu oleruceu chromosomes reveals islands of conserved organization. Generics 138, 499-5 10. Landau-Ellis, D. S., Angermuller, S., Shoemaker, R. C., and Gresshoff, P. M. (1991). The genetic locus controlling supemodulation in soybean (Glycine mar L.) co-segregates tightly with a cloned molecular marker. Mol. Gen. Gener. 228, 221-226. Landry, B., Hubert, N., Etoh, T., Harada, J. J., and Lincoln, S. E. (1991). A genetic map of Brussicu mpus based on restriction fragment length polymorphism detected with expressed DNA sequences. Genome 34, 543-552. Landry, B., Hubert, N., Crete, R., Chang, M. S., Lincoln, S., and Etoh, T. (1992). A genetic map of Brussicu oleruceu based on RFLP markers detected with expressed DNA sequences and mapping of resistance genes to race 2 of Plusmodiophoru brussicue (Woronin). Genome 35, 409420.

Lark, K. G., Weisemann, J. M., Matthew, B. F.. Palmer, R., Chase, K., and Malcalma, T. (1993). A genetic map of soybean (Glycine mar L.) using an intraspecific cross of two cultivars: ’Minsoy’ and ’Noir 1.’ Theoret. Appl. Gener. 86, 901-906. Laurie, D. A., and Bennett, M. D. (1986). Can. J . Genet. Cyrol. 28, 313-316. Laurie, D. A., and Bennett, M. D. (1988). Theorer. Appl. Genet. 76, 393-397. Laurie. D. A., and Bennett, M. D. (1989). Ann. Bor. 64, 675-681. Laurie, D. A., O’Donoughue, L. S., and Bennett, M. D. (1990). Srudler Gener. Symp. 19.95-126. Liu, Z., and Fumier, G. R. (1993). lnheritance and linkage of allozymes and RFLR in trembling aspen. J. Hered. 84, 419-424. Liu, A.. Li, H., Zhang, Q., Jiang, X.,Shi, S., and Yang, G. (1992). Mapping a wide compatibility gene of rice in relation to RFLP markers. J . Huurhong Agric. Univ. 11(3), 213-219. Lorenzen, L. L. (1994). “Soybean Cultivar Development: A Genome Perspective.” Ph.D. Dissertation, Iowa State University, Ames, IA. Lydiate, D., Sharpe, A., Lagercrantz, U., and Parkin, I. (1993). Mapping the Brussicu genome. Outlook Agric. 2. 85-92. Lynch, P. T., Finch, R. P., Davey, M. R., and Cocking, E. C. (1991). Rice tissue culture and its application. In “Rice Biotechnology” (G. Toenniessen and G. Khush, Eds.), pp. 135-155. CAB International in association with International Rice Research Institute, Oxon, UK. Mackill, D. J., Salam, M. A,, Wang, Z. Y.,and Tanksley, S. D. (1993). A major photoperiodsensitivity gene tagged with RFLP and isozyme markers in rice. Theorer. Appl. Gener. 85,536540.

Mansur, L., Lark, K. G., Kross, H., and Oliveira, A. (1993). Interval mapping of quantitative trait loci for reproductive, morphological and seed traits of soybean (Glycine m a L.).Theorer. Appl. Genet. 86, 907-913. McCarthy, J. C., Jenkins, I. N., Parrott, W. L., and Creech, R. G. (1979). The conversion of photoperiodic primitive race stocks of cotton to day-neutral stocks. MAFES Res. Rep. 4(19), 4 PP. McCouch, S. R., Abenes, M. L., Angeles, R., Khush. G. S., andTanksley, S. D. (1991). Molecular tagging of a recessive gene, xu-5, for resistance to bacterial blight of rice. Rice Genet. Newsl. 8, 143- 145. McCouch. S . R., Kochert, G., Yu, Z. H., Wang, Z. Y., Khush, G. S., Coffman, W. R., and Tanksley, S. D. (1988). Molecular mapping of rice chromosomes. Theorer. Appl. Gener. 76, 815-829.

McCouch, S. R., and Tanksley, S. R. (1991). Development and use of restriction fragment length polymorphism in rice breeding and genetics. In “Rice Biotechnology” (G. Toenniessen and G. Khush, Eds.), pp. 109-133. CAB International in association with International Rice Research Institute. Oxon, UK.

162

S. ALTENBACH ET AL.

McGrath, J. M., and Quiros, C. F. (1991). Inheritance of isozyme and RFLP markers in Brassica cumpesrris and comparison with E . oleracea. Theorer. Appl. Genet. 82, 668-673. McGrath, 1. M.. Jancso, M. M.,and Pichersky, E. (1993). Duplicate sequences with a similarity to expressed genes in the genome of Arabidopsis rhaliana. Theorer. Appl. Genet. 86, 880-888. Melake-Berhan, A., Hulbert, S. H., Butler, L. G., and Bennetzen, J. L. (1993). Theorer. Appl. Gener. 86, 598-604. Menancio-Hautea, D.,Fatokum, C. A., Kumar, L.,Danesh, D., and Young, N. D. (1993). Comparative genome analysis of mungbean (Vigna radiuia L. Wilczek) and cowpea ( V . unguicularu L. Walpers) using RFLP mapping data. Theorer. Appl. Genet. 86, 797-810. Meredith, W.R., Jr. (1984). Quantitative Genetics. I n “Cotton: Monograph Series Agronomy No. 24” (R.Kohel and C. Lewis, Eds.), Chap. 5, pp. 131-150. American Society of Agronomy, Madison, WI. Meredith, W. R., Jr., and Bridge, R. R. (1984). Genetic contributions to yield changes in upland cotton. In “Genetic Contributions to Yield Gains of Five Major Crop Plants” (W. Fehr, Ed.), pp. 75-86. Crop Science Society of America, Madison, WI. Meuhlbauer, G. J., Staswick, P. E., Specht, 1. E., Graef, G. L., Shoemaker, R. C., and Keim, P. (1991). RFLP mapping using near-isogenic lines in the soybean [Glycine max (L.) Merr.]. Theorer. Appl. Gener. 81, 189-198. Meyer, V. (1975). Male sterility from Gossypium harknessii. J . Hered. 66, 23-27. Meyer, J. R., and Meyer, V. G. (1961). Origin and inheritance of nectariless cotton. Crop Sci. 1, 167- 169.

Michelmore, R. W., Paran, I., and Kesseli. R. V. (1991). Identification of markers linked to diseaseresistance genes by bulked segregant analysis: A rapid method to detect markers in specific genomic regions by using segregating populations. Proc. Nurl. Acad. Sci. USA 88,9828-9832. Miksche, J. (1991). Strengthening Plant Genome Research Efforts-Goals of New USDA Program. Probe 1, 1-2. Mohan, M., Nair, S.,and Bennett, J. (1993). Mapping of a rice gene for resistance to biotype-I of gall midge (Orveoliu oryztie) by RFLP and RAPD analyses. In “Proceedings of the Sixth Annual Meeting of the International Program on Rice Biotechnology, February 1-5, 1993,” p. 13. Chiang Mai, Thailand. Muller, J. (1981). Fossil pollen records of extant angiosperms. Bor. Rev. 47, 1-142. Muller, J. (1984). Significance of fossil pollen for angiosperm history. Ann. MO Eor. Curd. 71,419443.

Murray, M. G., Ma, Y.,West, D. P., Romero-Severson, J., Cramer, J., Pitas, J. M., Kirschman, J.. and DeMars. S . (1989). Mol. Gen. Genet. 224, 5-9. Nagamura. Y., Antonio, B. A., Fukuda, A., Harushima, Y., Inoue, T., Lin, S . Y.,Shomura, A., Sue, N., and Yamamoto, K. (1993). A high density STS and EST linkage map of rice. Rice Genome 2(1), 3. Nance, W. L., ’hskan, G. A., Nelson, C. D., and Doudrick, R. L. (1992). Potential applications of molecular markers for genetic analysis of host pathogen systems in forest trees. Can. J . Foresr Res. 22, 1036-1043. Neale, D. B.,and Harry, D. E. (1994). Genetic mapping in forest trees: RFLPS, RAPDs and beyond. AgBiotechnol. News Inform. 6 , 107N-114N. Neale, D. B., and Williams, C. G. (1991). Restriction fragment length polymorphism mapping on conifers and applications to forest tree genetics and tree improvement. Can. J . Foresr Res. 21, 545-554.

Nelson, C. D., Nance, W. L., and Doudrick, R. L. (1993a). A partial genetic linkage map of slash pine (Pinus elliouii Engelm. var. elliorrii) based on random amplified polymorphic DNAS. Theorer. Appl. Genet. 87, 145-151.

USDA PLANT GENOME RESEARCH PROGRAM

163

Nelson, C. D., Doudrick, R. L., Nance, W. L., Hamaker, 1. M., andcapo, B. (1993b). Specificity of host: pathogen genetic interaction for fusifon-n rust disease on slash pine. In “Proceedings of the 22nd Southern Forest Tree Improvement Conference, Atlanta. GA, June 14-17, 1993,” pp. 403-410. National Technical Information Service, Springfield, VA. Nickell, A. D., Wilcox, 1. R., Lorenzen, L. L., Cavins, J. F., GufFy, R. G., and Shoemaker, R. C. (1994). The Fap2 locus in soybean maps to Linkage Group D. J. Hered. 85, 160-162. Nodari, R. 0.. Tsai, S. M., Gilbertson, R. L., and Gepts, P. (1993a). Towards an integrated linkage map of common bean. 11. Development of an RFLP-based linkage map. Theorer. Appl. Genet. 85, 513-520. Nodari, R. 0.. Tsai, S. M.,Guzmn, P., Gilbertson, R. L., and Gepts, P. (1993b). Toward an integrated linkage map of common bean. 111. Mapping genetic factors controlling host-bacteria interactions. Genetics 134, 341-350. O’Donoughue, L. S . , Wag, Z., Roeder, M.,Kneen, B., Leggett, M., Sorrells, M. E., and Tanksley, S. D. (1992). An RFLP-based linkage map of oats based on a cross between two diploid taxa ( A . venu utlunticu x A . hirrulu). Genome 35, 765-771. Osborn, T. C., Song, K. M., Kennard, W. C., Slocum, M. K., Figdore, S., Suzuki, J., and Williams, P. H. (1991). Genome analysis in Brussicu Using RFLR. In “Plant Molecular Biology II” (R. Hemnann and B. Larkins, Eds.), NATO AS1 Series 212, pp. 269-276. Plenum Press, New York. Park, Y.S.,Kwak, J. M., Kwon, 0.. Kim, Y.S., Lee, D. S.,Cho, M. J., Ize, H. H., and Nam, H. G. (1993). Generation of expressed sequence tags of random root CDNA clones of Brussicu nupus by single-run partial sequencing. Plant Phys. 103, 359-370. Paterson, A. H., Lander, E. S., Hewitt, J. D., Paterson, S.,Lincoln, S . E., and Tanksley, S. D. (1988). Resolution of quantitative traits into Mendelian factors by using complete linkage map of restriction fragment length polymorphisms. Nature 335, 72 1-726. Percival, A. E. (1987). The national collection of Gossypium germ plasm. So. Coop. Ser. Bull. 321, 1-362. Percy, R. G., and Wendel, J. F. (1990). Allozyme evidence for the origin and diversification of Gossypium burbudense L. Theoret. Appl. Genet. 79, 529-542. Pereira, M. G., Lee, M., Bramel-Cox, P., Woodman, W., Doebley, J., and Whitkus, R. (1994). Genome 37, 236-243. Polhill, R. M., Raven, P. H., and Stirton, C. H. (1981). Evolution and Systematics of the Leguminosae. In “Advances in Legume Systematics: Part I” (R. Polhill and P. Raven, Eds.), pp. 1-26. Royal Botanic Gardens, Kew, England. Prakash, S., and Hinata, K. (1980). Taxonomy, cytogenetics and origin of crop Erussicu. a review. O/M~U. Bot. 55, 1-59. Prince, J. P., Pochard, E., and Tanksley, S . D. (1993). Construction of a molecular linkage map of pepper and comparison of synteny with tomato. Genome 36, 404-417. Quiros, C. F., Hu, J., and TNCO,M. J. (1994). DNA-based marker maps of Brussicu. In “DNABased Markers in Plants” (R. Phillips and I. Vasil, Eds.). Kluwer Academic, Dordrecht, The Netherlands. Reinisch, A. R., Dong, J. M., Brubaker, C., Stelly, D., Wendel, J., and Paterson, A. H. (1994). A detailed RFLP map of cotton (Gossypium hirsutum x G . burbudense): Chromosome organization and evolution in a diso@c polyploid genome. Generics, (in press). Rines, H. W., Dahleen, L. S. (1990). Crop Sci. 30, 1073-1078. Robertson, D. S. (1989). Understandingthe relationship between qualitative and quantitative genetics. Plunr Cell Rep. 9, 81-87. Romero-Severson, J., Lotwer, J., Brown, C., and Murray, M. G. (1989). Use of RFLR for analysis of quantitative trait loci in maize. J . Genet. Breed. 43, 97-102.

164

S. ALTENBACH ET AL.

Ronald, P. C., Albano, B., Tabien, R.. Abenes, L., Wu, K., McCouch, S., and Tanksley, S. D. (1992). Genetic and physical analysis of the rice bacterial blight resistance locus, Xu-21. Mol. Gen. Gener. 236, 113-120.

Sadowski. J. P., Gaubier, P., Delseny, M., and Quires, C. (1%). Mapping of a gene complex formed by four Linked genes from Arabiubpsis in Brassicu genomes. In “Plant Genome n, The Second InternationalConfenme on the Plant Genome.” Scherago International, Inc., New York. Saito, A., Yano, M., Kishimoto, N., Nakagahra, M., Yoshimura, A., Saito, K., Kuhara. S . , Ukai, Y., Kawase. M., Nagamine, T., Yoshimura, S . , Ideta, 0..Ohsawa, R., Hayano, Y., Iwata, N., and Sigiura, M. (1991). Linkage map of restriction fragment length polymorphism loci in rice. Jpn. J. Breed. 41, 665-670. Shoemaker, R., and Specht, J. (1995). Integration of the soybean molecular and classical genetic maps. Crop Sci. 35, (in press). Shoemaker, R., Guffy, R.. Lorenzen, L., and Specht. J. (1992). Molecular genetic mapping of soybean: map utilization. Crop Sci. 32, 1091-1098. Slocum, M. K., Figdore, S. S . , K e ~ a r d W. . C., Suzuki, J. Y., and Osbom,T. C. (19%). Linkage arrangement of restriction fragment length polymorphism loci in Brassica oleracea. Theorer. Appl. Gener. 80, 57-64. Song, K., and Osborn,T. C. (1992). Polyphyletic origins of Brussicu napus: new evidence based on organelle and nuclear RFLP analyses. Genome 35, 992-1001. Song, K. M., Osborn,T. C., and Williams, P. H. (1990). Brassica taxonomy based on nuclear restriction fragment length polymorphisms (RFLR) 3. Genome relationships in Brussicu and related genera and the origin of B . olerucea and B. rupu (syn. cumpesrris). Theorel. Appl. Gener. 79, 497-506.

Song, K., Slocum, M. K., and Osbom,T. C. (1994). Molecular marker analysis of genes controlling morphological variation in Brussicu rupu (syn. cumpesrris). Theorer. Appl. Gener., (in press). Song, K. M.. Suzuki, J. Y., Slocum, M. K.. Williams, P. H., and Osbom,T. C. (1991). A linkage map of Brussicu rupu (syn. cumpestris) based on restriction fragment length polymorphism loci. Theorer. Appl. Gener. 82, 296-304. Staten, G. (1971). “Breeding Acala 1517 Cottons, 1926-1970.” New Mexico State Univ., College of Agric. and Home Econ., Memoir Series No. 4. Strauss, S. H., Lande, R., and Namkoong, G. (1992). Limitations of molecular-marker-aided selection in forest tree breeding. Can. J . Forest Res. 22, 1050-1061. Stuber, C. W., and Sisco, P. H. (1991). Proc. Annu. Corn Sorghum Res. Conf. 46, 104-1 13. Suenaga, K.,and Nakajima, K. (1989). Plunr Cell Rep. 8, 263-266. Tanksley, S . D. (1992). High density molecular maps of the tomato and potato genomes. Generics 132, 1141-1 160. Teutonico, R. A., and Osborn,T. C. (1994). Mapping of RFLP and qualitative trait loci in Brussicu rapa and comparison to linkage maps of B . M P U S , B. oleruceu and Arubidopsis thuliana. Theorer. Appl. Gener., (in press). Thieny-Mieg, J., and Durbin, R. (1991). ACEDB-A C. eleguns Database: 1. Users’ Guide. 11. Installation Guide. Ill. Configuration Guide. IV. Syntactic Definitions for the ACEDB Data Base Manager. By anonymous ftp from ncbi.nlm.nih.gov (130.14.20. I ) repositorylacedbl doc. 1-9.tar.Z. Thormann. C. E.. Ferreira, M. E., Camargo, L. E., Tivang, J. G., and Osbom, T. C. (1994). Comparison of RFLP and RAPD markers for estimating genetic relationships within and among cruciferous species. Theorer. Appl. Gener. 88, 973-980. Torres, A. M., Weeden, N. F., and Martin, A. (1993). Linkage among isozyme, RFLP and RAPD markers in Vicia f u h . Theorer. Appl. Gener. 85, 937-945. ’Tihieram, L. K., Glaubitz, J. C., Kiss, G., and Carlson, J. E. (1992). Single tree genetic linkage

USDA PLANT GENOME RESEARCH PROGRAM

165

mapping in conifers using haploid DNA from magagametophytes. BiolTechnology 10, 686690. ’hskan, G. A. (1992). Marker-aided selection: A tool for the improvement of forest tree species: Preface. Can. J. Forest Res. 22, 999-1000. Qler, F. J. (1908). ‘The Nectaries of Cotton,” Part V, pp. 45-54. U.S. Dept. Agr. Plant Ind. Bull. No. 131. Uchimiya, H., Kidou, S., Shiinazaki, T..Aotsuka, S., Takarnatsu, S., Nishi, R., Hashimoto, H., Matsubayashi, Y., Kidou, N., Umeda, M., and Kato, A. (1992). Random sequencing of CDNA libraries reveals a variety of expressed genes in cultured cells of rice (Oryza saliva L.). Planr J . 2,1005-1009. Wakamiya, I., Newton, R. J., Johnston, J. S., and Price, H. J. (1993). Genome size and environmental factors in the genus Pinus. Am. J. Bor. 80, 1235-1241. Wang. G . , Mackill, D. J., Bonman, I. M., McCouch, S. R., Champoux, M.C., and Nelson, R. J. (1994). RFLP mapping of genes confemng complete and partial resistance to blast in a durably resistant rice cultivar. Generics 136, 1421-1434. Weaver, D. B., and Weaver, J. B. (1977). Inheritance of ponen fertility restoration in cytoplasmic male-sterile upland cotton. Crop Sci. 17, 497-499. Webb, D. M., Baltazar, B. M., Rao-Arelli, A. P., Schupp, I., Clayton, K., and Keim, P. (1994). Genetic mapping of three soybean-cyst-nematode race-3 resistance loci in the soybean P1437654. Theorer. Appl. Genet., (in press). Weeden, N. F., Muehlbauer, F. J., and Ladizinsky, G. (1992). Extensive conservation of linkage relationships between pea and lentil genetic maps. J. Hered. 83, 123-129. Weisemann, J. M., Matthews, B. F., and Devine, T. V. (1992). Molecular markers located proximal to the soybean cyst nematode resistance gene, Rhg 4 . Theoret. Appl. Genet. 85, 136-138. Wendel, J. F. (1989). New World cottons contain Old World cytoplasm. Proc. Narl. Acad. Sci. USA 86, 4132-4136. Wendel, J. F., and Albert, V. A. (1992). Phylogenetics of the Cotton genus (Gossypium): Characterstate weighted parsimony analysis of chloroplast-DNA restriction site data and its systematic and biogeographic implications. Sysr. Bot. 17, 115-143. Wemer, J. E., Endo, T. R., and Gill, B. S. (1992). Toward a cytogenetically based physical map of the wheat genome. Proc. Narl. Acad. Sci. LISA 89, 11307-1131 1. Whitkus, R., Doebley, J., and Lee, M. (1992). Generics 132, 11 19-1130. Wilcox, P. L., Amerson, H. V., O’Malley, D., Carson, S., Carson, M. J., Kuhlman, G., and Sederoff, R. R. (1993). Fusiform rust-A model for marker assisted selection in loblolly pine. I n “Proceedings of the 22nd Southern Forest Tree Improvement Conference, Atlanta, GA, June 14-17, 1993,” pp. 174-182. National Technical Information Service, Springfield, VA. Williams, J. G., Kubelik, A. R.. Livak, K. J., Rafalski, J. A., and Tingey. S. V. (1990). DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. Nucleic Acids Res. 18, 6531-6535. Xiao, J., Fulton, T., McCouch, S., Tanksley, S., Kishimoto, N., Ohsawa, R., Ukai, Y., and Saito, A. (1992). Progress in integration of the molecular maps of rice. Rice Gener. Newsl. 9 , 124128. Xie, D. X . , Devos, K. M.,Moore, G., and Gale, M. D. (1993). RFLP-based genetic maps of the homoeologous group 5 chromosomes of bread wheat (Triricuni aesrivum L.). Theoret. Appl. Gener. 87, 70-74. Yik, C. P., and Birchfield, W. (1984). Resistant germplasm in Gossypium species and related plants to Rorylenchulus reniformis. J. Nematol. 16, 146-153. Young, N. D., Danesh, D., Menancio-Hautea, D., and Kumar, L. (1993). Mapping oligogenic resistance to powdery mildew in mungbean with RFLPS. Theorer. Appl. Gener. 87, 243-249.

166

S . ALTENBACH ET AL.

Young, N. D., Kumar, L.,Menancio-Hautea, D., Danesh, D., Talekar, N. S.. Shanmugasundarum, S., and Kim, D. H. (1992). RFLP mapping of a major bruchid resistance gene in mungbean (Vignu rudiara, L. Wilczek). Theorer. Appl. Genet. 84, 839-844. Yu, Z. H., Mackill, D. I., Bonman, J. M., and Tanksley, S. D. (1991). Tagging genes for blast resistance in rice via linkage to RFLP markers. Theoret. Appl. Genet. 81, 471-476. Yu, Y. G., Saghai-Maroof, M. A., Buss, G. R., Maughan, P. J., and Tolin, S. A. (1994). RFLP and microsatellite mapping of a gene for soybean mosaic virus resistance. Phytoparhology 84, 6064. Zamir, D., and Tanksley, S. D. (1988). Tomato genome is comprised largely of fast evolving, low copy-number sequences. Mol. Gen. Genet. 213, 254-261. Zhao, X., Lin, Y., and Paterson, A. H. (1994). Genetic mapping of DNA microsatellites from cotton. Plant Genome I 1 Proc., 74. Zheng, K., Shen, P.. Quian, H., and Wang, J. (1992). Tagging genes for wide compatibility in rice via linkage to RFLP markers. Chin. Rice Sci. 6(4), 145-150.

ANALYSISOF ORGANICMATTERIN SOILEXTRACTS AND WHOLESOILSBY PYROLYSIS-MASS SPECTROMETRY M. Schnitzerl and H.-R. Schulten2 'Centre for Land and Biological Resources Research, Agriculture and Agri-Food Canada Ottawa, Ontario, Canada K1A OC6 'Department of Trace Analysis, Fachhochschule Fresenius, 65 193 Wiesbaden, Germany

I. Introduction 11. Fundamentals of Pyrolysis-Mass Spectrometric Methods A. Soft Ionization Mass Spectrometry B. Pyrolysis-Field Desorption Mass spectrometry C. Pyrolysis-Field Ionization Mass Spectrometry D. Curie-Point Pyrolysis-Gas Chromatography/Electron Ionization Mass Spectrometry (Py-GUMS) 111. Analysis of SOM by Pyrolysis-Soft Ionization Mass Spectrometry A. Py-FIMS of n-Hexane-Chloroform Extracts of Humic Acids, Fulvic Acids, Paleosol Oh Horizon, and Soil Clay B. Py-FIMS of Supercritical n-Pentane Extracts of Humic Acids, Fulvic Acids, Paleosol Oh Horizon, and Soil Clay C. Py-FDMS of the n-Hexane-Chloroform Extracts of Humic Acids, Fulvic Acids, and Soil Clay D. FDMS of Supercritical n-Pentane Extracts of Humic Acids, Fulvic Acids, and Soil Clay E. Py-FIMS and Py-FDMS of Supercritical CO, Extracts of Whole Soils W. Summary of Data Obtained on the Extractions with Organic Solvents V. Curie-Point Py-GC/MS of Humic Acids and the Development of Novel Concepts for Their Chemical Structure A. Two-Dimensional Structures of Humic Acids B. Three-Dimensional Structures of Humic Acids and Soil Organic Matter VI. Analysis of Soil Organic Matter by Py-FIMS A. Py-FIMS of Annadale Humic Acid, Fulvic Acid, Humin, and Whole Soil B. Time-Resolved Py-FIMS VII. Effects of Minerals on the Py-FIMS of Fulvic Acid WI. Other Applications M.Conclusions References 167 Afvunca m Agnmoq, V d w 55 Copyright Q 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

168

M. SCHNITZER AND H.-R. SCHULTEN

I. INTRODUCTION The term soil organic matter (SOM) as used in this chapter refers to the sum total of all organic carbon-containing substances in soils. Chemically and physically, SOM consists of a mixture of plant and animal residues in various stages of decomposition, substances synthesized microbiologically and/or chemically from the breakdown products, and the bodies of live and dead microorganisms and small animals and their decomposing remains (Schnitzer and Khan, 1972). To simplify this very chemically complex and physically heterogeneous system, SOM is usually subdivided into nonhumic and humic substances. Nonhumic substances include those with still recognizable chemical characteristics (e.g., carbohydrates, proteins, fats, waxes, etc.). The bulk of SOM, however, consists of humic substances. These are amorphous, dark-colored, partly aromatic, polyelectrolyte-like materials that range in molecular weight from a few hundred to several thousand (Schnitzer, 1978). It is noteworthy that humic substances no longer exhibit the specific chemical and physical characteristics normally associated with well-defined organic compounds. In predominantly inorganic soils, which includes most agricultural soils, inorganic and organic soil constituents often are so closely associated that it is necessary to separate the two before either can be examined in greater detail. This separation is usually achieved by extracting the SOM. A vast amount literature exists on the extraction of SOM by a large number of different reagents under widely differing experimental conditions. SOM is usually partitioned into three fractions: (1) humic acid (HA), which is that fraction that coagulates when the alkaline extract is acidified to about pH 2; (2) fulvic acid (FA), which is the SOM fraction that remains in solution when the alkaline extract is acidified, that is, it is soluble in both alkali and acid; and (3) humin, which is that SOM fraction that remains behind, that is, it is insoluble in both alkali and acid. Over the years, many objections have been raised against the use of dilute alkaline solutions as SOM extractants. Stevenson (1982) suggested that the use of alkaline solutions could alter SOM through hydrolysis and oxidation. But because of their great efficiency, dilute alkaline solutions are still widely used for the extraction of SOM. To overcome some of the criticisms raised, the extractions are often done under a nitrogen atmosphere. Another serious difficulty with the extraction of SOM and its partitioning into HA, FA, and humin is that these are laborious and time-consuming procedures that are not suitable for the analysis of large numbers of soil samples. Also, aside from the separation of organic from inorganic soil components, extractions do not provide any chemical information on the organic materials that have been removed. It is in this area so far that SOM specialists have encountered many problems. Clearly, the need exits for new approaches not involving wet chemical

ANALYSIS OF SOIL ORGANIC MATTER

I69

methods. We have witnessed the rapid development of two analytical methods based on “high technology,” which appear to be suitable for the analysis of SOM. These are the combination of pyrolysis with soft ionization mass spectrometry and 13C NMR spectroscopy. Applications of the latter method to SOM analysis have been described by Wilson (1987) and Schnitzer (1991). ’ h o of the difficulties with applying 13C NMR to the analysis of SOM in whole soil and soil fractions are that the method requires the presence of more carbon than many agricultural soils contain and that paramagnetic Fe3+ interferes (Arshad et a f . , 1988). Also, in the case of SOM, signals in 13C NMR spectra tend to be broad and overlap, so that only information on carbon types (e.g., aliphatic, aromatic, C in CO,H groups) can be derived from these spectra. A promising procedure, capable of providing information on SOM at the molecular level, is pyrolysis-soft ionization mass spectrometry, a method pioneered by Schulten (1977). In this method, the sample is pyrolyzed directly under vacuum in the ion source of the mass spectrometer, and the volatile components are identified by soft ionization (field ionization or field desorption) mass spectrometry. Programmed linear heating of the sample from 50 to 750°C with a relatively slow heating rate of 10 K min-1 is generally employed, which permits the temperature-resolved evolution of thermal degradation processes. In this manner, the efficiency of pyrolysis-field ionization mass spectrometry (PyFIMS) is greatly improved in two ways. First, high mass resolution and accurate mass measurements provide information on the elemental composition of pyrolysis products. Second, by programming the temperature rise time and the starting and ending temperatures of the pyrolysis procedure, the direct introduction system can be utilized as a kind of “chromatographic”separation device for the fractionation of the mixtures. Under these conditions, little mass spectrometric fragmentation occurs, so that predominantly molecular ions are observed in the mass spectra. These features of Py-FIMS facilitate the identification of signals when analyzing complex mixtures. In this review, we trace the development of applications of this method to the analysis of organic extracts of whole soils, HAS, FAs, and humins, soil particle size fractions, whole soils, and claySOM complexes. Another pyrolysis-mass spectrometric method described in this chapter is Curie-point pyrolysis-gas chromatography/electron ionization mass spectrometry (Py-GUMS). This method is characterized by very high heating rates (flash pyrolysis) and raises the sample temperature on the order of 0.1 s to a preset final pyrolysis temperature [e.g., 510°C (alloy) or 770°C (pure iron)]. A comprehensive description of the method and collection of pyrolysis-mass spectra has been given by Meuzelaar et a f .(1982). Although limited by mass range (mostly <300 Da) and intricately lower temperature resolution, the advantages of this method are the combination of chromatographicparameters with mass spectrometry and the possibility of using computer-supported library searches for reliable identi-

170

M. SCHNITZER AND H.-R. SCHULTEN

fications (with quantitative data such as purity, fit). Applications of the latter, indirect method to structural analyses of HAS and FAs will be highlighted. The purpose of this review is to describe and evaluate applications of mainly Py-FIMS to the analysis of SOM and of Py-GUMS to structural studies on humic substances. Because so far the most systematic and comprehensive studies of the combined methods in this area have been pursued by the authors and their co-workers, examples of applications of these methods to SOM analysis are taken from our own work. Hopefully, this situation will change in the future as the use of these methods spreads and new types of applications are developed.

II. F U N D A M E N T A L S OF PYROLYSIS-MASS SPECTROMETRIC METHODS Mass spectrometry is one of the most expensive and sophisticated analytical techniques with respect to both the instrumental requirements and the skill and proficiency of the operating staff. Nevertheless, in the areas of agriculture, biochemistry, chemistry, environmental research, medicine, and soil science and in the control of industrial production, the extent of applications of mass spectrometry has increased during the past two decades and is still growing. This is due to the fact that mass spectrometric detection simultaneously can be highly sensitive, specific, and reliable for a large number of compounds with molecular weights up to approximately 100,OOO mass units, using special ionization modes. Sample amounts required for a mass spectrometric measurement can range from 1 pg (10-12 g) to the microgram (10-6 g) level. The analytical resolution of the method for complex mixtures is high, particularly when high mass resolution and/or temperature-resolved spectra are obtained or when coupled with chromatographic separations. These unquestionable advantages of mass spectrometry are of particular value when used in the aforementioned areas since the inherent analytical tasks often consist of a trace (ppm) or ultratrace (ppb) determination of organic or inorganic compounds. Perhaps the most limiting factor of mass spectrometry is inherent in the first step of the mass spectrometric procedure, during which free gaseous ions must be produced from the sample under investigation. Obviously, this conversion can cause difficulties for the investigation of compounds that are present in the solid state at ambient temperature and the high vacuum conditions in the ion source of the mass spectrometer. In the classical mode of ionization by electron impact, the compounds are first evaporated and then ionized by collision with 10-70 eV of electrons. Under these conditions, many polar organic compounds show thermal decomposition during the evaporation process and extensive muss spectrometric fragmentation due to the high energy transfer upon electron impact. Thus, mo-

ANALYSIS OF SOIL ORGANIC MATTER

171

lecular ions often are of small relative intensity or not even detectable. Of particular interest for the following investigationof very complex mixtures is the fact that even relatively intense molecular ions of a mixture component can be completely obscured by mass spectrometric fragments of other components. Therefore, alternate ionization modes for soft ionization were developed that reduce mass spectrometric fragmentation and enhance molecular ion intensities. It is clear that the larger the thermal fragments and the more reliable their identifications, the easier it is to derive the initial structures of the sample molecules.

A. SOFT IONIZATION MASSSPECTROMETRY Soft ionization mass spectrometry of organic molecules in the high electric field employed in field ionization (FI) and field desorption (FD) mass spectrometry (MS)has been pioneered by Beckey (1977). Early work in the 1960s showed the capacity of FIMS for analysis of mixtures. Despite the lower sensitivity of the

Oven

/

Counterelect rode

Adsorbed, Solid Sample

Field Strength

10' -10'

V/cm 7

Pot ent ial Difference approx. 10 OOOV

Wgure 1. Definition for distinguishing between Fl and FD.(a) The sample is supplied to the emitter via the gas phase: Fl. (b) At the beginning of the analysis, the sample is present at the surface of the emitter: FD [from Schulten (1979) with permission of the publisher].

172

M. SCHNITZER AND H.-R. SCHULTEN

method compared to electron ionization (El) MS, the outstanding features of field ionization mass spectra are strongly reduced mass spectrometric fragmentation and high molecular ion intensities for a wide range of mass and polarity of mixture components. The two techniques, FI and FD, are distinguished by different ways in which the sample reaches the emitter. Figure 1 illustrates the differences between the two techniques: If the sample molecules approach the emitter from the gas phase, via the gas inlet system or the direct probe, we refer to it as field ionization; if the sample is absorbed from a solution or suspension and is ionized from the absorbed state, we call it field desorption. Thus, the mode of sample supply allows for clear differentiation between FI and FD. However, when it comes to the ionization process at the emitter, there is a more fluid transition between the two techniques. It is important to remember these definitions when examining the FI and FD mass spectra in this chapter. With Py-FIMS, all volatile components generated during the heating program, e.g., between 50 and 700°C sample temperature, can be detected in extracts, whole soils, or soil particle size fractions. General experience indicates that the recorded mass range is about 3 times larger than that of spectra produced by conventional electron ionization. With Py-FDMS, the sample has to be dissolved in order to be adsorbed on the emitter surface. Thus, only compounds soluble in selected solvents are observed, but they can be detected in a mass range that is about 3-5 times larger than that obtained with Py-FIMS .

B. PYROLYSIS-FIELD DESORPTION MASSSPECTROMETRY In field desorption, positively charged ions are formed from neutral molecules

M in a very high electric field (field strengths 107-108 V/cm) either by loss of an electron, leading to odd electron (radical) molecular ions ([MI+.), or by attach"a]+, leading to even electron ment of positively charged ions such as [HI+, (more stable) molecular ions ([M+ HI+,[M + Na]+, etc.). For relatively volatile compounds, the production of FD ions starts at room temperature immediately after the electric field is applied to the emitter, which consists of a 10-pm tungsten wire that is activated by carbon dendrites for field enhancement and absorption surface for the sample. A convenient way to transfer the sample droplets to the FD emitter is to control the deposition of the solutions (concentration about 1 pg/ pl) by means of a stereomicroscope (approximate magnification X 8 0 to X 100) and a 10-1.1.1 syringe mounted on a small manipulator (Schulten, 1979). Desorption of the sample molecules in the high vacuum of the ion source is achieved by a direct heating current or indirectly by a laser beam. FDMS is particularly advantageous for two reasons. First, by using commercial double-focusing magnetic instruments, high mass resolution up to 100,OOO

ANALYSIS OF SOIL ORGANIC MATTER

173

m l h (10% valley definition) can be achieved. Second, access to high muss

molecular ions is available and has been reported at high mass resolution for >4000 (Schulten, 1980) and >15,000 Da by Rollins et al. (1990). Drawbacks are the low absolute intensities of the total ion currents produced and contaminations due to solvent impurities such as inorganic salts and plasticizers. For a more detailed review of the principle, methodology, and analytical applications of FDMS, see Schulten (1979).

C. PYROLYSIS-FIELD IONIZATION MASSSPECTROMETRY For temperature-resolved F'y-FIMS, about 100 pg of a humic substance such as HA, FA, humin, or 5 mg of whole soil samples is thermally degraded in the ion source of a MAT 731 (Finnigan, 28127 Bremen, Germany) modified high performance (AMD Intectra GrnbH, 27243 (Harpstedt, Germany) mass spectrometer. In Fig. 2 the instrumental setup for the FI/FD ion source is shown schematically (Schulten et al., 1987a). Note that alternative ionization modes such as chemical ionization (CI), fast atom bombardment (FAB) ionization, and laser ionization (LI) are available, and the registration of positive and negative ions is possible. The slotted cathode plate is the first part of ion optics, and any

Mass Spectrometer ,Slotted

Heating Coil ( 3 A )

n Laser Beam

Cathode Plate

/

-

.........

rTL

1

=

~11

siiatng Rod

Quartz Sample Thermocouple

Flpre 2. Schematic drawing of the modified experimental setup for pyrolysis-field ionization mass spectrometry [from Schulten er al. (1987a) with permission of the publisher].

174

M. SCHNITZER AND H.-R. SCHULTEN

magnetic or quadrupole mass spectrometer can be utilized. The samples are weighed before and after Py-FIMS (error kO.01 rng) to determine the pyrolysis residue and the produced volarile matter. A heatablekoolable direct introduction system with electronic temperature programming, adjusted at the +8 kV potential of the ion source, and a field ionization emitter are used. The slotted cathode plate serving as the counterelectrode is at -6 kV potential. Thus, at 2-mm distance between the emitter tips and the cathode a total potential difference of 14 kV is applied, resulting in an extremely high electric field strength, which is the essential basis for soft ionization. All samples are heated in high vacuum (1.3 X 10-4 Pa) from 50 to 700°C at a heating rate of approximately 10 K min-1. Depending on the volatility and thermal stability of the sample materials, 40-60 magnetic scans of the gaseous pyrolyzate components are recorded for the mass range 16- lo00 Da. In general, at least three replicates are performed for each sample. The total ion intensities (TII) of the single spectra are normalized to l-mg sample weight, averaged for replicate runs, and plotted against the pyrolysis temperature, which then results in Py-FIMS thermograms. For the selection of biomarkers and quantitative evaluations, particularly for whole soils and soil particle size fractions, detailed descriptions of the method have been published (Schnitzer and Schulten, 1992; Schulten er al., 1993; Sorge et al., 1994).

Sample Amount

F l p r e 3.

[pg]

-

Py-FIMS calibration curve of a cellulose standard for quantitative evaluations.

ANALYSIS OF SOIL ORGANIC MATTER

175

In principle, reliable identification is a prerequisite for quantification. As illustrated in Fig. 3, a standard such as crystalline cellulose (Avicell PH105) yields an acceptable calibration curve when the summed total ion intensities of Py-FI mass spectra are plotted against sample amounts between 10 and 320 pg. Basic Py-FIMS studies have shown a significant linear correlation between the C,, concentrations of whole soils, particle size fractions, litter materials, and humic fractions and the total ion intensity per milligram of sample (Sorge et al., 1993). Moreover, off-line pyrolysis under nitrogen of whole soils and particle size fractions and C,, and N, analyses of the residues indicated that about 3090%of C, and 6040% of N, are volatilized. Approaching the same volatilization for Py-FIMS, the mass spectra recorded for these samples represented, on average, 50 and 75% of sample C,, and N,, respectively (Leinweber and Schulten, 1995).

D. CURIE-POINT PYROLYSIS-GAS IONIZATION MASS CHROMTOGRAPHY/ELECTRON SPECTROMETRY (PY-GC/MS) Figure 4 displays the experimental setup for Py-GUMS (Schulten and Schnitzer, 1992). Samples (1) of humic substances, soils extracts, and soils are pyrolyzed in a q p e 03 16 Curie-point pyrolyzer (2) (Fischer, 53340 Meckenheim, Germany). The materials are not pretreated except for drying and, if necessary, milling. The final pyrolysis temperatures ( T , ) employed are 300,500, and 700”C, respectively. The total heating time (THT) was varied between 3 and 9.9 s. Following split injection (split ratio 1:3; flow rate 1 m1/20 s), the pyrolysis products are separated on a gas chromatograph (5) (Varian 3700, 64289 Darmstadt, Germany) equipped with a 30-m capillary column (DB5) coated with 0.25-pm film thickness and an inner diameter of 0.32 mm. The starting temperature for the gas chromatographic temperature program is 4O”C, and the end temperature is 250°C with a heating rate of 10 K min-1. The gas chromatograph was connected to a nitrogen-selective, thermosensitive detector (TSD) and a double-focusing Finnigan MAT 2 12 mass spectrometer. Conditions for mass spectrometric detection in the electron ionization mode are +3-kV accelerating voltage, 70-eV electron energy, 2.2-kV multiplier voltage, 1.1 s/mass decade scan speed, and a recorded mass range between mlz 50 and 500. A detailed description of the principle, potential, and limitations of Py-GUMS of humic fractions and soils has been presented (Schulten and Schnitzer, 1992). Furthermore, studies of organic nitrogen-containing compounds in soils using the TSD detector are in press. The latter expanded our knowledge of the structure of the “unknown” soil nitrogen (Schulten et al., 1995a).

176

M. SCHNITZER AND H.-R. SCHULTEN HELIUM

11I-11

J3 -I

I (HIGH MASS RESOLUTION

I

I

F l p r e 4. Schematic display of the instrumental setup for Curie-point Py-GC/MS combined with flame detection (FID), electron capture (ECD),and thennosensitive (TSD)detectors. These four independent detector systems for the pyrolysis products give on-line information on (1) all ions recorded by electron ionization or field ionization mass spectrometry, (2) organic materials, (3) halogentontaining compounds, and (4) NIP compounds, respectively [from Schulten and Schnitzer (1992) with permission of the publisher].

III. ANALYSIS OF SOM BY PYROLYSIS-SOFT MASS SPECTROMETRY I0"ION A. Py-FIMS OF N-HEXANE-CHLOROFORM EXTRACTS OF HUMIC Acms, F u L . .ACIDS, ~ PALEOSOL OHHORIZON, AND Son. CLAY The earliest studies by Schnitzer and Schulten (1989; Schnitzer et af., 1990a) focus on the Py-FIMS of organic solvent extracts of a HA, a FA, a Gleysolic Paleosol from the Arctic estimated to be 45,000,000 years old, a fine clay fraction separated from soil, and a whole soil. The solvents tested were n-hexane, chloroform, supercritical n-pentane, and supercritical COz. The HA was isolated from the surface horizon of a Udic Boroll, and the FA was extracted from the Bh horizon of a Haplaquod. The fine clay fraction was separated from the Oh horizon of a Qpic Haplorthod. The Q-FI mass spectrum of the combined hexane-chloroform extract of the HA is shown in Fig. 5. The presence of high mass n-alkanes is indicated by mlz 576 (C,,),842 (Ca),and 884 (C63). The most abundant component is the CZ4 n-fatty acid, which is shown by the molecular ion ([MI+.) at mlz 368 as base peak and an intense protonated molecule ([M + HI+) at m / z 369. Similarly protonated molecules are observed for dihydroxyflavone at m/z 241 and diols

177

ANALYSIS OF SOIL ORGANIC MATTER 1003

t5

368 241

80-

t7'

259

99

60-

396

40-

41 0

1 29

20 149

100

a , 0

-*

200

300

400

40

a, DL

20

500

600

m / z

700

800

830

900

Flpre 5. Py-Fl mass spectrum of the n-hexane and chloroform extracts of the HA [from Schnitzer and Schulten (1989) with permission of the publisher].

such as hexadecanediol (cl6)at m/z 259 and tetracosanediol (Cz4) at m / z 371. The presence of a homologous series of n-fatty acids is indicated by mlz 256 ( C I ~ )312 , (Czo), 326 ( C ~ I )340 , ( C Z ~ )368 , (CzJ, 396 (cz6),410 (C2gh 424 (C29), 452 (C31),and 480 (C33). An extended homologous series of molecular ions centered around m / z 676 and 704 is indicative of the presence of n-alkyl monoesters with the following carbon distribution: mlz 508 (c34), 536 (c3&564 (C3g), 592 (Cm), 620 (C4J9 648 (Cub 662 (C45h 676 (C46h 704 (c48),760 (c52), 788 (CS4),816 (c56h 830 ( W , and 844 (CSg). The m-Hmass spectrum of the combined hexane-chloroform extract of the FA is presented in Fig. 6. This spectrum is dominated by a homologous series of n-fatty acids, ranging from m / z 256 (cl6)and 284 (CIg)to 508 (C34), with m/z 368 (Cz4) and m/z 396 (c26) being the most abundant components. The extract also contains the following n-alkyl monoesters: mlz 732 (Cs0), 760 (C5z), 788 (C54), 802 (C55), 816 (c5.6)~ 830 (C57), 844 ( C S ~ )and , 900 (C62)- In addition, relatively weak signals of the following n-alkanes appear to be present among the higher-molecular-weight components: m / z 814 (CSg),828 (C5& 842 (Cm), 856 (C6J, 870 (c62), and 898 (W.

178

M. SCHNITZER AND H.-R. SCHULTEN 368

1

I 'Lo

40 20

3-4

100

300

200

400

1001

500

600

m / z

700

800

900

Flare 6. Py-FI mass spectrum of the n-hexane and chloroform extracts of the FA [from Schnitzer and Schulten (1989) with permission of the publisher].

The Py-FI mass spectrum of the n-hexane extract of the Gleysolic Paleosol Oh horizon is shown in Fig. 7. The spectrum shows the fragment ions mlz 157 and 297 and the molecular ion mlz 424 of 10-nonacosanol, which is typical of epicuticular waxes of conifers (Schulten et al., 1987b). Signals at mlz 270 and n-fatty acids, respectively, whereas rnlz 300 284 are assigned to the C,, and could be dehydroabietic acid and mlz 302 abietic acid. The presence of sterols is indicated by rnlz 386 (cholesterol), with the corresponding water elimination at rnlz 368, 396 (ergosterol), and 416 (stigmastanol). The n-C32 alcohol at mlz 448 is the most prominent component. The spectrum shows the presence of the following n-alkyl monoesters: mlz 620 (C42),648 (CM), 676 (c46), 704 (c48), 732 (C~O), 760 (CS~), 774 (C53h 788 (C541, 802 (C55h 816 (&), 830 (C57h 858 (c59h872 (ca),886 (C61)v 900 (c62), 928 (ctj&956 (CM),984 (c68)9 and 1012 (C,o). Molecular ions of n-alkyl diesters are mlz 986 (CM), 1168 (C79), 1196 (Cgl), 1224 (c83), and 1266 (c86). Weak signals of triesters are indicated in the signals at mlz 1406 and 1408. It is noteworthy that Py-FIMS of this hexane soil

ANALYSIS OF SOIL ORGANIC MATTER

4

'"q 80

,"

300

200

2

179

60-

750

400

500

600

700

830

850

950

1050

1150

1250

1350

1450

1550

m / z Flpre 7. Py-Fl mass spectrum of the n-hexane extract of the Gleysolic Paleosol Oh horizon [from Schnitzer er al. (1990a) with permission of the publisher].

extract produced molecular ions >1600 Da. Later work by Schnitzer et af. (1990b) showed the analytical potential of the combination of solid state 13C NMR and soft ionization mass spectrometry for direct investigations of complex organic matter in soils without pretreatment. Lipids identified in the hexane extract clearly indicated the presence of natural waxes (Murray and Schulten, 1981) from the coniferous and deciduous vegetation that dominated this area under a warm and moist climate 45 Ma ago. The F'y-FI mass spectrum of the combined hexane and chloroform extracts of the fine clay fraction is shown in Fig. 8. Fragment ions arising from (branched) saturated aliphatics are at mlz 57,7 1, 85,99, and 113 and from olefins at mlz 83, 97, 111, 125, and 153. The most prominent signals in the lower mass range are the molecular ions of the hexadecanediol at mlz 258, tetracosanediol at mlz 370, and the n-C,, fatty acid at rnlz 256. The signal at mlz 308 could be due to the [M-H20]+- ion of the Cz2n-alcohol. A homologous series of n-alkanes begins at mlz 212 (CIS)and continues to mlz 562 (C,). The presence of n-alkyl monoesters is indicated by mlz 620 (C42), 648 (C4), 676 (C4& 704 (C4& 732 (C5& 760 (C52)9 788 (C54h and 816 (C5d.

M.SCHNITZER AND H.-R. SCHULTEN

180

100

a,

0

200

300

400

676 I

0

80 3

n

60

-*

40

4

20 500

600

m / z

700

800

900

Figure 8. Py-Fl mass spectrum of the n-hexane and chloroform extracts of the soil clay [from Schnitzer and Schulten (1989) with permission of the publisher].

B. Py-FIMS OF SUPERCRITICAL N-PENTANE EXTRACTS OF HUMIC ACIDS,Fu~vrcACIDS,PALEOSOL OH HORIZON, AND SOIL CLAY

Cunent trends are to develop methods for the extraction of specific SOM components. One such method is the supercritical fluid extraction (SFE)of soils and SOM with supercritical CO,, n-pentane, methanol, and ethanol-H,O and acetone-H,O mixtures (Spiteller, 1985; Schnitzer et al., 1986; Schnitzer and Preston, 1987). Depending on the solvent system, these methods extract aliphatic, aromatic, N-containing, and other SOM components and so facilitate identifications of the major components of the extracts. This section shows applications of Py-FIMS to the analysis of supercritical n-pentane extracts of the same materials as those employed in Section A. The Py-FI mass spectrum of the supercriticaln-pentane extract of the HA (Fig. 9) is dominated by n-fatty acids, n-alkanes, sterols, and smaller amounts of n-alkyl monoesters. The presence of CI6-C,, n-fatty acids is indicated by signals ranging from m / z 256 to 620. Molecular ions of n-alkanes range from mlz 324

ANALYSIS OF SOIL ORGANIC MATTER

181 41 4

1

a) 0 rl

0’ Q

100

200

300

400

0°7

60$

45

It. .

m / z Figure 9.

-

Py-FI mass spectrum of the supercritical n-pentane extract of the HA.

(C23) to 814 (c58); alkyl fragments appear at m / z 71, 85, and 99. Strong sterol signals at m/z 396 (ergosterol, C28H,0), mlz 410 (dehydrostigmasterol, C2+-I6O), mlz 412 (stigmasterol, C2J14,0),and especially at mlz 414 (p-sitosterol, C29H500) are observed. Signals at m / z 482 and 496 could be due to C31 and C,, diols, while signals typical of monoesters are at mlz 704 (c48), 746 (C5,), 760 (C52). 774 (C53), 788 (C5J, 816 (&), and 830 (C5,). In addition, at m / z 846 a weak diester signal marks the end of the mass range. Plant-derived indicators such as dehydroabietic acid (C20H2802) at mlz 300 and abietic acid (C2oH3oO2) at rnlz 302 are of particular interest (Hempfling et af., 1991). The Fy-FI mass spectrum of the supercritical n-pentane extract of the FA (Fig. 10) is quite similar to that of the HA (Fig. 9). However, it shows the presence of a wider range of n-fatty acids from m / z 256 (c16) to 648 (C,) and more pronounced sterols at mlz 386 (cholesterol) and mlz 400 (ethylcholesterol). The sterols at mlz 410 (dehydrostigmasterol),412 (stigmasterol), and 414 (p-sitosterol) again show a similar pattern. The spectrum also displays the presence of n-alkyl monoesters at mlz 774 (C53), 816 (c56)9 and 830 (C5,) and a diester at mlz 846 (C5J. The Py-FI mass spectrum of the supercritical n-pentane extract of the gleysolic paleosol Oh horizon (Fig. 11) shows fragments of n-alkanes at m / z 71, 85, and

182

M. SCHNITZER AND H.-R. SCHULTEN

a, V

200

100

c 1002

300

400

I

n -

80 13

n 60 6 -. 40 0)

20

500

700

600

m / z

800

900

Flgure 10. Py-Fl mass spectrum of the supercritical n-pentane extract of the FA.

100

I

80 3

60

508

I

40

7114

20

a,

100

0

200

300

400

600

500

700

Q *

40

PI

20

1281 I

~

800

900

1000

m / z

"

~

'

'

1100

l

~

'

'

'

~

1200

~

~

- - 1500 --

' I ' ~I ~ ' ~~ ~ ' ['' ' -'- I.- 8~ -

1300

1400

.

~1

Ffgure 11. Py-Fl mass spectrum of the supercritical n-pentane extract of the gleysolic paleosol.

183

ANALYSIS OF SOIL ORGANIC MATTER

99. Ions at mlz 157, 297, and 406 are fragments, and mlz 424 is the molecular ion of 10-nonacosanol, a thermally very labile biomarker for coniferous epicuticular waxes (Schulten et al., 1987b). The presence of a homologous series of n-fatty acids is indicated by signals ranging from m / z 256 (&) to 536 (c36).Ions at m / z 324,338, 352 are the C23-CZ5 n-alkanes, respectively, while mlz 396 and 416 appear to be ergosterol and stigmastanol, respectively. The extract also contains a wide series of n-alkyl monoesters starting with m / z 620 (C42) and 648 (C,) and extending to mlz 900 (c62).Signals at m / z 987 and 1281 appear to be due to (protonated) C, and c g 7 n-alkyl diesters. The Py-FI mass spectrum of the supercritical n-pentane extract of the fine clay fraction (Fig. 12) shows n-alkyl fragments at rnlz 71, 85, 99, and 113. This extract contains two major components: n-fatty acids and n-alkyl monoesters. At m/z 340-480 are C22-C32 n-fatty acids, with the n-C24 (mlz 410) being the most prominent component; mlz 390,392, and 394 appear to be sterols, while m/z 448 and 476 are (M - H20]+. molecular ions of the C,, and C,, n-alcohols. A homologous series of n-alkyl monoesters starts at mlz 592 (C40) and extends to m/z 956 (C&. The most abundant monoesters are the c 4 g and CS7.The signal at m/z 1164 appears to be the n-Cg3 alkane and mlz 1310 is the n-Cs3 triester.

410

100 A

80

60 40

20 0

100 a,

u

c 80

0 -0

C

60

L3

-

a,

w

40

20 0 750

850

950

1050 m / z

1150

1250

1350

1450

>

figure 12. Py-H mass spectrum of the supercritical n-pentane extract of the soil clay [from Schulten and Schnitzer (1990) with permission of the publisher].

M. SCHNITZER AND H.-R. SCHULTEN

184

371

1001

80-

I

112

60-

259

279

39 1

.

590 620 648 I

l l

p!

100

800

200

900

300

1000

m / z

400

500

600

-

1100

1200

~

1300

741 761

I

l '

-

'

'

"

'

'

I

~

700

1400

1500

Elgure 13. Py-FD mass spectrum of the n-hexane and chloroform extracts of the HA.

C. PY-FDMS OF

OF THE N-HJZXANE-CCHLOROFORM EXTRACTS

HUMICACIDS, FULVIC ACIDS, AND Son, CLAY

The Py-FDmass spectrum of the combined hexane-chloroform extracts of the HA (Fig. 13) exhibits the presence of the n-C,, and n-C,6 fatty acids at m/z 242 and 256, respectively. A series of n-alkanes begins at m/z 296 (C21)and extends to m/z 590 (C4& Other n-alkanes are m/z 1052 (C,s) and 1416 (Clol).The most abundant components of the extract are m/z 370 and 258, the n-C24 and n-CI6 diols, respectively. The presence of small amounts of n-akyl monoesters is indicated by a homologous series that starts with m/z 648 (C,) and continues to m/z 968 (c68).A diester can be noted at m/z 986 (C,) and triesters at m/z 1310 (CS,) and 1394 G 3 ) . The Py-FD mass spectrum of the combined hexane-chloroform extracts of the FA (Fig. 14) exhibits signals for n-fatty acids at m / z 256 (c16)and 284 (C18)and for diols at mlz 258 (c16)and 370 (C24). The latter is the strongest signal in the spectrum. Signals of n-alkanes are m/z 324 (C23), 338 (C24), 352 (C25), and 492 (C35). For m/z 984 and 986, the n-C68 alkyl monoester and n-C& diester are tentatively assigned, respectively. The Py-FDmass spectrum of the combined hexane-chloroform extracts of the fine clay fraction is shown in Fig. 15. Molecular ions of n-alkanes are mlz 310

~

~

~

-

~

ANALYSIS OF SOIL ORGANIC MATTER

185

0

U 8 3

6 -. e,

80; 60-

985

4020-

986

1416

----+

m / z

Elpre 14. 4-FDmass spectrum of the n-hexane and chloroform extracts of the FA.

676

10080-

100

750

200

850

300

950

400

-

1050

m / z

500

1150

1250

600

1350

700

1450

Elgure 15. 4-FD mass spectrum of the n-hexane and chloroformextracts of the soil clay [from Schnitzer and Schulten (1989) with permission of the publisher].

M. SCHNITZER AND H.-R. SCHULTEN

186

394 (C28h 408 (c22), 324 (c231, 338 ((&), 352 (CZ~), 366 (c2& 380 (C29), and 436 (C3& and with decreasing abundances up to mlz 590 (C42). Signals [M-H20]+. of alcohols are m / z 336 (C24), 364 (C26), 378 (C27), 392 (c28), 406 (C29), 434 (C31), 448 (C3z), and 490 (C34. The most prominent components of this extract are saturated n-alkyl monoesters, ranging from m/z 592 (C40)to m / z 760 (C52), with m / z 676 (C,) and 704 (c48) being the most abundant monoesters. Other monoesters can be seen at mlz 788 (Cs4),8 16 (&), and 844 (c58). The presence of a homologous series of diesters is indicated by 846 (c56)~874 (c58), 902 (Ca), 930 (Cfj2h 958 (c&$),986 (Ccjfj), 1014 (c68)9 1042 (C70), 11 12 (C74, and 1168 (C7,&. A series of triesters can be seen ranging from mlz 1184 (c78) to 1408 (C9J. The most prominent components of this series are m/z 1380 (C92), 1352 (Cw), 1408 (C94),and 1324 (c88).

D. Py-FDMS OF SUPERCRITICAL N-PENTANE EXTRACTS OF HUMIC ACIDS, FULVIC ACIDS,AND Son. CLAY The Py-FD mass spectrum of the supercritical n-pentane extract of the HA (Fig. 16) is dominated by n-alkanes. Minor components are n-fatty acids, al-

100

408

III

1

4?6

40-

100

3

400

300

200

500

600

60 830 846

-

lto9 1

''W 1000 1100 I

700

800

m / z

900

I

I

1164 .

. r-

'

1248 1282

.

.

11

8

1200

Figure 16. Py-FD mass spectrum of the supercritical n-pentane extract of the HA [from Schnitzer and Schulten (1989) with permission of the publisher].

ANALYSIS OF SOIL ORGANIC MATTER

187

kanes, and n-alkyl di- and triesters. The presence of n-fatty acids is indicated by mlz 228 (Cl4), 242 (CIS), 256 (c16), 270 (C17), and 284 (c18). A homologous series of n-alkanes begins at mlz 296 (C21)and extends to 562 (C,). The most prominent n-alkanes in this series range from rnlz 352 (CZ5)to 436 (C3,). Other n-alkanes in this extract are mlz 898 (Cu), 1038 (C74), 1066 (c76), 1164 (c83), 1234 (c88), and 1248 (&). This spectrum also shows the presence of n-alkenes, which range from mlz 308 (C22)to 714 (C51).The presence of n-alkyl monoesters is indicated by signals at mlz 928 (Ca), 956 (Ca), and 970 (c67), that of diesters by mlz 972 ((3 and 986 (C& and that of triesters by mlz 1170 (C75), 1184 (c76), and 1282 (CS~).

The Py-FD mass spectrum of the supercritical n-pentane extract of the FA (Fig. 17) is dominated by the presence of a homologous series of n-alkanes that range from m / z 282 (C,) to 562 (Cm). The most prominent n-alkanes are the C,,, c,,, and Cz9.Other n-alkanes are mlz 604 (C43),828 (Cs9), 856 (c61), and 1038 (C74).Fatty acids produce signals mlz 256 (cl6) and 284 (Clg), whereas n-alkyl monoesters range from mlz 648 (C,) to 830 (C57).For n-alkyl diesters, signals at mlz 846 (c56), 930 (&), and 986 ( C,) are observed. The Py-FD mass spectrum of the supercriticaln-pentane extract of the fine clay fraction (Fig. 18) shows the presence of a homologous series of n-alkanes ex394

830

600

700

m / z

800

900

1000

8

1100

Figure 17. Py-FD mass spectrum of the supercritical n-pentane extract of the FA [from Schnitzer and Schulten (1989) with permission of the publisher].

188

M. SCHNlTZER AND H.-R. SCHULTEN

100A

am

80-

412

60-

4020704

0,

loo-,

8t 0

U

C

50

hy)

8o:r 60-

3

2

40-

m / z

>

Elgun! 18. Py-FD mass spectrum of the supercritical n-pentane extract of the soil clay [from Schulten and Schnitzer (1990) with permission of the publisher].

tending from mlz 282 (Cz0)to 534 (c38). Note the prominence of mlz 366 (nCZ6),380 (tt-C,7), and 394 (n-czs). The presence of significant concentrations of n-fatty acids is shown by the series from mlz 228 (C,4) to 340 (C24).A homologous series of n-alkyl monoesters is centered around mlz 704 (c48) and 830 (CS7).Signals at m / z 986 and 1322 appear to indicate the CM and C, n-alkyl diesters, and mlz 1254 is the c83 n-alkyl triester.

E. Py-FIMS AND PY-FDMS OF SUPERCRITICAL COz EXTRACTS OF WHOLE SOILS Supercritical fluid extraction of soils with CO, is of special interest to soil scientists because it can be done under relatively mild conditions. The supercritical pressure of CO, is 7.35 MPa, and the supercritical temperature is 31°C (Weast, 1979). The only known reported extraction of SOM with supercritical CO, is that published by Spiteller (1985), who identified a variety of fatty substances in the extracts. The supercritical fluid extraction with CO, described in this section is that of the Bh horizon of the Armadale soil. The Py-H mass spectrum of the supercriti-

189

ANALYSIS OF SOIL ORGANIC MATTER

cal CO, extract of the Armadale Bh horizon (Fig. 19a) shows weak fragment ions of saturated aliphatics at mlz 113, 127, 141, and 155. The most abundant component is mlz 427, whose molecular composition, as determined by high resolution mass measurements, is C,,HmO4. The signal at mlz 427 is, therefore, due to protonated decanedioic acid dioctyl ester. The diester eliminates octane (mlz 112) in two steps to produce mlz 315 (the c,, dioic acid) and mlz 203 (decanedioic acid). Loss of water from this dicarboxylic acid results in fragment ion mlz 185. Molecular ion mlz 424 and fragment ions rnlz 406,297, and 157 are typical of 10-nonacosanol, a common constituent and typical biomarker for coniferous, cuticular waxes (Schulten et al., 1987b). The presence of a homologous series of n-fatty acids is indicated by rnlz 228 (CI4)continuing to mlz 592 (Cm). Most prominent among these n-fatty acids are mlz 396 (&), 424 (C2& and 452 (C3,,). In addition, a homologous series of n-alkyl monoesters starts with mlz 564 (c38) and extends to mlz 914 (c63). The most prominent n-alkyl monoesters are rnlz 676 (C4& 704 (C4& 760 ( C 4 , and 788 (C54)The Py-FD mass spectrum of the supercritical CO, extract of the Armadale Bh horizon (Fig. 19b) shows protonated mlz 427, the c,, dioic acid monooctyl

100'6160

424c7 X)

80-

a>

452

396 x 5

60-

r--

704

40-

985

a, 0

r

100

200

300

4Otz7 500

600

700

ao 72 I a-. a,

cf

rn 1000

900

1

5

n 3

800

b)

604

x 5

r-I

424 -

40-112

20127

185 I

m / z

I

-

949

I977 1

1

'

1

1

Fppre 19. (a) Py-Fl mass spectrum of the supercritical CO, extract of Armadale Bh horizon soil and (b) Py-FD mass spectrum of the supercritical CO, extract of the Armadale Bh horizon soil [from Schulten and Schnitzer ( 1 9 9 1 ) with permission of the publisher].

190

M. SCI-NITZER AND H.-R. SCHULTEN

ester, as the most abundant component. A homologous series of n-fatty acids begins at m / z 228 (C14) and extends to m/z 578 (C39). The signal at m / z 436 could be the C30H,0 ketone. The presence of a homologous series of n-alkyl monoesters is indicated by m / z 620 (C42)and extends to m / z 844 (C58).The most abundant n-alkyl monoesters are m / z 676 (C&), 704 (C48), and 732.3(( Both FIMS and FDMS show that n-fatty acids, dioic acids, and n-alkyl monoesters are the major components of the supercritical CO, extract of the soil sample.

IV. SUMMARY OF DATA OBTAINED ON THE EXTRACTIONS WITH ORGANIC SOLVENTS Major components of all extracts are n-alkanes, n-fatty acids, diols, sterols, and n-alkyl mono-, di-, and triesters. Minor components are n-alcohols in the soil clay extracts and alkenes in the SF extract of the HA (Table I). Significant components of the HA extracts are the n-C17-n-C101 alkanes, CISC,, n-fatty acids, C22-C51 alkenes, c2&2, sterols, C16r C24, C31, and C,, diols, C,-c68 n-alkyl monoesters, c65 and c, n-alkyl diesters, and C7,-c9, n-alkyl triesters. Abundant constituents of the FA extracts are the CZo-C74 n-alkanes, cl6-c34 n-fatty acids, C24 diol, c27-c29 sterols, c&& n-alkyl monoesters, and c56C, n-alkyl diesters.

Table I Distribution of Major Components in Organic Extracts of the HA, FA, SOU Clay, and Armadale sdl Components n- Alkanes n- Alkenes

n-Fatty acids n- Alcohols Diols Sterols n-Alkyl monoesters n-Alkyl diesters n-Alkyl triesters Dioic acids (I

HA

FA c20-c74

Soil clay

Armadale soil c17-c34

nd'

CLl

c16-c34

c14-c30

nd

c29

c27-c29

nd nd

c44-c68

c38-cS2

C,-C66

nd nd

c24

nd nd

nd, not detected (relative abundances 6%).

c18* c26

ANALYSIS OF SOIL ORGANIC MATTER

191

In the soil clay extracts, n-alkyl monoesters centering around C , and ranging from C,, to C5, are the major components. Other significant constituents are the C, diol, C,,-C,, n-alcohols, and C23-C3t n-alkanes. These extracts contain more triesters (n-C,, to n-Cw) than diesters (n-C, to n-C,,). The most prominent triester is n-C,,, followed by n-C,. Major components of the supercritical CO, soil extract are n-fatty acids, n-alkyl monoesters, and dioic acids. While the nature of the components in all extracts is similar, they are present in different proportions. Extracts from the soil clay are richer in n-alkyl esters but poorer in n-alkanes and n-fatty acids than are the HA and FA extracts. The high-molecular-weight n-alkyl esters that were identified in the organic extracts of the HA, FA, and fine soil clay are similar to those detected by Schulten et al. (1986) in spruce wax and Carnauba wax. These observations suggest that the compounds identified and listed in Table I are components of natural waxes that appear to be associated with SOM.

V. CURIE-POINT Q-GC/MS OF HUMIC ACIDS AND THE DEVELOPMENT OF NOVEL CONCEPTS MIR THEIR CHEMICAL STRUCTURE Initial studies of soil humic acids by pyrolysis-gas chromatographywere first published by Nagar in 1963, and at about that time Giacobbo and Simon (1964) reported the first results of Curie-point pyrolysis. Applications of Curie-point pyrolysis-gas chromatography/mass spectrometry (Py-GUMS) to the analysis of SOM have been reported by Saiz-Jimenez and de Leeuw (1984, 1986); applications to polluted sediments were reported by de Leeuw et al. (1986), to forest and agricultural soils by Hempfling and Schulten (1988, 1990, 1991), and to aquatic humic substances by Abbt-Braun et al. (1989). This method is valuable for structural studies on humic materials of widely differing origins because the transfer of thermal energy to the sample is a fast process, with temperature rise times on the order of milliseconds (Irwin, 1982). The resulting thermal shock produces small, stable organic molecules.

A. TWO-DIMENSIONAL STRUCTURESOF HUMICACIDS The chemical structure of humic substances has been the subject of much speculation and research over the past 200 years. In spite of the availability of abundant experimental data in the literature, soil and environmental scientists so far have been unable to propose a valid concept for the chemical structure of

192

M.SCHNITZER AND H.-R. SCHULTEN

these materials. But because of the importance of this subject, more research on the structural chemistry of humic substances is needed. Curie-point pyrolysis-mass spectrometric analysis on two HAS [extracted from surface horizons of the Bainsville (an Haplaquoll) and Armadale (a Haplaquod) soils] was performed by Schulten and Schnitzer (1992) with the objective of obtaining structural information on these materials, which had previously been examined intensively by pyrolysis-soft ionization mass spectrometry (Schulten, 1987). Focal points were the confirmation of structural assignments and identification of common pyrolysis products by these complementary analytical tools. As shown in Table 11, the major compounds produced from the two HAS are benzenes and n-alkylbenzenes. Of special interest is the series of C,-C,, n-alkylbenzenes (Schulten et al., 1992). In addition, ethylmethylbenzene, methylpropylbenzene, methylheptylbenzene, methyloctylbenzene, and methylundecylbenzene are essentially members of the same series of alkylbenzenes that have important functions in the three-dimensional structure of humic substances. Other compounds of interest are trimethyl- and tetramethylbenzenes, alkylnaphthalenes, and alkylphenanthrenes. The alkyl substitution on naphthalene ranges from one to five methyls, while on phenanthrene it extends from one to four methyls. The alkylaryl compounds identified in Table I1 consist of aromatic rings that are linked covalently to aliphatic chains. Schulten et al. (1991) propose that these “building blocks” are released during pyrolysis from an alkylaromatic structural network that is made up of the constituents listed in Table 11. This preliminary two-dimensional structure contains voids of various dimensions that can trap and bind other organic and inorganic components. Note that the Armadale HA is richer in most compounds listed in Table I1 than the Bainsville HA. This may be due to differences in the origins of the two HAS. The Armadale HA was extracted from the Bh horizon of a Haplaquod, about 25 cm below the surface, while the Bainsville HA was extracted from the surface of the Bainsville soil, a Haplaquoll. One of the most striking features of the Bh horizon is its very low microbial activity, which may have led to better preservation of the compounds listed in Table 11. From the data in Table II, it appears that alkylaromatics are significant structures in HAS. In the past, these compounds have been largely overlooked (or reported as impurities, artifacts, etc.) because of the unavailability of adequate and sensitive combinations of analytical methods to detect and explain them. A chemical structure for the basic HA skeleton based on alkylbenzenes, alkylnaphthalenes, and alkylphenanthrenes is shown in Fig. 20. It is important to consider the frequently occumng f sign of a randomly continuous structure, which opens up a vast variety and complexity of different linkages that are characteristic for the three-dimensional networks of humic substances. A more complete version of the HA structure is presented in Fig. 21 in which

ANALYSIS OF SOIL ORGANIC MATTER Table II Building Blocks of Bainsville and Amadale HAS as Identified by Curie-Point Pyrolysis-Gas Chromatography/MassSpectrometry lntensitf Armadale

Bainsville

***** ***** **** * **

*** ***** **** * **

**

**

*

* * ** * ** ** * * * ** * * * **

* * *

*

*

* *

* * * * * * * *** * ** *** * ** **

*** * * ** * *

Compounds Benzene Toluene Ethylbenzene, xylenes Ethylmethylbenzene Propylbenzene Trimethylbenzenes Butylbenzene Methylpropylbenzene Tetramethylbenzene Pentylbenzene Hexylbenzene Heptylbenzene Octylbenzene Methylheptylbenzene Nonylbenzene Methyloctylbenzene Decylbenzene Methylnonylbenzene Undecylbenzene Meth y ldecylbenzene Dodecylbenzene Methy lundecylbenzene Tridecylbenzene Tetradecylbenzene Pentadecy lbenzene Hexadecylbenzene Heptadecy lbenzene Octadecylbenzene Nonadecy lbenzene Eicosylbenzene Hemicosylbenzene Docosylbenzene Styrene Methylstyrene Indene lndane Fluorene Naphthalene Methylnaphthalenes (continued)

193

194

M. SCHNlTZER AND H.-R. SCHULTEN Table I1 (conrinued) Intensity Armadale

** **

*

* *

* * * *

a

Bainsville

* * *

Compounds Dimethylnaphthalenes Trimethylnaphthalenes Tetramethylnaphthalenes Pentamethylnaphthalene Phenanthrene Methylphenanthrene Dimethylphenanthrene Trimethylphenanthrene Tetramethylphenanthrne

Intensity of peak height: *****, 80-I00%;

***, 40-60%; **, 20-40%; *, observed.

****, 60-808;

lQwe 20. Chemical structure for HA based on alkylammatic “building blocks” [from Schulten et al. (1991) with permission of the publisher].

ANALYSIS OF SOIL ORGANIC MATTER

195

oxygen, hydrogen, and nitrogen atoms have been inserted in conformity with analytical data obtained on many naturally occurring soil HAS. Oxygen is present in carboxyls, phenolic and alcoholic hydroxyls, and ether groups, and nitrogen is present in heterocyclic structures and as nitriles. The elemental composition of the HA structure in Fig. 21 is C,,,H,,,O~,, with a molecular weight of 5540 Da and an elemental analysis of 66.8% C, 6.0% H, 26.0% 0, and 1.3% N. There are different views in the literature on SOM of whether carbohydrates and proteinaceous materials are absorbed on or loosely retained by HA or whether they are bonded covalently to HA (Sowden and Schnitzer, 1967). But regardless of which mechanism is considered, carbohydrates and proteinaceous materials are HA components for analytical purposes because their presence affects the elemental analysis and functional group content of HAS. Carbohydrates have been reported to constitute about 10%of the HA weight (Lowe, 1978); a similar value has been suggested for proteinaceousmaterials in HA (Khan and Sowden, 1971). Thus, Schulten and Schnitzer (1993a) assume that a molecular weight of HA interacts with 10% carbohydrates and 10% proteinaceous materials. The resultwith a molecular ing HA has an elemental composition of C342H3880124N,2, weight of 6651 Da and an elementary analysis of 61.8% C, 5.9% H, 29.8% 0, and 2.5% N. When more carbohydratesand proteinaceous materials are added to the HA, the C content decreases but the 0 content increases.

Flpre 21. State of the art structural concept for HA [from Schulten and Schnitzer (1993a) with permission of the publisher].

196

M. SCHNITZER AND H.-R. SCHULTEN Table 111 Analytical Characteristics of HAS Extracted from Soils Belonging to Three MITerent Great Soil Groups and of the Proposed HA Structure

c (46) H (%) N (96)

s (%) 0 (%) TQtal acidity (meq/g) CO,H (meqlg), phenolic OH (meq/g), alcoholic OH (meqlg) OCH, (meq/g)

Udic Boroll

Haplaquod

Haplaquoll

Proposeda

56.4 5.5 4. I

58.2 5.4 3. I 0.7 32.6 5.7 3.2 2.5 3.2 0.4

54.2 6.0 6.0 0.9 32.9 6.4 3.5 2.9 3.0 0.4

61.8 5.9 2.5 29.8 5.8 4.4 I .4 I .4 0.3

1.1

32.9 6.6 4.5 2.1 2.8 0.3

aMW = 6651 Da.

The elemental composition and functional group content of HAS extracted from soils belonging to three different great soil groups, as well as similar data for the proposed HA structure, are shown in Table 111. A comparison of these data indicates that the analytical data for the proposed HA structure are in general agreement with those of HAS extracted from soils. The HA structure presented in Fig. 21 is in agreement with chemical (Schnitzer, 1978), oxidative and reductive degradation (Hansen and Schnitzer, 1966, 1969), colloid-chemical (Ghosh and Schnitzer, 1980). electron microscopic (Stevenson and Schnitzer, 1982), and I3C NMR and X-ray (Schnitzer et al., 1991a) investigations done on HASover many years and exhaustive consultation of the enormous amount literature on this subject. One of the more interesting features of the HA structure shown in Fig. 21 is the presence in the structure of voids of various dimensions that can trap or bind other organic components, such as carbohydrates, proteinaceous materials, lipids, and biocides, as well as inorganics such as clay minerals and hydrous oxides. The oxidative degradation of this structure would produce the benzenecarboxylic acids that have been isolated repeatedly as major oxidation products of humic substances (Schnitzer, 1978).

B. THREE-DIMENSIONAL STRUCTURES OF HUMIC Acms AND SOIL ORGANIC M.XTTER

Because practically all pyrolysis data so far have been obtained and published as two-dimensional plots, it was of interest to illustrate progress in commercially

ANALYSIS OF SOIL ORGANIC MATTER

197

available, relatively low cost software and personal computer equipment, which allows three-dimensional displays and computer-assisted design (CAD) of chemical structures and model reactions. In particular, the possibilities for molecular modeling and geometry optimizations of complex macromolecules, which are often the target of analytical pyrolysis, virtually open up a new dimension. This is demonstrated in the following for humic substances and soil organic matter, probably the most complex naturally occurring materials, as an example. For the humic acid structure described in Fig. 2 1, all three-dimensional work, model construction, chemical interaction studies, and semiempirical calculations used the HyperChem software (release 2 for Microsoft Windows 3.1) as described previously by Schulten ( 1995a). The personal computer employed consisted of an IBM-compatible PC system (486DX2 with a 66-MHz processor, VLB 34 in combination with 8 MB RAM, 250 MB disk, and a 17-in. color monitor). After modeling and geometry optimization of the handdrawn twodimensional HA (see Fig. 21) by HyperChem, the black and white threedimensional structure in Fig. 22a was obtained. This “skeleton” display is pro-

Elgure 22. Three-dimensional display of the proposed HA structure shown in Fig. 21 following molecular modeling by the HyperChem software. (a) Black and white plot of the structure in the “Sticks” mode; three hydrogen bonds are indicated by m w s . (b) (color plate 1) Color plot in the “Disk” mode. Elements are carbon (cyan), hydrogen (white), oxygen (red), and nitrogen (blue) [from Schulten (1995a) with permission of the publisher]; and (c) (color plate 2) Color plot of a HA oligomer complex illustrates the covalent bonds and intermolecular hydrogen bridges between five humic subunits resulting from the proposed HA model.

198

M. SCHNITZER AND H.-R. SCHULTEN

duced in a first approach by using default parameters for bond lengths, angles, torsions, bends, hydrogen bonds, and van der Wads forces. For simplicity and to allow a perspective overview, bonds are shown as lines and atoms as points (“Sticks” mode). Step by step geometry Optimization and energy minimization of the HA structure by semiempirical methods resulted in a total energy of 710.70 kJ nm-1 mol-1 and a convergence gradient of 0.037 kJ nm-I mol-I after 4700 calculation cycles (Schulten, 1995a). The corresponding structure as a color graph is shown in Fig. 22b (color plate 1) (“Disk” mode) and gives an illustration of the energy minimization process and spatial requirements. Rotation of the energy-minimized HA 3D version demonstrates the flexible network with voids and hollows that offer binding and trapping of biological and anthropogenic molecules. The space requirements for carbohydrates and peptides are fulfilled, and distances, angles, and interaction parameters, e.g., hydrogen bonds, clearly demonstrate that these substances can be trapped and bound in the voids of the threedimensional HA structure. Thus, the hypothesis of a sponge-like HA structure and occluding of biological molecules (Schulten et af., 1991; Schulten and Schnitzer, 1993a; Schnitzer, 1994; Schulten, 1994) are confirmed by independent mathematical methods. For humic particles the vast number of different structural variations and the high capacity of trapping and binding of inorganic (minerals, water, gases), biological (carbohydrates, peptides, lignins, etc.) and anthropogenic substances is even more pronounced as shown in Fig. 22C (color plate 2). The model of a HA pentamer was constructed using the HyperChem software as described previously (Schulten, 1995a). Four HA subunits (clockwise from the lower right side to the upper center) were connected by covalent bonds by three water eliminations. The fifth HA unit (upper right) is linked to this HA tetramer by two hydrogen bridges which were formed during the geometrical optimization (and thus energy minimization) process and stayed intact until the convergence limit [> 4.19 kJ (0.1 nm) mol-I] was reached. At this stage of the single point calculations using the algorithm of Pollak-Ribiere, the humic complex had a total energy of 10,606.447 ld (0.1 nm)-l mol-I and a convergence gradient of 3.769 kJ (0.1 nm) mol-1. The humic complex obtained by molecular modeling in Fig. 22C has the elemental composition of C,s,6Hl,,10,,N,s, the elemental analysis of 67.12% C, 6.22% H, 25.42% 0, and 1.24% N and the corresponding molecular mass of 28,202.31 g mol-1. The number of 3790 atoms results from the fact that seven f signs in the humic subunit of Fig. 21 have been completed by adding seven CH, groups and that the CH,OH group at the benzene ring (right, below in Fig. 21) has been added in one case for HA-HA bonding by water elimination (for details see Schulten, 1995a). The HA pentamer has approximately a width of 11.4 nm, height of 10nm, depth of 7.52 nm and has voids with diagonal diameters up to 11.3 nm. It is clear that

Color plate 1 (Figure 22b of Chapter 4)

Color plate 2 (Figure 22c of Chapter 4)

ANALYSIS OF SOIL ORGANIC MATTER

199

large molecules can be surrounded and trapped by this organic model structure. At this stage, however, the inorganic structure such as minerals, water, and charged particles (e.g., metal cations; anions; zwitter ions; etc.) have not been considered as the modeling calculation “in vacuo” and “no charges” allowed. Thus, the even more complex problem of organo-mineral bonds has to be tackled. For soil organic matter, a novel three-dimensional structural concept of organomineral complexes has been proposed (Schulten, 1995b), which is based on the HA structure described earlier and comprehensive investigations combining geochemical, wet chemical, biochemical, spectroscopic, agricultural, and ecological data with analytical pyrolysis. Direct, temperature-programmed pyrolysis in the ion source of the mass spectrometer and soft ionization (Py-FIMS) and PyG U M S were the main applied thermal methods. Emphasis was put on molecular modeling and geometry optimization of silica complexes attached to soil organic matter using modem PC software (HyperChem). As a first example of simulation experiments for soil processes with biological substances such as carbohydrates and peptides, trapping and binding of a trisaccharide, hexapeptide, and biocide (atrazine) in an organomineral complex were performed. In this manner, formation and decomposition of whole soil particles could be studied at the atomic level (nanochemistry)using exact data for bond lengths, angles, torsions, van der Waals interactions, and hydrogen bonds.

VI. ANALYSIS OF SOIL ORGANIC MATTER BY Py-FIMS In most agricultural soils, inorganic and organic soil constituents are so closely associated that it is necessary to separate the two before either can be examined in greater detail. This separation is usually achieved by extracting the SOM.The soil science literature contains a vast amount of information on the extraction of OM from soils by variety of reagents under widely differing experimental conditions (Kononova, 1966; Schnitzer and Khan, 1978; Stevenson, 1982). Serious difficulties with extracting SOM and then partitioning it into HA, FA, and humin are that these are laborious and time-consuming procedures that are not suitable for the analysis of large numbers of soil samples. A new approach is required to deal with these problems. A suitable method for the analysis of SOM is Py-FI mass spectrometry. Py-FI mass spectrometry has been used for the characterization and identification of major components of a variety of biomaterials. Especially noteworthy are applications to the study of soil humic substances (Schulten, 1987; Post et al., 1988; Bracewell et al., 1989; Zech et al., 1990; Schulten et al., 1991), natural waxes (Schulten et al., 1986, 1987b), lignins (Haider and Schulten, 1985), and aliphatics in clays and humic substances (Schnitzer and Schulten, 1989; Schulten and Schnitzer, 1990).

M.SCHNITZER AND H.-R. SCHULTEN

200

k PY-FUIS OF ARMADALE HUMIC ACID, FuLvrC ACID, HUMIN,AND WHOLESOIL 1. Armadale HA

The Py-FI mass spectrum of this HA (Fig. 23a) shows the presence in this material of four major components: carbohydrates, phenols, lignins, and n-fatty acids. Especially noteworthy is the prominence of the n-C2, (mlz 368), n-C26 (mlz 396), n-C2, (mlz 410), n-C,, (mlz 424). and n-C30 (mlz 452) fatty acids. The whole range of n-fatty acids extends from c,6 to C3,. Other components present in smaller amounts are monomeric lignins, the n-CIoto n-Czo diesters, and the n-C, to nX50 alkyl monoesters, of which the n-C,, monoester (mlz 662) is the most abundant. Relatively weak signals characteristic of N components are mlz 59, 79, 81, 93, 117, 131, and 167.

2. Armadale FA The Py-FI mass spectrum of the FA (Fig. 23b) is dominated by carbohydrates and phenols, followed by lignins. The most intense signals are rnlz 58 (acetone)

3r

1O O ~ '

424

80 60 40 20 a, 0

60

c 10011

100

200

300

400

500

600

700

200

300

400

500

600

700

I110

0 -0

C 3

A3

Q a, [r

100

m / z

_____L)

Flpre 23. Py-H mass spectra of (a) the Armadale HA and (b) the Armadale FA [from Schnitzer and Schulten (1992) with permission of the publisher].

ANALYSIS OF SOIL ORGANIC MATTER

201

and mlz 60 (acetic acid). Both compounds are thermally eliminated by methyl ethyl ketones, carbohydrates, and fatty acids at temperatures >30O0C. In addition, smaller amounts of n-fatty acids (mlz 256, 284,312,368, and 382), sterols (mlz 414), n-alkyl diesters, and monomeric and dimeric lignins are also present in the FA. No intense signals due to N-containing compounds can be detected.

3. Armadale Humin This spectrum (Fig. 24a) shows the strong presence in this material of carbohydrates, phenols, monomeric and dimeric lignins, alkyl-benzenes, and alkyl esters. The presence of a homologous series of n-fatty acids ranging from n-C,, to n-C,, is indicated. Of special interest is the series of n-alkylbenzenes with signals at mlz 316,330,344,358,372,386,400,414, and 428, which appear to alkylbenzenes, respecindicate the presence of C6H,.C,7H3, to C,H,C,,H,, tively. Molecular ions at mlz 206 and 220 could be due to di- and trimethylphenanthrene. Intense signals probably due to the n-C,, to n-C,, alkyl diesters are observed from mlz 202 to 342. Except for weak signals at mlz 67 (pyrrole) and mlz 8 1 (methylpyrrole), no signals due to N-containing compounds appear in this spectrum.

100

200

300

400

500

m / z lilgure 24. Py-n mass spectra of (a) the Armadale hurnin and (b) the Armadale soil [from Schnitzer and Schulten (1992) with permission of the publisher].

2 02

M. SCHNITZER AND H.-R. SCHULTEN 4. Annadale Soil

This Py-FI mass spectrum (Fig. 24b) is dominated by the presence of carbohydrates, phenols, monomeric and dimeric lignins, and alkyl esters. Molecular ions m/z394 and 408 appear to be due to n-czg and n-C,, alkanes, whereas the weak signals at m / z 442, 456, and 470 arise from the presence of C6H,C26H,, to C6H,.C2gH,, n-alkylbenzenes, respectively. It is noteworthy that this whole soil contains suberin-derived aromatic esters at m/z446,474, 502, and 530 (Hempfling et al., 1991). The signals at m/z 170 and 184 arise from tri- and tetramethylnaphthalene, respectively, while m/z 178, 192, 206, 220, and 224 are due to phenanthrene and methyl-, dimethyl-, trirnethyl-, and tetramethylphenanthrene, respectively. Similar to the Armadale humin, n-C,, to n-clg alkyl diesters are also present in the Py-FI mass spectrum of the whole Armadale soil. The Occurrence of N-containing compounds in the soil is indicated by m / z 59, 67, 79, 81, 93, 103, 117, 131, and 167. From the summary of the compounds identified in Table IV, it appears that, in the whole soil and in the humic fractions obtained from it, carbohydrates, phenols, lignin monomers and dimers, and, to a lesser extent, n-fatty acids are the major components. Minor components include n-alkyl mono- and diesters, n-alkylbenzenes, methylnaphthalenes, methylphenanthrenes, and N-containing compounds. Similar compounds are present in all materials, except that the HA tends to be enriched in n-fatty acids and humin in n-alkylbenzenes. The data in Table IV suggest that, from the analytical point of view, the most

Table IV Compounds Identified in the Initial Armadale Soil and in the HA, FA, and Humh Fractions Isolated fkom It ~

Compound identified Carbohydrates Phenols Lingin monomers Lignin dimers n-Fatty acids n-Alkyl monoesters n-Alkyl diesters n-Alkyl benzenes . Methylnaphthalenes Methy lphenanthrenes N compounds n-Alkanes

Soil ++a

++ ++ ++ +

++ + + + + +

HA

FA

Humin

++ ++ ++ ++ +++ +

++ ++ ++ ++ + + +

++ ++ ++ ++ ++ ++ ++ + + +

+

+, weak (relative intensity <20%); ++, intense (relative intensity 2040%); + + + , very intense (relative intensity >a%).

ANALYSIS OF SOIL ORGANIC MATTER

203

suitable material to be analyzed is the whole soil. The Py-FI mass spectrometry of the whole soil produces more identifiable compounds than that of any of the fractions and, at the same time, obviates the need for laborious and damaging extraction, separation, and purification procedures. Py-FIMS is possibly the first and only method currently available that allows soil chemists and other interested scientists to perform comprehensiveSOM analyses at the molecular level on airdry soils without any pretreatment. The assignments of the major signals in the presented mass spectra were made according to the identified compounds described in Table V.These identifications are based (among other methods) on determinationsof thermal properties (Schulten, 1987; Leinweber et al., 1992; Leinweber and Schulten, 1992), timeresolved Py-FIMS (Schulten and Schnitzer, 1993b), accurate mass measurements (Hempfling and Schulten, 1990), Curie-point gas chromatographyhass spectrometry (Hempfling and Schulten, 1991; Schulten and Schnitzer, 1992). extensive National Institute of Science and Technology and Wiley Library searches, and pyrolysis-mass spectrometric investigations of model polymers.

B. TIME-RESOLVED Py-FIMS With the aid of Py-FIMSat a heating rate of 10 K min-* and a temperature range of 5O-75O0C, approximately 40 magnetic scans of the supercritical carbon Table V Identification of Mqjor Signals in the Py-FI Mass Spectra mlz

60, 72, 82, 84, 96, 98, 110, 112, 114, 126, 132, 144, 162

94. 108, 110, 122, 124, 126, i38, 140, 154 124, 138, 140, 150, 152, 154, 164, 166, 168, 178, 180, 182, 194, 196, 208, 210, 212, 222, 236, 246, 260, 270, 272, 274, 284, 296, 298, 300, 310, 312, 314, 316, 326, 328, 330. 340, 342, 356 170, 184, 198, 202, 216, 230. 244, 254, 256, 258, 268, 270, 272, 284, 286, 298, 300, 312, 314, 326, 328, 340, 342, 354. 368, 380, 382, 394. 396, 408, 410, 422, 424, 438. 452, 466, 480, 494, 508, 648,662, 676. 704, 732 92, 106. 120, 134, 142. 148, 156, 162, 170, 176, 184, 190, 192, 198, 204, 206, 218, 220, 232, 234, 246, 260, 274, 288, 302, 316, 330, 344, 358, 372, 386 59, 67, 79, 81, 95, 103, 109, 111, 123, 125, 137, 139, 153, 161, 167, 181, 183, 195, 203, 233, 245, 255, 257, 271, 285, 333, 359, 363, 393

Identification Carbohydrates with pentose and hexose subunits Phenols Lignins

Lipids (alkanes, alkenes, fatty acids, dioic acids, and n-alkyl esters) Alky laromatics

N compounds

M. SCHNITZER AND H.-R. SCHULTEN

204

dioxide extract of an agricultural soil (Bainsville, Haplaquoll) can be recorded, and thus 40 mass spectra are produced. As shown near the bottom of Fig. 25, the total ion intensity (TII, in counts X lO3), with a maximum near 145°C of 1.6 X 106 counts, yields a stretched out curve between 50 and 450°C (Schulten and Schnitzer, 1991). The complete evolution profile of volatile and ionized substances can be divided into 23 single ion mass chromatograms with a mass 1 1

1 21 27

14 7 8

n

12 9

14 14

F

x

12

Y)

c

13

C

3

7

:

7

-

8

v

14

'-

26

c

45

0

*

25

C

-

14

7 3 1600 50

-

100 150 200 250 300 350 400 450 500 550 Temperature [

OC

]

FLpre 25. Mass chromatograms (abscissa, microoven (sample) temperature; ordinate, ion intensity) recorded by temperature-resolved Py-FIMS of the supercritical carbon dioxide extract of an agricultural soil (Haplaquoll). The total ion intensity ("II) and signal series of 28 mass units between mlz 368 and 984 are given [fmm Schulten and Schnitzer (1991) with permission of the publisher].

ANALYSIS OF SOIL ORGANIC MATTER

205

difference of 28 Da. The sequence of these FI signals for the Bainsville supercritical carbon dioxide extract ranging from mlz 368 to 984 illustrates how the different chemical constituents of this complex mixture are transferred into the gas phase, ionized by the high electric field, and detected. It is assumed that the signals mlz 368-564 represent the c24-c38 n-fatty acids. With rising temperature, the n-alkyl monoesters follow with mlz 592 (C,) to 816 (&). Finally, thermal ester cleavage above 350°C leads to cutin- or suberin-derived biomarkers (esters of o-hydroxy keto acids) between mlz 844 and 984. Because of the elemental composition (CaH132O2 = 957.0225 Da; C68Hl3& = 985.0560 Da), these ions appear at uneven masses in the Py-FI mass spectrum. As expected, the single ion chromatograms generally shift to higher evolution temperatures in line with the mass and polarity of the detected chemical species. Inspection of single ion chromatograms of homologues mlz 268-596 indicates the presence of two different classes of compounds. Although the chromatograms run almost parallel in temperature between 160 and 250"C, as displayed in Fig. 25 for the C24-C38saturated fatty acids, the low temperature section (50160°C) is much more pronounced and gives the first hint that material of high volatility (e.g., alkanes) could be present. One possibility to solve the identification problem is high muss resolurion. However, since only nominal masses are involved so far, the contributions of compounds with different elemental compositions and structures have not been considered. Solutions for these problems are high resolution and accurate mass measurements. For example, at temperatures between 50 and lOO"C, the ion with nominal mass 408 can be identified as the C2,H, alkane as it evolves from the extract mixture (measured, 408.4649 Da; calculated, 408.4695 Da). As result of Py-FIMS of the supercritical carbon dioxide extract of the agricultural soil, the presence of series of alkanes and unsaturated fatty acids can be confirmed (Schulten and Schnitzer, 1991). With Py-FDMS, the signals at mlz 550 (C37H&), 578 (C39H&), and 606 (C4,H,,O,) indicate thermally produced, olefinic subunits of n-alkyl monoesters. Furthermore, as shown in Fig. 26, the interpretation of in-source FIlFD MS investigations is well supported by integrating sections ("temperature windows") of the evolved thermal products. This step, which is calledfracfionuf volutilizution, yields seven distinct FI mass spectra for the Bainsville extract in temperature ranges between 50 and 550°C. The top spectrum (50-175°C) shows little except intense signals of the C,, octyl ester and, due to protonation, free hexacosanedioic acid. From 175 to 230"C, intense FI signals of saturated fatty acids are observed, followed by alkyl esters and thermal products of o-hydroxy keto acids (up to 420°C). Visual inspection of the shapes of the chromatograms (also called pyrograms) of the ion series and their shifts with rising temperature assists with correct assignments of the ion species. The FI mass spectrum integrated between 420 and 550°C provides no additional structural information on the

M. SCHNITZER AND H.-R. SCHULTEN

206

'"]

50

- 175 'C

100!112230

-

'i'

255 ' C 596

50:

.

185

648 676 620

C

100

200

300

400

500

600

700

800

900

1000

m/z FIpre 26. Fractional volatilization of the supercritical carbon dioxide extract of the agricultural soil (see Fig. 25) at seven temperature ranges of the temperature-resolved evolution steps is illustrated [from Schulten and Schnitzer (1991) with permission of the publisher].

extract. This observation is very helpful, however, for future designs of Py-FIMS parameters and for allowing one to focus the sensitivity and efficiency of the method on each individual class of compounds: extracts, whole soils, soil particle size and density fractions, or single soil particles (Schulten, 1993). The Py-Fl mass spectra of the Ap horizon (0-5 cm) of a Haplaquoll extracted

ANALYSIS OF SOIL ORGANJC MATTER

207

for 2 hr with distilled water at a constant pressure of 12.2 MPa and at temperatures of (a) 150, (b) 200, and (c) 250°C are shown in Fig. 27 (Schnitzer et al., 1991b). Here the question was whether Py-FIMS could be used to give information on the molecular-chemical composition of extracts corresponding to the temperatures of the extraction experiment. In general, the method showed that the extracts contained polysaccharides, n-fatty acids, n-alkanes, n-alcohols, sterols, N-containing compounds, and mono- and dilignins, all typical components of soil organic matter. Temperature resolution was achieved by analyzing soil extracts of different temperature intervals and demonstrating the selectivity of the proposed method. Intense mass signals in Fig. 27a,b at mlz 74, 84, 96, 98, 110, 112, 114, 126, and 162 were mainly due to thermal degradation products of polysaccharides. These intensities clearly dropped at the higher extraction temperature of 250°C. In contrast, only the 150°C water extraction (Fig. 27a) gave a series of C,,-C, alkenes between m / z 532 and 616. On the other hand, the PyFI mass spectrum of the 250°C extraction indicated the presence of an intense series of C,,-C,, n-fatty acids by signals at m / z 340-508.

806040-

e406

II

167

546 574

398

I 300

532

I

602 ,616

b)

n

6

50

100

150

200

250

300

350

400

450 500

550

650

c>

1 a, 100

E

600

80 60

368

40

20

50

100

150

200

250

300

m / z

350

400

450 500

550

600 650

>

Flgure 57. Pyrolysis-field ionization mass spectra of water extracts obtained from an agricultural soil at 17.2 MPa and (a) 150, (b) 200, and (c) 250°C [from Schnitzer eta!. (1991b) with permission of the publisher].

208

M. SCHNITZER AND H.-R. SCHULTEN

VII. EFFECTS OF MINERALS ON THE Py-FIMS OF FULMC ACID In previous investigations of FA-mineral interactions, the focus was on following changes in the minerals rather than changes in FA (Schnitzer and Kodama, 1966, 1967, 1969; Kodama and Schnitzer, 1969, 1971). This type of approach was the only option because of the lack of suitable methods for the analysis of FA. With recent developments in Py-FIMS, however, and its applica. tion to SOM and whole soils, it has become possible to investigate the effects of soil minerals on the thermal evolution of FA by this method. Schnitzer et af. (1994) examined FA-mineral interactions in physical mixtures and in chemical complexes by Py-FIMS. For the physical mixtures with FA, three minerals were selected: (a) quartz, a nonphyllosilicate; (b) sodium montmorillonite, an expandable phyllosilicate; and (c) kaolinite, a nonexpandable phyllosilicate. Chemical complexes were prepared from FA and kaolinite and from FA and sodium montmorillonite. To obtain insight into the chemical interactions between FA and the minerals in the physical mixtures and complexes, ion intensities (11) generated by molecular ions in the mass spectra from each major group of FA components (Table V) were plotted against pyrolysis temperatures. These plots, referred to as thermograms, were computed for each mixture and complex and are shown in Figs. 28-

Figure 28. Thermal profiles of six selected classes of compounds in FA [from Schnitzer et al. (1994) with permission of the publisher].

ANALYSIS OF SOIL ORGANIC MATTER

209

ul

b

150 2

Y 0)

100 2

'I20

3 0

-

10

100

50 200 300 Temperature

500

400

[

OC

]

600

700

+ -

Flgure 29. Thermal profiles of six major groups of components in mixtures of FA with (a) quartz, (b) montmorillonite. and (c) kaolinite [from Schnitzer et al. (1994) with permission of the publisher].

30. If we assume that the total area under each curve is related to the concentration of the particular group of compounds, then the plots in Figs. 28 and 29a exhibit the same order of evolution of components, namely, phenols > carbohydrates > alkylaromatics > lipids > lignins > N-containing compounds for both FA and FA-quartz mixture. This order, however, changes in Fig. 29b,c (the FAmontmorillonite and FA-kaolinite mixtures) and also in Fig. 30a,b (the FAmontmorillonite and FA-kaolinite complexes) in that alkylaromatics become very prominent along with carbohydrates and phenols. Thus, the order of component evolution changes with the type of mineral that is present and the type of association that is formed between the FA and the mineral. When the thermogram for FA (Fig. 28) is compared with the other thermograms, estimates can be made of how much of each component group is volatilized and how much is retained by the minerals. Comparisons of Fig. 29a-c with Fig. 28 (FA) show delays in the evolution of some FA components (even by quartz) but no significant retention. But when Fig. 30a,b is compared with Fig. 28, it becomes apparent that major FA components are retained by the clays.

M. SCHNITZER AND H.-R. SCHULTEN

2 10

:

I

20 n 0)

10 E

E 0

v)

40

(r

E

\

30 Y

X

20

2 3 0

10

100

200

300 Temperature

400

[

OC

500

]

600

2 F

700

Flgure 30. Thermal profiles of six major components in (a) FA-montmorillonite complex and (b) FA-kaolinite complex [from Schnitzer er al. (1994) with permission of the publisher].

Thus, while no significant retention of major FA components is detected by the physical mixtures, selective retention of some FA components by the clays in the FA-clay complexes is observed. The thermograms confirm close association of major FA components with clays in the chemical complexes. Sodium montmorillonite (Fig. 30a) retains lipids, lignins, and N-containing compounds more firmly than does kaolinite (Fig. 30b). On the other hand, kaolinite exhibits a greater affinity for carbohydrates and possibly phenols than does sodium montmorillonite. These experiments demonstrate the considerable potential of Py-FIMS for enhancing our knowledge and understanding of the effects of minerals on the thermal behavior of FA and on the nature of the organomineral complexes formed.

VIII. OTHER APPLICATIONS Applications of Py-FIMS to the characterization of SOM in different particle size fractions separated from agricultural surface soils (Leinweber and Schulten, 1992; Schulten et al., 1993) showed decreased abundances of phenols, lignin

ANALYSIS OF SOIL ORGANIC MATTER

211

monomers, alkylaromatics, carbohydrates, and N-containing compounds, but increasing concentrations of lignin dimers and lipids with increasing particle size (from clay to sand). In a Spadosol Bh horizon, from coarse silt to fine silt, increasing relative abundances of carbohydrates, phenols, lignin monomers, and lipids were observed, whereas from fine silt to clay, these biomarkers increased. Lignin dimers, alkylaromatics, and N-containing compounds do not show a general trend of relative abundances, depending on particle size. For the sand fraction, higher relative abundances of carbohydrates, phenols, and lignin monomers and lower relative abundances of dimeric lignin units and N-containing compounds were observed compared with the finer fractions (Sorge et al., 1994). Another investigation still in progress is concerned with the composition of soil nitrogen, with special emphasis on the nitrogen compounds that so far have not been identified. More than 50 N-containing compounds have been identified by Curie-point pyrolysis-gas chromatography mass spectrometry in four mineral soils. These compounds include pyrroles, imidazoles, pyrazoles, pyridines, pyrimidines, pyrazines, indoles, quinolines, N derivatives of benzene, alkyl nitriles, and aliphatic amines. N derivatives of benzenes and long-chain alkyl nitriles appear to be soil specific (Schulten et al., 1995a). Other applications of F'y-FIMS deal with the chemical composition of SOM in macro- and microaggregates separated from soils under different crop rotations, fertilization (Leinweber et al., 1993), the characterization of interlayer clayorganic complexes in an acid soil, the relation between SOM composition and soil quality, and effects of long-term cultivation on the chemical structure of SOM (Schulten and Leinweber, 1993; Leinweber et al., 1994; Schulten et al., 1995b).

M.CONCLUSIONS Both pyrolysis-soft ionization mass spectrometry (field ionization and field desorption pyrolysis-mass spectrometry) and 13C NMR demonstrate that aliphatic structures in HAS are often as important as, and at times even more important than, aromatic structures. This contrasts with the earlier view held by SOM chemists that the chemical structure of HAS was predominantly aromatic. The major components of all organic extracts from humic substances, soil clays, and whole soils are n-alkanes, n-fatty acids, n-alkyl mono-, di-, and triesters, diols, and sterols, while minor components are n-alcohols, dioic acids, and n-alkenes. Both pyrolysis-soft ionization mass spectrometric methods compare favorably with gas chromatographicprocedures. The latter, however, limits separations to compounds of molecular weight lower than about 550 mass units (Ogner and Schnitzer, 1970). By contrast, the highest-molecular-weight n-alkane

212

M. SCHNITZER AND H.-R. SCHULTEN

P N N

H

!Scheme I. Proposed structures of soil organic nitrogen constituentsas derived by Rash pyrolysis (Curie-point) GUMS. The displayed structures give a qualitative survey of the different classes of N-containing compounds and their contribution to total nitrogen (Nt).

detected by Py-FDMS in the hexane and chloroform extract of the soil clay is 1416 Da. Thus, we are witnessing enormous advances in the separation and identification of major SOM components. Another significant point is that pyrolysis-soft ionization mass spectrometry is especially effective for the detection and identification of high-molecular-weight alkanes, fatty acids, sterols, and esters, which are difficult to analyze by any other method. As far as the analysis of solid-state samples is concerned, Py-FIMS is the only analytical procedure currently available that allows soil chemists and other interested scientists to perform comprehensive SOM analyses at the molecular level on air-dried soils without any pretreatment. This obviates the need for laborious, time-consuming, and possibly damaging extraction, separation, and purification procedures and constitutes a major advance in SOM chemistry. Curie-point Py-GUMS is a valuable, complementary method for structural studies on humic materials. The Curie-point pyrolyzer performs flash pyrolysis, i.e., the transfer of thermal energy to the sample is fast. With the aid of this method and in combination with Py-FIMS, a novel structural concept for HAS was proposed in which alkylaromatics such as alkylbenzenes, alkylnaphthalenes,

ANALYSIS OF SOIL ORGANIC MATTER

213

and alkylphenanthrenes form the basic HA skeleton. This method is also very useful for the identification of the many soil nitrogen compounds that so far have not been identified and that constitute at least 50% of the weight of the total soil nitrogen. Recently, progress on molecular structures of N-containing soil constituents has been achieved and is expected to shed some light on the problem of unknown nitrogen. The most interesting results are illustrated in Scheme I (Schulten et al., 1995a). So far a variety of interactions of SOM with minerals and organics (herbicides and pesticides) have been studied by Py-FIMS [see, for example, Khan et al. (1993)l. One can look forward to many other applications for solving problems of immediate concern and for better understanding of the quality and role of SOM in soils. Finally, the high costs of purchasing the required equipment for pyrolysis-soft ionization mass spectrometry could be reduced by setting up national or regional laboratories that could process large numbers of samples by working on a 24-hr basis.

ACKNOWLEDGMENTS This work was supported by the Centre for Land and Biological Resources, Research Branch, Agriculture Canada, Ottawa (M.S.), the Deutsche Forschungsgemeinschaft (projects Schu 416/3, 416/18- l), the Ministerium fur Wissenschaft and Forschung, Bonn-Bad Godesberg, the Umweltbundesamt , Berlin, and the Institut Fresenius, Chemical and Biological Laboratories, Taunusstein. Germany (H.R.S.). The authors are very grateful to their technical co-workers and research associates who are cited in the list of references for their excellent contributions. We thank C. Sorge, Fachhochschule Fresenius, Wiesbaden, R. Miiller, Institut Fresenius, Taunusstein, and, in particular, Dr. habil. agr. P. Leinweber, Institute of Structural Analysis and Planning in Areas of Intensive Agriculture, Vechta, Germany, for close and constructive collaborations.

REFERENCES Abbt-Braun, G., Frimmel, F. H., and Schulten, H.-R. 1989. Structural investigations of aquatic humic substances by pyrolysis-field ionization mass spectrometry and pyrolysis-gas chromatographylmass spectrometry. Water Rex 23, 1579- 1581. Arshad, M. A., Ripmeester, J. A,, and Schnitzer, M. 1988. Attempts to improve solid-state 13C NMR spectra of whole mineral soils. Can. J. Soil Sci. 68, 593-602. Beckey, H. D. 1977. “Principles of Field Ionization and Field &sorption Mass Spectrometry.” Pergamon Press, Oxford, UK. Bracewell, 1. M., Haider, K., Larter, S. R.,and Schulten, H.-R. 1989.Thermal degradation relevant to structural studies of humic substances. In “Humic Substances 11. Search of Shucture” (M. H. B. Hayes, P. MacCarthy, R. L. Malcolm, and R. S. Swift, Eds.), pp. 181-222. Wiley, New York. de Leeuw, J. W., de Leer, E. W., Sinnighe, E.,Damstk, J. S., and Schuyl, P. J. W. 1986.Screening

2 14

M. S C W T Z E R AND H.-R. SCHULTEN

of anthpogenic compounds in polluted sediments and soils by flash evaporation/pyrolysisgas chmmatographylmassspectrometry. Anal. Chem. 58, 1852-1857. Ghosh, K., and Schnitzer, M. 1980. Macromolecular structures of humic substances. Soil Sci. 29, 266-276. Giacobbo, H., and Simon, W. 1964. Methodik zur Pyrolyse und auschliessenden gaschromatographischenAnalyse von Probenmengen unter einem Mikrogramm. Pharmacol. Acra Helv. 29, 162-167. Haider, K., and Schulten, H.-R. 1985. Pyrolysis-field ionization mass spectrometry of lignins, soil humic compounds, and whole soil. J . Anal. Appl. fyml. 8, 317-331. Hansen, E. H., and Schnitzer, M. 1966. The alkaline permanganate oxidation of Danish illuvial organic matter. Soil Sci. Soc. Am. Proc. 30, 745-748. Hansen, E. H., and Schnitzer, M. 1969. Zinc dust distillation and fusion of a soil humic and fulvic acid. Soil Sci. Soc. Am. Proc. 33, 75-81. Hempfling, R., and Schulten, H.-R. 1988. Characterization and dynamics of organic compounds in forest humus studied by pyrolysis-gas chromatography/electron ionization mass spectrometry and pyrolysis (high resolution) field ionization mass spectrometry. J. Anal. Appl. Pyrol. 13, 3 19-325. Hempfling, R., and Schulten, H.-R. 1990. Chemical characterization of organic matter in forest soils by Curie-point-pyrolysis-GC/MS and pyrolysis-field ionization mass spectrometry. Org. Geochem. 15, 131-145. Hempfling, R., and Schulten, H.-R. 1991. Pyrolysis-gas chromatography/mass spectrometry of agricultural soils and their humic fractions. Z . Pfinrenernaehr. Bodenk. 154, 425-430. Hempfling, R., Simmleit, N., and Schulten, H.-R. 1991. Characterization and chemodynamics of plant constituents during maturation, senescence and humus genesis in spruce ecosystems. Biogeochernistry 13,27-60. Irwin, W. J. 1982. “Analytical Pyrolysis.” Dekker, New York. Khan, S . U., and Sowden, F. J. 1971. Distribution of nitrogen in the black solonetzic and black chernozemic soils of Alberta. Can. J. Soil Sci. 51, 185-193. Khan, S. U., Schnitzer, M., and Schulten, H.-R. 1993. Fate of deltamethrin after nine years of incubation in an organic soil under laboratory conditions. J. Agric. Food Chem. 41, 11431151. Kodama, H., and Schnitzer, M. 1%9. Chemical characteristics of a fulvic acid rnontmorillonite complex. I n “Proc.Clay Conf., Tokyo, Vol.I” (L. Heller, Ed.), pp. 765-774. Israel University Press, Jerusalem. Kodama, H., and Schnitzer, M. 1971. Evidence for interlamellar adsorption of organic matter by clay in a podzol soil. Can. J . Soil Sci. 51, 509-512. Kononova, M. M. 1966. “Soil Organic Matter.” Pergamon, Elmsford, NY. Leinweber, P., and Schulten, H.-R. 1992. Differential thermal analysis, thermogravimetry,and insource pyrolysis-mass spectrometry studies on the formation of soil organic matter. Thermochim. Acra u)o, 151-167. Leinweber, P., and Schulten, H.-R. 1995. Composition, stability and turnover of soil organic matter. investigations by off-line pyrolysis and direct pyrolysis-mass spectrometry. J. Anal. Appl. Pyml. 32, in press. Leinweber, P., Schulten, H.-R., and Horte, P. 1992. Differential thermal analysis, thermogravimetry and pyrolysis-field ionization mass spectrometry of organic matter in particle-size fractions and bulk soil samples. Thermochim. Acta 194, 175-187. Leinweber, P., Reuter, G., and Schulten, H.-R. 1993. Investigations of organomineral clay fractions from long-term fertilization experiments in East Germany. Appl. Clay Sci. 8, 295-31 1. Leinweber, P., Schulten, H.-R., and Wrschens, M. 1994. Seasonal variations of soil organic matter in a long-term agricultural experiment. Planr Soil 160, 225-235.

ANALYSIS OF SOIL ORGANIC MATTER

215

Lowe. L. E. 1978. Carbohydrates in soils. I n “Soil Organic Matter” (M. Schnitzer and S. U. Khan, Eds.), pp. 65-94. Elsevier, Amsterdam. Meuzelaar, H. L. C., Haverkamp, J., and Hileman, F. D. 1982. “Pyrolysis-Mass Spectrometry of Recent and Fossil Biomaterials.” Elsevier, Amsterdam. Murray, K. E.. and Schulten, H.-R. 1981. Field desorption mass spectrometry of lipids I. The application of field desorption mass spectrometry to the investigation of natural waxes. Chem. Phys. Lipids 29, 1 I -2 I . Nagar. B. R. 1963. Examination of the structure of soil humic acids by Py-GC. Nature 199, 12131214. Ogner, G.. and Schnitzer, M. 1970. The Occurrence of alkanes in fulvic acid, a soil humic fraction. Geochim. Cosmochim. Acta 34, 392-3%. Post, B., Hempfling, R., Klamberg, H., and Schulten, H.-R. 1988. Zur Charakterisierung von Boden-Huminstoffen. Fresenius Z. Anal. Chem. 331, 273-281. Rollins, K., Scrivens, J. H., Taylor, M. J., and Major, H. 1990. The characterization of polystyrene oligomers by field desorption mass spectrometry.Rapid Commun. Mass Specmm. 4,355-359. Saiz-Jimenez, C., and de Leeuw, J. W. 1984. Pyrolysis-gas chromatography/massspectrometry of soil polysaccharides, soil fulvic acids and polymaleic acid. Org. Geochem. 6 , 287-293. Saiz-Jimenez, C., and de Leeuw, J. W. 1986. Chemical characterization of soil organic matter fractions by analytical pyrolysis-gas chromatography/massspectrometry.J. Anal. Appl. Pyml. 9, 99-1 19. Schnitzer, M. 1978. Humic substances: chemistry and reactions. I n “Soil Organic Matter” (M. Schnitzer and S. U. Khan, Eds.), pp. 1-64. Elsevier, Amsterdam. Schnitzer, M. 1991. Soil organic matter-the next 75 years. Soil Sci. 151, 41-58. Schnitzer, M. 1994. A chemical structure for humic acid. Chemical, 13C NMR. colloid chemical, and electron microscopic evidence. I n “Humic Substances in the Global Environment and Implications in Human Health” (N. Senesi and T. M. Miano, Eds.), pp. 57-69. Elsevier, Amsterdam. Schnitzer, M., and Khan, S. U. 1972. “Humic Substances in the Environment.’’ Dekker, New York. Schnitzer, M., and Khan, S. U. 1978. “Soil Organic Matter.” Elsevier, Amsterdam. Schnitzer, M., and Kodama, H. 1966. Effect of pH on adsorption of soil humic compound by montmorillonite. Science 153, 70-71. Schnitzer, M., and Kodama, H. 1967. Reaction between a podzol fulvic acid and sodiummontmorillonite. Soil Sci. SOC.Am. Pmc. 31, 632-636. Schnitzer, M., and Kodama, H. 1969. Reaction between fulvic acid, a soil humic compound and montmorillonite. Isr. J. Chem. 7, 141-147. Schnitzer, M., and Preston, C. M. 1987. Supercritical gas extraction of a soil with solvents of increasing polarities. Soil Sci. Soc. Am. J. 51, 639-646. Schnitzer, M., and Schulten, H.-R. 1989. Pyrolysis-field ionization mass spectrometry of aliphatics extracted from a soil clay, and humic substances. Sci. Total Envimn. 81/82, 19-30. Schnitzer, M., and Schulten, H.-R. 1992. The analysis of soil organic matter by pyrolysis-field ionization mass spectrometry. Soil Sci. Soc.Am. J. 56, 1811-1817. Schnitzer, M., Hindle, D. A., and Meglic, M. 1986. Supercritical gas extraction of alkanes and alkanoic acids from soils and humic materials. Soil Sci. SOC.Am. J. 50, 913-919. Schnitzer, M. Tamocai, C., Schuppli, P.,and Schulten, H.-R. 1990a. Nature of the organic matter in tertiary paleosols in the Canadian arctic. Soil Sci. 149, 257-267. Schnitzer, M., Tamocai, C., Schuppli, P., Hempfling, R., and Schulten, H.-R. 1990b. Palecenvironmental indicators in Eocene paleosols from Arctic Canada. Fresenius 1.Anal. Chem. 337,882884. Schnitzer, M., Kodama, H., and Ripmeester, J. A. 1991a. Determinationof the aromaticity of humic substances by X-ray diffraction analysis. Soil Sci. SOC.Am. J. 55, 745-750.

216

M.SCHNITZER AND H.-R. SCHULTEN

Schnitzer, M., Schulten, H.-R., Schuppli, P., and Angers, D. A. 1991b. Extractionof organic matter from soils with water at high pressures and temperatures. Soil Sci. SOC. Am. J. 55, 102-108. Schnitzer, M., Kodama, H., and Schulten, H.-R. 1994. Mineral effects on the pyrolysis-field ionization mass spectrometry of fulvic acid. Soil Sci. Soc.Am. J. 58, 1100-1 107. Schulten, H.-R. 1977. Pyrolysis field ionization and field desorption mass spectrometry of biomacromolecules. microorganisms, and tissue materials. In “Analytical Pyrolysis” (C. E. R. Jones and C. E. Cramer, Eds.), pp. 17-28. Elsevier, Amsterdam. Schulten, H.-R. 1979. Biochemical, medical, and environmental applications of field ionization and field desorption mass Spectrometry. Int. J. Mass Spectmm. Ion Phys. 32,97-283. Schulten, H.-R. 1980. Advances in field desorption mass spectrometry.I n “Soft lonjzation Biological Mass Spectrometry’’ (H. R. Moms, Ed.), pp. 6-38. Heyden, London. Schulten, H.-R. 1987. Pyrolysis and soti ionization mass spectrometry of aquaticlterrestrial humic substances and soils. J. Anal. Appl. Pyml. 12, 149-186. Schulten, H.-R. 1993. Analytical pyrolysis of humic substances and soils: Geochemical, agricultural and ecological consequences. J. Anal. Appl. Pyml. 25, 97-122. Schulten, H.-R. 1994. A chemical structure for humic acid. Pyrolysis-gas chromatography/mass spectrometry and pyrolysis-soft ionization mass spectrometry evidence. In “Humic Substances in the Global Environment and Implications in Human Health” (N.Senesi and T. M. Miano, Eds.), pp. 43-56. Elsevier, Amsterdam. Schulten, H.-R. 1995a. The three-dimensional structure of humic substances and soil organic matter studied by computational analytical chemistry. Fresenius J. Anal. Chem. 351, 62-73. Schulten, H.-R. 1995b. The three-dimensional structure of soil organo-mineral complexes studied by analytical pyrolysis. J. Anal. Appl. Pyml. 32, in press. Schulten, H.-R., and Leiinweber. P. 1993. Influence of the inorganic matrix on the formation and molecular composition of soil organic matter in a long-term experiment. Biogeochemistry 22, 1-22. Schulten, H.-R., and Schnitzer, M. 1990. Aliphatics in soil organic matter in fine-clay fractions. Soil Sci. SOC.Am. J. 54, 98-105. Schulten, H.-R., and Schnitzer, M. 1991. Supercritical carbon dioxide extraction of long-chain aliphatics from two soils. Soil Sci. SOC. Am. J. 55, 1603- 161I . Schulten, H.-R., and Schnitzer, M. 1992. Structural studies of soil humic substances by Curie-point pyrolysis-gas chromatography/massSpectrometry. Soil Sci. 153, 205-224. Schulten, H.-R., and Schnitzer. M. 1993a. A state of the art structural concept for humic substances. Naturwissenschaften 80, 29-30. Schulten, H.-R., and Schnitzer, M. 1993b. Temperature-resolved pyrolysis-soft ionization mass spectrometry of soil humic acids. Org. Geochem. 20, 17-25. Schulten, H.-R., Simmleit. N., and Rump, H. H. 1986. Soft ionization mass spectrometry of epicuticular waxes isolated from coniferous needles. Chem. Phys. Lipids 41, 209-224. Schulten, H.-R., Simmleit, N.,and Miiller, R. 1987a. High-temperature, high-sensitivity pyrolysisfield ionization mass spectrometry.Anal. Chem. 59, 2903-2908. Schulten, H.-R., Murray, K. E., and Simmleit, N. 1987b. Natural waxes investigated by softionization mass spectrometry. 2. Naturforsch. 42C, 178- 190. Schulten, H.-R., Plage, B., and Schnitzer, M. 1991. A chemical structure for humic substances. Naturwissenschaften 78, 3 1 1-3 12. Schulten, H.-R., Leinweber, P., and Sorge, C. 1993. Composition of organic matter in particle-size fractions of an agricultural soil. J. Soil Sci. 44, 677-691. Schulten, H.-R., Sorge, C., and Schnitzer, M. 1995a. Structural studies on soil nitrogen by Curiepoint pyrolysis-gas chromatography/mass spectrometry with nitrogen-selectivedetection. Biol. Fertil. Soils, in press. Schulten, H.-R., Monreal, C. M.M., and Schnitzer, M. 1995b. Effect of long-term cultivation on the chemical structure of soil organic matter. Natunvissenschaften 82, 42-44.

ANALYSIS OF SOIL ORGANIC MATTER

217

Sorge, C., Miiller. R., Leinweber, P., and Schulten, H.-R. 1993. Pyrolysis-mass spectrometry of whole soils, soil particle-size fractions, litter materials and humic substances: statistical evaluation of sample weight, residue, volatilized matter and ion intensity. Fresenius J. Anal. Chem. 346, 697-703. Sorge, C., Schnitzer, M., Leinweber, P., and Schulten, H.-R. 1994. Molecular-chemical characterization of organic matter in whole soil and particle-size fractions of a Spodosol by pyrolysis-field ionization mass spectrometry. Soil Sci. 158, 189-203. Sowden, F. G., and Schnitzer, M. 1967. Nitrogen in illuvial organic matter. Can. J. Soil Sci. 47, 111-116.

Spiteller, M. 1985. Extraction of soil organic matter by supercritical fluids. Org. Geochem. 8, 1 I I 113.

Stevenson, F. J. 1982. “Humus Chemistry.” Wiley, New York. Stevenson, I. L., and Schnitzer, M. 1982. Transmission electron microscopy of extracted fulvic and humic acid. Soil Sci. 133, 179-185. Weast. R. C. 1979. “Handbook of Chemistry and Physics,” 59th ed. CRC Press, Boca Raton, n. Wilson, M. A. 1987. “NMR Techniques and Applications in Geochemistry and Soil Chemistry.” Pergamon Press, Oxford, UK. Zech, W., Hempfling, R., Haumeier, L., Schulten, H.-R., and Haider, K. 1990. Humification in subalpine Rendzinas: Chemical analyses, IR and 13C NMR spectroscopy, and pyrolysis-field ionization mass spectrometry. Geoderma 47, 123- 138.

This Page Intentionally Left Blank

ROLEOF METAL-ORGANIC COMPLEXATION INMETAL SORPTION BY SOILS Robert D. Harterl and Ravendra Naidd 'Department of Natural Resources University of New Hampshire, Durham, New Hampshire 03824 2CSIRO Division of Soils and CRC for Soil and Land Management, Glen Osmond, SA 5064,Australia

I. Introduction U. Organics in the Soil Solution III. Metals in the Soil Solution A. Metal-Water Interactions B. Metal-Ligand Interactions C. Stability Constants IV. Effect of Low-Molecular-Weight Organics on Metal Ion Reactions with Organic Surfaces V. Effect of Organics on Reactions of Metal Ions and Complexes with Inorganic Surfaces A. Iron Oxides B. Manganese Oxides C. Aluminum Oxides D. Silica Oxides E. Clays F. Soils VI. Environmental Implications A. Contaminant Transport B. Soil Genesis and Fertility C. Metal Toxicity VII. Summary and Research Needs References

I. INTRODUCTION Dissolved organic carbon (DOC)is an important component of soil solution. It plays a major role in numerous soil chemical processes in the ecosystem. Pro219 &ma in Agrvnmy, Vdumr f f Copyright 0 1995 by Academic Press,Inc. All rights of reproductionin any form reserved.

220

R. D. HARTER AND R. NAIDU

cesses such as mineral weathering (Thurman, 1985), metal leaching and toxicity (Ugolini et al., 1977; Dawson et al., 1978; Driscoll et al., 1980), acid-base chemistry in natural waters (Driscoll and Schecher, 1990), solubility control of metal ions in soils (Pohlman and McColl, 1988; Fox et al.. 1990; Tipping and Hurley, 1992; McBride, 1989), and dissolution and plant availability of nutrient ions such as P (Deb and DeDatta, 1967; Appelt et al., 1975; Traina et al., 1986; Fox et al., 1990; Bolan et al., 1994) are all influenced by the nature and amounts of dissolved organics in soil solution. The effects of dissolved organics on anion dissolution and mobilization have been well established (Hue, 1991; Fox et al., 1990). This review focuses on the nature of interaction between trace metals in soil solution, dissolved organics in soil solution, and solid surfaces. The interaction between metal cations and dissolved polyfunctional organic compounds of low molecular weight is important because of its role in mineral-weathering and soil-forming processes (Schalscha et al., 1967; Henderson and Duff, 1963; Pohlman and McColl, 1986) and its potential role in heavy metal contamination of soil and groundwater. The principles underlying the interactions between metal ions and organic compounds are well understood, and there have been many papers published on organometallic interactions involving extractable humic and fulvic acids and metal ions beginning with those of Bloomfield (1953a,b, 1954a,b,c, 1955) who first showed the probable role of fulvic acid in spodic horizon development. However, only limited information is available on interactions between low-molecular-weight organics and metal ions.

II. ORGANICS IN THE SOIL SOLUTION Dissolved organics that interact with soil constituents and trace metal ions are of two major kinds: (i) a range of low-molecular-weightorganic acids, including polyphenols, simple aliphatic acids, amino acids, sugar acids, and hydroxamate siderophores (Stevenson, 1967; Graustein et al., 1977, Lynch, 1978; Cromack et al., 1979; Fox and Comerford, 1990), and (ii) a series of soluble humic/fulvic acids. The low-molecular-weight organic acids of the first group are produced continuously in soils through microbial activity. These are also sensitive to microbial degradation, however, and are therefore short-lived in soil solution (Stevenson, 1967). Moreover, under natural nonsterile conditions many of the simple aliphatic acids will not diffuse far before being adsorbed or modified by the microflora. Consequently, their concentrations in soil solution can vary considerably with time. Since these acids are short-lived, it is often assumed that they are of little importance in soils. Presumably for this reason, there is a paucity of information on the role of low-molecular-weight organic acids in metal solubilization and their availability in soil-plant systems, although there are a

METAL-ORGANIC COMPLEXATION IN SOILS

22 1

number of papers on the influence of such organic acids on the dissolution and plant availability of nutrient phosphorus. The continual production of such acids within the rhizosphere and during the plant residue decomposition process does, however, make them and their conjugated anions chemically important (McColl and Pohlman, 1986; Fox and Comerford, 1990). The sources and amounts of soluble organics depend on the nature of ecosystems. High amounts would be expected to be present within the rhizosphere (Rovira and Davey, 1974) and around decomposing plant residues (Whitehead et al., 1981, 1983). where biological activity is intense. In a forest ecosystem, the sources of soluble organics are both the natural depositions in and decomposition of plant residues such as leaves, branches, and other organic debris, as well as organic substances derived from the decomposition of roots (Pohlman and McColl, 1988). Naidu er al. (1993) found that the thick decomposing litter layer under forest systems was accompanied by higher concentrations of DOC than those under pasture, with resultant enhanced Fe and A1 mobility. In general, the concentrationsof DOC beneath organic forest floors are higher than those underlying mineral soils (McDowell and Likens, 1988). In soils supporting legumes, the concentration of low-molecular-weightorganics should be high within the rhizosphere (Rovira, 1969; Rovira and Davey, 1974). These acids have been shown to play a major role in weathering processes, metal detoxification, and enhanced plant nutrient availability. Fox and Comerford ( 1990) determined the concentration and nature of low-molecularweight organic acids in selected forest soils of the southeastern United States and in the rhizosphere of pine seedlings, reporting that the suite of organic acids in the rhizosphere was more complex than that in the bulk soil. In the rhizosphere, high concentrations of both oxalate and formate were detected along with traces of citric, acetic, and aconitic acids. Farm management practices such as organic matter retention and application of manures and organic wastes can also enhance the concentrations of organic acids. Lynch (1978) studied the production of acetic acid in anaerobic soils containing plant residues, reporting phytotoxic levels of the dissolved acids. Bolan er al. (1994) compared the nature and amounts of soluble organic acids in soils, pine litter, and poultry manure, reporting 20-200 times higher organic acid concentrations in poultry manure than are present in soils. The nature and amounts of low-molecular-weight organic acids commonly reported in soil solutions, forest litter leachates, and water extracts of poultry manure are listed in Table I. A wide variety of low-molecular-weight acids, including oxalic, citric, formic, acetic, malic, succinic, malonic, maleic, lactic, aconitic, and fumaric acids, have been identified in the soil system (Stevenson, 1967; Huang and Schnitzer, 1986; Bolan et al., 1994). The concentration of the acids in cultivated soils is extremely low (Stevenson and Ardakani, 1972) compared to those in soils under pine forest systems (Muir er al., 1964; Naidu er al.,

Table I Natore and Amounts (pmol I-')of Low-Molecular-Weight Organic Acids Identified in Soil Solutions and Poultry Manure Vegetation Forest

Acetic acid

Aconitic acid

120-280

Forest Cultivated field Forest

5- 12

Formic acid

Fumaric acid

tr

0.33 72.33

tr

tr

tr

7.14 x 10-4

44.19

Lactic acid

Maleic acid

Malic acid

109-189

tr

90

12-52 74.5

101-137 3-6.2

tr

tr

8-8.4

Forest Pasture Pwlhy manure

Cihic acid

tr-

174 (7.75- 15) x 10-4 2.98

(6.3-18.4)

6.71

X

Malonic acid

20-74 4.2-8.0

Oxalic acid

3-22 0.75-16.5 60-1043

10-4

tr

Succinic acid

Reference

tr

Schwarte er 01.. 1954

125-282 10

Hue er 01.. 1986

tr

Foxand comerford. 1990 Grierson. 1992

x 10-3 0.028 22.52

0.15 3.54

Bolan ef 01.. 1994 16.76

METAL-ORGANIC COMPLEXATION IN SOILS

223

1994) and poultry manure extract (Bolan et al., 1994). In all systems, however, low-molecular-weight aliphatic acids usually comprise only a small fraction of the total organic acids present in soil solution. Generally, the concentration of individual low-molecular-weight organic acids may range from micromolar to millimolar concentrations (Table I). The nature and concentration of these acids determine the extent to which soil processes are affected. For example, acetic and formic acids encourage metal release through dissolution processes, while di- and tricarboxylic acids enhance dissolution through chelation. Root exudates consist of a variety of aliphatic acids, such as citric, oxalic, and tartaric acids, many of which are capable of forming complexes with metal ions (Rovira and Davey, 1974). The most abundant organic acids identified in tree root exudates include citric, fumaric, malic, malonic, and succinic acids (Smith, 1976). A wide range of low-molecular-weight aliphatic organic acids, including phenolic acids and aldehydes, occur in soils and are derived from decomposing plant residues (Katase, 1981; Whitehead et al., 1981, 1983). The quantities of chelating organics present in soil solution vary with time, and their concentrations presumably are the net result of the amounts produced and transformed through either microbial or other chemical processes to low-molecularweight products or polymeric humic substances. While the effect of these low-molecular-weight organic acids on phosphorus solubility has been the subject of considerable research, there is a paucity of information on metal chelation/trace element chemistry, particularly in soil systems.

III. METALS IN THE SOIL SOLUTION k MET~-WATER INTERACTIONS Metal ions undergo a series of reactions involving both the solid and solution phases. The concentration and the nature of the ions present in the soil system are a result of these interactions. Thus, the chemical composition of soil solution is dynamic (Fig. 1) and is determined by multiphase equilibria involving (a) the solid phase, i.e., clay minerals, poorly ordered inorganic minerals, and organic materials, (b) the liquid phase, comprising water and dissolved solids, (c) the gaseous phase, comprising mainly oxygen and carbon dioxide, and (d) the complex exchange phase. The interactions between metal ions and the solid phases involving clay minerals and poorly ordered inorganic phases have been the subject of numerous reviews, and readers are referred to books edited by Adriano (1986, 1992). This section therefore is limited to the chemistry of metal ions in soil solution and its practical implications for both plant uptake and the environment.

224

R. D. HARTER AND R. NAIDU Plants

Fertiiizers

Sludge Atmospheric deposition

F l

t

desorption-

immobilization

aqueous phase

-

-

mineralization

Oxides carbonates

Leaching

t

Ground water Flgure 1. Chemical composition of the soil solution.

Once introduced into the soil system, metal (Mn+) ions rapidly undergo a series of reactions in both the aqueous and solid phases, and the nature of these reactions depends on the configuration of outer electron orbitals, reflected in the position of the metals within the periodic table. Although there are about 80 metal ions in the periodic table, only 20-30 ions are capable of independent existence as aquometal ions or as partially hydrolyzed metal ions in aqueous solutions. Of these, only 12 or so are frequently encountered as essential metal ions in biological systems. In the case of electropositive metals such as those from group I, contact with the aqueous phase results in the formation of aquated metal ions:

Divalent and trivalent metal ions such as those from groups I1 (e.g., Ca2+) and I11 (e.g., AP+) and transition metal ions (e.g., Cu2+) can, however, undergo a series of hydrolytic reactions that can be written as M(HZO),"+ % [M(H,O)x-y(OH)y](m-y)++ yH+

(2)

Generally, the higher the positive charge on the metal, the more dissociated (acidic) the hydrogen atoms of the coordinated water molecules. However, the extent of hydrolysis is pH dependent and therefore influenced by dilution. The metal ions that hydrolyze range from hard metal ions, i.e., those that include the alkali and alkaline earth metal ions, to moderately soft metal ions such as Cd2+, Pb2+, and Hgz+. Since the product of the metal hydrolysis reaction is H+, metal ions are generally classed as Lewis acids (after G. N. Lewis). Pearson (1963) further classified metal ions into two groups, depending on whether they formed their

METAL-ORGANIC COMPLEXATION IN SOILS

225

most stable complexes with C, N, or 0, the first ligand atoms from groups V, VI, and VII (hard acids, class a), or whether they formed their most stable complexes with the second or a subsequent member of each group (soft acids, class b). This classification, named “the hard and soft acid base theory (HSAB)”, states that hard Lewis acids more readily react with hard Lewis bases and soft acids more readily react with soft bases. Hard acids are usually small in size, with high electropositivity, low polarizability, and do not contain unshared pairs of electrons in their valence shells, whereas soft acids are large in size, have lower electropositivity, have high polarizability, and do contain unshared pairs of electrons in their valence shells (Pearson, 1963). Trivalent metal ions such as Fe3+, AP+ , and Si4+, for example, are hard because of their high ionic charge, while the monovalent ions H+, Li+, Na+, and K+ are hard because of their low ionization potential and relatively small size. Hard bases have low polarizability and are difficult to oxidize. Soft bases, on the other hand, are readily polarizable and easy to oxidize. Soft acids form stable complexes with bases that are highly polarizable and good reducing agents. Table I1 lists the metal ions and inorganic ligand anions according to the hard and soft acid-base concept. Generally, the hard metal ions include the plant macronutrient and secondary nutrient metal cations (Ca, Mg), while the borderline acids include the micronutrient elements such as Mn(II), Fe(II), Co, and Zn (Table 11).

B. METAL-LIGANDINTERACTIONS In addition to the hydrolytic reactions discussed in the preceding section, the metal ions also react with inorganic and organic ligands commonly present in soil

Table I1 Some Hard and Soft Acids and Bases Hard H + , Li+, Na+, K+,Be,+, Mg2+, CaZ+,SrZ+, Sn2+, A13+, Sc3+. Ga3+, In3+, La3+. Cr3+, C O ~ +Fe3+, , As3+, I r 3 + , Si4+, Ti4+, Zr4+ Th4+ pU4+ H20, OH, 0, ROH, COO-, C03’-, NO3-, PO4>, SO4’-, CI04-, F-

Soft

Borderline

Acids Cu+, Ag+, Au+, TI+, Hg+, Cs+, Pb2+, Cd2+, R2+, Hg2+, Moo. Th3+

Fez+, Co2+, Ni2+, Cu2+, Zn2+, Pb2+

Bases H-, S”, SH, I-

CI-, NO,-,

Br-

226

R. D. HARTER AND R. NAIDU

solutions. Some of these ligands include the halides, group VB elements and organics such as those released in root exudates, decomposing organic matter, etc. Often these interactions lead to metal-ligand ion pairs and frequently to soluble metal-ligand complexes. Chelation also results in the presence of multidentate ligands. The nature of these interactions and the stability of the metalligand complex vary with the nature of both the ligand and the metal ion, the properties of the soil solution, and whether the ligands can compete with water in the primary hydration shell, which is present in a high and effectively constant concentration. Given the competing hydration [Eq. (3)] and metal-ligand formation [Eq.(4)]reactions,

* M(H20),"+6,,

(3)

+ L l i , % ML[r,r) + XHZO

(4)

MTZq) + xH2O M(H,O),"+e,)

the relative affinity of the metal cation for water and the ligand will determine whether complexation can occur. The extent to which an aquo cation combines with ligands to form complex ions is a thermodynamic problem and can be treated in terms of appropriate expressions for equilibrium constants and enthalpies of the competing reactions, AHhydration [Eq. (3)] and AHfomacion[@. (4)].Thermodynamically, the relationship is simple: If the energy of bonding to the ligand is greater than the hydration energy of one or more water molecules that must be replaced from the hydration shell, then the complex will form. If the hydration energy is greater, the waters will not be replaced and the complex will not form. The relationship, however, becomes more complex given that water itself can take two forms, depending on the proton activity (pH) of the solution. Often the stability of metal-ligand interactions can be deduced from the hard and soft acid-base behavior of metal and ligand ions through the Misono softness parameter (Misono er al., 1967), which is defined by the equation:

Y = 10IzR(Z)Iz+,

(5)

where R is the radius of a metal ion with valency 2 and whose ionization potential is I,. Those metal ions that have Y greater than 0.32 nm generally have low electronegativity and high polarizability; such metals tend to form covalent chemical bonds and are characterized as soft Lewis acids. Values of Y less than 0.25 nm correspond to metal ions that have high electronegativity and low polarizability; these metal ions are classified as hard acids and form ionic rather than covalent chemical bonds. Those metal ions that have Y values between 0.25 and 0.32 nm correspond to borderline cases. The organic ligand groups commonly encountered in natural soil systems are listed in Table III. On the basis of the Misono softness parameter and the hard and soft acid-base concept, it may be predicted that many interactions between transition metal ions and soft bases such as acetates, formates, etc. do not result

227

METAL-ORGANIC COMPLEXATION IN SOILS Table IIi Hardness Classification of Some Organics That May Be Encountered in Nature Decreasing hardness Enolate /O>C=C \

Phosphate 0

II / o -

RO- P

\O-

Ether R

II / o -

R- P

Amine R'

I

ooCarboxylate 0

II

Merceptide

R -S-

R-C-O-

\O-

Carbonyl R R,>=O

O \ R'/

R -N

Phosphorate 0

b

Phenoxide

Oxinate R R , 'C=N /

deprotonated amide 0

\o-

Aromatic amines N

I

R"

in significant formation of metal-ligand complexes. However, dicarboxylates and enolates, which are examples of hard bases, can form stable complexes with a number of transition hard metal ions. The sequence of decreasing hardness (Table 111) indicates decreasing affinity for hard metal ions. The softer, more polarizable donors toward the right of Table I11 have higher affinity for soft metal ions such as Cu*+ and Hg*+. Martell (1960, 1967, 1978) has published detailed reviews on the factors affecting the affinity between donor and acceptor atoms in the formation of coordinate bonds. He has also delineated the factors affecting the relationship between multidentate ligands and their affinity for various types of metal ions. Most stable complexes commonly encountered in natural systems can involve bidentate combinations of monodentate donor groups such as hydroxy acid anions (Table 111), hydroxamates, and catchecols. These donor groups are very effective complexing agents for hard metal ions like AP+ and Fe3+, and all are found among the natural siderophores produced in microbial systems

228

R. D. HARTER AND R. NAIDU

(Martell el al., 1988). Other examples include salicylic acid and oxalic acid. More complex examples include the polymeric “humic” and “fulvic” acids, which contain multiple ligand groups that are oriented sufficiently that they can simultaneously bind to a single metal ion. The Misono softness parameter has been related to the metal toxicity sequences and their tendency to form oxyanions. Sposito (1989) reports that soft metal cations, i.e., 1, < 0.28, are more toxic to plants than borderline metal cations (28 < 1, < 0.35), which are more toxic ( I , > 0.35) than hard metal cations. This is primarily because the “soft” metal cations lack the ability to form strong complexes with organic ligands.

C. STABILITY CONSTANTS The energy released during a metal-ligand interaction is defined as the stability constant of the reaction. The equilibrium constant of the corresponding metal complex in aqueous solution provides an indication of the affinity of the metal ions for ligands, as shown in Eqs. (6) and (7). For interactions involving bi- or polydentate ligands, the equilibrium constant is usually determined for each successive metal-ligand interaction, and the net stability constant is determined as the product of the successive K values. Allowing K, to symbolize the equilibratium constant, we can write

Ideally, determination of the equilibrium constant of the complexation process should include the chemical reaction involving protonated ligands. This is because most of the ligands are basic and therefore are usually protonated in aqueous solutions [Eq.(lo)] within physiological and environmental pH ranges (Martell et al., 1988).

H+ + H,-,L;l

* H,L,

(10)

Martell et al. (1988) discuss the details of the processes involved in the determination of stability constants of the complex reactions that involve protonated ligands. In soil solutions, a wide range of metals and ligands are present; this leads to a

METAL-ORGANIC COMPLEXATION IN SOILS

229

wide range of chemical reactions, which produce a range of metal-ligand complexes. Ligands that form soluble complexes with metal ions in solution may have a significant effect on the solubility and plant availability of metals in soils. The complexation reaction is particularly important in the case of both heavy metals and group 111 metals, such as Al, which have significant environmental and plant growth implications. Stability constant data are generally used in computer models (e.g., Minteq, Geochem) to predict the speciation of metal ions in the soil solution. However, the stability constant data commonly used in these models are limited by the absence of data on metal-humate or metal-fulvate complexes. This causes considerable constraint because humic acids from different sources can vary in molar mass and number of functional groups. Consequently, the nature and stability of the complex formed can vary widely. Moreover, pH will have a profound effect on the active functional groups through ionization of the carbonyl groups, thereby changing the number of binding sites. In addition, there is no provision for the range of metal-ligand complexes formed by interaction with low-molecular-weight organic acids.

IV. EFFECT OF LOW-MOLECULAR-WEIGHT ORGANICS ON METAL ION REACTIONS WITH ORGANIC SURFACES Soil organic matter is perhaps the single most important component for the retention of heavy metal ions. Organic complexes are particularly important for the retention of copper by soils under a variety of conditions (McLaren and Crawford, 1973; Petruzzelli et al., 1978; Tobia and Hanna, 1958; Singh, 1971; Borah et al., 1992; Holmgren et al., 1993; McLaren et al., 1981). Interactions with soil organic matter are, however, also very important in the retention of other metal ions (Petruzzelli et al., 1981; Keefer and Singh, 1986; Bergseth and Stuanes, 1976), and Basta et al. (1993) indicated that complexation is more important for retention than are organic matter exchange sites. It is therefore improbable that low-molecular-weightorganics will appreciably alter the mechanism of metal ion retention by the soil organic matter. Rather, competition for solution phase metal ions is apt to inhibit retention by organic matter. The extent to which such inhibition occurs will depend on the relative stability of the two metal-organic forms. Schnitzer and Skinner (1967) note that the fate of metal ions in the soil will often depend on the relative stability of the complex they form with soil organic matter. To this, one can add that whether an ion is retained by soils also depends on the relative stability of solution phase ligand-metal and soil organic mattermetal bonds. Thus, for example, Schnitzer and Skinner (1967) indicate that, at

Table IV Some Reported Conditional StabiUty Constants for Metal-Fulvic Acid and Metal-Humic Acid Compared to Stability Constants of Some Commonly Used Soil Extraction Ligands' Complexation ion (Log K) Ligand Hurnic acid

Fulvic acid

source of ligand

pH

29 soils

7

4 soils 3 soils soil

5 6 5

soil

5 7

4.42 7.65 3.66 6.52

Manure com- 5 7 post

3.72 4.02

Dhilon er al., 1975

Poultry litter

5 7

2.93 4.13

Prasad and Sinha, 1980

Commercial HA

6

31 soils

7

Soil

4

5.60, 3.95=

5

6.00,4.08 6.30,3.78

6

Cd(I1)

CO(I1)

Cu(I1)

8.1-9.1 6.2-6.8,4.9-5.6b

Fe(I1)

Mn(I1)

Ni(I1)

Pb(1I)

mu

Reference

4.20-10.33

Matsuda and Ito, 1970 Rosell er al., 1977 Taga er al., 1991 Dhilon et of., 1975

4.5-9.7

RasadandSinha, 1980

Taga et al., 1991

7.0, 5.6b

3.88-9.30

Matsuda and Ito, I970 Breenahan er al., 1978

Soil

Soil

5.7 6.7 7.7 3.5 5

2.02, 2.04, 3.01d 2.02, 3.00, 3.01 2.03, 3.00, 3.01

Brady and Pagenkopf, 1978 2.2 3.69

5.79 8.69

5.06 5.77

1.47 3.78

3.47 4.14

3.09 6.13

1.73 2.34

Schnitzer and Skinner, 1966, 1%7

4.88 7.51

Dhilon et al., 1975

I Soil

5.9 7

3.64 4.54

h a d and Sinha, 1980

Soil

3 5

2.2 3.6

Schnitzer, 1978

Soil

5.7 6.7 7.7

Soil

4 5 6 7 8

Soil

Lake water Water

5

8 4

4.7 5 6

2.8 4.1

3.3 4

2.1 3.7

3.1 4.2

2.6 4

5.3, 9.8, 14.0d 5.6. 10.6, 15.5 6.0, 10.7, 15.4

Brady and Pagenkopf, 1978

3.23 3.8 4.08 4.32 4.63

Saar and Weber,

1979

8.05-8.80 5.48, 4.WE 6.00, 3.86 5.95, 3.70 6.11, 3.86

5.14

5.14

Shuman, 1978 Breenahan et al., 1978

(continues)

Table IV (continued) Complexation ion (log K ) Ligand

source of ligand

pH

Water

4 5 6 7 8

Supemumine

7

Garden peat

8

Farmyard manure

5

W N

Cd(I1)

Col(I1)

Cu(I1)

Fe(I1) Mn(I1)

Ni(I1)

Pb(I1)

=(In

3.15 3.48 3.68 3.91 4.08

Reference Saar and Weber, 1979

N

5.33

Matsuda and Ito,

4.83

Shuman. 1979

4.8

Pandeya, 1993

1970

8.5

7.16-8.51

4.325.64 12 12.8

5.1

Farmyard manure

3.5 5

Poultry manure

5 8.5

14 15.2

Pandeya, 1993

Sewage sludge

5

11.5

Pandeya, 1993

4.68 6.98

Aggarwal and Sastry, 1993

Acetic acid Oxalic acid Citric acid M-EDTA M-DTPA N

w W

Sewage sludge

5 8.5

Sugar industry waste

5

3.04, 2.27b

3.88,2.116

Sposito er al., 1981

4.22,2.62b 13 13.8 14.7

8.5 1.93

1.46

2.22

3.89

4.12

5.36

4.83f

20.56 20.27

17.26 20.42

Pandeya, 1993

1.4

1.43

2.68

1.57

6.23

3.95

5.16

4.91

4.87

5.908

3.7of

5.11f

19.7 22.65

1.4

15.27 14.81 17.67 16.78

19.52 21.44

4.7of 18.88 19.93

17.44 19.56

Martell and Smith, I977 Martell and Smith, 1977 Martell and Smith, 1977 Lindsay, 1979 Lindsay, 1979

a The ionic strength under which the data were obtained varies from author to author, but is generally in the range 0.01-0.1. Stability constants of the extractant ions were obtained at zero ionic strength unless otherwise noted. K, and K2 from Scatchard plots; possibly same as footnote c. Values are for bonding with two and one phenolic carboxylate group(s). respectively. Values for Cd-FA, Cd(II)-FA, and Cd(II1)-FA, respectively. A soil conditioner. f I = 0.16. 8 I = 0.1,T = 20.

234

R. D. HARTER AND R. NAIDU

pH 5.0, logK for Pb- and Ni-fulvic acid complexes are 6.13 and 4.14, respectively (Table IV), while log K of Pb- and Ni-oxalate complexes are 5.16 and 4.91, respectively (Martell and Smith, 1977). Thus, on the basis of the values of Schnitzer and Skinner (1967), it is expected that oxalate can displace Ni but not Pb from fulvic acid. On the other hand, ligands such as EDTA and DTPA are capable of extracting all listed ions from organic bonding sites. This is, of course, an oversimplification, and even “weak” ligands such as acetate and oxalate are capable of removing a certain amount of metal ions from soil organic bonding sites. Stability constants for metal-organic matter complexes can be obtained by titrating a metal solution with a solution of the organic matter, and the stability constant, Ki, is obtained from the relationship

where [Mf] is the concentration of the free metal ion, [La] is the concentration of ligand in the acid form, and [ML] is the metal-ligand concentration (Mantoura and Riley, 1975; Bresnahan et al., 1978). Assigning the v to the ratio of metalcomplexed to undissociated ligand,

Scatchard plots (Fig. 2) are obtained by plotting u/[Mf] as a function of u. The slope of the plot is the reciprocal of the stability constant, and the x-axis intercept indicates the average number of binding sites per molecule, ni. Strengths of

-v Flgure 2. Illustration of a Scatchard plot. Y is the ratio of metal-complexed ligand to undissociated ligand molecules ((ML]/[L,]), [M,] is the concentration of free metal ions, Ki is the stability constant, and ni is the average number of binding sites per ligand molecule.

METAL-ORGANIC COMPLEXATION IN SOILS

235

a

Flpre 3. ‘‘Strong’’four-coordinate (a) and “weak” bidentate (b) Cu bonding sites on fulvic acid molecules, as suggested by Bresnahan er al. (1978).

metal-organic bonds are variable, depending on the functional group, the type of bond formed, and the coordinating ability of the metal ion. As a result, Scatchard plots of metal bonding to soil organic materials are typically bimodal, indicating two distinct bond energies, and can sometimes be better described as curvilinear, indicating a continuum of energy sites. Taga et al. (199 1) concluded that strong Cu bonding sites on humic acid were largely carboxyl groups and that the weak sites were amino groups and phenolic hydroxyls. Bresnahan et al. (1978) noted, however, that bonding sites tended to be carboxylate, phenolate, and carbonyl. They felt that the “strong” fulvic acid Cu bonding sites, which are evident at low solution Cu activity, were the result of four-coordinate bonding of the Cu ions (Fig. 3a). As the amount of Cu in the system is increased, they indicated that the quadridentate site is resolved into two weaker bidentate sites (Fig. 3b). While Bresnahan et al. (1978) used two salicylic acid groups as their model, they explained that any combination of phthalic acid and salicylic acid groups would be equally justified. It is obvious that

236

R. D. HARTER AND R. NAIDU

metal ions bonded into the two different configurations could have substantially different stability constants and so have quite different susceptibility to complexation by solution phase ligands. By the same token, whether soil organic matter fixes metal ions complexed with organic ligands in solution can depend on the extent of quadrivalent site saturation.

V. EFFECT OF ORGANICS ON REACTIONS OF METAL IONS AND COMPLEXES WITH INORGANIC SURFACES In addition to reactions with organic matter, sorption by iron oxides (Johnson, 1986; Benjamin and Leckie, 1981; Kabata-Pendias, 1980; Okazaki et al., 1986; Stahl and James, 1991a; Manceau etal., 1993) and manganese oxides (Shuman, 1988; Traina and Doner, 1985; Kabata-Pendias, 1980; Manceau et al., 1993; Burau, 1973; Zasoski and Burau, 1988; Fu et al., 1991; Stahl and James, 1991b) has been identified as the major mechanism by which metal ions are retained in the soil. Retention on clay exchange sites (Schlichting and Elgala, 1975; Basta et al., 1993; Kabata-Pendias, 1980) and aluminum oxides (Okazaki et al., 1986; Cavallaro and McBride, 1984; Schulthess and Huang, 1990), while a valid mechanism, appears to be less important under most conditions. These observations support Jenne’s (1968) conclusion that iron and manganese oxides provide the primary control of heavy metal concentrations in soil and water. The presence of organic molecules or ions in the system can, however, alter metal sorption characteristics. Huang and Lin (1981) noted that three outcomes are possible: (a) metal ion sorption may be inhibited due to complexation of the metal ion or competition at the surface for sorption sites; (b) metal ion retention may be enhanced if the ligand both forms a strong complex with the ion and has a strong affinity for the surface; and (c) there may be no perceptible change if the ligand only weakly complexes the metal and has a weak affinity for the surface. The nature of the effect will greatly depend on the mechanism and location of both the metal and ligand bonding to the surface. For example, a metal ion that is bonded by inner-sphere complexation (specific sorption) will be less susceptible to ligand effects than one bonded by outer-sphere complexation (nonspecific sorption, generally ion exchange). To a significant extent, both the effect of metal sorption by inorganic surfaces and the effect of organics on sorption can be predicted on the basis of HSAB theory. Inorganic surfaces with high permanent charge tend to be hard, while those having high variable charge tend to be soft in nature (Xu and Harsh, 1990a,b). Thus, both soft metals and soft organics should have greater affinity for the variable charge surfaces, while hard metals and hard organics should have affinity for high charge clays such as vermiculite. Soft metals have an affinity for soft organics and hard metals have an affinity for hard organics.

METAL-ORGANIC COMPLEXATION IN SOILS

237

Therefore, the presence of polarizable (soft) organics in solution should enhance the sorption of soft metals by variable charge surfaces. On the other hand, the presence of nonpolarizable (hard) organics in solution should retard the sorption of hard metals by variable charge surfaces and enhance the sorption by permanent charge surfaces. By using this argument, competition for variable charge surface sites should be expected primarily between soft organics and hard metals, and competition for permanent charge surface sites should be between hard organics and soft metals. It is therefore obvious that an understanding of the effect of organic ligands on the reactions of metals with soil requires an understanding of metal reactions with both the organic and the solid surfaces present in the system. Establishment of the role of organic ligands in metal retention furthermore requires an understanding of the chemical nature of both the metal ion and the organic, as well as the effect of the ligands on both the reactions and the solid surfaces.

A. IRONOXIDES Soils having similar iron contents can differ appreciably in their ability to retain heavy metals (Harter, 1979), depending on the nature of the iron oxide that is present in the soil. Iron oxide reactive sites occur primarily at deprotonated surface hydroxyls. The higher oxides such as hematite (a-Fe,O,) have two-thirds of their octahedral sites filled (Schwertmann and Cornell, 1991). which means that their surface structures will contain relatively few hydroxyl sorption sites. On the other hand, hydrous minerals such as goethite (a-Fe00H) have only about one-half of their octahedral sites filled (Schwertmann and Cornell, 1991). A larger fraction of the goethite surface will, therefore, contain hydroxyls, providing relatively more metal reaction sites. This difference in density of surface hydroxyls can be readily seen in infrared spectra of the iron oxides. Due to the widespread Occurrence of goethite in nature, metal reactions with both natural and synthetic goethite have been studied extensively. b o n e r (1993) has reported that the Cu and Pb sorption edges (the pH at which sorption increases rapidly) of goethite occur between pH 4 and 7, with the Zn edge Occurring about 1.5 pH units higher. He found no effect of ionic strength on sorption, indicating inner-sphere complexation. Mehadi ( 1993) similarly reported no effect of ionic strength on Ni retention by a synthetic goethite due to inner-sphere complexation. Kalbasi er al. (1978) reported zinc to be both specifically and nonspecifically sorbed at iron oxide surfaces. About 60-90% of the zinc sorbed was accompanied by the release of two hydrogen ions, and the authors proposed an olation bridge structure, with each zinc ion being attached to two hydroxyls at the iron oxide surface (Fig. 4a). The remaining 10-40% of the zinc sorbed was accom-

238

R. D.HARTER AND R. NAIDU

Elpre 4. Potential inner-sphere (specific) (a) and outer-sphere (nonspecific) (b) metal bonding sites at the surface of iron oxides (Kalbasi et al., 1978).

panied by the release of one hydrogen ion and was thought to be associated with a counterion at the surface (Fig. 4b). By using X-ray absorption fine structure (EXAFS),Manceau et al. (1993) found that Pb bonded to both goethite and fresh hydrous iron oxide was surrounded by 2 subshell 0 at 0.222 and 0.242 pm, 1 Fe at 0.32 pm, and 0.4 Fe at 0.34 pm. This arrangement indicates that the Pb ion was sharing an iron octahedral edge. According to Schwertmann and Cornell (1991), such sites are available only at steps of the goethite (1 10) surface. Manceau et al. (1993) also noted that hydrous iron oxide has a higher Pb sorption potential than does goethite, which is in agreement with the greater number of iron octahedral edge sites in the less crystalline material. While there is relatively little information on the role organic molecules play in metal ion sorption by iron oxides, there has been enough investigation to establish some general principles. Retention of organics by iron oxide surfaces can decrease the zero point of charge (ZPC)of oxide minerals, as has been shown to occur with inorganic anions such as phosphate (Kuo and McNeal, 1984). For example, EDTA is adsorbed by iron oxide surfaces at low pH and has been shown to make the surface more negative (Rueda et al., 1985). Simple organic acids such as lactic, tartaric, citric, and oxalic acids can be adsorbed by iron oxides, with sorption increasing as solution pH decreases (Schwertmann and Taylor, 1989). Phenolics will coordinate directly with iron oxide surfaces (McBride, 1987), possibly at sites similar to those bonding phosphate. The o-diphenols, having adjacent hydroxyl groups, appear to have a particular affinity for the iron oxide surface and are readily oxidized to form larger humic acid-like polymers (McBride, 1987; Shindo, 1992). Not all low-molecular-weight organics will be similarly polymerized, but complexation of either simple or polymeric organics will alter the nature of the surface, in most cases increasing the net

METAL-ORGANIC COMPLEXATION IN SOILS

239

surface negative (or decreasing positive) charge density. On the other hand, considering the structure of many aromatics, competition of the organic molecule for metal bonding sites at iron octahedral edges is also possible. Zeltner et al. (1986) have noted the retention of salicylate by goethite and have suggested that the ions displace two surface hydroxyls from single iron atoms. These sites are relatively easily deprotonated and therefore are probable sites for heavy metal retention. Several researchers have evaluated the effect of organics on metal ion sorption by iron oxides, and, depending on the system pH and the nature of the ligand, both increases and decreases in metal sorption have been reported. As noted later in this chapter, Chairidchai and Ritchie (1992) have reported a relationship between the zero point of charge and metal retention by a soil high in hydrous iron oxides. The presence of certain ligands tended to enhance sorption when the pH was below the ZPC and to reduce sorption at pH levels above the ZPC. In evaluating the effect of several organics on the retention of Cu by amorphous iron oxide surfaces, Davis and Leckie (1978) reported no effect of either salicylic acid or protocatechuic acid (PCCA) on Cu retention, even though both organics were sorbed below pH 7.0. They further reported a slight enhancement in Cu sorption in the presence of 2,3-pyrazinedicarboxylicacid (2,3-PDCA), a significant increase in Cu sorption in the presence of glutamic acid, and almost total exclusion of Cu from the surface in the presence of picolinic acid. It is apparent that neither salicylic acid nor PCCA competes with Cu for sorption sites, nor do they form complexes of any type. In both picolinic acid and 2,3PDCA, carboxyl groups are adjacent to ring N’s (Fig. 5 ) ; picolinic acid has one carboxyl/ring N pair and 2,3-PDCA has two. Both organics effectively complex Cu, but the difference in response rules out metal bridging between the organic and the iron oxide surface. In the absence of Cu, and apparently in the presence as well, the molecule attaches to the iron oxide surface both through the carboxylate group and the N lone pair electrons. Since this is also the site of Cu complexation, the only picolinic acid molecules capable of complexing Cu are those remaining in solution, and Cu sorption by the iron oxide is excluded. 2,3-

a

C O O H ( 0 1

COOH HOOC -(CH2)2-

COOH

I COOH

Plcollnlc acld

2,SPDCA

NH~+

Glutamlc acld

Schematic molecular structures of picolinic, 2,3-pyrazinedicarboxylic(2.3-PDCA). and glutamic acids.

Rgure 5.

240

R. D. HARTER AND R. NAIDU

PDCA adsorbed by the surface still has a carboxyl/ring N pair directed toward the solution and capeble of Cu complexation. Thus, 2,3-PDCA enhances Cu sorption by the iron oxide surface, while picolinic acid retards Cu sorption. Glutamic acid, a nonaromatic molecule, also contains two carboxylate groups, one of which is adjacent to an amine group, and this molecule also enhances Cu retention by the surface. The degree of organic complexation of an ion also can affect sorption. For example, in evaluating methylated tin sorption by synthetic iron oxide in the presence of fulvic acid, Donard and Weber (1985) noted that nearly all (88-98%) of the MeSnC1, was removed from solution, while about half (28-66%) of the Me2SnC1, and one-quarter (15-28%) of the Me,SnCl were removed from solution. Since all three methyltin species should be neutral, Donard and Weber (1985) attributed these observations to differing polarities; monomethyltin has the highest polarity and trimethyltin the lowest. They felt that retention reactions occurred primarily with fulvic acid on the iron oxide surface. In contrast, when studying the response of butylated tin, Randall and Weber (1986) observed the least removal from solution for Bu2SnC1, (0-56%), and only slightly more BuSnC1, (72-100%) than Bu,SnCl (57-95%) was removed from solution. They felt that the butyltin species similarly reacted with fulvic acid coatings on the iron oxide, rather than the iron oxide surfaces themselves, and that nonpolar forces were important to the retention of butyltin. The hydrophobicity of alkyltin compounds increases with the increased number of carbon atoms, so that the mixed results apparently resulted from tributyltin sorption being favored by hydrophobicity and monobutyltin sorption being favored by polarity. It is possible that the sorption of organic reductants on iron oxide surfaces could create sorption sites on the oxide surface by reduction of the iron. McBride (1987), however, found only trace amounts of Fe(I1) in solution and rapid uptake of O2 when reductants such as hydroxyquinone were complexed at the iron oxide surface. It would therefore appear that the instability of reduced iron in an oxidized system would mitigate against reduction as a mechanism for creating metal ion sorption sites on the iron oxide surface.

B. MANGANESE OXIDES Hydrous manganese oxides appear to be an important source of heavy metal ion sorption sites in soils (Shuman, 1988), and Fu et al. (1991) noted that the manganese oxides sorb more Cd than do hydrous iron oxides. Zasoski and Burau (1988) proposed high-energy and low-energy Cd and Zn sorption sites, with the high-energy sites showing greater Cd selectivity. Depending on the type of manganese mineral present and the pH, exchange sorption can be an important

METAL-ORGANIC COMPLEXATION IN SOILS

241

mechanism (Stahl and James, 1991b). In addition, McKenzie (1972) found that certain metal ions, particularly Co, are capable of displacing Mn from manganese oxide crystal surfaces due to higher crystal field stabilization energies. It appears, however, that a more complex interaction of oxidation-reduction and sorption-replacement is involved in the retention of most ions at manganese oxide surfaces. For example, Traina and Doner (1985) demonstrated that Mn(I1) ions appear in solution when Co(II) is added to manganese oxides, indicating that the Co(l1) is reducing Mn(1V) at the crystal surface. When Cu was used as the sorbate, they were unable to confirm the sorption of Cu(I1) and a similar increase of Mn(I1) in solution unless Mn(II) was present on the surface. A similar reaction seems to occur when Pb(I1) is placed in the presence of manganese oxides (Dillard et al., 1981), with the Pb(I1) being oxidized to Pb(IV) at the surface. Wakatsuki et al. (1993) reported inner-sphere complexation of Cr on manganese oxides. They also indicate that the retention mechanism is reduction of Mn(IV) to Mn(I1) by oxidation of Cr(II1) to Cr(V1) and replacement of the Mn(I1) in the crystal structure by Cr(II1). In fact, since Mn(I1) has zero crystal field stabilization energy (McKenzie, 1972), if a reductant capable of reducing the crystal Mn(1V) is present, any metal ion can replace Mn(I1) from manganese oxide surfaces. This leads to a potentially important role of organics in the sorption of metal ions by manganese oxides. McBride (1987) reported that phenolic compounds, many of which are strong reducing agents, reacted vigorously with Mn oxides, resulting in a system with high oxygen demand. Furthermore, Bartlett (1990) demonstrated that in the presence of manganese oxides tannic acids are converted to hurnin, and Shindo (1992) verified humic acid synthesis from several phenolic compounds in the presence of manganese oxides. This suggests, therefore, that the presence of low-molecular-weightorganics should encourage metal ion sorption by manganese oxides. In the process of oxidation and polymerization of the organic compound, Mn(1V) will be reduced to Mn(I1) at the surface, and an ion retention site is thereby produced. Stone and Morgan (1984b) tested the reactivity of 15 aromatics and 12 aliphatics toward manganese oxide and found that catechols, hydroquinones, methoxyphenols, and resorcinols among the aromatics, as well as ascorbate, oxalate, and pyruvate, increased Mn(I1) in the solution. Stone and Morgan (1984a) further demonstrated that the reductant ion must be complexed onto the surface before electron transfer can occur. Thus, not only will a potential retention site be created within the manganese oxide surface but the oxidized organic molecule or polymerization product may provide cation exchange or metal complexation sites as well. We were unable, however, to find any work evaluating metal retention by manganese oxides in the presence of organic molecules or ions and, thus, cannot verify the accuracy of this predicted effect.

242

R. D. I-IARTER AND R. NAIDU

The surface structure of aluminum oxides consists of hydroxyl ions bonded to two and one aluminum ion(s) in a ratio of approximately 2:l (Hsu, 1989). The singly bonded hydroxyls can be protonated or deprotonated, depending on the solution hydrogen ion activity. The reactivity of the surface depends on the nature of the crystal. For example, gibbsite [y-Al(OH),] tends to form as hexagonal plates, whereas bayerite [a-Al(OH),] frequently takes a pyramidal form (Hsu, 1989). Thus, while the primary reaction sites of gibbsite are predominantly located at the crystal edges, the latter will tend to form more steps on the surface, which, when deprotonated, will provide a greater density of surface reaction sites. Boehmite (y-A100H) also appears to be present in some soils. It tends not to form well-developed crystals (Hsu, 1989), and its more amorphous nature means that it will contain even more active sorption sites. From electron spin resonance (ESR),studies McBride (1985b) concluded that noncrystalline aluminum hydroxide has at least two Cu complexation sites. He suggests that these are discrete surface sites and are evident only at low Cu loading. At high Cu loading, chemisorption was replaced by nucleation or precipitation of hydroxy Cu at the surface. Gibbsite sorbs substantially less Cu (about 1OX less; Shuman, 1977) than does noncrystalline aluminum hydroxide. The Cu sorbed by gibbsite is primarily oriented with the z-axis perpendicular to the (001) plane of the mineral (McBride et al., 1984), indicating that sorption occurred on steps of the planar surface. These observations led McBride (1982) to suggest that the two chemisorption sites were similar, occumng at hydroxyls bonded to a single A1 ion, but that the Cu ion could react with either one or two of the sites. Such a mechanism is consistent with the observed Cu sorption capacity of noncrystalline alumina > boehmite > gibbsite, since with increased crystallinity the 0 ion is associated with two, rather than one, A1 ions. Kalbasi et al. (1978) has reported that the Zn reaction with aluminum oxide surfaces is similar to that proposed for iron oxides (Fig. 5), which is also consistent with the proposed Cu reaction mechanism. Likewise, Vordonis et al. (1992) noted that Co2+ and Ni2+ were sorbed via inner-sphere complexation at these same sites. It is, in fact, probable that the mononuclear AlOH at crystal edges serves as the retention site for most, if not all, heavy metals. The effect of organics on metal sorption by the aluminum hydroxides therefore depends in part on whether the organics are bonded similarly to the surface. McBride and Wesselink (1988) reported catechol to be strongly and selectively sorbed by aluminum hydroxide surfaces and suggested that the dominant sorption mechanism was the formation of a bidentate complex with AlOH groups on crystal edges. Sorption of salicylic acid, benzoic acid, and phthalic acid also occurs at these sites (Kummert and Stumm, 1980), as does sorption of oxalate (Parfitt et al., 1977). This means that low-molecular-weight organics can com-

METAL-ORGANIC COMPLEXATION IN SOILS

243

pete with metals for surface sites and will likely decrease metal ion retention unless metal-organic complexes are formed. High-molecular-weight organics such as humic acid and fulvic acid are also known to be strongly sorbed to aluminum oxide surfaces (Schulthess and Huang, 1991), even at very high pH levels. These molecules are more apt to be H-bonded to the surface, thereby providing increased metal sorption sites. Girvin et al. (1993) reported that both Co(I1) and Co(I1)-EDTA were readily sorbed by 6-Al,03 but, as expected, the former only at pH levels above about 7.0 and the latter at pH levels below 7.0. The response to variations in ionic strength indicated the formation of outer-sphere complexation of the Co(I1)-EDTA on the 6-A1203. They suggested that sorption of the Co(I1)-EDTA occurred at single AlOH sites. Since sorption of Co(I1) and Co(I1)-EDTA occurred at different pH levels, however, there would be no surface competition. Rather, at high pH the EDTA should retard sorption by keeping the Co in solution. McBride (1985a) noted that the presence of glycine inhibits Cu sorption by gibbsite and boehmite at pH > 5, a phenomenon they attributed to the reduction of Cu hydrolysis. They found that complexed copper-glycine was sorbed at crystal steps on gibbsite and suggested a metal bridging structure as illustrated in Fig. 6. Two sorbed species were detected on the boehmite surface. One was as illustrated in Fig. 6, and the other had a second glycine molecule replacing the water molecule attached to the Cu ion. Water washing removed the second glycine, and a large excess of glycine caused desorption of Cu. Elliott and Huang (1979) reported enhanced sorption of Cu by y-A1203when in the presence of chelating agents such as NTA, glycine, and aspartic acid, at least part of which was attributed to specific sorption of negatively charged metal-organic complexes. Conversely, Chubin and Street (1981) reported decreased Cd in the presence of EDTA at all pH levels and a 25% reduction in Cd sorption in the presence of acid citrate systems. Inhibition in acid environments is apt to indicate competition for sorption sites.

Flgure 6. Schematic representation of Cu bridging between glycine and gibbsite surfaces (McBride, 1985a).

244

R. D. HARTER AND R. NAIDU

D. SILICAOXIDES The ability of silica oxide surfaces to retain heavy metals is well-known and is the reason dilute heavy metal solutions must be stored in either plastic or acidified containers. Metal retention by the silica oxide surface occurs because partially deprotonated orthosilicate groups on silanol surfaces can act as either monodentate or bidentate ligands (Schindler et al., 1976). According to Schindler et al. (1976), the ligand properties of silanol surface OH groups are not basically changed from those of free OH-, explaining the coincidence of sorption and hydrolysis. In comparison to the previously discussed iron, manganese, and aluminum oxide surfaces, however, silanol surfaces are not highly sorptive toward heavy metals. These surfaces tend to be even less sorptive toward organic molecules. Schulthess and Huang (1991), for example, found no sorption of humic acid above pH 4, except through a hypothesized metal bridging mechanism. Fulvic acid also was not significantly sorbed by Si sites except in the presence of some metals. Below pH 3.5, fulvic acid could be adsorbed by Si oxide, but few soils are found to be this acidic. When SiO, has been included in low-molecular-weight organic sorption studies, no affinity of either organic or metal-organic complexes for the silanol surface has been noted.

E. CLAYS As noted by Oades (1989), organic-clay complexes and reactions have been studied extensively, and their properties are well established. Likewise, the properties of clays homoionic to various metal ions are well characterized, and the effect of the saturation ion on clay-organic reactions is generally known. The converse, the effect of organic molecules on the retention of metals by clay, has received less attention. The presence of complexing agents such as EDTA or DTPA has been shown to substantially decrease the rate and extent of Zn sorption by smectite (Asher and Bar-Yosef, 1982) and soil clays (Elsokkary, 1980). Stadler and Schindler (1993) similarly noted that p-alanine inhibited Cu sorption by a calcium smectite. The presence of malonate, however, tended to shift the adsorption edge to a higher pH, which they attributed to the formation of ternary surface complexes. Conversely, the presence of ethylenediamineenhanced sorption in the acid pH range, but inhibited sorption at pH >6.0. The former was attributed to stabilization of the cationic species Cu(ethylenediamine),2+and the latter to ligand competition. The presence of EDTA has also been shown to decrease Cd sorption by kaolinite (Chubin and Street, 1981; Haas and Horowitz, 1986). Puls etal. (1991) reported that p-hydroxybenzoic acid and 0-toluic acid decreased Pb and Cd

METAL-ORGANIC COMPLEXATION IN SOILS

245

retention by kaolinite, but 2.4-dinitrophenol increased Cd sorption. The latter was attributed either to formation of a positively charged 1: 1 Cd-organic complex or to preferential sorption of dinitrophenol and subsequent Cd sorption. Considering the structure of 2,4-dinitrophenol, the former seems more probable. The decrease in metal ion sorption in the presence of organic molecules can be attributed to the formation of a soluble complex that maintains the metal ion in solution. Somewhat greater attention has been given to the effect of soil organic matter on the retention of heavy metal ions by clays, but results are often apparently contradictory. The addition of humic acid to kaolinite has been shown to sharply decrease Cu sorption (Gupta and Harrison, 1982). Haas and Horowitz (1986), however, reported increased Cd sorption by kaolinite when either algenic acid or humic acid was added. They suggested the formation of an adsorbed organic layer, which served as a solid phase ligand. The relationship was dependent upon ligand concentration, with sorption decreasing as concentration increased. They felt that two independent reactions were occurring: (1) complexation of the ion by humic acid and (2) sorption of the humic acid by kaolinite. Thus, retention of Cd at the surface only occurred because the sorbed organic had already complexed Cd. When the humic acid concentration in solution increased, the probability of Cd association with an adsorbed molecule decreased. The humic acid:clay ratio used by Gupta and Harrison (1982) was approximately 100 times that used by Haas and Horowitz f1986), which explains the apparent contradiction. Campbell et al. (1987) reported that the addition of a small amount of humic acid to smectite also substantially increased Cd sorption. The Cd was not fixed, however, since it could be removed by decreasing the Cd concentration of the solution. Conversely, Levy and Francis (1976) found that the presence of humic acid does not seem to affect Cd sorption by smectite unless the surface contains Fe or A1 coatings. When the coatings were present, humic acid caused a decrease in Cd sorption. The authors suggested that Cd and humic acid adsorption sites were identical in oxide-coated clays. When no coating was present, the humic acid apparently was retained on planar surfaces, whereas the Cd was retained at edge A1 bond sites. Bar-Tal et al. (1988) examined the influence of fulvic acid on the sorption of Zn by smectite. Below pH 7.5 the presence of fulvic acid in the solution resulted in the Zn remaining in solution, but above pH 7.5 an increasing amount of fulvic acid was necessary to keep the Zn in solution. At all pH levels sorption was decreased by the formation of a Zn-FA complex, but at high pH the relationship shifted due to the formation of ZnOH+ and the high bonding energy of this ion to the surface. Inskeep and Baham (1983) found that the presence of water-soluble extracts of a forest litter layer, dried Chicago sewage sludge, or peat soil all dramatically decreased Cu retention and slightly decreased Cd retention by a

R. D. HARTER AND R. NAIDU smectite as pH increased. The effect was attributed to the formation of a soluble complex in solution, which prevented hydrolysis of the metal ion.

F. SOILS Soil systems provide heterogeneous surfaces with a range of binding affinities to metal ions. Therefore, sorption phenomena within the soil are the cumulative result of individual component reactions. While symbiotic and competitive effects do occur, on the basis of the foregoing discussion some general relationships can be established for predicting whether retention will occur. To a significant extent, sorption reactions will tend to depend on the charge nature of the components. For example, most variable charge soils with low organic matter have a zero point of charge (ZPC) in the range pH 5-5.5. Therefore, at low pH, the colloid surface becomes positively charged and attractive to negatively charged organic ligands. These ligands may in turn be associated with a metal ion, and the metal can be retained through complexation with an adsorbed organic molecule. The result is an observed enhancement in metal sorption in the presence of the organic. At pH levels above the ZPC, the soil should react similarly to one dominated by permanent charge colloids. When the soil is dominated by permanent charge colloids, three basic possibilities exist: (a) If the ligand carries a negative charge, it is less likely to react with surfaces but can complex the metal ions, effectively competing with sorption sites. In such cases, adsorption of the metal ion will be reduced due to complexation with solution phase organic molecules. (b) An uncharged ligand may interact nonionically with the surface (e.g., H-bonding or van der Waals forces) as reported by El-Sayed et al. (1971), who found that Cu(NH,)$+ was retained by smectite in excess of the CEC due to orderly arrangement of the ligands on the surface. (c) Metal ions complexed with organic molecules may still react with the surface, creating a bridge between the inorganic surface and the organic. In both cases (b) and (c), organic sorption by the surface will be enhanced, but metal ion retention may be enhanced, reduced, or unchanged, depending on such factors as bond strength and stereochemistry. Our results (unreported laboratory data) indicate that, in soils having high affinity for metal ions, many of the ligands have little effect on sorption. Chairidchai and Ritchie (1990, 1992,1993) have evaluated the effect of organics on the sorption of Zn by an Australian entisol having a significant variable charge component. Of the organics used, catechol had no effect on sorption. Acetate, tricarballylate, and salicylate tended to decrease metal sorption, with the effect being enhanced in the presence of added humate. Citrate and, to a

METAL-ORGANIC COMPLEXATION IN SOILS

247

lesser effect, oxalate, however, either enhanced or reduced sorption, depending on system parameters. Below the ZPC, citrate increased sorption by as much as 51%, an observation consistent with a decrease in ZPC upon the reaction of citrate with positively charged oxide surfaces. Alternatively, the observation is also consistent with the sorption of negatively charged zinc citrate on positively charged surfaces. (The resultant zinc citrate surface complex would be similar in either case.) At pH levels above the ZPC, citrate reduced Zn retention by up to 38%, probably indicating that Zn was present as a negatively charged zinc citrate complex in solution and therefore could not be bonded to the negatively charged oxide surface. Elliott and Denneny ( 1982) evaluated acetate, oxalate, nitrilotriacetate (NTA), and EDTA complexation of Cd by the surface horizon of three northeastern U . S . soils, one of which (Spodosol) had a significant variable charge component. There was a very slight enhancement in Cd retention by the Spodosol in the presence of acetate, oxalate, and NTA at low pH. Oxalate increased Cd sorption by the other two soils at pH values less than about 5.0. Otherwise, Cd sorption by the soils was generally reduced in the presence of the organics. On the basis of the Cd-ligand stability constant, the authors suggested that reduced sorption was the result of organic ligands holding Cd in solution. Similarly, Prasad and Sarangthem (1993) found that DTPA effectively prevented Zn sorption by a calcareous soil and that EDTA substantially retarded sorption. They did find that fulvic acid was much less effective than the two chelates in holding Zn in solution. Likewise, Jardine, et al., (1993) noted that Co2+transport through the soil was enhanced in the presence of EDTA. Considering the effect of high-molecular-weight organics, McLaren et al. (1981), after studying Cu sorption by a number of soil components, cautioned that sorption would likely be decreased in the presence of soluble organics. In confirmation, Neal and Sposito (1986) reported that sorption of Cd by an Aridisol, an Alfisol, and two English soils, all dominated by permanent charge materials, was inhibited by the presence of soluble organic matter from sewage sludge. They felt that soluble complexes were formed. Likewise, Sinha et al. (1977) noted that fulvic acid reduced Zn sorption by several alkaline soils. Mittal et al. (1984), on the other hand, found virtually no effect to a slight increase in Cu retention by an Indian soil to which up to 2% humic acid had been added, perhaps because the humic acid had been largely sorbed by the soil surface. Evaluating Cd, Cr, and Cu retention by dissolved organic carbon in a Spodosol, Guggenberger et al. (1994) found that the mobile hydrophilic fraction complexed substantially more Cr and Cu than did the hydrophobic fraction, which resulted in significant amounts of Cr and Cu in B horizon leachate. Cd remained in the inorganic form, thus susceptible to the same retention reactions that would occur in the absence of dissolved organics.

248

R. D. HARTER AND R. NAIDU

VI. ENVIRONMENTAL IMPLICATIONS Numerous environmental issues arise in relation to the interaction of metal ions with soluble organics. Some of these include the phytoavailability of metals, plant nutrient availability, toxicological effects of coordinated metal ions on aquatic and marine organisms, and transport of contaminants, particularly in relation to implications for surface and groundwater quality and soil genesis. All of these issues are highly dependent on the nature and concentration of the contaminant in the soil solution phase. The soil solution is the medium from which plants extract ions and water, and, being a potentially mobile phase, it is most susceptible to leaching. Dissolved organics can increase the solution phase affinities of non-ionic compounds and significantly complex ionic compounds under a variety of experimental conditions (Chiou ef al., 1986; Carter and Suffet, 1982; Sheppard ef af., 1980). The interactions illustrated in Fig. 7 emphasize the role of dissolved and particulate organics in affecting the distribution of trace metals in the soil environment’s solid and solution phases. Reactions with metals may range from simple ion pair effects to more complex chelation type reactions involving the formation of ionic bonds. Organic carbon levels (OC > 0.1% by weight) in soils and sediments control the sorption processes of nonionic species by increasing the contaminant sorption (Sheppard et al., 1980; Means et al., 1978, Chiou ef al., 1979). Similar effects are expected of ionic compounds. It follows that dissolved and solid phase organic carbon can be a dominant factor controlling the fate and transport of contaminants in ground water (Fig. 7). The environmental implications of the interactions between organics and metal ions are briefly discussed in this section.

Ground water

Flpre 7. Fate and transport of contaminants in groundwater.

METAL-ORGANIC COMPLEXATION IN SOILS

249

A. CONTAMINANT TRANSPORT Numerous investigators have demonstrated that the association of pollutants with naturally occumng organics can be a major factor controlling the fate and distribution of contaminants in soils, surface water, and sediments (Sheppard et af.,1980; Schwarzenbach and Westall, 1981; Carter and Suffett, 1982; Dunnivant et af., 1992). In addition to the potential for increased solubility of the contaminants (Means et af., 1978; Champ et af., 1984; Kim et af., 1987a,b; Enfield et af., 1989; Abdul et al.. 1990), these organic phases (dissolved and particulate organics), together with highly reactive mineral colloid particles in soil pore water, can enhance the transport of contaminants through porous media (McCarthy and Zachara, 1989). Such transport processes may occur either as soluble metal-organic complexes, stabilized mineral colloid particles with adsorbed contaminants, or particulate organic matter-metal associations (Fig. 7). Trace metals that are typically immobile due to strong binding to soil particles or low water solubility may move to or through the subsurface evironment in association with mobile colloids (Kaplan et af., 1993). On the basis of laboratory column studies, Kaplan et af. (1993) demonstrated that the rate of colloid-associated arsenate transport was over 21 times that of dissolved arsenate. Colloid mobility as a vehicle for contaminant transport often is not discussed, presumably because of the scarcity of information on surface chemical and mineralogical characteristics and conditions conducive to the generation of stable colloid suspensions. According to Puls and Powell (1992), colloid transport is highly dependent upon colloid stability. They found that iron oxide colloids not only were mobile but under some hydrogeochemical conditions they were transported faster than tritiated water, a conservative tracer. Colloid stabilization is governed by particle mineralogy and surface charge density and by the extent of the thickness of the electrical double layer (Matijevic, 1973). The adsorption of dissolved organics and humic substances can impart a negative surface charge to colloids (Fig. 8a), such as oxides, layer silicates, and calcium carbonate, with positively charged surface sites, thereby increasing the stability and mobility of these particles (Fig. 8b). Ryan (1988) postulated that coatings of adsorbed humic substances develop a negative charge on mobile layer silicate and iron oxide colloids in anoxic Atlantic coastal plain groundwater. Kaplan et af.(1993) observed that mobile colloid surface charge was greatly enhanced by organic carbon coating; interestingly, concentrations of organic carbon associated with mobile colloids were equal to or higher than the organic carbon concentrations in the bulk soils from which the mobile colloids were derived. The highly mobile nature of soluble organics has been demonstrated by numerous investigators. Hoffmann et al. (1981), for example, noted that trace metals such as Cd, Cu, and Pb in river water are associated primarily with intermediate-

R. D. HARTER AND R. NAIDU

250

b +I

2

6

10

PH Flgre 8. Schematic diagrams of typical organic ligand impact on surface charge (a) and electrophoretic mobility (b) as affected by system pH.

molecular-weight organics (1OOO- lO,OOO), which readily pass through 0.4-p,m membrane filters. Humic substances have also been found to migrate rapidly in some aquifers. The migration of colloid-sized tannin and lignin (molecular weight >2OOO D) from a waste pulp liquor migrated through a sand aquifer was found to be at the same rate as the groundwater flow (Robertson et al., 1984). Rapid movement of both layer silicate clays and organic matter through the vadose zone has been observed (Jardine et al., 1989), and Naidu et al. (1993) reported the presence of high concentrations of dissolved organics in subsurface water in catchments in the Mount Lofty Ranges, Adelaide, Australia. These

METAL-ORGANIC COMPLEXATION IN SOILS

251

latter investigators demonstrated the transport of metal ions, Fe and Al, in association with dissolved organics. They also reported a strong relationship between dispersed colloid particles and dissolved organics, concluding that DOC enhanced the dispersion of colloids in subsurface water. The mobility of some heavy metals through Indian red soils was investigated by Khan et al. (1982), who reported that the mobility followed the order Cr > Ni > Pb > Cd. The greater mobility of Cr and Ni was attributed to their ability to form highly stable metal-soil organic matter complexes, which facilitated their transport. The lower Cd mobility was attributed to both a lower Cd-soil organic matter stability constant and the high rate of water molecule exchange, decreasing the hydrated ion size and facilitating its interaction with mineral colloids. Madak et al. (1992) also reported the presence of high concentrations of mobile trace metals, such as Cd, Cu, Ni, Cr, and Pb, and low levels of Fe and Mn in association with humic and fulvic acids in the Indian Ganges River water. Dissolved organic matter can influence the adsorption characteristics of mineral particles (Salomons and Forstner, 1984). The humate coatings that mobilize layer silicate clays and oxide colloids alter their surface properties, making them more or less reactive with contaminant ions (Dalang et al., 1984). Moreover, the reaction of metal ions with the mineral surface may be altered through competition with organic molecules for surface binding sites on the particles. This is also apparent from the report by Salomons and Forstner (1984), who found that, under lake water conditions, i.e., with particle concentrations of 2-16 mg I-* and dissolved organic carbon concentrations of 1-4 mg 1-1 at pH 8, adsorption of Cu and Zn was reduced significantly by the presence of natural organic matter. They attributed these observations to the competition between organic molecules and metal ions for the surface binding sites on the particles. Inskeep and Baham (1983) reported that the addition of natural water-soluble organic ligands from forest litter, sewage sludge, or soil had little effect on Cd sorption by smectite, but Farah and Pickering (1976) found a significant difference between sorption of Cd from a landfill leachate and sorption from a pure sodium nitrate solution. This indicates that the nature of organics probably influences the extent of reaction with soil material. The presence of ligands causes the threshold pH, at which precipitation-sorption of hydroxy species occurs, to be shifted to higher values for Pb and Cd on kaolinite, illite, and smectite. The magnitude of the effect depends on the stability of the metal-ligand complex (Farah and Pickering, 1976). This suggests that changes in the surface charge density of the colloid particles in the presence of dissolved organics can have a major influence on the sorption reactions and subsequent transport of metal ions. Chelates such as DTPA and EDTA have been found to be effective in keeping metal ions in soil solution (Prasad and Sarangthem, 1993; James and Bartlett, 1983) and in increasing diffusion rates within the soil (Gupta and Deb, 1984).

252

R. D. HARTER AND R. NAIDU

The results obtained, however, are dependent on pH (Prasad and Sarangthem, 1993; Norvell and Lindsay, 1972; James and Bartlett, 1983) and the electrolytic makeup of the matrix solution (Lahav and Hochberg, 1975; Norvell and Lindsay, 1972), due to the relative stability constants of hydrogen and various cationic forms of the chelates. Similar observations have been made by using a variety of high-molecular-weight natural organics, although these materials are usually less effective in preventing sorption by the soil. Prasad et al. (1990) reported that the majority of Zn moving in calcareous soils is organically bound, and this is also the major source of plant available Zn. Evaluating Cd, Cr, and Cu movement through a spodosol, Guggenberger et al. (1994) found that Cd moved almost entirely in the inorganic form, while Cr and Cu formed stable complexes with hydrophilic acids. The mobility of the complex increased as pH decreased, and the Cr complex did not dissociate even at pH 2.0. The Cu complex did dissociate at lower pH levels. The effect of organic ligands on the sorption of Cu(I1) by standard clay minerals kaolinite, illite, and smectite has been studied extensively in the laboratory (Farah and Pickering, 1977). The results of these investigations show that the nature of interaction varies with the pH and the mineral type. For instance, with kaolinite in alkaline medium, the clay acts as a nucleation site for the formation of hydroxy-bridged copper species, and the major role of many ligands is to mask this precipitation reaction, since uncharged and negatively charged complexes are not sorbed to any measurable extent. By allowing kaolinite to come into contact with the metal prior to the addition of the ligand, greater retention of the metal results than when reacted in the presence of the ligand. The behavior of illite was similar to that of kaolinite, with the controlling process apparently being the formation of polymeric hydroxy species on particular surface sites of the clay. Dissolved organics exhibit both polar and nonpolar chracteristics, depending on the nature of the ligand. Pohlman and McColl(l988) quantified the concentrations of hydrophobic and hydrophilic acids in soluble organics from the forest litter, reporting that both fractions can associate with Fe and Al. The mobility of dissolved organics could facilitate the cotransport of both metal and organic contaminants. Dunnivant et al. (1992) investigated the cotransport of metal ions and nonionic organic contaminants by naturally occumng dissolved organic carbon (DOC) using columns containing aquifer material. Contaminant mobility was found to increase as solution DOC increased. Desorption processes were similarly affected by the presence of mobile DOC. Their results supported the hypothesis that contaminants can be cotransported by mobile DOC in groundwater. They concluded that the transport of Cd was controlled by two competing processes: uptake of Cd by mobile DOC and adsorption to immobile or solid phase DOC. Increased DOC concentrations (while solid phase DOC and solution concentrationsremain constant) resulted in increased Cd mobility in soil columns because of an increased affinity in the solution phase.

METAL-ORGANIC COMPLEXATION IN SOILS

253

B. SOILGENESIS AND FERTILITY The genesis of a number of soil horizons, particularly spodic horizons, has been explained in terms of the formation, migration, and precipitation of soluble organic matter, particularly humic substances complexed with aluminum and, in some cases, with both aluminum and iron. The subsequent release of metal cations in the form of complexes or chelates has an important bearing on soil formation and nutrient supply to plant roots. The role of soil organic matter on the weathering of primary minerals has been demonstrated by Schalscha et al. (1967), who concluded that the weathering products are removed by the chelation of both solution and solid products. Such mobilization and precipitation of the metal chelates, in addition to leading to horizon differentiation giving rise to different kinds of soils (de Connick, 1980), are thought to provide carrier mechanisms by which depleted nutrients at the root surface can be replenished (Lindsay, 1974). The extent of the effect depends on chelate stability. Reviewing the major processes involved in the formation of podzol, Stobbe and Wright (1959) concluded that polyphenols, organic acids, and other complexing substances leached from the litter layer dissolved sesquioxides. Such organic acid-induced weathering of minerals and rocks, through the formation of metal-organic complexes, can enhance the concentrations of both macro- (Song and Huang, 1988) and micronutrients in soils. Laboratory studies by Fox and Comerford (1990) show that a relatively complex suite of organic acids, including oxalic, formic, citric, malic, acetic, and aconitic acids, are present in southeastern U.S. spodosols. Such low-molecular-weight organic acids, which have been found in soils, manure, and sludge materials, can participate in ligand exchange reactions that release P from mineral surfaces (Martell et al., 1988; Bolan ef al., 1994). For further information on the role of organics in soil formation, readers are directed to reviews by Stevenson (1967), Flach et al. (1981), and Stumm et al. (1985).

C. METALTOXICITY Metal toxicity to both plants and other living organisms has long been related to the activity of the species present in the aqueous phase. Chelation to organic ligands present in soil solution can considerably reduce the toxicity of metals through a reduction in the activity of the species in the aqueous phase. Hue et al. (1986) reported that a decrease in dissolved organic carbon in the soil solution can lead to a reduction in the rate and degree of organometal chelation, allowing increased reaction time between metal ions such as A1 and plant root hairs. This could damage the plant root system. Thus, the presence of dissolved organics in the soil solution may prove beneficial to plants, especially in soils with phytotoxic levels of soluble metal ions such as A1 or Mn.

254

R. D. HARTER AND R. NAIDU

There is considerable evidence that the bioavailability of many trace elements present in the aquatic system is influenced by the free ionic forms of several trace elements. For instance, the activities of trace metals such as Cu, Cd, Pb, and Zn have been found to be the most sensitive indicators of toxicity to phytoplankton (Sunda and Guillard, 1976; Wong et al., 1978) and fish (Brown et al., 1970; Davies et al., 1976). However, in many natural waters, the presence of organic acids has been found to reduce the toxic effects of metals (Sunda and Guillard, 1976; Davies et al., 1976) on aquatic organisms. The role of metal-organic interactions in the bioavailability of metals has been reviewed by numerous researchers, and for a more detailed report readers are directed to “The Importance of Chemical Speciation in Environmental Processes” by Bernhard et al. (1986).

MI.SUMMARY AND RESEARCH NEEDS Although the importance of dissolved organics in soil reactions has been long recognized, information on the role of organics in metal reactions with the inorganic components of soil is still scarce. While sorption and complexation of metals by humic and fulvic acids has been investigated extensively, the role of these high-molecular-weightorganics in retention of metals by inorganic surfaces is imperfectly known. Even less understood is the effect of low-molecular-weight organics on metal retention by inorganic surfaces. Soil chemists have often assumed that the low-molecular-weight organics are readily metabolized in the soil, are short-lived, and, therefore, are of limited significance to inorganic reactions. As the ability to accurately measure organics in soil solution improves, however, it is becoming apparent that the low-molecular-weight organics are far more prevelant in soil solution than heretofore assumed. Extant research indicates that low-molecular-weight ligands in soil solution may either enhance or retard reactions with solid surfaces, depending on the functional groups on the organic molecule, soil surface properties, and soil solution conditions. It is therefore imperative that increased research efforts be devoted to evaluating the effects of these organics on metal reactions in the soil.

REFERENCES Abdul. A. S., Gibson, T. L., and Rai, D. N. (1990). Use of humic acid solution to remove organic contaminants from hydrogeologic systems. Envimn. Sci. Technol. 24, 328-333. Adriano, D. C. (1986).‘Trace Elements in the Temstial Environment.” Springer-Verlag. New York. Adriano, D. C. (1992). “Biogeochemistry of Trace Metals.” Lewis, New York. Agganval, B., and Sastry, T. G. (1993). Stability constants of iron and zinc chelates. J. I d . SOC.Soil Sci. 41, 243-256.

METAL-ORGANIC COMPLEXATION IN SOILS

255

Appelt, H.. Coleman, N. T., and Pratt. P. F. (1975). Interactions between organic compounds, minerals, and ions in volcanic-ash derived soils. 11. Effects of organic compounds on the adsorption of phosphate. Soil Sci. Soc.Am. Proc. 39, 628-630. Asher, L. E., and Bar-Yosef, B. (1982). Effects of pyrophosphate, EDTA, and DTPA on zinc sorption by montmorillonite. Soil Sci. SOC. Am. J. 46, 271-276. Bar-Tal, A . , Bar-Yosef, B., and Chen, Y. (1988). Effects of fulvic acid and pH on zinc sorption on montmorillonite. Soil Sci. 146, 367-373. Bartlett, R. J. (1990). An A or an E Which will it be? I n “Proceedings of the Fifth International Soil Correlation Meeting (ISCOM IV) Characterization, Classification, and Utilization of Spodosols” (J. M. Kimble and R. D. Yeck, Eds.), pp. 7-18. USDA, Soil Conservation Service, Lincoln, NE. Basta, N. T., Pantone. D. J., and Tabatabai. M. A. (1993). Path Analysis of Heavy Metal Adsorption by Soil. Agron. J. 85, 1054-1057. Benjamin, M. M.,and Leckie, J. 0. (1981). Multiple site adsorption of Cd, Cu, Zn,and Pb on amorphous iron oxy-hydroxide. J. Colloid Inrerfiace Sci. 79, 209-221. Bergseth, H., and Stuanes, A. (1976). Selectivity of humus material with respect to heavy metal ions. Acra Agric. Scand. 26, 52-58. Bernhard, M., Brinckman, F. E., and Sadler, P. J. (1986). “The Importance of Chemical “Speciation” in Environmental Processes.” Springer-Verlag, New York. Bloomfield, C. (1953a). A study of podzolization. Part 1. The mobilization of iron and aluminum by scots pine needles. J. Soil Sci. 4, 5-16. Bloomfield, C. (1953b). A study of podzolization. Part 11. The mobilization of iron and aluminum by the leaves and bark of Agarhis australis (kauri). J. Soil Sci. 4, 17-23. Bloomfield, C. (1954a). A study of podzolization. Part 111. The mobilization of iron and aluminum by rimu (Dacrydium cupressinum). J. Soil Sci. 5, 39-45. Bloomfield, C. (1954b). A study of podzolization. Part IV. The mobilization of iron and aluminum by picked and fallen larch needles. J. Soil Sci. 5, 46-49. Bloomfield, C. (1954~).A study of podzolization. Part V. The mobilization of iron and aluminium by aspen and ash leaves. J. Soil Sci. 5, 50-56. Bloomfield, C. (1955). A study of podzolization. Part VI. The immobilization of iron and aluminium. J. Soil Sci. 6, 284-292. Bolan, N. S.. Naidu, R., Mahimairaja, S., and Baskaran, S. (1994). Influence of low molecularweight organic acids on the solubilization of phosphorus. Biol. Ferril. Soils., (in press). Borah, D. K., Rattan, R. K., and Banerjee, N. K. (1992). Effect of soil organic matter on the adsorption of Zn, Cu and Mn in soils. J. fnd. SOC. Soil Sci. 40, 277-282. Brady, B., and Pagenkopf, G.K. (1978). Cadmium complexation by soil fulvic acid. Can. J. Chem. 56, 2331-2336. Bresnahan, W. T., Grant, C. L., and Weber, J. H. (1978). Stability constants for the complexation of copper(I1) ions with water and soil fulvic acids measured by an ion selective electrode. Anal. Chem. 50, 1675-1679. Brown, V. M., Shurben, D. G., and Shaw, T. L. (1970). Studies on water quality and the absence of fish from some polluted English rivers. Warer Res. 4, 363-381. Burau, R. G. (1973). Sorption of heavy metals by hydrous MnO. Geochim. Cosmochim. Acra 37, 1277-1293. Campbell, G. D., Galicia. H. F., and Schindler, P. W. (1987). Binding of cadmium by montmorillonite-humic acid mixtures: miscible-displacement experiments. Ausr. J. Soil Res. 25, 391-403. Carter, C. W., and Suffet, 1. H. (1982). Binding of DDT to dissolved humic materials. Envimn. Sci. Technol. 16, 735-740. Cavallaro, N., and McBride, M. B. (1984). Zinc and copper sorption and fixation by an acid soil clay: Effect of selective dissolutions. Soil Sci. SOC. Am. J. 48, 1050-1054.

256

R. D. HARTER AND R. NAIDU

Chairidchai. P., and Ritchie, G.S. P. (1990). Zinc adsorption by a lateritic soil in the presence of organic ligands. Soil Sci. SOC. Am. J. 54, 1242-1248. Chairidchai, P., and Ritchie, G. S. P. (1992). The effect of pH on zinc adsorption by a lateritic soil in the presence of citrate and oxalate. J. Soil Sci. 43, 723-728. Chairidchai. P., and Ritchie, G. S. P. (1993). Zinc adsorption by sterilized and non-sterilized soil in the presence of citrate and catechol. Commun. Soil Sci. Plunf Anal. 24, 261-275. Champ, D.R., Young, J. L., Robertson, D. E., and Abel, K. H. (1984). Chemical speciation of long lived radionuclides in a shallow ground water flow system. Wuterfolluf.Res. J . Can. 19,35-54. Chiou, C. T.,Peters, L. J., and Freed, V. H. (1979). A physical concept of soil-water equilibria for non-ionic organic compounds. Science 206, 83 1-832. Chiou, C. T., Malcolm, R. L., Brinton, T. I.. and Kile, D. E. (1986). Water solubility ehancement of some organic pollutants and pesticides by dissolved humic and fulvic acids. Environ. Sci. Technol. 20, 502-508. Chubin, R. G., and Street, J. J. (1981). Adsorption of cadmium on soil constituents in the presence of complexing ligands. J. Environ. Quul. 10, 225-228. Cromack, K., Jr., Sollins, P., Graustein, W. C., Speidel, K., Todd, A. W., Spycher, G., Li, C. Y., and Todd, R. L. (1979). Calcium oxalate accumulation and soil weathering in mats of the hypogeous fungus Hysrerungium crussum. Soil Biol. Biochem. 11, 463-468. Dalang, F., Buffle, J., and Haerdl, W. (1984). Study of the influence of fulvic substances on the adsorption of copper(l1) ions at the kaolinite surface. Environ. Sci. Technol. 18, 135-141. Davies, P. H., Goettl, J. P., Sinley, J. R., and Smith, N. F. (1976). Acute and chronic toxicities of lead to ranbow trout, Sulmo guirdneri, in hard and soft water. Wuter Res. 10, 199-206. Davis, J. A., and Leckie. J. 0. (1978). Effect of adsorbed complexing ligands on trace metal uptake by hydrous oxides. Environ. Sci. Technol. 12, 1309- 1315. Dawson, H. J., Ugolini, F. C., Hrutfiord, B. F., and Zachara, J. (1978). Role of soluble organics in the soil processes of a podzol, central Cascades, Washington. Soil Sci. 126, 290-296. Deb, D. L.. and De Datta, N. P. (1967). Effect of associated anions or phosphorus retention in soil. 11. Under variable anion concentrations. PIunr Soil 26, 432-444. de Connick, F. (1980). Major mechanisms in formation of spodic horizons. Geodermu 24, 101-128. Dhillon, K. S.. Sinha, M. K., and Randhawa, N. S. (1975). Organo-metallic phosphates. V. Complexation of zinc and phosphorus by humic compounds. PIunr Soil 43, 317-326. Dillard, J. G., Koppelman, M. H., Crowther, D. L.. Schenck, C. V., Murray, J. W., and Balistrieri, L. (1981). X-ray photoelectron spectroscopic (XPS) studies on the chemical nature of metal ions adsorbed on clays and minerals. I n “Adsorption from Aqueous Solutions” (P. H. Tewari, Ed.), pp. 227-240. Plenum Press, New York. Donard, 0. F. X., and Weber, J. H. (1985). Behavior of methyltin compounds under simulated estuarine conditions. Environ. Sci. Technol. 19, 1104-1 110. Driscoll, C. T., and Schecher, W. D. (1990). The chemistry of aluminium in the environment. Environ. Geochem. Heulrh 12, 28-49. Driscoll, C. T., Baker, J. P., Bisogni, J. J., and Schofield, C. L. (1980). Aluminium speciation and its effect on fish in dilute acidified waters. Nurure 284, 161. Dunnivant, F. M., Jardine, P. M., Taylor, D. L., and McCarthy, J. F. (1992). Cotransport of cadmium and hexachlorobiphenyl by dissolved organic carbon through columns containing aquifer material. Environ. Sci. Technol. 26, 360-368. Elliott, H. A., and Denneny. C. M. (1982). Soil adsorption of cadmium from solutions containing organic ligands. J. Environ. Quul. 11, 658-662. Elliott, H. A,, and Huang, C. P. (1979). The adsorption characteristics of Cu(l1) in the presence of chelating agents. J. Colloid lnrerjuce Sci 70, 29-45. El-Sayed, M. H., Burau, R. G.,and Babcock, K. L. (1971). Reactions of copper tetrammine with bentonite clay. Soil Sci. SOC. Am. Pmc. 35, 571-574.

METAL-ORGANIC COMPLEXATION IN SOILS

257

Elsokkary, 1. H. (1980). Reaction of labeled zinc-65 dichloride, zinc-65 EDTA and zinc-65 DTPA with different clay-systems and some alluvial Egyptian soils. Plant Soil 54, 383-393. Enfield, C. F., Bengtsson, G., and Lindquist, R. (1989). Influence of macromolecules on chemical transport. Envimn. Sci. Technol. 23, 1278-1286. Farah, H., and Pickering, W. F. (1976). The sorption of copper by clays. II. Illite and montmorillonite. A m . J. Chem. 29, 1177- 1184. Farah, H . , and Pickering. W. F. (1977). The sorption of lead and cadmium species by clay minerals. Ausr. J. Chem. 30, 1417-1422. Flach, K. W., Holzhey, C. S . . de Connick, F., and Bartlett, R. J. (1981). Genesis and classification of andepts and spodosols. In “Soils with Variable Charge” (B. K. G. Theng, Ed.), pp. 41 1-426. Offset Publications, Palmerston North, New Zealand. Fox, T. R., and Comerford, N. B. (1990). Low molecular weight organic acids in selected forest soils of the southeastern USA. Soil Sci. SOC. Am. J. 54, 1139- 1144. Fox, T. R., Comerford, N. B.,and McFee, W. W. (1990). Phosphorus and aluminium release from a spodic horizon mediated by organic acids. Soil Sci. Sac. Am. J . 54, 1763-1767. Fu, G., Allen, H. E., and Cowan, C. E. (1991). Adsorption of cadmium and copper by manganese oxide. Soil Sci. 152, 72-81. Girvin, D. C., Gassman, P. L., and Bolton, H. (1993). Adsorption of aqueous cobalt ethylenediaminetetraacetate by delta-Al,O,. Soil Sci. SOC. Am. J. 57, 47-57. Graustein, W. C . , Cromack, K.,Jr., and Sollins, P. (1977). Calcium oxalate: Occurrence in soils and effect on nutrient and geochemical cycles. Science 198, 1252-1254. Grierson, P. F. (1992). Organic acids in the rhizosphere of Banksia inregrijioriu L. f. Plant Soil 144, 259-265. Guggenberger, G.. Glaser, B., and Zech, W. (1994). Heavy metal binding by hydrophobic and hydrophilic dissolved organic carbon fractions in a spodosol A and B horizon. Warer Air Soil Pollur. 72, I 1 1- 121. Gupta, G. C., and Harrison, F. L. (1982). Effect of humic acid on copper adsorption by kaolin. Warer Air Soil Pollur. 17, 357-360. Gupta, G. N., and Deb, D. L. (1984). Effect of chelating agents on zinc diffusion in two soils. 2. Pjanz. Bodenk. 147, 533-539. Haas, C. N., and Horowitz, N. D. (1986). Adsorption of cadmium to kaolinite in the presence of organic material. Warer Air Soil Pollur. 27, 131-140. Harter, R. D. (1979). Adsorption of copper and lead by Ap and B2 horizons of several northeastern United States soils. Soil Sci. SOC. Am. J. 43, 679-683. Henderson, M. E. K.,and Duff, R. B. (1963).The release of metallic and silicate ions from mineral, rocks, and soils by fungal activity J. Soil Sci. 14, 236-246. Hoffmann, M. R., Yost, E. C., Eisenreich, S. J., and Maier, W. J. (1981). Characterization of soluble and colloidal-phase metal complexes in river water by ultrafiltration: A mass-balance approach. Environ. Sci. Technol. 15, 655-661. Holmgren, G. G. S., Meyer. M. W., Chaney, R. L., and Daniels, R. B. (1993). Cadmium, lead, zinc, copper, and nickel in agricultural soils of the United States of America. J . Environ. Qual. 22, 335-348. Hsu, P. H. (1989). Aluminum hydroxides and oxyhydroxides. I n “Minerals in Soil Environments” (J. B. Dixon and S. B. Weed, Eds.), 2nd ed.. pp. 331-378. Soil Science Society of America, Madison, WI. Huang, C. P., and Lin, Y. T. (1981). Specific adsorption of Co(ll) and (Co(III)EDTA]- complexes on hydrous oxide surfaces. I n “Adsorption from Aqueous Solutions” (P. H. Tewari, Ed.), pp. 61-91. Plenum Press, New York. Huang, P. M.. and Schnitzer, M. (1986). “Interactions of Soil Minerals with Natural Organics and Microbes,” SSSA Special Pub. No. 17. Soil Scociety of America, Madison, WI.

2 58

R. D. HARTER AND R. NAIDU

Hue, N. V. (1991). Effect of organic acid/anion on P sorption and phytoavailability in soils with different mineralogies. Soil Sci. 152, 463-47 I . Hue, N.V., Craddock, G. R., and Adams, F. (1986). Effect of organic acids on aluminium toxicity in subsoils. Soil Sci. SOC. Am. J, 50, 28-34. Inskeep, W. P., and Baham, 1. (1983). Competitive complexation of cadmium(I1) and copper(I1) by water-soluble organic ligands and sodium-montmorillonite. Soil Sci. SOC. Am. J. 47, 11091115. James, B. R., and Bartlett, R. J. (1983). Behavior of chromium in soils. V. Fate of organically complexes Cr(II1) added to soil. J. Envimn. Q u l . 12, 169-172. Jardine, P. M.,Wilson, G. V., Luxmoore, R.J., and McCarthy, 1. F. (1989). Transport of inorganic and natural organic traces through an isolated pedon in a forest watershed. Soil Sci. SOC.Am. J . 53, 317-323. Jenne, E. A. (1968). Controls on Mn, Fe. Co, Ni, Cu, and Zn concentrations in soils and water; the significant role of hydrous Mn and Fe oxides. Adv. Chem. 73, 337-387. Johnson, C. A. (1986). The regulation of trace element concentration in river and estuarine waters contaminated with acid mine drainage: The adsorption of Cu and Zn on amorphous Fe oxyhydroxides. Geochim. Cosmochim. Acta 50, 2433-2438. Kabata-Pendias, A. (1980). Heavy metals sorption by clay minerals and oxides of iron and manganese. Mineral. Polon. 11, 3-13. Kalbasi, M., Racz, G. J., and Loewen-Rudgers, L. A. (1978). Mechanism of zinc adsorption by iron and aluminum oxides. Soil Sci. 125, 146-150. Kaplan, D. I., Bertsch, P. M.,Adriano, D. C., and Miller, W. P. (1993). Soil-borne mobile colloids as influenced by water flow and organic carbon. Environ. Sci. Technol. 27, 1193-1200. Katase, T. (1981). Distribution of different forms of p-hydroxybenzoic, p-coumaric, and ferulic acids in forest soils. Soil Sci. Plant Nutr. 27, 365-37 I . Keefer, R. F., and Singh, R. N. (1986). Correlation of metal-organic fractions with soil properties in sewage-sludge-amendedsoils. Soil Sci. 142, 20-26. Khan, S., Nandan, D., and Khan, N. N. (1982). The mobility of some heavy metals through Indian red soils. Envimn. follut. (Ser. B ) 4, 119-125. Kim, J. I., Buckua, G., and Zhuang, W. (1987a). Humid colloid generation of transuranic elements in ground water and the migration behaviour. Mazer. Res. Soc.Symp. Proc. 84, 747-756. Kim, J. I., Buckua, G., and Klenze, R. (1987b). I n “Natural Analogues in Radioactive Waste Disposal” (B. Come and N. A. Chapment, Eds.), pp. 289-299. Graham and Trotman, London. Kooner, Z. S. (1993). Comparative study of adsorption behavior of copper, lead, and zinc onto goethite in aqueous systems. Envimn. Geol. 21, 242-250. Kummert, R., and Stumm, W. (1980). The surface complexation of organic acids on hydrous a-Al,O,. J. Colloid Interface Sci. 75, 373-385. Kuo, S., and McNeal, B. L. (1984). Effects of pH and phosphate on cadmium sorption by a hydrous femc oxide. Soil Sci. SOC. Am. J. 48, 1040-1044. Lahav, N., and Hochberg, M.(1975). Fixation of iron and zinc applied as chelates into a soil column during leaching. Soil Sci. Soc.Am. fmc. 39, 1213-1215. Levy, R., and Francis, C. W. (1976). Adsorption and desorption of cadmium by synthetic and natural organo-clay complexes. Geodenna 15, 361-370. Lindsay, W. L. (1974). Role of chelation in micronutrient availability.In ‘The Plant Root and Its Environment” (E. W. Carson, Ed.), pp. 507-524. University Press of Virginia, Charlottesville. VA. Lindsay, W. L. (1979). “Chemical Equilibria in Soils.” Wiley, New York. Lynch, J. M.(1978). Production and phytoxicity of acetic acid in anaerobic soils containing plant residues. Soil Biol. Biochem. 10, 131-135. Madak, D. P., Singh, K. P.,Chandra, H., and Ray, P.K. (1992). Mobile and bound forms of trace metals in sediments of the lower Ganges. Worer Res. 26, 1541-1548.

METAL-ORGANIC COMPLEXATION IN SOILS

2S9

Manceau, A,, Charlet, L., Boisset, M. C., Didier, B., and Spadini, L. (1993). Sorption and speciation of heavy metals on hydrous Fe and Mn oxides: From microscopic to macroscopic. Appl. Clay Sci. 7 , 201-223. Mantoura, R. F. C., and Riley, J. P. (1975). The use of gel filtration in the study of metal binding by humic acids and related compounds. Anal. Chim. Acra 78, 193-200. Martell, A. E. (1960). The relationship of chemical structure to metal-binding action. I n “MetalBinding in Medicine” (M. Seven, Ed.), pp. 1-8. Lippincott, Philadelphia, PA. Martell, A. E. (1967). The chelate effect. Adv. Chem. Ser. 62, 272-294. Martell, A. E. (1978). Chelating agents for metal buffering and analysis in solution. Pure Appl. Chem. 50, 813-829. Martell, A. E., and Smith, R. M. (1977). “Critical Stability Constants. Vol. 3. Other Organic Ligands.” Plenum Press, New York. Martell, A. E., Motekaitis, R. J., and Smith, R. M. (1988). Structure-stability relationships of metal complexes and metal speciation in environmental aqueous solutions. Environ. Toxicol. Chem. 7 , 417-434. Matijevic, E. (1973). Colloid stability and complex chemistry. J. Colloid Interface Sci. 43, 217-244. Matsuda, K., and Ito. S. (1970). Adsorption strength of zinc for soil humus. 111. Relation between stability constants of zinc-humic and -fulvic acid complexes, and the degree of humification. Soil Sci. Plant Nurr. 16, 1-10, McBride, M. B. (1982). Cu2+ adsorption characteristics of aluminum hydroxide and oxyhydroxides. Clays Clay Miner. 30,21-28. McBride, M. B. (1985a). Influence of glycine on Cuz+ adsorption by microcrystalline gibbsite and boehmite. Clays Clay Miner. 33, 397-402. McBride, M. B. (1985b). Sorption of copper(I1) on aluminum hydroxide as affected by phosphate. Soil Sci. Soc. Am. J. 49, 843-846. McBride, M. B. (1987). Adsorption and oxidation of Phenolic compounds by iron and manganese oxides. Soil Sci. Soc. Am. J . 51, 1466-1472. McBride, M. B. (1989). Reactions controlling heavy metal solubility in soils. Adv. Soil Sci. 10, I 56. McBride, M. B., and Wesselink, L. G. (1988). Chemisorption of catechol on gibbsite. boehmite, and noncrystalline alumina surfaces. Environ. Sci. Technol. 22, 703-708. McBride, M. B., Fraser, A. R., and McHardy, W. J. (1984). Cuz+ interaction with microcrystalline gibbsite: Evidence for oriented chemisorbed copper ions. Clays Clay Miner. 32, 12-18. McCarthy, J. F., and Zachara, J. M. (1989). Subsurface transport of contaminants: Role of mobile colloid and particles. Environ. Sci. Technol. 23, 496-504. McColl, J. G., and Pohlman, A. A. (1986). Soluble organics and their chelating influence on A1 and other metal dissolution from forest soils. Water Air Soil follur. 31, 917-927. McDowell. W. H., and Likens, G. E. (1988). Origin, composition and flux of dissolved organic carbon in the Hubbard Brook Valley. Ecol. Monogr. 58, 177-195. McKenzie, R. M. (1972). The sorption of some heavy metals by the lower oxides of manganese. Geoderma 8 , 29-35. McLaren, R. G., and Crawford, D. V. (1973). Studies on soil copper 11. the specific adsorption of copper by soils. J. Soil Sci. 24, 443-452. McLaren, R. G., Swift, R. S., and Williams, J. G. (1981). The adsorption of copper by soil materials at low equilibrium solution concentrations. J. Soil Sci. 32, 247-256. Means, I. L., Crerar, D. A,, and Dubuid, J. 0. (1978). Migration of radioactive wastes: Radionuclide mobilization by complexing agents. Science 200, 1477- 1481. Mehadi, A. A. (1993). “Reaction of Nickel with Soils and Goethite: Equilibrium and Kinetic Studies.” Ph.D. Dissertation, University of New Hampshire, Durham, NH (DA9400395). Misono, M., Ochiai, E., Saito, Y., and Yoneda, Y. (1967). A new dual parameter scale for the

2 60

R. D. HARTER AND R. NAIDU

strength of Lewis acids and bases with the evaluation of their softness. J. Inorg. Nucl. Chem. 29, 2685-2691. Mittal, S . B., Mehta, S. C., and Rastogi, S. K. (1984). Influence of humic acid and calcium carbonate on copper adsorption in soils. J. I d . SOC. Soil Sci. 32, 366-367. Muir, 1. W., Morrison, R. I., Brown, C. I., and Logan, 1. (1964). The mobilization of iron by aqueous extracts of plants. J. Soil Sci. 15, 220-225. Naidu, R., Williamson, D. R., Fitzpauick. R. W., and Hollingsworth, 1. (1993). Effect of land use on the composition of throughflow water immediately above clayey B horizons in the Warren catchment. Ausf. J. Exp. Agric. 33, 239-44. Neal, R. H., and Sposito, G. (1986). Effects of soluble organic matter and sewage sludge amendments on cadmium sorption by soils at low cadmium concentrations. Soil Sci 142, 164-172. Norvell, W. A,, and Lindsay, W. L. (1972). Reactions of DTPA chelates of iron, zinc, copper, and manganese with soils. Soil Sci. Soc. Am. Proc. 36, 778-783. Oades, J. M. (1989). An introduction to organic matter in mineral soils. In “Minerals in Soil Environments,” (J. B. Dixon and S . B. Weed, Eds.), 2nd ed., pp. 89-159. Soil Science Society of America. Madison, WI. Okazaki, M., Takamidoh, K., and Yamsne, I. (1986). Adsorption of heavy metal cations on hydrated oxides and oxides of iron and aluminum with different crystallinities. Soil Sci. Planr Nurr. 32, 523-533. Pandeya, S. B. (1993). Ligand Competition Method for Determining Stability Constants of Fulvic Acid Iron Complexes. Geoderma 58, 219-231. Parfitt, R. L., Fraser, A. R., Russell, 1. D.,and Farmer, V. C. (1977). Adsorption on hydrous oxides. 11. Oxalate, benzoate, and phosphate on gibbsite. J. Soil Sci. 28, 40-47. Pearson, R. G. (1963). Hard and soft acids and bases. J. Am. Chem. SOC. 85, 3533-3539. Petruzzelli, G., Guidi. G., and Lubrano, L. (1978). Organic matter as an influencing factor on copper and cadmium adsorption by soils. WaferAir Soil Pollur. 9, 263-269. Petruzzelli, G., Guidi, G., and Lubrano, L. (1981). Interactions among heavy metals and organic matter in soil. In “Heavy Metals in the Environment,” pp. 686-689. Commission of the European Communities and World Health Organisation, Amsterdam. Pohlman, A. A., and McColl, J. G. (1986). Kinetics of metal dissolution from forest soils by soluble organic acids. J. Environ. Quai. 15, 86-92. Pohlman, A. A., and McColl, J. G. (1988). Soluble organics from forest litter and their role in metal dissolution. Soil Sci. SOC. Am. J. 52, 265-271. Prasad, B . , Mehta, A. K., and Sinha, M. K. (1990). Zinc fractions and availability of applied zinc in calcareous soil treated with organic materials. J. I d . SOC. Soil Sci. 38, 248-253. Prasad, B . , and Sarangthem, I. (1993). Adoption of zinc as affected by its sources in calcareous soils. J. I d . SOC. Soil Sci. 41, 261-265. Prasad, B . , and Sinha, M. K. (1980).Physical and chemical characterization of soil and poultry litter humic and fulvic metal complexes. Planr Soil 54, 223-232. Puls, R. W., and Powell, R. M. (1992). Transport of inorganic colloids through natural aquifer material. Implications for contaminant transport. Environ. Sci. Techno1 26, 614-621. Puls, R . W., Powell, R. M., Clark, D.,and Eldred, C. J. (1991). Effects of pH, solid/solution ratio, ionic strength, and organic acids on Pb and Cd sorption on kaolinite. Wurer Air Soil Pollur. 5758,423-430. Randall, L., and Weber, I. H. (1986). Adsorptive behavior of butyitin compounds under simulated estuarine conditions. Sci. Total Environ. 57, 191-203. Robertson, W. D., Barker, J. F., LeBeau, Y.,and Marcoux, S. (1984). Contamination of an unconfined aquifer by waste pulp liquor: A case study. Ground Water 22, 191-197. Rosell, R. A., Miglierina, A. M., and Novilla, L. Q. (1977). Stability constants of some complexes of Argentine humic acids and micronutrients. In “Soil Organic Matter Studies. Proc. Symp.

METAL-ORGANIC COMPLEXATION IN SOILS

261

Organized by IAEA, FA0 and Agrochimica, Braunschweig, September 1976,” Vol. 11, pp. 1521. International Atomic Energy Agency, Vienna, Austria. Rovira, A. D. (1969). Plant root exudates. Bor. Rev. 35, 35-57. Rovira, A. D., and Davey, C. B. (1974). Biology of the rhizosphere. In “The Plant Root and Its Environment” (C. W. Carson, Ed.), pp. 153-204. University of Virginia Press, Charlottesville. VA . Rueda, E. H., Grassi, R. L., and Blesa, M. A. (1985). Adsorption and dissolution in the system goethitelaqueous EDTA. J. Colloid Interface Sci. 106, 243-246. Ryan. J. N. (1988). “Ground Water Colloids in Two Atlantic Coastal Plain Aquifers: Colloid Fonnation and Stability.” M.S. Thesis. Dept. of Civil Eng., Massachusetts Institute of Technology, Cambridge, MA, 250 pp. Saar, R. A., and Weber, J. H. (1979). Complexation of cadmium(I1) with water- and soil-derived fulvic acids: Effect of pH and fulvic acid concentration. Can. J. Chem. 57, 1263-1268. Salomons, W., and Forstner, U. (1984). “Metals in the Hydrocycle.” Springer-Verlag, Berlin. Schalscha, E. B., Appelt, H., and Schatz, A. (1967). Chelation as a weathering mechanism. I. Effect of complexing agents on the solubilization of iron from minerals and granodiorite. Geochim. Cosmochim. Acra 31,587-596. Schindler, P. W., Furst, B., Dick, R., and Wolf, P. U. (1976). Ligand properties of surface silanol groups I. Surface complex formation with Fe, Cu, Cd, and Pb. J. Colloid Interface Sci. 55, 469-475. Schlichting, E., and Elgala, A. M. (1975). Heavy metal distribution and clay contents in soils. Z . Pjlanz. Bodenk. 1975, 563-571. Schnitzer, M. (1978). Humic substances: chemistry and reactions. In “Soil Organic Matter” (M. Schnitzer and S. U. Khan, Eds.), pp. 1-64. Elsevier, New York. Schnitzer, M., and Skinner, S. 1. M. (1966). Organo-metalic interactions in soils. 5 . stability constants of Cu++-, Fe++-, and Zn++-fulvic acid complexes. Soil Sci. 102, 361-365. Schnitzer. M., and Skinner, S. 1. M. (1967). Organo-metalic interactions in soils. 7. stability constants of Pb++-, Ni++-, Mn++-, Co++-, Ca++-, and Mg++-fulvic acid complexes. Soil Sci. 103, 247-252. Schulthess, C. P., and Huang, C. P. (1990). Adsorption of heavy metals by silicon and aluminum oxide surfaces on clay minerals. Soil Sci. SOC. Am. J . 54, 679-688. Schulthess, C. P., and Huang, C. P. (1991). Humic and fulvic acid adsorption by silicon and aluminum oxide surfaces on clay minerals. Soil Sci. SOC. Am. J . 55, 34-42. Schwarte, S. M., Vamer, I. E.. and Martin, W. P. (1954). Separation of organic acids from several dormant and incubated soils. Soil Sci. SOC.Am. Proc. 18, 174-177. Schwarzenbach, R. P., and Westall, J. (1981). Transport of nonpolar organic compounds from surface water to groundwater. Laboratory sorption studies. Environ. Sci. Technol. 15, 13601367. Schwertmann, U., and Cornell, R. M. (1991). “Iron Oxides in the Laboratory: Preparation and Characterization.” Verlagsgesellschaft mbH, Weinheim, Germany. Schwertmann, U.,and Taylor, R. M. (1989). Iron oxides in the laboratory. In “Minerals in Soil Environments”(J. B. Dixon and S. B. Weed, Eds.), 2nd ed., pp. 379-438. Soil Science Society of America, Madison, WI. Sheppard, J. C., Cambell, M. J., Change, T.,and Kittrick, J. A. (1980). Retention of radionuclides by mobile humic compounds and soil particles. Environ. Sci. Technol. 14, 1319-1353. Shindo, H. (1992). Relative effectiveness of short-range ordered Mn(IV), Fe(III), Al, and Si oxides in the synthesis of humic acids from phenolic compounds. Soil Sci. Plant Nurr. 38, 459-465. Shuman, L. M. (1977). Adsorption of Zn by Fe and Al hydrous oxides as influenced by aging and pH. SoilSci. SOC. Am. J . 41,703-706. Shuman, L. M. (1979). Zinc, manganese, and copper in soil fractions. Soil Sci. 127, 10-17.

262

R. D. HARTER AND R. NAIDU

Shuman, L. M. (1988). Effect of removal of organic matter and iron- or manganese-oxides on zinc adsorption by soil. Soil Sci. 146, 248-254. Singh, M. (1971). Retention of added copper by two soils as affected by organic matter, CaCO, and exchangeable ions. Geoderma 5, 219-227. Sinha, M. K., Dhillon, S. K., and Dhillon, K. S. (1977). Zinc chelate reactions in alkaline soils. Ausr. J . Soil Res. 15, 103-113. Smith, W. H. (1976). Character and significance of forest tree root exudates. Ecology 57,324-331. Song, S. K., and Huang. P. M. (1988). Dynamics of potassium release from potassium-bearing minerals as influenced by oxalic and citric acids. Soil Sci. SOC.Am. J. 52, 383-390. Sposito, G. (1989). ‘“The Chemistry of Soils.’’ Oxford Univ. Press, Oxford, England. Sposito, G., Holtzclaw, K. M., and LeVesque-Madore,C. S. (1981). Trace metal complexation by fulvic acid extracted from sewage sludge. I. Determination of stability constants and linear correlation analysis. Soil Sci. SOC.Am. J. 45, 465-468. Stadler, M., and Schindler, P. W. (1993). The effect of dissolved ligands upon the sorption of Cu(I1) by Ca montmorillonite. Clays Clay Miner. 41, 680-692. Stahl, R. S., and James, B. R. (1991a). Zinc sorption by iron-oxide-coated sand as a function of pH. Soil Sci. SOC. Am. J. 55, 1287-1290. Stahl, R. S., and James, B. R. (1991b). Zinc sorption by manganese-oxide-coatedsand as a function of pH. Soil Sci. SOC.Am. J. 55, 1291-1294. Stevenson, F. J. (1967). Orgnaic acids in soil. In “Soil Biochemistry” (A. D. McLaren and G. H. Peterson, Eds.), Vol. 1 ., pp. 119- 146. Dekker, New York. Stevenson, F. J., and Ardakani, M. S. (1972). Organic matter reactions involving micronutrients. In “Micronutrients in Agriculture” (J. J. Mordvedt er al.. Eds.),pp. 79-1 14. American Society of Agronomy, Madison, WI. Stobbe, P. C., and Wright, J. R. (1959). Modern concepts of the genesis of Podzols. Soil Sci. SOC. Am. Pmc. 23, 161-164. Stone, A. T., and Morgan, J. J. (1984a). Reduction and dissolution of manganese(I11) and manganese(1V) oxides by organics. 1. Reaction with hydroquinone. Environ. Sci. Technol. 18,450456. Stone, A. T., and Morgan, J. J. (1984b). Reduction and dissolution of manganese(II1) and manganese(1V) oxides by organics. 2. Survey of the reactivity of organics. Envimn. Sci. Technol. 18, 617-624. Stumm, W., Furrer, G., Wieland, E., and Zinder, B. (1985). The effects of complex-formingligands on the dissolution of oxides and aluminosilicates. In ‘“The Chemistry of Weathering” (J. I. Drever, Ed.), pp. 55-74. Reidel, Boston, MA. Sunda, W., and Guillard, R. R. L. (1976). The relationship between cupric ion activity and the toxicity of copper to phytoplankton. J. Mar. Res. 34, 51 1-529. Taga, M., Tanaka, S., and Fukushima, M. (1991). Evaluation of copper(I1)-bindingability of humic acids in peat using sulphopropyl-SephadexC-25 cation exchanger. Anal. Chim. Acra 244, 281287. Thurman, E. M. (1985). “Organic Geochemistry of Natural Waters.” Nijhoff-Junk. Boston, MA. Tipping, E., and Hurley, M. A. (1992). A unifying model of cation binding by humic substances. Geochim. Cosmochim. Acra 56, 3627-3641. Tobia, S. K., and Hanna, A. S. (1958). Effect of copper sulfate added to imgation water on copper status of Egyptian soils. 1. amount of copper retained by soils. Soil Sci. 85, 302-306. Traina, S. J., and Doner, H. E. (1985). Heavy metal induced releases of manganese(l1) from a hydrous manganese dioxide. Soil Sci. SOC.Am. J. 49, 317-321. Traina, S. J., Sposito, G., Hesterberg, D., and Kafkafi, U. (1986). Effects of pH and organic acids on orthophosphate solubilty in acidic, montmorillonitic soil. Soil Sci. SOC.Am. J. 50, 45-52.

METAL-ORGANIC COMPLEXATION I N SOILS

263

Ugolini, F. C., Minden, R., Dawson, H., and Zachara, J. (1977). An example of soil processes in the Abies amabilis zone of Central cascades, Washington. Soil Sci. 124, 294-302. Vordonis, L., Spanos, N., Koutsoukos, P.G., and Lycourghiotis, A. (1992). Mechanism of adsorption of Co2+ and Ni2+ ions on the pure and fluorinated ~-alumina/electrolytesolution interface. Langmuir 8, 1736- 1743. Wakatsuki, T., Rasyidin, A., and Naganawa, T. (1993). Multiple regression method for estimating rates of weathering and soil formation in watersheds. Soil Sci. Plant Nurr. 39, 153-159. Whitehead, D. C., Dibb, H., and Hartley, R. D. (1981). Multiple regression method for estimating rates of weathering and soil formation in watersheds. Soil Biol. Eiochem. 13, 343-348. Whitehead, D. C., Dibb, H., and Hartley, R. D. (1983). Bound phenolic compounds in water extracts of soils, plant roots and leaf litter. Soil Biol. Biochem. 15, 133-136. Wong, P.T.S..Chau, Y.K.,and Luxon, P. L. (1978). Toxicity of a mixture of metals on freshwater algae J. Fish. Res. Board Can. 35, 479-481. Xu, S.,and Harsh, J. B. (1990a). Monovalent cation selectivity quantitatively modeled according to hardlsoft acidlbase theory. Soil Sci. SOC.Am. J. 54, 357-363. Xu, S., and Harsh, J. B. (1990b). Hard and soft acid-base model verified for monovalent cation selectivity. Soil Sci. SOC. Am. J. 54, 1596-1601. Zasoski, R. J., and Burau, R. G. (1988). Sorption and sorptive interaction of cadmium and zinc on hydrous manganese oxide. Soil Sci. SOC. Am. J. 52, 81-87. Zeltner, W. A,, Yost, E. C., Machesky, M. L.,Tejedor-Tejedor. M. I . , and Anderson, M. A. (1986). Characterization of anion binding on goethite using titration calorimetry and cylindrical internal reflection-Fourier transformation infrared spectroscopy. In “Geochemical Processes at Mineral Surfaces” (J. A. Davis and K. F. Hayes, Eds.), pp. 142-161. American Chemical Society, Washington, DC.

This Page Intentionally Left Blank

DNA MARKERS AND PLANTBREEDING PROGRAMS Michael Lee Department of Agronomy, Iowa State University, Ames, Iowa 5001 1

I. Introduction A. Components of Plant Breeding Methods B. Sources of Genetic Gain for Increased Crop Productivity C. DNA Markers as Fundamental Links between Plant Breeding and Plant Biology 11. Assessing Genetic Diversity and Merit A. Germ Plasm Identification, Classification, and Management B. Parent Selection C. Limits of Assessing Genetic Diversity via DNA Markers 111. Genome Architecture: Genetic and Physical Characterization of Crop Plant Genomes A. Development of Integrated Maps B. Relating Genetic and Physical Distances in Crop Plant Genomes C. Insights into Recombination and Its Role in Generating Genetic Variation D. Map-Based Cloning Comes of Age IV. Analysis of Complex Traits and Phenomena A. Quantitative Inheritance Patterns B. Assessing and Introgressing Exotic Germ Plasm C. Response to Selection in Plant Breeding Programs V. Marker-Assisted Selection A. Deterministic and Simulation Studies of MAS in Plant Breeding Programs B. Empirical Results C. Integrating MAS into Plant Breeding Programs VI. Survey of the Status of DNA Markers in Cultivar Development Programs VII. Summary and Conclusions References

I. INTRODUCTION “One of the great success stones of U.S. agricultural research and development is the contribution of enhanced genetic potential to grain yields in major 265 Aduanca m Agronomy, Vdmr YY Copyright Q 1995 by Academic Press, Inc. All rights of reproductionin any form reserved

2 66

M.LEE

crops” (Huffman and Evenson, 1993). On a worldwide basis, plant breeding has been one of the most successful technologies developed in modem agriculture: its methods are opportunistic and adaptable to myriad production schemes, they require relatively inexpensive input, and their products have pervasive social benefits. Imminent global developments will demand continued and perhaps greater success because agricultural systems will be required to maintain or increase production with fewer resources: the human population is projected to grow from 5.7 to 10 billion by the year 2050, the supply of high quality land is diminishing, and concerns about environmental quality will encourage production practices that require less energy (e.g., chemicals and fuel). In many circumstances, plant genetics will substitute for extant production and management practices, and plant breeders will mediate the substitution.

A, COMPONENTS OF PLANT BREEDINGMETHODS Most successful plant breeding programs consist of several components that are often conducted as reiterative procedures [from Schnell (1982) with slight modification]: ( I ) identify clear and reasonable goals for selection regarding type of cultivar, traits, and their levels; (2) define the target environment for production; (3) understand the plant, production system, and clients; (4) carefully select and create test environments representative of the target environment; ( 5 ) survey and choose germ plasm (DNA sequences, genes, pure lines, clones, accessions, cultivars, populations, and scientists); (6) identify and create genetic variation; (7) assemble genes into genotypes; (8) match genotypes with environment to optimize production; (9) hire good scientists. When practiced on a continuous basis, these components of plant breeding have achieved impressive results.

B. SOURCES OF GENETICGAINFOR INCREASED CROPPRODUCTIVITY The contributions of plant breeding to improved productivity of several major crops have been well documented for U.S. agriculture (Duvick, 1984; Fehr, 1984). For most grain crops, yields have increased continuously since the 1930s, and nearly 50% of the gains may be attributed to the enhanced genetic potential of the cultivars. Similar patterns, albeit often at lower magnitudes and rates, may be observed for several other regions of the world. Besides increased productivity, plant breeding has been capable of remarkable transformations in quality (e.g., canola from rapeseed), growth habits and use (e.g., in the U.S.,1930s soybeans as forage crop to 1960s soybeans as grain crop), and adaptation (e.g., short stature small grain cultivars, adapting tropical sorghum germ plasm to

DNA MARKERS AND PLANT BREEDING PROGRAMS

267

temperate latitudes). Certainly a parallel list of less successful endeavors could be prepared, but over all, the genetic modifications of crop species achieved through plant breeding have been very positive and important. To assess and develop strategies for continued progress, several questions should be addressed. How have the gains been achieved? Can the present rates of gain be maintained or improved? What are the sources and costs of future gains? Substantive answers to these questions become progressively difficult to obtain. The first question is rather easy to answer, superficially. Genetic gains of major crops have two major sources: the infrastructure inherent to the crop and that endowed to the plant breeding program. With only a rudimentary understanding of a few aspects of crop plant biology, such as gross morphology, primary mode and mechanism of reproduction, transmission genetics, physiology, and interactions with biotic and abiotic stresses, plant breeding has molded germ plasm for increased productivity. Important modifications in crop form and function have included reductions in the anthesis-silk emergence interval and barrenness in maize (Duvick, 1984), partition of photoassimilate in small grains (Austin, 1994), and incorporation of resistance to biotic and abiotic stresses in rice (Khush, 1993). Whereas the changes have been well documented at the phenotypic level, very little is known about the response to selection at the genotypic level and less about its biological basis. If crop genetics is expected to substitute for other technologies and become a more prominent source of productivity gains, expanded knowledge of crop biology will be prerequisite. Critical developments for material infrastructure have included plot combines, planters, computers, near-infrared reflectance analyzers, and off-season nurseries. These innovations have permitted increased sampling of the genetic and target environments, production of more sexual generations per year, faster cycle times, and more opportunities for selection and development of other methods to increase the ratio of genetic to environmental variation. Collectively, these practices have allowed plant breeders to assess the merits of myriad genotypes in a very comprehensive manner. The value of these innovations cannot be underestimated because plant breeding remains a numbers game in which time is of the essence for growers, industry, and consumers. Can the present rates of gain be maintained or improved? Sinclair (1993) and others have argued that only marginal opportunities remain for genetic improvements in crop yield potential. Instead, greater benefits may be realized through genetic improvements in crop tolerance to abiotic and biotic stresses. Thus, target environments characterized by stress-related yield losses or limits might realize improved rates of gain. What are the sources and costs of future gains? Certainly these will vary with the crop and production system. Expected sources of increased rice productivity include enhanced yield potential, heterosis, modified plant types, improved yield stability, gene pyramiding, and exotic and transgenic germ plasm (Khush, 1993).

268

M. LEE

The results of a survey presented by Phillips (1983) projected that U.S. maize yields through the year 2000 would continue to increase largely due to conventional plant breeding (1 bushel/acre/year) and emerging biotechnologies (1.7 bushels/acre/year by the year 2000). Contributions from the former component have been realized, while those of the latter have lagged behind prediction in the U.S. and other regions. This situation does not condemn new biotechnologies; rather, it illustrates the difficulty of predicting results on the basis of very meager experience and understanding of crop biology. Perhaps when accompanied by improved knowledge of plant biology, plant biotechnology shall become a significant source of genetic gain. One thing is certain: the law of diminishing returns looms in the future so that the cost per gain will be higher.

c. DNA MARKERS AS FUNDAMENTAL LINKSBETWEEN PLANT BREEDINGAND

PLANTBIOLOGY

Much, if not most, of the success of plant breeding has been accomplished in a virtual vacuum of basic knowledge of plant biology. Obviously, information was available, but the vast majority of it was either irrelevant or incapable of being incorporated into breeding programs. Today, most important biological phenomena utilized by plant breeding programs (e.g., heterosis, epistasis, host-pest interaction, response to abiotic stress) are described in abstract or anecdotal concepts. Obviously, the dearth of facts has not prohibited progress in or resourcefulness of plant breeding during the preceding decades. But in an era of heightened expectations for crop genetics, the demand should increase for firm data and more complete understanding of genetic gains. Basic plant biology will be the source of much new information about genomes, genes, pathways, and interactions of direct relevance to crop improvement. In many instances, DNA markers will be the vital link between a nascent tributary and mainstream plant breeding. The topics contained herein emphasize the realized and potential utility of nuclear DNA polymorphisms and their detection for plant breeding programs of annual crops. Whereas many of the principles and inferences should apply as well to genetic improvement of long-lived perennial species, a number of important differences in tree biology and breeding methodology justify a separate review. Nuclear DNA polymorphisms have been considered solely in this review; however, one should not overlook the utility of organellar DNA polymorphisms for crop improvement programs (Douches et al., 1991; Steinborn et al., 1992; Grabau et al., 1992; Havey, 1993; Rajeshwari et al., 1994; Lorenze et al., 1994; Fauron and Casper, 1994). This subject also merits separate consideration. Various aspects of DNA markers and crop improvement have been reviewed in detail since 1991. Those specific reviews have been cited, updated, and comple-

DNA MARKERS AND PLANT BREEDING PROGRAMS

269

mented throughout this text as needed. At least two related books have also been published (Beckmann and Osborn, 1992; Phillips and Vasil, 1994). These books consist of independent and specialized chapters summarizing methods, applications, and maps of specific crops. The last comprehensive review of the status of DNA markers in crop improvement was prepared in 1991 (Paterson et d.,1991a). At that time, many of the applications eloquently conceptualized and described in that treatise awaited empirical evaluation and maturation. Much has happened since 1991, but it has not been considered from the holistic perspective of plant breeding. With that in mind, the present review was prepared.

11. ASSESSING GENETIC DIVERSITY AND MERIT One consequence of modem agricultural practices, which generally emphasizes maximum productivity with acceptable quality and uniformity, has been a reduction in the genetic diversity of the primary gene pool under cultivation, with similar fates for the secondary and tertiary gene pools of most major crops. Even though the extent of the reduction may be largely unquantifiable, it is generally assumed that valuable and irreplaceable genes have been lost or ignored, that plant genetic resources have been shrinking at accelerated rates, and that cropbased agriculture has become more vulnerable to the vagaries of climate and associated biotic and abiotic stresses. Undoubtedly there is considerable merit, validity, and controversy associated with each point. Facts and anecdotes aside, the consequences of a narrow genetic base of major crops have been experienced sporadically throughout history, often with significant human and economic costs. Therefore, an awareness of genetic diversity and management of crop genetic resources have been important components of plant improvement programs. The foundation of crop-based agriculture largely rests on the availability and knowledge of extant plant genetic resources in germ plasm collections and at successive stages of development in breeding programs. Specifically, knowledge of genetic diversity and relationships among sets of germ plasm and the potential merit of the genetic diversity would be beneficial to all phases of crop improvement. For example, assumptions regarding the distribution of genetic diversity among samples of crop germ plasm and relatives have been primary concerns for conservation and introgression programs. Such concerns have been also requisite for the efficient search for the elusive unique and favorable allele(s) by plant breeders. Assessments of the genetic composition of crop germ plasm and relatives have been, for the most part, conducted on the basis of a carefully developed rationale

270

M. LEE

and methodology, but limited genetic information derived directly from the plant. Historically, inferences have been based on reproductive biology, ecogeographic data, morphology, pigmentation, ontogeny, social history, pedigree records, breeding behavior, in siru and ex situ evaluation of agricultural traits, chromosome structure and behavior, and protein markers (storage proteins, isozymes, and other anonymous proteins), among others. Each perspective has provided valuable information and, depending on one’s objectives, might be sufficient. However, the progression of perspectives and the continuous preoccupation with issues pertaining to genetic diversity suggest that there is opportunity, and need, for improvement. Among the many facets of DNA marker technology, DNA fingerprinting has been the most pervasive application. The attractions of DNA fingerprinting have been its increased power of resolution and the potential for absolute objectivity. Perspectives provided by DNA fingerprinting have constituted critical evidence in forensics and paternity cases and population maintenance of endangered animal species, and DNA fingerprinting has been proposed as a means of voluntary genetic testing of humans for mutations related to various conditions (Nowak, 1994).

Parallel applications of DNA fingerprinting for plant improvement have been forecast (Soller and Beckmann, 1983; Tanksley, 1983). Subsequently, there has been considerable collection of preliminary survey data and, to a lesser degree, exploratory investigations into the utility of DNA markers for characterizing the genetic diversity and composition of crop germ plasm (Smith and Smith, 1992). In this section, I shall attempt to survey the status of selected applications of DNA fingerprinting for activities of common interest to plant breeding programs.

A. GERMPLASMIDENTIFICATION, CLASSIFICATION, AND MANAGEMENT Virtually every assemblage of crop germ plasm has been characterized by a system of descriptors and subsequently organized into categories. Regardless of the (de)merits and myths of each system, the resulting order has served a useful purpose in crop improvement by providing organization, structure, standards, context, and direction. These systems have been influential in activities important to the complete spectrum of crop improvement-from sampling strategies of germ plasm collections through registration of cultivars. Given the significance and scope of these activities, and the often ambiguous genetic foundations of the organizational systems, periodic reviews and revisions have been provoked by advances that promise to elucidate and clarify. Such is the situation for DNA markers and plant genetic resources.

DNA MARKERS AND PLANT BREEDING PROGRAMS

271

-

1. Germ Plasm Collections The Gene Banks

The prospect of utilizing DNA marker technology for managing germ plasm collections has been the subject of at least two comprehensive reviews (Kresovich and McFerson, 1992; Bretting and Widrlechner, 1995). Germ plasm management is a multifaceted endeavor involving acquisition, maintenance, and characterization such that the plant genetic resources are conserved and utilized for crop improvement. In the long term, the maintenance of collections probably deserves our greatest attention as the number of accessions and difficulties of preserving in situ reserves has increased for most crops, whereas the risks to be managed have remained ambiguous, unpredictable, and very serious. Maintenance is likely to become more difficult as the financial cost of maintaining collections, especially collections of large, long-lived, perennial species, increases. However, the primary concerns of plant breeding programs involve issues of greater significance in the near term. Those of most immediate interest to plant breeding programs involve knowledge of the current genetic content of the collections-acquisition and distribution of genetic diversity among accessions, relationships of collections (new and old) to elite germ plasm, and characterization of their potential genetic merit. There are a number of ways in which DNA markers could improve the management of plant genetic resources for the benefit of plant breeding programs and, ultimately, crop improvement. One of the constant tasks of germ plasm managers has been to assess the degree to which a collection’s gene pool overlaps with nature (estimates) or other collections. Traditionally, this has been accomplished mostly on the basis of morphological variations in concert with ecogeographic information. How effective have these methods been at complementing the genetic diversity of extant collections? Critical data are lacking on this point for many crops but the process obviously becomes more difficult with larger collections, presumably with fewer and smaller gaps. Surveys of germ plasm collections with DNA markers have revealed ecogeographic distributional patterns of “genetic” variation that could be used to develop sampling strategies for curators and breeders of annual (Lubbers et al., 1991; Goffreda et al., 1992; Kresovich et al., 1992) and perennial crop species (Besse et al., 1994; Laurent er al., 1994). With the increased availability of DNA sequence data with connections to known functions and advancements in technology, it may soon be possible to conduct molecular assessments of diversity among large samples of germ plasm. When related to reference or core sets, these data could be helpful guides for conducting data-driven acquisitions from geographic regions and other collections more likely to contain unique variants, much in the same way that isozymes have been used (Rick, 1979; Marshall, 1990). Also, as the size of a collection increases, it becomes more difficult to avoid

272

M. LEE

the inclusion of duplicate or at least very similar accessions. This situation gradually becomes a considerable waste of resources for maintenance, evaluation, and enhancement. Evaluation of numerous, highly similar accessions not only wastes plant breeding resources but likely reduces the chance of identifying the truly unique and valuable accessions. For many traits, wholesale surveys may be impractical because of either the expense of evaluation or adaptation of the accessions (Edwards, 1992). Therefore, additional sources of information upon which to base sampling strategies might increase the efficiency and effectiveness of the search and, ultimately, the utilization of germ plasm collections by plant breeding programs. The genetic basis and potential merit of an accession’s phenotype have been critical considerations for its utilization in breeding programs, as novelty alone may not suffice in many situations. This aspect has become especially acute when the level of performance and quality between elite germ plasm and accessions in collections have differed considerably. In these situations, introgression of the trait into elite germ plasm usually has been considered a long-term, risky, and perhaps unreasonable venture depending on the genetic complexity of the trait (Edwards, 1992). Empirical methods for identifying sources of favorable alleles have been developed (Dudley, 1987; Gerloff and Smith, 1988), but they do not provide information for a priori selection of accessions. Others (Beer et al., 1993) have suggested that the initial selection of accessions should be based on the trait@)of interest with subsequent choices based on DNA marker data to maximize the chance of recovering transgressive segregants. This suggestion may have some merit as plant breeding studies and DNA markers have indicated that the exotic source parents have contributed positive factors for the target trait as well as nontarget traits in maize (Lee et al., 1990), tomato (de Vicente and Tanksley, 1993), and oats (Lawrence and Frey, 1976; M. Lee, unpublished data). DNA marker technology has not yet made a significant impact on the management and utilization of germ plasm collections. At this point in time, it is important to note that any suggestion for the use of DNA markers in these areas has emphasized caution and complementation of established methods. However, given the magnitude of the issue for crop improvement, it must be viewed as a situation ripe with opportunities and substantial benefits for genetic gain and certainly tempting enough for further investigation. 2. Elite Germ Plasm-Hard Currency

Assessments of the genetic diversity of elite crop germ plasm have been sought and used by plant breeders for numerous reasons-genetic relationships, parent selection, germ plasm management and sampling, and germ plasm protec-

DNA MARKERS AND PLANT BREEDING PROGRAMS

273

tion, among others. Initial investigations have established that DNA markers provide superior discriminatory power relative to protein and morphological markers (Smith and Smith, 1992), with a few exceptions such as muskmelon (Neuhausen, 1992). The availability of a repertoire of methods for detecting DNA polymorphisms suggests that they could be advantageous for allogamous as well as autogamous annual crops that have more restricted gene pools, such as soybean (Akkaya et al., 1992), wheat (He et al., 1992), and tomato (Williams and St. Clair, 1993), and especially for perennial and other crops with far fewer options for assessing genetic diversity, such as bananas and plantain (Howell et al., 1994), cacao (Wilde er al., 1992), poplar (Castiglione et al., 1993), cranberry (Novy et al., 1994), sweet potato (Connolly et al., 1994), and papaya (Stiles et al., 1993). The actual and potential advantages of DNA markers have stimulated review and revision of long-held methods used to assess genetic diversity and relationships at several stages of crop improvement-parent selection (Sections 1I.B and IV.B), progeny selection (Sections 1V.B and V), and cultivar identification. This review will address the former two stages, as the third has been the subject of a comprehensive review (Smith and Smith, 1992). No doubt, many of the established methods will emerge essentially unscathed because they provide adequate efficiency and effectiveness for plant breeding programs. Likewise, a few methods and their supporting concepts will be modified, cosmetically or substantially, because DNA markers have contributed previously unavailable insight and information. How do plant breeders assess genetic diversity and relationships among elite germ plasm for the purpose of genetic gain? Many of the methods used by germ plasm managers have been used by plant breeders (Section 1I.A). In addition, plant breeders often have access to pedigree information, performance records (e.g., combining ability, progeny evaluation, selection, and breeding history), and inferences gleaned from various mating designs (Dudley, 1987;Troyer et al., 1988). The strength of some of the methods is that they are often based on direct assessments of what the breeder needs to know about the germ plasm. Such methods will be extremely difficult to improve. However, some methods and concepts have relied on weak genetic foundations, if any. For some plant breeding practices, that may constitute a weakness that reduces their efficiency. At least some of these deficiencies may be satisfied in part by DNA markers. One of the most pervasive measures of genetic relationships in elite crop germ plasm has been Maltcot’s (1948) coefficient of coancestry cf), which provides an estimate of the degree of genetic similarity between two individuals. This measure estimates the probability that two randomly drawn, homologous genes (alleles) from each of two individuals are identical by descent. The measure has been based on Mendelian inheritance and probability and has been calculated

2 74

M. LEE

under several assumptions: (1) absence of selection, mutation, migration, and drift; (2) regular diploid meiosis; and (3) no relationship (f = 0) for individuals without verified common ancestors (Melchinger, 1993). Several common features of plant breeding programs have represented departures from these assumptions: (1) intense selection; (2) drift due to small sample sizes; (3) irregular nondiploid meiosis for some crops; and (4) unknown or incorrect pedigree records (Bernardo, 1993; Messmer et al., 1993). Nevertheless, this method of estimating the degree of similarity will create information each generation and has been used widely by crop breeding programs (Martin et al., 1991). Several studies have evaluated and compared estimates of genetic diversity based on the coancestry coefficient and DNA markers (primarily RFLR). The studies unanimously concluded that DNA markers provide a more accurate portrayal of genetic diversity among sets of elite germ plasm of maize (Smith et al., 1990; Messmer et al., 1993),Brassica oleracea (Nienhuis et al., 1992),Brassica campestris (McGrath and Quiros, 1992), European wheat (Siedler et al., 1994), European barley (Graner et al., 1994), spring barley (Tinker et al., 1993), oats (O’Donoughue et al., 1994), sorghum (Ahnert et al., in press) and sweet corn (Gerdes and Tracy, 1994). In general, genetic distance measures based on DNA markers andfhave been positively correlated and, thus, have placed entries into the same general groupings. The value of DNA markers has been realized when entries may be too closely related, when ancestry has been obscured through generations of selection (Siedler et al., 1994), and when pedigree records have been inaccurate (Messmer et al., 1993; Graner et al., 1994). Even when pedigree records have been acceptable indicators of genetic relationships, DNA-based estimates have provided additional useful information (Smith et al., 1990). As with any method, DNA-based estimates of genetic diversity have an inherent potential for error and bias. With DNA markers, there has been justifiable concern about laboratory technique, standards, and data interpretation (Lander, 1989; Smith and Smith, 1992). With a few precautions, errors may be minimized or may be available for further evaluation. Estimates based on DNA markers have an upward bias for f (Cox et al., 1985), but the bias may be a significant problem only for natural populations (Lynch, 1988). Also, estimates of genetic similarity based on the number of DNA fragments in common between two individuals do not necessarily portray similarities based on common ancestry; the bands may merely reflect genes that are identical in state (i.e., alike in state) and not identical by descent (Smith and Smith, 1992). Despite these and perhaps other potential limitations, DNA markers have represented a significant improvement in plant breeders’ perception of genetic diversity. On the basis of the number of methods available for detecting DNA polymorphism and relatively comprehensive coverage of the genome, DNA markers have become a standard tool for this aspect of plant breeding programs.

DNA MARKERS AND PLANT BREEDING PROGRAMS

275

B. PARENTSELECTION Despite empirical and theoretical investigations, the development of methods for identifying, choosing, and using parents that routinely produce superior progeny and cultivars has been an unsuccessful struggle. The concept of progeny testing, reportedly first used by Louis De Vilmorin during the later stages of the 19th century, has served agriculture well (Briggs and Knowles, 1967). Certainly breeders have been routinely successful once they identify their core, elite germ plasm, but much trial and error and some good luck were probably encountered during the search for and development of good parents. Methods of parent selection may be divided into two broad categories, a priori (direct evaluation of the parents) and a posteriori (some form of progeny testing; Baenziger and Peterson, 1992). Plant breeding programs of annual crops have relied predominantly on the latter category, especially for the development of Fl hybrid cultivars. A priori methods have been used more commonly for simply inherited traits, by experienced breeders with core germ plasm, and for evaluating exotic germ plasm. However, the latter use should be reconsidered in light of theoretical examples (Knapp, 1994) and empirical studies in maize (Lee et al., 1990) and tomato (de Vicente and Tanksley, 1993), which demonstrate the recovery of unexpected, favorable alleles from donor parents, alleles masked by a priori evaluation but detected by DNA markers in segregating populations. In contrast, to plant breeding, animal breeding programs (e.g., dairy cattle) have benefited from methods such as the best linear unbiased predictor (BLUP) for parental evaluation and prediction of progeny performance. Extensions of that model have been proposed to include DNA markers to improve estimates of breeding values (Goddard, 1992). Bemardo (1994) suggested that BLUP, utilizing RFLP-based estimates of genetic relationships of maize inbred parents and performance data (grain yield of F1 progeny) of a subset of all hybrid combinations, could predict the performance of the much larger subset. Presumably, a similar rationale could be adapted and tested for other crops and traits. Given the time and expense of evaluating some traits, and the fate of most crosses, such clairvoyance would be welcomed by plant breeders. 1. Source (Base) Populations

The ideal characteristics of source populations may vary among crops, but a few features have been common to most breeding programs concerned with economic yield: (1) mean performance of progeny relative to target; (2) sufficient genetic variation to provide a basis for gain from selection; (3) inclusion of unique and favorable alleles to improve the elite gene pool; and (4) promotion of favorable combinations of extant positive alleles in elite gene pools. Given the

276

M. LEE

complexity and number of traits evaluated in breeding programs, a posteriori methods have a very secure future regarding their role in the identification of good source populations. However, a few theoretical and empirical studies have indicated ways in which DNA markers may facilitate a priori selection of parents for source populations. Previous studies with protein markers (isozymes or seed storage proteins) have provided a few clues about the utility of DNA markers for selecting parents for source populations, especially with regard to item (1) in the preceding paragraph. The pioneering report of the sufficiently tight genetic linkage between a gene for nematode resistance and an Aps isozyme allele in tomato (reviewed in MedinaFilho and Tanksley (1983)l has established that genetic markers could facilitate a priori parent and progeny selection. The utility of this approach depends, of course, on the frequency with which genes with highly qualitative effects have been used in breeding programs and the ability to find indicative DNA markers in elite germ plasm. Whereas the frequency of use of such genes may vary considerably among crops, it is certain to increase with the advent and widespread use of transgenic germ plasm. Evaluation of high-molecular-weight (HMW) glutenins and their role in breadbaking quality in wheat may provide a good case study for some applications of marker-assisted selection of parents and their progeny for several reasons: ( 1 ) direct evaluation of bread-baking quality is expensive and requires special facilities; (2) components of bread-baking quality have been established; (3) the time from planting to evaluation is considerable, especially for winter wheat (ca. 1 year); (4) the functional role of HMW glutenins in bread-baking quality has been established; ( 5 ) the transmission and molecular genetics of HMW glutenins have been well characterized; (6) bread-baking quality has been studied extensively with biometric methods and special chromosome stocks useful for partitioning its genetic (quantitative and qualitative), environmental, interactive, and correlative components of variation; (7) the quantitative genetic components of break-baking quality have low to intermediate heritabilities; (8) through much trial and error, several decades of breeding and selection have established a record of regular productivity; and (9) the complexity of the case may be at the simple end of the biological spectrum because the role of metabolism and biochemical pathways should be less prominent than for many other traits, even though it has several well-known components and is itself a component of the ultimate trait, grain yield (Sears and Cox, 1993). Under these conditions, wheat researchers have been motivated to develop methods of efficiently predicting baking quality and have the background and history needed for a meaningful analysis. HMW glutenins have been used to aid the selection of parents and progeny, but their use has varied with the breeding program and objectives. In cultivar development programs, their use has been routine when levels of quality and

DNA MARKERS AND PLANT BREEDING PROGRAMS

277

genetic variation for quality in breeding populations have been low and high, respectively (Payne, 1987). However, as favorable alleles have approached fixation, the utility of the markers has diminished. Although they are still used, breeders in these circumstances have maintained and enhanced quality predominantly through reliance on a core of elite germ plasm and more direct tests of advanced generation progeny (Baenziger and Peterson, 1992; Sears and Cox, 1993). Introgression programs have utilized knowledge of allelic variation of HMW glutenins to select parents (Lubbers et al., 1991) and have proposed using the markers to identify progeny that carry unique and favorable alleles for quality (Sears and Cox, 1993). Absolute reliance on HMW glutenins as indirect predictors of quality has not been advised for several reasons. First of all, when all of the favorable glutenin and gliadin alleles were surveyed and assessed in a similar genetic background, they accounted for less than half of the genetic variation in baking quality in several studies [cited in Sears and Cox (1993)l. Mansur et al. (1990) reported that chromosomes 3A, 3B, and 7B have significant effects on baking quality, even though they are not known to carry expressed genes for gliadin and glutenin proteins. Perhaps these chromosomes carry modifier genes that could be monitored by DNA markers to account for a higher proportion of the genetic variation. Also, epistatic interactions and differential effects on various aspects of baking quality among alleles may limit the efficacy of marker-based selection (Dong et al., 1991, 1992). Finally, quantitative as well as qualitative differences in specific HMW glutenin subunits have been reported to have important effects on quality (Halford et al., 1992). Thus, the presence or absence of an allele, as well as its level of expression and interactions, must be considered. This brief case analysis does not necessarily condemn the use of markers as indirect selection criteria. Instead, it suggests strategies for reasonable deployment and expectations. The markers are likely to be relatively efficient in more divergent crosses, and as the parents become more elite (spectrum), their utility might decline for some objectives. These expectations are in good agreement with a simulation study that reported greater efficiency of marker-assisted selection for coupling phase linkages (Gimelfarb and Lande, 1994). However, as the biological basis of the traits is revealed and the marked alleles acquire a breeding history, it should be possible to devise strategies and assume reasonable expectations for this approach. In addition to prediction of mean performance, breeders have sought methods of parent selection that provide assurance of sufficient genetic variance within populations for the traits of interest. For cultivar development, the overall goal typically has been to identify progeny that maintain performance standards for most traits and exceed levels of the target cultivar(s) for a few others. Given the unknown nature of the underlying crop biology, and the errors inherent in estimating genetic variance, the complexity of this challenge cannot be overesti-

278

M.LEE

mated and may be beyond the realm of predictive methods for the near future. However, the potential benefits of such methods would be substantial and have been sufficient motivation for undertaking this potentially quixotic pursuit. The essential history of and salient issues for developing such predictive plant breeding methods have been reviewed in several empirical studies involving small grains (Cowen and Frey, 1987a,b; Cox and Murphy, 1990; Souza and Sorrells, 1991). These studies evaluated several estimates of genetic distance or divergence on the basis of coefficient of parentage, variation of traits among parents and their progeny, and heterosis for association with performance and variation in segregating generations. Estimates of genetic distance based on coefficient of parentage have been most closely related with estimates of genetic variance in the populations in two studies (Cowen and Frey, 1987a; Souza and Sorrells, 1991), although the relationship has been too weak to be of predictive value. In each study, the authors recommended combining distance measures as the most effective approach. There have been few reported attempts of using DNA-marker-based estimates of genetic relationships among parents for predicting genetic variation among their progeny. In a preliminary study using RFLPs in oats, Moser and Lee (1994) observed that genealogical and RFLP-based estimates of genetic distance were correlated positively with each other, but they were not correlated with distance estimates based on quantitative traits. A few weak correlations between genetic variances and only two estimators (genealogical and RFLP-based) were observed for two of six traits. In this sample of oat germ plasm, the estimators did not have any utility, but the authors suggested that the utility of DNA-based marker predictions of genetic variation could be increased when (1) DNA marker alleles are sufficiently linked to well-characterized quantitative trait loci (QTL) and when (2) inferences are limited to well-defined genetic reference populations. The desire to recover favorable transgressive segregants has been among the primary motivations for creating populations with adequate genetic variation. The results of several QTL mapping studies have indicated that transgressive segregation may be due, in part, to the accumulation of complementary alleles (Tanksley, 1993; Schon et al., 1993; Veldboom et al., 1994a,b). In each study, most of the desirable alleles at QTL for a given trait were derived from the expected parent: the parent’s phenotype has been a good predictor of the source of most of the favorable alleles. However, the other parent has contributed alleles with positive effects at other loci, such that the phenotypes of segregants carrying combinations of complementary factors exceed those of the parents. Similar results have been obtained with marker-assisted backcrossing and selection for grain yield in maize (Stuber, 1994b; Section V.B). Presumably, these combinations might produce similar patterns of segregation in closely related populations (perhaps using the same donor parent and a relative of the first recurrent parent). This could provide a basis for a priori selection of parents, progress toward

DNA MARKERS AND PLANT BREEDING PROGRAMS

2 79

predictions of genetic variation, and suggestions for combinations of favorable alleles in similar genetic backgrounds. The search for favorable alleles has motivated the development of several methods designed to (1) identify carriers of unique and favorable alleles relative to a defined genetic reference population and (2) suggest approaches for incorporating the germ plasm regarding breeding method and recipient parent(s) (Dudley, 1987; Gerloff and Smith, 1988;Troyer et al., 1988). Empirical evaluation of Dudley’s method@)has indicated that donor germ plasm may be ranked with sufficient accuracy for parent selection. Subsequently,the method has been used to identify maize synthetic population BS 11(FR)C7 as a donor of alleles to improve inbred Mo17 as a parent of Fl hybrid cultivars with inbred B73 (Dudley, 1988). Because the method does not provide estimates of gene number and location, RFLP markers have been used to provide that information as a basis of marker-assisted enhancement of Mo17 (Zehr et al., 1992). 2. Hybrid Combinations

Breeders of hybrid cultivars have often used a predominantly a priori method of parent selection for hybrid combinations: heterotic groups. A heterotic group is a collection of germ plasm that, when crossed to germ plasm external to its group (usually another heterotic group), tends to exhibit a higher degree of heterosis (on average) than when crossed to a member of its own group. Of course, important exceptions to this tendency have been observed, but the concept has provided a simplifyingand convenient tool for germ plasm management and utilization for activities such as the maintenance of core germ plasm, creation of source populations, selection of testers for identifying hybrid combinations, and classification of exotic germ plasm. Maize breeders have made the greatest use of this concept in the development of F1 hybrid cultivars from inbred lines. The establishmentof heterotic groups in maize has been a somewhat haphazard process requiring decades to evolve in U.S.breeding programs. Without the benefit of systematic breeding methods, many combinations of parents were tested until dominant pedigrees emerged to establish an identity for a heterotic group or a subset thereof (Hallauer et al., 1988). Whereas this approach has sufficed for maize and a few other crops with facile methods of pollen control and a growth habit conducive to testing many combinations relatively quickly (e.g., beets, sunflowers), other crops may benefit from alternative strategies. Conversion to hybrid cultivars has been a goal of many crop breeding programs convinced of the merits of heterosis, uniformity, and the economics of seed production. Often, the primary obstacle to developing hybrid cultivars has been a system for controlling pollen production (e.g., sorghum, wheat, rice, carrot). With some crops, notably trees (cacao, rubber, oil palm), the main

280

M.LEE

problems have been the time and space required to test the potentially large number of hybrid combinations and the lack of pedigree and performance data to help decrease the possible number of combinations for testing. However, once pollination control has been achieved, for nearly all crops researchers have been compelled to address the same questions regarding parent selection and crossing to find the best hybrid combinations: (1) How should we identify or establish heterotic groups? (2) What genotypes should we use as testers for hybrid performance? Relative to U.S. maize breeding in the 1920s, some of these crops have a few potential advantages: (1) decades of selection have produced a highly productive germ plasm base, and (2) pedigree records may be helpful in estimating genetic distances on the basis of alleles that are identical by descent. These resources (often lacking for tree crops), particularly item (2) supplemented with DNA marker data, could help breeders develop crossing schemes of maximal efficiency by avoiding crosses between closely related parents and by focusing on crosses likely to yield hybrid progeny having the desired degree and quality of heterozygosity. In some regards, the supplemental information provided by DNA markers could represent an effective substitution of information for time needed to mold heterotic groups for these crops. Crops with these circumstances also have disadvantages for maximizing potential benefits of heterosis: (1) several decades of breeding and many generations of selection may have produced highly productive, but genetically narrow, germ plasm (e.g., wheat and soybean), and (2) the heterotic group(s) used as a source of seed parents may have a limited and restricted genetic base due to the constraints imposed by the method of achieving pollen control in hybrid seed production fields (e.g., sorghum and rice). In these instances, breeders may need to develop strategies for incorporating genetic diversity into the seed parents and build heterotic groups, as proposed by Zhang et al. (1994) for rice. According to this strategy, DNA markers linked to photoperiod-sensitive, male sterility factors will be used to backcross them into new genetic backgrounds and expand the gene pool used as seed parents for hybrid cultivars. Presumably, this strategy could be adapted for crops with similar methods of pollen control. Several studies using RFLP-based estimates of genetic similarity among elite maize inbreds have demonstrated the utility of DNA markers for placing the lines into their respective heterotic groups (Lee et al., 1989; Melchinger et al., 1991; Dudley et al., 1991; Livini er al., 1992; Messmer et al., 1993). Similar observations have been made for a set of 105 elite U.S. sorghum inbreds (Ahnert et al., in press). One surprising feature of the maize studies has been the high degree of similarity between the heterotic groups (based strictly on the number of DNA fragments common to pairs of inbreds from each group). These assessments have suggested that the flint X dent heterotic groups used by European maize breeders may be slightly more divergent than the Reid Yellow Dent (represented by Iowa

DNA MARKERS AND PLANT BREEDING PROGRAMS

281

Stiff Stalk Synthetic) X Lancaster Sure Crop heterotic groups common to U.S. programs. These and related studies (e.g., Smith et a f . , 1991) have indicated that sufficient quantities of DNA marker data may have considerable utility for assigning lines to heterotic groups with a degree of precision adequate for breeding programs. An implicit purpose for establishing and using heterotic groups has been the desire to predict the performance of hybrids created by intergroup crosses. Such ability has been needed because parental performance per se has not been sufficiently correlated with the performance of its hybrid progeny for important traits (e.g., economic yield), and the possible number of hybrid combinations has exceeded the capacity of field testing. Initially, predictive methods based on the performance of single-cross hybrids were developed and then used for developing three-way and double-cross hybrids of maize and other crops (Sprague and Eberhart, 1977). The development of vigorous inbred lines permitted the emergence of commercial single-cross hybrids. Subsequently, breeders have become quite adept at creating and improving inbred lines and, by doing so, have gained considerable experience with certain inbred lineages and intergroup combinations. That experience and information have permitted maize breeders to select parents and predict intergroup cross performance with acceptable accuracy for the purposes of establishing source populations. However, significant challenges for predicting hybrid performance have remained because ( 1) progeny derived from source populations must be tested in intergroup combinations; (2) a heterotic group may have important and perhaps unperceived genetic substructure (Melchinger er a f . , 1991; Livini er a f . , 1992); and ( 3 ) inbred lines of new, uncharacterized, or mixed origin may not fit into established groups. The initial attempts of using protein and DNA markers for predicting hybrid performance in maize have been summarized by Stuber (1994a). Associations between hybrid performance and predicted heterozygosity have been stronger for crosses between lines of similar pedigrees (Frei er a f . , 1986; Lee et a f . , 1989; Smith et al., 1990). These observations may be indicative of the relative predictive powers provided by the knowledge of allelic states: identical by descent versus alike in state (i.e., identical in stare; Smith and Smith, 1992). However, the strength of the associations diminished at the upper ranges of predicted heterozygosity and genetic dissimilarity. Also, the corresponding associations for intergroup crosses and lines of unknown pedigree have been too weak to be of significant predictive value (Melchinger et al., 1990; Godshalk er a f . , 1990; Dudley et a f . , 1992; Boppenmaier et a f . , 1992). Thus, the utility of the markers for this aspect of maize breeding has been and shall remain limited until the validity of Bernardo’s proposals (1994) has been established for intergroup crosses. Nevertheless, the markers should provide additional useful information and guidance for the development of hybrid cultivars for crops lacking wellestablished heterotic groups.

282

M. LEE

3. Research Needs In general terms, the primary research need for improving our understanding and utilization of genetic diversity in plant breeding programs is to establish and strengthen the connection between molecular variations (e.g., qualitative and quantitative polymorphisms in proteins and nucleic acids) and variations in trait expression (genetic and environmental). Obviously, this is a formidable and perhaps quixotic undertaking because the underlying biology is largely unknown and certainly complex. However, plant breeders have a history of considerable achievement in the absence of detailed knowledge regarding basic biology; thus, we might expect that gains in the insight and utility of DNA markers could be achieved in stages. For example, breeding values might be assigned to DNA marker alleles for certain reference populations as a result of mapping studies and examinations of the distribution of alleles among sets of elite germ plasm. Some breeding schemes might be modified slightly to make use of pangenerational genetic information. Molecular characterization of genes might provide the basis for DNA-based surveys of allelic diversity of known important genes, thus focusing and facilitating certain aspects of germ plasm surveys. Of all the topics treated in this review, the understanding and use of genetic diversity is perhaps the most complex because it represents the climax of other, interdependent endeavors. Improvement in current methods of assessing and utilizing genetic diversity is likely to be one of the most difficult objectives to achieve because the improvements will be partially based on information derived from mapping experiments with their inherent vagaries (Sections IV and V). Also, some of the current methods are either so fundamental or so ingrained that the implications of indicated changes may be considered to be unacceptable risks and too controversial. For those reasons, as well as the complex nature of plant phenotypes and their interactions with the environment, plant breeders shall remain somewhat perplexed concerning assessments of genetic diversity and germ plasm selection for the foreseeable future. However, astute incorporation of information acquired from DNA markers will contribute to an improvement in this situation for many crops.

C. LIMITSOF ASSESSINGGENETICD ~ V E RVIA S ~DNA ~Y MARKERS

Any technique and derived information are subject to errors, mistakes, misinterpretation, and misconception. These issues have been the subject of considerable debate regarding the use of DNA markers in forensics (Lander, 1989) and genetic testing of humans (Nowak, 1994). Parallel discussions and concerns regarding similar applications in basic and applied plant sciences have been

DNA MARKERS AND PLANT BREEDLNG PROGRAMS

283

addressed in several reviews (Doebley and Wendel, 1989; Smith and Smith, 1992; Whitkus er al., 1994; Bretting and Widrlechner, 1995; among others) and are summarized briefly here: (1) What is known about the allelic relationships of the DNA fragments being used as data? Are they homologous? Orthologous? Paralogous? (2) What are the allelic frequencies in my reference populations? What are appropriate sample sizes of subjects and markers? (3) Under what conditions does a DNA polymorphism reflect a meaningful genetic difference? What are the relationships between genetic diversity detected by DNA markers and other types of genetic markers? (4) Do different DNA markers assess different portions of the genome? Does this represent a serious bias for genome sampling and data interpretation? ( 5 ) Has the target genome been saturated with DNA markers? What is the best way to determine whether the genome is saturated? (6) What are the limits of resolution of the method@)for detecting DNA polymorphism? What internal standards and checks were used in the analysis? This list of questions and issues is rather lengthy, albeit incomplete, because it does not include another galaxy of issues pertinent to assessing genetic diversity components and aspects of gene expression. Despite these concerns and limitations, DNA markers have provided unprecedented opportunities for genetic resolution, and therefore, their use will expand, not as a panacea, but as a complement to the existing methods and their inherent limitations.

III. GENOME ARCHITECTURE: GENETIC AND PHYSICAL CHARACTERIZATION OF CROP PLANT GENOMES Many of the limitations of important plant breeding methods have been rooted in the status of the technical infrastructure for conducting genetic analyses. Breeders and geneticists of all crops have lacked an informative and integrated genetic context to aid in the interpretation and conciliation of perspectives provided by seemingly different approaches to genetic improvement. The result has been a situation resembling the Tower of Babel, with breeders, geneticists, cytogeneticists, taxonomists, molecular biologists, plant pathologists, and other factions contributing to the confusion. A key component of the infrastructure and context of future plant breeding programs will be genetic maps. The maps, when fully integrated, will have several roles: (1) to provide a focal point and hub for data derived from the perspectives of myriad disciplines for each crop; (2) to constitute a vital two-way avenue connecting plant breeding and plant molecular biology; (3) to contribute essential information for the positional cloning of genes; (4) to facilitate considerable and directed expansion of a crop’s gene pool through comparative mapping of related and unrelated taxa; (5) to accelerate

2 84

M. LEE

identification and incorporation of useful genes into cultivars; and (6) to contribute important clues toward understanding the biological basis of complex traits and phenomena important to crop improvement. The significance of these and other roles and their implementations will vary with the repertoire of genetic technologies available to the crop, breeding methods and goals, and the nature of the crop’s nuclear genome. However, the foundation provided by the maps will have a positive impact on the genetic improvement of crop species in many instances. Various aspects of genetic map development have been presented previously (Paterson et al., 1991a) and in a book that includes summary reports of maps and their uses in several crops (Phillips and Vasil, 1994). This section of the present review will emphasize information from post-1991 literature and will use data from several crops to illustrate how maps have been used to reveal the features of crop genomes with implications for crop improvement.

A. DEVELOPMENT OF INTEGRATED MAPS Historically, genetic maps have provided very few advantages for plant breeding programs and crop improvement, even for species such as maize and tomato with relatively well-developed maps. The primary problems have been the types of markers predominantly used to create maps (macromutations and cytological markers), the existence of poorly integrated maps, each based on a different type of marker, the lack of informative markers in germ plasm used by breeders, and the polyploid nature of many crop genomes. The current and future generations of maps ameliorate these problems in significant ways. The advent of DNA markers has enhanced the relevance of genetic maps to plant breeding and improved the prospect of using linkage information as an important element in crop improvement schemes. The maps have provided a new source of information and raw materials (genes) for plant breeding, as well as an impetus for modifying some plant breeding methods. 1. Types of Markers

Integration of linkage information derived from various types of markers has significantly improved the resolution of crop genome architecture and created opportunities for improved interpretation of the genetic bases of crop improvement. Prior to the availability of DNA markers, maps (if they existed) provided few opportunities for identifying the sources of genetic variation in germ plasm and traits manipulated by plant breeders. Likewise, the maps based solely on macromutations (identified by alleles with highly qualitative effects) and cytological markers created very few opportunities for geneticists and molecular

DNA MARKERS AND PLANT BREEDING PROGRAMS

285

biologists to assess the agricultural significance of their favorite genes, much less devise directed strategies for locating and characterizing genes presumed to be highly significant. There is considerable effort being devoted to integrating crop genetic maps such that genetic information derived from DNA markers, macromutations, cytogenetic markers, and quantitative trait loci may be related to each other. This resource will provide, for the first time, the essential context for objective dialogues between plant breeding and many branches of basic science. One of the initial stages of recent map integration has involved cytological and DNA markers. When adequate cytogenetic stocks and manipulations have been available, their union with DNA markers has significantly enhanced our perception of genome architecture for crops such as tomato and potato (Tanksley et al., 1992), maize (Weber and Helentjaris, 1989), wheat (Werner er al., 1992), and barley and rye (Devos et al., 1993a). These investigations have revealed patterns of genome duplication, recombination, and cytogenetic-genetic distances along chromosomes. This information is essential for efficient deployment of a wide spectrum of genetic technologies, from targeted cloning of important genes through introgression of exotic germ plasm (Devos et al., 1993b). Another phase of map integration has involved DNA markers and macromutations. Reports of genetic linkage between DNA markers and macromutations have increased at a seemingly exponential rate. To the extent that such alleles have been used in breeding programs, these reports include a considerable array of expanded opportunities for using markers as indirect selection criteria. An area of potentially more pervasive significance for crop improvement has been the integrative mapping of partially sequenced cDNA clones in crops such as maize (e.g., Chao et al., 1994). Especially for macromutation-rich maps, this activity will provide many opportunities for matching mutants collected and characterized over several decades with molecules (across taxa). Eventually, this process will provide a basis for determining the biophysical bases of genetic variation and phenotypic expression for many traits. To the extent function and DNA sequence have been conserved across plant taxa (Helentjaris, 1993), integrated maps and their markers may represent a very important plant genetic resource for crop breeding. The third and least complete phase of integrative mapping involves polygenes or quantitathe trait loci (Tanksley, 1993). Despite the inherent ambiguities of the process, QTL mapping (SectionsIV and V) will provide vital informationfor basic and applied aspects of crop improvement. QTL mapping conducts a pangenomic assessment of gene location and action for potentially any phenotype (i.e., trait). Several aspects of QTL mapping make this approach especially powerful for adding important genes and regions to maps: (1) relatively comprehensive coverage of the genome provided by DNA markers; (2) provided adequate DNA polymorphism may be detected the choice of mapping parents may be extensive. Therefore, important genetic regions may be added to maps in a directed manner

286

M.LEE

depending on the ability to assess genetic diversity and variation for the trait($. (3) Unlike mapping macromutations, this approach may provide a more sensitive survey of the genome because it does not require fortuitous observation and recovery of alleles with highly qualitative effects on the trait(s) of interest. For any given crop species, what proportion of the loci has been identified by alleles with highly qualitative effects? By mutants? Analyses of nearly isogenic lines and subsequent QTL mapping in maize have identified a region with major effects on components of maturity (e.g., days to flowering and number of internodes; Kim et a f . , 1993; Phillips et a f . , 1992) and tissue culture response (Armstrong et a f . , 1992) for chromosomes 8 and 9, respectively. For these traits, the reports represent initial evidence of genetic factors on the chromosomes and characterization of their effects. Of course, the identified QTL usually are not located on the maps with the same precision and accuracy as other markers. However, the limitations created by the positional ambiguities may often be insignificant relative to the value of identifying the regions for the first time. The benefits of integrated maps have already been realized in a few investigations. Previous mapping with chromosome translocations has provided independent verification of QTL in maize for resistance to an insect (Schon ef al., 1993), a virus (McMullen and Louie, 1989), and a fungal pathogen (Freymark et a f . , 1993). Comparisons of genetic positions of QTL and loci defined by previously identified macromutations (Beavis et al., 1991) have supported the hypothesis that quantitative and qualitative genetic variation may often originate from alleles at common loci (Robertson, 1985). In maize, positive tests of allelism between a QTL and a macromutation for lateral branching (Doebley and Stec, 1994) and sequence analysis of alleles at the Sh2 locus have supported the hypothesis (Alrefai et al., 1994). In these situations, the supporting evidence was gathered in a very direct manner due to the availability of an integrated map.

2. Comparative Mapping The use of common sets of DNA probes to detect and map homologous sequences across sexually isolated species has revealed a surprisingly high degree of conservation in terms of copy number and homology of low copy probes, linkage, and locus order. Recognition of the considerable conservation of these features within sets of plants such as rice, wheat, and maize (Ahn et a f . , 1993), sorghum and maize (Pereira et a f . , 1994), wheat, barley, and rye (Devos et a f . , 1993a). tomato, potato, and pepper (Tanksley et a f . , 1988, 1992), and Arubidupsis and Brassica (Teutonic0 and Osborn, 1994) has inspired the suggestion of considering such groups as single genetic systems (Helentjaris, 1993; Bennetzen and Freeling, 1993). This concept should have considerable merit and mutual advantages for breeders and geneticists. Often, the genome size of one member of the group is

DNA MARKERS AND PLANT BREEDING PROGRAMS

287

severalfold smaller than other members. The smaller genome size should accelerate positional cloning of orthologous genes. Once the gene in the source species has been cloned and sequenced, this information may be used to quickly isolate the orthologous gene in the target species, as demonstrated by the isolation of the gene for chalcone flavonone isomerase in maize using sequence information from Petunia, snapdragon, and bean (Grotewold and Peterson, 1994). Also, the repertoire and number of mapped and characterized genes may vary considerably between members of a group (e.g., tomato versus potato, maize versus sorghum). In these instances, map information from the “gene-rich species may provide important clues about a map region’s genetic content in the “gene-poor” species and vice versa. Comparisons of locus order and distribution of recombination events may also elucidate barriers and suggest strategies to incorporate germ plasm in wide crosses (Devos et al., 1993b). For plant breeding programs, this information represents an opportunity for considerable, directed expansion and an improved definition of a crop’s gene pool. Comparative mapping with DNA clones has provided the basis for parallel investigations of other markers. For example, a region containing a locus that conditions the absence of ligules has been conserved among rice, wheat, and maize (Ahn et al., 1993). Similar inspections of linkage data of other taxa should reveal many other examples, such as the parallel linkage between genes for resistance to leaf rust (Puccinia spp.) and prolamines in oats, wheat, and maize (Rayapati et al., 1994a,b). The pattern of conserved linkage and function has been extended to include QTL. The initial report of orthologous QTL noted that the RFLP loci with the greatest effects on seed weight in mung bean and cowpea were detected by the same clones (Fatokun et al., 1992). In a similar manner, comparative mapping in maize and sorghum has revealed three putatively orthologous regions for plant height (Fig. 1; Pereira and Lee, 1995). In sorghum, each region has a major effect on that trait and on a unique suite of other traits (e.g., tillering, panicle dimensions, and leaf length and width), much like some of the dw loci in sorghum. Interestingly, plant height mutants at maize genetic loci in related regions have pleiotropic effects on some of the same combinations of traits as the sorghum QTL and the candidate dw loci.

3. Benefits for Crop Improvement The benefits of comparative and integrated maps for plant breeding programs are substantial for the short and long term. These have been interspersed throughout the preceding section and may be summarized as follows: (1) To the extent macromutations are utilized by breeders, integrated maps increase opportunities for indirect selection methods. (2) The ability to share genetic information between sexually isolated species should accelerate the isolation of targeted genes. (3) The definition of crop gene pools should become broader and more precise

288

M. LEE Maize(1) A

Sorghum(A)

Sorghum(H)

Maize(9)

Short arm P I 0 10005

Short arm ISU152 ISU074

.

BNL3.06

-

ISU136

-

PI010005

ISU136

BNW.59 ISU116 UMC33 NPI236 UMC37

LongI arm

ISU116

ISU123

BNL3.06 ISUO88 i- d_ 3 UMCll4 ISUllO

- ISU032B

""4'......,....,.,, 1 UMC33

Long arm

Malze(6)

C

' ' dw2

- ISUO88 -

UMC114

Sorghum(E)

Short arm

PI02007 14

- PI010016

NPI280 pyl UMC62

dw4

1 1 0 cM

ISU140

Long arm

figure 1. Conserved regions of maize and sorghum genomes having the same RFLP loci (joined by broken lines). Numbers and letters in parentheses indicate maize chromosome numbers and sorghum linkage groups, respectively. The shaded areas are the confidence intervals (1 .O log unit as indicated by MAPMAKER-QTL) for plant height QTL. In A and B, the maize QTL were identified by Beavis etal. (1991), and in C, the maize QTL was identified by Veldboom et al. (1994). brl, a n ] , d ,and P y l are maize plant height loci defined by alleles with qualitative effects and located in that position of the maize genome (Maize Newsletter, 1992). dw2. dw3. and dw4 are sorghum plant height loci defined by alleles with qualitative effects and possibly correspond to sorghum plant height QTL. Positions of dw2, dw3, and dw4 are hypothetical (Pereira and Lee, 1995).

for specific genes. (4) Understanding of the biological basis of complex traits should improve by providing a common language to various branches of biology. ( 5 ) Important genes may be localized by a variety of increasingly complementary methods. (6) An element of objective hypothesis testing has become available for plant breeding. This approach, like others, has some limitations. One limitation may relate to the observation that 10-20% of the low copy DNA clones from one species seem to be specific, or at least much more homologous, to the source species (Fatokun et al.,1992, Pereira et al., 1994). That could represent a substantial number of genes. At least some of those genes might confer unique or neomorphic functions in the source species. Undoubtedly, there will be many examples of species-

DNA MARKERS AND PLANT BREEDING PROGRAMS

2 89

specific low copy sequences that turn out to be very important genes. Nevertheless, there is so much to learn about so many shared sequences that the approach is easily justified.

B. RELATINGGENETIC AND PHYSICAL DISTANCES IN CROPPLANT GENOMES The demand for specific knowledge of the ratio, DNA base pairs/centimorgan, has increased in several areas related to crop improvement. This information has been especially important for those interested in positional cloning, as a high ratio could mean the difference between a successful chromosome walk and a seemingly endless forced march. Knowledge of this ratio could be useful for introgression strategies as well. For example, a favorable chromosome position of the target gene (i.e., a relatively highly recombinogenic region) or a priori knowledge of patterns of the distribution of recombination might provide clues regarding the most efficacious approach. Investigations conducted solely with cytological techniques, with classical genetic markers, or nonplant systems (Meagher et al., 1988) have portended many of the recent observations in plants made with DNA markers and methods. Some examples include the following: ( I ) the ratio may vary 10,000-fold among different regions of the genome (Meagher er al., 1988); (2) the distribution of recombination may be very uneven and seemingly nonrandom along the cytological length of chromosomes, with a tendency to be higher in distal regions and lower in centromeric regions (Lukaszewski and Curtis, 1993); (3) recombination rates may vary considerably with genetic background (Lukaszewski, 1992), between related taxa (Rick, 1969), and with the sexual origin of the gamete (Phillips, 1969); and (4) the rate of recombination seems to be much higher in expressed genes (Thurieaux, 1977; Dooner, 1994). Subsequent refinement of our perception of the relationship between physical and genetic distances with DNA markers has revealed features of chromosome organization of direct consequence for gene isolation and manipulation. 1. Macroscale

Integration of RFLP and cytogenetic maps has substantially increased the resolution of gross chromosome structure as exemplified by analyses in wheat (Werner er al., 1992; Ogihara er al., 1994), tomato (Tanksley er al., 1992), and maize (Weber and Helentjaris, 1989). Not only have the cytogenetic methods provided a rapid, and often the sole, means of locating DNA marker loci (e.g., monomorphic bands), they have been a primary source of information for fine mapping important genes as a prelude to map-based cloning (Gill er al., 1993).

290

M. LEE

Refinements in complementary techniques for chromosome-scalephysical detection and mapping of unique, low copy sequences using variations of in siru hybridization (hitch and HeslopHarrison, 1993) should offer additional power for this approach to gene isolation and resolution of chromosome form and function for some crop spp. To what extent has macroscale cytogenetics facilitated plant breeding and cultivar development over the preceding decades? Of course, the response depends somewhat on the nature of the crop’s genome, stage and method of breeding, and objectives for selection. In crops with small, numerous, and undefined chromosomes, cytogenetic analyses and manipulations may be precluded entirely. Breeding programs that rely on interspecific and intergeneric crosses have often benefited from cytogenetic techniques. Also, the contributions of cytogenetics to basic research have been invaluable. However, once the genes have entered the elite gene pool on their way to cultivar development, the role of cytogenetics has often diminished. However, coincident developments in several areas have raised the prospect of introducing cytogenetics into the foundation of cultivar development in several crops, as well as strengthening its role where it has been used more routinely, such as in wheat (Schwarzacheret al., 1992). Advancements in in siru hybridization should improve capabilities for physical mapping of low copy sequences. This technology would provide useful information regarding the cytogenetic positions of native and nonnative (transgenic) DNA markers, as well as the status of adjacent chromatin (eu- or hetero-). Some of this information may be inferred from extant maps, but direct inspection may be preferred in many situations where the genetic map has not been saturated or integrated. Why is cytogenetic position of potential importance? For map-based cloning, the position of the target locus relative to the centromere and heterochromatin may be an important determinant of the probability of success because of their well-known effects on recombination (Tanksley et al., 1992; Carland and Staskewicz, 1993). Also, variations in the expression and stability of transgenes may be due, in part, to position effects attributable to heterochromatic regions. Even if transgene expression has been determined to be adequate, insertion into regions known to be recombination “hot spots” might facilitate their subsequent sexual transfer among elite cultivars by reducing linkage drag. The same rationale may be applied to genes introduced through wide crosses. In some species, cytogenetic analysis might facilitate the identification and selection of more favorable transgenic events or at least eliminate those likely to be problematic in the future. In these situations, cytogenetics may have an expanded role in and substantial impact on cultivar development and genetic gain by allowing a quick and comprehensive overview of genome organization and location of important DNA markers.

DNA MARKERS AND PLANT BREEDING PROGRAMS

291

2. Microscale DNA markers have permitted localized assessments of genome organization and dynamics, with important implications for all phases of crop improvement. Information derived from investigations of interspersion patterns of sequence classes, the ratio of physical and genetic distances, and fine mapping has a bearing on the prospects for chromosome walks, the nature and putative products of recombination, and the sources of genetic variation. Detailed physical and genetic studies of specific regions have been reported for several crops and revealed the following estimates of physicaVgenetic distance ratios (number of base pairs per centimorgan): (1) tomato, Tm-2a (4 Mb; Ganal et a f . , 1989),12 (43 kb; Segal et a f . , 1992),Mi (>1 Mb; Messeguer et a f . ,199 l), Pro (200 kb; Martin et a f . , 1993),j t (200 kb; Wing et a f . , 1994); (2) rice, Xu21 (270 kb; Ronald et a f . , 1992); (3) wheat, alpha-Amy-1 ( 1 Mb; Cheung et a f . , 1991);and (4) maize, bzl (43 kb for the region within the locus; Dooner, 1986), a1 (217 kb within the locus; Civardi er a f . , 1994), and the al-sh2 interval (1460 kb; Civardi et a f . , 1994). As expected, the ratio varies greatly and seems to have been an important factor in some of the initial direct isolations of genes from crops via chromosome walks. Likewise, these studies and related determinations of interspersion patterns of DNA classes (Springer et a f . , 1994) have predicted bleak prospects for chromosome walking in crops with large, complex genomes using methods currently available. These investigations have also elucidated patterns and perhaps bases for variation in genetic distances. In maize, molecular and genetic fine structure analysis of several loci (bz, wx, Adhl, a l , and R ) has indicated that the physical/genetic ratio, averaged over the entire genome, is approximately 100 times higher than those values for regions within genes [summarized in Dooner (1994)l. This pattern has been confirmed through direct genetic and molecular analysis of the a1 locus and the a1 -sh2 interval in maize (Civardi et al., 1994). Clearly, recombination occurs at much greater rates within genes and may even be restricted to genes in eukaryotes (Thurieaux, 1977).

C. INSIGHTSINTO RECOMBINATIONAND ITS ROLE IN

GENERATING GENETIC VARIATION

The ratio [(importance of recombination to plant breeding)/(understanding recombination and its consequences)] must be close to infinity. The pervasive importance of recombination may be indicated by a few questions: (1) What is the basic unit of inheritance in plant breeding programs? Is it the chromosome? Chromosome segment? Gene? Parts of genes? (2) What are the linkage blocks?

292

M. LEE

How are they established and maintained? (3) What are the effects, benefits, and detriments of enhancing opportunities for recombination? In elite germ plasm? In introgression programs (transgenes and wild germ plasm)? For QTL mapping (Section IV and Fig. 2)? (4) What are the restrictions on recombination for incorporating exotic germ plasm? ( 5 ) How does recombination influence genetic variation? Although a clear understanding of recombination mechanisms is still in the future, investigations with DNA markers have revealed features of the localized chromosome environment (structure and content) that suggest mechanisms with important implications for plant breeding and crop improvement. At this point, we have the equivalent of the “smoking gun”. Patterns of pangenomic and localized recombination frequencies and positional distributions have been observed during the compilation of DNA-based marker

c

umc37 an2.6/ /’

umc86A Flpre 2. Comparison of plant height QTL detection in maize on chromosome 1 in FZ3 and F6:, generations of a single-cross population of inbreds H99 (short) and Mo17 (tall). Names of DNA marker loci are on the left. In the F2:3generation a significant region (LOD score threshold of 2.0) was attributed to a single QTL, identified by umc37, with the genetic effect for increased plant height derived from Mo17. The region could not be further resolved with the FZ. generation. This region has been resolved into three independent QTL in the F6:, generation, identified by P1, umc37, and umc86A (significance levels 0.05, 0.01, and 0.001 denoted by *, **, and ***) on the.basis of evaluation in multiple QTL models. Genetic effects for increased plant height for all three QTL are derived from Mo17.

DNA MARKERS AND PLANT BREEDING PROGRAMS

293

maps. One important trend has been an inverse relation between recombination frequency and genetic divergence of mapping parents; genetic maps created with interspecific crosses have been smaller than those created with intraspecific crosses (Rick, 1969; McCouch et al., 1988; Gebhardt el al., 1991; Doebley and Stec, 1993; Rayapati et al., 1994a). Gross chromosomal structural polymorphisms may account for some of the reduced map distances, but at least for tomato maps, other sources must be affecting recombination or the recovery of the products of recombination. Also, as noted for the tomato maps, the overall distribution of recombination in interspecific and intraspecific crosses did not differ significantly (Tanksley et al., 1992), but that may not be the situation for other crops such as wheat and triticale (Lukaszewski, 1992). If those reports are representative of their respective crops, the general trend for their respective crops, that information has implications for introgression programs. By assuming that restrictions to recombination are prevalent at some level of genetic and taxonomic divergence (to be defined), and recombination between exotic and domestic chromosomes is desired, then additional opportunities for recombination should be provided either through larger sample sizes or random mating (Lonnquist, 1974). In other circumstances, one may wish to expedite introgression of a gene from a wild parent while minimizing linkage drag. In this situation, one strategy may be to select candidate exotic parents initially for their trait(s) and subsequently for evidence of prior hybridization events with the domestic species using DNA markers. Presumably, combinations of parents that exhibit the least genetic divergence would provide a more recombinogenic environment and, thus, potentially less linkage drag around the target gene(s) coming from the wild parent. Regional assessments of recombination in tomato have illustrated important consequences of and considerations for gene introgression. For example, the genetic distance around the Mi region varies nearly 8-fold (0.4-2.9 cM) in accordance with the source of the region: the segment containing the Mi region from Lycopersicon peruvianum, a wild relative of cultivated tomato, seems to suppress recombination (Tanksley et al., 1992). With this in mind, it should be interesting to monitor the fates of eurkaryotic and noneukaryotic transgenes (Ingelbrecht et al., 1994) and their influence on recombination in crop species. Evaluation of linkage drag around the Tm-2a locus revealed wide variation in retention of the donor parent chromosome after several generations of backcross breeding (Young and Tanksley, 1989b). In that set of germ plasm there does not seem to be a very strong relation between the number of backcross generations and the size of donor chromosome segment recovered in the backcross progeny. One explanation may be that region may have been recalcitrant to recombination in some pairs of donors and recurrent parents. Alternatively, the breeders were less fortunate with some backcross populations and missed the preferred recombinants every generation. If the former explanation has credence, it may be

294

M. LEE

possible to identify the genetic basis of novel mechanisms, resulting in restnctions to recombination. The indications of highly elevated rates of recombination within genes have important implications for generating genetic variation. In plant breeding programs, one of the major roles of recombination has been as a means of exchanging alleles at loci (or, perhaps, equivalent segments of chromosomes). Presumably, the exchanges have been complete in the sense that one allele (segment) has been substituted, in toto, for another. That may well be the prevalent product, but observations from fine structure analysis suggest the possibility of another type of exchange capable of producing “hybrid” alleles. Depending how the recombination process is resolved, one can imagine the derivation of new combinations of regulatory and structural regions of genes. The genetic effects of the resulting hybrid alleles could vary considerably depending upon the function of the gene product and the tolerance of it and the organism for such changes. Such a mechanism could be a very important source of genetic variation and may help account for some of the continuous genetic gains achieved in plant breeding programs that rely on a putatively narrow genetic base (Rasmusson, 1991; Hallauer, 1990). The additional resolution provided by DNA markers has increased our awareness of the prevalence of localized gene or sequence duplication in crop plants: examples include 27-kDa zein (Das and Messing, 1987), Rpl (Sudupak et al., 1993), rhm (Peterson, personal communication, 1994) in maize, Mla in barley (Wise and Ellingboe, 1985), pro (Carland and Staskawicz, 1993) and Cf-4 and Cf-9 in tomato (Baht-Kurti et al., 1994), seed storage proteins, and numerous other multigene families. In situations where careful inspection has been possible, localized sequence duplication has threatened to become a law of plant genome organization. Insights into the dynamics and perhaps origins of such regions have been obtained through analysis of the Rpl locus in maize (Sudupak et al., 1993). Allelic constitutions of flanking RFLP alleles in susceptible progeny have provided strong evidence indicating that unequal meiotic exchange (i.e., crossing over) occurs in this region and may generate alleles with new genetic effects. Such a mechanism could also affect the copy number of sequences, as demonstrated for RFLP loci in that region (Hong et al., 1993). To what extent should we expect to observe similar phenomena at other loci? Do plants have mechanisms for regulating such mechanisms and gene copy number (Flavell, 1994)? Do some regions, structures, or patterns of sequence organization have a greater propensity for unequal exchange and tolerance for its products? Finally, do such mechanisms contribute to genetic variation in plant breeding programs? Given the number of possibilities, the answer to the latter question must be affirmative. However, the more important considerations have to do with the merits of the variation and the frequency at which it is generated. Several studies

DNA MARKERS AND PLANT BREEDING PROGRAMS

295

of elite maize inbreds, exemplified by Bogenschutz and Russell (1986), have reported genetic changes in several traits (plant height, number of tassel branches, grain yield, number of days to flowering) after maintaining the lines for several generations by self-pollination. By starting with highly homozygous and homogeneous inbreds, the average values of one or several traits have exhibited a trend maintained over several generations, consistent with an overall decrease in plant vigor. The degree of change varies with the inbred background and the trait. The changes have been very gradual and have not been associated with the production of any obvious macromutations. This phenomenon has been reported for other elite maize inbreds maintained by one or several geographically dispersed programs (W. A. Russell, personal communication, 1994) and for the barley cultivar “Atlas” [reviewed in Allard (1960)l. The underlying mechanism(s) has (have) not been determined, but it has become clear that maize inbred lines have an inherent genetic error rate within their genomes. Perhaps a mechanism similar to that proposed for the generation of flax genotrophs (Walbot and Cullis, 1985) or variation at the Rpl locus may be a source of variation within these highly inbred lines.

D. -BASED

CLONINGCOMES OF AGE

Targeted isolation of plant genes based strictly on the map position of a phenotype has been strengthened substantially by the advent of DNA markers. With further technical advancements and adaptation to idiosyncrasies/challenges inherent to a crop and phenotype, numerous important genes should be isolated in a directed manner (Ellis et al., 1988). Genetic maps based on DNA markers have improved the efficiency of established approaches, such as transposon tagging (Briggs and Beavis, 1994), and provided the critical missing link between the target locus and large inserts of cloned DNA used for chromosome walking to the locus of interest. In plants with relatively small genomes, the triumvirate of DNA markers, large-scale DNA cloning, and production of transgenic plants has made it possible to clone genes in a directed manner. Together, these and other approaches to gene identification and isolation will gradually elucidate some of the biological complexities of important phenotypes and create unforeseen opportunities for their manipulation and utilization in plant breeding strategies. 1. Positional Cloning

The general steps for this procedure were presented in a previous review (Paterson et al., 1991). Following that scheme, positional cloning with yeast artificial chromosomes (YACs) was first used successfully in plants in the model

296

M.LEE

species Arabidopsis (Arondel et al., 1992) and subsequently in tomato to isolate genes for disease resistance (Martin et al., 1993). Certainly other genes will be isolated in this manner, and the rate of success should increase with advances in DNA cloning (e.g., bacterial artificial chromosomes, BACs; O’Connor et af., 1989), accelerated DNA sequencing, and plant transformation. Many well-defined loci are available for this approach, and probably more will be identified as new macromutations are induced and discovered and as genetic mapping of complex traits becomes more refined (Section IV). No doubt, most of the isolated genes will represent true revelations for plant biology, and some of them may even make measurable advancements to agricultural productivity. However, there are some very substantial limitations to this and similar approaches due to the technology, crop species, phenotype, nature of crop productivity, or a combination thereof.

2. Isolating Genes through Transposon Tagging Transposon tagging as a means of gene isolation in plants was first demonstrated in maize (Fedoroff et al., 1984). Subsequently, maize transposable element systems were modified and introduced into other plant species to facilitate gene tagging and isolation (Ellis et a f . , 1988). Comprehensive discussions of approaches and rationale have been the subject of review (Walbot, 1992). While the tagging process may not be subject to absolute control regarding the genetic location of inserts, the related problems may be reduced if a good genetic map is available for the reference population and region of interest. For example, elements of the Ac-Ds transposon system tend to move to linked sites in maize; thus, by assuming that the system behaves similarly in the new species (e.g., tomato; Healy et af.,1993), one may enhance the chances of tagging a locus by monitoring genetic linkage between it and the transposon. The utility of DNA markers for this purpose and related aspects of transposon tagging in maize have been reviewed (Briggs and Beavis, 1994) and exemplified by Schmidt et af. (1987) in cloning the opaque2 locus and by Chandler et a f . for the B locus ( 1989). The mapped DNA marker loci have been especially useful for monitoring linkage in maize because DNA probes of the elements often display homology with several fragments in the reference population. The markers help discriminate linked and unlinked fragments and perhaps identify those inserted into the target locus. Transposon tagging has been demonstrated as an effective means of isolating targeted genes in maize and other crops. In maize, examples include opaque-2, Hml (Gohal and Briggs, 1992),and@ (Schnableand Wise, 1994).Inothercrops, highlights have involved important resistance genes for tobacco and the N locus for tobacco mosaic virus (Whitham et af., 1994), flax and the L6 locus for the rust

DNA MARKERS AND PLANT BREEDING PROGRAMS

297

pathogen Melumsporu lini [J. G. Ellis cited in Moffat (1994)], and tomato and the Cf-9 locus for Cfadosporiumfulvum [J. D. G. Jones cited in Moffat (1994)]. Unlike chromosome walking, transposon tagging does not necessarily require a transformation system for proof via complementation (although that can be very efficient); rather, one usually relies on the recovery of wild-type revertants from putative transposon insertion mutant alleles. Also, the method may have some relative utility for crops of large genome size or when highly efficient transformation has not yet been established. Once a recognizable element has inserted into a locus of interest (clearly, the weakest aspect of this strategy), successful gene cloning may be expected with a reasonable degree of certainty. These two strategies of gene isolation have considerable power, but they also have several limitations some of which have been discussed previously (e.g., genome size, need for a transformation system, transposon insertional pattern and frequency, need for the introduction and development of a system for most species). In addition, the implementation of either approach has been limited largely to traits that may be assessed by simple and direct visual inspection and usually early very early in the life cycle of the plant. Such traits and their underlying genes are certainly important, but so are many others that may be difficult to approach solely with either strategy. Despite their limitations, these two approaches will certainly lead to the isolation of many interesting genes for plant biology and some useful genes for crop improvement. There is no doubt that the genes will compel revisions of plant biology textbooks, given the present level of ignorance and the rate at which genes may be isolated and their functions assessed. However, it is much more difficult to predict the impact on crop improvements for three reasons: the history of discovery, the complexity of plant biology, and the growing challenges of crop production. First, the history of inventions and technology development does not bode well regarding humanity’s ability to forecast and transfer discoveries along uncluttered linear paths toward progress (Rosenberg, 1994). Even with machines there have been too many unanticipated developments and unforeseen utilities (e.g., computers). Thus, the lessons of history have been, first, to create as many opportunities as possible as foundations for technology development and, second, to prepare our minds to recognize and seize them. The new genes will suffice largely for the first step. Given the second reason, the unperceived complexity of plant biology, the only certainties are surprise and more complexity. Even “simple” organisms such as the virus lambda (Ptashne, 1987) have evolved complex and finely tuned sets of controls of their lifestyles. What should we expect of plants? Many of the newly isolated plant genes will merely define some step or aspect of a pathway (e.g., metabolic or signal transduction), so that their deployment for crop im-

298

M. LEE

provement may require an understanding of other elements of pathways and their roles in the plant’s biology (Peterson, 1992; Strohman, 1994). Certainly, viable strategies based on single nonplant and plant genes (e.g., viral coat protein, insecticidal proteins, native and nonnative enzymes; Voelker et al., 1992) or small assemblages of genes (e.g., engineered male sterility for producing hybrid seed; Mariani er al., 1992) have been developed. Some of those approaches will prove to be truly effective and valuable by protecting the present production capacity and increasing the efficiency of certain aspects of production by substituting the expanded repertoire of crop genetics for other technologies. But how will these reductionist strategies contribute solutions to the pervasive and holistic challenges encountered by crop improvement programs in their production environments (the third reason)? Among the many challenges to consider, two are sufficiently universal: ( 1) increase the inherent production capacity of a crop (i.e., its yield potential) and (2) improve tolerance to abiotic stresses (water deficits, temperature extremes, and unfavorable soil chemistry) at various stages of the plant’s life cycle. The greatest strengths and weaknesses of plant breeding programs relate to the plant breeders having been forced to consider the complete life cycle of the plant in diverse production environments and to seek crop genotypes capable of achieving acceptable compromises. This struggle has resulted in some remarkable, albeit unplanned, manipulations of crop plants inherent physiology and morphology. For example, the harvest index of wheat and other small grain cultivars has increased from 30% to nearly 50% over the past 80 years of selection, whereas total biomass production has remained essentially constant (Austin, 1994). With temperate maize, less dramatic shifts have been observed for dry matter production, but substantial improvements have been realized for hybrid cultivars’ tolerance to heat and water stress (Russell, 1993) by reducing the tendency toward barrenness (i.e., no ear development) and decreasing the interval between anthesis and silk emergence under stress conditions. Obviously, crop plants have demonstrated remarkable plasticity and resiliency for some rather complex and highly synchronized biological processes. Given the number of challenges facing crop production, especially in developing countries (Fischer, 1993), insights into the underlying processes would be welcomed. Surely some enlightenment and perhaps genetic gain will be realized as the “new” genes are utilized by plant breeding programs.

3. Contributions to Crop Improvement Programs For the first time in 10,OOO years, crop breeders and biologists have the opportunity to peer into the structure, content, and dynamics of crop genomes. The situation is in many ways analogous to one’s first blind date: whereas a few very good clues were available before the actual encounter, the critical details are

DNA MARKERS AND PLANT BREEDING PROGRAMS

299

revealed, in part, through much closer examination (afforded by DNA markers and associated methods in the present context). Thus, perhaps the greatest contribution will be a greater appreciation of the elements and overall architecture of crop genomes. Ultimately, the improved understanding will compel scientists to reconsider long-standing concepts and dogma, which upon review shall emerge either unscathed and perhaps stronger or undergo serious modifications, if they remain at all. Plant breeding continues to enter the era of plant molecular biology. To do so and to realize maximum benefits, plant breeding programs must have access to genes confemng the traits of interest and must learn how the genes and genotypes were shaped by nature and artificial selection. The genes are needed not only for revealing the biological basis of genetic variation but as a direct means of genetic gain. It is clear that macro- and microscale genome analyses, within limits, will be able to supply at least some of the genes long manipulated by breeders as components of complex biological systems, plants. The predicted genetic gain afforded by single genes must have an exceptionally large standard error. Certainly single genes affecting macromutations have been and will continue to be used, but several decades of plant breeding history suggest that it would be naive to get too excited about any one gene solely for its potential to provide significant genetic improvement. Before many, if not most, of the isolated genes make important contributions to crop improvement, their signals for gene expression must be understood because the signals may be as important and as interesting as, or more so, the structural regions. For example, deposition of leaf cuticular waxes in maize and sorghum follows a reverse pattern with respect to leaf age; accumulation is heaviest on juvenile and adult leaves in maize and sorghum, respectively. Presumably very similar metabolic pathways must be involved in both crops, but timing and synchrony with the plant’s phase seems very different. Understanding the causes of such differences based on gene expression will be vital to devising and deploying crop improvement strategies in the future. The first step begins with the gene. In addition to gene isolation, investigations into genome architecture eventually may provide information sufficient to construct a component of the genome, a plant artificial chromosome (PAC). Admittedly, this is far in the future, but the elementary pieces are slowly being identified and characterized: telomeric sequences (Kilian and Kleinhofs, 1993), centromeric sequences (Alfenito and Bkhler, 1993), and origins of replication. Plants are capable of accommodating supernumerary chromosomes (e.g., B chromosomes) without the detrimental effects often realized with changes in dosage (Birchler, 1993), although gross aneuploidy has severe consequences for growth and development (Carlson, 1988). In maize, supernumerary chromosomes exhibit a highly heterochromatic appearance, and they are devoid of known active genes essential to the viability of the plant (Carlson, 1988). If the plant’s chromosome structure and function

3 00

M. LEE

and regulatory mechanisms may be understood sufficiently, gene and crop manipulation with PACs may be a reasonable and attractive consideration for the future.

IV. ANALYSIS OF COMPLEX TRAITS AND PHENOMENA Plant breeding has been a very successful endeavor despite limited basic understanding of several universal “black boxes” of fundamental importance to its methods and goals; genetic elements of quantitative inheritance patterns, G X E, heterosis, epistasis, and the genetic basis of response to artificial selection constitute a brief list. Presumably, a more comprehensive understanding of the biological basis of any of these topics for traits of interest to plant breeding programs would contribute positively and efficiently to genetic gain. There has been adequate motivation to use DNA markers in attempts to elucidate genetic aspects of those topics for numerous traits and crops. The theoretical basis and results of many initial investigations of those topics have been the subject of general (Paterson er al., 1991a; Tanksley, 1993) and more specific reviews on detecting QTL (Dudley, 1993; Knapp, 1994) and heterosis (Stuber, 1994a) and on analysis of genotype by environmental interaction in QTL analysis (Beavis and Keim, 1995) noted throughout this section. However, the advent of DNA markers has fostered a rather dynamic research environment in these areas. Thus, numerous nascent and significant reports have compelled me to append and complement portions of some of those reviews.

A. QUNTI-ITATIVEINHERITANCE PATTERNS Characterization of genetic elements that affect quantitative genetic variation (i.e., QTL) has considerable appeal to basic and applied research interests. Genetic mapping of QTL with DNA markers will provide an important complementary perspective on genetic variation and genome structure to those afforded by classical, cytological, and molecular approaches. Macromutations may not be tolerated by some genes, their products, or phenotypes because of their sensitive biophysical structures or vital roles in plant development. Therefore, those loci may not be accessible to certain means of genetic analysis. Such mutants may also be difficult to identify due to the need for special screening procedures and gene expression during later stages of plant development. With the emergence of genetic maps integrated across taxa, QTL mapping may expedite determinations of order and biological function to genomic regions of crops. As discussed in

DNA MARKERS AND PLANT BREEDING PROGRAMS

301

Section 111, such genome cross-referencing creates several opportunities for gene cloning strategies and investigations of genome evolution. Given the opportunistic nature of basic research, QTL mapping will certainty provide important information and contribute to advancements. The enthusiasm for QTL mapping for applied research is strongly associated with the hope of using genetically linked DNA markers as the means of indirect selection for genomic regions affecting quantitative traits in plant breeding programs. QTL mapping and DNA markers may also provide insights into facets of quantitative inheritance patterns and other complex processes and phenomena that have been exploited but not yet understood by plant breeders or anyone else.

1. Detecting and Locating Quantitative Trait Loci (QTL) Classical and cytogenetic markers have been used to locate QTL for several decades in crops such as maize and wheat. However, the advantages of DNA markers, improved resolution, coverage, and codominance, make them the method of choice for this pursuit of the genome. Development and assessment of methods for detecting and genetically mapping QTL have been particularly active areas: (1) population sampling strategies (Weller and Wyler, 1992; Darvasi and Soller, 1992; Schmitz et a f . , 1993; Wang and Paterson, 1994); (2) type of population and progeny (Dudley, 1992; Knott and Haley, 1992a;Carbonell et a f . , 1993; Moreno-Gonzalez, 1993; Haley et a f . , 1994); (3) determining the genetic location(s) of QTL (Knott and Haley, 1992b; Darvasi and Weller, 1992; Luo and Kearsey, 1992; Martinez and Curnow, 1992; van Ooijen, 1992; Jansen, 1992, 1993; Luo and Woolliams, 1993; Zeng, 1994; Jansen and Stam, 1994); (4) threshold levels for detecting QTL (Dudley, 1993;Rebai et a f . ,1994);( 5 ) missing data (Martinez and Cumow, 1994); and (6) estimating genetic effects (MorenoGonzalez, 1993; Carbonell et a f . , 1992; Darvasi e t a f . ,1993; Hoeschele and Van Raden, 1993;Hayashi and Ukai, 1994).A good summary and discussion of major issues for QTL mapping with DNA markers as related to plant breeding and genetics may be found in two reviews (Weller, 1992; Knapp, 1994). A common theme permeating the largely theoretical treatments cited in the preceeding paragraph is the uncertainty of a QTL's genetic location relative to DNA markers. In most circumstances, a QTL may be placed within a genetic region of 15-20 cM with an acceptable degree of certainty. Simulation studies have indicated that the degree of genetic resolution does not improve very much with dense maps (1-2 cM) and large population sizes (loo0 gametes): confidence intervals for QTL with large effects remain near 10 cM (Darvasi et al., 1993). The limits of resolution have several sources: lack of recombinant gametes, genetic heterogeneity for regional restrictions to recombination, missing data for markers and traits, mistakes in data collection, linked QTL, and proba-

3 02

M. LEE

bly more for QTL of truly minor effects. The lack of resolution will be a fact of life for the foreseeable future. The important distinction to make is whether the level of genetic resolution constitutes a fatal flaw or an inconvenience. Of course, the answer depends on the objectives for locating the QTL. Mapbased cloners will need much better resolution than 10 cM unless cloning and transformation methods improve quickly and substantially and the region of interest has an unusually low physical to genetic distance ratio. Here, 10 cM is close to a fatal flaw unless there are other means of confidently achieving more precise and accurate placement. For those interested in MAS for breeding or genetics, 15-20 cM may be an acceptable (and practical) limit of resolution. Smaller regions (1-5 cM), although ultimately necessary for maximum efficiency according to simulated MAS (Gimelfarb and Lande, 1994), would be difficult to recover because of limited opportunities for recombination with normal breeding procedures and standard population sizes for most crops. Here, assessment of the region could be further refined through recombination in subsequent generations and by progeny testing. These processes could reveal options for MAS of the target region(s). In the absence of such resolution, those interested in MAS may be forced to transfer larger regions than necessary to be reasonably confident that the desired QTL have been recovered. Obviously, a region of this size (15-20 cM) could contain several QTL of opposite (+/ -) effects for the traits of interest, thus limiting the genetic gain and breeding value of the region. For exploratory mapping research, resolution of 15-20 cM or greater may be quite acceptable because it may represent the first clue about gene location and suggest strategies for the next step of the investigation. One of the most rewarding and potentially deceptive aspects of QTL mapping is that putative QTL of large and small effects are detected with comforting regularity. Would the same QTL (major or minor) be detected in another sample of the population? Simulation studies (van Ooijen, 1992; Carbonell et al., 1993; Beavis, 1994) and limited empirical investigations (Beavis, 1994) have provided ample evidence for concern. With small sample sizes (
DNA MARKERS AND PLANT BREEDING PROGRAMS

303

progeny (N = 96, 112, and 400 for generations F3, F2:4, and F2:3 derived from F2 after four generations of random mating, respectively) for QTL. Each sample was grown in a different set of years and locations and evaluated by different research groups. Sets of common RFLP loci and traits were included in all samples. Briefly, there was very little agreement among the samples for location and effects of QTL, suggesting that sampling has considerable influence on their estimation. Several other explanations for the differences were also discussed. In contrast, QTL mapping of insect resistance (Lee, 1993) and morphological traits in maize (Veldboom et af., 1994a; P. J. Freymark and M. Lee, unpublished results) and maize X teosinte populations (Doebley and Stec, 1993) detected several QTL in similar genomic regions between populations involving the same parents. Also, a summary of QTL detected in several mapping populations by one research group identified genomic regions, albeit large, common to several genetic backgrounds (Abler e? af., 1991; Helentjaris, 1992).

2. Estimating Genetic Effects and Expected Genetic Gain A potential use of QTL mapping is to improve methods for predicting the mean phenotype of progeny produced by individuals selected in breeding programs (Paterson er d . ,1991a). Procedures for estimating genetic effects for QTL share many of the same problems encountered for mapping QTL. In most experiments, the models used are overparametrized (too many marker loci for the types and sizes of populations). Also, estimates of recombination and effects are inherently confounded (Edwards et af., 1987). More accurate assessments would be obtained with independent samples of progeny from the same population (Lande and Thompson, 1990), as opposed to the typical practice of using the same sample for estimating location and effects. Therefore, most estimates of the genetic effects of QTL will be biased. The type and degree of bias will depend on the genetic and environmental designs and models used to detect and characterize the QTL. The direction and degree of bias are important because the estimates will be used to forecast genetic gain and merit (Hoeschele and Van Raden, 1993), assess gene action for breeding programs, and investigate complex phenomena such as heterosis (Stuber, 1994a). Also, they might be used by molecular biologists to select targets for map-based cloning. Most QTL mapping designs permit estimates of additive effects. These will usually be biased upward to a degree dependent on numerous variables (Carbone11 et d . , 1992; Darvasi e? d . , 1993; Hoeschele and Van Raden, 1993; Beavis, 1994). The sources of the bias include, but are not limited to, the following: deficiency of recombinant gametes, G X E, and underestimation of epistasis. One source of bias, recombination, is illustrated in Figure 2 (Austin and Lee, 1994; D. F. Austin and M. Lee, in press) for plant height QTL on the long arm of chromosome 1 of maize. In the F3 generation, the QTL composed

304

M. LEE

one large genomic region. In the F6 generation, the region was resolved into three distinct QTL defined by the same RFLP loci as in the F3. Additive effects at the locus an 2.6 were 15 and 7 cm for the F3 and F6 generations, respectively (data not shown). Possibly, the additional opportunities for recombination and the increased homozygosity of the F6 lines facilitated resolution of the region. Similar observations have been made for F3 lines that were derived from maize populations before and after random mating (Beavis et al., 1992; W. D. Beavis and M. Lee, unpublished results): estimates of genetic effects were usually smaller for the sample of lines derived after random mating. Estimates of dominance effects, already hindered because they are often based on a deviation from a hypothetical midpoint, are expected to be biased upward because of recombination (for those linked in repulsion). Empirical evidence supporting this expectation may be found in a classical quantitative genetics study of maize (Moll et al., 1964). Epistatic effects, if they are perceived with popular methods, may be underestimated, perhaps grossly (Lamkey and Lee, 1993; Tanksley, 1993). From a strict molecular perspective, epistasis is important for the expression of most, if not all, genes; therefore, it must be a significant determinant and source of variation for their associated phenotypes, unless the agents of epistasis (e.g., transcription factors) do not exhibit much variation. Also, phenotypes are achieved through interactive and interrelated pathways (metabolic and ontogenetic) involving external and internal signals. Therefore, epistasis should be expected. When rigorous genetic and molecular analyses have been possible, epistasis has been detected and defined clearly. In maize, examples include the anthocyanin pathway (Coe et al., 1988), the opaque-2 locus and levels of different zein classes (Schmidt et al., 1987), and sucrose synthase genes (Chourey and Taliercio, 1994). The type and magnitude of epistasis are important to breeding programs because they influence estimates of expected gain from selection and breeding strategies. Detection of epistasis through biometric analysis has been hindered by the very nature of the methods, typically taking a pangenomic perspective as opposed to specific loci and alleles. Also, methods commonly assume that epistasis does not exist (Lamkey and Lee, 1993), in part because simple two-locus epistatic models involving linkage are obstinate mathematically (Weir and Cockerham, 1977). With biometric approaches, the perception of epistasis and its relative importance as a source of genetic variation and effects has depended on the method and materials [i.e., reference population(s)]. In maize, generation means analysis has usually detected significant epistatic effects, whereas analysis of variance (covariance of relatives) typically has revealed less significant or nonsignificant epistatic effects (Lamkey and Lee, 1993). The reference population is also an important consideration as estimates of epistasis (and dominance) from complex synthetics or open-pollinated populations have tended to be lower than

DNA MARKERS AND PLANT BREEDING PROGRAMS

305

those obtained from populations created from a few elite inbred lines (Lamkey et al., 1993). Significant epistatic effects have been reported in QTL mapping studies, but they are relatively infrequent. Several reasons have been presented to account for this pattern (Tanksley, 1993): (1) mapping populations have been too small to provide sufficient representation of the numerous possible genotypic classes at interactive loci (linked or unlinked); (2) the number of possible multilocus combinations is too large for standard statistical tests; and (3) recombination between QTL and DNA markers (the basis of the tests) will underestimate any epistatic effects. Suggested solutions for improved assessments of gene effects have included much larger populations sizes (increase from <300 to >500 progeny), denser marker maps for detecting epistasis, and restricted analysis of phenotypic extremes (Lark er al., 1994). In general, proposals for reducing bias and increasing genetic resolution have included the construction of genetic stocks (e.g., NILS, deletions, random-mated populations) and improved models for simultaneous treatment of all QTL (Knapp er a!., 1992). In addition, sampling of the target environment should be more extensive in some situations to reduce upward bias due to G X E. The efficacy and utility of these solutions remain to be tested for nearly all QTL reported to date. Collection of definitive data for some traits (e.g., grain yield, root and stem strength) and crop species should prove challenging. The status and influence of G X E in QTL analysis have not been considered in this review. Interested readers may find a comprehensive treatment of that subject in a forthcoming review by Beavis and Keim (1995). A quote from Allard (1960) summarizes the current status of G X E in QTL studies: In practice, then, the difference between qualitative and quantitative characters depends not so much on the magnitude of the effect of individual genes as on the relative importance of heredity and environment in producing the final phenotype. It is therefore apparent that the key to progress in the analysis of quantitative characters lies in evaluating the relative contributions of these two causal agents to the variability, and it is to this subject that we must now turn. To date, the surprising observation has been the lack of the crossover type of interaction for QTL even in the presence of highly significant G X E, as indicated through a pangenomic analysis of variance. Obviously, genes interact with and respond to environmental cues (e.g., temperature-sensitive rust resistance, vernalization requirements, photoperiod response). Perhaps the present experimental designs only permit reliable detection of QTL with major and environmentally robust effects. Given that many of the mapping populations were constructed to provide maximum divergence (for markers and phenotypes), this may be a reasonable expectation. Perhaps more carefully planned, designed, and

306

M. LEE

controlled experiments will be needed to investigate G X E. A more thorough appreciation of this phenomenon will be important for more realistic appraisals of QTL and their potential for advancing genetic gain. The uncertainty related to QTL position and effects will hinder the effectiveness of MAS for breeding and positional cloning in many circumstances. For MAS, the consequences could involve less genetic gain than anticipated due to inflated estimates of additive and dominance effects and unperceived epistatic effects. Also, the ambiguous genetic position of QTL might require MAS of excessively large regions to guarantee transfer. All this must be weighed against the merits of conventional selection: depending on one’s knowledge of expected and realized genetic gain achievable through non-MAS methods and comparable time and resources required to make the gain, the results from MAS may or may not be a disappointment. For positional cloning, the mapping ambiguities represent a formidable obstacle for currently available technology. As mentioned previously, the hurdles may be lowered somewhat through construction of special genetic stocks, but the confidence interval of the target region is predicted to remain large, (10 cM; Darvasi et al., 1993). Given the known variation in the physical/genetic distance ratio, the number of base pairs in such regions could be unreasonably high for map-based cloning under the best circumstances (Section 111.B .2). The problems and consequences of estimating QTL positions and effects, as they relate to genetic gain, are not new to plant breeding. Substantial efforts have been made to predict accurately genetic gain, but they have not been entirely successful. Predicted and observed genetic gain have agreed to the extent that both have been positive. However, predicted gain has exceeded observed gain in most circumstances, especially in the short-term. The difference in predicted and observed gain may be attributed to several sources: sampling errors of predictors (genetic, environmental, and error variances), G X E effects, genetic drift and inbreeding, linkage disequilibrium, and epistasis (Hallauer and Miranda, 198I), a very familiar list. Also, backcross-breeding has been successful at transferring genes for qualitative and quantitative traits, but in many instances the level of trait expression in the final backcross progeny is less than that observed in the donor parent. This has been especially true for traits with quantitative inheritance patterns and lower heritability (Fehr, 1987). Perhaps DNA markers could be used to improve the relation between predicted and observed gain in situations where it has been unacceptably weak and realized gains have been small and slow to achieve. The theoretical and more limited empirical assessments of QTL detection and characterization have been very instructive regarding descriptions of types of problems and suggestions for minimizing them. While empiricists should heed such warnings, pragmatic concerns will compel them to assume the associated risks, proceed with due caution, and attempt to use the mapping information. At

DNA MARKERS AND PLANT BREEDING PROGRAMS

3 07

this point, assessment of relative risk (e.g., types and rates of errors, the cost of being wrong versus the benefits of being correct) and comparative advantage (e.g., the history of conventional versus the potential of MAS, economics, efficiency, and the value of time) become important considerations. Fortunately, plant breeders and geneticists do not assume the same direct risks as human genetic counsellors; thus, they can afford to be more opportunistic and creative in their use of QTL mapping information. However, given the inherent and circumstantial ambiguities of QTL mapping, the best approach may be to treat initial assessments of most QTL as hypotheses to be tested rather than genes to be selected for genetic gain.

3. Genetic Heterogeneity A phenotype is considered to be genetically heterogeneous if it may be caused by mutations at one of several loci (Lander and Botstein, 1986). Obviously, the potential for genetic heterogeneity to influence assessments of phenotypes in plant breeding and QTL mapping is enormous for several reasons: polyploidy and intragenomic duplication, complexities and interrelationships inherent to metabolism and ontogeny, unknown allelic frequencies and constitutions at QTL in reference populations, and the hierarchical structure of important traits (e.g., grain yield). We should expect to observe many examples of genetic heterogeneity to the delight and disappointment (initial) of biologists and breeders, respectively. For example, the classical maize genetic map contains at least 40 loci affecting components of plant height (Coe et al., 1988). How many loci should we expect to find for yield and determinants of yield? (How many angels may dance on the point of a pin?) For practical plant breeding, the issue here is the degree to which estimates of QTL effects and location may be transferred from one population to another. Observations based on empirical studies provide contrasting perspectives. Several investigations have reported few if any QTL in common between populations: soluble solids in tomato (Tanksley and Hewitt, 1988). plant height in maize (Beavis et af., 1991), resistance to gray leaf spot disease in maize (Bubeck et al., 1993), and protein content in maize (Schon et al., 1994). Alternatively, evidence for consistent QTL detection across populations has been observed in maize for insect resistance (Lee, 1993), morphological traits in elite maize (Abler et al., 1991) and maize x teosinte (Doebley and Stec, 1993) populations, flowering ( K w t e r et al., 1993; Phillips et af., 1992) and has also been observed from comparative QTL mapping between mung bean and cowpea for seed weight (Fatokun et al., 1992) and between maize and sorghum for plant height (Pereira and Lee, 1995). Analysis of this issue is complicated by many factors described for QTL detection (Section 1V.A. 1-2). For example, small sample sizes ( C 150 progeny)

3 08

M. LEE

were used in several studies, so sampling variation could be important. Nevertheless, it is probably safe to assume that considerable genetic heterogeneity will be evident for QTL and to a lesser extent for more qualitative factors. Does this constitute a fatal flaw or an inconvenience? The answer depends on the inference one wishes to make and the particular circumstances of the breeding programs. Here, the concept and definition of the genetic reference population become important. Should the estimates of QTL position and effects be relevant to the entire species or subsets of the elite gene pool? More focused definitions and limited scope (gene pool) for inferences probably will ameliorate frustrations caused by genetic heterogeneity for some traits and procedures. QTL mapping studies involving common parents suggest that this may be a reasonable expectation (Abler et al., 1991; Koester er al., 1993; Phillips et al., 1992; Doebley and Stec, 1993; Lee, 1993). Given the presumed conservative and narrow genetic foundation of elite crop germ plasm, the actual reference populations might be smaller than realized and conducive to acceptably accurate inferences for some traits. For some circumstances, the issue of genetic heterogeneity may be less significant. For example, the identification and spread of novel races of insects and fungi may render elite gene pools useless as a source of resistance (e.g., race 24 of barley stripe rust and North American spring barley; Chen et al., 1994). Often, resistance may be available in very few representatives from another gene pool. Here, the breeder has few options. Ultimately, one genetic source of the trait may be selected and become predominant in the elite gene pool, and this should minimize the degree of genetic heterogeneity for some phenotypes. 4. Genetic Nature of QTL

Quantitative genetics has been described by Lewontin (1977) as an attempt to produce knowledge through a systemization of ignorance: in most instances, nothing has been known about number and function of genes, linkage, and underlying biology of the trait(s) or process(es) being assessed as reduced statistical entities. Nevertheless, quantitative genetic principles have fostered the development of useful plant breeding practices such as methodical progeny testing schemes and objective approaches for comparing genetic gain from selection for different breeding schemes (Lamkey and Lee, 1993). QTL mapping, a scion of quantitative genetics, has continued the tradition of ignorance but has managed to reduce it to more approachable packages. The true biological nature of any QTL in plants has yet to be determined. At least five concepts have been proposed regarding this topic (Lamkey and Lee, 1993). QTL are (1) major genes with pleiotropic effects on other traits [reviewed by Barton (1990)l; (2) fundamentally different from major genes in that QTL alleles are limited to small effects (Mather, 1941); (3) modifiers (epistatic) of major genes

DNA MARKERS AND PLANT BREEDLNG PROGRAMS

3 09

(Mukai and Cockerham, 1977); (4) loci with an allelic series with a range of effects leading to their recognition as macromutations (mutants) and micromutations (Allard, 1960; Thompson, 1975; Robertson, 1985); and (5) tightly linked multilocus clusters (Allard, 1960, 1988). Distinction among these concepts may be important to the design of subsequent investigations and deployment of the QTL. Several reports have revealed important features of the genetic nature of specific QTL. Two related reports on Drosophila have provided important paradigms for plant QTL mappers. DNA insertion polymorphism of wild-type alleles at the achaete-scute locus were demonstrated to be strongly associated to the quantitative trait, bristle number (Mackay and Langley, 1990). Similar observations were made for the same trait and another locus, scabrous (Lai et al., 1994). Both studies support the concept that a series of wild-type alleles may exist at loci definable by macromutations. The wild-type alleles with a range of effects may constitute an important source of genetic variation. One potential benefit of this relationship between quantitative and qualitative variations is that the macromutations, because of their major effects on the phenotype, may be amenable to gene cloning methods. Once fully characterized, they may be used to isolate homologous wild-type alleles. Molecular analysis of the wild-type alleles may provide insights into the biological basis of quantitative genetic variation and related subjects such as the genetic basis of response to artificial selection. Genetic and molecular analyses of QTL in plants have tended to support this concept. QTL mapping studies in several crops (e.g., maize, tomato, barley) have located QTL to regions known to contain loci defined by macromutations. Linkage relationships inferred from integrated maps have provided the basis for proposals for direct analysis of QTL for morphology and kernel starch concentration. Doebley and Stec (1993) identified a QTL on the long arm of chromosome 1 that had major effects on tillering and lateral branch development. The QTL was estimated to be within 10 cM of a genetic locus, tbl, defined by alleles of qualitative effects on the same trait(s). The QTL was back-crossed into a uniform inbred background and test-crossed as a heterozygote to a genetic stock heterozygous for the macromutation, tbl. One-fourth of the progeny exhibited a weak mutant phenotype indicating that the original QTL (at least part of that region) was allelic to the macromutation (Doebley and Stec, 1994). QTL for kernel starch concentration were localized to the vicinity of the Sh2 locus (Goldman et al., 1993), which codes for an enzyme known to be involved in the rate-limiting step of starch biosynthesis in plants. Macromutations at Sh2 drastically decrease the starch levels in the endosperm. Sequence analysis of Sh2 alleles from the reference population revealed a novel transposon insertion (Alrefai et al., 1994). Further tests are in progress to test the relationship and characterize the biological basis for quantitative variation associated with this region (Rocheford, personal communication, 1994). Obviously, such analyses are not possible for most QTL,

310

M.LEE

but the opportunities should increase with further integration of maps, characterization of cDNA clones, and development of routine transformation systems. QTL do not always coincide with regions identified by known macromutations. For example, QTL for quantitative sources of disease resistance were not linked to known genetic loci in barley (Heun, 1992) and maize (Freymark et af., 1993). Also, well-characterized QTL in maize for glume architecture (Dorweiler er al., 1993) and flowering (Phillips ef al., 1992; Kim et af., 1993) seem to define new genetic loci. In the latter two situations, macromutations were resolved clearly from the milieu of continuous phenotypic variation through the development of NILs. As mentioned in Section 111, such analyses represent some of the power of QTL mapping and its potential as a very important source of information for revealing the genomic architecture of phenotypes and processes. Evidence of clustered multilocus QTL has been reported in tomato (Paterson et af.,1991b), in barley (Hayes et al., 1993), maize (Veldboom et al., 1994), and sorghum (Pereira and Lee, 1995) for several traits. In each situation, it was not possible to attribute solely the clustering to linkage, pleiotropy, or correlations due to ontogeny. Therefore, further analysis of such regions clearly is warranted. Obviously, no single category or theory will suffice for describing the genetic nature of QTL.

5. Research Needs for QTL Analysis At this time, perhaps the greatest need for QTL analysis is the verification of initial mapping results regarding positions, effects, and localized content of QTL. Some of these features may become more evident through improved methods for scanning genomes, but nothing will substitute for direct tests of putative QTL. In some situations, verification could be achieved by detecting heterozygotes at QTL in later generations of self-pollination and comparing opposite alleles in homozygotes in sister lines [Atkins and Mangelsdorf (1942) cited in Allard (1960)], with MAS in independent samples of carefully chosen reference populations or additional sampling strategies of the same population (Lark et af., 1994). Other objectives may require the production of additional recombinants or genetic stocks for more focused analyses and as preparation for molecular scrutiny and more complete integration of genetic maps. Collectively, the results thus far indicate that real genes are being detected by QTL mapping methods. However, numerous concerns about the methods lead to the following questions. Exactly what are we detecting? And perhaps of greater importance, what are we NOT detecting? Breeding programs interested in using assessments of QTL in selection programs will need to make a long-term commitment to developing and implementing pangenerational genetic information similar to animal breeding methods.

DNA MARKERS AND PLANT BREEDING PROGRAMS

311

This will require evaluation of breeding records and values for pedigrees and QTL alleles. For some crops, the development of the necessary databases may not be unreasonable because of extensive prior knowledge of the genetic merit of germ plasm, availability of breeding records and materials, and ease of developing and evaluating progeny from carefully chosen crosses representative of key reference populations.

B. ASSESING AND INTROGWSSING EXOTIC GERMPLASM This aspect of crop improvement is positioned to benefit the most from applications of DNA markers. Not only do molecular markers provide an unprecedented glimpse into the distribution of genetic diversity (Section 11), they also provide an opportunity to improve the documentation of actual contributions from exotic germ plasm to improvements in elite germ plasm. Certainly, the role of exotic germ plasm has been undervalued, in part due to a very limited ability to perceive and prove it. Assessments with DNA markers indicate that exotic, donor parents contribute more genes with positive effects that could not have been predicted from their phenotypes or pedigrees of progeny in maize (Lee et al., 1990), tomato (de Vicente and Tanksley, 1993; Eshed and Zamir, 1994), oats (Lawrence and Frey, 1976; M. Lee, unpublished results), and wheat (Rogowsky et al., 1991; Schwarzbacher et al., 1992). Simulation studies also suggest that the efficiency of MAS should be greatest under conditions typical of introgression programs, the predominance of coupling phase linkage and restrictions to recombination (Section V.A.2). Exotic germ plasm has been an important source of genes with highly qualitative effects, usually for defensive traits (resistance to abiotic and biotic stresses). Extraction of such genes should be enhanced through marker-assisted selection with backcrossing (MAS BC; Section V.A. 1) and an increased awareness of restrictions to recombination (Section 111). In contrast, the role of exotic germ plasm in improving quantitative traits has been less prominent. Edwards (1992) has discussed several reasons for this differential use: (1) the presumption that exotics had little to offer elite germ plasm (i.e., low frequency of favorable alleles); (2) the desire and need to demonstrate short-term gains in cultivar development; (3) the overwhelming size and underwhelming characterizationof some germ plasm collections; and (4) the widening gap in performance of elite and exotic germ plasm. With the advent of DNA markers, Edwards suggests that it may be possible to develop efficient strategies for rapidly identifying and incorporating favorable exotic alleles into elite backgrounds to realize a net improvement in trait performance. DNA markers could also increase the efficiency of germ plasm conversion

312

M. LEE

programs such as those used for sorghum (Duncan et al., 1991) and wheat (spring X winter crosses; Fischer, 1993) and proposed for maize (A. R. Hallauer, personal communication, 1994). The goal of the conversion programs for maize and sorghum is to adapt tropical germ plasm such that it may be grown and evaluated in temperate regions. The major barrier to direct evaluation is the photoperiod response of the tropical germ plasm (and stature for sorghum). The tropical germ plasm is adapted by crossing to an elite temperate-adapted parent, producing a segregating population, selecting adapted progeny, and often back-crossing them to the tropical parent in a cyclical manner (up to five BC generations) to promote maximal recovery of exotic germ plasm. Once the adapted growth habit has been achieved, the merit of the exotic genes may be assessed in breeding programs. Such conversion programs might incorporate DNA markers at several stages. First, selection of exotic parents should promote maximum diversity (and minimum duplicates). DNA markers could assist with the selection of exotic parents for conversion. When segregating progeny are selected for back-crossing, markers could be used to identify progeny significantly deviating from average expectations of allelic composition to use those with the highest proportion of exotic alleles (Pereira and Lee, 1995). This identification could be made prior to flowering and would reduce the number of backcross generations (Section V.A. 1) and facilitate maximum recovery of exotic alleles. Elimination of some BC generations might permit a reallocation of resources into the conversion of additional populations. Thus, breeders would have more opportunities for assessing the merits of truly exotic alleles with unique and favorable effects.

C. RESPONSE TO SELECTION IN PLANT BREEDING PROGRAMS The genetic basis of response to artificial selection remains a mystery (Barton, 1990). To what extent is the observed response attributable to extant genetic variation? To variation generated de now? What mechanisms are capable of generating genetic variation beneficial to crop improvement programs? What mechanisms stabilize the content and expression of plant genomes? Answers to these questions have important and direct implications for many phases of plant breeding, such as effective population sizes, selection response models, and production of transgenic crops. Slowly, DNA markers are revealing some features of underlying mechanisms. Retrospective analysis of selection programs with DNA markers has provided important clues about the transfer and maintenance of genes in plant breeding programs. Perhaps the first big surprise was the variable sizes of chromosome segments retained in BC programs, as exemplified by Young and Tanksley (1989b) in tomato. Once detected, it would be important to determine whether such regions were the products of bad luck on the part of the breeder or real

DNA MARKERS AND PLANT BREEDING PROGRAMS

313

restrictions to recombination. Surveys of long-term recurrent selection programs have found high frequencies of fixed RFLP alleles in advanced generations (Sughroue and Rocheford, 1994; K. R. Lamkey and M. Lee, unpublished data). On the basis of effective population size and variation revealed by RFLPs, the genetic variation in these generations should be greatly depleted, yet biometric analysis and further selection response suggest the contrary. What is the source of genetic variation in these situations? Analysis of one selection program has found a QTL for low levels of endosperm starch associated with a transposon insertion of the Sh2 locus (Goldman et al., 1993; Alrefai et al., 1994). Did the insertion occur during the lifetime of the selection study? This preliminary evidence indicates that transposons could be an important source of genetic variation in plant breeding programs, in accordance with previous suggestions (Schwarz-Sommer et al., 1985; Nelson, 1990; Lamkey et al., 1991). Surveys of plant genes and mutant alleles have identified numerous retrotransposon families and their remnant sequences associated with genes, some of which are the cause of mutations. The proximity and ubiquity of the retrotransposons suggest an important role in the evolution of regulatory and structural sequences with profound influences on gene expression and duplication (White et al., 1994). Detailed molecular and genetic analyses have revealed a strong tendency in plants for localized sequence duplication (Section 1II.B .2). In such instances, unequal crossing over has been invoked as a mechanism leading to the production of tandem arrays of genes and pseudogenes. The origin of such organization remains unresolved but some of it may be mediated by retrotransposons (White et al., 1994). The creation of such regions potentially complicates gene expression (Flavell, 1994); thus, investigation of how plants have reconciled this complexity should provide important clues for stable expression of transgenes and, perhaps, the creation and preservation of pseudogenes as sources of cryptic variation. Larger-scale sequencing projects and related investigations will undoubtedly reveal more surprises about mechanisms of plant genome fluidity and maintenance. Plant breeding programs may be an important source of materials and guidance for activities such as retrospective analysis of the products and progress from selection in elite, core groups of germ plasm used so successfully by cultivar developers.

V. MARKER-ASSISTED SELECTION Success in basic research is often easier to achieve than success in the subsequent development and application. This is certainly true for QTL mapping and

3 14

M. LEE

MAS for cultivar development. The challenges loom large for MAS; in many crops, conventional selection has had several decades to evolve into a very effective technology. Plant breeding programs in most crops conduct simultaneous selection for several (>lo) traits, in many (e.g., 50-200) populations, across diverse environments. Genetic advancements may be rapid as many annual crop species complete three sexual generations each year (most complete two) with off-season nurseries (with or without selection). Therefore, the development of a new cultivar requires 5-10 years for most annual crop species. However, plant breeding programs have not solved every problem facing crop production systems. In some situations, genetic advance has been limited due to the complex and ambiguous nature of the trait(s) (e.g., water stress, salinity, grain quality, heterosis for grain yield) and its response to unperceived or uncontrollable environmental cues. It may be possible to create a selection environment capable of discriminating genetic and environmental variation, but widespread and routine use of that environment may be impractical for the number of populations and progeny used in breeding programs and some traits (e.g., quality factors requiring end-user tests, insect resistance, stage-specific water stress). Therefore, selection for such traits may be delayed until a later generation, after the sample size has been reduced on the basis of other traits and less costly evaluation procedures. This limits opportunities for genetic gain. Also, the list of selection criteria may grow because of new objectives (e.g., quality standards), stress factors (e.g., new pathogens or races thereof), and technology (e.g., transformation). Under these circumstances, breeding methods and selection criteria may require revision if not establishment de novo. MAS will be capable of producing genetic gain in many circumstances. The more important concerns are related to the relative costs, rates of gain (short and long term), and ultimate levels of gain for MAS and non-MAS. Reconciliation of these issues is strongly dependent on the crop, trait(s), and other circumstances. However, a few questions may apply to most situations: (1) For a given breeding method, how long does it take to develop a new cultivar? What is the expected duration of a cultivar’s utility? What is the record of productivity for your breeding program regarding the release of new and useful cultivars? (2) For a given breeding method, how much does it cost to produce a new cultivar? (3) For a given trait, what are the financial costs of selection? The genetic costs? The history of response to selection? The frequency of success and failure? (4) What is the economic value of time? To the breeding program? Seed company? Grower? Consumer? Firm answers to these questions should be available for most breeding programs and non-MAS methods. Answers, or at least educated guesses, are becoming available for MAS through simulation studies and some empirical investigations. Their comparison might help focus and resolve the debate.

DNA MARKERS AND PLANT BREEDING PROGRAMS

315

A. DETERMINISTIC AND SMUTION STUDIESOF MAS IN

PLANTBREEDINGPROGRAMS

The development of meaningful simulation models for artificial selection in plant breeding programs must be one of the most formidable challenges. Initially, the models must attempt to reduce vaguely defined and complex biological processes and their products to very simple terms; ultimately, the models must account for an equally complex component, economics. Nevertheless, such modeling is valuable because it promotes focused thought and attempts objective analyses.

1. Monogenic Inheritance DNA markers offer considerable advantages for backcross (BC) breeding: (1) indirect selection of desirable gene(s) from donor parents, (2) selection for regions of recurrent parent genome unlinked to the introgressed region, and (3) reduction of linkage drag of unwanted donor parent genome near the introgressed region(s). Collectively, these capabilities provide a means of reducing the number of generations needed for recovery of the converted recurrent (recipient) parent. These features and their implementation have been discussed from the perspective of reducing linkage drag (Young and Tanksley, 1989b), comprehensive introgression strategies (Hospital er d.,1992; Openshaw er d.,1994), and linkage phase of markers (Haley er af., 1994). Simulation studies and empirical reports have concurred that MAS assists BC breeding by decreasing the number of generations for reducing linkage drag and recovery of the recurrent parent. However, the magnitude of the expected benefits has varied in different simulation studies. Openshaw er af. (1994) reported that the number of BC generations could be reduced from seven to three by using a modest sample size (
316

M. LEE

advantage of MAS should decline with additional regions for BC (Edwards and Page, 1994). The utility of MAS for achieving and improving genetic gain through BC breeding depends on the current and potential role of that breeding method. BC breeding has been widely used for introducing monogenic characters and less so for polygenic traits. Perhaps the utility of this method could be increased for the latter situation through QTL mapping. In the short term, the role of BC breeding is likely to increase whenever genes have been introduced through asexual transformation methods. In most scenarios, the transformation methods exhibit some degree of genotype specificity, such that methods must be adapted to each genotype or the efficiency of methods is enhanced for a few selected genotypes (Armstrong et af., 1992). Therefore, the transgene (the product of considerable development costs at this point) will be introduced into inferior germ plasm through several independent transgenic events. With present technology, the transgene will insert into a different genomic region in each event. Also, the transformed germ plasm might be an F1 hybrid (e.g., elite line recalcitrant to tissue culture regime crossed to line amenable to tissue culture regime). The task will be to move the transgene into the most elite germ plasm as quickly as possible. DNA markers may greatly increase the efficiency at several stages under these circumstances. By assuming acceptable and reliable levels of transgene expression, markers could help initially by identifying inserts within chromosome regions of the elite parent. Within that subset of events, it might be possible to select inserts located in more desirable chromosome regions (eu- versus heterochromatin; centro- versus telomeric regions). Subsequent backcrossing will be achieved at accelerated rates as previously discussed. Also, it might be possible to backcross several additional independent transgene lineages as insurance against unpredictable transgene silencing events in later generations. Similar strategies could be applied to genes derived through introgression of wild germ plasm. Obviously, the value of the reduced number of backcross generations, less linkage drag, and more complete recovery of the recurrent parent genome are best assessed within the context of circumstances peculiar to the crop, selection criteria, agricultural system, and cost of acquiring the marker data. The paradox of MAS BC is that its reductionist nature may actually lower the level of maximum genetic gain possible for a population. A tendency of BC schemes, likely to be exacerbated in MAS BC, is to assume that the genetic content of the donor genome is inferior to that of the recurrent parent genome at all regions except for the location of the target gene(s). Yet RFLP markers in maize (Lee et af., 1990) and tomato (de Vincente and Tanksley, 1993) and evaluation of interspecific oat crosses (Lawrence and Frey, 1976; M. Lee, unpublished results) have demonstrated the donor parent usually has positive factors for traits unrelated to the main objective(s). Further, breeders have recovered

DNA MARKERS AND PLANT BREEDING PROGRAMS

317

these unanticipated favorable alleles in BC programs without MAS. Such regions would be routinely eliminated by MAS BC unless prior knowledge of their merit was available. Given these circumstances, the products of MAS BC might be expected to be somewhat inferior to those obtained through conventional BC. If true, implementation of MAS BC may depend on the relative values of the reduced time for product development, the recurrent parent phenotype, and the potential units of genetic gain. 2. Polygenic Inheritance

The relative efficacies of MAS and traditional selection for improving quantitative traits of domestic animals and plants have been considered in several simulation studies (Lande and Thompson, 1990; Lande, 1992; Zhang and Smith, 1993; Edwards and Page, 1994; Gimelfarb and Lande, 1994; de Koning and Weller, 1994). Except for de Koning and Weller (1994), economic considerations were not included. Also, the situations apply to single-trait selection, although some consider multiple QTL (Edwards and Page, 1994) and traits (de Koning and Weller, 1994). Despite their use of somewhat different conditions and assumptions, the studies generally concur on several points. The efficiency of MAS is enhanced and may be more efficient than traditional selection under the following circumstances: (1) the trait(s) under selection has low heritability; (2) tight linkage between QTL and markers (<5 cM), with additional efficiency realized when coupling linkages predominate (Gimelfarb and Lande, 1994); (3) in earlier generations of selection prior to fixation of alleles at or near marker loci and recombinational erosion of marker-QTL associations; and (4) larger sample sizes for mapping and selecting QTL are used to improve estimates of QTL effects and to avoid rapid fixation of alleles, respectively. Some notable deviations from the general trends include the insignificant role of sample size reported by Edwards and Page (1994) and the inefficiency of MAS except for special circumstances noted by Zhang and Smith (1993). Under these circumstances, MAS is expected to result in greater rates of genetic gain during the early generations of selection. At later generations (e.g., lo), the gain afforded by traditional selection is expected to surpass or approach that of MAS. The simulation studies indicate that MAS should have some advantages for artificial selection of annual crop species for some traits and circumstances. For example, the increased efficiency of MAS attributed to coupling phase linkage and very close linkage between markers and QTL suggests that MAS should be much more efficient for crosses between exotic and elite germ plasm as opposed to crosses between elite parents. Therefore, MAS may be a more important tool for introgression and germ plasm conversion programs, although with the cost and capability of current marker technology the advantages of MAS may be diminished further (Sections VI; Lande, 1992) for the near term. Presumably, the

3 18

M. LEE

advantages of MAS would be greater for long-lived perennial crops with long generation times and limited opportunities for selective breeding. However, one assessment of the utility of MAS for these situations identified numerous limitations for this approach to forest tree breeding (Strauss et al., 1992) and suggested that markers would contribute to genetic advances by providing an improved understanding of the genetic architecture of complex traits (QTL mapping) and as a complement to traditional selection under special circumstances. Clearly, there are substantial impediments to implementing and benefiting from MAS of polygenic traits for some breeding methods, crops, and traits. Certain traits (e.g., economic yield) may well be beyond the realm of influence of any technology other than direct selection and widespread testing. As integrated maps, related QTL databases, and marker detection systems improve, some of the impediments might diminish. Likewise, unforeseen opportunities shall emerge. For the moment, it is important to remember that MAS for polygenic traits is at the nadir of its efficiency relative to direct selection.

B. EMPIRICAL RESULTS Indications from the survey (Section VI) and simulation studies provide sufficient reason to believe that MAS BC of monogenic traits will provide adequate advantages in the short term to justify its use in cultivar development. Advancements in DNA marker methods and mapping strategies have made it possible to identify sufficiently tight linkage with macromutations in a directed manner. Thus, if such genes are needed for cultivar development, one may be fairly confident that suitable markers may be developed to accelerate selection and genetic gain. The prospects for MAS of polygenic traits are less certain. Predictions from simulation studies and concerns about procedures for detecting and characterizing QTL should provide adequate cause to pause and reconsider the anticipated benefits of MAS. At best, the predictions were cautiously optimistic about the general utility and efficiency of MAS and suggest that benefits may be limited to specific situations. The few empirical investigations and attempts at MAS of polygenic traits have demonstrated a positive response to selection. The initial MAS studies using isozyme markers have been summarized and discussed in reviews (Stuber, 1992, 1994a; Edwards, 1992). Even with the sparse genome coverage afforded by isozymes, linkage disequilibrium was sufficient for a positive response to MAS, similar to the gains realized through mass selection (Stuber and Edwards, 1986). Presumably, the improved map coverage with RFLPs should increase the response to MAS.

DNA MARKERS AND PLANT BREEDING PROGRAMS

3 19

MAS using RFLPs has been demonstrated to be an effective means of transferring genes (regions) for hybrid grain yield in elite maize inbreds (Stuber, 1994a,b). Chromosome regions affecting hybrid grain yield were identified in donor inbreds Tx303 and Oh43 through QTL mapping (Stuber and Sisco, 1991). When the regions were expected to enhance the grain yield of the hybrid, B73 x Mo17, they were backcrossed by using MAS for three generations into inbreds B73 and Mo17, respectively. Up to six regions were back-crossed into each recurrent parent. Each backcross-derived B73 and Mo17 was top-crossed to the normal Mo17 and B73 as a means of comparing their grain yields with that of the normal hybrid. Replicated evaluations in 1 year indicated that 45 of 141 enhanced B73 conversions and 51 of 114 converted Mol7’s exceeded the grain yield of the normal hybrid by at least one standard deviation, while only 15 and 10, respectively, of the converted lines had lower grain yields than the control. Information on predicted and expected gains, genotype X environment interactions, grain moisture, maturity, and lodging were not included in the report. Also, parallel conventional selection was not considered; nevertheless, the study demonstrated that MAS could be used to manipulate complex traits such as grain yield in elite germ plasm. Empirical comparative studies of MAS and conventional selection have been reported in maize for processing traits (Edwards and Johnson, 1994) and topcross performance (grain yield, moisture, and lodging; Stromberg er al., 1994) in hybrids. Phenotypic recurrent selection (PRS; one cycle in one population, B) and marker-assisted recurrent selection (MARS; three or four cycles in populations A and B) were conducted in independent samples of the same elite base population (B; Edwards and Johnson, 1994). MARS was conducted on the basis of QTL estimates from 160 unselected F4 families of each population with 61 and 52 RFLP loci. Selection of families for recombination was based on a breeding value calculated by summing the assumed directional response of allelic substitution across all loci. PRS was based on top-cross performance and visual selection of inbred progeny. The evaluation of response to selection for each method was conducted in 1993 in one location in Minnesota. MARS achieved positive changes in most traits in both populations; however, the response of each trait was greater for PRS. Stromberg et al. (1994) reported that MAS for early generation testing of hybrid performance was as effective as conventional selection, but the methods neither achieved a significant response to selection nor predicted with accuracy the performance of advanced generations. Evaluations of response to selection in both studies were conducted under poor growing conditions, whereas the selection environments were either more benign (Edwards and Johnson, 1994) or suboptimal due to contrasting conditions (dry versus wet; Stromberg e? al., 1994). Therefore, it would be somewhat premature to make firm conclusions regarding the efficacy and efficiency of MAS.

320

M. LEE

C. INTEGRATING MAS INTO PLANTBREEDING PROGRAMS MAS will gradually earn its place in plant breeding methodology. For reasons discussed previously, the first generation of MAS will involve BC breeding of monogenic factors and limited MARS of factors for polygenic traits as a means of pyramiding genes. Examples of breeding schemes for these approaches have been presented by Melchinger (1990) and Edwards (1992). Additional guidelines and modifications have been suggested in the simulation studies. Choice of trait and situation will be important in the early days of MAS for polygenic traits. Therefore, a necessary prerequisite will be a clear understanding of the relative efficiencies and comparative advantages of traditional selection methods, their accomplishments, and costs. Traits requiring special environments (e.g., abiotic and biotic stresses), methods, or large-scale processing of samples (e.g., milling, baking, malting quality) might present good opportunities. These situations often hinder breeding programs because the expensive or special environment may be fairly limited in terms of throughput regarding the number of progeny; thus, breeders may delay selection for such traits until later generations. By that stage, the sample size has been reduced considerably so that the opportunities for genetic gain have been limited. There is no doubt that MAS will work in any situation in which conventional selection has achieved progress. However, the real issues are relative efficiency (including costs) and rates and ultimate levels of genetic gain in the short and long term. These issues will be decided by the set of circumstances peculiar to each crop and breeding program.

VI. SURVEY OF THE STATUS OF DNA MARKERS IN CULTNAR DEVELOPMENT PROGRAMS The gap between principles and practices may be substantial in any field of science and technology. The literature may be outdated, incorrect, biased in favor of positive reports, and often irrelevant to the goals and objectives of those primarily interested in applications. Therefore, considerable experience and knowledge go unreported. This is especially true for plant breeding and DNA markers because so much of the research occurs in the private sector. Also, plant breeders, public or private, are not acknowledged or rewarded on the basis of their methods, but rather on the results of their methods (cultivars and germ plasm). In an attempt to present a current and comprehensive portrayal of DNA markers’ roles in plant breeding programs, a survey was prepared for the purpose of assessing the present and near-term utility of DNA markers (e.g., RFLR, RAPDs, and other PCR-based markers) for cultivar development programs. Cul-

DNA MARKERS AND PLANT BREEDING PROGRAMS

32 1

tivar was defined as “cultivated variety of a plant” (Fehr, 1987). Transgene was defined as “a segment of DNA of defined function from any source that has been introduced into crop germ plasm using transgenic technology (e.g., biolistics, A. turnefaciens). The survey was conducted during the period October 7, 1994, through November 25, 1994. A copy of the survey is available upon request. The survey was sent to 80 individuals on the basis of their research experience with DNA markers and affiliation with or role within a plant breeding program, typically for an annual, seed-propagated, agronomic crop species. Fifty-five responses were received (Table 1) from scientists located in the following countries: Germany (2), England (l), Mexico (2), United States (46), Switzerland (l), France (I), Brazil (I), and Argentina (I). The distribution between the private and public sectors was nearly equivalent. Maize and wheat were most commonly listed as the crop used as the primary basis for responses. The frequency of breeding systems (i.e., type of cultivar) among the crops was 15 hybrid, 6 pure line, 2 synthetic, and 1 clonal. As with most endeavors in plant breeding, the survey has some obvious limitations due to sampling and resource allocation. The responses lack representation of certain agronomic crops (e.g., rice and barley), most horticultural crops, and long-lived perennials and of scientists working in developing countries. The lack of representation certainly introduces some bias due to differences in infrastructure (e.g., the acquisition and disposal of chemicals) and breeding methods (annuals versus long-lived perennials). Thus, interpretations may have

Table I Overview of Survey Respondentsa and Crops Included in the Survey Crop

Frequencyb

Crop

Alfalfa Barley Beans Table beets Brassica (canola) Broccoli Cacao carrot Cotton Cucumber Maize Melon

1

Millet Oats Onion Rice Sorghum Soybeans Sugarcane Sunflower Sweet corn Sw itchgrass Tomato Wheat

1 1

2 2 1 1 1

2 I 33 I

Frequencyb 1

3 2 I 5 5

I 2 1

1

2 10

a 32 public sector respondents; 23 private sector respondents.

Respondents were asked to indicate the crop species that are the primary basis for their responses to the survey. Respondents could list up to four crops.

322

M.LEE

the most relevance to a rather specific reference population of breeding programs and crops. Respondents were asked to classify the importance of DNA markers for cultivar development for the present and future (Table 11). Most responses indicated that DNA markers were either fairly unimportant (32) or had no utility (8) at the present time. However, there was considerably more optimism for the next 5-10 years. Over the next 5 years, all respondents expected DNA markers to become more important. In 10 years, DNA markers were expected to be fairly (31) or very (15) important for the production of cultivars. DNA markers have been involved in cultivar development: 41 responses indicated usage at some stage of cultivar development, although only 16 noted (3 others abstained) that the markers were used for cultivars released for commercial production (data not shown), The 16 respondents were asked to indicate how the markers were used in cultivar development (Table 111). Specific information about traits and breeding methods was not requested. Five categories of usage were indicated in the responses and most ( 1 9 16) selected more than one category. As expected, transfer of qualitative factors (native or transgenic) accounted for most of the activity, although transfer of QTL was indicated as frequently as transfer of transgenes. Within the next 5 years, expected use of DNA markers was substantially higher for all categories except “Other” [responses in this category included germ plasm protection (2), patenting (3), and retrospective analysis (l)]. For the near future, the more important applications involved transfer of transgenes or native monogenic factors, followed by parent selection and transfer of QTL. Considering the time required to develop a new cultivar or inbred line for commercial release and the recent origin of DNA markers and maps, the present level of utilization may be quite reasonable. Most of the usage probably involved some form of backcross breeding. In situations involving QTL, transgenes, and even some native monogenic factors, the linkage between DNA markers and genes of interest had to be determined before the initiation of the markerTable II Importance of DNA Markers in Cultivar Development No. of responses

Present Futurea

Very important

Fairly important

Fairly unimportant

No impact

No. opinion

1

14 31

32 8

8 0

0 1

15

a Respondents were asked to assess the importance of DNA markers for cultivar development programs in 10 years.

DNA MARKERS AND PLANT BREEDING PROGRAMS

323

Table IIl Present and Projected Applications of DNA Marker Technology for Cultivar Development

No. of responses Projected rankingb Applications

Present

ProjectedD

1

2

3

4

5

I . Transferring native monogenic factors using marker-assisted selection and any breeding method 2. Transferring quantitative trait loci using marker-assisted selection and any breeding method 3. Transferring transgenes using markerassisted selection and any breeding method 4. Parent selection (includes germ plasm surveys and prediction of progeny performance) 5. Other (please describe)

12

50

16

17

6

4

I

7

41

4

2

17

15

I

7

47

20

10

10

2

0

5

45

6 1 5

9

8

3

I

6

1

1

3

0

1

Projection within the next 5 years. Respondents were asked to rank items in order ( I through 5, I being most important). Most respondents assigned ranks to three or four items. a

facilitated program. By assuming this necessary calibration process required as long as 2 years (production of segregating generations, progeny development, data collection, and analysis), additional time would be required for the transfer. For most crops on the list, 3-4 backcrosses, often without selection, may be achieved in 1 year. By allowing 2 years for wide-scale testing to verify recovery of desired attributes of the donor and recurrent parents, the entire procedure may have required 3-6 years. In the most advanced situations (e.g., tomato, maize, rice, soybean), DNA-based maps using RFLP markers were developed by the mid-late 1980s, so that the first cultivars developed with their assistance would only be available for commercial release by the early 1990s (perhaps sooner for backcrossing monogenic factors) in most instances. For most other crops, maps and the initial mapping experiment are either only just concluded or under development and, therefore, not ready to be used in cultivar development. Respondents were asked to rate 11 applications of DNA markers for their present and potential (within 10 years) utility in cultivar development (Table IV). Regarding their present utility, applications of greater importance included, in approximate order, backcrossing transgenes, genetic mapping of qualitative factors (native and transgenes), transfer of native qualitative factors, mapping

3 24

M.LEE Table IV Ratinga of Present and Potential UUUty of Applications of DNA Markers in Cultlvar Development Programs No. of responses Present rating

Potentialb rating

Application

1

2

3

4

5

1

2

3

4

5

1. DNA fingerprinting parents for cre-

7

15

19

14

0

I7

19

15

2

0

2

10

I8

24

1

7

24

14

8

0

2 6 1 8 19 22

3 I1

7 1 3 7 1 0 3 0 30 17

4 6

1 0

1 0

2.

3. 4.

5. 6. 7. 8.

9. 10.

11.

ating source (i.e., base) populations DNA fingerprinting parents for predicting performance of progeny in hybrids Backcrossing transgenes Transferring native, qualitative (monogenic) factors (includes backcrossing) Transferring native quantitative factors (includes backcrossing) Genetic mapping of quantitative trait loci Genetic mapping of qualitative loci (native and transgenes) Map-based cloning (includes transposon tagging, chromosome walking) Monitor homozygosity in progeny (includes quality or purity control) DNA fingerprinting progeny in recurrent selection programs Other

4

11

22

15

3

15

22

10

5

1

16

16

14

8

I

23

16

7

6

I

23

20

10

I

1

27

21

2

2

I

10

II

15

12

7

16

17

10

4

6

13

16

13

12

1

25

14

7

7

I

3

7 1 7 2 3 5

6 1 5 2 0 7 5

3

0

3

0

0

0

0

0

0

0

Rating system: 1 = very important, 2 = somewhat important, 3 = somewhat unimportant, 4 = no utility, 5 = no opinion. Potential utility within the next 10 years.

QTL, monitoring homozygosity in progeny, map-based cloning, and DNA fingerprinting parents for source populations. The remaining items provide marginal, if any, utility for cultivar development. Perhaps one surprise in the ratings was the relatively high utility ascribed to the acquisition of genetic linkage information for qualitative as well as quantitative factors. The rating of the latter item is particularly interesting given the low ratings assigned to the transfer of quantitative factors. The prospects for the potential utility of most applications indicated increased importance with their relative order in close agreement with the ratings based on

325

DNA MARKERS AND PLANT BREEDING PROGRAMS

the present. A few notable differences included a substantial reduction in the number of ratings assigned to the category “no utility” for most items. Also, positive shifts in the ratings were evident for several applications, particularly DNA fingerprinting parents for source populations, transfer of QTL, and monitoring homozygosity in progeny. Ranking of the items regarding their present utility (Table V) revealed the same trends established in the ratings with one exception: the utility of map-based cloning declined considerably with a distribution of rank selections nearly identical to fingerprinting progeny in recurrent selection programs. Some of the decline in appraised utility for map-based cloning may be related to the frequency of crops listed as the primary basis for the responses (Table I). There is a preponderance of responses based on crops with very large genome sizes (e.g., maize, small grains, and soybean) and a dearth of input from researchers using crops of small genome size (e.g., rice and tomato), situations in which some methods of map-based cloning (i.e., chromosome walking) presumably would have their least and greatest utility, respectively. Map-based cloning has had limited opportunity to isolate important genes because of its currently Table V Rankingu of Applications of DNA Markers for Present Utility in Cultivar Development Programs No. of responses

Application

1

2

3

4

5

6

7

8

9 1 0 1 1 A v g

1 . DNA fingerprinting parents for

2

7

6

4

5

5

8

7

2

1

3

5.6

I

2

4

3

4

7

6 1 0

5

7

2

6.9

17 8 1 1 4 6 1 2 5 1 5

4 7

2 6

2 I

0 2

2 0

2 0

0 0

3.2 3.7

2.

3. 4.

5. 6. 7. 8. 9. 10. I I.

creating source populations DNA fingerprinting parents for predicting performance of proge“Y Back-crossing transgenes Transferring qualitative (monogenic) factors Transferring quantitative factors Genetic mapping of quantitative trait loci Genetic mapping qualitative loci Map-based cloning Monitoring homozygosity in progeny Fingerprinting progeny in recurrent selection programs Other Ranking system: I

=

1

0

6

6

3 6 1 3 1 0 1 5 6 1 1

6 3

4 3

6 2

1 6 . 7 1 4 . 7

8 1 3 1 2 5 1

7 5 8

7 I 6

6 5 3

6 2 2 4 9 1 0

3 0 3 1 3 5 0

0 8 5

0 7 0

3.6 7.6 5.5

2

1

1

2

3

1

3

7 1 4 1 3 3

8.0

1

2

0

1

0

0

0

0

2 7

greatest utility relative to other items on list.

0

0

0

326

M. LEE

incomplete development (e.g., chromosome walking in any plant spp. and transposon tagging systems in most plant spp.) or the dependency on other nascent technologies such as transformation and large-scale DNA cloning. However, one common variant of map-based cloning, transposon tagging, has been available for nearly 10 years in one large-genome crop, maize, and now in other crops. Perhaps the genes isolated to date have not contributed to significant genetic changes in crops, but reports of cloned genes for disease resistance in several crops (Ellis et al., 1992; Moffat, 1994) and male sterility in maize (Albertsen et al., 1993) may change some opinions in favor of this application. An assessment of the importance of several issues as limitations to using DNA markers for cultivar development is summarized in Table VI. First, each item was rated regarding the degree to which it was perceived to be a limitation. Next, items were selected and placed into one of two categories, least and most limiting, with a maximum of four items per category. Most (19/22) issues were identified by a majority of the respondents as being very or somewhat limiting, and 12 of those had a relatively high ratio of the two categories (No. very limiting/No. somewhat limiting): the number of genes affecting traits in populations used by breeders (28/16, 1.8); the cost of DNA marker technology (31/19, 1.6); throughput capability of DNA marker technology (24/20, 1.2; human resources required for the collection of DNA marker data (23/20, 1.2); lack of empirical evidence regarding the efficiency of MAS (21/20, 1.1); high efficiency of current breeding methods (18119, 0.95); lack of basic information on the genes (15120, 0.8); human resources required for analysis of trait and marker data (15125, 0.6); accuracy of genetic positions of DNA marker-trait associations (14/24,0.6); genotype by environment interaction (G X E; 18/28,0.6); the number of DNA marker loci (10/18, 0.55); and information management (12123, 0.5). On the basis of the number of responses and ratios of the rating categories, the first six issues would seem to represent common challenges for using DNA markers in cultivar development. Selection of the least and most limiting issues provided some confirmation and contrast relative to the previous ratings. Several issues (2, 3, 15, 17, and 21) remained among the most limiting and were joined by G X E as common, major concerns. Three issues (4, 7, and 22), previously rated among the more significant limitations were frequently selected for the category least limiting. Also, items pertaining to human resources were selected much less frequently in relation to the ratings. There is some evidence of substantially divided opinions for one issue, the number of DNA marker loci. Overall, the most prominent limitations relate largely to DNA technology (cost and capacity), plant biology (G X E, genes, and traits), and relative research experience (lack of empirical evidence for MAS and efficiency of current breeding methods). Of these, issues pertaining to plant biology probably would

Table VI Rating of Issues Regarding Their Present Importance as Limitations to Using DNA Markers in Cultivar Developmenta No. of responses Extremes Ratingb

~

Issue

1

2

3

4

Most limiting

Least limiting

I . Genetic heterogeneity among reference populations (i.e., same phenotype but different genotype) 2. Throughput capability of DNA technology (from DNA extraction through detection of DNA polymorphisms) 3. Cost of DNA marker technology 4. Number of DNA marker loci 5. Distribution (Genetic) of DNA marker loci 6. Collection of trait data (includes material and physical resources) 7. Infomation management 8. Methods of estimating genetic parameters 9. Recovery of desired recombinants 10. Genotype by environment interaction 1 I . Information on gene action (e.g., epistasis and dominance) 12. Accuracy of the genetic positions of DNA markertrait associations 13. Types of source populations used by breeders 14. Obtaining and disposing of chemicals and reagents used for detecting DNA polymorphisms 15. High efficiency of current breeding methods 16. Resistance to modifying established breeding methods 17. Number of genes affecting the trait(s) in populations used by breeders 18. Human resources required for trait assessment 19. Human resources required for collection of DNA data 20. Human resources required for analysis of trait and DNA data 21. Lack of empirical evidence regarding efficiency of DNA marker-assisted selection 22. Lack of basic information on the gene(s) (includes expression and function of gene product) 23. Other

8

18

22

6

8

12

24

20

10

0

20

7

31

19 18 27 25

4 25 20 21

0 I

28

3 21

I 1

4 6

15

11

23 30 30 28 30

18 14 16 7 11

I 3 2 I 2

7 7 3 21 5

15 6 4 I 3

14

24

14

2

8

5

7 I

20 12

25 35

2 6

5 I

15 16

18 5 28

19 23 16

15

26 9

2 0 I

12 4

18

18

5

12 23 15

28 20 25

13 11 14

I 0 0

6 9 5

9 6 7

21

20

13

0

16

6

15

20

18

1

6

15

I

0

0

0

I

0

10 6

7 12 7 6 18

10

14

6

a Respondents were asked to rate each issue regarding its present importance as a limitation to using DNA markers in cultivar development programs. Respondents were asked to rate the issues under the assumption that they would be able to use their favorite PCR-based detection method with a number of DNA marker loci equal to that of the current system. Respondents were asked to assign a rating according to the described system and, second, to assign a maximum of four items to each of two categories, most or least limiting. Rating system: 1 = very limiting, 2 = somewhat limiting, 3 = not limiting, and 4 = no opinion.

328

M.LEE

be selected as major limitations for any plant breeding method and program. Issues in the first and third categories are more specific to the purpose of the survey. These issues have a common theme, efficiency: the relative efficiencies of a very new tool and a very old technology-DNA markers and plant breeding. Advancements in DNA technology, basic knowledge of plant biology, and experience will increase the absolute efficiency of DNA markers. However, an efficiency gap may persist due to its original magnitude and improvements in plant breeding methodology. What is the potential for significant advancement in extant plant breeding methods? Obviously, the relative potential for advancements depends on myriad variables, such as crop, trait, environment, changes in extant technologies (e.g., plot combines, near-infrared reflectance analysis, etc.), and infrastructure. The keys to success, of course, will be identification of comparative advantages, appropriate integration of methods, and careful analysis of information.

VII. SUMMARY AND CONCLUSIONS For DNA markers and plant breeding programs, it is the best of times, and the worst of times. DNA markers have become established as another tool for many phases of crop improvement, but the utility of this technology varies considerably with the application and context of the crop and culture. For some applications, rigid assessments may be premature. Consider some of the following examples. In the 1940s, the prototype mainframe computer, Eniac, with its 18,000 vacuum tubes was envisioned to have a very limited market size (5?) by IBM. Yet subsequent innovations have made computers common items and essential components of plant breeding programs. There are numerous examples of inventions and technologies that required time to mature and contribute. The process involves vision, innovation, and patience (Rosenberg, 1994). Similar examples may be found in plant breeding, and one need only review the history of the early stages of the development of the inbred-hybrid concept for maize breeding. The ratio believer/skeptic was lower than today. Also, in the 1940s, many prominent maize breeders and geneticists were concerned that the genetic variation for grain yield had been exhausted (Hallauer and Miranda, 1981). Obviously, grain yields have increased in response to artificial selection. The world of DNA technology is changing rapidly, whereas plant breeding methodology has remained relatively stagnant. Thus, there is some reason to believe that innovations in the former may help it adapt to the speed and comprehensive scope of the latter. Prototype DNA combines have been constructed (Rafalski and Tingey, 1993). Methods for cloning and transferring very large

DNA MARKERS AND PLANT BREEDING PROGRAMS

329

DNA segments (Strauss et af., 1993) are developed and refined on a regular basis, and electronic catalogs of abstracted crop plant biology are being constructed. Further refinements in genome research organization should be considered and adapted from highly successful models (Moffat, 1992; Anderson, 1992; Stevens, 1994). No doubt, some of these efforts are in their “Eniac” stage of development and have a very bright future in the wake of large-scale genome projects (Copeland et af., 1993). Is this the dawn of the “biology” era of plant breeding? Previous innovations and technologies that had major impacts on the efficiency and productivity of plant breeding may be inherently different from today’s biotechnologies and related basic information. Computers, combines, and NIR analyzers were easily inserted into existing plant breeding methods. They allowed plant breeders to look at more genotypes in more environments. The newer biotechnologies are different. They will require plant breeders and their colleagues to look within the plant and understand its architecture before routine, beneficial, and predictable advancements. How will we assess the efficacy of DNA markers as a tool for crop improvement? In some situations, the accounting process may be easy. Important benchmarks have been established for several crops, breeding programs, and traits: rate and cost of gain, genetic component of gain, and time and cost required for cultivar development. Situations lacking such history and databases will be more difficult to assess. The first phase of the encounter between plant breeding and DNA markers has concluded. While the markers have unquestioned benefits for basic research, their utility for plant breeding remains to be established and verified in several aspects. Plant breeders have been justifiably skeptical given the history of hype associated with previous biotechnologies and other derivatives from basic biology (Simmonds, 1991). The efficiency of plant breeding programs, often with several decades of organized activity, will be difficult to enhance in a direct manner, but there is sufficient evidence to maintain an optimistic forecast (Section VI). Also, the comprehensive scope and myriad challenges of breeding programs will require an equally broad and critical assessment of new approaches and sources of information.

ACKNOWLEDGMENTS I thank members of the Lee Laboratory for their assistance with the literature review, colleagues for taking a few minutes from their busy schedules to respond thoughtfully to the survey, M. Lents and C. Wirth for typing the references, D. Austin for summarizing survey results and preparing figures, W. Beavis for sharing preprints and ideas, A.E. Melchinger and K . R. Lamkey for discussions, and members of the editorial board of Advances in Agronomy for the invitation to contribute to this

330

M. LEE

review. The review of literature for this chapter was concluded in December, 1994. Preparation of the review was supported partially by NRICGP Award 93-37300-8770 and Project 3134 of the Iowa Agriculture and Home Economics Experiment Station. This is journal paper no. 1-16237 of the Iowa Agriculture and Home Economics Experiment Station, Ames, IA.

REFERENCES Abler, B. S. B., Edwards, M. D., and Stuber, C. W. 1991. Isonenzymatic identification of quantitative trait loci in crosses of elite maize inbreds. Crop Sci. 31, 267-274. Ahn, S., Anderson, J. A., Somlls. M. E., and Tanksley, S . D. 1993. Homeologous relationships of rice, wheat and maize chromosomes. Mol. Gen. Genet. 241, 483-490. Ahnert, D., Austin, D., Lee, M., Livini, C., Woodman, W. L., Openshaw, S. J., Smith, J. S. C., Porter, K., and Dalton, G. 1995. Molecular genetic diversity among elite sorghum inbred lines. Crop Sci. in press. Akkaya, M. S., Bhagwat, A. A., and Cregan, P. B. 1992. Length polymorphisms of simple sequence repeat DNA in soybean. Generics 134, 1131-1 139. Albertsen, M. C., Fox, T. W., and Trimmel, M. R. 1993. Cloning and utilizing a maize nuclear male sterility gene. In “Proceedings of the Forty-Eighth Annual Corn and Sorghum Industry Research Conference” (D. Wilkinson, Ed.), Vol. 48, pp. 224-233. American Trade Seed Assoc., Washington, D.C. Alfenito, M. R., and Birchler, I. A. 1993. Molecular characterization of a maize B chromosome centric sequence. Genetics 135, 589-597. Allard, R. W. 1960. “Principles of Plant Breeding.” Wiley, New York. Allard, R. W. 1988. Genetic changes associated with the evolution of adaptedness in cultivated plants and their wild progenitors. J. Hered. 79, 225-238. Alrefai, R., Orozco, B., and Rocheford, T. R. 1994. Detection and sequencing of the transposable element ILS-I in the Illinois long-term selection maize strains. Plant Physiol. 106, 803-804. Anderson, C. 1992. US genome project does it the French way, conceding that size matters after all. Nature 360,401. Armstrong, C. L., Romero-Severson, J., and Hodges, T. K. 1992. Improved tissue culture response of an elite maize inbred through backcross breeding, and identification of chromosomal regions important for regeneration by RFLP analysis. Theorer. Appl. Genet. 84, 755-762. Arondel, V., Lemieux, B., Hwang, I., Gibson, S., Goodman, H. M., and Somerville, C. R. 1992. Map-based cloning of a gene controlling Omega-3 fatty acid desaturation in Arabidopsis. Science 258, 1353- 1354. Austin, D. F., and Lee, M. 1994. Comparative linkage analysis of RFLP loci and QTL in F2:3 and F6:7 recombinant inbreds. Maize Genet. Coop. Newsl. 68, 7-8. Austin, R. B. 1994. Crop breeding opportunities. In “Physiology and Determination of Crop Y i e l d (K. J. Boote, et al., Eds.). ASA, CSSA, and SSSA, Madison, WI. Baenziger, P. S., and Peterson, C. J. 1992. Genetic variation: Its origin and use for breeding selfpollinated species. In “Plant Breedings in the 1990s” (H. T. Stalker and J. P. Murphy, Eds.), pp. 69-92. CAB International, Wallingford, UK. Baht-Kurti, P. J., Dixon, M.S., Jones, D. A., Norcott, K. A., and Jones, J. D. G. 1994. RFLP linkage analysis of the Cf-4 and Cj-9genes for resistance to Cladosporiumfulvum in tomato. Theoret. Appl. Genet. 88, 691-700. Barton, N. H. 1990. Pleiotropic models of quantitative variation. Genetics 124, 773-782. Beavis. W. D. 1994. The power and deceit of QTL experiments. I n “Forty-Ninth Proc. of the Annual Corn & Sorghum Industry Research Conference” (D. Wilkinson, Ed.), in press. American Seed Trade Association, Washington, DC.

DNA MARKERS AND PLANT BREEDING PROGRAMS

33 1

Beavis, W. D., and Keim, P. 1995. Identification of QTL that are affected by environment. In “New Perspectives on Genotype-by-Environment Interaction” (M. Kang, Ed.), in press. CRC Press, Boca Raton. FL. Beavis. W. D., Grant, D., Albertsen, M., and Fincher, R. 1991. Quantitative trait loci for plant height in four maize populations and their associations with qualitative genetic loci. Theorer. Appl. Genet. 83, 141-145. Beavis, W. D., Lee, M., and Grant, D., Hallauer, A. R., Owens, T., Katt, M., and Blair, D. 1992. The influence of random mating on recombination among RFLP loci. Maize Genet. Coop. N e d 66,52-53.

Beckmann, J. S . , and Osborn, T. S. (eds.). 1992. “Plant Genomes: Methods for Genetic and Physical Mapping.” Kluwer Academic, Dordrecht, The Netherlands. Beer, S . C.. Goffreda. J., Phillips, T. D., Murphy, J. P., and Somlls, M. E. 1993. Assessment of genetic variation in Avenu srerilis using morphological traits, isozymes. and RFLR. Crop Sci. 33, 1386-1393. Bennetzen, J. L.. and Freeling, M. 1993. Grasses as a single genetic system-genome composition, collinearity, and compatibility. Trends Genet. 9, 259-261. Bernardo, R. 1993. Estimation of coefficient of coancestry using molecular markers in maize. Theorer. Appl. Genet. 85, 1055-1062. Bernardo, R. 1994. Prediction of maize single-cross performance using RFLR and information from related hybrids. Crop Sci. 34, 20-25. Besse, P., Seguin, M., Lebrun, P., Chevallier, M. H., Nicolas, D., and Lanaud, C. 1994. Genetic diversity among wild and cultivated populations of Hevea brasiliensis assessed by nuclear RFLP analysis. Theorer. Appl. Genet. 88, 199-207. Birchler. 1. A. 1993. Dosage analysis of maize endosperm development. Annu. Rev. Genet. 27, 181204.

Bogenschutz, T. G., and Russell, W. A. 1986. An evaluation for genetic variation within inbred lines maintained by sib-mating and self-pollination. Euphyrica 35, 403-412. Boppenmaier, J., Melchinger, A. E., Brunklaus-Jung, E., Geiger, H. H., and Hemnann, R. G. 1992. Genetic diversity for RFLR in European maize inbreds. I. Relation to performance of flint X dent crosses for forage traits. Crop Sci. 32, 895-902. Bretting, P. K., and Widrlechner, M. P. 1995. Genetic markers and plant genetic resource management. In “Plant Breeding Reviews” (J. Janick, Ed.), 13, 11-86. Briggs, F. N., and Knowles, P. F. 1967. “Introduction to Plant Breeding.” Reinhold, New York. Briggs, S . P., and Beavis, W. D. 1994. How RFLP loci can be used to assist transposon-tagging efforts. In ‘The Maize Handbook (M. Freeling and V. Walbot, Eds.), pp. 653-659, SpringerVerlag, New York. Bubeck, D. M., Goodman, M. M., Beavis, W. D.. and Grant, D. 1993. Quantitative trait loci controlling resistance to gray leaf spot in maize. Crop Sci. 33, 838-847. Carbonell, E. A., Gerig, T. M.. Balansard, E., and Asins, M. J. 1992. Interval mapping in the analysis of nonadditive quantitative trait loci. Eiornetrics 48, 305-315. Carbonell, E. A., Asins, M. J., Baselga, M., Balansard, E., and Gerig, T. M. 1993. Power studies in the estimation of genetic parameters and the localization of quantitative trait loci for backcross and doubled haploid populations. Theorer. Appl. Genet. 86, 41 1-416. Carland, F. M.. and Staskewicz, B. J. 1993. Genetic characterization of the Pro locus of tomato: semi-dominance and cosegregation of resistance to Pseudonwnus syringae pathovar tomato and sensitivity to the insecticide Fenthion. Mol. Gen. Gener. 239, 17-27. Carlson, W. R. 1988. The cytogenetics of corn. I n “Corn and Corn Improvement”(G. F. Sprague and J. W. Dudley, Eds.), 3rd ed., pp. 259-343. Am. Soc. Agron., Madison, WI. Castiglione, S . , Wang. G., Damiani, G., Bandi, C., Bisoffi, S . , and Sala, F. 1993. RAPD fingerprints for identification and for taxonomic studies of elite poplar (Populus spp.) clones. Theorer. Appl. Genet. 87, 54-59.

M. LEE

332

Chandler, V. L., Radicella, J. P., Robbins, T. P., Chen, J., and lbrks, D. 1989. n o regulatory genes of the maize anthocyanin pathway are homologous: Isolation of B utilizing R genomic sequences. Pfanr Cell 1, 1175-1 183. Chao, S., Baysdorfer, C., Heredia-Diaz, 0.. Musket, T., Xu, G., and Coe, E. H.,Jr. 1994. RFLP mapping of partially sequenced leaf cDNA clones in maize. Theorer. Appl. Genet. 88,7 17-72 I . Chen, F. Q., Prehn, D., Hayes, P. M., Mulrooney, D., Corey, A., and Vivar. H. 1994. Mapping genes for resistance to barley stripe rust (Puccinia striiformis f. sp. hordei). Theorer. Appl. Genet. 88, 215-219.

Cheung, W. Y.,Chao, S., and Gale, M. D. 1991. Long-range physical mapping of the a-amylase-l (a-Amy-I) loci on homoeologous group 6 chromosomes of wheat. Mol. Gen. Genet. 229,373379.

Chourey, P. S., and Taliercio, E. W. 1994. Epistatic interaction and functional compensation between the two tissue- and cell-specific sucrose synthase genes in maize. Proc. Narl. Acud. Sci. USA 91, 7917-792 I . Civardi, L., Xia, Y. Edwards, K. J., Schnable, P. S., and Nikolau, B. J. 1994. The relationship between genetic and physical distances in the cloned al-sh2 interval of the Zea mays L. genome. Proc. Natl. Acad. Sci. USA 91, 8268-8272. Coe, E. H., Jr., Neuffer, M. G., and Hoisington, D. A. 1988. The genetics of corn. In “Corn and Corn Improvement” (G. F. Sprague and J. W. Dudley, Eds.), 3rd ed., pp. 81-258. Am. Soc. Agron., Madison, WI. Connolly, A. G., Godwin, I. D., Cooper, M., and DeLacy, I. H. 1994. Interpretation of randomly amplified polymorphic DNA marker data for fingerprinting sweet potato (Ipomoea bataras L.) genotypes. Theorer. Appl. Genet. 88, 332-336. Copeland, N. G., Jenkins, N., Gilbert, D., Eppig, J., Maltais, L., Miller, J., Dietrich, W., Weaver, A., Lincoln, S . , Steen, R., Stein, L., Nadeau, J., and Lander, E. 1993. A genetic linkage map of the mouse: Current applications and future prospects. Science 262, 57-66. Cowen, N. M., and Frey, K. J. 1987a. Relationship between genealogical distance and breeding behavior in oats (Avenn saliva L.). Euphyrica 36, 413-424. Cowen, N. M., and Frey, K. J. 1987b. Relationship between three measures of genetic distance and breeding behavior in oats (Avenn saliva L.). Genome 29, 97-106. Cox, T. S., and Murphy, J. P. 1990. The effect of parental divergence on F, heterosis in winter wheat crosses. Theoret. Appl. Genet. 79, 241-250. COX, T. S., Kiang, Y. T., Gorman, M. B., and Rodgers, D. M. 1985. Relationship between coefficient of parentage and genetic similarity indices in the soybean. Crop Sci. 25, 529-532. Dale, E. C., and Ow, D. W. 1991. Gene transfer with subsequent removal of the selection gene from the host genome. Proc. Narl. Acad. Sci. USA 88, 10558-10562. Darvasi, A., and Soller, M. 1992. Selective genotyping for determination of linkage between a marker locus and a quantitative trait locus. Theorer. Appl. Genet. 85, 353-359. Darvasi, A., and Weller, J. 1. 1992. On the use of moments method of estimation to obtain approximate maximum likelihood estimates of linkage between a genetic marker and a quantitative locus. Heredig 68,4346. Darvasi, A., Weinreb, A., Minke, V., Weller. J. I., and Soller, M. 1993. Detecting marker-QTL linkage and estimating QTL gene effect and map location using a saturated genetic map. Genetics 134, 943-95 1. Das, 0. P, and Messing, I. W. 1987. Allelic variation and differential expression at the 27-kilodalton zein locus in maize. Mol. Cell. Biol. 7, 490-4497. de Koning, G. J., and Weller, J. 1. 1994. Efficiency of direct selection on quantitative trait loci for a two-trait breeding objective. Theoret. Appl. Genet. 88, 669-677. de Vincente, M. C., and Tanksley, S. D. 1993. QTL analysis of transgressive segregation in an interspecific tomato cross. Generics 134, 585-596.

DNA MARKERS AND PLANT BREEDING PROGRAMS

333

Devos, K. M., Millan, T.,and Gale, M. D. 1993a. Comparative RFLP maps of the homoeologous group-2 chromosomes of wheat, rye and barley. Theoret. Appl. Genet. 85, 784-792. Devos, K. M., Atkinson, M. D., Chinoy, C. N., Francis, H. A., Harcourt, R. L., Koebner, R. M. D., Liu, C. J., Masojc, P., Xie, D. X.,and Gale, M. D. 1993b. Chromosomal rearrangements in the rye genome relative to that of wheat. Theoret. Appl. Genet. 85, 673-680. Doebley, J., and Stec. A. 1993. Inheritance of the morphological differences between maize and teosinte: Comparison of results for two F, populations. Genetics 134, 559-570. Doebley. J., and Stec, A. 1994. Teosinte branched1 and the origin of maize. Maize Cener. Coop. Newsl. 68, 88-89. Doebley, J. F., and Wendel, J. F. 1989. Applications of RFLR to plant systematics. In “Current Communications in Molecular Biology: Development and Application of Molecular Markers to Problems in Plant Genetics” (T. Helentjaris and B. Burr, Eds.), pp. 57-67. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Doebley, J. Stec, A., Wendel, J., and Edwards, M. 1990. Genetic and morphological analysis of a maize-teosinte F, population: Implications for the origin of maize. Proc. Natl. Acad. Sci. USA 87, 9888-9892. Dong, J., Cox, T. S . , Sears, R. G., and Lookhart, G. L. 1991. High molecular weight glutenin genes: Effects on quality in wheat. Crop Sci. 31, 974-979. Dong, H., Sears, R. G., Cox, T. X.,Hoseney, R. C., Lookhart, G. L., and Shogren, M. D. 1992. Relationships between protein composition and mixograph and loaf characteristics in wheat. Cereal Chem. 69, 132-136. Dooner, H. K. 1986. Genetic fine structure of the bronze locus in maize. Genetics 113, 1021-1036. Dooner, H. K. 1994. Genetic fine structure from testcross progeny analysis. In “The Maize Handbook” (M. Feeling and V. Walbot. Eds.), pp. 303-306. Springer-Verlag, New York. Donveiler, J., Stec, A,, Kermicle, J., and Doebley, J. 1993. Teosinte g l u m architecture I : A genetic locus controlling a key step in maize evolution. Science 262, 233-235. Douches, D. S . , Ludlam, K., and Freyre, R. 1991. lsozyme and plastid DNA assessment of pedigrees of nineteenth century potato cultivars. Theoret. Appl. Genet. 82, 195-200. Dudley, J. W. 1987. Modification of methods for identifying populations to be used for improving parents of elite single crosses. Crop Sci. 27, 940-943. Dudley, J. W. 1988. Evaluation of maize populations as sources of favorable alleles. Crop Sci. 28, 486-491. Dudley, J. W. 1992. Theory for identification of marker locus-QTL associations in population by line crosses. Theoret. Appl. Genet. 85, 101-104. Dudley, J. W. 1993. Molecular markers in plant improvement: Manipulation of genes affecting quantitative traits. Crop Sci. 33, 660-668. Dudley, J. W., Saghai-Maroof. M. A,. and Rufener, G. K. 1991. Molecular markers and grouping of parents in maize breeding programs. Crop Sci. 31, 718-723. Dudley, J. W., Saghai-Maroof, M. A., and Rufener, G.K. 1992. Molecular marker information and selection of parents in corn breeding programs. Crop Sci. 32, 301-304. Duncan, R. R., BrameLCox, P. J., and Miller, F. R. 1991. “Contributions of Introduced Sorghum Germplasm to Hybrid Development in the USA” CSSA Special Publication No. 17, pp. 69101. Crop Sci. Soc. Am., Madison, WI. Duvick, D. N. 1984. Progress in Conventional Plant Breeding. In “Gene Manipulation in Plant Improvement” (J. P. Gustafson, Ed.),pp. 17-30. Plenum Press, New York. Edwards, M. 1992. Use of molecular markers in the evaluation and introgression of genetic diversity for quantitative traits. Field Crops Res. 29, 241-260. Edwards, M., and Johnson, L. 1994. RFLR for rapid recurrent selection. In “Proc. of the Symposium Analysis of Molecular Marker Data,” pp. 33-40. Am. Soc.Hort. Sci. and Crop Sci. Soc. Am., Madison, WI.

334

M. LEE

Edwards, M.D., and Page, N. J. 1994. Evaluation of marker-assisted selection through computer simulation. Theorer. Appl. Genet. 88, 376-382. Edwards, M. D., Stuber, C. W., and Wendel, J. F. 1987. Molecular-marker-facilitated investigations of quantitative-trait loci in maize. I. Numbers, genomic distribution and types of gene action. Generics 116, 1 13- 125. Ellis, I. G., Lawrence, G. J., Peacock, W. J., and Pryor, A. J. 1988. Approaches to cloning plant genes conferring resistance to fungal pathogens. Annu. Rev. Phyroparhol. 26, 245-263. Ellis, J. G., Finnegan, E. J., and Lawrence, G. J. 1992. Developing a transposon tagging system to isolate rust-resistance genes from flax. Theorer. Appl. Genet. 85, 46-54. Eshed, Y., and Zamir, D. 1994. Introgressions from Lycopersiconpennellii can improve the solublesolids yield of tomato hybrids. Theorer. Appl. Genet. 88, 891-897. Fatokun, C. A., Menancio-Hautea, D. I., Danesh, D., and Young, N. D. 1992. Evidence for orthologous seed weight genes in cowpea and mung bean based on RFLP mapping. Generics 132, 841-846. Fauron, C. M.-R., and Capser, M. 1994. A second type of normal maize mitochondria1 genome: An evolutionary link. Generics 137, 875-882. Fedoroff, N. V., Furtek, D. B., and Nelson, 0. E., Jr. 1984. Cloning of the bronze locus in maize by a simple and generalizable procedure using the transposable element Acrivarion (Ac.). Proc. Narl. Acad. Sci. USA 81, 3825-3829. Fehr, W. R. (Ed.). 1984. “Genetic Contributions to Yield Gains of Five Major Crop Plants,” Special Publication No. 7. Crop Sci. Soc. of Am., Madison, WI. Fehr. W. R. 1987. “Principles of Cultivar Development,” Vol. I , Crop Species. Macmillan, New York. Fischer, R. A. 1993. Cereal breeding in developing countries: Progress and prospects. I n “lnternational Crop Science I” (D. R. Buxton, R. Shibles, R. A. Forsberg, B. L. Blad, K. H. Asay, G. M. Paulson, and R. F. Wilson, Eds.), pp. 201-209. Crop Sci. Soc. Am., Madison,

WI. Flavell, R. B. 1994. Inactivation of gene expression in plants as a consequence of specific sequence duplication. Pmc. Narl. Acad. Sci. USA 91, 3490-3496. Frei, 0. M., Stuber, C. W., and Goodman, M. M. 1986. Use of allozymes as genetic markers for predicting performance in maize single-cross hybrids. Crop Sci. 26, 37-42. Freymark, P.J., Lee, M., Woodman, W. L., and Mattinson, C. A. 1993. Quantitative and qualitative trait loci affecting host-plant response to Exsemhilum rurcicum in maize (Zea mays L.). Theorer. Appl. Genet. 87, 537-544. Ganal, M. W., Young, N. D., and Tanksley, S. D. 1989. Pulsed field gel electrophoresis and physical mapping of large DNA fragments in the Tm-2a region of chromosome 9 in tomato. Mol. Gen. Genet. 215, 395-400. Gebhardt, C., Ritter, E., Baroue, A., Debner, T., Walkmeier, B., Schachtschabel, U., Kaufmann, H., Thompson, R. D., Bonierbale. M. W., Ganal, M. W., Tanksley, S. D., and Salamini, F. 1991. RFLP maps of potato and their alignment with the homeologous tomato genome. Theorer. Appl. Genet. 83, 49-51. Cedes, J. T., and Tracy, W. F. 1994. Diversity of historically important sweet corn inbreds as estimated by RFLR, morphology, isozymes, and pedigree. Crop Sci. 34, 26-33. Gerloff, J. E.,and Smith, 0. S. 1988. Choice of method for identifying germplasm with superior alleles. I. Theoretical results. Theorer. Appl. Genet. 76, 209-216. Gill, K. S., Gill, B. S., Endo. T. R., and Mukai, Y. 1993. Fine physical mapping of P h l . a chromosome pairing regulator gene in polyploid wheat. Generics 134, 1231-1236. Gimelfarb, A., and Lande, R. 1994. Simulation of marker assisted selection in hybrid populations. Genet. Res., Camb. 63. 39-47.

DNA MARKERS AND PLANT BREEDING PROGRAMS

335

Goddard, M. E. 1992. A mixed model for analyses of data on multiple genetic markers. Theore?. Appl. Genet. 83, 878-886. Godshalk, E. B., Lee, M.,and Lamkey, K. R. 1990. Relationship of restriction fragment length polymorphisms to single-cross hybrid performance of maize. Theorer. Appl. Genet. 80, 273280. Goffreda, J. C., Bumquist, W. B., Beer, S . C., Tanksley, S. D., and Sorrells, M. E. 1992. Application of molecular markers to assess genetic relationships among accessions of wild oat, Avenu sterilis. Theorer. Appl. Genet. 85, 146- 151. Gohal, G. S., and Briggs, S. P. 1992. Reductase activity encoded by the HMI disease resistance gene in maize. Science 258, 985-987. Goldman, 1. L., Rocheford, T. R., and Dudley, J. W. 1993. Quantitative trait loci influencing protein and starch concentration in the Illinois long term selection maize strains. Theorer. Appl. Genet. 87, 217-224. Grabau, E., Davis, W. H., Phelps, N. D., and Gengenbach, B. G. 1992. Classification of soybean cultivars based on mitochondria1 DNA restriction fragment length polymorphisms. Crop Sci. 32, 271-274. Graner, A., Ludwig, F.. and Melchinger, A. E. 1994. Relationships among European barley germplasm. 11. Comparison of RFLP and pedigree data. Crop Sci. 34, 1199-1205. Grotewold, E., and Peterson, T. 1994, Isolation and characterization of a maize gene encoding chalcone flavonone isomerase. Mol. Gen. Genet. 242, 1-8. Haley, C. S . , and Knott, S . A. 1992. A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. Heredity 69, 315-324. Haley, S. D., Afanador, L., and Kelly, J. D. 1994. Selection for monogenic pest resistance traits with coupling- and repulsion-phase RAPD markers. Crop Sci. 34, 1061-1066. Halford, N. G., Field, J. M., Blair, H., Unvin, P., Moore, K., Robert, L., Thompson, R., Flavell, R. B., Tatham, A. S., and Shewry, P. R. 1992. Analysis of HMW glutenin subunits encoded by chromosome 1A of bread wheat (Triricum uestivum L.) indicates quantitative effects on grain quality. Theorer. Appl. Genet. 83, 373-378. Hallauer, A. R. 1990. Methods used in developing maize inbreds. Muydicu 35, 1-16. Hallauer, A. R., and Miranda, J. B. 1981. “Quantitative Genetics in Maize Breeding.” Iowa State Univ. Press, Ames, IA. Hallauer, A. R., Russell, W. A., and Lamkey, K. R. 1988. Corn breeding. In “Corn and Corn Improvement” (G. F. Sprague and J. W. Dudley, Eds.), 3rd ed., pp. 463-564. Am. Soc. Agron., Madison, W1. Havey, M. J. 1993. A putative donor of S-cytoplasm and its distribution among open-pollinated populations of onion. Theorer. Appl. Gene?. 86, 128-134. Hayashi, T., and Ukai, Y. 1994. Detection of additive and dominance effects of QTLs in interval mapping of F2 RFLP data. Theoret. Appl. Genet. 87, 1021-1027. Hayes, P. M., Liu, B. H., Knapp. S. J., Chen, F., Jones, B., Blake, T.,Franckowiak, J., Rasmusson. D., Sorrells, M., Ullrich, S.E., Wesenberg, D., and Kleinhofs, A. 1993. Quantitative trait locus effects and environmental interaction in a sample of North American barley germ plasm. Theorer. Appl. Gene?. 87, 392-401. He, S . , Ohm, H., and Mackenzie, S. 1992. Detection of DNA sequence polymorphisms among wheat varieties. Theorer. Appl. Genet. 84, 573-578. Healy, J., Corr, C., DeYoung, J., and Baker, B. 1993. Linked and unlinked transposition of a genetically marked Dissociation element in transgenic tomato. Genetics 134, 57 1-584. Helentjaris, T. G. 1992. RFLP analyses for manipulating agronomic traits in plants. I n “Plant Breeding in the 1990s” (H. T. Stalker and J. P. Murphy, Eds.), pp. 357-372. CAB International, United Kingdom.

336

M. LEE

Helentjaris, T. 1993. Implications for conserved genomic structure among plant species. Proc. N d . Acud. Sci. USA 90,8308-8309. Heun, M. 1992. Mapping quantitative powdery mildew resistance of barley using a restriction fragment length polymorphism map. Genome 35, 1019- 1025. Hoeschele, I . , and Van Raden, P. M. 1993. Bayesian analysis of linkage between genetic markers and quantitative trait loci. I. Prior knowledge. Theorer. Appl. Gener. 85, 953-960. Hong, K. S., Richter, T. E., Bennetzen, J. L., and Hulbert, S. H. 1993. Complex duplications in maize lines. Mol. Gen. Genet. 239, 115-121. Hospital, F., Chevalet, C., and Mulsant, P. 1992. Using markers in gene introgression breeding programs. Generics 132, I I 99- I2 10. Howell, E. C., Newbury, H. J., Swennen, R. L.. Withers, L. A., and Ford-Lloyd, B. V. 1994. The use of RAPD for identifying and classifying Musu germplasm. Genome 37, 328-332. Huffman, W. E.. and Evenson, R. E. 1993. “Science for Agriculture: A Long Term Perspective.” Iowa State Univ. Press, Ames, IA. Ingelbrecht, I., Van Houdt, H., Van Montagu, M., and Depicker, A. 1994. Posttranscriptional silencing of reporter transgenes in tobacco correlates with DNA methylation. Proc. Nutl. Acud. Sci. USA 91, 10502-10506. Jansen, R. C. 1992. A general mixture model for mapping quantitative trait loci by using molecular markers. Theorer. Appl. Genet. 85, 252-260. Jansen, R. C. 1993. Interval mapping of multiple quantitative trait loci. Generics 135, 205-211. Jansen, R. C., and Stam, P. 1994. High resolution of quantitative traits into multiple loci via interval mapping. Generics 136, 1447-1455. Khush, G. V. 1993. Breeding rice for sustainable agricultural systems. .In “International Crop Science I” (D. R. Buxton, R. Shibles, R. A. Forsberg, B. L. Blad, K. H. Asay, G. M. Paulson, and R. F. Wilson, Eds.), pp. 189-199. Crop Sci. Soc. Am., Madison, WI. Kilian, A., and Kleinhofs, A. 1992. Cloning and mapping of telomere associated sequence from Hordeum vulgure L. Mol. Gen. Genet. 235, 153-156. Kim, T. S., Phillips, R. L., and Eun, M. Y. 1993. Identification of genomic regions controlling maturity in maize (Zea mays L.). In “Crop Production and Improvement Technology in Asia,” pp. 379-397. Korean Society of Crop Science, Korea. Knapp, S. J. 1994. Mapping quantitative trait loci. In “DNA-Based Markers in Plants” (R.L. Phillips and 1. K. Vasil, Eds.), pp. 58-96. Kluwer Academic, Dordrecht, The Netherlands. Knapp, S. J., Bridges, W. C., and Liu, B. H. 1992. Mapping quantitative trait loci using nonsimultaneous and simultaneous estimators and hypothesis tests. In ‘‘Plant Genomes: Methods for Genetic and Physical Mapping” (J. S. Beckman and T. S. Osborn, Eds.), pp. 209-237. Kluwer, Dordrecht, The Netherlands. Knott, S. A., and Haley, C. S. 1992a. Aspects of maximum likelihood methods for the mapping of quantitative trait loci in line crosses. Genet. Res. 60, 139-151. Knott, S. A., and Haley, C. S. 1992b. Maximum likelihood mapping of quantitative trait loci using full-sib families. Generics 132, 121 1- 1222. Koester, R. P., Sisco, P. H., and Stuber, C. W. 1993. Identification of quantitative trait loci controlling days to flowering and plant height in two near isogenic lines of maize. Crop Sci. 33, 1209- 1216. Kresovich, S., and McFerson, J. R. 1992. Assessment and management of plant genetic diversity: considerations of intra- and interspecific variation. Field Cmp Res. 29, 185-204. Kresovich, S., Williams, J. G . K., McFerson, J. R., Routman, E. I., and Schaal, B. A. 1992. Characterization of genetic identities and relationships of Erussicu oleruceu L. via a random amplified polymorphic DNA assay. Theorer. Appl. Gener. 85, 190-196. Lai, C., Lyman, R. F., Long, A. D., Langley. C. H., and Mackay, T. F. C. 1994. Naturally

DNA MARKERS AND PLANT BREEDING PROGRAMS

337

occurring variation in bristle number and DNA polymorphisms at the scabrous locus of Drosophila melanogasrer. Science 266, 1697- 1702. Lamkey, K. R., and Lee, M. 1993. Quantitative genetics, molecular markers, and plant improvement. In “Focused Plant Improvement: Towards Responsible and Sustainable Agriculture: (B. C. Imrie and J. B. Hacker, Eds.), Proc. 10th Australian Plant Breeding Conf., Gold Coast, 18-23 April 1993, pp. 104-1 15. Organising Committee, Australian Convention and Travel Service, Canberra, Australia. Lamkey, R., Peterson, P. A., and Hallauer, A. R. 1991. Frequency of the transposable element Uq in Iowa stiff stalk synthetic maize populations. Gener. Res., Camb. 57, 1-9. Lamkey, K. R., Schnicker, B. J., andcocken, T. L. 1993. Choiceofsource populations for inbred line improvement. I n “Proc. of the Forty-Eighth Annual Corn & Sorghum Industry Research Conference” (D. Wilkinson, Ed.), pp. 91- 103. American Seed Trade Association, Washington, DC. Lande, R. 1992. Marker-assisted selection in relation to traditional methods of plant breeding. I n “Plant Breedings in the 1990s” (H. T. Stalker and J. P. Murphy, Eds.), pp. 437-451. CAB International. United Kingdom. Lande, R., and Thompson, R. 1990. Efficiency of marker-assisted selection in the improvement of quantitative traits. Generics 124, 743-756. Lander, E. S. 1989. DNA fingerprinting on trial. Nature 339, 501-505. Lander, E. S., and Botstein, D. 1986. Strategies for studying heterogeneous genetic traits in humans by using a linkage map of restriction fragment length polymorphisms. Proc. Nail. Acad. Sci. USA 83, 7353-7357. Lark, K.G., Orf, J., and Mansur, L. M. 1994. Epistatic expression of quantitative trait loci (QTL) in soybean [Glycine max (L.) Merr.] determined by QTL association with RFLP alleles. Theoret. Appl. Gener. 88, 486-489. Laurent, V.. Ristemcci, A. M., and Lanaud, C. 1994. Genetic diversity in cocoa revealed by cDNA probes. Theoret. Appl. Genet. 88, 193-198. Lawrence, P. L., and Frey. K. J. 1976. Inheritance of grain yield in oat species crosses (Avena saliva L. X A. srerilis L.) Egypt. J . Genet. Cytol. 5 , 400-409. Lee, E. A., Lee, M.. and Lamkey, K. R. 1990. RFLP analysis of isogenic lines B14 and B14A. Maize Genet. Coop. Newsl. 64, 20. Lee, M. 1993. Genetic analysis of resistance to European corn borer and northern corn leaf blight in maize. I n “Proc. of the Forty-Eighth Annual Corn & Sorghum Industry Research Conference” (D. Wilkinson, Ed.), pp. 213-223. American Seed Trade Association, Washington, DC. Lee, M., Godshalk, E. B., Lamkey, K. R., and Woodman, W. W. 1989. Association of restriction fragment length polymorphisms among maize inbreds with agronomic performance of their crosses. Crop Sci. 29, 1067-1071. Leitch, 1. J., and Heslop-Harrison. J. S. 1993. Physical mapping of four sites of 5s rDNA sequences and one site of the a-amylase-2 gene in barley (Hordeum vulgare). Genome 36, 517-523. Lewontin, R. C. 1977. The relevance of molecular biology to plant and animal breeding. I n “Proc. of the Int. Conf. in Quantitative Genetics,” 16-21 August 1976 (E. Pollak, 0. Kempthorne, and T. B. Bailey, Jr., Eds.), pp. 55-62. Iowa State Univ. Press, Ames, IA. Livini, C., Ajmone-Marsan, P., Melchinger, A. E., Messmer, M. M., and Motto, M. 1992. Genetic diversity of maize inbred lines within and among heterotic groups revealed by RFLR. Theorer. Appl. Genet. 84, 17-25. Lonnquist, J. H. 1974. Consideration and experiences with recombinations of exotic and corn belt maize germplasms. Annu. Corn Sorghum Res. Conf. Proc. 29, 102-1 17. Lorenz, M., Weihe, A., and Borner, T. 1994. DNA fragments of organellar origin in random amplified polymorphic DNA (RAPD) patterns of sugar beet (Eera vulgaris L.). Theorer. Appl. Gener. 88, 775-779.

M. LEE

338

Lubbers, E. L., Gill, K. S., Cox, T. S., and Gill, B. S. 1991. Variation of molecular markers among geographically diverse accessions of Triricum tauschii. Genome 34, 354-361. Lukaszewski, A. J. 1992. A comparison of physical distribution of recombination in chromosome IR in diploid rye and in hexaploid triticale. Theorer. Appl. Genet. 83, 1048-1053. Lukaszewski, A. J.. and Curtis, C. A. 1993. Physical distribution of recombination in B-genome chromosomes of tetraploid wheat. Theoret. Appl. Genet. 86, 121-127. Luo, Z. W., and Kearsey, M. J. 1992. Interval mapping of quantitative trait loci in an F2 population. Heredity 69, 236-242. Luo, Z. W., and Woolliams, J. A. 1993. Estimation of genetic parameters using linkage between a marker gene and a locus underlying a quantitative character in F2 populations. Heredity 70, 245-253.

Lynch, M. 1988. Estimation of relatedness by DNA fingerprinting. Mol. Biol. Evol. 5 , 584-599. Mackay, T.F. C., and Langley, C. H. 1990. Molecular and phenotypic variation in the achaere-scure region of Drosophila melanogaster. Nature 348,64-66. Maltcot, G. 1948. “Les Mathtmatiques de I’Hbt6ditt.” Masson et Cie, Paris. Mansur, L. M., Qualset, C. O., Kasarda, D. D., and Moms, R. 1990. Effects of ‘Cheyenne’ chromosomes on milling and baking quality in ‘Chinese Spring’ wheat in relation to glutenin and gliadin storage proteins. Crop Sci. 30,593-602. Mariani, C., Gosselb, V., De Beuckeleer, M., De Block, M., Goldberg, R., De Greef, W., and Leemans, J. 1992. A chimeric ribonuclease-inhibitor restores fertility to male sterile plants. Nature 357, 384-387. Marshall, D. R. 1990. Crop genetic resources: current and emerging issues. In “Plant Population Genetics, Breeding, and Genetic Resources” (A. H. D. Brown, M. Clegg, A. Kahler, and B.S. Weir, Eds.), pp. 367-388. Sinauer Associates, Sunderland, MA. Martin, G. B., Brommonschenkel, S. H.,Chunwongse, J., Frary, A,, Ganal, M. W., Spivey, R., Wu, T., Earle, E. D., and Tanksley, S. D. 1993. Map-based cloning of a protein kinase gene conferring disease resistance in tomato. Science 262, 1432- 1436. Martinez, O., and Curnow, R. N. 1992. Estimating the locations and the sizes of the effects of quantitative trait loci using flanking markers. Theorer. Appl. Genet. 85,480-488. Martinez, O . , and Cumow, R. N. 1994. Missing markers when estimating quantitative trait loci using regression mapping. Heredity 73, 198-206. Mather, K. 1941. Variation and selection of polygenic characters. J. Genet. 41, 159-193. McCouch, S. R., Kochert, G., Yu, Z. H., Wang, Z. Y.,Khush, G. S., Coffman, W. R., and Tanksley, S. D. 1988. Molecular mapping of rice chromosomes. Theorer. Appl. Genet. 76,815829.

McGrath, J. M., and Quiros, C. F. 1992. Genetic diversity at isozyme and RFLP loci in Brassica campestris as related to crop type and geographical origin. Theoret. Appl. Genet. 83, 783-790. McMullen, M. D., and Louie, R. 1989. The linkage of molecular markers to a gene controlling the symptom response in maize to maize dwarf mosaic virus. Mol. Planf-MicrobeInteracr. 2, 309314.

Meagher, R. B., McLean, M. D., and Arnold, J. 1988. Recombination within a subclass of restriction fragment length polymorphisms may help link classical and molecular genetics. Genetics 120, 809-818. Medina-Filho, H..and Tanksley, S. D. 1983. Breeding for nematode resistance. I n “Handbook of Plant Cell Culture, Vol. I. Techniques for Propagation and Breeding” (D. A. Evans, W. R. Sharp, P. V. Ammirato, and Y.Yamada, Eds.), pp. 904-923. MacMillan Co., New York. Melchinger, A. E. 1990. Use of molecular markers in breeding for Oligogenic disease resistance. Planr Breeding 104, 1-19. Melchinger, A. E. 1993. Use of RFLP markers for analysis of genetic relationships among breeding materials and prediction of hybrid performance. In “International Crop Science I” (D. R.

DNA MARKERS AND PLANT BREEDING PROGRAMS

339

Buxton, R. Shibles, R. A. Forsberg. B. L. Blad, K.H. Asay, G. M. Paulson, and R. F. Wilson, Eds.), pp. 621-628. Crop Science Society of America, Madison, WI. Melchinger, A. E.. Lee, M., Lamkey, K. R.,Hallauer, A. R., and Woodman, W. L. 1990. Genetic diversity for restriction fragment length polymorphisms and heterosis for two diallel sets of maize inbreds. Theorer. Appl. Genet. 80, 488-496. Melchinger, A. E., Messmer, M. M., Lee, M., Woodman, W. L., and Lamkey, K. R. 1991. Diversity and relationships among U.S. maize inbreds revealed by restriction fragment length polymorphism. Crop Sci. 31, 669-678. Messeguer, R., Canal, M., de Vicente, M. C., Young, N. D., Bolkan, H., and Tanksley, S. D. 1991. High resolution RFLP map around the root knot nematode resistance gene (Mi) in tomato. Theorer. Appl. Gener. 82, 529-536. Messmer, M. M., Melchinger, A. E., Boppenmaier, J., Brunklaus-Jung, E., and Herrmann, R. G. 1992. Relationships among early European maize inbreds. I. Genetic diversity among flint and dent lines revealed by RFLR. Crop Sci. 32, 1301-1309. Messmer, M. M., Melchinger, A. E., Herrmann, R., and Boppenmaier, J. 1993. Relationships among early European maize inbreds: 11. Comparison of pedigree and RFLP data. Crop Sci. 33, 944-950.

Moffat, A. S . 1992. New plant institute recommended. Science 256, 1755-1756. Moffat, A. S. 1994. Mapping the sequence of disease resistance. Science 265, 1804-1805. Moll, R. H., Lindsey, M. F., and Robinson, H. F. 1964. Estimates of genetic variances and level of dominance in maize. Generics 49, 41 1-423. Moreno-Gonzalez, J. 1993. Efficiency of generations for estimating marker-associated QTL effects by multiple regression. Generics 135, 223-23 I . Moser, H., and Lee, M. 1994. RFLP variation and genealogical distance, multivariate distance, heterosis, and genetic variance in oats. Theorer. Appl. Genet. 87, 947-956. Mukai, R., and Cockerham, C. C. 1977. Spontaneous mutation rates at enzyme loci in Drosophila melanogasrer. Proc. Narl. Acad. Sci. USA 74, 2514-2517. Nelson, 0. E. 1990. Quantitative changes in maize loci induced by transposable elements. In “Plant Population Genetics, Breeding, and Genetic Resources” (A. H. D. Brown, M. T. Clegg, A. Kahler, and B. S. Weir, Eds.), pp. 116-127. Sinauer Associates, Inc. New York. Neuhausen, S. L. 1992. Evaluation of restriction fragment length polymorphism in Cucumis melo. Theorer. Appl. Genet. 83, 379-384. Nienhuis, J., Slocum, M. K., DeVos, D. A., and Muren, R. 1992. Genetic similarity among Erassica oleracea genotypes as measured by restriction fragment length polymorphisms. J. Am. Soc. Hori. Sci. 118, 298-303. Novy, R. G., Kobak, C., Goffreda, J., and Vorsa, N. 1994. RAPDs identify varietal misclassification and regional divergence in cranberry [Vaccinium macrocarpon (Ait.) Pursh]. Theorer. Appl. Genet. 88, 1004-1010. Nowak, R. 1994. Genetic testing set for takeoff. Science 265, 464-467. O’Connor, M., Peifer, M., and Bender, W. 1989. Construction of large DNA segments in Escherichia coli. Science 244, 1307-1312. O’Donoughue, L. S., Souza, E., Tanksley, S. D., and Sorrells, M. E. 1994. Relationships among North American oat cultivars based on restriction fragment length polymorphisms. Crop. Sci. 34, 1251-1258. Ogihara, Y., Hasegawa, K.,and Tsujimoto, H. 1994. High-resolution cytological mapping of the long arm of chromosome 5A in common wheat using a series of deletion lines induced by gametocidal (Gc) genes of Aegilops spelroides. Mol. Gen. Genet. 244, 253-259. Openshaw, S. J., Jarboe, S . G.. and Beavis, W. D. 1994. Marker-assisted selection in backcross breeding. In “Proc. of the Symposium Analysis of Molecular Marker Data,” pp. 41-43. Am. Soc. Hort. Sci. and Crop Sci. Soc. Am., Madison, WI.

340

M. LEE

Paterson, A. H., Tanksley, S. D., and Sorrells. M. E. 1991a. DNA markers in plant improvement. Adv. Agron. 44, 39-90. Paterson, A. H., Damon, S., Hewitt, J. D.. Zamir, D., Rabinowitch, H.D., Lincoln, S. E., Lander, E. S., and Tanksley, S. D. 1991b. Mendelian factors underlying quantitative traits in tomato: Comparison across species, generations, and environments. Generics 127, 181- 197. Payne, P. I. 1987. Genetics of wheat storage proteins and the effect of allelic variation on bread baking quality. Annu. Rev. Plant Physiol. 38, 141-153. Pereira, M. G., and Lee, M. 1995. Identification of genomic regions affecting plant height in sorghum and maize. Theorer. Appl. Genet. 90, 380-388. Pereira, M. G., Lee, M., Bramel-Cox, P., Woodman, W., Doebley, J., and Whitkus, R. 1994. Construction of an RFLP map in sorghum and comparative mapping in maize. Genome 37, 236-243.

Peterson, P. A. 1992. Quantitative inheritance in the era of molecular biology. Muydicu 37, 7-18. Phillips, R. L. 1969. Recombination in Zeu m y s L. 11. Cytogenetic studies of recombination in reciprocal crosses. Generics 61, 117-127. Phillips, R. L. 1983. Genetic engineering of plants: some perspectives on the conference, the present, and the future. In “Genetic Engineering of Plants” (T. Kosuge, C. Meredith, and A. Hollaender, Eds.). Plenum Publishing Corp., New York. Phillips, R. L., and Vasil, 1. K. (eds.). 1994. “DNA-Based Markers in Plants.” Kluwer Academic Publishers, Dordrecht, The Netherlands. Phillips, R. L., Kim, T. S., Kaeppler, S. M., Parentoni, S. N., Shaver, D. L., Stucker, R. E., and Openshaw, S. J. 1992. Genetic dissection of maturity using RFLR. In “Proc. of the FortySeventh Annual Corn & Sorghum Industry Research Conference” (D. Wilkinson, Ed.), pp. 135-150. American Seed Trade Association, Washington, DC. Ptashne, M. 1987. “A Genetic Switch: Gene Control and Phage X.” Cell Press and Blackwell Scientific, Cambridge, MA. Rafalski, J. A., and Tingey, S. V. 1993. Genetic diagnostics in plant breeding: RAPDs, microsatellites and machines. Trends Genet. 9, 275-280. Rajeshwari, R., Sivaramakrishnan, S., Smith, R. L., and Subrahmanyam, N. C. 1994. RFLP analysis of mitochondria1 DNA from cytoplasmic male-sterile lines of pearl millet. Theoret. Appl. Genet. 89, 88, 441-448. Rasmusson, D. C. 1991. Barley breeding at present and in the future. In “Barley Genet. XI, Vol. 11. Proc. 6th Int. Barley Genet. Symp.,” Helsingborg, Sweden. Rayapati, P. J., Gregory, J. W.. Lee, M., and Wise, R. P. 1994a. A linkage map of diploid Avenu based on RFLP loci and a locus conferring resistance to nine isolates of Pucciniu coronuru var. ‘avenue’. Theoret. Appl. Genet. 89, 831-837. Rayapati, P. J., Portyanko, V. A., and Lee. M. 1994b. Placement of loci for avenins and resistance to Puccinia coronuru to a common linkage group in Avenu sfrigosu. Genome 37, 900-903. Rebai, A., Goffinet, B., and Mangin, B. 1994. Approximate thresholds of interval mapping tests for QTL detection. Genetics 138, 235-240. Rick, C. M. 1969. Controlled inrrogression of chromosomes of Solunum pennellii in Lycopersicon esculenrumi segregation and recombination. Generics 62, 753-768. Rick, C. M. 1979. Biosystematic studies in Lycopersicon and closely related species of Solunum. I n “The Biology and Taxonomy of the Solanaceae” (J. G. Hawkes, R. N. Lester, and A. D. Skelding, Eds.), pp. 667-678. Academic Press, New York. Robertson, D. S. 1985. A possible technique for isolating genic DNA for quantitative traits in plants. J. Theorer. Biol. 117, 1-10. Rogowsky, P. M., Guidet, F. L. Y., Langridge, P.,Shepherd, K.W., and Koebner. R. M. D. 1991. Isolation and characterization of wheat-rye recombinants involving chromosome arm IDS of wheat. Theorer. Appl. Genet. 82, 537-544. Ronald, P. C.. Albano, B., Tabien, R., Abenes, L., Wu, K., McCouch, S., and Tanksley, S. D.

DNA MARKERS AND PLANT BREEDING PROGRAMS

341

1992. Genetic and physical analysis of the rice bacterial blight disease resistance locus, XdI. Mol. Gen. Genet. 236, 113-120. Rosenberg, N. 1994. Inventions: Their unfathomable future. The New York Times. August 7, p. 9. Russell, W. A. 1993. Achievements of maize breeders in North America. I n “International Crop Science I” (D. R. Buxton, R. Shibles, R. A. Forsberg. B. L. Blad, K. H. Asay, G. M. Paulson, and R. F. Wilson, Eds.), pp. 225-233. Crop Science Society of America, Madison, WI. Schmidt, R. J., Burr, F. A., and Burr, B. 1987. Transposon tagging and molecular analysis of the maize regulatory locus opaque-2. Science 238, 960-963. Schmitz, S.-F. H.. Schwager, S. J.. and Pollak. E. J. 1993. Determining the minimum sample size required to obtain sufficient progeny with a desired genotype at two quantitative trait loci. Theoret. Appl. Genet. 87, 136- 144.

Schnable, P., and Wise, R. P. 1994. Recovery of heritable, transposon-induced, mutant alleles of the rj2 nuclear restorer of T-cytoplasm maize. Generics 136, 1171-1 185. Schnell, F. W. 1982. A synoptic study of the methods and categories of plant breeding. Z. Pjlunrenzuch. 89, 1- 18. Schon, C. C., Lee, M., Melchinger, A. E., Guthrie, W. D., and Woodman, W. L. 1993. Mapping and characterization of quantitative trait loci affecting resistance against second generation European corn borer in maize with the aid of RFLR. Herediry 70, 648-659. Schon, C. C., Melchinger. A. E., Boppenmaier, J., Brunklaus-Jung, E., Hemnann, R. G., and Seitzer, J. F. 1994. RFLP mapping in maize: Quantitative trait loci affecting testcross performance of elite European flint lines. Crop Sci. 34, 378-389. Schwarzbacher, T., Anamthawat-J6nsson, K., Harrison, G. E., Islam, A. K. M. R., Jia, J. Z., King, I. P., Leitch, A. R., Miller, T. E., Reader, S. M., Rogers, W. J., Shi, M., and Heslop-Harrison, J. S. 1992. Genomic in situ hybridization to identify alien chromosomes and chromosome segments in wheat. Theoret. Appl. Genet. 84, 778-786. Schwarz-Sommer, Z., Gierl, A., Cuypers, H., Peterson, P. A., and Saedler, H. 1985. Plant transposable elements generate the DNA sequence diversity needed in evolution. EMBO J . 4,591-597. Sears, R. G., and Cox, T. S. 1993. Improving milling and baking quality of wheat. I n “International Crop Science I” (D. R. Buxton, R. Shibles, R. A. Forsberg, B. L. Blad, K. H. Asay, G. M. Paulson, and R. F. Wilson, Eds.), pp. 665-669. Crop Science Society of America, Madison, WI. Segal, G., Sarfatti, M., Schaffer, M. A,, On, N., Zamir, D., and Fluhr, R. 1992. Correlation of genetic and physical structure in the region surrounding the I2 Fusurium oxysporum resistance locus in tomato. Mol. Gen. Genet. 231, 179-185. Siedler, H., Messmer. M. M., Schachermayr, G. M., Winzeler, H., Winzeler, M., and Keller, U. 1994. Genetic diversity in European wheat and spelt breeding material based on RFLP data. Theorer. Appl. Genet. 88, 994-1003. Simmonds, N. W. 1991. Bandwagons I have known. Trop. Agric. Newsl. (December), 7-10. Sinclair, T. R. 1993. Crop yield potential and fairy tales. In “International Crop Science I” (D. R. Buxton, R. Shibles, R. A. Forsberg, B. L. Blad, K. H. Asay, G. M. Paulson, and R. F. Wilson, Eds.), pp. 707-71 1. Crop Science Society of America, Madison, WI. Smith, J. S. C., and Smith, 0. S. 1992. Fingerprinting crop varieties. Adv. Agron. 47, 85-140. Smith, 0. S., Smith, J. S. C., Bowen, S. L.,Tenborg, R. A., and Wall, S. J. 1990. Similarities among a group of elite maize inbreds as measured by pedigree, F, grain yield, grain yield, heterosis, and RFLR. Theoret. Appl. Genet. 80, 833-840. Smith, 0. S . , Smith, J. S. C., Bowen, S. L.,and Tenborg, R. A. 1991. Numbers of RFLP probes necessary to show associations between lines. Muize Genet. Coop. Newsl. 65, 66. Soller. M., and Beckmann, J. S. 1983. Genetic polymorphism in varietal identification and genetic improvement. Theoret. Appl. Gener. 67, 25-33. Souza, E., and Sorrells, M. E. 1991. Prediction of progeny variation in oat from parental genetic relationships. Theoret. Appl. Genet. 82, 233-241.

342

M. LEE

Sprague. G. F., and Eberhart, S. A. 1977. Corn breeding. In “Corn and Corn Improvement” (G. F. Sprague, Ed.),Monograph No. 18, pp. 305-362. Am. Soc. Agron. Madison, WI. Springer, P. S.,Edwards, K. J., and Bennetzen, J. L. 1994. DNA class organization on maize Adhl yeast artificial chromosomes. Proc. Narl. Acud. Sci. U.S.A. 91, 863-867. Steinborn, R., Weihe, A., and Boerner, T. 1992. Mitochondria1genome diversity within a cultivarof Daucus curoru (ssp. surivus) revealed by restriction fragment analysis of single plants. Plant Breed. 109,75-77. Stevens, J. E. 1994. Japan picks a winner in the rice genome project. Science 266, 1186-1 187. Stiles, J. I., Lemme, C., Sondur, S., Morshidi, M. B., and Manshardt. R. 1993. Using randomly amplified polymorphic DNA for evaluating genetic relationships among papaya cultivars. Theorer. Appl. Genet. 85, 697-701. Strauss, S. H., Lande, R., and Namkoong, G. 1992. Limitations of molecular-marker-aided selection in forest tree breeding. Can. J. Foresr Res. 22, 1050-1061. Strauss, W. M., Dausman, J., Beard, C., Johnson, C., Lawrence, J. B.. and Jaenisch, R. 1993. Germ line transmission of a yeast artificial chromosome spanning the murine a, (I) collagen locus. Science 259, 1904-1907. Strohman, R. 1994. Epigenesis: The missing beat in biotechnology? BiolTechnology 12, 156-164. Stromberg, L. D., Dudley, J. W., and Rufener, G. K. 1994. Comparing conventional early generation selection with molecular marker assisted selection in maize. Crop Sci. 34, 1221-1225. Stuber, C. W. 1992. Biochemical and molecular markers in plant breeding. In “Plant Breeding Reviews” (J. Janick, Ed.), pp. 37-61. Wiley, New York. Stuber, C. W. 1994a. Heterosis in plant breeding. I n “Plant Breeding Reviews” (J. Janick, Ed.), pp. 227-251. Wiley, New York. Stuber, C. W. 1994b. Enhancement of grain yield in maize hybrids using marker-facilitated introgression of QTLs. I n “Proc. of the Symposium Analysis of Molecular Marker Data,” pp. 44-46. Am. Soc. Hort. Sci. andCropSci. Soc. Am. Stuber,C. W., and Edwards, M. D. 1986. Genotypic selection for improvement of quantitative traits in corn using molecular marker loci. I n “Proc. 41st Annual Corn and Sorghum Research Conf. ,” pp. 70-83. American Seed Trade Assoc., Washington, DC. Stuber, C. W., and Sisco, P. H. 1991. Marker-facilitated transfer of QTL alleles between elite inbred lines and responses in hybrids. Annu. Corn Sorghum Res. Con!. P m . 46, 104-113. Stuber,C. W., Lincoln, S. E., Wolff, S.W.,Helentjaris, T., and Lander, E. S. 1992. Identification of genetic factors contributing to heterosis in a hybrid from two elite maize inbred lines using molecular markers. Generics 132, 823-839. Sudupak, M. A., Bennetzen, 1. L., and Hulbert, S. H. 1993. Unequal exchange and meiotic instability of disease-resistance genes in the RpI region of maize. Generics 133, 119-125. Sughroue, J. R., and Rocheford, T. R. 1994. Restriction fragment length polymorphism differences among Illinois long-term selection oil strains. Theorer. Appl. Gener. 87, 916-924. Tanksley, S. D. 1983. Molecular markers in plant breeding. PIunr Mol. B i d . Rep. 1, 3-8. Tanksley, S. D. 1993. Mapping polygenes. Annu. Rev. Genet. 27, 205-233. Tanksley, S. D., and Hewitt, J. 1988. Use of molecular markers in breeding for soluble solids content in tomato-A re-examination. Theorer. Appl. Gener. 7 5 , 81 1-823. Tanksley, S. D., Bernatzky, R., Lapitan, N. L., and Prince, J. P. 1988. Conservation of gene repertoire but not gene order in pepper and tomato. Pmc. Narl. Acud. Sci. LISA 85,6419-6423. Tanksley, S. D., Ganal, M. W.. Prince, J. P., de Vicente, M. C., Bonierbale, M. W., Broun, P., Fulton, T. M., Giovannoni, J. I., Grandillo, S., Martin, G.B., Messeguer, R.,Miller, J. C., Miller, L., Patterson, A. H., Pineda, O., RGder. M. S., Wing, R. A., Wu, W., and Young, N. D. 1992. High density molecular linkage maps of the tomato and potato genomes. Generics 132, 1141-1 160. Teutonico, R. A., and Osborn,T. C. 1994. Mapping of RFLP and quantitative trait loci in Brussica

DNA MARKERS AND PLANT BREEDING PROGRAMS

343

rupu and comparison to the linkage maps of B. M P U S . B. oleruceu, and Arubidopsis rhaliana. Theorer. Appl. Gener. 89, 885-893. Thompson, J. N., Jr. 1975. Quantitative variation and gene number. Nurure 258, 665-668. Thurieaux, P. 1977. Is recombination confined to structural genes on the eukaryotic genome? Nuture 268,460-462. Tinker, N. A., Fortin. M. G.,and Mather, D. E. 1993. Random amplified polymorphic DNA and pedigree relationships in spring barley. Theoret. Appl. Genet. 85,976-984. Troyer, P. F., Openshaw, S. J., and Knittle, K. H. 1988. Measurement of genetic diversity among popular commercial corn hybrids. Crop Sci. UI, 481-485. van Ooijen, 1. W. 1992. Accuracy of mapping quantitative trait loci in autogamous species. Theorer. Appl. Genet. 84, 803-8 1 I . Veldboom, L. R., Lee, M., and Woodman, W. L. 1994a. Molecular marker-facilitated studies in an elite maize population. 1. Linkage analysis and determination of QTL for morphological traits. Theoret. Appl. Genet. 88, 7- 16. Veldboom, L. R., and Lee, M. 1994. Molecular-marker-facilitated studies of morphological traits in maize. 11. Determination of QTLs for grain yield and yield components. Theor. Appl. Genet. 89, 451-458. Voelker, T. A,, Worrell, A. C., Anderson, L., Bleibaum, J., Fan, C., Hawkins, D. J., Radke, S. E., and Davies, H. M. 1992. Fatty acid biosynthesis redirected to medium chains in transgenic oilseed plants. Science 257, 72-74. Walbot, V. 1992. Strategies for mutagenesis and gene cloning using transposon tagging and T-DNA insertion mutagenesis. Annu. Rev. Plunr Physiol. Plunr Mol. Biol. 43, 49-82. Walbot, V., and Cullis, C. A. 1985. Rapid genomic change in higher plants. Ann. Rev. Plunr Physiol. 36, 367-396. Wang, G.-L., and Paterson, A. H. 1994. Assessment of DNA pooling strategies for mapping of QTLs. Theoret. Appl. Genet. 88, 355-361. Weber, D., and Helentjaris, T. 1989. Mapping RFLP loci in maize using B-A translocations. Generics 121, 583-590. Weir, B. S . , and Cockerham. C. C. 1977. lko-locus theory in quantitative genetics. In “Proc. of the Int. Conf. on Quantitative Genetics,” 16-21 August 1976 (E. Pollak, 0. Kempthorne, and T. B. Bailey, Jr., Eds.), pp. 247-269. Iowa State Univ. Press, Ames, IA. Weller, J. I. 1992. Statistical methodologies for mapping and analysis of quantitative trait loci. In “Plant Genomes: Methods for Genetic and Physical Mapping” (J. S . Beckman and T. S . Osborn,Eds.), pp. 181-207. Kluwer, Dordrecht, The Netherlands. Weller. J. I., and Wyler, A. 1992. Power of different sampling strategies to detect quantitative trait loci variance effects. Theorer. Appl. Gener. 83, 582-588. Werner, J. E., Endo, T. R., and Gill, B. S. 1992. Toward a cytogenetically based physical map of the wheat genome. Proc. Nurl. Acud. Sci. USA 89, 11307-11311. White, S . E., Habera, L. F., and Wessler, S. R. 1994. Retrotransposons in the flanking regions of normal plant genes: A role for copiu-like elements in the evolution of gene StNCture and expression. Pmc. Nutl. Acud. Sci. USA 91, 11792-11796. Whitham, S., Dinesh-Kumar, S. P., Choi, D., Hehl, R., Corr, C., and Baker, B. 1994. The product of the tobacco mosaic virus resistance gene N: Similarity to toll and the interleukin-I receptor. Cell 78, 1-20. Whitkus, R., Doebley, J., and Wendel, J. F. 1994. Nuclear DNA markers in systematics and evolution. In “DNA-Based Markers in Plants” (R. L. Phillips and I. K. Vasil, Eds.), pp. 116141. Kluwer Academic, Dordrecht, The Netherlands. Wilde, J., Waugh, R., and Powell, W. 1992. Genetic fingerprinting of Theobmmu clones using randomly amplified polymorphic DNA markers. Theoret. Appl. Genet. 83, 871-877. Williams, C. E., and St. Clair, D. A. 1993. Phenetic relationships and levels of variability detected

344

M. LEE

by restriction fragment length polymorphisms and random amplified polymorphis DNA analysis of cultivated and wild accessions of Lycopersicon esculenrum. Genome 36, 619-630. Wing, R. A., Zhang, H.-B., and Tanksley, S . D. 1994. Mapbased cloning in crop plants. Tomato as a model system: 1. Genetic and physical mapping ofhintless. Mol. Gen. Genet. 242,681-688. Wise, R. P., and Ellingboe, A. H. 1985. Fine structure and instability of the M I - a locus in barley. Generics 111, 113-130. Young, N. D., and Tanksley, S . D. 1989a. RFLP maps and the concept of graphical genotypes. Theoret. Appl. Gene?. 77, 95-101. Young, N. D., and Tanksley, S. D. 1989b. RFLP analysis of the size of chromosomal segments retained around the Tm-2 locus of tomato during backcross breeding. Theore?.Appl. Genet. 77, 353-359.

Zehr, B. E.,Dudley, J. W., Chojecki, I., Saghai Maroof, M. A., and Mowers, R. P. 1992. Use of RFLP markers to search for alleles in a maize population for improvement of an elite hybrid. Theorer. Appl. Gener. 83, 903-9 I 1. Zeng, 2.-B. 1994. Precision mapping of quantitative trait loci. Generics 136, 1457-1468. Zhang, Q.,Shen, B. Z., Dai, X. K.,Mei, M. H.,Saghai Maroof, M. A., and Li, Z. B. 1994. Using bulked extremes and recessive class to map genes for photoperiod-sensitive genic male sterility in rice. Proc. Natl. Acad. Sci. USA 91, 8675-8679. Zhang, W., and Smith, C. 1993. Simulation of marker-assisted selection utilizing linkage disequilibrium: the effects of several additional factors. Theore?. Appl. Genet. 86, 492-496.

LONG-TERM PERSISTENCE OF ORGANIC CHEMICALS IN SEWAGE SLUDGE-AMENDED AGRICULTURAL LAND:A SOILQUALITY PERSPECTIVE A n g u s J. Beck, Ruth E. AIcock, Susan C. Wilson, Min-Jian Wang, Simon R. Wild, Andrew P. Sewart, and Kevin C. Jones Institute of Environmental and Biological Sciences Lancaster University Lancaster, LA14YQ United Kingdom

I. Introduction 11. Long-Term Experiments and the Compounds Investigated A. Long-Term Experiments

B. Organic Chemicals Investigated 111. Influence of Sewage Sludge Applications on the Concentration and Per-

sistence of Organic Chemicals in Soils A. Long-Term Trends in Cultivated Horizon Concentrations B. Relative Significance of Sewage Sludge versus Other Sources of Organic Chemicals to Soils C. Loss of Organic Chemicals from the Long-Term Sewage SludgeAmended Soils D. Need for Behavior Assessment Models W. Implications of Sewage Sludge Application to Farmland for Soil Quality Criteria A. Overview of Soil Quality Criteria B. Kinetically Constrained Soil Quality Limit (KCSQL) Concept V. Conclusions References

I. INTRODUCTION Sewage sludge is an inevitable byproduct of wastewater treatment. Approximately 1.2 and 7.0 million tons of sewage sludge solids are produced each year in the United Kingdom and the United States, respectively, resulting in annual disposal costs of approximately $375 and $844 million (Bruce and Davis, 1989). In the United States, 13.2% of the sludge produced is applied to agricultural 345 Advamca in A p q , V i e 55 Copyright 0 1995 by Aademic Prrop, Inc. All rights of reproduction in any form reserved.

346

A. J. BECK ET AL.

land, while an additional 10.1% is used in the composting industry (Anonymous, 1989). By contrast, in the United Kingdom and other European countries, including France, Denmark, The Netherlands, and Switzerland, over 30% is applied to agricultural land (Sauerbeck, 1987); in the United Kingdom, over 50% of sludge is now applied to agricultural land. The proportion is likely to increase further as a result of European Union (EU) Directives requiring that sewage dumped in coastal waters must receive at least primary treatment and that dumping at sea should cease by 1998. Land disposal is more popular than other means, including landfills, incineration and dumping at sea, because it is relatively inexpensive. Furthermore, application of sewage sludge to agricultural land may be desirable in that it can improve the physical (Epstein et al., 1976), chemical (Peterson et al., 1988), and biological properties of soils (Katterman and Day, 1989), which may enhance crop growth (e.g., see Frink and Hullar, 1985). In principle, the application of sewage sludge to agricultural land represents the most sensible environmental and economic option for its disposal. However, the practice needs to be subject to careful control because of the potential presence of unwanted constituents in the sludge, such as heavy metals and organic contaminants. Legislative controls operate in many countries to limit the addition of metals to agroecosystems in sludges. However, at present there are very few constraints on the presence/loadings of organic contaminants. This is partly because there is less information on the concentrationsof “priority organic pollutants” in sewage sludges and partly because of uncertainties over their fate and behavior in soils and food chains. The significance of organic chemicals will largely depend on the likelihood of transfer from sludged soil to crop plants, groundwater, grazing animals, and thence to humans. This in turn depends on the interaction between organic compounds and the sludged soil (i.e., their distribution between the solid, aqueous, and gaseous phases) and on the compounds’ concentration and persistence in sludge-treated soil. However, there is still a general lack of information on the long-term persistence of organic contaminants in sludge-treated soils, which hampers attempts at risk assessment. This is the topic addressed in this review. Characteristically, the overall loss of organic chemicals from soils is often biphasic, whereby a short period of rapid dissipation is followed by a longer period of slow chemical release (Pignatello, 1989). Dissipation processes, including leaching and volatilization, exhibit similar behavior. The primary ratelimiting factors governing this behavior are postulated to be fundamental sorption-desorption mechanisms, including intraparticle diffusion, intrasorbent diffusion, and chemisorption. which control the distribution of contaminants between the solid and aqueous or gaseous phases of soils and, hence, the supply of chemicals available to the various dissipation processes (Brusseau et al., 1991; Pignatello, 1993). Thus, these processes will also exert a major influence on the magnitude and residence time of resistant fractions of organic chemicals in sewage sludge-amended soils.

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

347

While legislation limiting the concentrations of metals in sewage sludges applied to agricultural land has now been in force.for some time (CEC,1986), no such controls exist for organic compounds. Maximum loading rates for some organic compounds in sewage sludge applied to agricultural land were proposed by the U.S. EPA some years ago, but these rates were unacceptable and a resubmission of the proposed regulations was required (Goldstein, 1989a-c). However, generic or site-specific soil quality limits for organic chemicals in all soils have been adopted or proposed in various countries in an effort to control or assess soil quality (Sheppard et al., 1992). Like all other soils, land that has received additions of sewage sludge must comply with these criteria. The most comprehensive generic limits, established in The Netherlands, distinguish between background concentrations for naturally occurring substances and analytical detection limits for manmade organic compounds, target values, and threshold concentrations above which remediation may be necessary-intervention values (Van den Berg et al., 1993). These are being employed to highlight contaminated sites that require treatment-remediation, ideally resulting in reductions in the soil concentrations of the compounds of concern (Beck et al., 1995). Over the last 6 years, our research group, in collaboration with Rothamsted Experimental Station and the Macaulay Land Use Research Institute, has taken advantage of a unique collection of archived soils, sewage sludges, and herbage, dating from the 1840s onward, to investigate the long-term persistence of nonionic organic chemicals in soils. Papers have been published reporting on the persistence and movement of polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (FCBs), and chlorobenzenes (CBs) in a range of soils subjected to single and multiple applications of sewage sludge and soils that have never received sewage sludge. In this review, we present a comprehensive overview of these studies and provide a review of the mechanisms that have been postulated to be responsible for the long-term persistence and dissipation of organic chemicals in sewage sludge-amended soils. We then employ this knowledge to predict the long-term background concentrations of organic chemicals that we might expect to persist in soils that both have and have not been amended with sewage sludge. In doing so, we have employed the kinetically constrained soil quality limit (KCSQL) concept developed by Beck et al. (1995). The KCSQL is synonymous with the concentration of a given chemical in a given soil when the residual or slow phase of dissipation is reached. Henceforth, further desorption represents an intrinsic kinetic constraint to the various dissipation processes, so that it is unlikely that any appreciable change in chemical concentration will occur over practical time scales under prevailing environmental conditions for a given soil (see Beck et al., 1995). Furthermore, it has been argued that many commonly used remediation strategies including pump-and-treat (McKay and Cherry, 1989) and soil venting (Travis and McInnis, 1992) may be only partially successful in overcoming these constraints. Finally, we compare KCSQLs derived for selected compounds investigated during our studies with

348

A. J. BECK ET AL.

generic soil quality limits adopted by The Netherlands, United Kingdom, and Canada in an effort to assess whether land that has received application(s) of sewage sludge complies with existing soil quality legislation.

II. LONG-TERM EXPERIMENTS AND THE COMPOUNDS INVESTIGATED A. LONG-TERM EXPERIMENTS The Luddington experimental site near Stratford-upon-Avon, Warwickshire, UK (ordnance survey map reference: SP 158529), and the Lee Valley experimental site at a rural location in Hertfordshire, UK (ordnance survey map reference: TL 370106), were jointly established by the Macaulay Land Use Research Institute (formerly Macaulay Institute for Soil Research) and ADAS (formerly the United Kingdom Agricultural Development Advisory Service) in I968 to investigate metal uptake by crops from sludge-amended soil (Berrow and Bumdge, 1990). The soil at the Luddington site is a sandy loam with 1.8% organic matter and pH 5.8, while the Lee Valley soil contains 4.9% organic matter and has a pH of 6.5. Here we discuss results from the top 0.15 m of six plots at each of these sites. Five plots received different sewage sludges at a rate of 125 tons (dry weight)/ha in 1968, and a control plot received no sewage sludge. One of the five sewage sludges was from a rural area, while the remainder were chosen such that they had “naturally” elevated concentrations of Cr, Cu, Ni, or Zn (in subsequent discussion, each plot is referred to as rural or on the basis of the elevated metal, e.g., Cr-rich). Following sludge application, the soils were dug or rotovated to a depth of 0.15 m and received no further amendments. Soil samples collected from the cultivated horizons at these sites were dried at 100°C prior to being archived. To investigate the effects of multiple applications of sewage sludge, an experimental plot was chosen from the Woburn Market Garden Experiment, situated in a rural location 27 km northwest of Rothamsted Experimental Station (ordnance survey map reference: TL 120137). This site was established by Rothamsted Experimental Station in 1942. The soil is a freely draining sandy loam with a pH of 6.5-7.0 and about 2% organic matter. The experimental plot discussed in this review received 25 annual applications of sewage sludge between 1942 and 1961, which were plowed in to a depth of 0.23 m. The sewage sludge was applied at rates of 20 tons/ha/yr in 1942.75 tons/ha/yr from 1943 to 1956, and 50 tons/ha/yr from 1957 to 1961. Since then, the experimental plot has only received inorganic fertilizers. Samples of each sludge applied to the plot were dried at 70°C and archived so that they were available for analysis. Soil samples

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

349

were bulked over the 0-0.23 m depth of the cultivated horizon, air-dried, and ground C2 mm before being archived. Samples were taken from different depths in this plot during 1982 to assess the downward migration of chemicals. A control plot receiving no sewage sludges was also investigated. More comprehensive details of the Woburn Market Garden Experiment can be found in McGrath (1 987). Although they received no sewage sludge amendments, we refer to data from the Park Grass, Broadbalk, and Hoosfield experiments at Rothamsted Experimental Station in a rural area 42 km northwest of London. These experiments are part of the Rothamsted Classical Experiments that were established by Lawes and Gilbert between 1843 and 1856 to investigate the effects of organic and inorganic fertilizers on crop yields, which have continued to the present day [see Warren and Johnston (1969) and Johnston and Gardner (1969) for further details]. One plot in each experiment, the control, received no such inputs, so that changes observed in this plot are the result of net air-soil exchanges (i.e., deposition-volatilization). They provide a unique opportunity to set the results of the sewage sludge experiments within the context of even longer term soil chemical changes, enabling us to distinguish between the significance of sewage sludges and other sources as contributors to changes in the organic chemical burden of UK soils (see Section 1II.B). The soils at the Rothamsted site are neutral or slightly calcareous silty clay loams containing 0.9-1.1% organic carbon. Samples from the top 0.23 m of untreated control plots from both the Broadbalk (sown annually with winter wheat since 1843) and Hoosfield (sown annually with spring barley since 1852) experiments were obtained for analysis. These samples were air-dried, ground, and sieved to <2 mm before being placed in glass containers to be stored in the archives. Grain samples from these experiments were also archived and available for analysis. The Park Grass experiment comprises permanent pasture and was started in 1856 with the site having been continuous pasture for at least 100 years before the experiment commenced. Soils were sampled in a manner similar to that used at Broadbalk and Hoosfield, with aftermath and hay samples also being collected and archived after the grass was cut in the autumn of each year.

B. ORGANIC CHEMICALS INVESTIGATED The organic chemicals considered here are all nonionic but have a wide range of physicochemical properties, so that some are volatile or leachable while others are strongly sorbed and persist in soils over prolonged periods. The principal chemical groups considered are the CBs, PCBs, and PAHs. The CBs comprise a group of 12 monocyclic aromatic compounds (Table I) in which between one and six of the hydrogen atoms of a benzene ring have been

3 50

A. J. BECK ET AL. Table I

Physieoehemid Properties of CBs, PCBs, end PAHs' ~

~

Compound CBs MCB I ,2-DCB I ,3-DCB 1.4-DCB 1,2,3-TCB I ,2,4-TCB I .3,5-TCB 1,2,3,4-TeCB 1,2,3,5-TeCB 1,2,4,5-TeCB PeCB HcB PCB congeners 6

8 14 18 28 52 44 40 61 66 101 99 1 I9 110 149 1 I8 188 153 105 I38 180 170 20 I 206 PAHse Naph Ace Flu Phen Anth

Solubility (mglliter. 25°C) 484 1%

307 170.19 21 40 5.3 7.8 3.6 I .27 0.65 0.005 0.706 NA 0.135 0.148 0.037 0.04 0.044

0.0193 0.056 0.013 0.0165 0.00485 0.0197 0.0095 NA NA 0.0018 0.0026 0.00495 0.002 0.000509 0.0024 O.oooO26 0.00011 32.2 4.36 I .71 1.1

0.0575

VPb (Pa, 25°C) 1580 118 I20 83 52.83 61 78.05 8.67 19.22 9.6 0.9565 0.2447 0.139 0. I03 0.19 0.026 0.034 0.0123 0.00576 0.00501 0.0042 0.00239 0.00285 0.00421 0.00188 0.001523 NA NA O.ooo9 0.00012 0.000474 0.000169 0.000126 O.oooO31 O.oooO387 O.ooOo38

10.8 1.16 0.45 0.0933 0.0116

HcC (Pa r n ~ r n o ~ ) 3.68 X 2.44 x 3.76 x 1.60 x 2.42 X 2.77 x 1.10 x 1.44 x 5.88 X 1.22 x 8.50 X 6.75 X

lo2 102

lo2 102

loz 103 103 102

lo2 102 101 10'

1.39 X lo-' 41.5 33. I4 48.3 26.97 43.45 36.2 17.75 14.49 3 1.65 21.63 22.34 1.88x 10-3 21.12 NA NA 9.00 x 10-4 17.64 4.74 x 104 24.88 34.16 7.23 3.87 x 10-5 13.83 51.74 17.13 10.82 5.37 21.79

log ~

, d

2.41 3.21 3.37 3.43 3.21 4.03 3.69 4.24 3.45 4.05 4.8 5.25

4.88 4.9 5.15 4.52 4.99 5.09 5.57 5.61 5.15 6.3 5.89 NA 6.45 NA NA NA 6.58 NA 6.24 6.36

5.99 NA 6.66 3.11 4.02 4.38 4.4 4.32

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND Table I Compound

Solubility (mglliter, 25°C)

351

(continued)

VPb (Pa, 25°C)

Hc (Pa m31mol)

log KWd

~

PAHse Fluor Pyrene WaIA

chrys WbIF BIklF B[alP D[ah]A B[ghi]P

0.243 0.132 0.01I 0.019 0.0015

0.00081 0.00535 0.00073 0.000328

0.205 0.00165 0.0001 0.0000015 O.ooOo295 O.ooOo176 0.00000379 0.00000000677 0.000000178

1.78 2.76 1.09 0.278 I .61 3.84 0.239 0.0075 0. I849

5.11 4.94 5.33 5.14

5.72 5.73 6.24 6.094 6.23

Values presented are the mean of a representative subset cited by MacKay et al. (1991,1992). Vp is the vapor pressure in Pa at 25°C. Hc is Henrys constant. K, is the organic carbon partition coefficient. Abbreviations: Naph, naphthalene; Ace, acenaphthene; Flu, tluorene; Phen, phenanthrene; Anth, anthracene; Fluor, fluoranthene; B[a]A, benz[a]anthracene; Chrys, chrysene; B[b]F, benzo(b1fluoranthene;BIRIF, benzo[klRuoranthene; B[a]P, benzo[a]pyrene; B[ghi]P, benzo[ghilperylene.

substituted with chlorine atoms (Fig. 1). Their water solubilities are generally low, while their vapor pressures are low to moderate and decrease with increasing substitution of the hydrogen atoms (Table I). They have high octanol/water partition coefficients so that they are strongly sorbed by soil organic matter. This, coupled with high electronegativity conferred upon them by the chlorine atoms, makes them extremely stable under prevailing environmental conditions. Consequently, they have been detected in soils, lake water and sediments, rivers, sea water, drinking water, animals, plants, and humans [see Wang and Jones (1994a) for comprehensive references]. The dichlorobenzenes (DCBs), 1,2,4-trichlorobenzene (1,2,4-TCB), and hexachlorobenzene (HCB) have been classified as priority pollutants by the U.S. EPA and EU (Jones and Wild, 1991). The PCB family of compounds comprises a biphenyl compound with 10 possible positions (labeled 2-6 and 2’-6’ in Figure 1) where chlorination may occur. Thus, there are 10 groups of isomers, from the 3 monochlorobiphenylsto the single decachlorobiphenyl. Collectively, 209 PCB congeners are possible, and around 100 congeners have been reported in various commercial preparations and in environmental samples. Commercial PCBs have been manufactured in several countries, including the United Kingdom, United States, France, Japan, Germany, and Russia, under various trade names (e.g., Aroclors in the United States, Clophen in Germany) and have been marketed and used worldwide (Safe, 1994). Total UK production between 1954 and 1977 was 66,748 tons (Harrad et

352

A. J. BECK ET AL.

rn Monochlorobenzene

Ci

Hexachlorobenzene

CI

Pi& 2,4,4'-hichlorobiphenyl

2,2',3,4,4',5'-hexachlorobiphenyl

PBflS Phenanthrene

FIpre 1.

Benzo(a)pyrene

Chemical structures of selected CB, K B , and PAH molecules.

al., 1994). On a world scale, out of a total global commercial production of 2 x 106 tons, it is estimated that about 31%, i.e., 0.37 X 106 tons, is still present in the environment and has been widely dispersed throughout the biosphere (Tanabe, 1988). The physicochernical properties, degradability, and toxicity of PCBs are all related to their molecular structure. The lower chlorinated congeners such as 18 and 28 possess comparatively high water solubilities and vapor pressures and are therefore less persistent in soils than higher chlorinated PCBs such as 138 and 180, which are more lipophilic (Table I). PCB congeners containing chlorine substitution on the 2 and 5 positions of the biphenyl ring have been reported to be preferentially degraded in soils (Bedard ef al., 1987). Toxicological studies on individual congeners indicate that PCB toxicity is dependent not only on degree of chlorination but also on molecular configuration. Planar (or coplanar) PCBs,

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

353

which have nonorrho substitution and are heavily substituted at the meta and para positions, appear to have the greatest toxic potential. The three coplanar congeners, PCBs 77, 126 and 169, are the most biochemically active and also the most toxic (Safe, 1994). Atmospheric deposition was the primary source of PCBs to the terrestrial environment and has resulted in trace levels of PCBs being present throughout the biosphere. Global redistribution of PCBs continues through revolatilization and deposition and is augmented by the disposal of electrical equipment still in use and from landfill operations. As a result, PCBs have been identified in almost every compartment of the global ecosystem. The lipophilic nature and persistence of PCBs result in high bioaccumulation potentials and subsequent biomagnification in higher trophic levels of both terrestrial and marine food chains. Consequently, concentrations (ppb) in water and air lead to concentrations orders of magnitude higher in species at the top of marine and terrestrial food chains. The PAHs comprise two or more fused benzene rings (Fig. 1). Several hundred homo- and heterocyclic PAHs can potentially be formed, but usually a subset of 14, designated by the U.S. EPA, are routinely analyzed. These compounds generally become more lipophilic, less soluble, and less volatile with increasing molecular weight (Table I). PAHs are formed by the incomplete combustion of organic materials such as coal, oil, and gas, so that environmental loads have increased enormously over the last century in tandem with increased large-scale consumption of fossil fuels (Jones et al., 1989a).Consequently, PAHs are ubiquitous in the environment as a result of atmospheric redistribution and deposition of these emissions (McVeety and Hites, 1988). They have been widely detected in soils, water, air, sediments, animals, plants, and humans [see Wild and Jones (1995) for comprehensive references]. Many PAHs have been included in U.S. EPA and EU priority pollutant lists because of their known, e.g., benzo[a]pyrene, or suspected carcinogenicity and/or mutagenicity (IARC, 1983).

111. INFLUENCE OF SEWAGE SLUDGE APPLICATIONS ON THE CONCENTRATION AND PERSISTENCE OF ORGANIC CHEMICALS IN SOILS A. LONG-TERM TRENDS IN CULTIVATED HORIZON CONCENTRATIONS 1. CBs The concentration of CBs in sewage sludges applied to the Woburn sludgeamended plot varied by more than an order of magnitude between 1942 and 1961 (Table 11). DCBs were present at the highest concentrations and 1,2,3-TCB at the

3 54

A. J. BECK ET AL. Table I1 CBs, PCB, and PAH Concentrations in the Sewage Sludges Applied at Woburn Concentration Compound CBs ( M k g ) 1.3-DCB 1,4-DCB 1.2-DCB 1,3,5-TCB 1,2,4-TCB 1,2,3-TCB 1.2.4.5-TeCB I ,2,3,4-TeCB PeCB HCB PCBs PCB congeners (pg/kg) 18 28 52 66 101

149 153 138 180 PPCBs PAHs (mglkg) Naph Ace/F Phen Anth Fluor Pyrene Benz/Chry B[blF WIP B[ghi]P PPAHs

Mean 10.7 29.8 17.4 1.22 2.63 0.21 0.33 1.89 0.69 2.51 67.4 84 223 I I4 525 108 73 57 75 75 I573 0.19 0.46 4.65 0.61 8.13 6.81 7.89 8.74 2.8 10.52 50.8

Std dew 21.6 17.6 27.9 I .09 3.32 0.36 0.33 2.56 0.57 1.27 65.4 62 218 83 71 97 62 53 78 70 1201 0.2 0.2 I .4 0.4 5.8 4.2 5.3 3.4 1.3 6 23.6

lowest concentration, while monochlorobenzene (MCB) and 1,2,3,5-TCB were not detected. In 1942, 1947, 1948, 1952, 1954, and 1959 the CB addition to the soil was less than 1 g/ha, while it exceeded 6 g/ha in 1943 and 1953. The mean annual soil load was 2.60 g/ha/year, with a total application of approximately 52

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

-

o ! 1950

-

,

1955

.

,

1960

.

.

1%

.

,

-

1

1970

.

1975

r

-

1980

355

1.2-DCB

I,z,4-KB

,

.

1985

.

-

19W

d

1995

Year

Flpre 2. Loss of selected chlorinated benzenes in the cultivated horizon of the Woburn experimental plots.

g of CBs per hectare between 1942 and 1961. The concentrations of I ,2-DCB, I,3-DCB, and the three TeCBs in the control soil and of 1,2,3,5-TeCB and 1,2,4,5-TeCB in sludged soil were all below or near their detection limits. The concentrations of individual CBs in the sludge-amended soil were less than 1 pg/kg, except for I ,4-DCB. The concentrations of individual CBs in the control plot were stable from 1942 until the present day, with the exception of 1,4-DCB. Most of the CBs added to the soil in sewage sludge had disappeared by the time of sampling in 1951 and 1960 (Fig. 2). The only exception was HCB, which dissipated more slowly. The CB residues remaining in the soil in 1960 were very persistent, with little loss occurring in the following 30 years. During the 50 years of the experiment, 1,2-DCB was the least persistent residue and HCB was most persistent, with 6 and 22%, respectively, of the total amount of these compounds applied from 1942 to 1961 remaining in the soil in 1991 (Fig. 2). Excluding 1,4-DCB, about 10% of the X B s applied originally was still present in the soil 30 years after sludge application stopped. During the 1960s, 1,4-DCB concentrations increased sharply in both the control and sludge-amended plots, reaching maximum concentrations of 10 and 17 pg/kg, respectively, around 1967. This increase was not due to sludge amendment, because that ceased in 1961 and the control plot received no sludge additions at any time. It was possible that 1,4-DCB entered the soil as an impurity in some pesticides because it has been widely used in their manufacture (IPCS, 1991). Another possible source was atmospheric deposition. Oliver and

3 56

A. J. BECK ET AL.

Nichol(l982) reported that peak inputs of CBs to sediment cores taken from the Niagara River and Lake Ontario occurred during the 1960s, and Rappaport and Eisenreich (1988) reported that the atmospheric input of HCB to peat cores in midlatitude United States also peaked around the 1960s. These all matched the peak period of CB production in the United States (Oliver and Nichol, 1982; Rappaport and Eisenreich, 1988). Anaerobic reduction of HCB to TCBs and DCBs in sewage sludges has been reported by Fathespure et al. (1988). Wang et al. (1995) have evaluated these three possible causes and concluded that pesticide impurities and/or atmospheric deposition was most likely responsible for the increase in 1,4-DCB observed at Woburn in the 1960s. 2. PCBs

Chronological trends in ZPCB concentrations in the Zn and Cr plots at Luddington are shown in Fig. 3. No soil samples were collected at the time of sludge amendment, so the amount of PCBs added to the soil in 1968 in sewage sludge has been calculated so that the PCB concentration at this time could be estimated. The estimated increase in the ZPCB concentration due to sludge amendment on each plot was added to corresponding control plot values to give theoretical soil concentrations. The Cr-rich plot exhibited the greatest elevation in soil PCB concentrations following sludge amendment, with ZPCB concentrations of 439 kg of ZPCB/kg in 1972 (Fig. 3). Thereafter, PCB concentrations in all of the plots declined rapidly until the late 1980s, but since then concentrations in all of the plots have remained at or near those of the control plots, i.e., 30 to 60 kg of ZPCB/kg of soil. In both the sludge-amended and control plots, the relative proportions of tri- and tetra-chlorinated congeners generally declined while the proportion of higher molecular weight compounds, i.e., hexa- and heptachlorobiphenyls, generally increased through time (Table 111). By 1972 the sludged plots had lost negligible amounts of the ZPCB applied in 1968, but by 1985-1986 losses were about 60%. The loss of sludge-applied PCB congeners varied depending on the sludge applied (see Table 111). In the Zn-rich plot 2 1% of ZPCB remained in 1990, while between 30 and 42% remained in the Cu-, Cr, and Ni-rich plots, i.e., congeners declined more rapidly in the Zn-rich plot. This probably reflects differences in the retentive characteristics of the different sludge matrices for PCBs. A first-order model successfully predicted the dissipation of sludge-derived PCB congeners and ZPCB in the Zn-, Cu-, and Ni-rich plots at Luddington over the time scale of this experiment. By using these models, estimated half-lives for selected congeners and XPCB have been calculated for selected plots at Luddington (Table IV). It should be emphasised that these half-lives represent net losses of PCBs in the field, subject to marked temporal variations in temperature and precipitation, due to the combination of possible loss mechanisms including

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

O+ 1%8

8

1978 Year

357

1988

mim 0 ,

I

1988

1978 Year

1968

Figure 3. Comparison of theoretical (open circles) and measured (closed circles) ZPCB concentrations in the cultivated horizons of the Zn-rich plot (a) and the Cr-rich plot (b) with ZPCB concentrations in the control plot (dotted line) at the Luddington site.

Table 111 Percentage of Selected PCB Congeners Remaining (Based on Theoretical versus Actual Concentrations) in the Cultivated Horizon of Selected Sludge-Amended Plots at Luddington XPCB (8)

Congener 28 (%)

Congener 180 (8)

Year

Zn

Cu

Cr

Ni

Zn

Cu

Cr

Ni

Zn

Cu

Cr

Ni

1972 1976 1985 1990

100

87 79 42 30

100

100 100

100 96 53

63 83 58 3

100

84 100 89 31

67 70 48 28

100 100

58 26 29 28

74 73 58 35

99 41 21

100 39 30

44 42

20

81 86 14

64 69

A. J. BECK ET AL.

3 58

Table IV Half-Life Estimates for Sludge-AppliedPCB Congeners in the Cultivated Horizon at Luddington

Plot

PCB 18

PCB 28

ZPCB

Cr-rich Zn-rich Cu-rich Ni-rich

3.0 4.9 5.5

2.1 5.3 4.4 1.2

0.2 (5.3)O 5.3 8.5

a

7.0

1.1

Values in brackets are r,,, for the second phase of decline.

volatilization, biological degradation, plant offtake, and transboundary movement. ZPCB loss rates differ between the various metal-treated plots at Luddington. The Cu- and Ni-rich plots have the longest half-lives of 6.5-8.5 years. The shortest half-lives for both ZPCB and congeners 18 and 28 (<2-5.5 years) were found in the Cr-rich plot. PCB half-lives in the Luddington plots are significantly longer than those reported previously by others (Moza et al., 1979; Berthouex and Gan, 1991; Gan and Berthouex, 1993). This is probably due to the unique time scale of this experiment (25 years), which was about 5 times longer than most other experiments (typically ranging from 1 to 5 years), which focus on short- to medium-term loss processes (Diercxsens and Timadellas, 1987; Berthouex and Gan, 1991; Fairbanks et al., 1987). For example, Gan and Berthouex ( 1993) calculated half-lives for di- and trichlorinated homologues and ZPCB of 6, 9, and 19 months, respectively. The heavier homologues (pentachlorinated and upward) did not disappear from the surface soil in their study, supporting our observations of congener 153 at Woburn (see the following). The range, mean, and standard deviation of the PCB concentrations in the sludges applied at Woburn are given in Table 11. Concentrations varied by more than an order of magnitude between 1942 and 1961. Over the 20 years, about 1 kg of ZPCB was applied per hectare in sludge, with congeners 28,52, 101, and 18 comprising the largest proportion. Chronological trends for cultivated horizon ZPCB , congener 28 and 153 concentrations in the control, and sludged plots are shown in Fig. 4, while concentrations of selected other individual congeners are given in Table V. The trend in the ZPCB burden of the sludge-amended plot closely follows that of the control plot. This was because air-soil exchanges of PCBs exerted a strong influence on the PCB concentrations in the cultivated horizon of the sludge-amended plot. The Z K B concentration in the sludge-amended plot increased through time to reach a maximum of 817 p.g of ZPCB/kg in 1972, despite the cessation of sludge

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

'"1

3 59

a

1910

1950

1960

1m

1980

1990

Year

1910

1950

1960

1970

1980

1990

Year

Figure 4. Comparison of theoretical (closed circles) and measured (dotted squares) ZPCB (a), congener 28 (b), and congener 153 (c) concentrations in the cultivated horizon of the sludge-amended plot with their concentrations in the control plot (dotted line) at the Wobum site.

amendment in 1961. PCB concentrations remained relatively constant between 1960 and 1980 and then fell sharply to contemporary concentrations of about 70 pg of ZPCB/kg. By 1992, 31 years after the last sludge was applied, the sludgeamended plot contained over 5 times more ZPCB than the control plot. The lighter congeners (e.g., 28, 52, and 66) generally dominated the congener com-

3 60

A. J. BECK ET AL. Table V Cultivated Horizon PCB Congener Concentrations (pglkg) in the Sludge-Amended Plot at Woburn

PCB congener 8 18

28 52 66 101

149 153 I38 I80

ZPCB

1942

1951

1960

1967

1972

1980

1984

1992

2.2 7.7 9.6 10.1 2.1 1.7

18.7 14.9 62.9 18.0 7.5 2.3 3.7 2.2 2.6 4.0 178

7.8 37.3 240.2 76.0 40.9 16.4 12.0 6.8 11.3 8.8 640

8.0 55.4 209.7 72.4 59.7 14.4 10.4 6.2 12.4 10.3 652

15. I 54.9 270.1 88.6 87.3 20.7 12.9 7.2 14.9 10.9 817

28.4 48.3 131.1 37.3 17.9 11.9 10.8 6.5 15.0 11.2

1.7 9.6 39.3 10.5 7.6

ND ND ND 0.5 4.8 10.4 5.3 3.2 7.6 7.9 70.1

1.0

0.3 0.8 ND 62.6

449

5.8

8.7 5.3 12.8 10.2 183

position of the sludge-amended plot (Table V). Trichlorinated congeners (e.g., 18 and 28) dominated the congener composition in each sample from 1951 to 1984, constituting more than 30% of the CPCB content. However, in the most recent sample (1992) pentachlorinated congeners (101, 110, and 118) were most abundant. There has been a shift in the relative proportion of individual congeners, with a move toward greater proportions of the heavier homologue groups in the two more recent samples (Fig. 4). Interestingly, lower molecular weight PCBs only declined in concentration after 1980 (Table V). Concentrations of congeners 28 and 52, for example, remained relatively constant from 1951 to 1972 and then declined sharply in the most recent sample. By contrast, concentrations of heavier congeners (e.g., 101, 149, 138, and 180) were much lower following the cessation of sludge additions in 1961, but have remained relatively constant since, decreasing only slightly in the most recent sample. Given the sludge PCB concentrations (Table 11), known sludge application rates, calculated soil bulk density, and soil PCB data for the control plot, theoretical soil concentrations were calculated that would have resulted from each successive sludge application. Theoretical elevations in soil concentrations in 1951, 1960, and 1961 (i.e., observed control plot concentrations + estimated elevations) for selected congeners are given in Table V1 and plotted in Fig. 4 for ZPCB and for congeners 28 and 153. As noted earlier, concentrations of the lower chlorinated congeners (e.g., 28 and 66; Fig. 4) in both the control and sludged plots increased dramatically during the 1960s, when they received only inorganic fertilizers and known amounts of specified pesticides. Interestingly, during the 1960s (i.e., after sludge additions ceased), levels of the lower chlorinated congeners in the sludge-amended plot increased by more than those in the

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

361

Table VI Theoretical Elevation of PCB Congener Concentrations (pg/kg) in the Cultivated Horizon of the Sludge-Amended Plot at Woburn PCB congener

I942

1951

I960

1961

8 18

2.2 7.7 9.6 10 2. I I .7 I .o 0.3 0.8 0.0 63

9.4 (>loo) 21.3 (70) 59.2 (>loo) 17. I (>loo) 5.9 (>loo) 4.4 (54) 3.5 (>loo) 2.1 (80) 1.8 (>loo) 3.0 (>loo) 195 (95)

10.0 (63) 43.0 (80) 184.6 (>LOO) 65.3 (>loo) 61.7 (64) 26.8 (53) 16.4 (62) 12.0 (47) 12.8 (69) 12.2 (55) 717 (81)

9.3 (86) 50.3 (>loo) 189 (>loo) 63.1 (>100) 53 (>loo) 30.8 (47) 16.8 (62) 13.1 (47) 16.4 (74) 16.4 (62) 717 (91)

28 52 66 101

149 I53 I38 I80

ZFCB

control plot (Fig. 4). Indeed, the measured concentrations of some of the congeners that dominate the PCB content of air (e.g., 28, 52) actually exceed the theoretical elevations projected for the sludge-amended plot in Table VI. This is believed to be indicative of the greater sorptive capacity of the higher organic matter sludged soil, which limited subsequent losses, primarily by volatilization. PCB retention has been reported to be strongly dependent on the organic matter content of surface soils (Moza et al., 1979). To illustrate the differences in the behavior of different homologues, the trends in measured and theoretical concentrations of congeners 28 and 153 are also plotted in Fig. 4. By the time of sampling in 1960, over half of the theoretical concentration of congener 153 had disappeared. The remaining residue appears to be very persistent and reduced only slightly in the following 30 years. Problems in estimating sludge-derived PCB half-lives at Woburn are compounded by multiple application rates and also by the likelihood that compound loss rates from sludge-amended soils varied over time (Fig. 4). Presumably at Woburn the ZPCB concentration became elevated for a time following each successive application of sewage sludge (between 1942 and 1961) and then declined. Such a pattern would have been repeated throughout the experimental period after each sludge addition, so that by 1960 an accumulated fraction had built up with an increased proportion of the more recalcitrant congeners. For these reasons, it is not possible to calculate a simple, single half-life for PCBs added in sludge to the Woburn soil between 1942 and 1960. Given the uncertainties, it is only possible to estimate a crude half-life of ZPCBs, left in later years, ca. .14-19 years is typical of the period 1967-1992, allowing for dilution and transboundary losses (see Section III.C.6).

3 62

A. J. BECK ET AL.

3. PAHS The PAH content of the sewage sludges applied to the Lee Valley soil plots is given in Table VII. The Cu-rich sludge was the most contaminated, containing 89 mg of XPAH/kg of soil, while the rural sludge, containing 15 mg of ZPAH/kg of soil, was the least contaminated. At Luddington the most contaminated soil applied was the Zn-rich sludge, which contained 57 mg of ZPAH/kg of soil (Table VII). Fluoranthene was the most abundant individual compound in most of the sludges, although in both of the Cr-rich sludges, benzo[ghi]perylene was present at the highest concentrations. PAH concentrations in the control plot at Lee Valley steadily increased with time from 415 to 934 pg of ZPAH/kg of soil as a result of atmospheric deposition. Between 1968 and 1988 there was a net gain of approximately 3.26 mg of ZPAH/mz/year at Lee Valley. No evidence was found to suggest that contamination may have occurred from surrounding plots. Naphthalene, acenaphthene/fluorene, and anthracene concentrations in the control plots at both sites have remained relatively constant through time (Fig. 5). Sewage sludge amendment resulted in increases in soil PAH concentrations in all plots. At Lee Valley the Nirich plot had the highest ZPAH concentrations in 1972, about 3000 pg/kg of soil, with ZPAH concentrations decreasing in the sequence Ni > Zn > Cu > Cr > rural (Fig. 5). This sequence still prevailed in 1988 and broadly reflects the amount of PAHs applied to each of the plots, with the exception of the Cu-rich plot, which received the highest ZPAH load. There was little change in the ZPAH burden of these soils after 1972, although concentrations of some individual compounds have continued to decline. At Luddington the ranking of the ZPAH content of the soil plots in 1972 followed the sequence Cu > Zn > Ni > rural > Cr, with the Cu-rich plot containing 2300 pg of ZPAH/kg of soil. By 1989, this sequence had changed to Zn > Ni > Cu > rural > Cr (Fig. 5). Soil PAH concentrations were slightly lower at Luddington than at Lee Valley because the sludges were less contaminated with PAHs and the soil had a higher bulk density. The higher molecular weight PAHs were more persistent at both sites. At Luddington, by 1972 the sludge-amended plots had lost between 40% and 60% of the PAHs applied in 1968. By 1989, these losses had increased to about 90%. The plot treated with the Ni-rich sludge exhibited different behavior, with only 10% of the applied PAHs being lost by 1972 and 35% still remaining in 1989. Benzo[ghi]peryleneand coronene were the most persistent, with 50- 100% of these compounds remaining in most plots by 1972. By contrast, naphthalene and acenaphthene/fluorene were virtually depleted by 1972. Half-lives at Luddington ranged from under 2 years for naphthalene to over 7 years for fluoranthene and over 9 years for benzo[ghi]perylene and coronene. The mean half-life for ZPAH in all plots was 8 4 years. The longest half-lives were observed in the Ni-rich plot, e.g., about 13 years for ZPAH, while the

*

Table W

PAH Congener Concentrations (mglkg) in the Metal-Rich Sludges Applied at Luddington and Lee Vdey Luddington PAH' congener

Lee Valley

Rural

Zn-rich

Ni-rich

Cr-rich

Cu-rich

Rural

Zn-rich

Ni-rich

Cr-rich

Cu-rich

0.01 0. I 1.1 0.1 3.0 2.8 1.8 1.4 0.5 0.7 0.9 0.4 12.8

0.1 0.3 2.6 0.5 8.1 5.7 9.5 9. I 2.1 6.2 11.5 1.2 56.9

0.04 0.1 2.6 0.3 4.4 3.7 2.9 2.0 0.7 1.2 1.5 0.5 19.9

0.03 0.1 0.9 0.2 2.8 2.3 2.9 2.9 0.9 1.9 4.4 1.3 20.6

0.2 0.3 1.4 0.3 9.8 8.7 9.9 7.7 3.0 3.6 6.0 0.5 51.4

0.1 0.8 1.2 0. I 3.2 1.8 2.0 1.8 0.7 0.8 1.9 0.4 15.1

0.3 1.8 4.1 0.6 11.1 6.8 7.4 5.7 2. I 3.7 5.4 1.4 50.8

0.4 2.0 12.5 1.5 13.6 10.1 7.6 4.8 1.9 3.2 3.4 0.8 61.9

0.1 0.6 1.9 0.3 5.5 5.1 4.4 6.4 1.7 5.7 8.6 2.8 43.1

0.6 2.4 4.3 0.5 19.1 15.1 13.1 11.0 4.5 5.7 9.4 3.0 89.0

See Table I for abbreviations.

3 64

A. J. BECK ET AL.

-

I-

--O-

Cu-rlch NI-rieh

$ m

I8

ma

0 1968

1978

1963

1988

YUr

(a) Luddington

-

Pm 3

J m

cu-rkh Cr-rich

B 1

Irn 2

am

0 1968

1978

1983

1988

Year

(b)kValley FLpre 5. Loss of PAHs from the cultivated horizons of the sludge-amendedand control plots at Luddington (a) and Lee Valley (b).

shortest half-lives generally occurred in the Cu-rich and Cr-rich plots (typically 3-6 years). At Lee Valley there was a greater loss of PAHs between 1968 and 1972 than at Luddington. By 1972, the plots amended with the rural, Zn-rich, and Cr-rich sludges had all lost 60-70% of the PAHs originally applied, and the

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

365

Cu-rich plot had lost 80% of ZPAHs. By 1988, PAH concentrations in the rural sludge plot at Lee Valley had returned to control plot levels, while the others contained 10-20% of the PAH burden added in 1968. As at Luddington, PAHs have been more persistent in the Ni-rich plot. The structure-related persistence of PAH compounds at Luddington also appeared in the Lee Valley soils. Large amounts of many PAHs were lost at Lee Valley by 1972 (no samples were taken prior to that date), with first-order kinetics models failing to predict their dissipation so that half-lives cannot be calculated accurately. Average halflives for benzo[ghi]perylene and coronene at Lee Valley were estimated to be about 10 years. The loss of PAHs was significantly quicker at Lee Valley than at Luddington. This is probably due to differences in soil type and/or differences in the active microbial communities at the two sites. The PAH half-lives from Luddington were significantly longer than those reported in the literature. At both Lee Valley and Luddington, PAH losses from the Ni-rich plots were lower than from the other plots. The reasons for this are not clear, but metal suppression of microbial processes may have been responsible. The ZPAH concentrations of the 25 sewage sludges applied at Woburn ranged from 18 to 121 mg/kg, with a mean value of approximately 50 mg/kg (Table 11). Generally, benzo[ghi]perylene was the most abundant individual PAH compound, and it and phenanthrene, fluoranthene, pyrene, benzanthracene/chrysne, benzo[b]fluoranthene, and benzo[a]pyrene typically ranged from 1 to 30 mg/kg. Between 1942 and 1961, approximately 1420 tons (fresh weight) of sewage sludge/ha (equivalent to 766 tons dry weight/ha) was applied to the experimental plot, so that about 44 kg of ZPAHs was added to the soil over this period. In 1942 the Woburn soil contained approximately 200 pg ZPAHs/kg soil, which was reported by Jones et al. (1989b) to be typical of rural UK soil. There was some evidence for an increase in the ZPAH content of the control plot with time (Fig. 6), which was most likely caused by atmospheric deposition (Jones et al., 1989a). In the sludge-amended plot, ZPAH content increased to 5500 pg/kg by 1960, and by 1984, 23 years after the last sludge amendment, the sludgeamended plot still contained over 3 times as much ZPAH as the control plot (Fig. 6). Between 1942 and 1961 the sludge-amended plot plow layer concentrations of benzo[ghi]perylene, benzo[b]fluoranthene, benzo[a]pyrene, and benzanthracene/chrysene all increased by approximately 1 order of magnitude, while acenaphthene/fluorene, phenanthrene, and anthracene concentrations increased only slightly (Fig. 6). Following the cessation of sewage sludge amendments, phenanthrene, acenaphthene/fluorene, and anthracene concentrations in the sludge-amended soil all declined rapidly (Fig. 6). Interestingly, by 1984 the concentrations of acenaphthene/fluorene and phenanthrene were lower in the sludge-amended plot than in the control plot. By contrast, 7 times more benzo[a]pyrene and 5 times

A. J. BECK ET AL.

3 66

1942

1951

I W

1%7

1972

1980

1981

Year

Flpre 6. Loss of PAHs from the cultivated horizons of the sludge-amended and control plots at Woburn.

more benzo[ghi]perylene were detected in the sludge-amended plot than in the control plot. By 1960, 70% of the XPAHs applied to the sludged plot were lost, and by 1984 only 16% remained. Losses of fluoranthene, pyrene, benzanthracene/chrysene, benzo[b]fluoranthene, and benzo[ghi]perylene ranged from 78% to 87% by 1984, while benzo[a]pyrene was the most persistent, with 36% of that applied still present in 1984. The complexities of multiple sludge additions prevent the derivation of accurate half-life values from the time of the first sludge application. However, halflives were estimated for the loss of PAHs after 1960. The first-order half-life for ZPAH from 1960 to 1988 was approximately 19 and 15 years for phenanthrene, 8 for anthracene, 17 for fluoranthene, 18 for pyrene, 28 for benzanthracenelchrysene, 27 for benzo[b]fluoranthene, 26 for benzo[a]pyrene, and 25 for benzo[ghi]perylene. Poor temporal trends in the data for naphthalene and acenaphthene/fluorene prevented half-life estimation. If sludge-bound PAHs at Woburn were subject to transboundary movement because of plowing, similar to that for heavy metals (see Section III.C.6), then the estimated half-life values due to biodegradation and abiotic losses increase markedly. The half-life for ZPAH increases to approximately 28 and to 65 years for benzo[a]pyrene. These halflives are significantly greater than previously reported PAH half-lives derived from laboratory studies, which are generally less than 2 years (e.g., see Bossert and Bartha, 1986; Heitkamp and Cerniglia, 1987).

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

367

B. RELATIVESIGNIFICANCE OF SEWAGE SLUDGE VERSUS OTHER SOURCES OF ORGANIC CHEMICALS TO SOILS

Sewage sludge is only applied to a small proportion of agricultural soil internationally. In the United Kingdom, for example, a little over l% of agricultural land receives sludges. By contrast, all soils receive atmospheric deposition inputs, and a substantial proportion will receive additions of organic chemicals in pesticide formulations and agricultural wastes, such as animal slurries. For the compounds we discuss in this review, such as the VOCs, CBs, PAHs, and PCBs, atmospheric deposition inputs to UK agricultural land will be greater on a national basis than inputs in sludges. Alcock et al. (1995)estimated the national input of PCBs to UK soils from atmospheric deposition to be -465 kg/year, with only -200 kg/year being applied in sludge. By contrast, nonvolatile compounds such as detergents and their breakdown products (e.g., nonylphenol) enter wastewaters following domestic and industrial usage and can be applied to agricultural soils in sludges, but are unlikely to be applied via other routes. Jacobs et al. (1987)sought to put organic chemical loadings to agricultural soils in context by comparisons with the inputs in pesticides. They noted that many pesticides used today are organic chemicals that are added to soil-plant systems at rates of 0.2-4.0kg of active ingredient per hectare. By assuming an agronomic rate of sludge application of 10 tons/ha (dry weight), the organic chemical loadings expected for concentrations in sludges of 1, 10, and 100 mg/kg are 0.01,0.1,and 1 .O kg/ha. At rates used to reclaim drastically disturbed land, 100 tonslha, the compound loadings for sludges containing 1, 10, and 100 mg/kg would be 0.1,1 .O, and 10 kg/ha, respectively. For agronomic rates, “biologically active” compound concentrations in sludges approaching 100 mg/kg therefore could perhaps be viewed as potentially having an impact on the soil-plant system, depending on the chemical/toxicological properties of that organic compound. At high sludge rates (e.g., 100 tonsiha), concentrations approaching 10 mg/kg in sludge could be expected to add amounts comparable to quantities of pesticides added in agricultural operations. On the basis of these simple observations and data on the prevalence of organics in sludges and potential loadings to soils, agronomic or environmental risks due to the application of domestic sewage sludge to agricultural land would appear to be minimal. In addition, of course, many organics will be bound by soil/sludge organic matter and biologically degraded by soil microorganisms. However, two cautionary notes should be added here. First, pesticides are added deliberately to soils to perform a specific (ostensibly beneficial) role, to improve crop yield and/or quality, whereas sludge application to agricultural soils serves two primary purposes-that of improved soil quality and nutrient recycling but also that of a cost-effective disposal option for the water industry. Second, most

368

A. J. BECK ET AL.

of the modem generation of pesticides and herbicides considered by Jacobs et al. (1987) do not bioaccumulate; it is probably therefore legitimate to consider their primary potential adverse effects to be solely on soil microbial processes or invertebrates (and, hence, potentially on soil fertility) or on plants. However, many of the persistent organochlorines present in sludges, such as PCBs and polychlorinated dibenzo-p-dioxins and -furans (KDDs/-Fs), do bioaccumulate in the food chain. Consequently, when sludge is applied to grassland pasture, they may transfer very effectively to livestock and the human food chain, particularly if the sludge is just surface applied and not subsurface injected. The risk they pose is therefore "removed" from the sludge-treated soil itself.

C. Loss OF ORGANIC CHE~~ICALS FROM THE LONG-TERM SEWAGE SLUDGE-AMENDED SOILS Studies to separate specific chemical loss processes, such as biodegradation, leaching, and volatilization, on the experimental plots themselves is not possible. Any investigation of such dissipation processes must be by indirect inferences from the change in soil and herbage chemical concentrations and/or by laboratory experiments with contemporary soils collected from areas adjacent to experimental plots or with the archived soils themselves. We have used all of these approaches in undertaking the experiments presented in this and the preceding section. However, this work is still in its early stages, so only a brief overview of results is possible. Where possible, we direct readers to other sources of information on dissipation processes in sludge-amended soils. 1. Degradation

Wang and Jones (1994b) investigated the behavior and fate of chlorobenzenes in spiked and sewage sludge-amended soils in laboratory microcosms. While this study suggests that biotic and abiotic losses of CBs were possible, these processes were insignificant in comparison to volatilization. The predominant biodegradation mechanisms are oxidative, leading to the formation of hydroxylated aromatics (Wang and Jones, 1994a). For example, MCB, DCB, TCB, and TeCBs have been reported to be transformed into chlorophenols and catechols in the presence of Pseudornonus sp. (Spain, 1990). Anaerobic degradation to chlorophenols and chlorocatechols in sewage sludge has also been reported (e.g., see Marinucci and Bartha, 1979), while Fathespure er al. (1988) have reported that HCB can be degraded step by step to lower chlorinated CBs under anaerobic conditions. No information is available on the abiotic degradation of CBs in sewage sludge-amended soils, mainly due to the difficulty of isolating such

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

369

losses from those by other processes (Wang and Jones, 1994a,b). However, Kaufman (1983) has suggested that CBs may be subjected to some photochemical degradation at the time of sludge application to soil. Biodegradation was likely to be an important loss pathway for the two- and three-ring PAHs. However, PAHs with four or more benzene rings are particularly resistant to microbial degradation, and to date no microorganisms have been found that can use these compounds as sole sources of carbon and energy (Heitkamp et af., 1988). Microbial degradation of 1%-labeled PAHs was reported to be more rapid in sediments collected from contaminated environments than those from pristine conditions due to the development of adapted microflora (Herbes and Schwall, 1978; Heitkamp and Cerniglia, 1987). Such a mechanism may have been responsible for the lower concentrations of acenaphene/fluorene and phenanthrene in the sludge-amended plot than in the control plot at Wobum in 1984. Like biodegradation, abiotic losses are only likely to be important for two- and three-ringed PAHs in soils. Park et af. (1990) reported that only 2- 18% of the loss of naphthalene, anthracene, and phenanthrene could be attributed to abiotic processes, while the loss of compounds with over three rings was insignificant, Alcock et af. (R. E. Alcock, A. J. Beck, J. Bacon, C. A. Parker, and K. C. Jones, in preparation) conducted a laboratory microcosm experiment to assess the significance of aerobic and anaerobic degradation of PCBs in soils from the Luddington and Lee Valley sites. They found that under sterile, aerobic conditions no loss occurred for the majority of congeners, but under natural, aerobic conditions 20-30% of the di- and trichlorinated congeners was lost. The importance of volatilization must also be implicated in relation to these results (see Sections 1II.B and III.C.4). Alcock et al. (R. E. Alcock, A. J. Beck, J. Bacon, C. A. Parker, and K. C. Jones, in preparation) report that PCB losses under anaerobic conditions occurred for both natural and sterile soils, but attributed this to possible flaws in the design of the anaerobic experiments. They were unable to conclude whether anaerobic degradation occurred during the 3-month experiment. 2. Leaching Profile distributions of CBs at the Woburn site have not yet been investigated. Chang and Page (1985) reported that compounds with a K , of 200-300 liters/kg or higher were expected to be effectively immobilized in soil under most imgation conditions (91-122 cm/year). All of the CBs have K , values higher than 300 liters/kg (Wang and Jones, 1994a,b), and the average amount of water that percolated through the soil at Woburn (precipitation-evapotranspiration) was 30 cm/year (McGrath, 1987), so that leaching of CBs in the aqueous phase appears unlikely. However, the detection of CBs in groundwater (Howard, 1989) would

3 70

A. J. BECK ET AL.

0

lm,

m

-

~~

I

3Ooo

ZPAH concentration Cpg/kg)

Flgure 7. Profile distribution of PAHs in the control and sludge-amended plots at Wobum.

suggest that leaching is possible. It is possible that this may have been due to transport facilitated by mobile or dissolved colloids. In 1982, samples were taken from the Woburn control and sludge-amended plots to a depth of 0.76 m to assess the downward migration of organic chemicals and metals. The results of Wild (1991), see Fig. 7, show greater ZPAH concentrations in the sludge-amended plots than in the control plot between depths of 0.23 and 0.30 m. suggesting that PAHs moved from the cultivated horizon to the subsoil in the sludge-amended plot. However, below 0.3 m the ZPAH concentrations were similar in all plots, indicating that the penetration of sludge-applied PAHs was restricted to shallow depths. Migration with colloidal material (Vinten ef al., 1983), dissolved organic matter (Beck and Jones, 1993), and rapid movement down macropores (Jones et d.,1989c) were suggested as mechanisms for small losses of ZPAH from unsludged soils, so that it is possible that such mechanisms were responsible for the profile migration reported by Wild (1991). Despite this, Wild (1991) considers the evidence for downward migration of PAHs at Woburn to be tenuous, suggesting that the elevated PAH concentrations between 0.23 and 0.3 m were most likely caused by an increase in the mixing depth to 0.27 m due to changes in plowing techniques during the latter years of the experiment.

3. Plant Uptake The proportion of CBs lost due to plant uptake is small compared to the total amounts lost. Wang and Jones (1994~)reported that only 0.03-0.72% of the CBs added to soils in the form of sewage sludge was taken up by carrots. Both carrot

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

371

foliage and roots took up CBs from sewage sludge-amended soil, PAH-spiked soil, and a control soil receiving no addition of sewage sludge. There was no evidence of significant transport of CBs between carrot roots and stems. Uptake of CBs by carrot foliage was mainly of volatilized CBs from the air and was dependent on the foliage yields. Uptake by roots increased with increasing CB concentration. TeCBs were most effectively taken up by carrot foliage and root peel, while TCBs were taken up most effectively by carrot cores. These results support an earlier study by Smelt and Leistra (1974), who found that PeCB and HCB were taken up by carrots and grass. Ryan et al. (1988) and Jones et al. (1989d) have asserted that most PAHs do not possess the appropriate physicochemical properties to facilitate substantial foliar or root uptake. Kampe and Leschbar (1989) could find no evidence for elevated PAH concentrations in the stems and leaves of vegetation grown in sewage sludge-amended soils, while root crops such as sugarbeets, potatoes, and carrots have been shown to contain slightly increased PAH concentrations. Similarly, Wild and Jones (1992a,b) reported that carrot foliage PAH concentrations were unaffected by sewage sludge applications (PAH loadings), but root peel concentrations increased to a plateau concentration with increasing soil PAH levels. Low-molecular-weight compounds dominated individual components of the ZPAH load in the carrot root tissues. About 70% of the PAH burden in carrots was found in the peel, but carrot core ZPAH concentrations were unaffected by sludge applications. This suggests that no transfer of PAHs from the peel to the core occurred. Fresh weight carrot core concentrations were all <42.2 pg/kg. These laboratory investigations are generally in agreement with our longterm field experiments. Wild (1991) analyzed a wide range of archived root and foliage crops from the Woburn, Luddington, and Lee Valley experiments for PAHs. He could find little evidence for any difference in the PAH concentrations of plants grown on the control plots and sludge-amended plots. None of the foliage samples from sludge-amended plots had significantly different PAH concentrations from the foliage samples from the control plots; thus, Wild (1991) concluded that foliage PAHs were not strongly related to soil PAH concentrations and were most likely derived from the atmosphere. The only root crops that had elevated PAH concentrations in the sludge-amended plots relative to the analogous control plots were red beets grown at Woburn in 1963 and 1983. 4. Volatilization

It is likely that the main fate of CBs added to the Woburn plot soil was volatilization. This was evident from differences in the persistence of CBs in the soil. Compounds such as 1 ,2-DCB, which has a relatively high vapor pressure and Henry’s constant, were lost more rapidly than those CBs with lower vapor pressures and Henry’s constants, such as HCB. Controlled laboratory experiments also implicate volatilization as the predominant loss mechanism for CBs

372

A. J. BECK ET AL.

(Wang and Jones, 1994b). For example, CB persistence in microcosm chambers allowing and preventing volatilization were compared; little or no loss was observed in the latter over 250 days. These results are supported by Goodin and Webber (1991), who found that volatilization was the main loss process for 1,2,4-TCB, PeCB, and HCB incubated in laboratory aerated chambers. Naphthalene and phenanthrene may be expected to be lost rapidly from the soil to the atmosphere by volatilization because of their relatively high vapor pressures and Henry’s constants. Volatilization, along with biodegradation, was likely to be the main loss process for the low-molecular-weight PAHs at Woburn, Luddington, and Lee Valley. This assertion is supported by Park ef af. (1990), who reported that loss of the lower molecular weight PAHs is mainly by biodegradation and volatilization, whereas the persistence of the higher molecular weight PAHs is probably largely controlled by biodegradation. However, it should be stressed that sludge has been reported to contain high concentrations of highly volatile compounds such as benzene and toluene, even after digestion and lagoon drying (Crathorne ef af., 1988). The sludges applied at Woburn had been digested and dried in lagoons for several years prior to application. Thus, much of the relatively labile volatile fraction of PAHs was likely to have been lost from the sludge at this time. The difference between the persistence of the different PCBs in the soils from the experimental sites considered here (see Section III.A.2) suggests that volatilization is the dominant loss process for PCBs in sludge-amended soil. At all of the sites, the dissipation of those congeners with high vapor pressures and Henry’s constants was faster than that of those that were less volatile. Other studies (e.g., see Moza et al., 1979; Gan and Berthouex, 1991) support our findings that the tetra homologues and below are likely to be lost by volatilization, while the penta homologues and above are much more stable.

5. Ingestion by Livestock None of the experimental plots discussed in this review have been grazed by livestock. However, our screening models (Wild and Jones, 1992b; S. R. Wild, A. J. Beck, and K. C. Jones, in press) suggest that many compounds discussed here have a high potential to be transferred to grazing livestock by direct ingestion of soil-sorbed residues (see also Section 1II.D). Readers may find the following detailed reports on livestock transfers useful (Fries, 1991; McLachlan, 1993; Wild et af., 1994).

6. Transboundary Losses Due to Plowing Soil movement due to plowing has been highlighted as a mechanism for the loss of metals from the sludge-amended Woburn plots (McGrath, 1987). If

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

373

particle-bound PAHs behave in a manner similar to particle-bound metals, then transboundary losses could account for a substantial proportion of the 84% ZPAHs lost from the sludge-amended plot by 1984 at Woburn. This was unlikely to be a major loss mechanism for the CBs because of their rapid loss from soils following sludge application, but it may have a substantial impact on the less volatile PCBs, i.e., the pentahomologues and above.

D. NEEDFOR BEHAVIORASSESSMENT MODELS From the foregoing discussion, it should be apparent that the behavior of organic chemicals in sewage sludge-amended soils is complicated by the number of loss processes involved and the complexity of their dependence on sitespecific environmental and anthropogenic factors. Furthermore, the compounds discussed here represent only a small proportion of the many compounds that have been reported to be present in sewage sludges. Given the vastly different modes of behavior of these compounds, it is unlikely that experimental means alone will achieve the level of understanding required to institute appropriate environmental legislation and management practices that will be protective of wildlife and human health. One possible alternative is to use simple behavior assessment or screening models based on organic compound physicochemical properties to predict their likely environmental fate and behavior. Among the most useful physicochemical properties are the water solubility and vapor pressure of the compound. Wild and Jones (1992b) employed compound octanokwater partition coefficients, Henry’s constants, and water solubilities to screen a large number of organic chemicals on EU and U.S. EPA priority pollutant lists for their potential to transfer from sludge-amended soils into crop plants and grazing livestock. The approach also illustrated which transfer processes were the most important with respect to human exposure. Later, Wild et al. (S. R. Wild, A. J. Beck, and K. C. Jones, in press) refined this screening approach and applied it to the organic chemical groups discussed here. Their primary objective was to identify selected compounds that may be typically representative of each group or that may behave in exceptional ways and also their main loss pathways from sludge-amended soils so that they can be targeted as a high priority for further experimental investigation. The behavior assessment procedure of Wild et al. (S. R. Wild, A. J. Beck, and K. C. Jones, in press) provides a simple preliminary estimate of a compound’s susceptibility to soil sorption, volatilization, degradation, leaching, plant root bioaccumulation, foliar uptake, bioconcentration in animal fatlmilk via soil/sludge, and herbage ingestion. A summary of the potential environmental fate and behavior of the PCBs, CBs, and PAHs is given in Table VIII. With regard to human exposure to sludge-applied PCBs, livestock ingestion of sludge-

H

d d -I d d d d 1

1 1 1 7

1 1 1 1

H d

H H H H

H

H H H H

H H H H H H H H

H

H H H d d 1 d d d d 1

H d

H

H H H

1N

H

H H

1N 1N

1

H

-IN

1 1 7

1 1 1 1 1 1 -I

1 1 1 1 1 1 -I

-IN -IN -IN -IN -IN

-IN

-IN -IN

-IN -IN -IN

-IN -IN -IN

d d d d d d d d d d d d d d

d d 1 d d

d d 1

H d

H H H H

yv yv yv yv

H H H H H H H H H H H H H H

ax+ I a x - cI ax-z‘I a3 sa3 9OZ 08 I OLI

69 1 ES I 8E I I01 LL

99 19

PP

zs

w M P-

SZ 81

sax

w U u l

1.2.3-TCB 1.2.4-TCB 1,3,5-TCB 1,2,3,4-TeCB 1,2,3,5-TeCB 1,2.4,5-TeCB PeCB HCB PAHs Naphthalene Acenaphthene Flourene Phenaothrene Antluacene Pyrene Flouranthene Chrysene Benz[a]anthracene Benzo[alpyrene Benzo[blflouranthene Benzo[k]flouranthene Benzo[ghilperylene Dibenz[ahlantluacene

* Key:

M M H H H H H H H H H

H H H

H H H H H L L L L L L L L L

MP MP MP MP MP MP P P

NL

MP MP MP P P P P P P P P P P P

NL NL NL

NL NL NL NL NL NL NL

NL NL NL NL

NL NL NL NL NL NL NL

L H H H H

H H L L L L L L L

H H H H H L L L L L L L L L

L

L

M M

M M

H H H H H H H H H H H

H H L L L L L L L L L

bioconcentration A, via soillsludge ingestion by livestock; bioconcentration B, via herbage ingestion by livestock; H, high; M, moderate; L, low;

P, possible; NL, no appreciable leaching. Adapted from R. Wild, A. J. Beck, and R. C. Jones (in press).

3 76

A. J. BECK ET AL.

amended soil followed by human consumption of meat and/or dairy products is the main exposure pathway. PCBs also exhibit semivolatile behavior, so that there is a potential risk to human exposure from the atmosphere and the consumption of crop plants. However, consumption of meat and dairy products is likely to be the most important exposure pathway because the concentrations of PCBs in these products will be significantly higher than those in herbage or the air we breathe (Duarte-Davidson and Jones, 1994). The environmental behavior of the CBs was the most complicated of the compounds discussed here. All of the CBs were identified as susceptible to volatilization, while their potential to accumulate in crop plants and livestock varied. Research on PAHs continues to be afforded high priority on the basis of their toxicity and widespread distribution in the environment.

IV. IMPLICATIONS OF SEWAGE SLUDGE APPLICATION TO FARMLAND M)R SOIL QUALITY CRITERIA

Soil quality criteria have been established in some countries to assess the effectiveness of remediation strategies employed in the cleanup of contaminated soil and as a means of limiting future contamination. These quality criteria must be both practicable and scientifically defensible in order that they may be widely recognized as cost effective and protective of environmental quality and human health. Soil quality criteria may be either generic or site specific. Generic guidelines are uniform permissible contaminant concentrations that apply to all sites, while site-specific criteria relate to complex assessments of restoration goals for individual contaminated sites (Sheppard er af., 1992). Generic criteria have been established countrywide in The Netherlands, United Kingdom, and Canada. In the United States, the U.S. EPA has employed site-specific restoration goals at many Superfund sites. However, arguments have recently been presented in favor of developing a hybrid approach to setting soil quality criteria that would incorporate an element of uniform federal cleanup goals in the United States (e.g., see Dzombak er af., 1993). Soil contamination has recently been afforded a higher priority in the EU. However, much of the contemporary legislation pertaining to soil protection is to be found in EU Directives governing waste and water protection. The first directive to specifically address the issue of soil protection was the directive on sewage sludge in agriculture (CEC, 1986), which set standards for contamination with heavy metals in agricultural soils amended with sewage sludge. Since then, the Fourth Environmental Action Programme 1987-1992 (OJ C328 7.12.1987) and Fifth Environmental Action Programme

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

377

1993-2000 (CEC, 1992) have been launched, which encapsulate an environmental protection strategy as a multimedia and multi-sectoral pollution control problem, implicitly including soil protection. Some countries, e.g., United Kingdom, The Netherlands, and Canada, have set quantitative generic guidelines (Siegrist, 1990; Sheppard et al., 1992). The generic soil cleanup guidelines employed in the United Kingdom since 1987 are called trigger concentrations and they are specific to land end use. These limits were predominantly derived from professional judgment, but also considered a range of factors including ingestion, skin exposure, phytotoxicity, corrosion, and explosion (ICRCL, 1987). A new approach, based on a quantified risk assessment methodology with site- and population-specific parameters being replaced by probability density functions representing typical scenarios for a chosen afteruse, is currently under development (Ferguson and Denner, 1993). The generic criteria employed in The Netherlands comprise target values that represent background concentrations for naturally occurring substances and analytical detection limits for manmade organic compounds and intervention values, the thresholds above which remediation may be necessary (MHSPE, 1994; Beck et al., 1995). These standards, corrected for soil organic matter and clay content, were derived by using a behavior and exposure assessment model that considered ecotoxicological data on plants, soil fauna, and microorganisms and human toxicological data. National assessment and remediation guidelines for contaminated land were published by the Canadian Council of Ministers for the Environment in 1991 (CCME, 1991). These guidelines were derived on the basis of experience and professional judgment and were considered generally to be protective of human and environmental health for specified uses of soil. The Canadian limits are largely approximate analytical detection limits or background chemical concentrations observed in “clean” soils. Only those quality limits relevant to the case studies presented in Section IV.B.3 are considered in this review. Those readers who may be interested will find a comprehensive summary of generic soil quality limits adopted by The Netherlands, United Kingdom, and Canada in Beck et al. (1995).

B. KINETICALLYCONSTRAINED SOILQUALITY LIMIT (KCSQL) CONCEPT 1. Overview

The overall loss or dissipation of organic chemicals from soils is a biphasic process with an initial period of rapid loss usually being observed, followed by a longer period of slow release (Pignatello, 1989; see also Fig. 8). The fraction of the total organic chemicals comprising the rapidly lost phase (frequently referred

3 78

A. J. BECK ET AL.

Flgure 8. Kinetics of organic chemical dissipation in soils.

to as the rapidly reversible or labile fraction) and the residual fraction (also referred to as resistant, slowly reversible, or diffusion-limitedfraction) subject to slower release is dependent on a number of factors. These include the physicochemical properties of the chemicals themselves, physicochemical characteristics of the soil, environmental factors, and anthropogenic influences that affect the major chemical loss processes. For example, in a given soil, we might expect those chemicals with low water solubilities, low volatilities, and high sorption constants (e.g., high-molecular-weight PAHs and highly chlorinated PCBs) to have a larger residual fraction, which will be approached more slowly than the comparatively small residual fraction frequently observed for more volatile, water-soluble compounds such as naphthalene and 1,2-DCB, which are more readily lost. As we have seen earlier in this review, loss processes include leaching, volatilization, and biological and chemical degradation, each of which also exhibits biphasic patterns of behavior over time. Furthermore, within the soil microporosity, sorption-desorption mechanisms that control the distribution of organic chemicals between the solid, aqueous, and gaseous phases behave likewise (see Beck et al., 1995). This is because the kinetic constraints on the rate of supply of the compound out of the “residual pool” (i.e., soil-sorbed compounds) into the soil bulk solution to allow these dissipation processes to occur are limited by diffusion along concentrationgradients. Thus, it is now well established that the ultimate rate-limiting factor controlling the persistence of organic chemicals at the field scale is the kinetics of the compounds’ sorptiondesorption between the solid and aqueous phases within the soil microporosity. The sorption-desorption mechanisms postulated to be the primary ratelimiting phenomena can be classified into physically and chemically based categories. Chemically based mechanisms involve the formation and breaking of strong chemical bonds such as covalent bonds at specific binding sites onlwithin soil organic matter and/or mineral components. Organic chemicals bound to soils by these interactions are very stable and may be irreversibly sorbed under

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

379

prevailing environmental conditions (e.g., see Fuhr et al., 1991; Calderbank, 1989). Thus, they are generally considered to be unavailable to loss by volatilization or any other dissipation process. There is some controversy over the susceptibility of nonionic organic chemicals to such mechanisms (e.g., see Beck and Jones, 1993), with some researchers believing that these compounds are much more likely to interact with soils by a mechanism analogous to the partitioning of a nonionic chemical between water and an immiscible organic phase such as octanol (e.g., see Chiou et al., 1979). By contrast, both partitioning and specific site interactions can be reconciled with diffusive mass transfer rate-limiting mechanisms. These mechanisms can be further subdivided into film diffusion, sorption-retarded intraparticle diffusion, and intraorganic matter diffusion categories (see Brusseau et al., 1991). Film diffusion relates to the resistance encountered by an organic chemical when crossing a thin film of water surrounding soil solids, when moving from the solid phase into the bulk solution or gaseous phase. Film diffusion has been reported to be generally insignificant in comparison to the other processes discussed here (Brusseau and Rao, 1989). Retarded intraparticle diffusion has been defined as the “aqueous-phasediffusion of solute within pores of microporous particles (e.g., sand grains) mediated by retardation resulting from instantaneous sorption to pore walls,” while intraorganic matter diffusion is similar, with the exception that sorption-retarded diffusion occurs within organic matter matrices (Brusseau et al., 1991). A considerable amount of evidence has now been accumulated in favor of each of these mechanisms, which has been critically evaluated by Pignatello (1993) and Beck et af. (1995). Although this issue has yet to be completely resolved, we strongly support the arguments in favor of intraorganic matter diffusion as the predominant ratelimiting factor governing the persistence of nonionic organic chemicals in soils, but we nevertheless acknowledge that intraparticle diffusion may be important for some classes of ionic organic contaminants, e.g., chlorophenols. As a result of this biphasic phenomenon, when the residual phase of dissipation is reached, the concentrationof a given compound in a given soil is unlikely to decrease substantially further under prevailing environmental conditions. Furthermore, it has also been suggested that this phenomenon represents an intrinsic kinetic constraint on the remediation of contaminated soil (McKay and Cherry, 1989; Alexander and Scow, 1989; Travis and McInnis, 1992). Thus, we define the contaminant concentration of this residual phase as a kinetically constrained soil quality limit (KCSQL). To date, KCSQLs have not been deliberately and specifically determined by experimentation, so we have had to derive them from the data obtained in our experiments. KCSQLs for selected contaminants in sewage sludge-amended soils and control soils are compared with generic soil quality criteria adopted by The Netherlands, United Kingdom, and Canada to assess whether the practice of sludge amendment complies with existing soil quality legislation for selected organic chemicals. It must be emphasized that the

:

3 80

A. J. BECK ET AL.

KCSQL for a given organic chemical will be site specific, but because existing soil quality criteria are generic they do not recognize the potential for markedly differing behavior of chemicals in different soils. Ongoing refinement of soil quality criteria in the United Kingdom and The Netherlands is likely to result in the introduction of limits normalized on the basis of soil organic matter and/or clay content in an effort to overcome this problem (see previous discussion). A paper exploring the relationships between KCSQLs and the physicochemical properties of both organic chemicals and soils is currently being prepared by A. J. Beck and K. C. Jones (in preparation).

2. Numerical Derivation of KCSQLs The numerical derivation of KCSQLs is illustrated in Fig. 9. Chemical concentrations (C), expressed in units of pg or mglkg, are plotted against time (t) in units of days (d), months (m), or years (yr), as appropriate for the compound of interest. These data are then fitted with the best fit empirical model [see Beck et al. (1995) for further details]. To overcome the problem of varying sampling intervals, comprehensive data sets are generated from the raw data by interpolating measured data (using the best fit model in each case), such that a nominal time interval of 1 was obtained between individual data points regardless of the units used. The time, r,, at which dCldr = 10-1, was determined and substituted into the best fit model to calculate the KCSQL, i.e., C, (pglkg), such that C, is equal to C at time t,. Since dC/dt = 0 only when r+ 00, it was necessary to define the time t, for an arbitrary rate of change in contaminant concentration, i.e., dC/dt = 10-1, which can be recognized to be insignificant. Where dCldt > 10-1 during the time scale of the experiment, the best fit model can be used to forecast t, (see Fig. 9).

Ky:

a a 0

Omeaeuffdconanhationdstb o Interpolated wncenlraliondab derived from thebeet fitempirlcal model

foreuapddataderiwdhrnthebest fit empirical d e l

= dUdt = 10-1

4

t

h

Flgure 9. Derivation of KCSQLs (all notation is discussed in the text). With permission of Beck er al. 1995 and CRC Press.

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

381

For the model y = a + (b/&), the KCSQL was considered to be synonymous with parameter a. Adoption of this approach to deriving KCSQLs enables each data set to be characterized by the best fit empirical model providing the most reliable predictions possible. Furthermore, all of the different data sets remain directly comparable because each model is differentiated and, ultimately, only contaminant concentrations and dissipation times are compared. There was no requirement for complex direct comparison of the goodness of fit data from dissimilar model parameters. In using this approach, we are defining the KCSQL as the contaminant concentration (C,)in soil at a given time (t,) such that the rate of change in concentration is practically insignificant thereafter. Of course, our selection of dC/dt = 10-1 as the practically insignificant rate of change was largely subjective. It proved most useful for the data we present here, but much further research and scientific discussion may be needed before a consensus on the appropriate rate(s) of change is to be reached. A more comprehensive discussion of the numerical basis of the KCSQL concept is given by Beck et al. (1995).

3. Selected Case Studies for Organic Chemicals In this section, we present a number of case studies for selected compounds from the long-term studies discussed earlier and also examples of volatile organic chemicals (VOCs) from another experiment conducted by our research group. These groups of compounds represent a broad spectrum of environmental behavior ranging from nonpersistent, volatile, relatively water-soluble compounds (e.g., solvents) to recalcitrant, relatively insoluable, nonvoltile contaminants (e.g., high-molecular-weight aromatic hydrocarbons). Furthermore, these groups of compounds are included in priority pollutant lists adopted by both the EU and the U.S. EPA. We begin by considering simple examples where the dissipation of organic chemicals has been monitored following a single sewage sludge application or following the cessation of multiple sludge amendments. Later, we discuss the effects of multiple/periodic sludge amendments from a soil quality perspective. Volatile organic compounds such as toluene are frequently detected in sewage sludges applied to soils. Following sludge amendment, most of these compounds are volatilized within short time periods. However, we must stress that some volatile compounds have been reported to reside in soils over many years (e.g., see Steinberg et al., 1987; Pignatello, 1989), albeit at extremely low concentrations that are frequently less than typical contemporary analytical detection limits. Such residues are unlikely to exceed Dutch target values. This has been confirmed by our own investigations of toluene, ethylbenzene, mpxylene, and o-xylene in a clay loam soil under grass ley which received 50 tons (dry weight) of sludge/ha. The concentrations of these four compounds in the cultivated

3 82

A. J. BECK ET AL.

0

o

a

16

a4

31

40

48

5s

64

Days after sludge amendment Flpre 10. Derivation of KCSQL from the loss of mpxylene from a clay. With permission of Beck et al. 1995 and CRC Press.

horizon (20 cm) exceeded the Dutch target value for them of 50 pg/kg soon after sludge amendment, but declined below this within 24 hr (e.g., see Fig. 10). None of these compounds were detected above 3.0 pg/kg in the control plot. KCSQLs of 5 , 3 , 5 , and 1 pg/kg (to the nearest microgram) reached after 2 , 3 , 2, and 2 days (to the nearest day) for toluene, ethyl-benzene, mpxylene and o-xylene, respectively, were at least 10 and 20 times lower than the Dutch target value and the Canadian assessment (and also agricultural remediation) values, respectively. No UK trigger concentrations have been set for these compounds. Dutch target values have been set at 1.0 Fg/kg (e.g., carbon tetrachloride, tetrachloroethane, trichloroethane, trichloroethylene, and trichloromethane) or the detection limit (e.g., I ,3-dichloropropene, dichloromethane, dichloroethane, and vinyl chloride) for many halogenated aliphatic compounds. Although such low target values for these compounds may be scientifically defensible, it is inappropriate to set a quality limit at a compound’s detection limit. This may encourage contention of the legislation on the basis of what constitutes appropriate compound detection limits. Furthermore, because detection limits are continuously refined and/or decreased with improvements in sampling and analysis, it will become increasingly difficult, even for highly volatile compounds, to realize target values in soils. If typical, the results for CBs from our Woburn experiment would suggest that sewage sludge amendment is unlikely to be a problem so far as contemporary soil quality legislation is concerned. None of the individual chlorinated benzenes, mono- through hexachlorobenzene, were found to be above Dutch target values of 10 (mono- to tetrachlorinated benzenes) or 2.5 pg/kg (penta- and hexachlorobenzene), even following 25 successive annual additions of sewage sludge. Contemporary soil concentrations, over 20 years after the cessation of sludge amendment, are less than 1 pg/kg for all of the CBs (see Section 1II.A.1).

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

0

u

16

39

51

383

65

Yeara after sludge amendment

Rgure 11. Derivation of KCSQLs from the loss of XFCBs from control plots at five long-term sites and from the sludge-amended plot at Wobum.

For our case study of PCBs, we have integrated the results of Alcock et al. (1993) for ZPCB long-term persistence in the control plots at all of the experimental sites in the south of England (i.e., Woburn, Luddington, Lee Valley, Park Grass, Broadbalk, and Hoosfield) on the basis that these authors consider these results to be typical of the region as a whole (see Section II.A.2). Maximum soil concentrations in control plots were 298-555 pg ZPCBs/kg soil during the late 1960s and early 1970s (see Section II.A.2). These were between 24 and 52 times greater than the Dutch target value of 20 pg ZPCBs/kg soil. ’benty-six years later (1992) the concentration of ZPCBs in the control plot soil had fallen to around this level (Fig. 11) as a result of natural loss processes, primarily believed to be volatilization to the atmosphere (see Section III.C.4). This result is of great environmental significance because if, as Alcock et al. (1993) suggest, these results are typical contemporary background levels, then “normal” soils are currently falling below soil quality limits. However, we should stress that, although this may be so, evidence remains that concentrationscurrently present in the environment may be having an adverse impact on wildlife and human health. This may be due to the lag time required for reductions in biota at the top of food chains to take effect [Section II.A.2; see also Loganathan and Kannan (1994)l. This clearly illustrates the usefulness of an integrated multimedia (i.e., soil-airwater-biota) approach to setting environmental standards. The ZPCB concentrations in the cultivated horizon of the sludge-amended plot at the Woburn site (which received multiple sludge amendments) can be seen to be elevated relative to their concentrations in the control plot. However, the rate of loss was slower, and Fig. 11 suggests that the XPCB concentration is unlikely to fall below the Dutch target value until about the year 2030, should dissipation continue at current rates. At Luddington and Lee Valley, XPCB concentrations were lower and their rate of dissipation in sludge-amended plots was more rapid than at

3 84

A. J. BECK ET AL.

Woburn, so that the Dutch target value is likely to be reached earlier. Indeed, PCB concentrations in some of the sludge-amended plots have already reached control plot concentrations (see Section III.A.2). Further declines in the ZPCB concentrations in both the sludged and control plots appear likely. The trends in Fig. 11 could only be fitted with single exponential models, suggesting that a residual phase, strongly rate-limited by diffusive mass transfer, has not yet been reached, so that a KCSQL cannot yet be assigned to these sites. However, monitoring at the sites will continue, so that it should be possible to identify the point at which the loss of PCBs becomes more strongly diffusion limited. The persistence of ZPAHs following the cessation of sludge amendment in 1961 at the Woburn site was more discouraging than the behavior of the CBs or PCBs. Over a 23-year period (approximately the same as for successive sludge additions), only about half the ZPAHs added to the soil were lost, and a KCSQL of 2231 pg of ZPAHs/kg of soil (95% CI = 1297-3166 pg/kg) suggests that further reduction in the concentration was likely to be minimal and would require prolonged periods without further sludge applications (Fig. 12). Furthermore, there is some evidence for an increase in the background concentrations of ZPAHs in soils that received no sewage sludge amendments as a result of continuous atmospheric deposition (see Section III.A.3), although Wild et al. (1990) could not rule out the possibility that this may have been due to transfer of PAHs during the plowing of sewage sludge-amended plots nearby (see Section III.C.6). Despite this, the difference in the KCSQL for the ZPAHs in the sludgeamended soil (i.e., 2231 pg ZPAHs/kg soil) and their contemporary concentration in the control plot receiving no sewage sludges (approximately 700 p,g of ZPAHs/kg of soil) is significant.

d

io

n'0

30

40

50

Yeam after sludge amendment

Rgure 12. Derivation of KCSQL from the loss of XPAHs from the sludge-amended plot at Wobum. With permission of Beck et al. 1995 and CRC Press.

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

385

Contemporary agricultural practice normally involves multiple applications of sewage sludge to the same farmland year after year, similar to the Woburn experimental plot discussed in this review. Thus, for many agricultural soils, increasing contaminant loadings will be the norm. One of the pertinent questions, so far as soil quality legislation is concerned, is how many sludge applications will be possible under typical contemporary practice without contravening existing soil quality legislation? Although set up to meet other objectives, the Woburn experiment provides valuable insight into this problem. The ZPAH concentration in the cultivated horizon of the Woburn sludgeamended plot increased from 216 pg/kg in 1942 to 5527 pg/kg in 1960, the year before cessation of sludge amendments. Although these concentrations are greater than the ZPAH Dutch target value (1000 pg XPAH/kg soil), they are nevertheless encouraging insofar as they demonstrate that prolonged application (in this case 25 successive years) of sewage sludge to agricultural land was possible without any contravention of existing UK quality criteria or the Dutch intervention value (200,000 pg ZPAHs/kg soil). This would suggest that it would have been possible to make further annual applications of sewage sludges to this soil and remain safely within existing quality legislation without any legal obligation for soil cleanup. As noted earlier, CBs did not exceed Dutch target values following 25 successive sludge applications. We have refrained from extrapolating the available data for any of these compound groups to predict how many successive applications may be possible without contravening existing soil quality legislation because the data are very limited (essentially two points), and we are unsure of how typical this trend may be for other sludges/soils/environmental conditions. The critical factor affecting the buildup of residual fractions of organic chemicals in soils is the half-life of the contaminant relative to the time interval between applications. Where the time elapsed between applications is significantly less than the half-lives of the contaminants, such as for the PAHs in the Woburn sludge-amended plot, an elevation in the residual fraction will occur. By contrast, consider the effect of multiple applications of the herbicide atrazine (0.8 kg/ha 30 September 1986, 2.4 kglha 7 May 1987, and 2.0 kg/ha 10 May 1989) to a corn crop on a clay loam soil in Canada (see Frank ef al., 1991). Although a fraction of the atrazine added in each application was carried over to supplement that added with subsequent applications, there was no evidence for a buildup in the residual fraction with time. The half-lives for each atrazine application ranged from 42 days during the growing season of 1987 to 327 days for the winter of 1986-1987, while the residual fraction in the top 0.15 m of soil just prior to each application ranged from 60 to 260 pg/kg [X(parent compound metabolite desethylatrazine)]. Such behavior may also be typical of many compounds occurring in sewage sludges that we have not yet investigated.

+

386

A. J. BECK ET AL.

V. CONCLUSIONS A number of general conclusions can be drawn from the detailed discussions of organic chemical persistence and dissipation chronologies presented in Section III. The application of sludges to agricultural land generally resulted in elevated concentrations of the compounds in the cultivated horizons relative to soils that received no sewage sludge. For most of the compounds considered, volatilization to the atmosphere was the most significant loss pathway from the soil. Furthermore, atmospheric deposition was identified as an important source of some compounds, principally the PCBs and l,CDCB, to the soil. Thus, it is of great importance to consider the effects of other processes such as air/soil exchanges when investigating the effects of sludge applications to land. Other loss processes such as leaching and degradation were believed to be of little significance when compared to volatilization. The biphasic phenomenon characteristic of diffusion-limited dissipation was observed for most compounds, so that first-order half-lives presented may underestimate the long-term persistence of many compounds. Despite this, the half-lives presented here are significantly longer than those that have frequently been reported for CBs, PCBs, and PAHs in other field and laboratory experiments. This clearly indicates the value of such long-term studies to environmental risk assessment procedures. Elevated concentrations of some organic contaminants resulting from sewage sludge applications are unlikely to contravene existing soil quality limits in countries such as The Netherlands, United Kingdom, and Canada. Where limits were exceeded following sewage sludge amendment for some compounds (e.g., volatile industrial solvents), natural loss processes, principally volatilization, quickly reduced the soil concentrations to acceptable levels. It may be argued that in some cases these studies may not be typical or representative of contemporary agricultural practices. However, this in itself is not a problem because the sludges used were often the “dirtiest” possible, and application rates were sometimes up to 5 times greater than typical contemporary rates. Thus, it is possible that the results presented here may represent more extreme contamination scenarios. Furthermore, the extent of the differences between the concentrations of many contaminants in sludge-amended soils reported here and those concentrations designated as intervention values by The Netherlands leaves ample potential for an increase in typical sludge application rates in the future should this become necessary as a result of a need for increased sludge disposal to land following the ban on disposal at sea starting in 1998. If increased application of sludge to land is considered to be a viable option in the impending disposal problem, then we recommend that further research will be necessary to determine how many multiple applications of sludge can be made to land such that soil quality limits are not breached. Although the natures of sludges, soils,

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

387

agricultural practices, and environmentalconditions vary from place to place, we believe that, with the appropriate research to characterize these problems, it will be possible to draw up a series of practical guidelines that would allow for increased but legally acceptable sludge application to land in the future.

ACKNOWLEDGMENTS We are grateful to the United Kingdom Ministry of Agriculture, Fisheries and Food, Department of the Environment, Natural Environment Research Council, and the Water Research Centre for funding our research on organic chemicals in soils. We are also grateful to Rothamsted Experimental Station and Macaulay Land Use Research Institute for allowing us access to the long-term experiments and archived soil, sludge, and herbage collections, without which this work would not have been possible. Grateful thanks are also extended to Johnny Johnston and Steve McGrath for their invaluable support in these studies.

REFERENCES Alcock, R. E., Johnston, A. E., McGrath, S. P., Berrow, M. L., and Jones K. C. (1993).Long-term changes in the polychlorinated biphenyl (PCB) content of United Kingdom soils. Environ. Sci. Technol. 27, 1918-1923. Alexander, M., and Scow, K. M. (1989). Kinetics of Biodegradation in Soil. In “Reactions and Movement of Organic Chemicals in Soils” (B. L. Sawhney and K. Brown, eds.), pp. 243-269. Soil Sci. Soc. Am. Inc., Am. Soc. Agron. Inc., Madison, WI. Anonymous. (1989).The sludge dilemma. Operations Forum 4, 15-16. Beck, A. J., and Jones, K. C. (1993). Natural organic substances and contaminant behaviour: progress, conflicts and uncertainty. In “Organic Substances in Soil and Water: Natural Constituents and Their Influences on Contaminant Behaviour” (A. J. Beck, K. C. Jones, M. B. H. Hayes, and U. Mingelgrin, eds.), pp. 184-194. Royal Society of Chemistry, Cambridge. Beck, A. J., Wilson, S. C., Alcock, R. E., and Jones, K. C. (1995).Kinetic constraints on the remediation of soils contaminated with organic chemicals: implications for soil quality limits. Crii. Rev. Environ. Sci. Technol. 25(1), 1-43. Bedard, D. L., et al. (1987).Extensive degradation of Aroclors and environmentally transformed polychlorinated biphenyls by Alcaligenes eurrophus H850.Appl. Environ. Microbiol. 53,10941102. Berrow, M. L.. and Burridge, J. C. (1990).Persistence of metal residues in sewage sludge treated soils after seventeen years. Int. J. Environ. Anal. Chem. 39, 173-177. Berthouex, P., and Gan, D. R. (1991). Loss of PCBs from municipal sludge-treated farmland. J. Environ. Eng. 11, 5-24. Bossert, 1. D., and Bartha, R. (1986).Structure-biodegradability relationships of polycyclic aromatic hydrocarbons in soil. Bull. Environ. Contam. Toxicol. 37, 490-495. Bruce, A. M., and Davis, R. A. (1989).Sewage sludge disposal: current and future options. Wat. Sci. Technol. 21, I 113-1 128. Brusseau, M. L., and Rao, P. S. C. (1989). Sorption nonideality during organic contaminant transport in porous media. CRC Crit. Rev. Environ. Control. 1, 33-99. Brusseau, M. L., Jessup, R. E., and Rao, P. S. C. (1991).Nonequilibrium sorption of organic chemicals: elucidation of rate-limiting processes. Environ. Sci. Technol. 25, 134- 142.

3 88

A. J. BECK ET AL.

Calderbank, A. (1989). The Occurrence and significance of bound pesticide residues in soil. Rev. Envimn. Conrum. Toxicol. 1oS, 71-103. CCME. (1991). Interim Canadian Environmental Quality Criteria for Contaminated Sites. Report CCME EPC-CS34, September, 1991. Canadian Council of Ministers of the Environment, Winnipeg, Manitoba. CEC. (1986). Directive on the protection of the environment and in particular of the soil when sewage sludge is used in agriculture. Official Journal L 187/6, 4 July 1986. Council of the European Communities. CEC. (1992). COM (92) 23 Final, Volume 11, Towards Sustainability, Brussels, 27 March 1992. Chang, A. C., and Page, A. L. (1985). Fate of wastewater constituents in soil and groundwater: Trace organics. In “Irrigation with Reclaimed Municipal Wastewater-A Guidance Manual” (P. G. Stuart and T.Asano, eds.), pp. 15-1-15-20. Lewis Publishers, Chelsea, MI. Chiou, C. T., Peters, L. J., and Freed, V. H. (1979). A physical concept of soil-water equilibria for nonionic organic compounds. Science 206, 831-832. Crathorne, B., Donaldson, K., James, H. A,, Rogers, H. R. (1988). The determination of organic contaminants in U.K. sewage sludges. I n “Organic Contaminants in Wastes, Water, Sludge and Sediment-Occurrence, Fate and Disposal” (D. Quaghebeur, I. Temmerman, and G. Angeletti, eds.), pp. 45-65. Elsevier Applied Science, London. Diercxsens, P., and Tarradellas. J. (1987). Soil contamination by some organic micropollutants related to sludge spreading. fnr. J. Environ. AM^. Chem. 28, 143-159. Duarte-Davidson, R., and Jones, K. C. (1994). Polychlorinated biphenyls in the UK population: estimated intake, exposure and body burdens for the general population. Sci. Torul Environ. 151, 131-152. Dzombak, D. A., Labieniec, P. A., and Siegrist, R. L. (1993). The need for uniform soil cleanup goals. Envimn. Sci. Technol. 27, 765-766. Epstein, E., Taylor, J. M., and Chaney, R. L. (1976). Effects of sewage sludge and sludge compost applied to soil on some soil physical and chemical properties. J. Envimn. Quul. 5, 423-426. Fairbanks, B. C., O’Connor, G. A.. and Smith, S. E. (1987). Mineralisation and volatilisation of PCBs in sludge-amended soils. J. Envimn. Qwl. 16, 18-24. Fathespure, B. Z., Tiedje, J. M., and Boyd, S. A. (1988). Reductive dechlorination of hexachlorobenzene to tri- and dichlorobenzenes in anaerobic sewage sludge. Appl. Envimn. Microbiol. 2, 327-330. Ferguson, C., and Denner, J. (1993). Soil guideline values in the UK: New risk-based approach. In “Contaminated Soil” (F. Arendt, G. J. Annokkee, R. Bosman, and W. J. Van den Brink, eds.), pp. 365-372. Kluwer Academic Publishers, Dordrecht. The Netherlands. Frank, R., Clegg, B. S., and Patni, N. K. (1991). Dissipation of atrazine on a clay loam soil, Ontario, Canada, 1986-90. Arch. Envimn. Conrum. Toxicol. 21, 41-50. Fries, G. F. (1991). Organic contaminants in terrestrial food chains. I n “Organic Chemicals in the Environment” (Jones, K. C., ed.), pp. 207-236. Elsevier Applied Science Publications, London. Frink, C. R., and Hullar, (1985). Criteria and recommendationsfor land applications in the northeast. Penn. Agric. Exp. Srn. Bull., 851. Fuhr. F., Steffens, W., Mittelstaedt, W., and Brumhard, B. (1991). Lysimeterexperimentswith Wlabelled pesticides-an agroecosystem approach. Pesr. Chem., 37-47. Can, R., and Berthouex. P. M. (1993). Disappearance and crop uptake of PCBs from sludgeamended farmland. War. Environ. Res. 66, 54-69. Goldstein, N. (1989a). EPA sludge disposal regulation proposed. BioCycle 3, 44-50. Goldstein, N. (1989b). EPA tackles sludge rules comments. EioCycle 9, 56-59. Goldstein, N. (1989~).Scientists assess 503 proposal. BioCycle 9, 60-61. Goodin, J. D.. and Webber, M. D. (1991). Greenhouse studies with 14C-labelledorganic contaminants. Preliminary report. Environment Canada, Burlington, Ontario.

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

389

H m d , S. J., Sewart, A. P., Alcock, R. E., Boumphrey, R., Burnett. V., Duarte-Davidson, R., Halsall, C., Sanders, G., Waterhouse, K., Wild, S. R., and Jones, K. C. (1994). Polychlorinated biphenyls (PCBs) in the British environment: sinks, sources and temporal trends. Environ. Pollut. 85, 131-147. Heitkamp, M. A., and Cerniglia, C. E. (1987). Effects of chemical structure and exposure on the microbial degradation of polycyclic aromatic hydrocarbons in freshwater and estuarine systems. Envimn. Toxicol. Chem. 6, 535-546. Heitkamp, M. A., Franklin, W., and Cerniglia, C. E. (1988). Microbial metabolism of polycyclic aromatic hydrocarbons: isolation and characterisation of a pyrene degrading bacterium. Appl. Environ. Microbiol. 54, 2549-2555. Herbes, S. E., and Schwall, L. R. (1978). Microbial transformation of polycyclic aromatic hydrocarbons in pristine and petroleum contaminated sediments. Appl. Environ. Microbiol. 35, 306316. Howard, P. H. (1989). “Handbook of Environmental Fate and Exposure Data for Organic Chemicals,” Vol. 1. Lewis Publishers, Chelsea, MI. IARC. (1983). Monographs on the evaluation of the carcinogenic risk of chemicals to humans: polynuclear aromatic hydrocarbons, Vol. 32. WHO, Lyons, France. ICRCL. (1987). Interdepartmental Committee on the Redevelopment of Contaminated Land. Guidance on the Assessment and Redevelopmentof Contaminated Land. 59/83, 2nd Ed., July 1987. HMSO, UK. IPCS. (1991). Chlorobenzenes other than hexachlorobenzene. World Health Organisation, Geneva. Jacobs, L. W., O’Connor, G. A., Overcash, M. A., Zabik, M. J., and Rygiewicz, P. (1987). Effects of trace organics in sewage sludges on soil-plant systems and assessing their risk to humans. I n “Land Application of Sludge” (A. L. Page, T. G. Logan, and 1. A. Ryan), pp. 101-143. Lewis Publishers, Chelsea, MI. Johnston, A. E., and Gardner, H. V. (1969). Broadbalk: Historical introduction. Report of Rothamsted Experimental Station for 1986. Part 2. Jones, K. C., and Wild, S. R. (1991). “Organic Chemicals Entering Agricultural Soils in Sewage Sludges: Screening for Their Potential to Transfer to Crop Plants and Livestock.” Foundation for Water Research, Report No. FR 0169, 143 pp. Marlow, Buckinghamshire, UK. Jones, K. C., Stratford, J. A., Waterhouse, K. S., Furlong, E. T., Giger, W.. Hites, R. A,, Schdner, C., and Johnston, A. E. (1989a). Increases in the polynuclear aromatic hydrocarbon content of an agricultural soil over the last century. Environ. Sci. Technol. 23, 95-101. Jones, K. C., Stratford, J. A., Waterhouse, K. S., and Vogt, N. B. (1989b). Organic compounds in Welsh soils: polynuclear aromatic hydrocarbons. Environ. Sci. Technol. 23, 540-550. Jones, K. C., Stratford. J. A,, Tidridge, P., Waterhouse, K. S., and Johnston, A. E. (1989~). Polynuclear aromatic hydrocarbons in an agricultural soil: long-term changes in profile distribution. Environ. Pollut. 56, 337-351. Jones, K. C., Grimmer, G., Jacob, J., and Johnston, A. E. (1989d). Changes in the polynuclear aromatic hydrocarbon (PAH) content of wheat grain and pasture grassland over the last century from one site in the UK. Sci. Total Environ. 78, 117-130. Kampe, W., and Leschbar, R. (1989). Occurrence of organic pollutants in soils and plants after intensive sewage sludge applications. I n “Organic Contaminants in Wastewater, Sludge and Sediment: Occurrence, Fate and Disposal” (D. Quaghebeur, I. Temmerman, and G. Angeletti, eds.), pp. 35-41, Elsevier Applied Science, London. Katterman, F. R. H., and Day, A. D. (1989). Plant growth factors in sewage sludge. BioCycle 3,6465. Kaufman, D. D. (1983). Fate of toxic organic compounds in land-applied wastes. In “Land Treatment of Hazardous Wastes” (Pam, F., ed.), pp. 77-151. Noyes Data C o p , Park Ridge, NJ. Loganathan, B. G., and Kannan, K. (1994). Global organochlorine contamination trends: An overview. Ambio 23, 187-191.

3 90

A. J. BECK ET AL.

Marinucci, A. C., and Bartha, R. (1979).Biodegradation of 1,2.3-trichlorobenzene in soil and in liquid enrichment culture. Appl. Envimn. Microbial. 11, 81 1-817. McGrath, S. P. (1987).Long-term studies of metal transfers following application of sewage sludge. In “Pollutant Transport and Fate in Ecosystems” (P. J. Coughtrey, M. H. Martin, and M. H. Unsworth, eds.), pp. 301-317. British Ecological Society, Spec. Pub. No. 6.Blackwell Scientific Publications, Oxford. McKay, D. M., and Cheny, J. A. (1989).Groundwater contamination: pump-and-treat remediation. Environ. Sci. Technol. 23, 630-636. McLachlan, M. S. (1993).Mass balance of PCBs and other organochlorine compounds in lactating cows. J. Agric. Food Chem. 4,474-480. McVeety, B. D., and Hites, R. A. (1988).Atmospheric deposition of polycyclic aromatic hydrocarbons to water surfaces: A mass balance approach. Armos. Environ. 22, 51 1-536. MHSPE. (1994).Environmental quality objectives in The Netherlands. Ministry of Housing, Spatial Planning and the Environment, The Hague, The Netherlands. Moza, P., Scheunert, I., Klein, W., and Korte, F. (1979).Studies with 2,4‘.5-tri~hIorobiphenyl-~~C and 2,2’,4,4‘,6-penta~hlorobiphenyl-’~C in carrots, sugar beets and soil. J . Agric. Food Chem.

27, 1120-1124. Oliver, B. G . , and Nichol, K. D. (1982).Chlorobenzenes in sediments, water and selected fish from Lakes Superior, Huron, Erie, and Ontario. Environ. Sci. Technol. 16, 532-536. Park, K. S., Simms, R. C., Dupont, R. R., Doucette, W. J., and Matthews, J. E. (1990).Fate of PAH compounds in two soil types: influence of volatilisation, abiotic loss and biological activity. Environ. Toxicol. Chem. 9, 187-195. Peterson, A. E., Speth, P. E., and Schletcht, P. (1988).Effect of sewage sludge application on groundwater quality. I n “Eleventh Annual Madison Waste Conference,” pp. 272-285. Madison, WI. Pignatello, J. J. (1989).Sorption dynamics of organic chemicals in soils and sediments. I n “Reactions and Movement of Organic Chemicals in Soils” (B. L. Sawhney and K. Brown, eds.), pp. 271304. Soil Sci. Soc.Am. Inc., Am. Soc.Agron. Inc., Madison, WI. Pignatello, J. J. (1993).Recent advances in sorption kinetics. In “Organic Substances in Soils and Water” (A. J. Beck, K. C. Jones, M. H. B. Hayes, and U. Mingelgrin, eds.), pp. 128-140. Royal Society of Chemistry, Cambridge, UK. Rappaport, R. A., and Eisenreich, S. J. (1988). Historical atmospheric inputs of high molecular weight chlorinated hydrocarbons to eastern North America. Environ. Sci. Technol. 22, 931-

941. Ryan, J. A., Bell, R. M., Davison, J. M., and O’Connor, G. A. (1988).Plant uptake of nonionic organic chemicals from soils. Chemosphere 17, 2229-2323. Safe, S. (1994).Polychlorinated biphenyls (PCBs): Environmental impact, biochemical and toxic responses and implications for risk assessment. Crir. Rev. Toxicol. 24, 87-149. Sauerbeck, D. (1987). Effects of agricultural practices on the physical, chemical and biological properties of soils: Part 11-use of sewage sludge and agricultural wastes. In “Scientific Basis for Soil Protection in the European Community” (H. Barth and P. L‘Hermite, eds.), pp. 181210.Elsevier Applied Science Publishers, London. Sheppard, S. C., Gaudet, C., Sheppard, M. I., Cureton, P. M., and Wong, M. P. (1992).The development of assessment and remediation guidelines for contaminated soils, a review of the science. Can. J. Soil Sci. 2, 359-394. Siegrist, R. L. (1990). Development and implementation of soil quality and clean up criteria for contaminated sites. In “Contaminated Soil 1990” (K. Wolf, J. Van den Brink, and F. J. Colon, eds.), pp. 149- 156. Kluwer Academic Publishers, Dordrecht, The Netherlands. Smelt, J. H.,and Leistra, M.(1974).Hexachlorobenzenes in soils and crops after soil treatment with pentachloronitrobenzne. Agric. Environ. 1, 65-71.

ORGANIC CHEMICALS IN SLUDGE-AMENDED LAND

391

Spain, J. C. (1990). Metabolic pathways for biodegradation of chlorobenzenes. I n “PseudomonasBiotranformations, Pathogenesis, and Evolving Biotechnology” (S. Silver, A. M. Chakrabarty, B. Iglewski, and S. Kaplan, eds.), pp. 197-206. American Society for Microbiology, Washington, D.C. Steinberg, S. M., Pignatello, J. J., and Sawhney, B. J. (1987). Persistence of 1.2-Dibromoethane in Soils: Entrapment in Intraparticle Micropores. Environ. Sci. Technol. 21, 1201- 1207. Tanabe, S. (1988). PCB problems in the future: foresight from current knowledge. Environ. Pollur. 50, 5-28. Travis, C. C., and McInnis. J. M. (1992). Vapor extraction of organics from subsurface soils. Environ. Sci. Technol. 26, 1885-1887. Van den Berg, R., Denneman, C. A. J., and Roels, J. M. (1993). Risk assessment of contaminated soil: Proposals for adjusted, toxicologically based Dutch soil clean up criteria. I n “Contaminated Soil 1993” (F. Arendt, G. J. Annokkee. R. Bosman, and W. J. Van den Brink, eds.), pp. 349364. Kluwer Academic Publishers, Dordrecht, The Netherlands. Vinten, A. J., Y m n , B., and Nye, P. H. (1983). Vertical distributions of pesticides into soil when adsorbed on suspended particles. J . Agric. Food Chem. 31, 662-664. Wang, M., and Jones, K. C. (1994a). Behaviour and fate of chlorobenzenes introduced into soil-plant systems by sewage sludge application: a review. Chemosphere 28, 1325-1360. Wang. M., and Jones, K. C. (1994b). Behavior and fate of chlorobenzenes in spiked and sewage sludge-amended soil. Environ. Sci. Technol. 28, 1843- 1852. Wang, M., and Jones, K. C. (1994). Uptake of chlorobenzenes by carrots from spiked and sewage sludge amended soil. Environ. Sci. Technol. 28, 1260-1267. Wang, M., McGrath, S. P., and Jones, K. C. (1995). Chlorobenzenes in field soil with a history of multiple sewage sludge applications. Environ. Sci. Technol., in press. Warren, R. G., and Johnston, A. E. (1969). Hoosfield continuous barley. Report of Rothamsted Experimental Station for 1968. Part 2, pp. 12-25. Wild, S. R. (1991). The fate and behaviour of PAHs in sewage sludge amended agricultural soils and their uptake by plants. Ph.D. Thesis, Lancaster University. Wild, S. R., and Jones, K. C. (1992a). Uptake of polynuclear aromatic hydrocarbons (PAHs) by carrots (Daucus carora) grown on freshly sewage sludge amended agricultural soils. J. Environ. Q ~ a l21, . 217-225. Wild, S . R., and Jones, K. C. (1992b). Organic chemicals entering agricultural soils in sewage sludges: screening for their potential to transfer to crop plants and livestock. Sci. Total Environ. 119, 85-119. Wild, S. R., and Jones, K. C. (1995). Polynuclear aromatic hydrocarbons in the United Kingdom environment: an assessment of sources and sinks. Environ. Pollur. 87, in press. Wild, S. R., Waterhouse, K. S., McGrath, S. P., and Jones, K. C. (1990). Organic contaminants in an agricultural soil with a known history of sewage sludge amendments: polynuclear aromatic hydrocarbons. Environ. Sci. Technol. 24, 1706-171 1. Wild, S. R., Harrad, S. J., and Jones, K. C. (1994). The influence of sewage sludge application to agricultural land on human exposure to polychlorinateddibenzo-p-dioxins and furans. Environ. Pollur. 83, 357-369.

This Page Intentionally Left Blank

PEANUT BREEDINGAND GENETICS David A. Knauft and Johnny C. Wynne Department of Crop Science North Carolina State University Raleigh, North Carolina 27695

I. Introduction

II. Diversity of Peanut A. Distribution

B. Taxonomy C. Production D. Breeding 111. Genetic Variability A. Natural B. Genetic Transformation IV. Genetics A. Qualitative B. Quantitative V. Breeding A. General B. Hybridization C. Early Generation Testing D. Breeding Methods E. Genotype x Environment Interaction F. Sustainable Systems VI. Research Related to Breeding A. Physiology B. Soil Deficiencies and Toxicities C. Biotic Stresses D. Expanded Uses VII. Summary and Conclusions References

I. INTRODUCTION Progress in peanut (Aruchishypogaeu L.) breeding and genetics lagged behind that of many major crops primarily because of the relatively few scientists assigned to the crop and the scarcity of financial resources to support research in countries where the crop is grown (Wynne et al., 1991). However, considerable progress in breeding, genetics, and related disciplines has been made during the 393 &mrs m Agmnomy, V d m YI

Copyright 0 1995 by Aademic

Press,Inc. All rights of reproductionin any form reserved

394

D. A. KNAUFT AND J. C.WYNNE

last two decades. Breeding has received high priority from scientists at the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), which adopted the peanut as a mandate crop in 1976 (Gibbons, 1980). Funding of the Peanut Collaborative Research Support Program (CRSP) in 1982 by the United States Agency for International Development (USAID) also increased the worldwide effort on breeding of the peanut. This review primarily summarizes accomplishments in peanut breeding and genetics and in related disciplines that have occurred since peanut research has received expanded efforts in the United States and other countries during the last two decades.

II. DIVERSITYOFPEANUT A. DISTRIBUTION The genus Aruchis contains a rich diversity of plant types. Both annuals and perennials are known, and although most species reproduce by seed, some are rhizomatous and reproduce largely through vegetative means. The species occur in regions as different as poorly drained, swampy areas near sea level, to drought conditions, to mountainous regions at elevations up to 1600 m. The genus Aruchis is naturally restricted to the countries of Brazil, Argentina, Bolivia, Paraguay, and Uruguay in South America (Singh and Simpson, 1994). It is generally believed that the peanut cultivated for food and oil around the world, Aruchis hypogueu L., originated in southern Bolivia or northern Argentina (Gregory et ul., 1980).

B. TAXONOMY The taxonomy of the Arachis species has been confusing. Different species have been assigned the same name, and the same species have received different names. Many collected Arachis species have not been named or have received unauthorized names in the literature. A recent monograph named and described 69 Arachis species, and this work should improve peanut taxonomy (Krapovickas and Gregory, 1994). Sexual reproduction in the genus is distinctive in the plant kingdom and helps characterize the genus. After pollen travels through the stylar tissue and fertilization takes place in the ovule, a meristematic region basal to the embryo begins active cell division. The fertilized embryos are pushed away from the branches on which the flowers are borne in a structure called the gynophore, more commonly called the peg. The peg is geotropic and can grow more than 30 cm before

PEANUT BREEDING AND GENETICS

395

reaching the soil surface. The peg continues to grow into the soil, and once it reaches its maximum depth it becomes horizontal. The pod begins to form through rapid growth of endocarp tissue between the ovule and shell layers. As the ovules grow, the endocarp tissue becomes smaller and disappears after seeds are mature. Under ideal conditions of temperature, moisture, and nutrition, mature seeds of A. hypogaea are formed 60 days after fertilization. The cultivated peanut (A. hypugueu) is a tetraploid. The species is divided into two subspecies, and each subspecies is divided into two botanical varieties. The subspecies hypogaea is characterized by alternate branching, a lack of inflorescences on the main stem, and the first branch on the cotyledonary lateral always being vegetative. Seed dormancy is usually present. The variety hypogaea has a main axis usually under 50 cm, with stems having minimal pubescence. The variety hirsura has main axes often longer than 1 m, with stems pubescent and very late maturing compared with the variety hypoguea. The subspeciesfasrigiara has sequential branching, with inflorescences always present on the main axis and the first branch on the cotyledonary lateral always being reproductive. Seed dormancy is usually absent. The varietyfasrigiara has vegetative branches on primaries absent or usually placed at the distal nodes, simple inflorescence, and pods with two or more seeds. The variety vulgaris has vegetative branches occasionally and irregularly placed, compound inflorescence, and two-seeded pods. In the United States, three of the four botanical varieties are grown comrnercially. A. hypogaea subsp. hypogaea var. hypogaea is divided into two market classes, mnner and Virginia. The two classes are characterized by differences in pod size, with Virginia pods being larger than runner pods. Virginia types are historically grown in North Carolina and Virginia and are used for in-shell products and roasted nuts. The runner types are grown in the southeastern United States, primarily Georgia, Alabama, and Florida, and in the Southwest (Texas and Oklahoma), with the major use being for peanut butter or confectionery purposes. The Spanish market class corresponds to the fasrigiara subspecies, variety vulgaris, grown in the Southwest for confectionery and roasted nut uses. Valencia peanuts (ssp. fasrigiura var. fasrigiara are grown in New Mexico for confectionery uses and in the southeastern United States where they are harvested immature and consumed after boiling in a salt brine. Although A. hypogaea is tetraploid, most members of the genus are diploid. Plausible origins of A. hypoguea have been proposed for many years, on the basis of cytological and genetic studies. Gregory and Gregory (1976) postulated that Arachis duranensis and Aruchis cardenasii intercrossed to produce A. hypogaea. However, A. hypogaeu has a distinct pair of chromosomes present only in the B genome, and the B genome is present only in Aruchis batizocoi. No successful hybrids between species carrying the A genome and A. batizocoi have been made, causing researchers to examine other possible origins for cultivated

3 96

D.A. KNAUFT AND J. C. WYNNE

peanut. Recent genetic studies of molecular variability among Aruchis species has confirmed the large genetic distance between A. hypogueu and A . butizocoi (Ro et ul., 1992; Halward et ul., 1992). These same studies have shown considerable similarity between A. hypogueu and A. durunensis (Ro et ul., 1992) and between A. hypogueu and A . durunensis, Aruchis ipuensis, and Aruchis speguzzinii (Kochert et ul., 1991).

C. PRODUCTION Peanuts are grown in tropical and warmer temperate regions throughout the world. Global production averages nearly 24 million metric tons from almost 20 million hectares. Average yields vary from approximately 0.50 to nearly 3.0 metric tons per hectare. China and India together account for over half the world’s production. Other major producing regions are southeastern Asia, western and southern Africa, Argentina, Brazil, and the United States. Average yields in the United States frequently exceed 2.7 metric tons per hectare, while yields in Argentina and China often average more than 2 tons per hectare (USDA, 1994). In the United States, production levels are controlled by the government under a price support and quota system. The system allows only certain levels of production to occur, but guarantees a minimum price to the grower. The economics of this price support has made it profitable for growers to adopt production systems that attempt to maximize yield per unit land area through high levels of management and input. In the last few years, the price support system for peanuts in the United States has been under increased scrutiny by government officials and the agricultural industry. Should major changes take place in this system, corresponding changes in production practices in the United States would need to occur. Because the U.S. peanut price is higher than the world price, lower cost production systems would need to be developed for domestic production to be cost-competitive with imported peanuts. Cultivar development would need to address this change in production, and the effects on peanut cultivar development are discussed in this review. Smartt (1994) gave a global view of production practices and noted that they vary considerably. In the United States, Australia, and portions of South America, the crop is grown with intense management, generally with high levels of mechanical and chemical input. The crop is grown in mixtures with other species, mainly to provide food and cooking oil for the farmer, in parts of Africa and Southeast Asia. In many countries the crop is grown in monoculture as a cash crop, primarily for export. The intensity of management varies considerably around the world, depending on the economic return for the crop or the role of peanuts in farm subsistence.

PEANUT BREEDING AND GENETICS

397

Over one-half of the peanuts grown in the United States are used for food, primarily as peanut butter, roasted peanuts, or in candies.

D. BREEDING The procedures and objectives used in breeding improved cultivars of peanut will depend on the use of the crop, whether for oil or food, and on the amount of inputs used in production. As with all domesticated plants, peanuts are not grown under the same conditions they experience in the wild. Very low levels of pod production will take place unless pests are controlled and the physical environment is manipulated to provide adequate conditions for peanuts to express their genetic potential for pod yield. When peanut market prices are high, such as in the United States, pest control and environmental manipulation can include significant amounts of purchased inputs. The most effective cultivars for these conditions are those that have high yield potential when stresses are managed. The development of pest resistance and tolerance to abiotic stresses such as drought and temperature extremes is not essential in these systems. In fact, genetic tolerance to these stresses frequently includes a reduced yield potential, whether the stress is present or absent. When grown as a subsistence crop by low-resource farmers or where market prices are lower, many measures to manipulate stresses are no longer costeffective. In these conditions, the crop itself must be tolerant of biotic and abiotic stresses if adequate yields are to be obtained. Breeding has the potential to provide cultivars with these tolerances. A consideration with increasing significance is the need of the consumer. In Europe and the United States in particular, peanut consumers are concerned about food quality and want to be certain that their food contains the minimum amount of pesticide residues possible and the minimum amount of naturally occurring toxins. They also are concerned about the effect of peanut production practices on the quality of the environment. These consumer demands require the development of cultivars with higher levels of biotic and abiotic stress tolerance to reduce the risk of food and environmental contamination.

III. GENETIC VARIABILITY

A. NATURAL Successful cultivar development requires genetic variability for exploitation. Systematic germ plasm collections, particularly of A. hypogaeu, have been made by several national organizations and by the International Center for Research in

3 98

D. A. K N A m AND J. C. WYNNE

the Semi-Arid Tropics (ICRISAT) in Hyderabad, India. Two major collections exist in the United States under the auspices of the United States Department of Agriculture (USDA). The first is at the National Seed Storage Laboratory in Ft. Collins, CO. This is a long-term storage facility meant to safeguard the collection. A second collection exists at the Southern Regional Plant Introduction Station in Griffin, GA. The material at Griffin is the working collection of Arachis germ plasm designed for use by peanut scientists. Nearly 8000 accessions of A. hypogaea are present at this facility. The purpose of the regional center is not only to provide safe storage of seeds for researchers but to grow materials to ensure adequate supplies of high-quality seeds, to identify needs for new material and to coordinate its acquisition, and to characterize the collection for more efficient use by researchers. A crop advisory committee, composed of peanut scientists from a range of disciplines, provides input into the management of this collection. As accessions within collections are characterized, the data gathered are compiled into catalogs. This information can be used by scientists, and plant introductions containing desired characteristics may be obtained for research purposes or cultivar development. Development of the Germplasm Resources Information Network (GRIN), a database of descriptor information for each plant introduction in the USDA system, has made it more efficient to access information regarding this collection. A version of GRIN is available for use on microcomputer (USDA, 1992). The entire 8000-accession collection can be unwieldy for researchers starting to assess variability. An efficient method of subsampling the peanut germ plasm collection has been developed through the use of a core subset. Holbrook et al. (1993) stratified the accessions in the collection by country of origin and available morphological data. After multivariate analyses, accessions were clustered into groups and 10% of the groups were randomly sampled. Analysis of this subset has been made, and on the basis of six phenotypic traits, Holbrook et al. (1993) concluded that the genetic variation expressed for each trait in the entire collection has been preserved in the core subset. Late leafspot-resistant lines were identified in the core subset, and through sampling only 10% of the collection, 54% of the known resistant accessions were found (Holbrook et al., 1993). The core collection has also been used to screen for variations in oil content (J. Bruniard, C. C. Holbrook, and D. A. Knauft, 1993, unpublished data), white mold resistance (Sclerotiurn rolfsii), and CBR (Cylindrocludiurn black rot) resistance (T. G. Isleib, et al., 1994, unpublished data). The peanut collection at ICRISAT is the largest in the world, containing over 12,000 accessions. Researchers with the USDA and at ICRISAT have exchanged information on their respective collections, and it has been estimated that approximately half the collection at ICRISAT is not present in the USDA collection (R. N. Pittman, 1994, personal communication). The ICRISAT collection is

PEANUT BREEDING AND GENETICS

399

intended both to give adequate storage conditions for the collection in a tropical environment and to act as a source of material for breeding, genetic, and other scientific studies with peanut, particularly in the developing countries. Other germ plasm collections exist in countries including Argentina, Australia, China, Israel, Malawi, Niger, Nigeria, Senegal, South Africa, Taiwan, Thailand, and Zimbabwe. The germ plasm flow between countries at times is unhindered, yet phytosanitary regulations and political considerations frequently delay or prevent movement across political borders. The relative importance of national peanut germ plasm collections relative to other economic considerations also varies among countries and through time. Germ plasm collection and exchange are expected to undergo major changes since the United Nations Convention on Biological Diversity went into effect December 29, 1993. This agreement recognizes the control nations have over their genetic resources and places rights to these resources subject to the primacy of intellectual property. Agreements are to be developed between countries that are the source of germ plasm and the organization that uses the material. Before this convention, recent peanut germ plasm collection trips of U.S. origin operated from the basis that the collected germ plasm was a shared resource. Each accession was collected in sufficient quantity to provide the genetic resource to the host country and the United States. In this manner, both the country of origin and the collecting country were provided with identical material, and the host country was provided resources to supplement its own germ plasm collection efforts. The new United Nations Convention creates a different working environment as it attempts to provide appropriate compensation for the natural resources (in the form of germ plasm) of a source country. However, the convention provides for continued compensation based on the value of the product derived from the initial resource, which is unlike other natural resource compensation for such commodities as minerals and foodstuffs. Beyond the difference from other natural resources, dramatic changes will occur in the way germ plasm compensation occurs. While designed to provide developing countries with some compensation for their natural resources, the convention could have the opposite effect of reducing the flow of germ plasm among countries. While germ plasm has limited intrinsic value, its realized value comes from breeding efforts used in the development of cultivars. Until recently, most peanut germ plasm has been freely exchanged, including released cultivars coming from countries that traditionally collect germ plasm back to source countries that have lacked the resources to develop their own cultivars. In the past, the new cultivars have been used essentially free of charge. The new convention calls for separate agreements made between individual countries of origin and utilization. While the convention was designed to allow compensation to countries with indigenous genetic resources, it could have the negative effect of making the return of the value-added product

400

D. A. K N A U n AND J. C.WYhWE

of these resources, the advanced breeding lines and cultivars, prohibitively expensive. ICRISAT and other organizations with an international scope are working to develop protocols for peanut germ plasm exchange among nations based on this convention. Besides the systematic collections available through ICRISAT, the USDA, and other national collections, individual breeders maintain materials within their own program resources. Large collections exist in the breeding programs of Texas A&M University (College Station and Stephenville), Oklahoma State University, North Carolina State University, the University of Georgia (both state and USDA), the University of Florida (Marianna), and AgraTech Seeds, Inc. (Ashburn, GA). Lines in these breeding programs include noncommercializable breeding lines, genetic stocks, materials from genetic studies, and germ plasm developed to incorporate tolerance to various biotic and abiotic stresses. One concern in the industry is the loss of desired germ plasm when programs are closed down. Breeding programs at New Mexico State University, the University of Florida (Gainesville), and the USDA program in Virginia have closed in the last few years. While the material closest to commercialization from these programs may be exploited by other programs, limited additional resources prevent substantial evaluation of the material in these programs that is not destined for immediate release. The genetic variability in the peanut is extensive. Much of this variability has been described by Murthy and Reddy (1993). In addition, Isleib et al. (1994) have listed sources of variability for resistance to pest problems of significance worldwide. Despite the extensive research reported in these works, the knowledge of peanut genetics lags behind that of other major crops. For example, only eight linkage groups are known in cultivated peanut (Murthy and Reddy, 1993). Most linkages are not close, limiting their usefulness in any breeding efforts. The closest linkages noted so far are a 10% crossing over between genes for variegated testa color and nonnodulation (Dashiell, 1984) and 7.8% recombination between orange corolla color and purple testa (Knauft et al., 1991). In both cases the traits involved have negative economic value. Molecular markers have been used in other crop species to provide linkage maps that can be of utility in genetic and breeding studies. Little variation has been found at either the molecular or nucleic acid level in cultivated peanut. Several workers have found little isozyme variability in cultivated species (Grieshammer and Wynne, 1990a; Lacks and Stalker, 1993). A similar lack of variability was found when restriction fragment length polymorphisms were studied by using genomic probes (Kochert er al., 1991; Ro et al., 1992) and single primers of arbitrary sequence (Halward et al., 1991a, 1992). More extensive molecular variability has been identified in wild species, particularly in the progeny from interspecific hybridization. A high proportion of polymorphism was found between the A. hypogaea cultivar TMV-2 and a synthetic amphidiploid from A. batizocoi X Arachis chacoense (Lanham er al., 1992).

PEANUT BREEDING AND GENETICS

401

Progeny from the cross of Arachis stenosperma X Arachis cardenasii have been used to develop a linkage map based on differences in DNA sequences (Halward et al., 1993). Although this map may have some utility in tracing the introgression of genes from these wild species into cultivated peanut, there is uncertainty as to the significance of this map for breeding efforts in cultivated peanut. The use of linkage maps will be greatly aided by the identification of specific chromosomes in peanut. However, little work has been done to identify aneuploid series that could be used for such work. Aneuploid studies have been reviewed by Murthy and Reddy (1993). Aneuploids have been noted to occur either spontaneously or through interspecific hybridization. Most aneuploids are reported to have unstable chromosome numbers and reduced vigor. The monosomic and nullisomics identified have been highly sterile, suggesting the limited utility of haploidy through anther culture or other techniques. Extensive genetic variability exists in species related to A . hypogaea (Singh and Simpson, 1994). The USDA and ICRISAT germ plasm collections contain accessions of many of these species, which include both annuals and perennials. Many of the accessions produce limited seed and are propagated vegetatively. Vegetative propagation is also used to maintain many of the wild species, and the large cost limits the number of accessions that can be maintained in germ plasm collections. Accessions of wild Arachis species have been found with greater variations than cultivated peanut for many traits. Singh and Simpson (1994) provided a synopsis of research identifying wild species with high levels of resistance to early and late leafspot, rust, Cylindrocladium black rot, groundnut rosette virus, tomato spotted wilt virus, peanut stunt virus, northern rootknot nematode, and peanut rootknot nematode. They also reported work identifying wild species accessions with resistance to insects, including thrips, leafhoppers, corn earworm, armyworm, lesser cornstalk borers, and spider mites. Some screening of nutritional quality has also taken place, with high tryptophan content, high oil, and high linolenic acid reported (Singh and Simpson, 1994). Although the genetic variability within wild species includes many important characteristics, it has been difficult to incorporate this variability into A . hypogaea. Most Arachis species are either cross incompatible or produce hybrid seeds with low frequency. In addition, interspecific hybrids that contain desired genetic information from a wild species also contain many undesirable traits that are difficult to remove while maintaining the desired trait. Tissue culture has been used to introgress germ plasm from other species. Hybridization of a number of species combinations can take place, with subsequent abortion of the embryo. Embryo culture has been used to rescue hybrids from several different interspecific hybrids. Success has depended on many factors, including the specific genotypic combination (Ozias-Akins et al., 1992c) and the growth stage of the culture. A quiescent stage of embryo development has been identified. Once this stage is reached, further differentiation in some

402

D. A. K N A W AND J. C . WYNNE

hybrids is not possible (Pattee el al., 1988). Some researchers have been able to rescue embryos by avoiding events partially correlated with quiescence (Moss and Stalker, 1987; Halward and Stalker, 1987a,b; Pattee and Stalker, 1992). It may be possible to change medium or culture techniques or explant tissue and improve the success of embryo rescue, although there are few reports of the efficacy of these procedures in peanut. Tissue culture also may be used to produce haploids, through either anther or ovary culture. Doubling of these haploids could produce plants to bridge introgression of wild germ plasm and could serve as a source for homozygous breeding lines. To this point there has been limited success in the culture of anthers or ovaries in peanut. Further uses of tissue culture include the study of metabolic processes in suspension cultures and in vitro plant regeneration for somatic hybridization, somaclonal variation, and genetic transformation. The tissue culture literature on peanut has been reviewed (Murthy and Reddy, 1993; Knauft and Ozias-Akins, 1995). Much of the work has focused on shoot and root formation that have relied on the intrinsic competency of the cultured explant tissues. De novo shoot formation from totally undifferentiated tissues of peanut is still difficult. Other Arachis species have shown a greater capacity for regeneration from mature tissues than A . hypogaea. Shoots have been obtained from cultured expanded leaves of Arachis villosulicarpa (Hoehne) (Dunbar and Pittman, 1992) as have roots (Pittman et al., 1984). Suspension cultures from mature leaf callus of Arachis paraguariensis (Chodat & Hassler) were totipotent for at least 2 months (Li et al., 1993),as were suspension cultures initiated from anthers of the same species (Still et al., 1987).A . paraguariensis has also been used to regenerate plants from suspension cultures of protoplasts of this species (Li et al., 1993). Most tissue culture work in peanut relies on plant regeneration through embryogenesis or organogenesis for studying metabolism and development or generating genetic variations. In the peanut, long-term, organized cultures can be maintained (McKently et al., 1990; Durham and Parrott, 1992; Ozias-Akins er al., 1992b). Prolific organogenesis can be obtained from mature seed parts, (McKently et al., 1990; Daimon and Mii, 1991), immature cotyledons (OziasAkins, 1989), immature embryo axes (Hazra et al., 1989), whole immature embryos (Sellars et al., 1990), mature embryo axes (McKently, 1991), or young leaves (Baker and Wetzstein, 1992; Chengalrayan et al., 1994). Some genetic differences in the frequency and magnitude of embryogenesishave been reported (Sellars et al., 1990; Ozias-Akins et al., 1992a).

B. GENETICTRANSFORMATION One system of DNA delivery for peanut genetic transformation includes infection by wild-type strains of Agrobacterium rhizogenes (Mugnier, 1988) and

PEANUT BREEDING AND GENETICS

403

Agrobacterium tumefaciens (Lacorte et al., 1991; Mansur et al., 1993). Cocultivation of peanut hypocotyl callus with A. tumefaciens has resulted in kanamycin-resistant, P-glucuronidase (gus)-positive callus (Franklin et al., 1993). The same techniques also were used to transform callus with a functional peanut stripe virus coat protein gene (Franklin et al., 1993). Plant regeneration has not been achieved from these Agrobacterium-transformed tissues. Preliminary evidence suggests that peanut embryo axes cocultivated with Agrobacterium have been stably transformed (McKently et al., 1993). Direct DNA delivery with microprojectile bombardment is an alternative to Agrubacterium-mediated transformation. Stably transformed callus lines (Clemente et al., 1992), regenerable embryogenic cultures, (Ozias-Akins et al., 1992b, 1993), and regeneration of transgenic peanut plants from these cultures (OziasAkins et al., 1993) have been reported. Particle bombardment was used to transform the cultivars Horunner and Florigiant with the tomato spotted wilt virus-nucleocapsid protein gene and the gene for phosphinothricin acetyl transferase (Brar er al., 1994). The latter gene was shown to provide resistance to the herbicide Basta.

IV. GENETICS A. QUALITATIVE Genetic studies have been conducted to determine the inheritance of many traits in the peanut, particularly those with economic value. An exhaustive review of peanut genetics and cytogenetics has been published by Murthy and Reddy (1993). The following discussion will be limited to research since 1993 on the inheritance of traits valuable to breeding efforts.

1. Vegetative Traits Peanut growth habit varies from upright to prostrate. The growth habit nomenclature varies in the literature and in the lists of descriptors (ICRISAT, 1992), and it can be difficult to distinguish between categories. Despite these difficulties, many researchers have identified a simple inheritance pattern for growth habit. The pattern has varied depending on parental materials used. Monogenic (Varman et al., 1986), digenic (Coffelt, 1974; Varman et al., 1986), trigenic (Ashri and Levy, 1978), and tetragenic (Essomba et ul., 1987) inheritance has been reported, with prostrate dominant to upright. Dwarf plants frequently occur in breeding programs (Gupta, 1988). In general, dwarfs are caused by a single recessive gene (Patil and Mouli, 1975; Radha-

404

D. A. KNAUET AND J. C . WYNNE

krishnan et al., 1991; Tiwari and Khanorkar, 1984; Branch and Hammons, 1983), with an intermediate type identified by Branch and Hammons (1983) as being incompletely dominant. Often, the dwarf mutants are accompanied by complete sterility. One of the largest studies of this type involved 10 crosses with 13 genotypes (Gupta, 1988). The dwarf, sterile types were controlled by either two, three, or four genes. Peanut can be colonized by Bradyrhizobium from the cowpea inoculation group to form nitrogen-fixing nodules. Several nonnodulating lines of peanut have been identified, and although the lack of nodulation is genetically controlled, the trait is complexly inherited. Gorbet and Burton (1979) reported nonnodulating types, and on the basis of segregation in advanced generations from initial crosses, they found that nonnodulation was not simply inherited. Nigam et al. (1980a,b, 1982) and Dashiell and Gorbet (1982) found that nonnodulation was controlled by two recessive genes. Dutta and Reddy (1988) found a trigenic inheritance pattern for the trait, which could be explained if crosses from earlier studies did not differ for one of the three genes involved. Inheritance patterns are more difficult to study with this trait because the intensity of nodulation can be affected by genetics of the host, the Bradyrhizobium strain, and environmental conditions. Many leaflet shapes have been identified in the peanut. They include a monogenic, dominant, crinkled leaflet (Hammons, 1964), a digenic, recessive, drooped leaflet called flop (Branch and Hammons, 1981), a narrow leaflet that has been determined to be partially dominant in some studies (Matlock et al., 1970) and monogenic dominant in others (Tiwari et al., 1990), a monogenic, recessive leaflet called cup (Hammons, 1953). leaflets with ribbed parallel lateral veins called corduroy (Loesch and Hammons, 1968), and a partially crinkled leaflet with yellow margins designated as puckered (Dwivedi and Nigam, 1989). Many leaflet mutants have been produced through induced mutations or identified in the progeny from wide crosses. The inheritance of leaflet size itself has had limited study, although it would appear to be quantitatively inherited (Essomba et al., 1993). Mutations for leaflet pigmentation have been found by many scientists [reviewed by Murthy and Reddy (1993)] to be controlled by one or two recessive genes. Branch and Kvien (1992) showed that variegated leaflets in peanut were cytoplasmically controlled. 2. Reproductive Traits

The peanut flower is a typical legume flower with standard, wings, and keel. Flowers of most cultivars are yellow, with darker orange lines radiating from the base of the standard. Scientists have identified many variations in flower color and pattern of lines on the standard. Differences in color terminology among

PEANUT BREEDING AND GENETICS

405

researchers has made it difficult to make comparisons among studies. However, most variants are simply inherited. White has been found to be controlled by duplicate recessive genes [Patil(1965), cited in Murthy and Reddy (1993)], light yellow is monogenic recessive (Patil and Mouli, 1984), lemon yellow is dominant to orange and is controlled by two genes (Banks and Pittman, 1986), deep orange is dominant to orange and is controlled in an epistatic manner (Radhakrishnan et af., 1991). and deep orange is recessive to normal yellow (Knauft et af., 1991). Many pod and seed characteristics of peanuts are quantitatively inherited. One major exception is testa color, which is probably the most studied characteristic in peanut. Two sets of duplicate genes (designated F1F2 and DlD2) are responsible for the basic presence or absence of pigmentation in the testa. A dominant allele of one of the F and one of the D genes must be present to condition pigmentation. If either the F or the D gene series is recessive, white testa results. When the F and D genes are present to allow pigmentation to occur, then several colors are possible, including red, pink, purple, tan, or variegated. Many colors are controlled by duplicate loci, as expected for the tetraploid species. Frequently additional genes are found. Norden et af. (1988) and Branch (1989) each identified white testa controlled by a single dominant gene. Other work showed that the two sources [Pearl Early Runner by Norden er af. (1988) and PI 408735 by Branch (1989)l contained different dominant genes (Knauft er af., 1991). Two genes for red testa color are insufficient to account for the results obtained in crosses, and a third gene has been identified (Holbrook and Branch, 1989; Branch and Holbrook, 1988). A complete summary of testa color inheritance has been published by Murthy and Reddy (1993). Peanut seed composition is generally observed to be quantitatively inherited. However, duplicate recessive genes are responsible for increasing the oleic acid composition in peanuts by approximately 25%. The mutant causes a corresponding reduction in linoleic acid content to near 2% (Moore and Knauft, 1989). One of the genes appears to be common in peanut germ plasm, effectively causing segregation from high oleic acid by normal oleic acid crosses to be monogenic (Knauft et af.,1993b). Isozymes have been identified for three enzymes in peanut: phosphohexose isomerase, isocitrate dehydrogenase, and glutamate oxaloacetate transaminase (Grieshammer and Wynne, 1990b).

B. QUANTITATIVE Many economically important traits in peanut are quantitatively inherited (Murthy and Reddy, 1993). Specific characteristics associated with yield or stability of production will be discussed in that section. However, yield, yield components, and characteristics associated with market quality have been stud-

406

D.A. K N A W AND J. C. wY”E

ied. These traits have generally been found to have low narrow-sense estimates of heritability.

V. BREEDING A. GENERAL Under most conditions, the peanut is highly self-pollinated (Knaufi et al., 1992), and breeding procedures in use for cultivar development are those generally used for self-pollinated crops. Much of the work on breeding techniques has been summarized previously (Knauft er al.. 1987; Reddy, 1988; Coffelt, 1989b; Nigam et al., 1991; Isleib er al., 1994; Knauft and Ozias-Akins, 1995). The choice of parents is critical to the overall success of any breeding program. The U.S. market tolerates only minimal variations in pod and seed size, pod appearance, testa color and seed shape, seed composition, processing traits, and flavor. Because of these narrow limitations, and the difficulty of combining these traits in a single genetic entity, parents are generally chosen that already combine as many of these traits as possible. At one time, only three cultivars dominated U.S.peanut production. These cultivars were Florunner (U.S. runner market type, A. hypogaea subsp. hypogaea var. hypogaea), Florigiant (U.S. Virginia market type, A. hypogaea subsp. hypogaeu var. hypogaea), and StarrKomet (US.Spanish market type, A. hypogaea subsp. fasrigiara var. vulgaris; Comet was a selection from Starr). Florunner was grown on virtually the entire runner market production area, Starr and Comet were produced on 82.5% of the area devoted to Spanish production (Hammons, 1976), and over 80% of the Virginia area was devoted to Florigiant (T. A. Coffelt, personal communication, 1993). These cultivars were the predominant types for the manufacture of food products, and both consumers and the food industry came to rely on these types as standards. New cultivars needed sufficient similarity to existing cultivars so that processing of any single cultivar or any cultivar combination would produce the same quality product. This need for uniformity within the peanut industry led to the use of these dominant cultivars as parents in peanut breeding programs to maximize the probability that new cultivars would contain similar processing and consumer properties. These properties are complexly inherited, and the probability of recovering all the properties in the progeny of crosses between parents without this combination is unlikely. Many cultivars released after the dominance of these cultivars in the early 1970s have been closely related to the three original cultivars (Knauft and Gorbet, 1989; Isleib and Wynne, 1992). The pedigrees of runner market types have had a more narrow genetic base than those of the

PEANUT BREEDING AND GENETICS

407

Virginia types, and the four ancestral lines from the background of Florunner (Small White Spanish, Basse, Spanish 18-38, and Dixie Giant) have made up much of the ancestry of runner peanut cultivar acreage (Isleib and Wynne, 1992). As public and private breeding programs in the southeastern United States have matured, more runner cultivars have been released. The runner cultivars released since the late 1980s, including Marc I, Andru 93, AT 127, AT 108, Georgia Runner, and Georgia Browne, have a broader genetic base than earlier lines. Although the backgrounds of Virginia and Spanish cultivars are more complex than that of runner cultivars, their background is still narrow, reflecting the market demands for uniformity (Isleib and Wynne, 1992). Many techniques have been used to improve the efficiency of peanut breeding. Once breeding objectives are identified, determination of the parents most capable of producing progeny with the desired objectives is important. Techniques for the identification of appropriate parental lines have included use of diallel analysis (Holbrook, 1990) and test crosses (Isleib and Wynne, 1983b). Both techniques were found to identify desirable parents on the basis of parental performance per se. However, these studies included highly adapted and unadapted materials and may have limited utility in programs that either have a need to choose among adapted lines as parents or must include unadapted parents to obtain desired characteristics. A study using line X tester analysis came to the opposite conclusion, finding that lines with high mean pod yield had low general combining ability (Upadhyaya et al., 1992). A study used to choose diverse parents for a breeding program found that cluster analysis could identify appropriate cross combinations (Durga Prasad et al., 1985). Several studies have examined the performance of F, hybrids as indicators of parental potential (Isleib and Wynne, 1983a,b; Arunachalam et al., 1984a; Holbrook, 1990). These studies have had mixed results and usually would be impractical for breeding programs, given the difficulty of generating sufficient F, seed for testing purposes. Often, reciprocal cross differences in peanuts have been shown to affect the expression of characteristics. These differences have been observed for yield and other agronomic traits (Wynne and Emery, 1974; Dwivedi et al.. 1989), disease resistance (Kornegay et al., 1980; Coffelt and Porter, 1982), and nitrogen fixation (Phillips et al., 1989; Dwivedi et al., 1989). While many reciprocal cross differences have come in intersubspecific crosses (Wynne and Halward, 1989), this has not always been the case. Care must be taken to distinguish maternal effects from true reciprocal cross differences. Maternal effects have been noted for seed dormancy (Khalfaoui, 1991) and fatty acid composition (Mercer et al.. 1990) and may have been important in studies examining intersubspecific crosses for yield and other agronomic traits where only F, progeny were examined (Reddy et al., 1988). Reciprocal cross differences have been routinely observed in crosses used in

408

D. A. KNAUFT AND J. C. WYNNE

cultivar development in several programs. In breeding efforts at the former programs of the University of Florida in Gainesville and the USDA-ARS program in Virginia, it had been frequently noted that superior progeny were obtained from crosses made in one direction, while no material was maintained from crosses in the opposite direction. Unfortunately, no correlations were noted that could be used to detect the appropriate direction for crossing. While the inclusion of reciprocal crosses in breeding programs is common, the frequency with which cross differences take place must be balanced with the reduced numbers of parental combinations that can be made. Although it is commonly assumed that reciprocal cross differences are the result of cytoplasmic inheritance of organelle DNA, examples of paternal inheritance have been reported in peanuts, including the inheritance of isozyme variation (Grieshammer and Wynne, 1990b) and a low oil, shriveled seed mutation (Jakkula et al., 1993).

B. HYBRIDIZATION The first report of artificial hybridization was made by van der Stok [1910, cited in Coffelt (1989a)l. Procedures have been summarized by Norden (1980) and Nigam er al. (1990). Self-pollinationis prevented by removing anthers in the late afternoon or evening before maturation. Pollinations are made the following morning when stigmatic tissue is receptive and pollen is viable. Various methods of tagging are used to identify hybrid pods, and techniques have been developed to increase the probability of successful fertilization. The procedure is tedious and labor intensive and rarely results in more than three seeds for each pollination. The low yield from artificial hybridization limits the amount of hybrid seeds that can be obtained. Therefore, breeding methods requiring large amounts of seeds, including diallel selective mating and recurrent selection, are not commonly used. Few seeds from artificial hybridization rarely is the limiting factor in cultivar development programs. Rather, limitations are usually the cost and time involved in the evaluation of breeding lines (Knauft et ul., 1987). Out-crossing rates in peanut are generally low. Breeders and the seed industry have designed plot size and plot placement for generation advancement and seed increase of breeding lines, as well as cultivars, with the assumption of very low out-crossing rates. However, rates have varied in Virginia from 0 to 2.8% (Coffelt, 1989a) and in Florida from 1 to 8.1% (Knauft er ul., 1992), depending on the year and the genotype. The rates in Florida occurred despite weekly applications of insecticides to control Eemisia tubuci (silverleaf whitefly). The highest out-crossing rates occurred in a line that appeared to have delayed dehiscence, although normal floral biology was found in lines with over 3% out-crossing.

PEANUT BREEDING AND GENETICS

409

C. EARLY GENERATION TESTING The use of early generation testing to identify superior crosses and eliminate large amounts of material from a cultivar development program may increase breeding efficiency. However, there does not seem to be a clear advantage to this technique in peanut breeding. While Iroume and Knauft (1987a) predicted that such testing could be used to select crosses with superior yield and leafspot resistance, Halward et af. (1990) found that early generation testing for pod yield was not effective in a recurrent selection program. Both groups of researchers indicated that traits with higher heritabilities could be selected in early generations, but that other breeding procedures could be used to obtain the same results.

D. BREEDINGMETHODS 1. Mass Selection

Although there are many instances when mass selection is inefficient, its low cost contributes to its continued use in breeding programs. Holley and Wynne (1986) noted that it could be used to increase pod yield when intersubspecific crosses were made between adapted and unadapted germ plasm. They found less utility in intrasubspecific crosses and suggested that this may have been caused by negative correlations between seed yield and meat content. Negative correlations between pod yield and leafspot resistance were proposed as a cause for the failure of mass selection techniques as a tool for the identification of leafspotresistant lines with high pod yield (Knauft et af., 1993a). Mass selection for pod yield/plant was used in the F,-F, generations of a series of crosses, and mean pod yield, plant height, and harvest index were improved in the F, lines (Patra er af., 1992b). Here again a negative relationship existed between desired traits, including an association of pod yield with lower shelling percentage and smaller seed size. Another concern in mass selection is the genetic shifts that take place in the segregating populations. Although more competitive genotypes have been found to increase in frequency in mixed populations, these genotypes were not the highest yielding pure lines, suggesting a need to perform mass selection for relatively few generations to reduce the effects of this competition (Knauft and Gorbet, 1991). 2. Pedigree Method

Comparisons among the various methods for selection in early generations and advancement to near-homozygous lines have not been made in peanuts. Pedigree

410

D. A. KNAUFT AND J. C.WYNNE

selection is commonly used in breeding programs, and several studies have been made comparing the use of selection in different environments. When separate selections were made in Georgia and Zimbabwe from the same segregating population, lines were adapted only to the location where selections were made, and significant genotype x environment interaction was found (Branch and Hildebrand, 1989). A sequential method of selection at one location, followed by additional selection at a second and then a third location, was evaluated by Branch et al. (1991). Selections for leafspot resistance and pod yield made in sequence were generally equal to or better than pedigree selections from a single location. Unselected bulks developed with the single seed descent method generally performed as well as or better than the sequential selections and required considerably less cooperation and cost.

3. Back-Cross Method The inefficiency of hybrid seed production and the small number of simply inherited traits of economic importance limit the use of the back-cross method in peanut cultivar development. One of the few uses has been the successful incorporation of the high oleic acid-low linoleic acid trait into agronomically acceptable backgrounds (Moore and Knauft, 1989; Knauft er al., 1993b). Of further advantage in this instance were the limited undesirable linkages (Knauft et al., 1993b) and a rapid technique for the determination of fatty acid composition from as little as 15 mg of tissue from individual seeds (Zeile et al., 1993).

4. Single Seed Descent Method Single seed descent procedures are used infrequently by peanut breeders. Isleib et al. (1994) suggested that the spacings required to allow individual plant identification may be sufficiently great that the procedure has not attracted much attention. However, within-row plant spacings 6 times greater than that in normal breeding nurseries have been used in Florida with easy identification of individual plants at harvest, even when populations are segregating for plant type. Several economically important traits in peanut, including plant and pod characteristics, have high heritabilities. Segregating populations from crosses, particularly between subspecies, may contain sufficient variability that breeders have chosen to maintain the pedigree method for selection of these traits. Although little work has been done on the peanut to compare breeding methods, several researchers have utilized or recommended single seed descent procedures. Branch et al. (1991) found that unselected bulks derived from single seed descent performed as well as selections obtained from a sequential selection scheme at several locations. Anderson er al. (19Wb) recommended single seed descent for diallel crosses to obtain multiple foliar pest resistance after they

PEANUT BREEDING AND GENETICS

41 1

found that early generation assessment had limited usefulness because of environmental variations.

5. Recurrent Selection Recurrent selection has also had limited use in peanut breeding programs, most likely because of the difficulty of generating large numbers of seeds from artificial hybridization. The procedure has been used extensively in North Carolina (Wynne, 1976; Monteverde-Penso et al., 1987). Forty elite Virginia markettype peanut lines were randomly crossed to produce a population for study. After three cycles of recurrent selection, consistent gains in pod yield had been noted that were not correlated with any specific fruit traits (Monteverde-Penso and Wynne, 1988). The genetic variation for pod yield in this population decreased after each of four cycles of selection, although more lines were generated in later cycles that had pod yields greater than or equal to that of the check cultivar (Halward and Wynne, 1992). Further evaluation of the fourth cycle of selection showed that average improvement for most economically important traits had not progressed after the third cycle (Halward and Wynne, 1992). They found that significant genetic variability remained for all measured traits and felt the lack of progress may have been due to a low selection intensity (40%). A separate study using an interspecific hybrid between A. hypogaea and A. cardenasii Krap. et Greg. found that pod yield and market grade characteristics had been improved after two cycles of selection (Guok ef al., 1986). Significant genetic variability existed in this population after the two cycles of selection for morphological traits and leafspot resistance (Halward et al., 1991b).

6. Complex Crosses Complex crosses have been proposed as another method for obtaining desirable recombinants in peanut and would require less crossing than recurrent selection techniques. Arunachalam ef al. (1985) used three-way crosses to broaden the genetic base of peanut germ plasm. In the F, progeny from such crosses, the authors found plants that transgressed the better parent for yield components up to 370%. They also found that the general combining ability of early growth stage characteristics could be used to identify parents for complex crosses. Bandyopadhyay et al. (1985) identified three-way crosses as being superior to single crosses for improving physiological and yield components. Branch and Holbrook (1991) used a convergent hybridization scheme of crossing progressively more F, hybrids within cycles to develop a population with genetic contributions from 16 parental lines, with varying types of pest resistance. Knauft and Iroume (unpublished data, 1987) compared F5 lines obtained from pedigree selection of single and double crosses. The double crosses were made to

412

D. A. KNAUFT AND J. C.WYNNE

examine the utility of that procedure for incorporating multiple sources of resistance to late leafspot. However, no differences in pod yield or resistance were found. Other researchers have found that early generation intermating could be used to produce progeny superior for pod yield and yield components (Dutta et af., 1986). Selection in the early generation is made more difficult by the need to identify plants with a combination of desirable traits. Although index selection procedures exist that can be used in peanuts, little work has been done in this area. Bandyopadhyay et al. (1985) selected plants on the basis of an index that included physiological components. They found that it was more desirable than an index based solely on yield components, although they did not compare their results to those from direct selection. An index for simultaneousselection for late leafspot resistance and pod yield was developed by Iroume and Knauft (1987b). Anderson et al. (1990a) recommended the use of principal component analysis to examine components of multiple pest resistance in a 10-parent diallele. Unlike many other crops where the economic yield of the plant is visible during the growing season, pod yield in the peanut is only observable after plants have been harvested. Researchers have studied the relationship between pod yield and many vegetative characteristics in an attempt to allow breeders the ability to improve selection efficiency by conducting some evaluation before harvest, as well as using traits with higher heritabilities than pod yield. A summary of many correlation, or character association, studies has been published (Reddy, 1988). Nitrogen percentage and leaf area accounted for 75% of the variations in pod yield among 17 genetically diverse lines (Vinod Prabhu et al., 1990). Nigam et al. (1984) found only podlplant, pod weight, and seed/plant could be used effectively for indirect pod yield selection, although they examined 16 vegetative and reproductive traits. Plant canopy diameter at 60 days was found by Nagabhushanam and Prasad (1992) and Nagabhushanam et af. (1992) to be highly correlated to pod yield. Manoharan et af. (1990) found that pod yield was more affected by the environment than pod number, individual pod weight, and dry matter production. They suggested that selection for these traits, which were highly correlated with yield, could improve selection efficiency.

E. GENOTYPE x ENVIRONMENT INTERACTION Many studies have been conducted on the interaction between peanut genotypes and the environment (g X e), including those by Dashiell et al. (1982), Shorter and Hammons (1985), Knauft et al. (1986b), Norden er af. (1986), Anderson et al. (1989), Coffelt et al. (1993), and Raut et af. (1993). In general, these studies have found significant interactions, confirming the need for extensive multiyear and multilocation testing prior to cultivar release. Occasionally,

PEANUT BREEDING AND GENETICS

413

the g X e interactions are the result of magnitude differences among genotypes in environments, rather than significant differences in rank. Knauft and Gorbet (1993) found that shelling percentage and seed size both exhibited highly significant rank correlations when a broad range of genotypes was evaluated in a series of year and location comparisons. They indicated that selection for these traits could be made with minimal testing. Given the significant interaction between genotype and environment, several researchers have examined procedures that could be used in peanut breeding to evaluate the interaction. Procedures identified include pattern analysis (Shorter and Hammons, 1985), moving mean covariance adjustments (Shorter and Butler, 1986), and distance parameters and regression analyses (Anderson et al., 1989). Blending sister lines into a single cultivar has been proposed as a method for increasing genetic diversity and thus providing more stable production in different environments (Norden et al., 1982). Many peanut cultivars, including the popular Florunner and Florigiant cultivars, have been developed through blending of sister lines (Knauft er al., 1987). The component lines making up the peanut cultivars from the Florida program were chosen for inclusion on the basis of phenotypic similarities. It is not known whether the success of these cultivars is the result of genetic diversity from the component lines or from other properties. Because the lines had been combined without prior knowledge of their stability performance, it is interesting to examine studies on mixtures in peanuts. Although Schilling et af. (1983) found that blends from sibling lines with genetic variability were more stable than blends from various pure lines, they also found that pure lines could be identified that were as stable as the multilines. Norden et al. (1986) found that four mixtures and the four component lines making up each mixture had no significant differences in average pod yield or market grade traits. Knauft et al. (1986b) compared four multiline cultivars with two single genotype breeding lines and found that market grade stability was not improved by the multilines. None of the various combinations of two peanut genotypes with disparate growth habits tested by Rattunde et al. (1988) had higher pod yield than the best component line. Knauft and Gorbet (1991) found that the best component line had a greater pod yield and better market grade characteristics than a mixture of five lines. Patra et al. (1992a) obtained improved pod yield when a mixture was compared to the average of the component lines, but individual lines were identified with greater yield and yield stability than the mixture. The set of standard breeding procedures used in most peanut cultivar development programs today is similar to that used for the past 20 years (Norden, 1973; Norden et al., 1982; Knauft et af., 1987; Knauft and Ozias-Akins, 1995). The release of increased numbers of cultivars with desirable traits other than higher yield, such as pest resistance and improved chemical quality, is the result of cooperation among breeders, pathologists, entomologists, physiologists, and food

414

D.A. K N A U n AND J. C . WYNNE

scientists. Many advances are technical in nature, including instrumentation for screening large sample numbers for chemical composition and for data manipulation. Improvements in tissue culture techniques are providing greater access to wild species germ plasm, and molecular genetics techniques are providing greater access to genes outside Arachis.

F. SUSTAINABLE SYSTEMS Worldwide there is increased awareness that agricultural practices must be sustainable. While definitions of sustainability vary, they center on the need for providing a sufficient economic return to allow farmers a continued livelihood from fanning. A second component focuses on the need for fanning practices to sustain environmental quality, so that productive fanning will continue and the quality of the environment for nonagricultural purposes will not be degraded. Plant breeding has an important role in sustainable agriculture. The nature of that role depends on whether current programs are developing cultivars appropriate for sustainable systems. If they are not, either changes must take place in the emphasis placed on breeding objectives, or new breeding procedures must be put in place to achieve the desired goals. Duvick (1993) described the general needs of a sustainable agricultural system and included characteristics needed in cultivars for such a system. These characteristics were increased insect and pathogen resistance and adaptation to relatively unmodified soil and climatic conditions. Francis (1993) described the probability of successful incorporation of these characteristics into cultivars of any crop species on the basis of genetic variability, inheritance, ease of screening, and cost. He considered the likelihood relatively high for development of appropriate insect and pathogen resistance, moderate to low for nematode resistance, and variable for developing competition tolerance to weeds. Moderate success was predicted for development of cultivars with tolerance to drought and problem soils, improved nitrogen and phosphorus use efficiency, and root system modification. Implicit in sustainability is an increase, particularly in peanut, in the management intensity and a reduction in the level of the many inputs traditionally used in peanut fanning. The nature of cultivar development for such a system may be considerably different from current practice. Current breeding selection and testing procedures are carried out in nurseries with high input levels, similar to those used by growers who routinely obtain high yields. Exceptions to these breeding procedures occur when tolerance to specific stresses, either biotic or abiotic, is being selected. It will be important to determine whether sustainable production practices will require cultivars with general tolerance to lower inputs, or whether the develop-

PEANUT BREEDING AND GENETICS

415

ment of resistance to specific stresses will be more important in providing cultivan with consistent pod yields for farmers. It may be difficult to develop cultivars with a general tolerance to lower inputs. Some researchers have found that high input conditions allow greater realization of a genotype’s yield potential and that expression of desired traits for low input systems cannot be identified under high input conditions (Atlin and Frey, 1990). Genotypes identified for productivity in high input conditions are not necessarily suited for productivity in low input situations, and lines truly adaptable to low input conditions frequently must be selected in those environments. Sometimes, yields in poor environments can be related to one or a few factors limiting pod production. When this occurs, certain traits with moderate or high heritability (such as disease resistance or tolerance to specific soil conditions) can be bred into lines and will greatly increase the probability of developing cultivars that are tolerant to such conditions. Knauft and Gorbet (1990) examined this issue in peanut. They grew six genotypes in four production systems representing a range from low to high inputs. They found that genetic variance in an environment was not highly correlated with the productivity of that environment, suggesting that selection could be efficient in lower input environments. They were able to identify a genotype that was consistently productive in lower input environments by using modified stability analysis (Hildebrand, 1984). Although many input factors were changed in the four environments, the productive, stable genotype was characterized by a high level of resistance to late leafspot. The authors suggested that this resistance was likely the major trait responsible for the high yield in lower yielding environments. Kvien er al. (1993) conducted a study using a range of management systems with two cultivars, Florunner and Southern Runner. The latter has a broad range of pest resistance. The systems examined included preventative management, intensely monitored integrated pest management (IPM), IPM with the exclusion of compounds classified by the U.S. Environmental Protection Agency (EPA) as possible carcinogens, the previous IPM program with the exclusion of the more highly leachable compounds, the previous IPM program with the exclusion of both possible carcinogens and more highly leachable compounds, and management practices that are consistent with California certification for organic production. Both production and economic data were gathered. The intensely monitored IPM program provided an additional $180 return per hectare compared to the preventive management program for each cultivar. The elimination of possible carcinogens from the production scheme reduced economic returns significantly. Southern Runner had a $500 higher return per hectare than Florunner under these conditions. As in the Knauft and Gorbet study (1990), leafspot resistance was a major determinant of productivity in the various systems. Kvien and associates indicated that the greatest differences resulted from the elimination of the compound chlorothalonil, which is used for leafspot control.

416

D. A. KNAUFT AND J. C. WYNNE

VI. RESEARCH RELATED TO BREEDING k

PHYSIOLQGY

The development of peanut cultivars with improved pod yield and tolerance to abiotic stresses requires an understanding of the physiological responses that control these traits, the genetic variations within peanut germ plasm for these responses, and the inheritance of the characteristics themselves. The relationship between peanut physiologists and geneticists is such that the first two components of the process have been examined in some detail, but little work has been done on inheritance of physiological differences. Several studies have been conducted to determine the physiological traits contributing to the higher yield of more recently released peanut cultivars. Wells et af. (1991) examined Virginia market-type cultivars released in the United States since 1944 and found the newer, higher yielding cultivars allocated a greater proportion of dry matter to reproductive tissue than older cultivars. They also found that the allocation to reproductive growth occurred earlier in the growth cycle of the newer cultivars. Coffelt et af. (1989) and Seaton er af.(1992) also found that more recently released Virginia market-type cultivars had greater reproductive efficiency, whether measured as harvest index or percentage of flowers resulting in pods. Later releases also had a more spreading growth habit and greater seed and pod weight than earlier releases. It may be possible to increase pod yield potential in future cultivars by improving reproductive efficiency by selecting either for higher pod yield or for greater harvest index. Duncan et af. (1978) noted that the cultivar Early Bunch allocated nearly all photosynthate to pods once the linear stage of pod growth occurred. It is biologically difficult to increase photosynthate partitioning to pods beyond 100%. A complete lack of vegetative growth during pod fill may prevent genotypes from replacing lost vegetative tissue due to environmental stresses and may reduce yield stability. However, Bell et af. (1993b) found that redistribution of assimilates from vegetative dry matter to pods in peanut cultivars could take place during pod fill under less favorable conditions, including falling temperatures, defoliation by pathogens, or water stress. Some cultivars had greater redistribution than others, which may provide breeders with a selection tool for increased yield stability. Remobilization of reserves under conditions of source limitations is likely to be significant in maintaining yields, but may limit response to improved conditions, especially with high partitioning. This may increase environment to environment variation of high partitioning types. Although genotypes with low partitioning to photosynthate may have greater ability to replace leaves from disease or insect damage, Knauft and Gorbet (1990) have shown that it is possible to combine a moderately high partitioning

PEANUT BREEDING AND GENETICS

417

coefficient with disease resistance. Stirling and Black (1991) examined a range of environmental stresses to determine the growth stage most susceptible to stress. They found that the time between initiation of pegs and onset of rapid pod growth was the most sensitive. They indicated that breeding efforts to identify types with pod initiation phases that were less sensitive to adverse environmental conditions could improve yield stability. Bell et al. (1993a) found that variation among genotypes for dry matter production was accounted for by the effects of differing leaf area duration on cumulative, intercepted, photosynthetically active radiation. Their work also found genetic differences in sensitivity to night temperatures. They suggested that cultivars adapted to cool environments may provide opportunity for higher yields. Manipulation of peanut physiology to provide earlier onset of reproductive growth may be disadvantageous when the onset occurs before adequate vegetative growth. The definition of adequate vegetative growth may vary depending on the environment for which cultivars are being developed. In many parts of the world, the development of early maturing types can improve yield or yield stability by allowing the development of mature pods prior to environmental stress, usually drought or disease. Often, the production of mature pods before stress is more important than overall yield. The cultivar Chico is an early maturing line that has been used as a source of earliness in breeding programs, although other lines with more rapid pod development have been identified (N'Doye and Smith, 1992). The genetics of maturity in peanut is not well defined, largely because of the difficulty of defining maturity for an indeterminant, nonsenescent plant (Kvien and Ozias-Akins, 1991). Standard vegetative characteristics used in other crops, such as days to first flowering, are ineffective in identifying peanut lines with early maturity (N'Doye and Smith, 1992). Using the hull-scrape method to define pod maturity (Johnson, 1987), Holbrook er al. (1989) found high broadsense heritability estimates from a cross between parents with large maturity differences. Khalfaoui (1990) found maturity to be conditioned by few genetic factors, but highly influenced by the environment. The measures of maturity used in this study, including time to emergence, time from emergence to flowering, number of flowers, and proportion of mature pods (based on internal pod color), were poorly correlated with each other. N'Doye and Smith (1993) used days to emergence, number of days to lst, 5th, loth, 15th, 20th and 25th flower, number of full-size pods at early harvest, number of mature pods, and percentage of mature pods to measure early maturity in a genetic study with early maturing parents. They found that broad-sense heritability for these traits ranged from 36 to 45% and that general and specific combining abilities were not significant. Nigam et al. (1988) found that the time of seedling emergence to first flowering was conditioned predominantly by additive genetic variance, but the

418

D. A. KNAUFT AND J. C . WYNNE

relationship between first flowering and maturity was not discussed. Mutation breeding for early maturity as defined by yield at early harvest may be possible (Mouli and Kale, 1989). Despite the inherent difficulty of selection for early maturity and inconsistent results from genetic studies, several cultivars have been developed with early maturity. By using phenotypic traits of leaf yellowing, external and internal pod maturity traits, seed size, and shelling percentage, two cultivars that mature approximately 10 days earlier than standard runner market-type cultivars have been developed [Marc I (Gorbet et al., 1992) and Andru '93 (Gorbet and Knauft, 1994)l. Similar procedures were used to develop two Virginia market-type cultivars, VA 81B and VA 93B (Coffelt et al., 1982, 1994). Studies have been conducted with valencia peanuts to determine the utility of selection for early emergence and maturity when grown in the cool environment of Ontario, Canada (Michaels, 1988). Selection of early maturing peanuts based on percentage emergence and sound mature seed yield in this short-season environment was difficult. Maturity based on these parameters was complexly inherited. Another procedure used for the development of early maturing peanut lines is to harvest peanuts after a certain thermal time has passed. Thermal time is calculated on a base temperature above which peanut development occurs and a maximum temperature above which no development takes place. Vasudeva Rao et al. (1992) used a mean base temperature of 10°C with no maximum temperature to calculate thermal time. Rather than using a calendar basis, they harvested peanut lines after a minimum cumulative thermal time had been reached and then evaluated pod yield. The procedure was successful for identifying lines with higher yield at early harvest dates. Peanuts are exposed to a range of temperatures during growth. In some environments, the ability to germinate under a wide range of temperatures would be desirable. Large genetic differences in rate and percentage of germination have been found for a range of temperatures (Mohamed et al., 1988a). In many portions of the world, cool temperatures exist during planting seasons. Genetic differences have been noted for the ability to germinate in cool temperatures, as well as early season growth (Bhagat et al., 1988). Other scientists, working with different germ plasm, were unable to detect significant variations for seedling emergence or early season leaf growth in a range of temperatures (Mohamed et al., 1988b). The objective of much work on earlier maturing peanuts has been to develop genotypes that will escape drought stress. In addition to escape, some genotypes have been shown to have a synchrony of development that allows most pods to be initiated before periods of water deficit. A second approach has been to examine characteristics associated with tolerance to drought, as well as genetic variation for such characteristics.

PEANUT BREEDING AND GENETICS

419

In a study with four genotypes, Wright et al. (1991) identified harvest index as a major factor in improved pod yield in response to drought stress. It is possible that a higher harvest index is associated with a more synchronous development. Chapman et al. (1993a,t) have shown that, in these instances, more pods are initiated prior to water deficit, resulting in higher yields. There are some reported differences in photosynthesis among cultivars (Wright and Bell, 1992) that are related to the amount of carboxylase enzyme present (Nageswara Rao et al., 1994). The differences are thought to be associated with observed genetic variations in water use efficiency (Wright et al., 1994). Genetic differences in leaf folding as a response to stress have been observed, yet the differences are difficult to quantify. Matthews et al. (1988a,b) and Chapman et al. (1993a,b) have observed differences in the extent of leaf folding as a function of water potential. They also noted differences in leaf folding compared to leaf wilting in these conditions. Harris et al. (1988) showed that genotypic response to drought depended on differences in phenology. A tension exists with partitioning efficiency and sensitivity to drought stress, primarily through root growth. Khalfaoui and Havard (1993) found genetic differences in the rate of root extension. Nageswara Rao et al. (1993) and Watterott (1991) showed that the droughtresistant genotype ICGS 86707 had sustained root growth and higher root respiration during drought than other genotypes. Watterott (1991) also showed that genotypes with low partitioning had less stress during drought than high partitioning lines. Genotypes with low partitioning of photosynthate to reproductive tissue were found to sustain root extension during reproductive growth, while genotypes with high partitioning did not (Nageswara Rao et a / . , 1993). The development of low partitioning lines for drought tolerance will have a yield cost, especially when drought does not occur. Genotypes identified by Chavan et al. (1992) to be least affected by drought tended to be lower yielding when drought did not occur. Nageswara Rao ef al. (1989) compared high yield potential, drought-susceptible lines with low yield potential, drought-tolerant lines across seasons. They found, despite the drought susceptibility, that high yield potential lines averaged higher pod yields than the low yielding, droughttolerant types. Bailey and Boisvert (1991) identified risk efficient peanut cultivars that could be relied on to produce yields under a range of drought conditions. They found that, as risk aversion increased, cultivars with high average yields were replaced with those that had lower average yields, but a smaller range and higher minimum yields. In the Sahel, drought tolerance has not been associated with root growth. Greenberg et al. (1 992) have shown that, under dry conditions, the most productive genotypes are those that continued to grow at temperatures above those considered optimal for peanut growth, suggesting that this characteristic may be useful for the development of drought tolerance.

420

D. A. K N A m AND J. C. WYNNE

An additional mechanism identified in some genotypes is the greater ability to recover from drought. Harris et al. (1988) and Nageswara Rao et al. (1989) found that the differences were related to tolerances of the phenological processes at primordia levels, with tolerant types suspending expansion growth in leaf primordia rather than ceasing growth or dying. The peanut has been considered a photoperiod-insensitivespecies. Nevertheless, some response has been observed. Bagnall and King (1991) found that the number of flowers increased under short days in a phytotron study and that temperature affected the time to first flower, with Spanish genotypes having a higher base temperature for flowering than Virginia or valencia types. Bell and Harch (1991) found a slight decrease in flower, peg, and pod numbers under long-day conditions in a field study. These differences did not translate into decreased yields. The reactions occurred only under high temperatures (Bell et af., 1991a,b). They pointed out that, when selecting cultivars for adaptation to long-day, subtropical conditions, a temperature response may be important, since sensitive types may have higher critical temperatures than insensitive types.

B. SOILDEFICIENCIES AND TOXICITIES Peanut cultivars may also be developed with tolerance to soil deficiencies and toxicities. Mineral deficiencies can cause yield and quality problems in peanut production. Research on the mineral nutrition of peanuts has been summarized (Gascho and Davis, 1994; Gascho, 1995). One of the most common production problems occurs in soils with low calcium. When peanuts are grown in these soils without additional calcium, pod rot and poorly filled pods are common (Gascho and Davis, 1994). Genetic differences exist in response to effects of varying soil calcium concentrations, with large-seeded types generally requiring higher levels of calcium (Walker and Keisling, 1978). Gascho and Davis (1994) summarized many studies on genotypic response to soil calcium concentrations and found that large-seeded types required approximately twice the calcium concentration of small-seeded types. It has been accepted that calcium seed content depends on the diffusion of calcium from the soil solution. Several studies examining genetic differences in seed and pod calcium levels have identified the ratio of pod surface area to seed weight as the cause of the greater calcium requirements in large-seeded types, rather than a genetic difference in the amount of calcium required per unit weight of seed (Sumner et al., 1988; Gascho, 1992). Adams et al. (1993) identified genetic differences in requirements for ambient soil solution calcium concentrations. At present, the low cost of calcium supplementsto the soil have made genetic manipulation for this trait a low priority. Through a symbiotic relationship between the peanut plant and Bru-

PEANUT BREEDING AND GENETICS

421

dyrhizobium species, nitrogen fixation provides nitrogen throughout the growth cycle of the peanut plant. In many areas of the world, this process has sufficient inefficiency, either because of environmental conditions or a less than optimal relationship between the bacterium and the plant, that supplemental nitrogen is applied. Earlier studies suggested some promise for improved nitrogen fixation efficiency through genetic manipulation of either the Brudyrhizobium or the peanut plant (Arrendell et al., 1985; Arunachalam et al., 1984b; Joshi et af., 1990). However, subsequent studies showed that selection for improved nitrogen-fixating capabilities in peanuts would be difficult (Arrendell et al., 1988; Phillips et al., 1989). Although peanuts are grown in phosphorus-deficient soils in many parts of the world, phosphorus amendments are generally available, are inexpensive, and can easily be applied to alleviate phosphorus deficiency problems. Nevertheless, several studies have been conducted to examine genetic differences in the ability of peanut genotypes to respond to low-phosphorus soils. Variation has been reported among peanut genotypes for phosphate mobilization (Dwivedi et al., 1987; Kesava Rao et al., 1990) and for phosphorus use efficiency (Reddy et al., 1991). Relatively little work has been done on genetic differences in response to other specific soil nutrients in peanut, Reddy et al. (1993) found genetic variation in the expression of traits associated with iron stress tolerance, but in pot studies no genotypes were identified that could be classified as efficient in response to iron stress. Iron absorption efficiency has been found to be simply inherited (Gowda et al., 1993). In soils with iron deficiencies, differences in the efficiency of Bradyrhizobium strain to nodulate and the iron tolerance of the peanut cultivar were found to affect nitrogen fixation (O’Hara et al., 1993). Rhoads et al. (1989) identified Southern Runner to be more tolerant to zinc toxicity than Sunrunner. Hunshal et al. (199 1) found genetic variability among six cultivars in tolerance to salinity. In a study examining genetic tolerance to general low soil fertility, Branch and Gascho (1985) found differences among 24 cultivars. No correlations to a single soil or tissue nutrient could be found.

C. BIOTIC STRESSES Of the biotic stresses, insects are generally considered to cause less severe yield losses worldwide than either diseases or weeds. Nevertheless, the development of insect resistance is important for increased stability of yield, decreased costs of production due to reduced insecticide usage, decreased probability of environmental contamination from inappropriate use of insecticides, and specific environments where insect damage can create economic yield loss. A comprehensive report of peanut resistance to arthropod pests was made by

42 2

D. A. KNAUFT AND J. C . WYNNE

Lynch (1990). Many insects can cause yield reductions in peanut. Foliar feeders cause maximum yield loss when their feeding reduces photosynthetic area, especially during pod initiation and pod fill. Todd et al. (1991) screened commercially available genotypes for antibiosis to foliar feeders and found that NC 6 and Early Bunch had moderate antibiosis to corn earworm, fall armyworm, and velvetbean caterpillar. Several breeding lines have been shown to have resistance to peanut leafminer (Wightman and Amin, 1988; Gahukar, 1992). Resistance to termites has been reported by Amin et al. (1985). Resistance to leafhopper (Empoasca species) also exists in cultivated peanut (Wightman et al., 1990). Genetic analysis of this resistance showed that inheritance was nonadditive and that long trichomes contributed to the resistance. NC AC 2230 was found to have stable resistance and high combining ability (Dwivedi et al., 1986). Genetic variation for resistance to the stored insect pests Tribolium castaneum and Plodia interpunctella has been found in Nigeria (Mbata, 1992), and resistance to T. castaneum and Oryzaephilus surinamensis has been found in India (Singh, 1991).

Sometimes resistance to insects is important both for the damage the insect causes to the peanut plant and the disease vectored by the insect. This is particularly the case with thrips (Frankliniella species), since they cause direct damage to the plant and transmit tomato spotted wilt virus and bud necrosis. Because resistance to the insect vector does not necessarily confer resistance to the virus, screening for a combination of the two is important. An Indian cultivar, Robut 33-1, was found to have resistance both to thrips and to bud necrosis (Wightman et al., 1990). On the other hand, Southern Runner shows no resistance to thrips, yet is tolerant of tomato spotted wilt virus (Culbreath et al., 1992). Aphis craccivora causes leaf damage and transmits groundnut rosette virus, and important disease on the African continent. Several sources of resistance to the aphid have been identified (Wightman et al., 1990). A strong relationship between the amount of a condensed tannin, procyanidin, and the fecundity of aphids suggests that screening for procyanidin amount may identify genotypes with resistance to the aphid (Grayer er al., 1992). Nematode diseases are regionally important in peanut production. Kalahasti malady, caused by Qlenchorhynchus brevilineatus, is an important disease problem in India, where the disease can cause yield reductions of up to 50%. Several resistant types were found after screening 1599 genotypes. The resistant lines were, for the most part, poor yielding with undesirable pod and seed characteristics. One line showed promise for high yield (Mehan et al., 1993). A seed-borne nematode, Ditylenchus desrructor, has been identified in South Africa. While no resistance has been found, there is genetic variability for pod yield in response to nematode pressure. Fortunately, the high yielding cultivar Sellie was the most tolerant and highest yielding under infestation (Ventner et al., 1993). Peanut rootknot nematode, Meloidogyne arenuria, is the most destructive

PEANUT BREEDING AND GENETICS

423

nematode on peanuts in the United States. Part of the USDA peanut germ plasm collection has been screened in the greenhouse for resistance, and the material with the lowest gall index, egg mass index, number of eggs per plant, and number of eggs per gram of root was identified (Holbrook and Noe, 1992). When these lines were planted in the field, there was a high correlation with populations of nematodes in the rhizosphere of these genotypes. A greenhouse screening should provide good correlation to field resistance (Noe et al. 1992). Genetic differences have been shown when stored peanut seeds were subjected to infection by red flour beetle (Tribolium castuneum Herbst). Of 20 cultivars examined for reduction in kernel weight from infection, responses varied from 3.9 to 22.8%. The amount of frass left by the beetle was proportional to the number of beetles that emerged. Although none of the genotypes were immune, some resistance was identified, even using small sample sizes (Kalra and Singal, 1991). Lesser cornstalk borer resistance has been found in cultivated peanut germ plasm (Smith et al., 1980, 1990; Stalker et al., 1984). The cultivar NC 6 was released with resistance to southern corn rootworm and corn earworm (Campbell et al., 1977). Lynch and Mack (1995) have discussed the use of molecular genetics for engineering insect resistance in peanut. They indicated that major efforts will focus on incorporating endotoxins from Bacillus thuringiensis and protease inhibitors from other plants and improving the efficacy of baculoviruses that are pathogenic to insects that affect peanuts. They cautioned that, although insect resistance to genetically engineered peanuts is possible, a continued IPM approach is needed. Several foliar diseases cause serious problems worldwide to peanut production, including reduced yields and lower quality. The work on the development of pest resistance has been reviewed by Wynne et al. (1991). Specific sources of resistance to the most common disease organisms of peanut have been reported (Isleib et al., 1994), and details of inheritance studies have also been listed (Murthy and Reddy, 1993). A comprehensive discussion of peanut diseases can be found in Sherwood et al. (1995). Rust, caused by Puccinia arachidis Speg., is an important disease primarily in India and China. Several sources of resistance have been identified (Nigam et a f . , 1991; Subrahmanyam and MacDonald, 1987) in cultivated peanut. Depending on the source of resistance and the background in which it is placed, resistance has been found to be inherited either mono- or digenically (Knauft, 1987; Nigam et a l . , 1980; Tiwari et a l . , 1984; Reddy et al., 1987) or in some cases in a more complex fashion (Middleton and Shorter, 1987; Liao and Wang, 1988). Regardless, the incorporation of rust resistance into acceptable cultivars should not be difficult. No rust-resistant cultivars exist in the United States, where rust infections causing economic damage are sporadic.

424

D. A. KNAUFT AND J. C. WYNNE

Early leaf spot (Cercospora arachidicola Hori) and late leaf spot (Phaeoisariopsis personata Berk et Curt V. Ant) are the most damaging diseases of peanut worldwide. Yield reductions of up to 70% have been reported in India (Subrahmanyam et al., 1984), and reductions of over 80% have been reported in experimental plots in the United States (Knauft et al., 1988). Many studies have identified sources of resistance to either early or late leafspot used in breeding programs (Gorbet et al., 1982; Melouk et al., 1984; Subrahmanyam et al., 1985; Wells et al., 1994). Systematic evaluation of peanut germ plasm at ICRISAT (Nigam et al., 1991) and in Florida (Chiteka et al., 1988a,b) and Georgia (Anderson et al., 1993) has identified additional sources of resistance. Unlike rust diseases, the inheritance of resistance to leafspots has been consistently shown to be complex (Anderson et al., 1986, 1991; Green and Wynne, 1987; Kornegay et al., 1980; Walls and Wynne, 1985; Iroume and Knauft, 1987a,b; Jogloy et al., 1987). The resistance is the result of several components of resistance (Anderson et al., 1993; Chiteka et al., 1988a,b; Green and Wynne, 1986; Waliyar et al., 1993) and can be confounded with maturity (Friesen et al., 1992). Despite the complex inheritance, moderate resistance has been incorporated into a cultivar, Southern Runner (Gorbet et al., 1987). When crosses between resistant material and an adapted, susceptible parent have been made, occasionally lines can be recovered with high yield and greater resistance than either parent (Gorbet et al., 1990). While some reports have been made that resistant lines identified at one location are not resistant in other environments (Branch et al., 1991; ICRISAT, 1989), Chiyembekeza et al. (1993) found consistent reactions of peanut germ plasm to natural populations of leafspots in Florida and Malawi. Several soil-borne diseases affect peanuts, although their severity is generally localized. Aspergillus species cause one of the most serious problems in peanut producing areas of the world, primarily because the fungi produce the mycotoxin aflatoxin. The subject has been reviewed (Keenan and Savage, 1994). The development of genetic resistance to the disease has been difficult. Resistance to Aspergillus flavus has been identified in the pod, the seed coat, and the cotyledons (Isleib et al., 1994), as well as resistance to preharvest seed infection, in vitro seed colonization, and aflatoxin production per se. The types of resistance are not necessarily correlated, nor do laboratory, greenhouse, and field studies show consistent findings. Isleib et al. (1994) have published a list of peanut lines with one or more types of identified resistance. Because A . flavus is a weak pathogen, prevention of its initial infection into the peanut plant may be an important strategy for developing genetic resistance to the organism. Pods on a developing plant seem predisposed to A. flavus infection when plants are undergoing drought stress and soil temperatures in the podding zone are high (Cole, 1989). The development of drought-tolerant peanut lines may be useful in reducing the effects of A. flavus as well as of drought.

PEANUT BREEDING AND GENETICS

42 5

There are other soil-borne pathogens that cause production problems in specific peanut producing areas. Southern stem blight, Sclerotium rolfsii Sacc., causes multimillion dollar yield losses in the southeastern United States. Sources of resistance have been identified (Grichar and Smith, 1992; Smith et al., 1989; Branch and Csinos, 1987), and resistance has been incorporated into adapted, high yielding lines (Branch and Brenneman, 1993). The cultivar Southern Runner has a moderate level of resistance (Wells et al., 1994). Sclerotinia minor resistance has been identified (Akem et al., 1992; Melouk et al., 1989; Porter et al., 1992). Although inheritance of resistance to Sclerotinia minor is complex (Wildman et al., 1992), resistant cultivars and germ plasm have been released (Coffelt et al., 1982; Smith et al., 1990). Pod rot, caused by a complex of organisms, is a problem in certain parts of the world. Some genetic variation exists for resistance to this pest (Lewis and Filonow, 1990). Resistance to Cylindrocladium black rot is known, and the inheritance of resistance is complex (Green et al., 1983). Resistant types appear to delay the onset of epidemics rather than the rate of disease progress. Virus diseases are serious problems in many parts of the world, and variation for resistance has been reported (Isleib et al., 1994). Groundnut rosette virus resistance has been reported to be digenic (Olorunju et al., 1992; Nigam and Bock, 1990) and has been incorporated into cultivars. Some tolerance to bud necrosis and tomato spotted wilt virus has been found (Nigam et al., 1991). While genetic variation exists for the transmission of peanut stripe virus (Warwick and Demski, 1992), extensive screening for high levels of resistance to the virus has been unsuccessful (Nigam et al., 1991). Lines that fail to transmit peanut mottle virus have been identified (Isleib et al., 1994). On a worldwide basis, yield losses from weeds are significant. In the United States, where herbicides are commonly used, total losses associated with weeds and their control ranged from $132/ha in Texas to $391/ha in Florida (Wilcut et al., 1995). Yet the study of the interactions between peanuts and other plants has received far less attention than other biotic pest problems. Genetic differences in peanut competitiveness with weeds are known. Studies have identified genotypes that were less affected by competition from Xanthium strumurium L. than other lines (Fiebig et al., 1991), and genetic differences have been noted in response to horsenettle (Solanum carolinensis) and silverleaf nightshade (Solanum elaeagnifoliurn)(Hackett et al., 1987a,b). Fiebig et al. (1991) noted that the increased competitiveness with weeds came at a cost. Higher vegetative growth rates were needed to provide shading competition with weeds, and the increased vegetative growth rates reduced the partitioning rate to pods and caused an overall decrease in pod yields when expressed on a weed-free basis. The relative yield cost of weed competitiveness compared to the cost of weed control must be weighed in each system.

426

D. A. KNAUFT AND J. C . WYNNE

Manipulation of plant spacings would be expected to provide increased competitive ability of peanuts with weeds. However, most studies examining weed spacing and competition with weeds have found little or no response to increased spacing (Colvin er af., 1985a,b; Wehtje er af., 1984). Where responses have been reported (Bell er af., 1991b, 1993a,b), they have occurred with very early, erect Spanish types. In addition, no differences were noted for spatial arrangement. Therefore, it is not expected that significant advantages would be obtained from specific peanut genotypes. Differential responses to herbicides have been shown in peanut (Knauft et al., 1990), and there is potential for reducing herbicide inputs in peanut production through the use of more effective herbicides. Molecular manipulation for the development of peanut lines with resistance to broad-spectrum, safer herbicides may allow for further reductions in herbicide use, although peanuts engineered for herbicide tolerance may increase dependence on a single herbicide, which could increase the probability of weed resistance to the herbicide. Although the development of peanut genotypes with a high level of competitive ability is desirable for weed competition, in other instances, such as cases where peanut is intercropped, the ability to produce pod yields effectively without reducing the yield of companion crops is desirable. Several studies on the genotype response of peanut to intercropping with cereals have shown little genotype interaction with intercropping, suggesting that the highest yielding peanut line grown in pure stands would also be the most productive type in intercrop systems (Knauft, 1984; Ndunguru and Williams, 1993). Besides yield per se, the ability of the legume crop to provide nitrogen for the companion crops is important, especially in lower input systems. Senaratne and Ratnasinghe (1993) showed a more than 2-fold difference in the ability of intercropped peanut to provide nitrogen to a companion corn crop. The genotype fixing the largest amount of nitrogen did not supply the greatest amount to the companion crop, and corn yields were highest with the companion crop providing the greatest amount of nitrogen. Pod yields from this peanut cultivar were lower than that from other genotypes in the test.

D. EXPANDEDUSES In many parts of the world, peanut consumption has not increased. The crop has traditionally been used for two different purposes throughout the world. In the countries outside the United States, production has been for oil. Peanut oil is a high quality cooking oil that has a high flash point, allowing foods to be cooked rapidly, retaining flavor and nutrition. Other cooking oils exist in most markets, and peanut oil must compete successfully on the basis of quality or cost in local markets with other oils, including soybean, sunflower, palm, safflower, olive,

PEANUT BREEDING AND GENETICS

42 7

and rapeseed. Increased use of peanut oil requires either a cost or quality advantage. Cost advantages can be achieved through increased production of seed or increased oil content in the seed. Increased production and/or reduced costs of production have been previously described. Various researchers have examined the range of oil content in peanut genotypes, and values from 31 to 55% have been reported (Bovi, 1983; Jambunathan ef af., 1985, 1993b; Raheja et af., 1988; Dwivedi et af., 1990, 1993; Salunkhe et af., 1992). 'Ijlpically the oil content is influenced by the environment (Knauft et af., 1986b; Dwivedi et af., 1990). Many wider ranges of oil content were identified from single year or single location studies. Dwivedi et af. (1993) examined 13 genotypes that represented extremes from measuring the oil content of 8OOO germ plasm lines in the ICRISAT collection. They found that the original range of 34-54% oil content was not repeatable and that after multiple locations were used in evaluation the content ranged from 45 to 50%. Of the environmental conditions measured, they found higher soil pH and iron content were associated with higher oil content. A similar narrow range after repeated evaluation was found by Bansal et af. (1993). A curious variant in oil content has been described by Jakkula et af. (1993). Lines from the Florida breeding program were identified with variations in oil content within individual, true-breeding plants ranging from 20 to 50% for seed within a plant. Seeds with low oil tended to be produced later during the growing season than higher oil content seeds. As oil content decreased, the moisture content of fresh seed composition increased. On a dry weight basis, carbohydrate content, primarily sucrose, increased. Inheritance of this trait is paternal. When the low-oil lines are used as female parents, no low-oil segregants are found in F, or subsequent generations. When the low-oil lines are used as male parents, segregation follows a monogenic 3: 1 norma1:low-oil ratio. Generations following the initial F, segregating generation show inheritance patterns consistent with a monogenic inheritance pattern. Peanut seed composition includes approximately 50% oil, 25% protein, 20% carbohydrate, and 5% fiber and ash. Properties of peanut oil are determined by the fatty acid composition. Approximately 90% of peanut oil is composed of palmitic acid (16 carbons and no double bonds: 16:0), oleic acid (18:1), and linoleic acid (18:2). Although many studies have identified genetic differences in fatty acid composition in peanuts, most have examined a limited number of genotypes. The largest study examining differences in fatty acid composition among peanut genotypes was conducted by Norden et af. (1987). From 450 lines, they identified a breeding line with 80% oleic acid and 2% linoleic acid. The 80% oleic acid content was significantly higher than any previous report, and the 2% linoleic acid was a major deviation from previously known lower limits. Breeders and geneticists can easily manipulate the characteristic, as it is simply inherited (Moore and Knauft, 1989).

428

D. A. K N A m AND J. C.WYNNE

Incorporation of the trait through back-crossing further reduces the palmitic acid content of the seed, without other deleterious traits (Knauft et al., 1993b). The trait improves the keeping quality of peanut oil (O'Keefe et al., 1993), has some implications for health benefits, and can be used to improve the composition of pork products (Myer et al., 1992). Preliminary evidence has shown some association of the trait with reduced aflatoxin contamination (Holbrook et al., 1994). While changing the fatty acid composition of the peanut has been shown to improve the keeping quality of peanut oil, there may be additional methods of improving quality. Peanuts contain four forms of the antioxidant tocopherol. Contents of a-,p-, y-, and b-tocopherols vary with maturity and genotype. Nearly 2-fold differences existed for the individual and total tocopherol contents among eight commonly grown cultivars in the United States (Hashim et al., 1993). Given this type of phenotypic range within the relatively similar genetic background of runner and Virginia market-type cultivars in the United States (Knauft and Gorbet, 1989), genetic improvements in the antioxidant content of the peanut should be possible. A wide range of protein content has been reported in peanuts, from 16.2 to 36%. Most genotypes average near 25% protein (Dwivedi et al., 1990; Salunkhe et al., 1992). Protein and oil contents are inversely correlated (Dwivedi et al., 1990). Most of the protein is found in one of two storage proteins, arachin or conarachin. The amino acid composition of peanut also differs depending on the genotype, although major differences have been found in other Arachis species besides A. hypogaea (Basha and Pancholy, 1984; Bianchi-Hall et al., 1993). Some studies have reported differences in protein nutritional quality and composition (Ghuman et al., 1990; Kim et al., 1992), while others have shown little significant variation in protein quality among peanut genotypes (Basha, 1992; Jambunathan et al., 1992) or species (Jambunathanet al., 1993a).A methioninerich protein fraction has been shown to vary with peanut genotype (Basha, 1991). Relatively low levels of antinutritional factors, such as trypsin and chymotrypsin, are found in peanut. Higher levels of these factors have been reported in some accessions of wild species than in cultivated types (Jambunathan et al., 1993a). In the United States, peanuts are the major allergenic food among adults and are one of the major allergenic foods among children (Taylor, 1992). Reactions range in symptoms from skin rash to fatal anaphylactic shock. Most of the allergens have remained unidentified, although allergen activity is present in the two major storage proteins in peanut, arachin and conarachin (Burks et al., 1991, 1992). Characterization of the allergenic responses has not provided sufficient information to determine whether allergic responses could be reduced through either conventional or molecular genetic manipulation of peanut proteins. Carbohydrate contents between 22 and 33% have been reported for peanut

PEANUT BREEDING AND GENETICS

429

seeds (Savage and Keenan, 1994). Differences in extraction procedures as well as moisture and oil contents contribute to the variation in reported levels. Carbohydrates in the seed include disaccharides (primarily sucrose), oligosaccharides (primarily starch), and insoluble compounds. Variation in either the content or composition of peanut carbohydrates is not expected to have a significant effect on the utilization of peanut seed, and few studies have examined genetic differences. The peanut industry has long been interested in improving the roasted flavor of peanut products. Pattee and Giesbrecht (1990) and Pattee et ul. (1993) found genetic differences for the intensity of roasted peanut flavor, as well as other components that contribute to peanut taste. Broad-sense heritabilities for roasted peanut, sweet, and nutty sensory attributes ranged from 0.05 to 0.68 in the two studies. Although flavor is difficult to evaluate and therefore has received little emphasis in breeding programs, there may be potential to improve peanut flavor through breeding and selection. Most studies on peanut composition differences have been associated with the use of peanuts as a food or oil crop. These differences may be related to agronomic seed quality as well. Ketring (1992) has shown genotype differences in seed quality reduction and field emergence within and across seed storage periods. He suggests that there is a genetic potential to improve the longevity of seed quality during storage that would enhance the stability of field emergence.

VII. SUMMARY AND CONCLUSIONS Increased support for peanut breeding, genetics, and related research during the last two decades has led to the identification of variation for most abiotic and biotic stresses that affect peanut production. Peanut breeders, using traditional breeding procedures for self-pollinated species, are transferring these traits to improved germ plasm or cultivars. Cultivars with improvements for both abiotic and biotic stresses are beginning to be used in production systems, especially those systems using low inputs. Traits affecting consumer preference are receiving considerable attention in breeding programs in the United States. The many studies reported here suggest that conventional breeding procedures can be used to develop new and improved peanut cultivars with many desired traits not presently available to peanut growers. The use of molecular genetic techniques is at an early stage in peanuts compared to many other plant species of economic importance. Public concern about the safety of genetically engineered products will be particularly important for peanuts, as children are a large part of the domestic peanut market. Incorporation of resistances unavailable in Aruchis germ plasm, such as to various biotic

430

D. A. KNAUFT AND J. C. WYNNE

stresses as well as herbicide tolerance, may provide an opportunity for growers to produce peanuts at a lower cost with fewer inputs.

REFERENCES Adams, J. F., Hartzog, D. R., and Nelson, D. P. (1993). Supplemental calcium application on yield, grade, and seed quality of runner peanut. Agron. J. 85, 86-93. Akem, C. N.,Melouk, H. A., and Smith, 0. D. (1992). Field evaluation of peanut genotypes for resistance to sclerotinia blight. Crop Prorecr. 11, 345-348. Amin, P. W., Singh, K. N., Dwivedi, S . L., and Rao, V. R. (1985). Sources of resistance to jasid (Empwscu kerri Pruthi), thrips (Frunkliniellu schulrzei (Trybom)), and termites (Odonrorermes sp.) in groundnut (Aruchis hypogueu L.).Peanut Sci. 12, 58-60. Anderson, W. F., Wynne, J. C., Green, C. C., and Beute, M. K. (1986). Combining ability and heritability of resistance to early and late leaf spot of peanut. Peunur Sci. 13, 10-14. Anderson, W. F., Mozingo, R. W., and Wynne, J. C. (1989). Comparison of stability statistics as criteria for cultivar development in peanut. Peunur Sci. 16, 21-25. Anderson, W. F.. Beute, M. K., Wynne, J. C., and Wongkaew, S. (1990a). Statistical procedures for assessment of resistance in a multiple foliar disease complex of peanut. Phyroparhology 80, 145I - 1459. Anderson, W. F., Patanothai, A., Wynne, J. C., and Gibbons, R. W. (1990b). Assessment of a diallel cross for multiple foliar pest resistance in peanut. Oleuginew 45, 373-378. Anderson, W. F., Holbrook, C. C., and Wynne, I. C. (1991). Heritability and early-generation selection for resistance to early and late leafspot in peanut. Crop Sci. 31, 588-593. Anderson, W. F., Holbrook, C. C., and Brenneman, T. B. (1993). Resistance to Cercosporidium personarum within peanut germplasm. Peanut Sci. 20, 53-57. Arrendell, S., Wynne, J. C., Elkan, G.H., and Isleib, T.G.(1985). Variation from nitrogen fixation among progenies of a Virginia x Spanish peanut cross. Crop Sci. 25, 865-869. Arrendell, S., Wynne, J. C., Ekan. G. H.,and Schneeweis, T.J. (1988). Selection among early generation peanut progeny for enhanced nitrogen fixation. Peunur Sci. 15, 90-93. Arunachalam, V., Bandyopadhyay, A., Nigam, S . N., and Gibbons, R. W. (1984a). Heterosis in relation to genetic divergence and specific combining ability in groundnut (Aruchis hypogueu L.). Euphyricu 33, 33-39. Arunachalam, V., Pungala, G.D., Dutta, M., Nambiar, P. T. C., and Dart,P. (1984b). Efficiency of nitrogenase and nodule weight in predicting relative performance of genotypes assessed by a number of characters in groundnut (Aruchis hypogueu L.) Exp. Agric. 20, 303-309. Arunachalam, V., Bandyopadhyay,A., and Koteswara Rao, M. V. (1985). Performance of three-way crosses in groundnut. Ind. J. Agric. Sci. 55, 75-81. Ashri, A., and Levy, A. (1978). Natural and induced plasmon variation affecting growth habit in peanuts, A. hypogueu. In “Experimental Mutagenesis in Plants, Bulgarian Academy of Sciences, Sofia,” pp. 411-438. Atlin, G. N., and Frey, K. J. (1990). Breeding crop varieties for low-input agriculture. Am. J. Alrern Agric. 4, 53-58. Bagnall, D. J., and King, R. W. (1991). Response of peanut (Amchis hypogueu) to temperature, photoperiod and irradiance. I . Effect on flowering. Field Crops Res. 26, 263-277. Bailey, E., and Boisvert, R. N. (1991). Yield response of groundnuts to the amount and timing of water application. J. Prod. Agric. 4, 1-9. Baker, C. M.,and Wetzstein, H. Y. (1992). Somatic embryogenesis and plant regeneration from leaflets of peanut, Aruchis hypogueu. Plant Cell Rep. 11, 71-75.

PEANUT BREEDING AND GENETICS

43 1

Bandyopadhyay. A.. Arunachalam, V., and Venkaiah, K. (1985). Efficient selection intensity in early generation index selection in groundnut. Theor. Appl. Genet. 71, 300-304. Banks, D.J., and Pittman, R. N. (1986). Origin, inheritance and characteristics of a yellow flowered peanut from Bolivia. Proc. Am. Peunur Res. Educ. SOC. 18, 31 (Abstr.). Bansal, U. K.,Satija, D. R., and Ahuja, K. L. 0. (1993). Oil composition of diverse groundnut (Aruchis hypogueu L.) genotypes in relation to different environments. J. Sci. Food Agric. 63, 17-19. Basha, S . M. (1991). Deposition pattern of methionine-rich protein in peanuts. J. Agric. Food Chem. 39, 88-91. Basha, S. M. (1992). Soluble sugar composition of peanut seed. J. Agric. Food Chem. 40,780-783. Basha, S . M.,and Pancholy, S. K. (1984). Variations in the methionine-rich protein composition of the genus Aruchis. Peunur Sci. 11, 1-3. Bell, M. J., and Harch. G. (1991). Effects of photoperiod on reproductive development of peanut (Aruchis hypogueu L.) in a cool subtropical environment. I. Field studies. Ausr. J. Agric. Res. 42, 1133-1149. Bell, M. J., Bagnall, D. I., and Harch, G. R. (1991a). Effects of photoperiod on reproductive development of peanut (Aruchis hypogueu L.) in a cool subtropical environment. 11. Temperature interactions. Ausr. J. Agric. Res. 42, 1151-1161. Bell, M. 3.. Harch, G., and Wright, G . C. (1991b). Plant population studies on peanut (Aruchis hypogueu L.) in subtropical Australia. I . Growth under fully irrigated conditions. Ausr. J. Exp. Agric. 31, 535-543. Bell, M. J., Wright, G. C., and Harch, G. R. (1993a). Environmental and agronomic effects on the growth of four peanut cultivars in a sub-tropical environment. 1. Dry matter accumulation and radiation use efficiency. Expl. Agric. 29, 473-490. Bell, M. J., Wright, G. C., and Harch, G. R. (1993b). Environmental and agronomic effects on the growth of four peanut cultivars in a sub-tropical environment. II. Dry matter partitioning. Expl. Agric. 29, 491-501. Bhagat, N. R.,Ahmad, T., Lalwani, H. B., Patil, S. A,, Patra. G. J., and Acharya, D. (1988). Screening of bunch peanut (Aruchis hypogueu L. ssp. fusrigiutu) germplasm for cold tolerance in India. Trop. Agric. (Trinidad) 65, 109-1 12. Bianchi-Hall, C., Keys, R. D., Stalker, H.T., and Murphy, J. P. (1993). Diversity of seed storage proteins in wild peanuts (Aruchis species). Plunr Syst. Evol. 186, 1-15. Bovi, M. L. A. (1983). Genotypic and environmental effects on fatty acid composition, iodine value, and oil content of peanut (Aruchis hypogueu L.).Ph.D. Dissertation, University of Florida. Diss. Abstr. Int. B 44,406. Branch, W. D. (1989). Inheritance of dominant white peanut testa color. J. Hered. 80, 155-156. Branch, W. D., and Brenneman, T. B. (1993). White mold and rhizoctonia limb rot resistance among advanced Georgia peanut breeding lines. Peunut Sci. 20, 124-128. Branch, W. D., and Csinos, A. S. (1987). Evaluation of peanut cultivars for resistance to field infection by Sclerofium rolfsii. Plunr Dis. 71, 268-270. Branch, W. D., and Gascho, G. I. (1985). Screening for low fertility tolerance among peanut cultivars. Agron. J. 77, 963-965. Branch, W. D., and Hammons, R. 0. (1981). Disgenic inheritance for the flop trait in peanuts. Crop Sci. 21, 385-386. Branch, W. D., and Hammons, R. 0. (1983). Inheritance of micro phenotype in peanut. Crop Sci. 23, 1045-1046. Branch, W. D., and Hildebrand, G. L. (1989). Pod yield comparison of pure-line peanut selections simultaneously developed from Georgia and Zimbabwe breeding programs. Plunr Breed. 102, 260-263. Branch, W. D., and H o l b m k , C. C. (1988). Genetic relationship between RI, R2, and R3 for red peanut testa color. Peunur Sci. 15, 13-14.

432

D. A. K N A W AND J. C . WYNNE

Branch, W. D., and Holbrook, C. C. (1991). Registration of CPES peanut germplasm population. Crop Sci. 31,497-498. Branch, W. D., and Kvien, C. K. (1992). Cytoplasmically inherited albinism in peanut seedlings. J. Hered. 83,455-457. Branch, W. D., Kirby, J. S., Wynne, J. C., Holbrook, C. C., and Anderson, W. F. (1991). Sequential vs. pedigree selection method for yield and leafspot resistance in peanut. Crop Sci. 31, 274276. Brar, G. S.,Cohen, B. A.. Vick, C. L., and Johnson, G. W. (1994). Recovery of transgenic peanut (Arachis hypogaea L.) plants from elite cultivars utilizing ACCELL (r) technology. Plant J. 5 , 745-75 3. Burks, A. W., Williams, L. W., Helm, R. M., Connaughton, C., Cockrell, G., and O’Brien, T. (1991). Identificationof a major peanut allergen, Ara h I, in patients with atopic dermatitis and positive peanut challenges. J. Allergy Clin. Immunol. 88, 172-179. Burks, A. W., Williams, L. W., Connaughton. C., Cockrell, G., O’Brien, T. J., and Helm, R. M. (1992). Identificationand characterization of a second major peanut allergen, Ara h 11, with use of the sera of patients with atopic dermatitis and positive peanut challenge. J. Allergy Clin. Imm~n01.90,%2-%9. Campbell, W. V., Wynne, J. C., Emery, D. A., and Mozingo, R. W. (1977). Registration of NC6 peanuts. Crop Sci. 17, 346. Chapman, S. C., Ludlow, M. M., Blamey, F. P. C., and Fischer, K. S. (1993a). Effect of drought during early reproductive development on growth of cultivars of groundnut (Arachis hypogaea L.). 11. Biomass production, pod development and yield. Field Crops Res. 32, 21 1-225. Chapman, S. C., Ludlow, M. M.. Blamey, F. P. C., and Fischer, K. S. (1993b). Effect of drought during pod filling on utilization of water and on growth of cultivars of groundnut (Arachis hypogaea L.). Field Crops Res. 32, 243-255. Chavan, A. A., Dhoble, M. V.. and Khating, E. A. (1992). Effect of artificial water stress on different genotypes of groundnut (Arachis hypogaea) in dryland. Ind. J. Agric. Sci. 62, 376381. Chengalrayan, K., Sathaye, S. S., and Hazra, S. (1994). Somatic embryogenesis from mature embryo-derived leaflets of peanut (Arachis hypogaea L.). Plant Cell Rep. 13, 578-581. Chiteka. Z. A., Gorbet, D. W., Shokes, F. M., Kucharek, T. A., and Knauft, D. A. (1988a). Components of resistance to late leafspot in peanut. I. Levels and variability-implications for selection. Peanut Sci. 15, 25-30. Chiteka, Z. A., Gorbet, D. W., Knauft, D. A,, Shokes, F. M., and Kucharek, T. A. (1988b). Components of resistance to late leafspot in peanut. 11. Correlations among components and their significance in breeding for resistance. Peanut Sci. 16, 76-81. Chiyembekeza, A. J., Knauft, D. A., and Gorbet, D. W. (1993). Comparison of peanut resistance components to late leafspot in different environments. Crop Sci. 33, 994-997. Clemente, T. E., Robertson, D., Isleib, T. G., Beute, M. K., and Weissinger, A. K. (1992). Evaluation of peanut (Arachis hypogaea L.) leaflets from mature zygotic embryos as recipient tissue for biolistic gene transfer. Transgen. Res. 1, 275-284. Coffelt, T. A. (1974). Inheritance of growth habit in an infraspecific cross population of peanuts. J . Hered. 65, 160-162. Coffelt, T.A. (1989a). Natural crossing of peanut in Virginia. Peanut Sci. 16, 46-48. Coffelt, T. A. (1989b). Peanut. I n “Oil Crops of the World (G.Robbelen, R. K. Downey, and A. Ashri, eds.), pp. 319-338. McGraw-Hill Publ. Co., New York. Coffelt, T.A., and Potter, D. M.(1982). Screening peanuts for resistance to Sclerotinia blight. Plant Dis. 66, 385-387. Coffelt, T. A., Porter, D. M., and Mozingo, R. W. (1982). Registration of Virginia 81 Bunch Peanut. Crop Sci. 22, 1085-1086.

PEANUT BREEDING AND GENETICS

43 3

Coffelt, T. A., Seaton, M. L., and Van Scoyoc, S. W. (1989). Reproductive efficiency of 14 virginiatype peanut cultivars. Crop Sci. 29, 1217-1220. Coffelt, T. A., Wynne, J. C., and Monteverde-Penso, E. J. (1993). Genotype x environment interaction effects on cultivar development. Oleaginew 48(3), 133-137. Coffelt, T. A., Porter, D. M., and Mozingo, R. W. (1994). Registration of ‘VA 93B’ Peanut. Crop Sci. 34, 1126.

Cole, R. 3. (1989). Preharvest aflatoxin in peanuts. Inr. Biodeterior. 25, 253-257. Colvin, D. L., Walker, R. H., Patterson, M. G., Wehtje, G., and McGuire, J. A. (1985a). Row pattern and weed management effects on peanut production. Peanut Sci. 12, 22-27. Colvin, D. L., Wehtje, G. R., Patterson, M., and Walker, R. H. (1985b). Weed management in minimum-tillage peanuts (Arachis hypogaea) as influenced by cultivar, row spacing, and herbicide. Weed Sci. 33, 233-237. Culbreath, A. K., Todd, J. W., and Chamberlain, J. R. (1992). Disease progress of spotted wilt in peanut cultivars Florunner and Southern Runner. Phytoparhology 82, 766-771. Daimon, H.,and Mii, M. (1991). Multiple shoot formation and plantlet regeneration from cotyledonary node in peanut (Arachis hypogaea L.). Jpn. J. Breed. 41, 461-466. Dashiell, K. E. (1984). Inheritance of nodulation and its association with genes controlling testa color in Arachis hypogaea L. Ph.D. Dissertation University of Florida, Gainesville, FL. Diss. Abstr. 83, 24953.

Dashiell, K. E., and Gorbet, D. W. (1982). Genetic analysis of a non-nodulating peanut. Agron Abstr., 63. Dashiell, K. E., Kirby, J. S., and McNew, R. W. (1982). Genotype X environment interaction among peanut lines in Oklahoma. Peanur Sci. 9, 24-27. Dunbar, K. B., and Pittman, R. N. (1992). Adventitious shoot formation from mature leaf explants of Arachis species. Crop Sci. 32, 1353-1356. Duncan, W. G., McCloud, D. E., McGraw, R. L., and Boote, K. J. (1978). Physiological aspects of peanut yield improvement. Crop Sci. 18, 1015-1020. Durga Prasad, M. M. K., Arunachalam, V., and Bandyopadhyay, A. (1985). Diversity pattern elucidating choice of parents for hybridization in varieties of groundnut, Arachis hypogaea L. Trop. Agric. (Trinidad) 62, 237-242. Durham, R. E., and P m t t , W. A. (1992). Repetitive somatic embryogenesis from peanut cultures in liquid medium. Planr Cell Rep. 11, 122-125. Dutta, M., and Reddy, L. J. (1988). Further studies on genetics of non-nodulation in peanut. Crop Sci. 28, 60-62. Dutta, M., Arunachalam, V., Bandyopadhyay, A., and Prabhu, K. V. (1986). Early generation intermating for yield improvement in groundnut (Arachis hypogaea L.). Theor. Appl. Genet. 71, 662-666.

Duvick, D. N. (1993). The role of seed companies in crop improvement. I n “Crop Improvement for Sustainable Agriculture’’ (M. B. Callaway and C. A. Francis, eds.), pp. 46-58. University of Nebraska Press, Lincoln, NE. Dwivedi, S. L., and Nigam, S. N. (1989). Inheritance of a puckered leaf mutant in groundnut (Arachis hypogaea L.). Current Sci. 58, 1149-1 150. Dwivedi, S. L., Amin, P. W., Rasheedunisa, Nigam, S. N., Nagabhushanam, G. V. S.,Rao, V. R., and Gibbons, R. W. (1986). Genetic analysis of trichome characters associated with resistance to jasid (Empoasca kerri Pruthi) in peanut. Peanut Sci. 13, 15-18. Dwivedi, R. S., Joshi, Y.C.. and Shara, S. N. (1987). Modeling of peanut (Arachis hypogaea L.) for higher yield on phosphorus deficient soil. Oleagineux 42, 165-168. Dwivedi, S. L., Thendapani, K., and Nigam, S. N. (1989). Heterosis and combining ability studies and relationship among fruit and seed characters in peanut. Peanut Sci. 16, 14-20. Dwivedi, S. L., Jambunathan, R., Nigam, S. N., Raghunath, K., Shankar, K. R., and

434

D. A. KNAUIT AND J. C . WYNNE

Nagabhushanam, G. V. S. (1990). Relationship of seed mass to oil and protein contents in peanut (Arachis hypogaea L.). Peanut Sci. 17,48-52. Dwivedi, S. L., Nigam, S. N., Jambunathan, R., Sahrawat, K. L., Nagabhushanam, G.V. S., and Raghunath, K. (1993). Effect of genotypes and environments on oil content and oil quality parameters and their correlation in peanut (Arachis hypogaea L.). Peanut Sci. 20, 54-59. Essomba, N. B., Coffelt, T. A,, Branch, W. D., and Van Scoyoc, S. W. (1987). Inheritance of morpological traits in peanut. Pmc. Am. Peanut Res. Educ. Soc. 19, 16 (Abstr.). Essomba, N. B., Coffelt, T. A., Branch, W. D., and Van Scoyoc, S. W. (1993). Inheritance of leaflet size in peanut (Arachis hypogaea L.). Peanut Sci. 20, 90-93. Fiebig, W. W., Shilling, D. G., and Knauft, D. A. (1991). Response of peanut genotypes to interference from common cocklebur. Crop Sci. 31, 1289-1292. Francis, C. A. (1993). Crop breeding objectives and methods. In “Crop Improvement for Sustainable Agriculture” (M. B. Callaway and C. A. Francis, eds.), pp. 59-78. University of Nebraska Press, Lincoln, NE. Franklin, C. I., Shorrosh, K. M., Trieu, A. N., Cassidy, B. G.,and Nelson, R. S. (1993). Stable transformation of peanuts callus via Agrobacterium-mediated DNA transfer. Transgen. Res. 2, 321-324.

Friesen, B. E., Knauft, D. A., and Gorbet, D. W. (1992). The interaction between maturity and leafspot resistance in peanut. Abstr. Proc. Am. Peanut Res. Educ. SOC. 24, 25. Gahukar, R. T. (1992). Groundnut entomology: retrospect and prospect. Agric. Zool. Rev. 5, 139199.

Gascho, G. J. (1992). Groundnut (Peanut). In “IFA World Fertilizer Use Manual” (D. J. Halliday, M. E. Trenkel, and W. Wichmann, eds.), pp. 209-242. Intl. F e d . Ind. Assoc., Paris. Gascho. G.J. (1995). Soil fertility. In “Advances in Peanut Science’’ (H. E. Pattee and H. T. Stalker, eds.), in press. Am. Peanut Res. Educ. Soc.,Ahoskie, NC. Gascho, G. I., and Davis, J. G. (1994). Mineral nutrition. In ‘The Groundnut Crop: A Scientific Basis for Improvement” (J. Smartt, ed.), pp. 214-254. Chapman & Hall, London. Ghuman, P. K., Mann, S. K.,and Hira, C. K. (1990). Evaluation of protein quality of peanut (Arachis hypogaea) cultivars using Terrahymena pyrformis. J. Sci. Food Agric. 52, 137-139. Gibbons, R. W. (1980). The ICRISAT groundnut program. In “Proc. Intl. Workshop on Groundnuts, 13-17 Oct.,” pp. 12-16. ICRISAT, Patanchew, India. Gorbet, D. W., and Burton, J. C. (1979). A non-nodulating peanut. Crop Sci. 19, 727-728. Gorbet, D. W., and Knauft, D. A. (1994). Andru 93 Peanut. Univ. FL Agric. Expt. Stn. Special Publication S-391. Gorbet, D. W., Shokes, F. M., and Jackson, L. J. (1982). Control of peanut leafspot with a combination of resistance and fungicide treatment. Peanut Sci. 9, 87-90. Gorbet, D. W., Norden, A. J., Shokes, F. M., and Knauft, D. A. (1987). Registration of Southern Runner peanut cultivar. Crop Sci. 27, 817. Gorbet, D. W.. Knauft, D. A., and Shokes, F. M. (1990). Response of peanut genotypes with differential levels of leafspot resistance to fungicide treatments. Crop Sci. 30, 529-533. Gorbet, D. W., Knauft, D. A., and Norden, A. J. (1992). Registration of ‘MarcI’ peanut. Crop Sci. 32, 279.

Gowda, M. V., Kulkami, V. N., Nadaf, H. L., Habib, A. F., and Nadaf, S. K. (1993). Inheritance of iron absorption efficiency in groundnut. Crop Irnprov. 20, 197-200. Grayer, R. J., Kimmins, F. M., Padgham, D. E., Harborne, J. B., and Ranga Rao, D. V. (1992). Condensed tannin levels and resistance of groundnuts (Arachis hypogaea) against Aphis craccivora. Phytochemistry 31, 3795-3800. Green, C. C., and Wynne, J. C. (1986). Field evaluation of the components of partial resistance to early leafspot in peanut. Euphytica 35, 561-573. Green, C. C., and Wynne. J. C. (1987). Genetic variability and heritability for resistance to early leafspot in four crosses of Virginia-type peanut. Crop Sci. 27, 18-21.

PEANUT BREEDING AND GENETICS

43 5

Green, C. C., Wynne. J. C., and Beute, M. K. (1983). Genetic variability and heritability estimates based on the F2 generation from crosses of large seeded Virginia type peanuts with lines resistant to Cylindrochdium black rot. Peunur Sci. 10, 47-5 I . Greenberg, D. C., Williams, J. H., and Ndunguru, B. J. (1992). Differences in yield determining processes of groundnut (Aruchis hypogueu L.)genotypes in varied drought environments. Ann. Appl. Biol. 120, 557-566. Gregory, W. C., and Gregory, M. P. (1976). Groundnut Aruchis hypogueu (Leguminosae Papilionatae). I n “Evolution of Crop Plants” (N. w. Simmonds, ed.), pp. 151-154. Longman, London. Gregory, W. C.. Krapovickas. A., and Gregory, M. P. (1980). Structures, variation, evolution and classification in Aruchis. I n “Advances in Legume Science” (R. J. Summefield and A. H. Bunting, eds.), pp. 469-481. Royal Botanic Gardens, Kew, UK. Grichar, W. J., and Smith, 0. D. (1992). Variation in yield and resistance to southern stem rot among peanut (Aruchis hypogueu L.)lines selected for pythium pod rot resistance. Peunur Sci. 19,5558. Grieshammer, U.,and Wynne, J. C. (1990a). lsozyme variability in mature seeds of U.S. peanut cultivars and collections. Peunur Sci. 17, 72-75. Grieshammer, U.,and Wynne, J. C. (1990b). Mendelian and non-Mendelian inheritance of three isozymes in peanut (Aruchis hypogueu L.).Peunur Sci. 17, 101-105. Guok, H. P., Wynne, J. C., and Stalker, H. T. (1986). Recurrent selection within a population from an interspecific peanut cross. Crop Sci. 26, 249-253. Gupta, R. B. (1988). Genetics of brachytic sterility in groundnut (Aruchis hypogueu L.).I d . J . Agric. Res. 22, 41-50. Hackett, N. M., Murray, D. S . , and Weeks, D. L. (1987a). Interference of horsenettle (Solunum curolinense) with peanuts (Aruchis hypogueu). Weed Sci. 35, 780-784. Hackett, N. M., Murray, D. S., and Weeks, D. L. (1987b). Interference of silverleaf nightshade (Solunum elueugnifolium) on Spanish peanuts (Aruchis hypogueu). Peunur Sci. 14, 39-41. Halward, T. M., and Stalker, H. T. (1987a). Comparison of embryo development in wild and cultivated Aruchis species. Ann. Bot. 59, 9-14. Halward, T. M., and Stalker, H. T. (1987b). Incompatibility mechanisms in interspecific peanut hybrids. Crop Sci. 27, 456-460. Halward, T. M., and Wynne, J. C. (1992). Progress and variability after four cycles of recurrent selection in peanut. Peunur Sci. 19, 20-24. Halward, T. M., Wynne, J. C., and Monteverde-Penso, E. J. (1990). The effectiveness of early generation testing as applied to a recurrent selection program in peanut. Peunur Sci. 17,44-47. Halward, T. M., LaRue, E. A., and Kochert, G. (1991a). Genetic variation detectable with molecular markers among unadapted germplasm resources of cultivated peanut and related wild species. Genome 34, 1013-1020. Halward, T. M., Wynne. J. C., and Stalker, H. T. (1991b). Recurrent selection progress in a population derived from an interspecific peanut cross. Euphyricu 52, 79-84. Halward, T. M., Stalker, H. T., LaRue, E., and Kochert, G. (1992). Use of single-primer DNA amplifications in genetic studies of peanut (Aruchis hypogueu L.).Plunr Mol. Biol. 18, 315325. Halward, T. M., Stalker. H. T., and Kochert, G . (1993). Development of an RFLP linkage map in diploid peanut species. Theorer. Appl. Genet. 87, 379-384. Hammons, R. 0. (1953). Aruchis hypogueu: behavior of the induced mutant Cup. J. Elishu Mitchell Sci. SOC. 69, 84-85 (Abstr.). Hammons, R. 0. (1964). Krinkle, a dominant leaf market in the peanut, Aruchis hypogueu L.Crop Sci. 4, 22-24. Hammons, R. 0. (1976). Peanuts: Genetic vulnerability and breeding strategy. Crop Sci. 16, 527530.

436

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

Harris, D., Matthews, R. B., Nageswara Rao, R. C.. and Williams, J. H. (1988). The physiological basis for yield differences between four genotypes of groundnut (Arachis hypogaea L.) in response to drought. 3. Developmental processes. Exp. Agric. 24, 215-226. Hashim, 1. B., Koehler, P. E., and Eitenmiller, R. R. (1993). Tocopherols in runner and Virginia peanut cultivars at various maturity stages. J. Am. Oil Chem. SOC. 70,633-635. Hazra, S., Sthaye, S. S., and Mascarenhas, A. F. (1989). Direct somatic embryogenesis in peanut (Arachis hypoguea). BiolTechnology 7, 949-95 1. Hildebrand, P. E. (1984). Modified stability analysis of farmer managed, on-farm trials. Agron. J. 76,271-274.

Holbrook, C. C. (1990). Utility of early generation diallel analysis for predicting parental potential for yield and yield components in peanuts. Peanur Sci. 17, 9-1 I . Holbrook, C. C., and Branch, W. D. (1989). Additional locus with a recessive allele for red testa color in peanut. Crop Sci. 29, 312-314. Holbrook, C. C., and Noe, J. P. (1992). Resistance to the peanut root-knot nematode (Meloidogyne arenaria) in Arachis hypogaea. Peanut Sci. 19, 35-37. Holbrook, C. C., Kvien, C. S., and Branch, W. D. (1989). Genetic control of peanut maturity as measured by the hull-scrape method. Oleagineux 44, 359-364. Holbrook, C. C., Anderson, W. F.,and Pittman, R. N. (1993). Selection of a core collection from the U.S. germplasm collection of peanut. Crop Sci. 33, 859-861. Holbrook, C. C., Hunter, J. E., Knauft, D. A., Wilson, D. M.. and Matheron, M. E. (1994). Fatty acid composition as a possible mechanism for resistance to preharvest aflatoxin contamination of peanut. Proc. Am. Peanut Res. Educ. SOC. 26, (in press). Holley, R. N., and Wynne, J. C. (1986). Effectiveness of stratified mass selection for yield in intrasubspecific and intersubspecific crosses in peanut. Peanur Sci. 13, 33-35. Hunshal, C. S., Viswanath, D. P., Chimmad, V. P., and Cali, S. K. (1991). Performance of groundnut genotypes under saline water irrigation. j . Mahurashtra Agnc. Uniu. i b , I ! 6 - 1 !5 ICRISAT. (1989). Annual Report of the International Crops Research Institute for the Semi-And Tropics. ICRISAT, Patancheru, India. ICRISAT. (1992). “Descriptors for groundnut.” Intl. Board Plant Genet. Resour., Rome. Iroume, R. N., and Knauft, D. A. (1987a). Heritabilities and genetic, environmental, and phenotypic correlations for pod yield and leafspot severity in peanuts (Arachis hypogaea L.): Implications for early generation selection. Peanut Sci. 14, 46-50. Iroume, R. N., and Knauft, D. A. (1987b). Selection indices for simultaneous selection for pod yield and leafspot resistance in peanuts (Arachis hypogaea L.). Peanut Sci. 14, 51-54. Isleib, T. G., and Wynne, J. C. (1983a). Heterosis in testcrosses of 27 exotic peanut cultivars. Crop Sci. 23, 832-841. Isleib, T. G., and Wynne, J. C. (1983b). F4 bulk testing in testcrosses of 27 exotic peanut cultivars. Crop Sci. 23,841-846. Isleib, T. G., and Wynne, J. C. (1992). Use of plant introductions in peanut improvement. I n “Use of Plant Introductions in Cultivar Development, Part 2,” Publ. No. 20, pp. 75-116. CSSA. Madison, WI. Isleib, T. G., Wynne, J. C., and Nigam, S. N. (1994). Groundnut breeding. In’TheGroundnut Crop; A Scientific Basis for Improvement” (I. Smartt, ed.), pp. 552-623. Chapman & Hall, London. Jakkula, L. R., Knauft, D. A., and Gorbet, D. W. (1993). Characterization of a shriveled seed mutant in peanut. h c . Am. Peanur Res. Educ. SOC. 25, 36 (Abstr.). Jambunathan, R.. Raju, S. M., and Barde, S. P. (1985). Analysis of oil content of groundnuts by Nuclear Magnetic Resonance Spectrometry. J. Sci. Food Agric. 36, 162-166. Jambunathan, R., Gurtu, S., Raghunath, K., Kannan, S., Sridhar, R., Dwivedi, S. L., and Nigam, S. N. (1992). Chemical composition and protein quality of newly released ground (Arachis hypogaea L.) cultivars. J . Sci. Food Agric. 59, 161-167.

PEANUT BREEDING AND GENETICS

43 7

Jambunathan, R.. Singh, A. K.. Gum, S., and Ragbunath, K. (1993a). Amino acid composition, fatty acid composition, and levels of protease inhibitors in seeds of wild Arachis species. Oleagineu 48, 415-419. Jambunathan, R.. Sridhar. R., Raghunath, K., Dwivedi, S. L., and Nigam, S. N. (1993b). Oil quality characteristics and headspace volatiles of newly released groundnut (Arachis hypogaea L.) cultivars. J. Sci. Food Agric. 61, 23-30. Jogloy, S.. Wynne. J. C., and Beute, M. K. (1987). Inheritance of late leafspot resistance and agronomic traits in peanut. Peanur Sci. 14, 86-90. Johnson, W. C. (1987). The hull-scrape method to assess peanut maturity. Univ. of Georgia Coop. Ext. Ser. Bull. No. 958. Joshi, P. K., Kulkami, J. H.. and Bhatt, D. M. (1990). Interaction between strains of Brady rhizobium and groundnut (Arachis hypogaea L.) cultivars. Trop. Agric. (Trinidad) 67, 115118.

Kalra, V. K., and Singal, S. K. (1991). Screening of groundnut, Arachis hypogaeo Linn. cultivars/varieties against red flour beetle (Tribolium castaneum Herbst). J. Insect Sci. 4, 80-8 1. Keenan, J. I . , and Savage, G. P. (1994). Mycotoxins in groundnuts, with special reference to aflatoxin. In ‘The Groundnut Crop; A Scientific Basis for Improvement” (J. Smartt. ed.), pp. 509-551. Chapman & Hall, London. Kesava Rao, P. S., Tilak, K. V. B. R., and Arunachalam. V. (1990). Genetic variation for VA mycorrhiza-dependent phosphate mobilization in groundnut (Arachis hypogaea L.). Planr Soil 122, 137-142. Ketring, D. (1992). Physiology of oil seeds. X. Seed quality of peanut genotypes as affected by ambient storage temperature. Peanur Sci. 19, 72-77. Khalfaoui. J. L. (1990). Htrkditk de la pdcocitt extrEme dans le cas d’un croisement entre deux varittts d’arachide Spanish. Oleagineux 45, 419-436. Khalfaoui, J. L. (1991). Inheritance of seed dormancy in a cross between two Spanish peanut cultivars. Peanut Sci. 18, 65-67. Khalfaoui, J.-L. B., and Havard, M. (1993). Screening peanut cultivars in the field for root growth: a test by herbicide injection in the soil. Field Crops Res. 32, 173-179. Kim, N., Kim, Y.J., and Nam, Y.I. (1992). Characteristics and functional properties of protein isolates from various peanut (Arachis hypogaea L.) cultivars. J. Food Sci. 57, 406-410. Knauft, D. A. (1984). Genotype X cropping system interactions of peanuts grown as a sole crop and in association with maize. Trop. Agric. 61, 12-16. Knauft, D. A. (1987). Inheritance of rust resistance in groundnut. I n “Groundnut Rust Disease. Proceedings of a Discussion Group Meeting. ICRISAT,” pp. 183-187. ICRISAT, Patancheru, India. Knauft, D. A., and Gorbet, D. W. (1989). Genetic diversity among peanut cultivars. Crop Sci. 29, 14 17- 1422.

Knauft, D. A,, and Gorbet, D. W. (1990). Variability in growth characteristics and leafspot resistance parameters of peanut lines. Crop Sci. 30, 169-175. Knauft, D. A., and Gorbet, D. W. (1991). Agronomic performance and genetic shifts of genotype mixtures in peanut. Euphyrica 52, 85-90. Knauft, D. A.. and Gorbet, D.W. (1993). Consistency of rank correlations of peanut breeding lines for market grade characteristics and yield. Crop Sci. 33, 697-699. Knauft, D. A., and Ozias-Akins, P. (1995). Recent methodologies for germplasm enhancement and breeding. I n “Advances in Peanut Science” (H. E. Pattee and H. T.Stalker, eds.), in press. Am. Peanut Res. Educ. Soc., Ahoskie, NC. Knauft, D. A.. Norden, A. J., Gorbet, D. W., and Martin, F. G. (1986a). Stability of market quality factors in peanut (Arachis hypogaea L.). Soil Crop Sci. Soc. FL Proc. 46, 72-74. Knauft, D. A., Norden, A. I., and Gorbet, D. W. (1986b). The effect of three digging dates on oil

438

D.A. KNAUFT AND J. C . WYNNE

quality. yield, and grade of five peanut genotypes grown without leafspot control. Peanut Sci. 13, 82-86. Knauft, D. A., Norden, A. J., and Gorbet, D. W. (1987). Peanut breeding. I n “Principles of cultivar development” (W. R. Fehr, ed.), Vol. 2, pp. 346-384. Macmillan Publ. Co., New York. Knauft, D. A., Gorbet, D. W., and Norden, A. J. (1988). Yield and market quality of seven peanut genotypes grown without leafspot control. Peanut Sci. 15, 9-13. Knauft, D. A., Colvin. D. L., and Gorbet, D. W. (1990). Effect of paraquat on yield and market grade of peanut (Aruchis hypogueu) genotypes. Weed Technol. 4, 866-870. Knauft, D. A., Branch, W. D., and Gorbet. D. W. (1991). TWOdominant genes for white testacolor in peanut. J. Hered. 81, 73-75. K ~ u f t ,D. A., Chiyembekeza, A. J., and Gorbet, D. W. (1992). Possible reproductive factors contributing to outcrossing in peanut. Peunur Sci. 19, 29-31. Knauft, D. A,, H o l b m k , C. C., and Gorbet, D. W. (1993a). Use of mass selection for developing leafspot-resistant peanut lines. P m . Am. Peunur Res. Educ. SOC. 25, 27 (Abstr.). Knauft, D. A., Moore, K. M., and Gorbet, D. W. (1993b). Further studies on the inheritance of fatty acid composition in peanut. Peunur Sci. 20, 74-76. Kochert, G., Halward, T., Branch, W. D., and Simpson, C. E. (1991). RFLP variability in peanut (Aruchis hypogueu L.) cultivars and wild species. Theorer. Appl. Gener. 81, 565-570. Kornegay, J. L., Beute, M. K., and Wynne, J. C. (1980). Inheritance of resistance to Cercosporu uruchidicolu and Cercosporidiwn personarum in six Virginia type peanut lines. Peanut Sci. 7, 4-9.

Krapovickas, A., and Gregory, W. C. (1994). Taxonomia del genero Aruchis (Leguminosae). BOplandiU 8(1-4), 1-186. Kvien, C. K., and Ozias-Akins, P. (1991). Lack of monocarpic senescence in Florunner peanut. feunur Sci. 18, 86-90. Kvien, C. K., Culbreath, A. K., Wilcut, J. W., Brown, S. L., and Bell, D. K. (1993). Peanut production in systems restricting use of pesticides based on carcinogenicity or leachability. feunur Sci. 20, 118-124. Lacks, G. D., and Stalker, H. T. (1993). Isozyme analyses of Aruchis species and interspecific hybrids. Peunur Sci. 20, 76-81. Lacorte, C., Mansur, E., Timmerman, B., and Cordeiro, A. R. (1991). Gene transfer into peanut (Aruchis hypogueu L.) by Agrobucrerium rumefuciens. Plunr Cell Rep. 10, 354-357. Lanham, P. G., Fennell, S., Moss, J. P.,and Powell, W. (1992). Determination of polymorphic loci in Aruchis germplasm using randomly amplifieh polymorphic DNAs. Genome 35, 885889.

Lewis, P. I., and Filonow, A. B. (1990). Reaction of peanut cultivars to pythium pod rot and their influence on populations of Pyrhium spp. in soil. Peunur Sci. 17, 90-95. Li, Z., Jarret, R. L., Pittman, R. N., Dunbar, K. B., and Demski, J. W. (1993). Efficient plant regeneration from protoplasts of Aruchis puruguuriensis Chod. et Hassl. using a nurse culture method. Plunr Cell Tissue Orgun Culr. 34, 83-90. Liao, B. S., and Wang, Y. Y. (1988). Inheritance of resistance to rust (Pucciniu uruchidis Speg.) in peanuts (Aruchis hypogueu L.). Oil Crops China 3, 13-17. Loesch, P. J., Jr., and Hammons, R. 0. (1968). A radiation breeding experiment with peanuts. 111. Inheritance of macromutants (NC 4-18.5 kR). Rudiur. Bor. 8, 94-108. Lynch, R. E. (1990). Resistance in peanut to major arthropod pests. Florida Enromol. 73,422-445. Lynch, R. E., and Mack, T. P. (1995). Biological and biotechnical advances for insects associated with peanut. In “Advances in Peanut Science” (H. E. Pattee and H.T.Stalker, eds.), in press. Am. Peanut Res. Educ. Soc., Ahoskie, NC. Manoharan, V.,Ramalingam. R. S.. and Kalaimani, S. (1990). Genetic advance and path analysis in the F, generation of an intrasubspecific cross in groundnut. I n d . J. Gener. SO, 244-247.

PEANUT BREEDING AND GENETICS

439

Mansur, E. A., Lacorte, C., de Freitas, V. G., de Oliveira, D. E., Timmerman, B., and Cordeiro, A. R. (1993). Regulation of transformation efficiency of peanut (Aruchis hypogueu L.) explants by Agrobucferium tumefuciens. Plant Sci. 89, 93-99. Matlock. R. S., Banks, D. J., Tai, Y. P., and Kirby, J. S. (1970). Inheritance of a narrow-leaflet character in peanuts, Aruchis hypogaeu L. Agron. Absrr., 15. Matthews, R. B., Harris, D., Nageswara Rao, R. C., Williams, J. H.,and Wadia, K. D. R. (1988a). The physiological basis for yield differences between four genotypes of groundnut in response to drought. 1. Dry matter production and water use. Erp. Agric. 24, 191-202. Matthews, R. B.,Harris,D., Williams, J. H., and Nageswara Rao, R. C. (1988b). The physiological basis for yield differences between four genotypes of groundnut in response to drought. 11. Solar radiation and leaf movement. Exp. Agric. 24, 203-213. Mbata, G. N. (1992). The use of resistant crop varieties in the control of storage insects in the tropics and subtropics. Ambio 21,475-478. McKently, A. H. (1991). Direct somatic embryogenesis from axes of mature peanut embryos. In Vitro 27, 197-200. McKently, A. H., Moore, G. A., and Gardner, F. P. (1990). In v i m plant regeneration of peanut from seed explants. Crop Sci. 30, 192-196. McKently, A. H.,Moore, G. A., Wang, A., and Doostdar, H. (1993). Agrobucrerium-mediated transformation of peanut zygotic embryos and regeneration of transgenic plants. In Vim 29A, 65A (Abstr.). Mehan, V. K., Reddy, D. D. R., and McDonald, D. (1993). Resistance in groundnut genotypes to Kalahasti malady caused by the stunt nematode, Tylenchorhynchus brevilineurus. Inr. J. Pest M g t . 39, 201-203. Melouk, H. A,, Banks, D. J., and Fanous, M. A. (1984). Assessment of resistance to Cercosporu uruchidicolu in peanut genotypes in field plots. Plunr Dis. 68, 395-397. Melouk, H.A,, Akem, C . N., and Smith, 0. D. (1989). Reaction of peanut genotypes to Sclerotinia blight in field plots, 1986 and 1987. Eiol. Culr. Tesrs Conrrol Plant Dis. 4, 39. Mercer, L. C., Wynne, J. C., and Young, C. T. (1990). Inheritance of fatty acid content in peanut oil. Peanut Sci. 17, 17-21. Michaels, T. E. (1988). Effect of selection for emergency and maturity on yield of Ontario peanuts. Peunur Sci. 15, 69-72. Middleton, K., and Shorter, R. (1987). Occurrence and management of groundnut rust in Australia. In “Groundnut Rust Disease. Proceedings of a Discussion Group Meeting. 24-28 September, (1984);’ pp. 73-75. ICRISAT, Patancheru, India. Mohamed, H. A., Clark, J. A., and Ong; C. K. (1988a). Genotypic differences in the temperature responses of tropical crops. I. Germination characteristics of groundnut (Aruchis hypogueu L.) and pearl millet (Penniserum ryphoides S . & H.). J. Exp. Eor. 39, 1121-1 128. Mohamed, H.A., Clark, J. A., and Ong, C. K. (1988b). Genotypic differences in the temperature responses of tropical crops. 11. Seedling emergence and leaf growth of groundnut (Aruchis hypogueu L.) and pearl millet (Penniserum ryphoides S. & H.). J. Exp. Eot. 39, 1129-1135. Monteverde-Penso, E. J.. and Wynne. J. C. (1988). Evaluation of three cycles of recurrent selection for fruit yield within a population of Virginia-type peanut. Crop Sci. 28, 75-78. Monteverde-Penso, E. J . , Wynne, J. C., Isleib, T. G., and Mozingo, R. W. (1987). A comprehensive breeding procedure utilizing recurrent selection for peanuts. Peunur Sci. 14, 1-3. Moore, K. M., and Knaufi, D. A. (1989).The inheritance of high oleic acid in peanut. J. Hered. 80, 252-253. Moss, J. P., and Stalker, H. T.(1987). Embryo rescue in wide crosses in Aruchis. 3. In v i m culture of peg tips of A. hypogueu selfs and interspecific hybrids. Peunut Sci. 14, 70-74. Mouli, C., and Kale, D. M. (1989). Early-maturing mutants in ‘groundnut cultivar Phule-Pragati (JL24). Curt-. Sci. 58, 690-692.

440

D.A.KNAUF”ANDJ.C.WYNNE

Mugnier, J. (1988). Establishment of new hairy root lines by inoculation with Agmbacrerium rhizogenes. Plant Cell Rep. 7, 9-12. Murthy, T. G. K.. and Reddy, P. S. (1993). “Cytogenetics and Genetics of Groundnut.” Intercept Ltd., Andover, England. Myer, R. O., Johnson, D. D.. Knauft, D. A., Gorbet, D. W., Brendemuhl, 1. H.,and Walker, W. R. (1992). Effect of feeding high-oleic peanuts to growing-finishing swine on resulting carcass fatty acid profile, and on carcass and meat quality characteristics. J. Animal Sci. 70, 37343741.

Nagabhushanam, G. V. S., and Prasad, M. V. R. (1992). Selection criteria for groundnut (Arachis hypogaea L.)W i g . Oleagineux 47, 23-27. Nagabhushanam, G. V. S., prasad, M. V. R., and Jagadish, C. A. (1992). Studies on association among canopy and reproductive attributes in F, and F2 generations of intraspecific and intersubspecific crosses in groundnut (Arachis hypogaea L.). J. Gener. Breed. 46, 155-162. Nageswara Rao, R. C., Williams, J. H.,and Singh, M. (1989). Genotypic resistance to drought and yield potential of peanuts. Agmn. J. 81, 877-881. Nageswara Rao, R. C., Williams, J. H.,Wadia, K. D. R., Hubick, K. T., and Farquhar, G. D. (1993). Crop growth, water-use efficiency, and carbon isotope discrimination in groundnut (Arachis hypogaeu L.) genotypes under end-of-season drought conditions. Ann. Appl. Biol. 122, 357-367.

Nageswara Rao, R. C.. Udaykumar, M., Farquhar, G. D., Talwar, H.S., and Prasad, T. G. (1994). Variation in carbon isotope discrimination and its relation with specific leaf area and ribulose-1-5-bisphosphatecarboxylase content of groundnut genotypes. Ausr. J . P lanr Physiol., (in press).

N’Doye, 0.. and Smith, 0. D. (1992). Flowering pattern and fruiting characteristics of five short growth duration peanut lines. Oleagineux 47, 235-240. N’Doye, O., and Smith, 0. D. (1993). A note on the earliness of offspring from crosses among five short growth-duration peanut lines. Peanut Sci. 20, 132-137. Ndunguru. B. J.. and Williams, J. H.(1993). The impact of varying levels of competition from pearl millet on the yields of groundnut cultivars. Exp. Agric. 29, 29-37. Nigam, S. N., and Bock, K. R. (1990). Inheritance of resistance to groundnut rosette virus in groundnut (Arachis hypogaea). Ann. Appl. Biol. 117, 553-560. Nigam, S. N., Arunachalam, V., Gibbons, R. W., Bandyopadhyay, A., and Nambiar, P. T. C. (1980a). Genetics of non-nodulation in groundnut (Arachis hypogaea L.). Oleagineux 35,453455.

Nigam, S. N., Dwivedi, S. L., and Gibbons, R. W. (1980b). Groundnut breeding at ICRISAT. In “Roceedings of an International Workshop on Groundnuts,” pp. 62-80. ICRISAT, Patanchem, India. Nigam, S. N., Nambiar, P. T. C., Dwivedi, S. L.. Gibbons, R. W., and Dart, P. J. (1982). Genetics of non-nodulation in groundnut (Arachis hypogaea L.): Studies with single and mixed Rhizobium strains. Euphyrica 31, 691-693. Nigam, S. N., Dwivedi, S. L.,Sigamani, T.S. N., and Gibbons, R. W. ( I 984). Character association among vegetative and reproductive traits in advanced generation of intersubspecific and intrasubspecific crosses in peanut. Peanut Sci. 11, 95-98. Nigam, S. N., Dwivedi. S. L.,Nagabhushanam, G. V. S.. and Gibbons, R. W. (1988). Inheritance of period from seedling emergence to first flowering in peanut (Arachis hypogueu L.). J. Oilseeds Res. 5 , 101-106. Nigam, S. N., Vasudeva Rao, M. J., and Gibbons, R. W. (1990). Artificial hybridization in groundnut. Information Bulletin No. 29, International Crops Research Institute for the Semi-Arid Tropics, Patanchem, India. Nigam, S. N., Dwivedi, S. L., and Gibbons, R. W. (1991). Groundnut breeding: constraints, achievements, and future possibilities. Planr Breed. Absrr. 61, 1127-1 136.

PEANUT BREEDING AND GENETICS

441

Noe, J. P., Holbrook, C. C.. and Minton, N. A. (1992). Field evaluation of susceptibility to Meloidogyne arenaria in Arachis hypogaea plant introductions. J. Nemarol. 24, 712-716. Norden, A. J. (1973). Breeding of the cultivated peanut. I n “Peanuts-Culture and Uses,” pp. 175208. Am. Peanut Res. Educ. Soc.. Yoakum, TX. Norden, A. J. (1980). Peanut. I n “Hybridization of Crop Plants,’’ pp. 443-456. Am. Soc. Agron., Madison, WI. Norden, A. J., Smith, 0. D., and Gorbet, D. W. (1982). Breeding of the cultivated peanut. I n “Peanut Science and Technology” (H. E. Pattee and C. T. Young, eds.), pp. 95-122. Am. Peanut Res. Educ. Soc.,Yoakum, TX. Norden, A. J., Gorbet, D. W., Knauft, D. A., and Martin, F. G. (1986). Genotype X environment interactions in peanut multi-line populations. Crop Sci. 26, 46-48. Norden, A. J., Gorbet, D. W., Knauft, D. A., and Young, C. T. (1987). Variability in oil quality among peanut genotypes in the Florida breeding program. Peanur Sci. 14, 7- 11. Norden, A. J., Knauft, D. A., and Gorbet, D. W. (1988). A dominant gene for white seedcoat in peanut (Arachis hypogaea L.). J. Hered. 79, 212-214. O’Hara, G. W., Hartzook, A., Bell, R. W., and Loneragan, I. F. (1993). Differences between Bradyrhizobium strains NC92 and TAL lo00 in their nodulation and nitrogen fixation with peanut in iron deficient soil. Plant Soil 155-156, 333-336. O’Keefe, S. F., Wiley, V. A., and Knauft, D. A. (1993). Comparison of oxidative stability of high and normal oleic peanut oils. J. Am. Oil Chem. SOC. 70, 489-492. Olorunju, P. E., Kuhn, C. W., Demski, J. W., Misari, S. M., and Ansa, 0. A. (1992). Inheritance of resistance in peanut to mixed infections of groundnut rosette virus (GRV) and groundnut rosette assistor virus and a single infection of GRV. Planr Dis. 76, 95-100. Ozias-Akins, P. (1989). Plant regeneration from immature embryos of peanut. Planr Cell Rep. 8, 217-21 8. Ozias-Akins, P., Anderson, W. F., and Holbrook, C. C. (1992a). Somatic embryogenesis in Aruchis hypoguea L.: genotype comparison. Plant Sci. 83, 103- 1 I 1. Ozias-Akins, P., Anderson, W. F., Schnall, J. A., Singsit, C., Clemente, T. E., and Weissinger, A. K. (1992b). Regeneration of transgenic shoots from long-term embryogenic cultures of Aruchis hypogaea. P m . Am. Peanut Res. Educ. Soc. 24, 19 (Abstr.). Ozias-Akins, P., Singsit. C., and Branch, W. D. (1992~).Interspecific hybrid inviability in crosses of Aruchis hypogaea X A. srenospenna can be overcome by in virm embryo maturation or somatic embryogenesis. J. Plant Physiol. 140, 207-212. Ozias-Akins, P., Schnall, J. A., Anderson, W. F., Singsit, C . , Clemente, T. E., Adang, M. J., and Weissinger, A. K. (1993). Regeneration of transgenic peanut plants from stably transformed embryogenic callus. Planr Sci. 93, 185-194. Patil, S . H., and Mouli, C. (1975). Genetics of a dwarf mutant in groundnut. Theorer. Appl. Gener. 46,395-400. Patil, S . H., and Mouli, C. (1984). Preferential segregation of two allelic mutations for small leaf character in groundnut. Theorer. Appl. Gener. 67, 327-332. Patra, G. J., Singh, B. R., Mohanty, J. P., and Sahu, B. C. ( I992a). Effect of population diversity on yield stability of Spanish groundnut (Arachis hypogaea subsp. Jasrigiara var. vulgaris). Ind. J. Agric. Sci. 62, 41-46. Patra, G. J., Singh, B., and Patra, P. (1992b). Effectiveness of selection on character association in groundnut (Arachis hypogaea). Ind. J. Agric. Sci. 62,461-465. Pattee, H. E., and Giesbrecht, F. G. (1990). Roasted peanut flavor variation across germplasm sources. Peanur Sci. 17, 109-112. Pattee, H. E., and Stalker, H. T. (1992). Embryogenesis in reciprocal crosses of Arachis hypogaea cv NC6 with A. duranensis and A. srenospenna. Inr. J. Plant Sci. 153, 341-347. Pattee, H. E., Stalker, H.T., h d Moss, J. P. (1988). Embryo rescue in wide crosses in Arachis. 2. Embryo development in cultured peg tips of Arachis hypogaea. Ann. Bor. 61, 103-1 12.

442

D. A. K N A m AND J. C . WYNNE

Pattee, H.E., Giesbrecht, F. G., and Mozingo, R. W. (1993). A note on broad-sense heritability of selected sensory descriptors in Virginia-type Aruchis hypogueu L. Peunur Sci. 20, 24-26. Porter, D. M.. Coffelt, T. A., Wright, F. S., and Mozingo, R. W. (1992). Resistance to sclerotinina blight and early leafspot in Chinese peanut germplasm. Peanut Sci. 19, 41-43. Phillips, T. D., Wynne, J. C., Ekan, G. H.,and Schneeweis, T. J. (1989). Inheritance of symbiotic nitrogen fixation in two peanut crosses. Peanut Sci. 16, 66-70. Pittman, R. N.,Johnson, B. B., and Banks, D. J. (1984). In v i m differentiation of a wild peanut, Aruchis villosulicarpa Hoehne. Peunur Sci. 11, 24-27. Radhakrishnan, T.,Murthy. T. G. K., and Sen, P. (1991). Inheritance of five morphological traits in groundnut. J. Cyrol. Genet. 26, 5-8. Raheja, R. K., Batta, S. K., Ahuja, K. L., Labana, K. S., and Singh, M. (1988). Comparison of oil content and fatty acid composition of peanut genotypes differing in growth habit. Qual. P l m PI. Food Hum. NuW. 37, 103-108. Rattunde, H. F., Ramraj, V. M., Williams, J. H., and Gibbons, R. W. (1988). Cultivar mixtures: a means of exploiting morpho-developmental differences among cultivated groundnuts. Field Crops Res. 19, 201-210. Raut, S. S., Birari, S. P., and Jamadagni, B. M. (1993). Genotype x environment interaction in bunch-erect group of groundnut (Aruchis hypogueu). Ind. J. Agric. Sci. 63, 23-26. Reddy, K. B., Reddy, D. S., and Reddy, C. M. (1991). Evaluation of groundnut genotypes for phosphorus use efficiency. I d . 1. Planr Physiol. 34, 228-234. Reddy, K. B., Ashalatha, M., and Venkaiah, K. (1993). Differential responses of groundnut genotypes to Fe-deficient stress. J. Plunr Nurr. 16, 523-531. Reddy, L. I., Nigam, S. N., Dwivedi, S. L., and Gibbons, R. W. (1987). Breeding groundnut cultivars resistant to rust (Pucciniu uruchidis Speg.). I n “Groundnut Rust Disease. Proceedings of a Discussion Group Meeting. 24-28 September (1984),” pp. 17-25. ICRISAT, Patancheru, India. Reddy, M. V., Reddy, D. C., Reddy, B. J., Reddy. D. R., and Mahalakshmi, A. (1988). Reciprocal cross differences in a 6 X 6 diallel set of groundnut (Aruchis hypogueu L.) in the F, generation. Euphyricu 38, 24 1-245. Reddy, P. S. (1988). Genetics, breeding and varieties. I n “Groundnut” (P. S. Reddy, ed.), pp. 200317. Indian Council of Agric. Res., New Delhi. Rhoads, F. M., Shokes, F. M., and Gorbet, D. W. (1989). Response of two peanut cultivars to soil zinc levels. Univ. FL Inst. Food Agric. Sci., Res. Rpt. NF-89-2. Ro, 0. G., Smith, R. L., and Knaufi, D. A. (1992). Restriction fragment length polymorphism evaluation of six peanut species within the Aruchis section. Theom. Appl. Gener. 84,201-208. Salunkhe, D. K., Chavan, J. K., Adsule, R. N., and Kadam, S. S. (1992). Peanut. In “World Oilseeds. Chemistry, Technology, and Utilization,” pp. 140-216. Van Nostrand Reinhold Publ. Co., New York. Savage, G. P., and Keenan, J. I. (1994). The composition and nutritive value of groundnut kernels. In ‘The Groundnut Crop; A Scientific Basis for Improvement” (J. Smartt, ed.), pp. 173-213. Chapman & Hall, London. Schilling, T. T., Mozingo, R. W., Wynne, J. C., and Isleib, T. G.(1983). A comparison of peanut multilines and component lines across environments. Crop Sci. 23, 101-105. Seaton, M. L., Coffelt, T. A,, and Van Scoyoc, S. W. (1992). Comparison of vegetative and reproductive traits of 14 peanut cultivars. Oleugineux 47, 471-474. Sellars, R. M., Southward, G. M.. and Phillips, G. C. (1990). Adventitious somatic embryogenesis from cultured immature zygotic embryos of peanut and soybean. Crop Sci. 30, 408-414. Senaratne, R., and Ratnasinghe, D. S. (1993). N supply by groundnuts to maize in a maize plus groundnut intercropping system, as affected by the genotype. Biol. Ferril. Soils 15, 215-219. Shenvood, J. L., Beute, M., Dickson, D. W., Elliott, V.,Nelson, R. S., Opperman, C., and Shew,

PEANUT BREEDING AND GENETICS

443

B. (1995). Biological and biotechnological control advances in Aruchis diseases. I n “Advances in Peanut Science” (H. E. Pattee and H. T. Stalker, eds.), inpress. Am. Peanut Res. Educ. Soc., Ahoskie, NC. Shorter, R., and Butler, D. (1986). Effect of moving mean covariance adjustments on error and genetic variance estimates and selection of superior lines in peanuts (Aruchis hypogueu L.). Euphyricu 35, 185-192. Shorter, R., and Hammons, R. 0. (1985). Pattern analysis of genotype adaptation and genotype x environment interactions in the uniform peanut performance tests. Peunur Sci. 12, 35-41. Singh, V. (1991). Effect of natural infestation of Oryzuephilus surinurnensis Linn. and Tribolium custuneum Herbst on the kernel quality of groundnut varieties. J. Insecr. Sci. 4, 82-84. Singh, A. K., and Simpson, C. E. (1994). Biosystematicsand genetic resources. I n “The Groundnut Crop: A Scientific Basis for Improvement” (J. Smartt, ed.), pp. 96-137. Chapman and Hall, London. Smartt, J. (1994). The groundnut in farming systems and the rural economy-a global view. I n “The Groundnut Crop: A Scientific Basis for Improvement” (J. Smartt, ed.), pp. 664-699. Chapman and Hall, London. Smith, J. W., Posada, L., and Smith, 0. D. (1980). Greenhouse screening for resistance to the lesser corn stalk borer. Peanut Sci. 7, 68-71. Smith, 0. D., Boswell, T. E., Grichar, W. J., and Simpson, C. E. (1989). Reaction of select peanut (Aruchis hypogueu L.) lines to southern stem rot and pythium pod rot under varied disease pressure. Peanut Sci. 16, 9-14. Smith, 0. D., Aguirre, S. M., Boswell, T. E., Grichar, W. J., Melouk, H.A., and Simpson, C. E. (1990). Registration of TxAG-4 and TxAG-5 peanut germplasm. Crop Sci. 30, 429. Stalker, H. T., Campbell, W. V., and Wynne, J. C. (1984). Evaluation of cultivated and wild peanut species for resistance to the lesser cornstalk borer (Lepidoptera: Pyralidae). J. Econ. Enromol. 77,53-57. Still, P. E., Plata, M. I., Campbell, R. J., Bueno, L. C., Chichester, E. A., and Niblett, C. L. (1987). Regeneration of fertile Aruchis puruguuriensis plants from callus and suspension cultures. Plunr Cell Tissue Culr. 9, 37-43. Stirling, C. M., and Black, C. R. (1991). Stages of reproductive development in groundnut (Aruchis hypogueu L.) most susceptible to environmental stress. Trop. Agric. (Trinidud) 68, 296300. Subrahmanyam, P., and McDonald, D. (1987). Groundnut rust disease: Epidemiology and control. I n “Groundnut Rust Disease. Proceedings of a Discussion Group meeting. 24-28 September (1984),” pp. 27-39. ICRISAT, Patancheru, India. Subrahmanyam, P., Williams, J. H., McDonald, D., and Gibbons, R. W. (1984). The influence of foliar diseases and their control by selective fungicides on a range of groundnut (Aruchis hypogueu L.)genotypes. Ann. Appl. Biol. 104, 476-476. Subrahmanyam, P., Ghanekar, A. M., and Nolt, B. L. (1985). Resistance to groundnut diseases in wild Aruchis species. In “Proceedings of the International Workshop on Cytogenetics of Aruchis,” pp. 49-55. ICRISAT, Patancheru, India. Sumner, M. E.,, Kvien, C. S., Smal, H., and Csinos, A. S. (1988). On the calcium nutrition of peanut (Aruchis hypogaea L.).I. Operational model. J. Fertil. Issues 5 , 97-102. Taylor, S. L. (1992). Chemistry and determination of food allergens. Food Technol. 46(5), 146-152. Tiwari, S. P., and Khanorkar, S. M. (1984). Colchicine-induced true breeding miniature mutant in groundnut. Curr. Sci. 53, 1262- 1263. Tiwari, S . P., Ghewande, M. P., and Misra, D. P. (1984). Inheritance of resistance to rust and late leafspot in groundnut (Aruchis hypogueu L.). J. Cyrol. Gener. 19, 97-101. Tiwari, S . P., Murthy. T. G. K., George, K. J., and Reddy, P. S . (1990). Genetics of some leaflet traits and development of marker stocks in groundnut. Legume Res. 13, 197-199.

444

D. A. KNAUFT AND J. C . WYNNE

Todd, J. W., Beach, R. M., and Branch, W. D. (1991). Resistance in eight peanut genotypes to foliar feeding of fall armyworm, velvetbean caterpillar, and corn eanvorm. Peanur Sci. 18, 38-40. Upadhyaya, H. D., Gopal, K.,Nadaf, H. L., and Vijayakumar, S. (1992). Combining ability studies for yield and its components in groundnut. I d . J. Gener. 52, 1-6. USDA. (1992). pcGRIN. Germplasm Resources Information Network. Data query system for the PC. ARS-108. U.S. Govt. Rint. Office, Washington, D.C. USDA. (1994). Peanut area, yield, and production. Production estimates and crop assessment division, FAS, USDA. Varman, P. V., Rathinaswamy, R., Ramalingam, R. S., and Bhat, M. V. (1986). Inheritance of growth habit in groundnut. Madras Agric. J. 73, 299-300. Vasudeva Rao, M. J., Nigam, S. N., and Huda, A. K. S. (1992). The thermal time concept as a selection criterion for earliness in peanut. Peanur Sci. 19, 7-10. Ventner, C., DeWaele, D., and Meyer, A. J. (1993). Reproductive and damage potential of Difylenchus destrucror on six peanut cultivars. J. Nemarol. 25, 59-62. Vinod Prabhu, K.,Arunachalam, V., and Bandyopadhyay. A. (1990). Nonparametric approach to mukitreit selection for yield in groundnut (Arachis hypogaea L.). Theorer. Appl. Gener. 80, 223-227. Waliyar, F.. McDonald, D., Subba Rao, P. V., and Reddy, P. M. (1993). Components of resistance to an Indian Source of Cercospora arachidicola in selected peanut lines. Peanut Sci. 20,93-96. Walker, M. E., and Keisling, T. C. (1978). Response of five cultivars to gypsum fertilization on soils varying in calcium content. Peanut Sci. 5 , 57-60. Walls, S. B., and Wynne, J. C. (1985). Combining ability for resistance to Cercosporidium personarum for 5 late leafspot resistant peanut germplasm lines. Oleagineux 40, 389-394. Wanvick, D. R. N., and Demski, J. W. (1992). Factors influencing peanut stripe virus transmission in peanut seeds. Firoparol. Bras. 17, 389-392. Watterott, J. (1991). Root respiration of groundnut as influenced by temperature, circadian rhythm, nitrogen source, drought and genotype, p. 85. Dr. Agric. Thesis, Rheinischen FriedrichWilhelms-Universitat zu Bonn, FRG. Wehtje, G., Walker, R. H., Paterson. M. G., and McGuire, J. A. (1984). Influence of twin rows on yield and weed control in peanuts. Peanut Sci. 11, 88-91. Wells, M. A., Grichar, W. J., Smith, 0. D., and Smith, D. H. (1994). Response of selected peanut germplasm lines to leafspot and southern stem rot. Oleagineu 49, 21-26. Wells, R., Bi, T., Anderson, W. F., and Wynne, J. C. (1991). Peanut yield as a result of fifty years of breeding. Agron. J. 83, 957-961. Wightman, J. W., and Amin. P. W. (1988). Groundnut pests and their control in the semi-arid tropics. Tmp. Pesr Management 34, 218-226. Wightman, J. W., Dick. K. M., Ranga Rao, G. V., Shanower, T. G . , and Gold, C. G. (1990). Pests of groundnut in the semi-arid tropics. I n “Insect Pests of Food Legumes” (S. N. Singh, ed.), pp. 243-322. John Wiley and Sons, New York. Wilcut, J. W., York, A. C., Grichar, W. J., and Wehtje, G. R. (1995). Biological and biotechnological control advances for weeds. I n “Advances in Peanut Science” (H. E. Pattee and H. T. Stalker, 4s.). in press. Am. Peanut Res. Educ. Soc.. Ahoskie, NC. Wildman, L. G . , Smith, 0. D., Simpson, C. E..and ’kber, R. A. (1992). Inheritance of resistance to Sclerotinia minor in selected Spanish peanut crosses. Peanut Sci. 19, 31-35. Wright, G. C., and Bell, M. J. (1992). Plant population studies on peanut (Arachis hypogaea L.) in subtropical Australia. 2. Water-limited conditions. Ausr. J. Exp. Agric. 32, 189-196. Wright, G. C., Hubick, K. T.,and Farquhar, G. D. (1991). Physiological analysis of peanut cultivar response to timing and duration of drought stress. A m . J. Agric. Res. 42, 453-470. Wright, G . C., Nageswara Rao, R. C., and Farquhar, 0 . D. (1994). Water-use efficiency and carbon isotope discrimination in peanut under water deficit conditions. Crop Sci. 34, 92-97.

PEANUT BREEDING AND GENETICS

445

Wynne, J. C. (1976). Use of accelerated generation increase programs in peanut breeding. An accelerated recurrent selection program. Pmc. Am. Peanut Res. Educ. Assoc. 8,44-47. Wynne, J. C., and Emery, D. A. (1974). Response of intersubspecificpeanut hybrids to photoperiod. Crop Sci. 14, 878-880. Wynne, J. C., and Halward. T.(1989). Cytogenetics and genetics of Aruchis. Crir. Rev. P h i Sci. 8, 189-220. Wynne, 1. C., Beute. M. K., and Nigam, S. N. (1991). Breeding for disease resistance in peanut (Arachis hypogaea L.). Annu. Rev. Phytoparhol. 29, 279-303. Zeile, W. L., Knauft. D. A., and Kelly, C. B. (1993). A rapid non-destructivetechnique for fatty acid determination in individual peanut seed. Peanut Sci. 20, 9- 11.

This Page Intentionally Left Blank

Index A Absorption edge, in XAS, 16-21 Accessions cotton, and Germplasm Resources Information Network, 136 peanut, multivariate analyses, 398-400 phenotype, potential merit, 272 ACEDB, based genomic databases, 152-154 Addresses, for plant genome database questions, I52 Adsorbates, at aqueous-mineral interface, 2225 Advanced Light Source soft X-ray source, 15 techniques for lighter elements, 37 Advanced Photon Source, hard X-ray source, 15 AGRICOLA bibliographic database, 149- 151 Agricultural land, sewage sludge-amended, 345-387 Agricultural nonpoint source model, 101 Agricultural Research Service, and National Research Initiative, 114-1 15 Agricultural wastes, as source of organic chemicals to soils, 367-368 Agriculture, crop-based, 269-270 Agronomy, GIs, 67-105 Airborne visiblehfrared imaging spectrometer, 77-81 n-Alkanes, detection by Py-HMS and PyFDMS, 176-191 n-Alkylbenzenes, series in Armadale humin, 20 1 Alkyltin, hydrophobicity, 240 Allergen activity, peanut, 428 Aluminum Al,, species, 49 XANES, 22 Aluminum oxides, surface hydroxyls, 242-243 Anonymous ftp, and plant genome database, 148-154 Aqueous-mineral interface, adsorbates at, 2225

Aquifer, migration of humic substances, 25025 1 Arubidopsis spp., and Erussicu, comparative mapping, 133- I35 Aruchis spp. distribution and taxonomy, 394-3% genetic variability, 397-403 production and breeding, 3%-397 vegetative and reproductive traits, 403-405 ARUINFO, GIS software, 103 Armadale Bh horizon, supercritical CO, extracts, 188- I 9 0 Armadale soil Py-FIMS, 202-203 Py-GUMS, 192-194 Arsenate adsorption, EXAFS studies, 24-25 Aspergillusflavus. affecting peanut species, 424 Atmospheric deposition 1 ,4-DCB, 355-356 as inputs of organic chemicals to soils, 367368 Atrazine applications, half-lives, 385 Auger electron yield, in XAS, 18

B Back-crossing, see ulso Out-crossing rates advantages of DNA markers, 315-317 genes for qualitative and quantitative traits, 306 incorporation of trait via, 428 marker-assisted, 278-279, 315-318 peanut, 410 Bainsville soil, Py-GUMS analysis, 192- 194 Barley clustered multilocus QTL, 310 genome analysis, 120-121 Beam lines facility access according to, 55-56 on hard X-ray rings, 14-15 in synchrotron, 3-6, 9 wiggler, in microdiffraction, 45 X-26A, at NSLS, 35

447

IND

448

Behavior assessment models, for organic chemic a l ~ 373-376 , Bending magnets energy distribution, 10-1 1 X-ray pattern produced by, 8-9 Benzo[a]pyrene, persistence, 365-366 Benzo[ghi]perylene, persistence, 362-366 Beryllium windows, in beamline, 14-15 Bioavailability, trace elements in aquatic system, 254 Biomolecules single crystal diffiction, 46-48 XAS, 26 Boehmite. sorption sites, 242-243 Bonding environments, deduction in XANES, 20 Bone fragments, contaminant distribution, SXRF, 36-37 Bradyrhizobium, peanut colonized by, 404, 420-42 1 Brassica spp. and Arabidopsis, comparative mapping, 133135 genetic maps, 129-135 Breaking radiation, see Bremssblahlung radiation Breeding, see also Plant breeding programs back-cross, advantages of DNA markers, 315-317 maize epistatic effects, 304-305 and hybrid combinations, 279-28 I peanut, 393-430 plant genetic heterogeneity in, 307-308 heterotic group in, 279-281 off-season nurseries in, 267 and plant biology, DNA markers as links, 268-269 target and test environment, 266-267 research related to, 416-429 Bremsstrahlung radiation, 7-9 Brighmess in synchrotron storage ring, 10 from wiggler sources, 12 Brilliance, in synchrotron storage ring, 10 Building blocks, alkylaryl aromatic, 192- 194

C Cadmium decrease in presence of EDTA, 243 sorbed by manganese oxides, 240-241

sorption by kaolinite and smectite, 244-245 uptake by mobile DOC. 252 Caenorhabditis elegans. ACEDB database, 152-154 Calibration curve, Py-FIMS, 174-175 Carbon dioxide, supercritical extracts Py-FIMS, 203-205 Py-HMS and Py-FDMS, 188-190 Carbon K edge, and obtaining chemical contrast, 51-52 CBs. see Chlorobenzenes Chlorine substitution, PCB congeners with, 349-353 Chlorobenzenes cultivated horizon concentrations, 353-356 degradation and leaching, 368-370 physicochemical properties, 349-35 1 potential fate in sewage sludge-amended soils, 374-375 volatilization, 37 1-372 Chloroform-n-hexane extracts Py-FDMS, 184-186 Py-FIMS, 176-179 Chromium distribution in plant tissue, 31-32 rich plots at Luddington, 356-358 Chromosomes Brassica and Arabidopsis, 133- 135 cotton, molecular map, 137 favorable position on target gene, 289 maize, very wide crosses, 119 plant artificial, 299-300 sugarcane, back-crossing, 123 yeast artificial, positional cloning with, 295296 Classification germ plasm, 270-274 in GIS, with mapping categories, 76-77 hardness, organics in nature, 227 metal ions, 224-225 Class probability maps, soil drainage, 86 Clays fulvic acid components retained by, 209-210 powder diffraction analysis, 44-45 retention of metals, 244-246 SAXS studies, 48-50 soil, see Soil clays Cloning map-based, 139, 295-300, 324-326 positional, 287-288, 295-2% Cobalt, second neighbor atoms, 25

INDEX Collimation undulator beam, 13 x-rays, 9 Colloids iron oxide, mobility, 249 permanent charge, soil dominated by, 246 Commercial growing, peanut in United States, 395 Competing reactions, hydration and metalligand formation, 226 Competition, peanut species with weeds, 425426 Complexation, metal-organic, role in metal s o p tion by soils, 219-254 Complex crosses, in peanut breeding, 41 1-412 Conductivity meters, based on electromagnetic induction, 83-84 Congeners, PCB, sludge-applied, loss, 356-361 Consortium for Advanced Radiation Sources, 56 Contaminants in bone fragments and turtle shells, SXRF, 36-37 KCSQLs for, 379-380 and nutrients, fate in soils, 21-25 transport in soils, 249-252 Conversion programs, germ plasm, 31 1-312, 317-31 8 Coordination environment, in XAS, 19-21, 27 Copper bonding sites on humic and fulvic acids, 235236 rich plots at Luddington, 356-358 rich sludge at Lee Valley, 362-365 sorption in presence of 2.3-PDCA, 239-240 Coronene, persistence, 362-366 Correction, in GIS, with multispectral imagery, 76 Cost factor, GIS applications, 74 cotton allotetraploid, linkage group ancestry, 138 chromosomes, molecular map, 137 cultivation, 135-136 improvement with DNA markers, 139-140 map-based gene cloning, 139 Cowpea, and mungbean, comparative mapping, 129 Crop condition maps, for site-specific farming, 95 Crop improvement benefits for, 288-289 peanut, 397

449

programs, 298-300 stages, 273-282 Crop plant genomes genetic and physical characterization, 283300 relating genetic and physical distances, 28929 1 Crop productivity, increased, sources of genetic gain for, 266-268 Crop yield maps, for site-specific farming, 9293 Cruciferae, PGRP, 129- 135 Crystal edges, gibbsite and boehmite, 242243 Cultivar development programs DNA markers in, status survey, 320-328 peanut, 409, 413-414 Cultivars peanut development, 4 13-430 improvement depending on crop use, 397 single, blending sister lines into. 413 Cultivated horizons, see also Soil horizons archiving of soil from, 348-349 movement of PAHs to subsoil from, 370 organic chemical concentrations, 353-366, 381-382 Curie-point pyrolysis-gas chromatography/ electron ionization MS, 175, 191-199, 21 1-212 Cytogenetics, and cultivar development, 290 Cytogenetic stocks, cotton, 137- 138

D Data DNA sequence, 271-272 in GIS retrieval and output, 71 spatial capture, 70-71 obtained on extractions with organic solvents, 190-191 source, remote sensing as, 75-76 Database, plant genome, 147-154 Data layers, GIS, for site-specific farming, 9296 1,4-DCB, see 1,4-Dichlorobe.nzene Decision support systems for site-specific farming, 90,92, 96-97 spatial, 103-104 Degradation, organic chemicals, 368-369

450

INDEX

Detection limits, refinedldecreased with improvements in sampling, 382 Detectors, energy-dispersive, in SXRF, 35-36 1,4-Dichlorobenzene, in control and sludgeamended plots. 355-356 D i h t i o n , see also Microdiffraction single crystal and time-resolved, 46-48 Diffusion, film and intraorganic matter, 379 Digital elevation models, in GIs, 85-86 Disease resistance genes, mapping, 127-128, 146 Diseases in peanut foliar, effect on production, 423-424 nematode, species resistance, 422-423 soil-borne. 424-425 Disposal, sewage on land, 346-348, 386-387 Dissipation and KCSQL concept, 377-380 organic chemicals in soils, 346-348 Dissolved organic carbon contaminant transport, 252 in decomposing litter layer, 221 Distribution maps, contaminated soil particles, 37-39 DNA fingerprinting, in plant breeding programs, 270-283 DNA markers advantages for back-cross breeding, 315-317 based estimates of genetic diversity, 274, 282-283 in cotton improvement, 139-140 genetic maps based on, 295-300 in genome analysis, 117-118 native and nonnative, 290 and plant breeding programs, 265-329 RFLP assisted selection, 120 for forest tree species, 144-145 RFLP and RAPD, 129- I30 status survey, 320-328 DNA polymorphisms, see also Restriction fragment length polymorphism methods for detection, 273-274 utility in plant breeding programs, 268 DOC, see Dissolved organic carbon DRASTIC, ranking pollution potential, 99- 100 Drought, peanut genotypes escaping, 418-420 Dutch target values, volatile compound residues, 381-385

E Early generation testing, peanut, 409 EDTA adsorbed by iron oxide surfaces at low pH, 238 complex with Co(II), 243 increase of diffusion rates within soil, 251252 Electromagnetic induction, conductivity meters based on, 83-84 Electromagnetic radiation, in synchrotron, 7-9 Electron microprobe techniques, sensitivity problem, 34-35 Electrons, circulated by synchrotron, 3-6 Elements low-z

studied by Auger electron yield, 18 XAS, 32 Z > 20, SXRF, 34-39 Energy distribution bending magnets, 10- 1 I insertion devices, 11-13 minimization, humic acid structure, 198 narrow range, in XAS, 16-21 versus wavelength, 9 Environment, target and test, in plant breeding, 266-267 Environmental applications. GIS,98- 104 Environmental condition maps, for site-specific fanning, 95 Environmental implications, interaction of metal ions with soluble organics, 248-254 EOS-A platform, multi-instrument, 80-82 Epistastic effects, underestimation, 303-305 Equilibrium constant, complexation process, 228-229 Eucalyprus, QTL mapping, 145 European Union priority pollutant lists, 35 1-353, 381 sewage dumping directives, 346 waste and water protection directives, 376377 Evolution, cotton, 138-139 Evolution profile, volatile and ionized substances, 204-205 EXAFS, see Extended X-ray absorption fine structure

INDEX Extended X-ray absorption fine structure Fe, 50 Pb, 238 quick technique, 33 region in XAS, 20-21 spatially resolved, 28

F Facilities, synchrotron, accessing, 54-56 Farming, site-specific, and CIS, 87-97 Farmland, sewage sludge applications, implications, 376-385 Fatty acid composition, peanut, 427-428 n-Fatty acids, detection by Py-FIMS and PyFDMS, 176-191 Fertilizers inorganic, and PCB congener increase, 36036 I and site-specific farming, 87-97 Field applications, management plan maps, 97 Fingerprinting, DNA, in plant breeding programs, 270-283 Fingerprint region, in XANES, 20 Flash pyrolysis, with Py-FIMS, 212-213 Fluorescent X-ray interference techniques, 4042 Flux, in synchrotron storage ring, 10 Foliar diseases, effect on peanut production, 423-424 Forest trees genetic and trait mapping, 143-145 limitations of MAS, 318 Fractal dimension, polymerized octahedral layers, 49-50 Fragmentation, mass spectrometric, 170- 171 Fresnel zone plates in soft X-ray microscopes, 51 in XANES, 33 Fulvic acid Armadale, Py-FIMS, 200-201 complex with metal, stability constants, 230233 Cu bonding sites, 235-236 definition, 168- 169 n-hexane-chloroform extracts Py-FDMS, 184-186 Py-FIMS, 176- 179 Py-FIMS, mineral effects, 208-210

451

role in Zn sorption by smectite, 245 supercritical n-pentane extracts Py-FDMS, 186-188 Py-FIMS, 180-183 Fungus, infected wheat m t s , micro-XANES, 30-3 1 Fuzzy searching, and plant genome database, 149-151

G Gene banks, for plant breeding programs, 271272 Genes comparative mapping, I18 disease resistance, mapping, 127- 128, 146 integration into legume maps, 127-128 isolation through transposon tagging, 296298 map-based cloning, I39 mapping, for forest tree species, 144-145 orthologous, positional cloning, 287-288 sequencing, for forest trees, 146 Genetic distance between Arachis spp., 396 and physical distance, in crop plant genomes, 289-291 RFLP-based estimates, 278 Genetic diversity assessment via DNA markers, 282-283 and genetic merit, assessment, 269-283 increase for peanut cultivars, 413 Genetic effects, QTL,estimation, 303-307 Genetic gain expected, estimation, 303-307 improvement through back-cross breeding, 316-317 produced by MAS, 314 sources for increased crop productivity, 266268 Genetic heterogeneity, in plant breeding, 307308 Genetic maps Brassicu spp., 130- I32 in plant breeding programs, 283-300 Genetic merit, and genetic diversity, assessment, 269-283 Genetics, and peanut breeding, 393-430 Genetic transformation, peanut, 402-403

452

INDEX

Genetic variability, Arachis spp., natural, 397402 Genetic variation estimation, 277-278 generation, role of recombination, 291-295 Genome architecture, 283-300 diploid, highly duplicated in Brassica, 130 maize and sorghum, PGRP, 115-120 small grains and sugarcane, 120-123 Genomic mapping, among legumes, 125- 129 Genotype x environment interaction peanut lines. 410, 412-414 as source of bias, 303-307 Geographic information systems advantages and disadvantages, 74 in site-specific farming, 91-97 technology overview, 68-74 Geometrical optimization, humic acid smcture, 198 Geostationary Operational Environmental Satellites, 79-80 Germ plasm collections, 271-272 crop, genetic composition, 269-270 elite, genetic diversity, 272-274 exotic assessing and introgressing, 31 1-312 introgression, 284 exotic and elite, crosses between, 317-318 Gossypiwn spp., 136 molding for increased productivity, 267 peanut, collection and exchange, 399-402 Germplasm Resources Information Network, 136, 398 Gibbsite, reaction sites at crystal edges, 242243 GIS, see Geographic information systems Gleysolic paleosol Oh horizon n-hexane-chloroform extracts, Py-FIMS, 176-179 supercritical n-pentane extracts, 180-183 Global positioning system, in site-specific farming, 91-97 Global production, peanut, 396-397 Glutenins, high-molecular-weight, 276-277 Goethite surface hydroxyls, 237-239 XAS applications, 22-25

Gopher, and plant genome database, 149-151 Cossypiwn spp., PGRP, 135-140 Graminae, PGRP, 115-124 GRASS, GIS software, 103 Greenhouse screening, for peanut diseases, 423 Ground-penetrating radar, for GIS, 83 Groundwater, contaminant fate and transport, 248-252 Groundwater ubiquity score, 99-100

H Half-life atrazine applications, 385 PAHs, 362-366 sludge-applied PCB congeners, 358, 361 Hard and soft acid base theory, 225, 236-237 Harvest index, in improved peanut pod yield, 419 Heritability, broad-sense, in peanut, 417-418, 429 Heterosis, mapping QTL for, 302-303 Heterotic group, in plant breeding, 279-281 Hexachlorobenzene, slow dissipation from soils, 355-356 n-Hexane-chloroform extracts Py-FDMS, 184-186 Py-FIMS, 176-179 High mass resolution Py-FDMS, 172-173 Py-FIMS, 205 Homozygosity, in progeny, monitoring, 324325 HSAB theory, see Hard and soft acid base theory Human exposure, to organic chemicals via diet, 376 Humic acid aliphatic structures, 21 1 Armadale, Py-FIMS, 200 complex with metal, stability constants, 230233 Cu bonding sites, 235-236 definition, 168-169 n-hexane-chloroform extracts Py-FDMS , 184- 186 Py-FIMS,176-179 Py-GC/MS. 191-199 role in Cd sorption by smectite, 245 supercritical n-pentane extracts

453 Py-FDMS, 186-188 Q-FIMS, 180-183 three-dimensional structure, 196- 199 Humin Annadale, Py-FIMS, 201 definition, 168-169 Hybrid combinations, n prion' parent selection for, 279-281 Hybridization, peanut artificial, 408 interspecific, 400-402 Hybrid rescue, peanut, 401-402 Hybrids, commercial single-cross, 281 Hydration, and metal-ligand formation, competing reactions, 226 Hydrolysis products, Al and Fe, SAXS studies, 48-50 Hydrophobicity, alkyltin, 240

I Illite, Pb and Cd on, 251-252 Information resources, plant genome database, 148- 152 Infrared microspectroscopy, at NSLS, 42 Ingestion, organic chemicals by livestock, 372376 Inheritance monogenic, 315-317 nonadditive, 422 polygenic, 317-318 qualitative, 405 quantitative patterns, 300-31 1 Inorganic solids, XAS, 21-22 Insertion devices energy distribution, 11-13 synchrotron, 3-7 Integrated pest management, and peanut cultivan, 415, 423 Intercropping, peanuts with cereals, 426 Interference patterns, constructive and destructive, 20-21 International Crops Research Institute for SemiArid Tropics, 394, 397-401, 427 Introgression exotic germ plasm, 284, 311-312 programs, 278, 292 traits into elite germ plasm, 272

Iron containing core of fenitin, measurement, 41 EXAFS studies, 50 and manganese, coordination in poorly ordered phases, 22 Iron oxides, surface hydroxyl sorption sites, 237-240

K Kaolinite Pb and Cd on, 251-252 sorption of Cu, 244-245 KCSQL, see Kinetically constrained soil quality limit Kinetically constrained soil quality limit and dissipation processes, 347 numerical derivation, 380-38 1 overview, 377-380 Ka radiation copper, 9 magnesium, 30-3 I

L Land disposal, sewage, 346-348, 386-387 Landstat Thematic Mapper, 77-78, 80-84 Leaching, organic chemicals, 369-370 Lead Pb(I1) in presence of manganese oxides, 241 sorption potential of iron oxide, 238 Leaf spot, early and late, peanut diseases, 424 Lee Valley site IPAH concentrations, 362-366 study of metal uptake by crops from sludgeamended soil, 348 Legislative controls, to limit addition of sludgemetals to agroecosystems, 346-347 Legume maps, 127-128 Leguminosae, PGW, 124- 129 Lentil, and pea, genetic maps from cross, 128I29 Ligand-metal interactions, 225-229 Light, synchrotron, 3-15 Linkage drag, reduction, 315-317 Linkage groups assignment to chromosomes, 137- 138 in cultivated peanut, 400-401

454 Litter layer, decomposing, high Doc concentrations, 221 Livestock, ingestion of organic chemicals, 372376 Locus Rpl, maize. 294-295

family Solanaceae, 140- 143 maize, 1 I8 disease resistance genes, 127-128, 146 genetic and trait, in woody species, 143-145 genomic, among legumes, 125-129 in CIS, 71 Sh2 oxidation state, 3 1-32 macromutations, 309-310 QTL, 128, 278, 285-286, 302-303 transposon insertion, 313 topographic, with RADARSTAT. 82-83 Logic inference systems, for CIS, 104 trait loci, 133 Luddington site Maps, see also specific types PAH contamination, 362-366 based gene cloning, 139, 324-326 ZPCB concentrations, 356-358 cytogenetic. integration with RFLP, 289-290 study of metal uptake by crops from sludgeintegrated, development, 284-289 amended soil, 348 RAPD, 144-145 RFLP, 121-122, 132 M for site-specific farming, 92-96 Marker-assisted recurrent selection, and QTL, Macromutations 319 gene tolerance for, 300 Marker-assisted selection at Sh2 locus, 309-310 for forest trees, 147 single genes affecting, 299 for plant breeders, 154 Magnets and QTL, 120. 313-320 bending, see Bending magnets target region, 302 injection, in synchrotron, 3-6 Markers Maize DNA, see DNA markers breeding types, and development of integrated maps, epistatic effects, 304-305 284-286 and hybrid combinations, 279-281 MAS, see Marker-assisted selection clustered multilocus QTL, 310 Mass selection, peanut breeding method, 409 fine structure analysis, 291 Maturity, early, peanut lines with, 417-418 genome, PGRP, 115-120 Metal ions QTL mapping, 286 classifications, 224-225 Rpl IOCUS, 294-295 reactions with organic surfaces, 229-236 US. yields through year 2000, 268 soil binding affinities, 246-247 Maltcot’s coefficient of coancesay, 273-274 sorption by manganese oxides, 240-241 Malvaceae, PGRP, 135-140 Metal-ligand interactions, 225-229 Management constraint maps, for site-specific Metals farming, 96 in soil solution, 223-229 Manganese, and iron, coordination in poorly orsoluble complexes, XAS, 26-27 dered phases, 22 toxicity, 253-254 Manganese oxides, sorption of metal ions, 240- Metal-water interactions, 223-225 24 1 Microclimate attribute maps, for site-specific Map locations, genes encoding key cell funcfarming, 94-95 tions, 118-1 I9 Microdiffraction, with synchrotron X-ray beams, Mapping 45 accuracy and classification, 76-80 Microsatellites, coupled with polymerase chain comparative, 286-288 reaction, 128 Arabidopsis and Brassica, 133- I35 Microtomography, X-ray computed, 52-54 cowpea and mungbean, 129 Mineral phases, poorly ordered, 16, 21-22, 50

45 5 Minerals effects on Py-FIMS of fulvic acid, 208-210 powder diffraction, 43-46 surfaces, redox reactions, 25-26 XAS, 21-22 Mirrors double elliptical Kirkpatrick-Baez, 33-34 focusing, in XAS, 18-19 Misono softness parameter, and metal-ligand complexes, 226-228 Mobility, soluble organics, 249-252 Molecular map, cotton chromosomes, 137 Monsoon '90 project, using remote sensing, 84 Mossbauer spectroscopy, isotope measurements, 42-43 Mungbean, and cowpea, comparative mapping, 129

N Naphthalene half-life at Luddington, 362-366 volatilization, 372 National Research Initiative and Agricultural Research Service, 114-1 15 plant genome awards, 116 National Synchrotron Light Source scanning transmission X-ray microscope, 5 I52 with two storage rings, 15 X-26A beamline, 35 Nematode diseases, in peanut species, 422423 NEXAFS, see X-ray absorption near edge structure NEXRAD radars, 79-80 Nickel displaced from fulvic acid by oxalate, 234 rich plots at Lee Valley, 362-365 at Luddington, 356-358 Nitrogen fixation, by peanut, 404,421 Nonpoint source pollution, watershed basis, 98, 100-104

NSLS, see National Synchrotron Light Source Nuclear magnetic resonance, I T , application to SOM, 169 Numerical derivation, KCSQLs, 380-381 Nurseries, off-season, in plant breeding, 267

Nutrients and contaminants, fate in soils, 21-25 leachate, and site-specific farming, 90-91 Nutritional profile, peanut, 427-429

0 Oats, genome analysis, 121 Oil, content in peanut genotypes, 426-427 Olation bridge structure, at iron oxide surface, 237-238 Organic acids, low-molecular-weight, 220-223 Organic chemicals behavior assessment models, 373-376 dissipation, rate-limiting factors, 346 loss from long-term sewage sludge-amended soils, 368-373 selected case studies, 38 1-385 in sewage sludge-amended soil, physicochemical properties, 349-353 Organics effect on reactions of metal ions with inorganic surfaces, 236-247 hardness classification, 227 low-molecular-weight, 229-236 in soil solution, 220-223 Organic solvents, extractions with, data summary, 190-191 Organogenesis, in peanut, 402 Out-crossing rates, see also Back-crossing peanut, 408 Oxalate displacement of Ni from fulvic acid, 234 increase of Cd sorption by soils, 247 Oxidation states Cr, 19-20 Fe, Mn, and U, 28-31 quantification and mapping, 28-32 Se, 21

P PAHs, see Polycyclic aromatic hydrocarbons Parasitic synchrotron X rays, 6 Parent selection hybrid combinations, 279-282 peanut cultivars, 406-408 source (base) populations, 275-279 Pathogens, soil-borne, affecting peanut species, 425

456

INDEX

K B s , see Polychlorinated biphenyls 2.3-PDCA, see 2.3-4.razinedicarboxyIic acid Pea, and lentil, genetic maps from cross, 128129 Peak resolution, powder diffiction, 43-44 Peanut breeding methods, 409-412 and production, 396-397 research related to, 416-429 cultivars, development, 413-430 distribution and taxonomy, 394-396 genetic variability, 397-403 genotype X environment interactions. 412414 reciprocal cross differences, 407-408 sustainable systems, 414-415 vegetative and reproductive traits, 403405 Pedigree method, peanut breeding, 409-410 Pedology, GIS applications, 85-87 Pegs, peanut species, 394-395.417 n-Pentane, supercritical extracts Py-FDMS, 186-188

Py-FIMS,180-183 Pepper, gene repertoire, 141-143 Persistence organic chemicals in sludge-amended soils,

353-376 PAHs, 362-366 Pest distribution maps, for site-specific farming,

95-96 Pesticide leachate. and site-specific farming,

90-91 Pesticides agricultural, 99-103 with 1.4-DCB, 355 as source of organic chemicals to soils, 367-

368 PGRP, see Plant genome research program pH, below zero point of charge, 239 Phenotypic xecurrent selection, trait responses,

319 Photoelectrons, production in XAS, 16-21 Physiology, peanut, 416-420 Plant artificial chromosome, in crop manipulation, 299-300 Plant biology complexity, 297-298 limiting DNA marker use in cultivar development, 326-328

and plant breeding, DNA markers as link,

268-269 Plant breeding programs and DNA markers, 265-329 and genome mapping, 125-127 integration of MAS, 320 response to selection in, 312-313 Plant genome database, 147-154 Plant genome research program Cruciferae, 129-135 Graminae, 115-123 history and program establishment, 113-1 14 Leguminosae, 124- 129 Malvaceae, 135-140 progress during 1991-1994, 1 I5 Solanaceae, 140-143 Plant uptake, CBs and PAHs, 370-371 Plots, sludge-amended, CB, PCB, and PAH concentrations, 353-366 Plowing transboundary losses due to, 372-373 transfer of PAHs during, 384 Pods, peanut varieties formation, 395 yield, 413-420 Polarization, radiation, 14 Pollutants fate in soils, 54 priority, classified by USEPA and EU, 351-

353 Pollution assessment, regional-scale, 98- 100 Polychlorinated biphenyls degradation, 369 ZPCB concentrations in experimental plots,

356-361 ZPCB long-term persistence, 383-384 physicochemical properties, 35 1-353 potential fate in sewage sludge-amended soils, 374-375 Polycyclic aromatic hydrocarbons cultivated horizon concentrations, 362-366 degradation and leaching, 368-370 ZPAH persistence, 384-385 physicochemical properties, 353 potential fate in sewage sludge-amended soils, 374-375 Polymerase chain reaction, coupled with microsatellites, 128 Populations reference, concept and definition, 308 source, for plant breeding programs, 275-279

457 Positrons, circulated by synchrotron, 3-6 Potato, gene repertoire. 141-143 Powder diffraction, minerals, 43-46 Preedge region, in XAS, 19 Profitability, site-specific farming, 91 PwFY homozygosity monitoring, 324-325 peanut, from interspecific hybridization, 400-

402 Proteinaceous materials, in humic acids, 195-

I96 Py-FDMS, see Pyrolysis-field desorption mass spectrometry Py-FIMS, see Pyrolysis-field ionization mass spectrometry Py-GUMS, see Curie-point pyrolysis-gas chromatography/electron ionization MS 2,3-Pyrazinedicarboxylic acid, Cu sorption in presence of, 239-240 Pyrolysis-field desorption mass spectrometry description, 172- 173 n-hexane-chloroform extracts, 184- 186 supercritical extracts COZ, 188-190 n-pentane, 186-188 Pyrolysis-field ionization mass spectrometry analysis of SOM, 199-207 description, 173- 175 fulvic acid, mineral effects, 208-210 n-hexane-chloroform extracts, 176- 179 supercritical extracts COZ, 188-190 n-pentane, 180-183 time-resolved, 203-207 Pyrolysis-mass spectrometric methods, fimdamentals, 170-175

Q QTL, see Quantitative trait loci Qualitative genetics, peanut, 403-405 Quantitative genetics, peanut, 405-406 Quantitative trait loci analysis, research needs, 310-31 I consistent detection across populations, 307-

308 detecting and locating, 301-303 gene frequency changes, 147 genetic effects, 303-307 genetic nature, 308-310 mapping, 128,278, 285-286

mapping in forest trees, 145 and MAS, 120 transfer. 322

R RADARSTAT, topographic mapping, 82-83 Radiation, see also specific types emitted in toroidal pattern, 8-9 synchrotron, properties, 7- 14 Rapidly lost phase, fraction of total organic chemicals comprising, 377-378 Raster formats, in CIS, 69, 101 Rate-limiting factors controlling persistence, 378 for dissipation of organic chemicals, 346 Ratioing, in CIS, 77 Rearrangements, chromosomal, 134-135, 141 Reciprocal cross differences, in peanut, 407-408 Recombination frequency among Erassicu spp., I32 role in generation of genetic variation, 291-

295 as source of bias in estimates of additive effects, 303-307 Recurrent selection, in peanut breeding, 41 I Redox reactions, on mineral surfaces, XAS, 25-

26 Remote sensing as data source in CIS, 75-76 supported CIS operations, 76-80 trends, 80-84 Reproductive traits, peanut, 404-405 Research facilities, synchrotron, 6 Research needs for genetic diversity in plant breeding programs, 282 for QTL analysis, 310-311 Resistance, peanut cultivars to biotic stresses, 421-426 to disease, 401-402 to specific stresses, 414-415 Resolution, genetic, limits, 301-302 Restriction fragment length polymorphism based estimates of genetic distance, 278 genetic similarity, 280-281 and cytogenetic maps, integration, 289-290 linkage mapping, 122- 123 markers, 120 MAS with, 318-319

458

INDEX

RFLP. see Restriction fragment length polymorphism Rhizosphere, low-molecular-weightorganics in, 22 1

Rice, genome analysis, 121-122 Risk assessment, site- and population-specific, 377

Rocks, computed microtomographystudies, 5354

Root exudates, aliphatic acids, complex with metal ions, 223 Root formation, de novo. peanut, 402 Root growth, and response to drought, 419 Rothamsted Classical Experiments, 349 Rpl locus, maize, 294-295 Rust leaf, resistance genes, 288 in pines, resistance gene mapping, 146

S Scales of measurement, synchrotron-basedtechniques, 15-54 Scanning transmission X-ray microscope at NSLS, 51-52 in XAS of low-Z elements, 32 Scattering, anomalous. in powder diffraction, 44-45

Selection in early generation, 412 in maize breeding, 279-281 marker-assisted for forest trees, 147 for plant breeders, 154 and QTL, 120. 313-320 marker-assisted recurrent, 3 19 parent hybrid combinations, 279-282 source (base) populations, 275-279 phenotypic recurrent, 319 in plant breeding programs, response to, 3 12313

recurrent, 41 1 utility for early emergence and maturity, 418 Selenium, distribution in plant tissue, 31-32 Self-pollination, peanut, 406-408 Sewage sludge applications to farmland, implications for soil quality criteria, 376-385 and organic chemical persistence, 353-376

sh2 locus macromutations, 309-3 10 transposon insertion, 313 Silica oxides, sorptive toward heavy metals, 244 Simple sequence repeats, in genotype identification, 128 Simulation for Water Resources in Rural BasinsWater Quality model, 101 Simulation models deterministic, MAS in plant breeding programs, 315-318 for site-specific fanning, 96 Single-cross hybrids, commercial, 281 Single seed descent method, peanut breeding, 410-41 1

Site-specific farming, and CIS, 87-97 Sludge-amended soil, organic chemical persistence, 353-376 Small angle scattering, studies of weakly scattering systems, 48-50 Smectite Cd sorption, 251 Cu and Cd sorption, 244-245 Soft ionization mass spectrometry, FI and FD, 171-172

Software ACEDB, 153-154 for CIS, 77-78, 99-100, 103 Soil, plant, and environmental sciences, applications of XAS, 21-34 Soil attribute maps, for site-specific fanning, 93-94

Soil-borne diseases, in peanut species, 424-425 Soil clays major components, 21 1-212 n-hexane-chloroform extracts Py-FDMS, 184-186 Py-FIMS, 176-179 supercritical n-pentane extracts Py-FDMS, 186-188 Py-FIMS , 180- 183 Soil deficiency, tolerance of peanut cultivars, 420-421

Soil drainage, class probability maps, 86 Soil horizons, see also Cultivated horizons Armadale Bh. 188-190 Gleysolic paleosol Oh, 176-183 reconstruction, 85-87 salinity, conductivity related to, 83-84 Spodosol Bh, 21 1

459

INDEX Soil-landscape modeling, by CIS, 84-87 Soil organic matter analysis by pyrolysis-soft ionization MS, 176-190 definition, 168-169 Py-FIMS analysis, 199-207 retention of heavy metal ions, 229-236 three-dimensional structure, 1%-199 Soil quality criteria, implications from sewage sludge application to farmland, 376-385 Soils, see also Sludge-amended soil Armadale, Py-FIMS, 202-203 Armadale and Bainsville, Py-GCIMS analysis, 192-194 binding affinities to metal ions, 246-247 computed microtomography studies, 53-54 fate of nutrients and contaminants, 21-25 genesis and fertility, 253 movement due to plowing, and organic chemical loss, 372-373 normal, below soil quality limits, 383-384 phosphorus-deficient, response of peanut genotypes, 421 poorly ordered mineral phases, 50 sources of organic chemicals, 367-368 whole, supercritical CO, extracts, 188190

XAS, 27-28 Soil solution DOC decrease in, 253 metals in, 223-229 organics in, 220-223 Solanaceae, PGRP, 140-143 SOM, see Soil organic matter Sorghum clustered multilocus QTL, 310 genome, PGRP, 115-120 Sorption at crystal edges, 242-243 effect of organics, HSAB theory-based predictions, 236-237 Sorption-desorption mechanisms, as ratelimiting phenomena, 378-379 Source population, for plant breeding programs, 275-279 Speciation, in family Solanaceae, 141 Spcdosol Bh horizon, Py-FIMS, 21 I SPOT, French satellite, 75, 78, 80-84 Stability constants Cd-SOM. ,~ 251

energy released during metal-ligand interaction, 228-229 for metal-fulvic acid and metal-humic acid, 230-236 Standing wave techniques, 40-42 Storage ring BESSY electron, 51 in synchrotron, 3-10 vacuum, isolated from beamline vacuum, 1415 stresses abiotic, tolerance in peanut cultivm, 416420 biotic, resistance of peanut cultivm, 421426 Substitution, chlorine, PCB congeners with, 349-353 Sugarcane, genome analysis, 123 Supercritical fluid extraction with CO,, 188-190 with n-pentane, 180-188 Super VHS technology, in airborne video, 82 Surface hydroxyls aluminum oxides, 242-243 iron oxides, 237-240 Surfaces inorganic and organic, metal ion reactions with, 229-247 silanol, sorptive toward heavy metals, 244 Surface-sorbed species, XAS, 22-25 Survey, status of DNA markers in cultivar development programs, 320-328 Sustainable systems, peanut cultivar development for, 414-415 SXRF, see Synchrotron X-ray fluorescence spectroscopy Synchrotron facility access, 54-56 general description, 3-7 X-ray, hard versus soft, 14-15 Synchrotron X-ray fluorescence spectroscopy, for elements Z > 20, 34-39 Synteny and genome structure, maize, 119 Graminae species, 123

T Taxonomy, peanut, 394-396 Temperature resolution, in Py-FIMS, 207

460

INDEX

Temperature response, peanut cultivars, 420 Terrain analysis, by GIS, 84-87 Thematic maps, generated by remotely sensed data, 76 Thermal profiles, compounds in fulvic acid, 208-210 Thermosensitive detector, in Py-GCIMS, 175 Time structure, pulsed, 14 Tomato clustered multilocus QTL, 310 gene repeltoire, 141-143 recombination, regional assessments, 293294 Total electron yield, in XAS, I8 Toxicity metal, 253-254 tolerance of peanut cultivars, 420-421 Trait loci, see also Quantitative trait loci mapping, 133 Traits complex, analysis, 300-313 incorporation through back-crossing, 428 introgression into elite germ plasm, 272 monogenic and polygenic, MAS, 318-319 qualitative and quantitative, gene backcrossing, 306 reproductive and vegetative, peanut, 403-405 Transgenes definition for DNA marker status survey, 320-328 introduced into germ plasm, 316 Transposons, source of genetic variation, 313 Transposon tagging in gene isolation, 2%-298 variant of map-based cloning, 326 Triangular irregular network, in GIs, 85-86 lbltle shell, contaminant distribution, SXRF, 36-37

metals, by crops from sludge-amended soil, 348 USDA Cotton Germplasm Collection, 136

V Vaccines, delivered by plant viruses, 47 Vacuum chamber, ring-shaped, in synchrotmn, 3-6 Variable management, in site-specific farming, 87-90 Vegetative traits, peanut, 403-405 Video, airborne. use of super VHS technology, 82 Viruses. plant, as vaccines, 47 Volatile organic compounds, in sewage sludges applied to soils, 381-382 Volatilization to atmosphere from soil, 386 CBS, PCBS, and PAHs, 371-372 fractional, in Py-FIMS and Py-FDMS,205206

W

WAIS, and plant genome database, 149-151 Waste disposal, and site-specific farming, 97 Wastes, agricultural, as source of organic chernicals to soils, 367-368 Water-metal interactions, 223-225 Watershed basis, nonpoint source pollution, 98. 100-103 Wavelength, versus energy, 9 Waxes coniferous epicuticular, 183 leaf cuticular, 299 natural, detection by Py-FIMS, 179 Weeds, and peanut yield losses, 425-426 Wheat U bread-baking quality, 276-277 genome analysis, 122-123 Undulators harvest index, 298 beam collimation. 13 roots, micro-XANES, 30-31 magnetic device in synchrotron, 3-6 Wigglers United Nations Convention on Biological Diverbeamline in microdiffraction, 45 sity, and germ plasm, 399-400 magnetic device in synchrotron, 3-6 Uptake sources of white light, 11-12 cadmium, by mobile DOC, 252 Woburn Market Garden Experiment, 348-349, CBs and PAHs, by plants, 370-371 353-365

461

INDEX Woody species, genetic and trait mapping, 143145

World Wide Web, and plant genome database, 149-151

X-ray intensity, definitions, 9- 10 X-ray microscopy, focusing and contrast, 5 1-52 X-ray sources hard versus soft, 14-15 synchrotron, generations, 6-7

X

Y XANES, see X-ray absorption near edge structure XAS, see X-ray absorption spectroscopy X-ray absorption near edge structure region in XAS, 19-20 S Kedge, 27 spatially resolved micro-XANES technique, 28-32

study of redox reactions, 25-26 X-ray absorption spectroscopy applications, 2 1-34 basic principles, 16-2 1 future developments, 32-34 X-ray diffraction, with synchrotron source, 4348

Yeast artificial chromosomes, positional cloning with, 295-296 Yields, peanut pod, 413-420

Z Zero point of charge oxide minerals, 238-239 variable charge soils, 246-247 Zinc organically bound in calcareous soils, 252 retention in soils, 246-247 rich plots at Luddington. 356-358, 362-363 sorption by smectite, 245-246

I S B N O-L2-000?55-X