consuulwtve Group an intemadonat Agricultural Rearch Oct.lqqoz Strategies for ausainig Crop G on, Enhncment, and Ute Grison Wkes Professor of Biology University of Massachusetts$oston ISSUES IN.. "Issues in Agriculture" is an evolving series of booklets on topics connected with agricultural research and devel- opment. The series is published by the Secretariat of the Consultative Group on InternationalAgricultural Research (CGIAR) as a contribution to informed discussion on issues that affect agri- culture. The opinions expressed in this series are those of the authors and do not necessarily reflect a consensus of views within the CGLAR system. Published by the Consultative Group on International Agricultural Research, CGIAR Secretariat, 1818 H St., N.W., i Washington, D.C., 20433. United States. (1 -202) 473-895 1; Fax (1 -202) 334-8750. October 1992. ~~~~~~~.............. .............. . ............. ..:..'....i.WW Strategies for Sustaining Crop Germplasm Preservation, Enhancement, and Use Garrison Wilkes Professor of Biology University of Massachusetts/Boston ISSUES IN AGRICULTURE, NO. 5 Consultative Group on International Agricultural Research Sg5X& WlgXSH'Xgg^'^''a"'t0}'a M;XESil' Wa' XM1gn'"'') N.P+2. g. i tr 9.k ter m=s=§§§SEX!i W2 hSg~~~~~~~~~~~~~~~~~~~1t |w$_ g~ Contents Introduction ......................................... 1 The Challenge ......................................... 3 Evolution under Domestication of Crop Plants ............4 Agricultural Treadmill ......................................... 6 Telescoped Evolution ......................................... 7 The Story of Agriculture ....................................... 9 History of Plant Breeding ................................... 10 Genetic Erosion ......................................... 12 Genetic Wipeout ......................................... 13 Genetic Vulnerability ......................................... 13 Genebanks and Germplasm ...................................... 16 In Situ Germplasm Conservation ....................... 21 Genebanks and Storage ..................................... 28 Genebanks Slow Down Crop Evolution .............. 34 The Use of Germplasm Resources ...................... 36 Germplasm for Varietal Development ................ 38 Training for Plant Breeding ....................................... 39 Plant Breeding Needs ......................................... 41 Differences between Primitive Populations and Modem Varieties ..................................... 44 Varietal Mixtures ......................................... 45 Plant Breeding in Developing and Developed Countries ......................................... 47 Future Needs and Priorities ..................... .......... 52 Agriculture, Germplasm Conservation, and Strategies for the Future ........................................ 53 Summary and Charge for Change .................. ........... 55 Acknowledgements ....................... .................. 56 References ....... 57 i Strategies for Sustaining Crop Germplasm Preservation, Enhancement, and Use Garrison Wilkes Introduction The idea comes flrst, then the accomplishment; in the same way genes are the blueprint and plant breeding is the yield improvement that feeds the expanding human populations. Genetically improved cultivated plants and domesticated animals provide an assured food supply, which liberates most humans from the daily quest for food, yet over half of the world's people still live off the production of their own fields. To meet the current demands of an increasing world population an ominous conflict exists between agricultural modernization to optimize production and the preservation of indigenous agriculture along with the genetic diversity found in those areas associated with agricultural origins and develop- ment. Agriculture is an early technology that has remained our most far-reaching impact on the planet. The quest for an assured food supply has done more to decrease biodiversity and physically alter the environment than any other activity in which we engage. Approximately 60 percent of the human population directly or indirectly makes their living from agriculture. Tragically, food production is population driven. As we produce more food, the human population becomes larger and the demand for increased yield creates an open spiral of greater impact on the land. Germplasm is the source of the genetic potential of living organisms. Among other things, diversified germplasm allows organisms to adapt to changing envi- ronmental conditions. No single individual of any species, however, contains all the genetic diversity of that species. ,., , . ' :, . : . ,:: : ' ' ' ' ' ' 4' ) ' 5~~. . ............. ........................................ : ' ' ' :' ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. ....-....a. .:.. ... . . :8B ''.:.::::B..: This means that the total genetic potential is represented only in populations made up of many individuals. Such genetic potential is referred to as the genepool. The potential represented in a genepool is the foundation for our crop plants in both agriculture and forestry. Germplasm is only maintained in living tissue, most often the embryo of seeds. When the seed dies the germplasm is lost. The limited number of plants that has historically fed the human population is approximately 1 percent of the flora of the world, and the number that have entered agriculture is a small fraction of that percent. As our human population has grown in number over the last two thousand years, and especially since the development of the science of genetics in this century, we have depended increasingly on a shorter list of the most productive and most easily stored and shipped crops. Today only about 150 plant species (Prescott-Allen and Prescott-Allen 1990) with about one-quarter million local landraces (Wilkes 1989) are important in meeting the calorie needs of humans. Extinction of a species or a genetic line represents the loss of a unique resource. This type of genetic and environmental impoverishment is irreversible. Throughout the world people increasingly consume food, take medicine, and employ industrial materials that owe their source to genetic resources of biological organisms. Given the needs of the future, genetic resources can be reckoned among society's most valuable raw material. Any reduction in the diversity of resources narrows society's scope to respond to new problems and oppor- tunities (Altieri, Anderson, and Merrick 1987). To the extent that we cannot be certain what needs may arise in the future (new plant diseases or pests, climatic change due to the greenhouse effect, and so forth): it makes sense to keep our options open. This conservation ra- tionale for future generations of humans applies to the Earth's endowment of useful plants more than to 2 almost any other category of natural resources. ..+eS 2+|g ........ *.<.a. 0+.0...... *E . S . . 0 The Challenge It is difficult to visualize a challenge more profound in its implications yet less appreciated by the general public than our food production system's dependence on sun, soil, water, and genetic resources (IUCN 1980). Food security and biodiversity will be our most obvious chal- lenges in the decade of the 1990s. How can we supply adequate nutrition to the 5.4 billion humans that exist now and the more than 6 billion that will exist by the year 2000? (United Nations 1989). To put the demands in perspective, in the first 20 years of the next century, one out of every ten humans that ever existed will be needing to eat. The food production of a single year will equal that of the entire century 1850-1950, and in the two decades 2000-2020, we will produce and consume as much food as we have since the beginning of agriculture eight to ten thousand years ago! Most of this population increase will take place in the developing countries where demand for food and agricultural products will double. The problems of a precarious food supply and rural poverty are ex- pected to aggravate pressure on scarce land for arable farming, deforestation for new lands will increase, and pushing systems beyond sustainable limits will result in increased habitat destruction. The human population will grow with short-term food increases and not be sustained by long-term "real" increases, so that by the year 2000 the FAO estimates (FAO 1981) there will be at least 600 to 650 million undernourished on the planet, most of them children! For the developed nations, population increases can be accommodated in part by eating lower on the food chain and consuming grains directly, but the developing world is already doing that. Because new arable land in the developing world will become steadily more scarce, higher yields mean using better agronomy dependent on a combination of more fertilizer, plowing, water lifting energy, and improved plant material. All but the last are agricultural inputs that compete for meager resources available in developing nations. Therefore, breeding for better crop plants will be the central focal point around 3 which all strategies to increase crop yields will develop. Human knowledge to grow and maintain the crop plants was essential in the early days of agriculture and so it will be in the high human population levels of the 21st century. It will be the positive responses of crops to agricultural systems, inputs, pests. pathogens, and so*- cial institutions that will determine the success of our attempt to feed ourselves and yet maintain agricultural systems that are sustainable (CGIAR 1989, 1990; Key-- stone 1991). Also agriculitural development must no1: come at the expense of any region or later generation,, nor threaten the remain}ing biodiversity of the plainet: as we progress toward a sustafinable society. Evolution under Domestication of Crop Plants The increase in the Earth's carrying capacity for humans has been made possible by the development of agriculture, which in turn has been dependent on the domestication of plants and their historical distribution well beyond their regions of origin. In the process of. domestication, cultivated plants have quite literally crossed a threshold from being wild in the natural veg- etation to being totally dependent on human care. In many cases these crops have been so genetically alteredi that they can no longer disperse their own seed Or- compete with weeds, that is they can't revert or escape to the wild. We call this selection process domestication. True domesticates have been so altered that they survive only under cultivation in agriculture or horti- culture. One example of a characteristic that has been altered is seed germnination. Farmers expect the seed of cultivated plants to germinate immediately after being placed in the soil. Such a characteristic usually would be lethal for a wild plant, which must have mechanisms to ensure germination only at the proper season, hence promoting survival. In cultivated plants survival is keyed to the farmer's commitment for seed bed preparation to .. . .. .s .N . .z.0.j.. ..t++.. +.* . . .0.. . +. ...* ++W decrease competition with other plants, the sowifig of seed, the protection of the plants during growth from pests and pathogens, and finally the collection of fruits, seeds, and other edible parts for human use. All major cultivated food plants have lost the ability to exist in the wild; in other words they are fully domes- ticated. These plants have been selected to produce unusually large plant parts, soft edible tissue, thick flesh with intense color, and fruits attached by tough stems. They are so altered genetically that they are dependent on us: to sow them in the proper season, to protect them from competition and predation, to supply them with water and nutrients, and to harvest their propagules in order to repeat the cycle. In the days of pre-agriculture, humans lived like any other animal in the sense that we daily hunted and gathered our food and on the days there was no success, we went hungry. Our density did not exceed 1 person per 25 square kilometers, and we were sustained on our forage territory. Today our density exceeds 25 persons per square kilometer, and in the urban zones around the world, human density approximates 600 to 1,200 persons per square kilometer with even higher densities in cities such as Mexico City, Rio, or Calcutta. Quite literally we are absolutely dependent on the yields of domesticated plants and animals and there is no turning back to hunting and gathering. Currently, we have synthetic fibers replacing cotton and linen, and synthetic rubber replacing natural rubber, however, as yet we do not have any synthetic foods replacing our basic crop plants: rice, wheat, corn, barley, sorghum, potato, sweet potato, sugar beet, sugar cane, common bean, soybean, peanut, banana, coconut, and cassava that account for three-quarters of all human energy worldwide. This list is primarily a calorie list and does not recognize the important role of low calorie vegetables and fruits in supplying vitamins, minerals, and protein to human nutrition. Also this list does not include regional foods that locally may supply more than 5 ..2 . S :{::y .......:. .:- :. v : v v . . . .. ... ::::>:: . ... " .." ...^ '. C '....." ' ... :' ' .' . ': ' W one-half of the calories consumed; and, in addition, both pasture forages and fiber crops are omitted. Still the point is clear, worldwide a very short list of food plants makes the critical difference. Agricultural Treadmill The human population has exceeded the carrying capacity for long-term sustainable agriculture without inputs in many parts of the world. Clearly, agriculture cannot continue on the treadmill of always producing more. In 400 human generations, we have progressed from wild food plants to high-yielding domesticates, but we appear at or near the limits. Two generalizations characterize the present world condition: * We are now in a state of diminished resources on a per capita basis. e Farming is no longer a subsistence activity where the cultivator eats and depends on the actual harvest, instead the cultivator depends on the money generated by the sale of the crop. With subsistence farming there are negative feed- back loops that limit production, but cash is a neutral (universal) commodity and cropping for cash has the tendency to be an open, ever expanding, feedback system that is clearly not sustainable on the planet in which we live. It is my personal contention that when the threshold was crossed from crops for subsistence to crops for sale agricultural becams destabilized. Selection for nutri- tional quality and respect forlocal habitat carrying capacity or sustainability were shed, and many of the problems we now experience began. This recent change to crops for cash has not been all negative. Cash crops can signifi- cantly increase the income and caloric intake of a farmer 6 with a smallholding. But the global exchange economy :.:.::,R::;,,..,.::: .:.. . v......... ................. ............a . z . . ... .. that has emerged, based on the principles of comparative advantage and specialization, has increased genetic uniformity at the expense of biodiversity. It also has created the reality of global interdependence based on money and not the quality of life or sustainable land use. The crossing of this threshold has created for us the imperative to anticipate future needs and better manage the crop plant genepools, because the natural mechanisms that once worked to create and maintain their diversity are no longer in operation. Genebanks and an increased cadre of plant breeders need to exist to replace the functions once carried out by the world's subsistence farmers. Telescoped Evolution Charles Darwin called artificial selection, telescoped evolution because what was normally a drawn-out affair in nature was condensed to a short period of time. Plants that are potential candidates for telescoped evolution or domestication must meet three criteria. They must be: * Capable of thriving in human rearranged environ- ments and often in full sun (helilophytes). * Yield productive over the short run in terms of land and labor and more recently for inputs: water, fertilizer, and improved seed. Excessive emphasis on yield ultimately leads to dense pure stands (monoculturesr and warfare with com- netitors (insects-insecticides, weeds-herbicides, and fungi-fungicides). * Genetically plastic and able to hybridize, mutate, and possess a range of genetic mechanisms that promote and maintain variability. The last criterion is especially important, because it is the one area where we can still make substantial improvements in crop plants (Harlan 1975). 7 Genepool Selection The genepool of crop plants and their wild relatives can be thought of as a bucket holding fluid.Under natural selection (normalizing selection mode) the genepool is comparable to a bucket sitting on its bottom. The surface expresses the phenotype, and the volume the genepool. The top surface is small (wild type genes) when expressed as a ratio to the total volume below (recessive multiple alleles). Under artificial selection by humans, the bucket is tipped, and some of the fluid is poured off. The top surface is larger, and the ratio of top area to volume is also greater. Many traits selected under artificial selection are recessive (non-brittle rachis in cereals for example), and once they are fixed, the allelic diversity of the genepool narrows to one. Artificial selection increases the pheno- typic diversity (easily observable traits-appearance) but often decreases the allelic diversity (actual genetic reser- voir). It is only on hybridization of two diverse forms that the allelic diversity is again broadened in the artificial selection process. This explains why plant breeders value good parents more than high-yielding progeny in their breeding plots. phenotypes ~~~~~~~~~~~~~~~~~~~.4N _ genepool t * *. .. ~ ~ ~ U~~) phenotypes Natural selection bucket of genes or genepool with limited "wildtype" dominant - "'i genes at surface. Artificial selection pours off many of the genes and the genepool is smaller but the top surface or phenotypes are more extensive. 8 1 *~ ~~~~~~~~~~~. .. .: ... ......... . . . . . :: . . . . The Story of Agriculture The story of agriculture over the last 10,000 years has been essentially the same in at least four independent siteswhere it began: Central Mexico/Guatemala, Andean South America, the Fertile Crescent in Southwest Asia, and Mid-Northern China (Heiser 1990). First there was human selection for unique traits and a proliferation of phenotypic variance as hidden recessive genes came to the surface through artificial selection and close inbreed- ing enforced by physical isolation and small population size (when compared to the more widespread wild popu- lations). Second the environment was restructured or rearranged, making it possible for the survival of geneti- cally lethal mutations and more importantly semilethals. In addition the seed was carried into new areas beyond its original native distribution as cultivation expanded, and alleles (diverse forms of a single gene) deep in the volume of the genepool bucket (hidden variation) rose to the surface, because they conveyed adaptation factors for the new conditions. The expanded cultivation following domestication brought the crop into the territory of wild relatives with which they were not genetically protected by isolating mechanisms, and inevitably hybridization occurred to produce a wealth of variation that has driven the selection process ever since (Simmonds 1976). Examples are the hexaploid bread wheats (6x) that can be thought of as three plants in one. Maize hybridized with teosinte, its closest relative, but unlike wheat kept the same chro- mosome number. Through back and forth or introgres- sive hybridization, genetic exchanges have occurred so that part of the teosinte genome is in maize and vice versa (CIMMYT 1988). Because of introgressive hybridization, teosinte is a 'little second plant" in the maize genepool (plant within a plant). Regionally distinct forms of cultivated rice hybridized and resulted in explosive bursts of variation (Chang 1985). Potatoes are a polyploid like wheat. Almost all major crop plants have developed a mechanism of one form or another to expand the gene 9 base and this has countered the tendency of artificial selection to narrow it (Wilkes 1989). History of Plant Breeding Historically, plant breeding as a human activity can be viewed as developing through three phases and we are currently on the threshold of the fourth. The earliest domesticated crops were probably not much more pro- ductive that the wild progenitors, but the act of cultiva- tion was a radical break with the past as artificial selec- tion was applied to small isolated populations. This was the first stage of plant breeding or human control over crop plant evolution. All the major world food crops were developed in this first stage (Hawkes 1983). The second stage* of plant breeding came with the discovery of the New World or Columbian Exchange that followedat e circumnavigation of the world. A rapid diffusion of crops, livestock, and farming techniques coupled with emigration resulted in tremendous genetic recombination as distinct and far-flung crops or distinct landraces were brought together and hybridized in farmers' fields. This era also saw the development of colonial empires in the tropics and cash plantation crops such as coffee, cotton, rubber, and sugar cane. Many of these cash crops, such as coffee, were moved from the Old Wisorld to the New, where free of their diseases, they became more productive: or the reverse, NewWorld to the Old, such as rubber. These plantation crops were the first modemn cash crops. Now all field crops, with the possible exception of kitchen gardens, are cash crops. The third stage of plant breeding and improvement began with the rediscovery of Gregor Mendel's classic experiments on the heredity of garden peas at the beginning of this century. For the first time the plant breeding community had a set of principles by which to proceed with the crop improvement process. Products of this era are hybrid corn, changes in the photoperiod response of 10 soybeans, and the dwarf stature wheat and rice from the Centro Internacional de Mejoramiento de Maiz y Trigo (CIMMYT) and the International Rice Research Institute (IRRI) respectively. These late 1960s Green Revolution cereals (dwarf stature wheat and rice) and the genes they hold now enter the food supply of 2 billion people and are directly responsible for feeding more than 800 million people by their increased yields. The fourth era of plant breeding, genetic engineer- ing, promises to have an impact equal to that of computers in the way we go about managing and structuring the productivity of the world around us. Traditional plant breeding is, in fact, genetic engineering, but this term is now being limited to biotechnologies such as in vitro cell culture, and recombinant DNA techniques. These bio- technologies can speed the breeding because genetic material is introduced directly into cell cultures (splicing), completely sidestepping the genetic recombination through the usual sexual processes orwhen cell cytoplasm is altered as in protoplast fusion. And, of course, all of these changes depend on cloning technologies where hundreds of identical plants are grown from units as small as a single cell. Already potatoes developed at the Centro Internacional de la Papa (CIP) using these tech- niques are having a significant impact worldwide and superior clones of cassava are leaving the Centro Inter- nacional de Agricultura Tropical (CIAT) for Africa and Asia. The advent of these technologies has ramifications that affect our current state regarding plant germplasm. Because these advances are produced using unnatural means they can become intellectual property and thus are potentially protected by the laws of ownership. The specter of patent protection has polarized the public views of germplasm as either public good (common heritage) or private good (rewards for value added). On a broad brush, the developed world has the skill to enter the biotechnology game, but because the developing world mostly lacks the resources (personnel, institutions, and finances) it is fearful of what these new technologies hold for it in the future (Juma 1989). Much of this fear is reflected in an Intemational Undertaking on Genetic 11 ~~~~~~~~~.. R'' .. ' . . .. ... .. .. :: ... . Resources of FAO (Witt 1985). This has been good because it has focused attention on why plant genetic resources have been undervalued, but unfortunately the discussions and solutions sought have been more politi- cal and have not led to an accelerated implementation of strategies necessary to conserve and use crop diversity (Wilkes 1988). These strategies center on five scientifi- cally based issues: 1) genetic erosion, 2) genetic wipeout, 3) genetic vulnerability, 4) genebanks, and 5) training for plant breeding. Genetic Erosion The technological bind of improved varieties Is that they have a tendency to eliminate the resource which they are based on and from which they have been derived. Current elite varieties yield better than the varieties they displace. Once a displaced variety is no longer planted, its genes are lost to future generations unless it is conserved. Primitive forms also are lost because of bad land use planning, environmental degradation, and ur- banization. Elite varieties also have a second force: they create market expectations. Once a highly uniform variety captures a large fraction of the market share, other varieties are bred to mimic or have the same attributes as this leading variety. This further eliminates diversity within the crop and even across crops. The market place also has the potential to influence and form a market focus for a single variety. These marketing forces of 'volume sales' and specialized handling of spe- cific varieties actually have promoted the decrease in the total number of crop plants that enter commerce. Long- distance transport focuses on a limited number of com- modities and on only certain crop varieties, with the result that local producers and small unique variety suppliers are forced out. In developing and developed countries the deep well for the genepool in landraces and in folk- or farmer-maintained varieties, which has been the foundation of the past breeding process, is disap- 12 pearing. When a tree falls in the forest there is a crash; when the seed of a unique variety is no longer planted it is a silent loss. Like soil erosion, it is gone with no drama of disappearing; genetic erosion leaves a void and a diminished genepool. Genetic Wipeout The wholesale loss of plant genetic resources is called by some, "genetic wipeout," a somewhat emotive term but useful in drawing attention to the problem. Genetic erosion is a slow gradual process based on the independent individual decisions of farmers while genetic wipeout is the rapid and one stroke destruction of genetic resources, usually by institutional failure. Social dis- ruptions such as political instability or crop failure and famine due to natural disasters can eliminate genetic resources rapidly and have done so repeatedly in the past. Quite literally the genetic heritage of a millennium in a particular valley can disappear in a single bowl of porridge if the seeds are all cooked and eaten instead of some being saved as seed stock. Equally dramatic is the discarding of a genetic collection because a curator retires and the collection is no longer of use to the institution. This is especially tragic if a comparable collection can no longer be recollected because of genetic erosion in the countryside. The processes of genetic erosion and genetic wipeout are not mutually exclusive but are, in fact, two ends of a spectrum interlocked by the demands of an increasing human population underwhich biodiversity impoverishment increases daily. Genetic Vulnerability To be vulnerable is to be at risk. One of the attributes of humans that sets us apart is the ability to learn from the past, to anticipate the future, and to be aware of risk. Another attribute is our ability to adapt and modify our behavior to minimize risk and decrease our vulnerability. We have been so efficient at food production that we tend 13 ..'X... .i":.".5.'.'. '"...; 'y to forget how much of the natural ecosystem we have rearranged and also forget there is vulnerability in that agroecosystem until a crisis-drought, crop disease, or pests-reminds us, that we are dependent on a food supply to sustain life (NAS 1972). What is meant by genetic vulnerability? 'Genetic" is clear, determined by genes, but "vulnerability" is more open to interpretation. The dictionary definition for vulnerability is "unprotected from danger." Thus genetic vulnerability means the extent to which crops are unpro- tected from danger by the genes they carry or in some cases lack them altogether. The danger is usually consid- ered to be the potential damage from unsuspected patho- gens and pests. These may be new strains or biotypes of a species already known to be able to attack "susceptible" varieties, or they may be pathogens and pest species that are encountering a new host species orvariety for the first time. The fact is epidemics hit the hardest if the host species is highly and uniformly susceptible and also present in dense stands of individuals. The host species may be either genetically very diverse overall (as varieties or a individuals), or genetically uniform overall, but if it is genetically uniform for susceptibility, then epi- demics may result . Such was the case for the single gene in the Texas male sterility cytoplasm of maize (1970 U.S. corn blight epidemic), which conditioned a highly suscep- tible reaction. No matter what genetic background that gene is in, an epidemic with race T of the corn blight pathogen will occur, given the right environmental condi- tions. In this case the gene giving susceptibility was an "agronomic gene." There was no Threakdown" of a resis- tance gene. No matter what the genetic basis of suscep- tLAJlity the three major hazards are pathogens (fungi, viruses, and so forth), pests (insects, nematodes, and so 14forth), and abiotic stresses. 14~~~~~~~~~~..... .. I ~ ~ ~ ~ ~ ~ ~ ... .. .... %", M M i*~~~~~~~ . . +++00+-+Z+++{W ................................. The key point about epidemics is that thek are promoted by dense stands of individuals containing the same gene (or genes) conditioning susceptibility, regardless of their overall genetic uniformity or diver- sity. The magnitude of the epidemic will be conditional to the geographic distribution of the gene (or genes) for susceptibility and the number of humans dependent upon the susceptible variety as a food source. Genetic vulnerability is the potentially dangerous condition of "thin ice" of having a narrow genetic base. Never before have there been such widespread monocultures (dense uniform stands of billions of identical or nearly identical plants) covering thousands of hectares. The narrowness of the genetic base is responsible for, on the one hand, the predictability of higher yields, and on the other hand, the greater risk of crop failure. Crop specialization and genetic uniformity do not create vulnerability. What they foster are the rapid inroads over vast areas if a vulnerability does develop. Genetic vulnerability is in fact a market phenomenon because farmers make the choice of which variety is best for them. If a widely planted high-yielding elite line crashes under the stress of a severe drought, the impact is large because the variety responding to the stress accounted for a sizeable percentage of the crop acreage. Neither the genetic uniformity of the variety nor the vast acreage caused the drought. Given the drought and the susceptibility of the variety to drought stress, the impact of vulnerability was magnified. Given a normal year the improved variety results in magnified yield, which is the desired outcome, and the reason certain varieties are so widely planted. Some widely planted and genetically uniform varieties have remained durable for decades and other varieties have folded the year they were released. This push for genetic uniformity over vast areas has increased the potential for genetic vulnerability as plant breeding of important commodities has become inter- 15 nationalized first from the International Agricultural Re- search Centers such as CIMMYT, IRRI, and CIP and now from private seed companies that sell seed around the world. Currently, the potential for a crop failure hinged to genetic vulnerability is considerable, and there is significant reason for concern. Genetic diversity in time (the sequential replacement of cultivars) is the chief substitute today for genetic diversity within cultivars (or among crop species) and it requires a bountiful support for dynamic plant breeding activities. Genebanks and Germplasm Genebanks are an institutional solution to genetic erosion and genetic wipeout. Although a degree of in situ conservation can preserve parts of the genepools of crop plants, to date this mode of conservation has not been implemented broadly enough to be as successful as ex situ genebanks. To be successful genebanks must have four distinct functions in addition to preservation of seeds and clones to counter loss due to genetic erosion. They must: (1) Be linked to exploration of undercollected zones to increase and maintain representation of the genepool as samples in the genebank. (2) Maintain the genetic integrity of samples, storage over time, and regenerate stocks periodically. (3) Evaluate and document to maintain a useful data base to guide management and use of the holdings. (4) Be linked to active evaluation, prebreeding, and early breeding for enhancement so that genebank mate- rials will be in a useful form for the plant community. Unless there are more plant breeders working with genebank materials to mix and evaluate the genes, agri- culture with elite varieties will undermine itself and fail. 16 [ARCs and Their Crop Responsibilities Year Crop Institute Established Location Responsibilitv International Rice 1960 Philippines Rice Research Institute (IRRI) International Maize 1966 Mexico Wheat, maize. and Wheat Improve- triticale ment Center (CIMMYI1 International Center 1967 Colombia Beans, cassava, for Tropical Agriculture rice, tropical (CIAT) pastures International Institute 1967 Nigeria Cassava, maize, of Tropical Agriculture plantain, cowpea, (IITA) soybean, yams International Potato 1970 Peru Potato, sweet Center (CIP) potato West Africa Rice 1970 Cote d'Ivoire Rice Development Association (WARDA) Asian Vegetable Research 1971 Taiwan Selected tropical and Development Center and subtropical (AVRDC) vegetables International Crops 1972 India Sorghum, pearl Research Institute for and finger millet, the Semi-Arid Tropics chickpea, pigeon- (ICRISAT) pea, groundnut International Livestock 1974 Ethiopia Browses, grasses, Centre for Africa (ILCA) legumes International Center for 1975 Syria Barley, lentil, Agricultural Research in faba bean, durum the Dry Areas (ICARDA) wheat, bread wheat, chickpea International Centre for 1977 Kenya Multipurpose Research in Agroforestry trees (ICRAF) International Network 1984 France Banana and for the Improvement of plantain Banana and Plantain (INIBAP) 17 .>~~~~~~~~~~~~~~~~~~~~~~~................ .s ...*... < Seed is the most common mode of genebank preser- vation. Seeds age in even the best genebank and ulti- mately die sooner or later depending on the species. Periodically the seed must be planted, populations maintained through controlled pollinations, and new fresh seed stock returned to the genebank. A seed genebank is generally the easiest to establish and does not incur a significant cost or management crisis until years later when the need for scientifically based regen- eration develops. To better understand genebanks requires an ap- preciation of the kinds of genetic materials that can be saved as seed. For convenience, germplasm can be organized in six distinct categories (Wilkes 1977) based on their station or advancement in the agroecosystem. * Varieties or cultivars in current use, often very advanced elite varieties. * O2bsolete cultivars, often the elite varieties of 20 to 50O years ago and usually found in the parentage of current cultivars. * Primitive cultivars or landraces of traditional agri- c>ulture. * Wild and weedyJ taxca near relatives of crop plants. * Special genetic stocks, including induced mu- tants. * Coadapted genetic stocks, in which two forms of a crop, two distinct crops, or a crop and symbiont such as a crop and its unrelated weed or nodule- forming bacterial are grown together. Varieties or cultivars in current use have generally undergone a vigorous selection process by plant breeders for plant type, response to input, predictability of yield and are more or less homogeneous genetically. The 18 released varieties possess a Thighly tuned" set of elite 4he~~. ... .. ....vA..... *AE s0 genes but a considerably narrowed gene base against the foundation parents from which they have come. Ulti- mately these foundation stocks trace back to landraces in diverse parts of the world. The released advanced vari- eties most favored by farmers are the ones most widely and frequently used as partners in current breeding programs for the next cycle of varieties. On average, varieties are replaced every 5 to 10 years, so they have a commercial lifetime of maybe 7 years. Some released varieties are knocked down by either pests or pathogens the year they are released, so there are no guarantees on the durability of new varieties. Obsolete cultivars are advanced cultivars from the most recent past that have been displaced by a newer release. Often specifically selected oldermaterials appear in the pedigree of a wide variety of releases. Some obsolete varieties are more useful as parents than they were as varieties at the time of their release. The Wheat Genebank at CIMMYT is holding a large collection of obsolete varieties that can be computer searched through a pedigree chart data base. Both special genetic stocks and induced mutational stocks, which are used in the breeding process. are comparable to obsolete varieties in the amount of genetic variation they possess. Prirnitive cultivars or landraces are the real treasure house because they are the largest depository of genes for a crop. They are also the largest unknown because they are usually heterogeneous genetically and few data exist on their morphological, biochemical, and genetic traits, or their responses to pests or environmental stresses. Landraces are a rich source for new traits but usually exhibit narrow local adaption. Landraces seldom, if ever, make acceptable broadly adapted varieties. Generally, landraces perform poorly under high inputs of fertilizer, water, and intensive cultivation and are replaced by the new elite seed. On the other hand, there is a fairly wide variation in the ability of landraces to survive fluctuating environments, to withstand cold, drought, insect damage, and other such variables. After all, most 19 landraces represent accumulated mutational events in- tegrated and balanced in the real world over centuries. It is with the genes landraces possess that modern plant breeding has formed the current elite varieties. The wild relatives and weedy taxa are poorly repre- sented in genebank collections and the most difficult to regenerate. This is because these ancestral forms and wild relatives closely related to the no longer extant ancestral taxa do not come from cultivated fields but come from unique and often highly specific wild habitats. These taxa self sow their seed and exhibit genetic adap- tation in an often very narrowly defined environment for day length, soil, and water relationship. They possess many dominant wild type alleles and carry rare recessive alleles at very low frequencies. It is almost impossible to regenerate the genetic integrity of the wild collected seed because the site of regeneration does not match the site of collection. There are regeneration site selection forces and it is difficult to regenerate a population of sufficient size to capture rare alleles with confidence. In most genebanks the wild and weedy forms are simply a form of wild populations and not a broad enough net to capture the genetic diversity of these unique genetic systems that have genes compatible with but not found in the crop plant. To compound the problem most plant breeders have limited interest in the wild and weedy relatives and their biology is not as well known as the crops. Most of the wild r-elatives (and there are certainly exceptions such as wheat, barley, and maize relatives) have only been collected along roadsides and the full extent of their populations is not fully known. These plants, which are neither culti- vated nor of interest to most herbarium taxonomists, have been understudied. Certainly they should be in genebank collections and it is important to conserve them ex situ as seed. The strongest case for in situ conservation can be made for wild relatives in their native habitat. The speciatl genetic stocks are usually carefully con- 20 structed genotypes such as seed lacking a specific >||s~~.. ++........... . chromosome or chromosome markers with recessive alleles on the chromosomes that are the tools of the plant breeder. These stocks make up a small fraction of the total genebank but they have been useful in the past and certainly have utility for the future. Many have been constructed through long selection processes and these tester stocks belong in the collections. The last category is seldom found in genebanks and comprises materials that the future with biotechnology tools will want to exploit in new and novel ways. These are two genetic systems that express mutual coadaption or coevolution. This can be two varieties of the same crop with the same growth habitat, height, planting, and so forth, but one is long season the other short; two distinct crops again with the same growth pattern (maize and sorghum grown together in Honduras where with sufficient water there will be both a maize and sorghum harvest and in dry years only sorghum). The most extreme pattern of these coadapted systems is a crop and its nodule-forming bacteria or unique pollinator insect. The tropics have numerous examples of these coadapted systems in relay planting, double cropping, and perennial agricultural systems that haven't been adequately captured in genebank collections. Some of these coevolutionary patterns are not even known to us and the hope is that insituconservationwill preserve them until we better understand how they work and can include them in genebanks. In Situ Germplasm Conservation The need for in situ conservation is really a relatively recent event. One hundred years ago humans were packed less densely, communication was localized, and our demands on our surroundings less harsh. All useful plants and animals were maintained in the open world of fields and forest (Vavilov 1935). Only recently have we engaged in ex situ conservation of crop plant seed in cold storage seed banks or in situ preserves. Before the world was so restructured by our activities these functions took 21 care of themselves (Altieri, Anderson, and Menrick 1987: Oldfield and Alcorn 1991). In situ conservation in perpetuity can take two somewhat distinct but overlapping pathways: * Entire Biomes -the entire preservation of vast tracts with the in situ conservation of animals and plants. This is a continuum from vast tracts of wilderness, totally devoid of humans, to unique habitats with considerable management and pro- tection by humans. This level of preservation will be extremely important in slowing the species extinction rate but will have little impact on regenerating degraded habitats or habitats that humans exploit as a resource base. Conservation at the biome level works to get humans out of the picture and does not promote humans at the pivot point of the system. * Specilc Sites - preservation of those useful plants and animals in a dynamic interaction that does not lead necessarily to a decreased carrying capacity of the human population. This also is a continuum from the tracts necessary to sustain a hunting-gathering society, to shifting-milpa agri- culture, to cultivated fields where the local crop landrace hybridized with a field margin, weedy, wild relative. The essential elements are that the plants and animals are found useful by humans and the human density does not increase to the point where the resource base is degraded. In truth these idealized systems will not be easy to achieve by in situ conservation but will deaccelerate their disappearance and give us more time to understand and appreciate how these systems evolved. Considerable potential exists for creative institutional arrangements including tourism for in situ conservation, especially for historical sites 22 and in the developing countries. .i.;. . ..... ..|......... . z FNg. .... ..... .. ....++ . zz+f^+. .e Levels Of In Situ Management So as not to lose sight of the full context of genetic resources, and the role of in situ conservation, I've broken the whole into fourmajorgroupings. These four are listed in terms of decreasing intensity of human management in the context of in situ preservation (Plucknett and others 1987: Alcorn 1991). * Cultivated plants and Living historic farms and sites domesticated animals rich in historical significance to the national heritage * Wild relatives of Evolutionary gardens and domesticated gardens of chaos in centers plants and animals of crop diversity and origin - often showing peasant farming at best, in situ monitoring of wild populations: this level might also include 'hot spots" for pests and pathogens of selected crop plants. * Wild species used These are the 'extended directly kitchens- of many peasant farmers that extend into the forest; the species may be wild but their survival managed. *Wild species currently Maybe they are still not used by people unrecognized. Because of the large number of accessions in the major crop collections (2 million), there is no way that an in situ system could be employed for all. In essence such a mandate would freeze all landrace in the genetic landscape of peasant farmers, placing much too high a social cost on peasant farmers. But a limited number of selected landraces in key sites would be both feasible, practicable, and appropriate. These key sites would be especially useful if they were located in "hot spots" where 23 the diseases, pests, and pathogens of the crop were also evolving under the pressures of natural and artificial selection. These would remain evolutionary gardens for continued dynamic genetic change (Hoyt 1988). These selected in situ stations would highlight the issue of genetic resources because of their location and educational function. To be effective, such in situ con- servation of landraces needs to be coupled with both tourism and research. A visit to the Mayan ruins at Chichen Itza along with a reconstruction of a milpa carefully researched by the Mexican Escuela Nacional de Agricultura would impress affluent tourists about the primacy of plants. The same could be done for potato and other tubers In the Peruvian highland agriculture at Inca Ayacucho. In the Old World, village India at the fertility temple sites around Khajuraho would be an excellent site. The elements these in situ programs have in com- mon is a living connection to the past. Living plants on display are a shared heritage from the past for all people. These are the crops that feed us and support our cities and our ability to live at high density. In making the connection between the past and present we make our identity to the past. Mexico can be proud of maize, its contribution to the world community; India its contribu- tion of sugar cane and rice; China its contribution of soybeans, rice, and tea; and the list goes on for about 500 significant food plants that support a human population of 5.4 billion. Yet the most significant service of in situ landrace preservation is not the heritage from the past, or the pride of the present, but the sense of direction. The living farms with their heritage germplasm of in situ landraces wouldn't preserve the genetic landscape as much as they would instill pride in the plant domestication accomplishments of the indigenous peoples in the past. These farms wouldn't focus onjust one crop but the mix of different crops and the cropping pattern. This is what 24 in situ can do best, whether it be the crop and its wild ... .. .. . F . . . . . . .. ... .. . .. . ..::: . < : :: S : : .. f:: S: : ): : . ~~~~. ..... .. :.'. relatives hybridizing or the interaction of two genotypes of the same crop or two distinct crops interacting in the same field. The second case for in situ conservation is evolu- tionary gardens. These are the great natural resource bases from which our useful crops have developed. These are the fields where the crop and its weedy wild relatives on the margin continue to hybridize and where there is the dynamic evolution through introgressive hybridiza- tion or changes in ploidy level. Teosinte in Mexico is a classic example of such a relationship. These systems are analogous to a still where the boiling of the mixture is hybridization and where the distillation by both artificial selection (the crop) and natural selection (the wild rela- tive) results in a dynamic generation of genetic refine- ments in the plants. In situ preservation of these sites would allow this system to continue for future evolution of the crop genetic system under full selection pressure of both pests and pathogens (Wilkes 1971). We know so little about the domestication process, which has taken place in the last 10,000 years, the pre- history, for most of the major crops. These in situ zones of crop origin are windows on the past that are being closed by changes in land use. Until a few years ago it was possible to find in the Nordic countries two distinct barleys, one early, the other late, with identical stature planted together in the same fields. Presumably these are in the Nordic Genebank but the knowledge of which two accessions go together is lost! Also there is a pea and barley combination that has been lost. In Mexico we still have the corn/bean/squash coadapted systems in place but they will not remain given the current economic priorities. Future generations might thank us immensely for 'holding the line" and preserving 500 of these systems worldwide for future study. Right now, when seed from these locations (weed and crop) go into exsitugenebanks, 25 ..... Bai.... :>:-W. . ....... many times the two are not cross-referenced. One of the few collectors who has been veiy sensitive to this problem is Professor Angel Kato-Y in Mexico (Kato-Y 1976). Now we come to the other side of the evolutionary gardens where the focus isn't on a single crop but on all the plants that the indigenous inhabitants use in the landscape. This is an evolutionary sequence also because some of the plants are full domesticates, some are incipient domesticates, others are gathered regularly, and some only occasionally. To find a society at stability using the vast knowledge oftheir surroundings is the ethnobotanist's dream. This way of life only exists in a non-cash society and so most have disappeared and the rest are disap- pearing. The information potential here is tremendous. These are the "Gardens of Chaos" of EdgarAnderson (Anderson 1952) where the agricultural extension officer considers it his moral responsibility to set the garden in straight rows and generally clean up the act. Edgar Anderson correctly saw everything as being in its place and balanced like a miniature ecosystem where rela- tionships were defined by webs rather than straight rows. Often these gardens have no clear line where native vegetation begins and indeed the wild plants are managed for useful products. Books can list the species used but they have a hard time drawing in all the lines of connection that is the web of these self-sustaining peoples and their plant and animal cohabitants of the "hearth." The evolution here is of the entire habitat and only in situ will capture this for genetic conservation. I think the strongest case for in situ conservation of useful plants in a self-sufficient peasant society is as a teaching tool to reestablish the primacy of plants to an industrial model/style of thinking urban dweller (you have seen one corn plant you have seen them all), who is entrained on the supermarket. These decisionmakers have forgotten that civilization depends on an assured food supply. They have forgotten or never really knew that yields and harvest are fraught with the variance of 26 weather, pests, and pathogens. In situ preservation of Row Crops and Garden Crops Field Crop -K> - Dooryard \ '.% Seed Agriculture Vegetative Culture Yield to Work Expanded High Low . I \ Area of World Temperate Tropical 3 Plant Life Cycle Seasonal 12 months and more Biomass Monoculture ' ; ; X Diversified Small 7 1 Large -'*. Stratification Open to sun Sun and shade mixed Mineral Cycle Prairiesoils / Lateritic soils ~ Bare ground f ' Harvest and planting . - Open 1 / same operation ' '-/ /S /1 More nearly closed -, , Use of detritus Predator Control Physical Biological Seasonal Peaking Morepronounced Less pronounced Periods Harvest storage Continuous harvest problems v - Plants Used Many with high - Fewer with high protein X protein yield yield . '., ,.:,........... .... . . ..................... ................. %................. .............. . .. self-sustaining agroecosystems is one of the best way's to demonstrate that environmental impoverishment is irre- versible. There is a saying I like that sums up this issue: "The law locks up he who steals the goose from off the common, but lets loose he who steals the common from the goose." We can't stop population growth and the simplifi- cation of ecosystems to planted monocultures but we don't have to have these conditions everywhere. In situ preserves are thresholds to what the world once was and hopefully they can be models to how it might be in the future. This doesn't mean that we all will go back to hunting and gathering. It means that by preserving some of the remaining human sustaining biological systems in place we will have models that work to guide us in the quest for sustainable agriculture. Dwarf stature wheat begot the idea of dwarf stature rice. Likewise gardens of chaos begot relay planting of crops now successfully deployed in Asia. The multilines with the same stature but differing maturities, pest resistance, or stress-tol- erance are all modeled in gardens of chaos. Coadapted different taxa are ideas that biotechnology talks about in, cell fusion across taxa lines yet these systems already exist in many village dooryard agroecosystems. These in situ "extended kitchens" have considerable information and their preservation is best justified if they are well- studied. Unfortunately, they are disappearing because much of the agricultural research work is following maximum yields and increasing dependence on the products of plant breeding (Wilkes 1971, 1991). Genebanks and Storage To be stored safely for a long time, living seeds and the genes they hold must be at a low moisture content. A large number of seed crops may be stored for up to 25 years or more at refrigerated temperatures just above the freezing point. These are called "active storage conditions." Longer storage periods are maintained at subfreezing 28 temperatures. At temperatures of - 18/209C safe storage *~~~~~~~~~~~~~~~ ~ ~ ~~~~~ ~ ~ ~ ~ ~ .. . . .. . v..;.......... g0 R -g for 50 years or more is the norm. These are called "base storage conditions." At the temperatures of liquid nitro- gen, super dry seed (approximately 6 percent moisture and in some cases as low as 3 percent) is expected to remain viable for up to 500 years, and certainly periods of 100 years is a reasonable time frame. The number of national and institutional genebanks with refrigerated active collection storage conditions exceeds one hundred. The number of genebanks with base collection storage facilities is less than forty. Not all genebanks have gotten around to actually regenerating their collections with rejuvenated fresh seed. Creating new genebanks does not correct the problem. It is only when genes increase the diversity in farm- ers' fields through the characterization, evaluation, and prebreeding process of growing the seed and searching for useful traits that we can say the tech- nology of genebanks is really dependable enough to replace the older system of farmer held seed. The concept of genebanks looks good on paper, but so far only the banks of the International Agricultural Research Centers, numerous developed country, and only a few developing country programs have the system in place and have a proven record of moving the genes back into farmers' fields (Table 1). Approximately 75 to 90 percent of the variation in the major crops and less than 50 percent for many minor crops is found in genebanks (Plucknett and others 1987; Cohen and others 1991). The largest genepool is found in the silently shrinking landraces and folk varieties of indigenous and peasant agriculture. Increasingly, the centers of genetic diversity for crop plants have become the mega-genebank storage facilities and not the coun- tryside. Genebanks are filled with seeds on the hopes that they will be useful, however, until these banks undertake the evaluation of their holdings I prefer to call them storage facilities. Worldwide there are far more storage facilities than genebanks at the present time. Using my concept of working genebanks with regenera- tion and evaluation capabilities few models can be found 29 5,.,,,. . ... . A.,. . .. .. .. .. .. Availability of Genebank Accessions to Potential Users, USA1 to be regenerated sufficient for available for per year distribution regeneration NPGS Barley: NSSL 721 436 2 NSGC 26,168 5,000 22,175 Maize: NSSL 21,671 1,865 2-7 NoRPIS 8,783 450 7,000 Potato: NSSL 3,262 86 1 IRI 4,272 300 4,000 Rice: NSSL 942 205 1 NSGC 16,010 5,000 13,899 Tomato: NSSL 1,984 1,202 1 NERPIS 5,615 300 5,500 Wheat: NSSL 1,597 619 2 NSGC 42.478 5,000 36.932 Totals NSSL 29,674 4,413 (14.9%) Regional 103,326 16,050 89,506 (86.2%) Table 1. Percentage of accessions from the National Seed Storage Laboratory (NSSL), Fort Collins, Colorado, and regional plant introduction conservation facilities available for distribution after regeneration. Values shown are num- bers (of accessions or locations) and percents of totals. Notes: 'Acronyms for regional plant introduction and collection facilities are as follows: NSGC-National Small Grains Collection; NCRPIS-North Central Regional Plant Introduction Station; IRI-Inter-Regional Potato Introduction Project: NERPIS-Northeast Regional Plant Introduction Station. Values for NSSL indicate number of unique accessions not yet in the regional plant introduction facilities. 'Numbers from Cohen and others 1991. The United States is used here illustratively because it is a coordinated multicrop system. The collection in Gatersieben, Germany. for instance, could have been used. 30 * .S... jo . 0.+++++. . .~~~~~ ~ ~~~~ ~ ~~~~ ~~~~~~~~~~ ..g .+ .... - - .. . .... outside the lARCs and the developed nations. The IARCs have the envious record of regenerating, evaluating, using, and maintaining their collections (Table 2). Looking at Tables 1 and 2, any genebank that can distribute 85 to 90 percent of its holdings at any given moment is fullyfunctional. Always there will be shortfalls for difficult to regenerate and/or widely called for seed and hopefully these can be supplied within a year or two. Remember also there are differences in field space and human effort due to pollination system differences. Not all crops are the same. To regenerate a mostly self- pollinating crop like wheat takes little space and only a few plants to regenerate an accession. To regenerate an outcrossing plant like maize requires at least 256 plants and handmade chain pollinations (different male and female plants) of at least 100 plants to catch all the alleles in the accession with a frequency of higher than 5 percent. Genebanks that can provide only 50 to 85 percent of their accessions at any given time are in need of upgrading. Genebanks that can provide less than 50 percent of their holding are in serious trouble. Many genebanks around the world are in this last category. The good part is that there is still time to upgrade the situation with many national collections before germplasm is lost. Realizing our dependence on landrace genetic re- sources creates a sense of humility, which in the arro- gance of our accomplishments, we have often ignored. In the words of Sir Otto Frankel: "To an unprecedented degree, this decade of vast consequence for the future of our planet is in the hands of perhaps two or three generations.. .no longer can we claim evolutionary innocence.. .we have acquired evolutionary responsibil- ity' (Frankel 1974). Sir Otto has been very blunt. If we know what we are destroying through negligence and inaction, then we are morally responsible. Plant breed- ers, conservationists, and international policymakers need to start pulling in the same direction to save our crop plant genetic heritage (Kloppenburg 1988: Frankel 1989). 31 .. h... i, ..: . .. . .. . .. . .. . .. . .... . . . . . ...... .. Availability of IARC's Genebank Accessions to Potential Users' Center Crop [Accessions Accessions able [Accessions Location to be regenerated sufficient for available for per year distribution regeneration AVRDC Mungbean 5,273 1,000 3,200 2-5 Pepper 3,471 250 1,000 2-5 Soybean 12,303 1,000 8,000 2-5 Tomato 3.814 350 3.000 2-5 Total 24,861 15,200 (61%) CIAT Phaseohis 25,000 1,800 20.000 3 (bean] Tropical pastures 21,000 1,800 14,000 3 Cassava 4.500 In vitro In vitro Total 50,500 34,000 (74%) CIMMYr Barley 7,200 2,000 6,480 3 Bread wheat 48,600 8,000 43,740 3-7 Durum wheat 15,300 3,000 13,770 3-7 Primitive wheat 4,320 4,000 3,888 3 Triticale 11,700 3,000 10,530 3 Wild relatives 2.700 500 2.430 3 Total 89,820 80,838 (90%) Maize 13,346 250 10,910 4-8 Teosintes (annual and perennial) 93 In situ Tripsac-n spp. 90 Clonal 1 Total 13,529 10,910 (81%) CIP Potato (Clonal, 4,500 3,500 clones 3,500 seeds 4-9 in vitro, seed) 300 seeds 650 clones Sweet potato 5,507 2,000 clones 50 clones 1 (donal, in 300 seeds vitro, seed) - Total 10,007 4,200 32 _ (40%) Availability of IARC's Accessions (cont'd.) Center Crop [Accessions Accessions able Accessions Location [to be regenerated sufficient for available for per year distribution regeneration ICARDA Food legumes 17,900 2,000 10,740 1 Total cereals 43,700 3,000 26,220 1 Forage legumes 22.000 2,000 13.200 1 Total 83,600 50,160 __ (60%) ICR1SAT Groundnut 12,841 1,500 12,500 10-15 Pearl millet 21,919 2,000 21,800 10-15 Pigeonpea 11,482 1,300 11,300 10-15 Other millets 7,082 1,000 7,000 10-15 Sorghum 32,890 2,600 32,600 10-15 Chickpea 15.995 1,400 15.600 10-15 Total 102,209 100,800 (9 9%]/ IITA Musa 412 412 412 2-7 Oryza species 12,500 4,500 9,000 3 Vigna 15,200 4,000 9,439 3 unguiculata Wild Vigna 1.450 1,000 1.000 Total 29,562 19,851 (67%) IRRI Rice 86,000 10,000 77,400 2 (90%)** WARDA Rice 5,430 1,000 3,000 3-16 ____ _________ ___ _ _ _ _ (55%] Table 2. Percentage of accessions from crop germplasm collections available for distribu- tion following regeneration. Values shown are numbers (of accessions or locations) and percents of totals: accessions usually available as seed except where noted. Notes: *AVRDC-Asian Vegetable Research and Development Center; CIAT-Centro Intemacional de Agricultura Tropical: CIMMYT-International Center for the Improvement of Maize and Wheat; CIP-Intemational Potato Center; ICARDA-International Center for Agricultural Research in the Dry Areas; ICRISAT-Intemational Crops Research Institute for the Semi-Arid Tropics; IITA-International Institute of Tropical Agriculture: IRRI- International Rice Research Institute; and WARDA-West Africa Rice Development Associa- tion. "Of its total 86,000 accessions. IRRI has successfully canned 43.500 accessions for medium- and long-term storage. 'Numbers from Cohen and others 1991. 33 .. ..,.. ... x: . .. ........... .g . There is a very strong case for the plant breeding community in the LARCs to take on a leadership role in training and doing evaluation prebreeding and moving genes from the bank to the breeders' plots. There are few other institutions other than LARCs as well positioned (including universities) to fill this obvious void in the plant breeding process. The world goal should be a cadre of plant breeders at the national and local levels in the ratio of 1 to 106 human population. To achieve this level means a tenfold increase in the skill of producing new plant varieties. Genebanks Slow Down Crop Evolution Genebanks have additional disadvantages otherthan their aging seed. Essentially they store seed and in that sense they draw genes out of circulation. For the seed to be useful requires documentation and evaluation, ac- tivities that very few genebanks are performing fully. Information management within genebanks is as im- portant as the physical arrangements for safekeeping of the seed. Genebanks slow down crop plant evolution, so both hybridizing and segregating populations of a breeding process become a necessary adjunct to the well-functioning genebank. Genebanks pass on the genes; they don't breed finished varieties. Until we improve our expecta- tions and funding of genebanks and their linkages to breeding, they will continue to function short of being truly useful to the plant breeding process. I emphasize the importance of germplasm for crop plants, because it is the area where I expect the greatest advances in the next two decades, and the greatest potential we stand to lose if management of plant genetic resources does not improve. For genetic advance for any crop there are two resources: (a) the gene or genes that control specific traits, and (b) the knowledge of how to exploit that trait in the total genetic background of the finished variety for farmers' fields. This loss of knowledge disturbs me because much of the international talk 34 about germplasm loss centers around genetic erosion :.:::: .b.t .. :B::B:. : b ::...............................:.... . . :.: and genetic wipeout and rarely stresses the equally important decrease in the number of plant breeders around the world. Compounding the trend is the privatization of plant breeding and the rush to biotech- nology that is drawing funds away from germplasm management in the public sector. Knowing what is in the world collections has been the major stumbling block to their use. A great deal of new and novel genetic variation is locked up for the major crops in the approximately 2 million genebank accessions worldwide (Marshall 1990). To be blunt, this variation is useless until we plant the seed and use it for crop improvement (Goodman 1990). If the function is solely storage the germplasm will never be evaluated and its potential will not be discovered. To evaluate samples means to look at plants in the field and discover traits or potential for crop improvement, not just counting the number of seed held and their vigor of germination. I would like to see germplasm being worked. To accomplish this goal the one hundred significant genebanks around the world need to add a minimum of a thousand evalu- ation/prebreeding plant breeders. Half ofthese genebanks are in developing nations, no single factor could increase food self-sufficiency more than an increase in regional plant breeding staff in the tropics and an infrastructure to support their proven elite seed. Only a few developing nations have an adequate cadre of plant breeders and interestingly those nations are food sufficient. Clearly, plant breeders are going to have to establish the efficacy of genebanks (Brown and others 1989). To do this there have to be more plant breeders and they must be linked to major genebanks. The number of plant breeding graduate students has decreased in recent years as the universities expanded into the new directions of biotechnology. This opens an excellent opportunity in my mind for the IARCs to expand their training role. I have yet to find a replacement for the insights of a practicing plant breeder who knows the crop well. When the number of working plant breeders decreases so will the amount of insight and the pace of crop improvement 35 will suffer. A development period of approximately 10 to 15 years exists between finding new useful germplasm and a finished variety in a farmer's field. The truth is the pace of finding new useful germplasm has already fallen off but is not evident because elite materials bred a decade ago are only now reaching the varietal release stage. What is not being done now will become evident at the end of the decade when there will be even more people and a greater demand for food. The Use of Germplasm Resources If a new crop-threatening hazard (disease, insect, and so on) appears for which genetic resistance is un- available in the breeders own germplasm, the breeder is forced to look elsewhere for the needed genes. In this event an active collection, if available, will be screened. Numerous examples are on record of sought-after genes being identified in active collections especially in cereals and legumes. An important source of improved germplasm in the major crop species can be found in university and national research plots or international screening nurs- eries and trials conducted by IARCs. Advanced lines and varieties undergoing evaluation in such trials are heavily utilized by many breeders as parents in new crosses. They do so because of the accumulated performance and evaluation data available to them that help to identify germplasm with the greatest potential value. The best of this germplasm eventually is made a part of base and active collections. Performance data on unfamiliar ma- terial, when available, are more useful to breeders than descriptor data because they provide productivity and adaptation information that is particularly sought by breeders in choosing parental materials. Almost no performance data are available on the germplasm held in the world network of genebanks. Proportionally the heaviest use of active collections is made by breeders of plant species that have little or no history of breeding improvement but for which collected 36 material is available. Generally, these are in the tropical :S -:::.... . .: 'l l '': R. : l : R X . gg o pulses, some root crops, vegetables, and some industrial crops. For these crops the breeder who has initiated an improvement program has no alternative but to use exotic (unknown) collected material for local evaluation and selection. Exotic germplasm is most often utilized in the pur- suit of long-range breeding goals. Some, but not all, of the exotic germplasm comes from active germplasm collec- tions. Exotic germplasm reaches the large programs via national, regional, and international evaluation networks and frequently by personal contacts of the scientists. Breeders point out that without better passport and descriptor data, they have no rational entry into the active collections and are forced to screen accessions more or less blindly in their search for useful material (Peeters and Williams 1984). Surely efficiency can be increased by better management and evaluation of collections and to achieve this the world cadre of plant breeds needs to increase and be actively involved. General agreement exists among breeders that ge- netic resources systems both nationally and interna- tionally should include comprehensive programs of en- hancement. They believe that enhancement may be one of the keys to maximizing future collection utilization by breeders. In general, genes identified in primitive materials exhibit considerable specific local adaptation (negative linkages) and need to be introgressed or repeatedly backcrossed to improved genotypes. Useful genes from these sources generally are used only when combined in genetic backgrounds that are more acceptable to breed- ers ("free of the negative baggage"). Public sector breeders conduct enhancement or prebreeding research moving desired genes into elite material but, unfortunately, this work is a relatively small part of their total breeding effort due to financial constraints, clientele pressure for new improved cultivars, and the time required to achieve useful results. This is sadly not going to change until the importance of enhancement is recognized and the service appropriately rewarded. The world community is ap- 37 - 2 -........ . .......... z. 0 .. proximately 750 to 1,000 public sector enhancement plant breeders short. Enhancement research is pursued in the private sector also, but to a much smaller extent. The successful products of such private research usually are kept as proprietary products and are available to the public breeders only after their release as genetic com- ponents of new commercial cultivars. Germplasm for Varietal Development Most cereal breeders do not make heavy use of germplasm of landraces and wild and weedy relatives existing in active collections for reasons as previously noted (Duvick 1984; Peeters and Galwey 1988). This is particularly true for breeders of species in which breeding improvement is well-established and has been underway for a long time. Breeders tend to be reluctant to use unadapted exotic germplasm for crossing purposes- particularly if they continue to make acceptable breeding progress without its use. This is sad because the potential for significant advances using new exotic material is foregone in favor of the slow advance. Better breeding for enhancement is a way around this bottleneck to genetic advancement. It is instructive to examine some of the specific traits that give new varieties their higher yield potential. The new maize hybrids are greatly improved in root strength (resistance to root-lodging), in resistance to stalk-rot fungi, in resistance to heat and drought, in ability to withstand poor nitrogen nutrition, in resistance to pests and pathogens, and in ability to withstand the deleterious effects of crowding due to dense planting. Breeders of the other cereals have shown that similar genetic improvements can be made. All crops have improved ratios of grain to straw. This causes propor- tionately less photosynthate to go into stems and leaves. All crops have better resistance to lodging, better drought resistance, and more capacity to handle excess (or defi- 38 ciency) of soil nutrients such as nitrogen (the cereals) or iron (sorghum, soybeans). The cereal crops are all greatly improved in resistance to premature death, a syndrome of unknown cause that results in poor grain development and/or stalk rot. Additionally, continual updating of resistance to new races (biotypes) of pathogens, insects, and nematode pests has been of vital importance to many crops. Varietal turnover in some advanced regions of crops production (Yaqui Valley, Mexico-wheat 5 tons/ hectare) are as short as 3 years while other regions (Punjab, Pakistan-wheat 1.7 tons/hectare) are much slower (12 years) to experience varietal improvement (Brennan and Byerlee 1991). This is an opportunity cost that could be reversed by increasing the number of plant breeders doing local adaption breeding for national or regional programs. Training for Plant Breeding Worldwide the ratio of total population to plant breeders is increasing. The needs are such that the reverse should be the case. In actuality, because the world population has increased, training of plant breeders has remained constant (actually a small decline) in the United States (Collins and Phillips 1991) and appears to parallel the condition worldwide over the decade (1980- 89). Of this total, approximately 40 percent of the United States' Ph.D. degrees awarded were to foreign students. Most of the United States' current elite crop varieties trace back to the product of breeders trained during the period 1930-75. This also applies to a remarkable degree to the developing world. Unfortunately in the United States, many smaller university plant breeding programs have shown a nearly 10 percent decrease over the 1980-89 period (Collins and Phillips 1991) as administrative shifts toward molecular and biotechnology programs have competed for limited funds. I have made the case elsewhere (Wilkes 1989) that the "productivity ratio" of public sector plant breeding full-time equivalents (FTE) is one to a million people fed. These FTE, approximately 500 plant breeders in number, 39 ........:'.1., 4 '' i'l l'*b' -, +g , , 6 s * "-- are active in bringing on new varieties and hybrids to feed a population of a half billion. The number of FTE in the private sector seed companies is estimated to be a very comparable number and interestingly the total number of FTE plant breeders in U.S. universities in 1989 engaged in 440 breeding programs was 458.6 individuals a little lower than the numbers in 1980 (Collins and Phillips 1991). In the United States, these numbers lead me to generalize that research and development of plant genetic resources (academic programs, public sector plant breeding, and programs of private seed companies) involves a breeder to population ratio of 1 to 500,000. Using this number there are obviously some regions of the world with serious deflciencies. The work of trained plant breeders around the world is one of the major reasons that the predicted famine of the 1970s never happened. One of the major developments in international agricultural research was the emphasis in this period by Rockefeller and Ford Foundations on training plant breeders for developing countries. These plant breeders are not being replaced by a second wave as well-trained. Funds to train for plant breeding skills are hard to come by and 90 percent of developing nations can't afford university tuition at academic institutions in the North. Even an advanced country, like Mexico, with a very well-developed agricultural research matrix, can- not afford to send nationals out for 3 to 5 years of Ph.D. training. Mexico has had to go to a program of short post- doc training for their Mexican trained key personnel. This is tragic because many of the leading plant breeders in the North have former students trained in the 1950s to 1970s around the world who have kept in touch and the mutual flow of germplasm has been significant. Even when the political situation has been cold, personal contact and mutual respect has resulted in a free and open exchange of South-South and North-South unimproved and elite material. This invisible infrastructure of goodwill has been one of the most powerful forces in plant breeding in. the 1960s and 1970s. Plant breeding is a human activity and without communication and exchange of seed the 40 process comes to a dead stop. That such a small cadre of st-b..*..^. W0.Zl* .+ +gg++++++ ~~. ....................jZg+++^ plant breeders have been so effective in doubling the world's food supply is a tribute to genetics and the art of plant breeding. Plant Breeding Needs The establishment was quick to rally round the concept of genebanks when the concern was genetic erosion, however, it has been very slow to realize that effective preservation of germplasm is only half the job (Williams 199 1). Nor could breeders be left to see to this: they were too busy with their own tasks. The plant genetic resources found in genebanks are the 'insurance' for future elite varieties only if they are used in the breeding process (a policy is underwritten). The impor- tance and effective use of plant diversity in a pro- ductive and sustainable agriculture is one of the least appreciated and poorly managed aspects of genebanking. Public sector research and development planners lack policies and management plans in this area and the general public is unaware of a lack of direction. Currently the major genebanks hold approximately 2 million accessions (about half of these are thought to be duplicates) worldwide. Trevor Williams has presented evidence that in fact the number of distinct and useful accessions is around a half million (Williams 1989, 1991). Yet these 500,000 accessions are useless until we grow out and use this variation for crop improvement; nor can they be considered valuable until they are used! This is an instant on and off modern world and there are those that think that germplasm in genebanks could be used like a parts catalog. All that has to happen is to evaluate all the traits for the accessions and enter this mass of information in a computer. Then all you do is "dial a trait"-plug into the database to find the desired trait-and then plug it in the plant. This plug-in idea is rather popular with the media accounts for the promise of biotechnology also. The real truth is that genetic factors that permit local adaptation and yield, the genetic margin of difference for the farmer of one variety or hybrid 41 over another, is practically never due to a single gene. About the only magic bullet single gene traits have been those for disease resistance and most of these become obsolete before a decade has passed. Progress in plant breeding of the major grain crops can be expressed quite simply as yield per unit area. Considering all of the major advances, there is no doubt that changes in farming practices such as increased use of nitrogen fertilizer, better weed control and more timely planting and harvesting, and pest control have had important effects on crop productivity. But the single most important input has been that of breeding in the production of improved hybrids and varieties (Duvick 1986). Repeated studies for each of fourmaj or crops (maize, wheat, soybeans, and sorghum) have shown that about 50 percent of yield gains over the past 50 years in the United States were due to varietal improvement. Pre- sumably the same would hold for other parts of the world. The rate of yield gain attributable to breeding has been estimated at about 1 percent per year. Gains have usually been linear and up to now show no sign of decreasing but there is some fearthat a flattening (logistic curve) might be ahead. The plant breeding community is the primary pro- vider of new varieties, yet ifbreeders don't supplyvarieties that find favor with farmers, they have no product. Therefore, to satisfy farmers, two concerns are always addressed by the breeder in any released variety: yield and predictability. Selection for superior genotype is primary in creating high-yielding varieties. Yield has many facets-the obvious tons/hectare, or farm gate value, cost of inputs, but also very important qualities such as nutritional, taste, texture, and so on and the long-term sustainability with current farming practices. Predictability means that the heritability must be high, and to achieve this standard there is a considerable pressure for genetic uniformity. Genetic uniformity can 42 be either in the form of homozygosity (often recessive) or the hybrid of inbred parents (hybrid maize). The kinds of uniformity desired in maize, for example, are: (1) rapid and uniform germination of seeds; (2) nearly simulta- neous flowering; (3) nearly simultaneous maturation of harvestable product; (4) stature that promotes mechani- cal harvest (ear height): (5) product uniformity for taste, flavor, and chemical composition; and of course (6) year- to-year stability of yield. This last point demands special mention because in the last 25 years plant breeders have developed selection techniques for buffered systems that yield well over a range of conditions across a wide range of habitats. The gains or increased yield capacity of the newest varieties and hybrids are exhibited, not only when inputs are optimal (high levels of soil fertility, plentiful water, and in the absence of disease and insect pests) but also when fertility, pest, and environmental conditions are unfavorable (Duvick 1986). The increased "toughness" of new hybrids and varieties is a fairly well-kept secret, at least outside plant breeding circles. It is generally claimed by nonplant breeders interested in criticizing the current state of food production that new hybrids and varieties are "weak" and disease-prone, and that they must be given plentiful supplies of water, fertilizer, and pesticides in order to do well. The facts are that modem varieties must be much tougherthantheirpredecessors in orderto survive modem cultural conditions. Inmaize for example, today's crowded growing conditions (due to high seeding rates to maximize yield) increase opportunities for individual plants to suffer water and nutrient shortages, and reduced avail- able sunlight. For all crops, monocropping facilitates the epidemic increases of disease, insect, and nematode pests and the pathogens and pests are generally more virulent over time. The older varieties cannot stand present growing conditions and exhibit female sterility, insufficient root growth, develop weak, lodging-prone stems, and become more prone to disease and insect attack (Duvick 1986). 43 . ..... .......... ...g Z .. - The adoption of these highly uniform, yet superior, hybrids and varieties means that plant breeders must bring on, in continuing fashion, new varieties with new resistances or tolerances. Varieties typically are replaced by new ones every 5 to 10 years (Brennan and Byerlee 1991). Those societies that grow modern plant varieties have, therefore, an indispensable need for modern stra- tegically planned plant breeding programs. Once steps are taken toward planting modern crop varieties. there is no alternative but to support energetic, well-- planned, and well-executed plant breeding programs. Differences between Primitive Populations and Modern Varieties Two maj or differences exist between the processes of early domestication of our crop plants and current breeding practices. First the early development of the crop in the area where it originated or diversified usually occurred in the presence of coevolving pests and pathogens. Second the genepools during the early evolution of the crop exhibited considerable variation because of hybridization with wild related species and also because of genetic recombination between primitive cultivars. Today few breeding programs use primitive cultivars in the breeding process ane most rely rather on a limited number oiF proven elite lines long ago derived from the primitive forms (Wilkes 1989). In peasant farming cultures excessive genetic uni- formity was virtually never present, no matter whether the crop was annual or perennial, self-pollinated or cross-pollinated. Although self-pollinated annuals in general consisted of small populations of homozygous individuals they usually contained a large range of different homozygous genotypes and thus, as a population, were highly heterogeneous. Naturally cross-pollinating crop species invariably were maintained in an outcrossedl condition and were both heterogeneous and heterozy- prgouss and perennial plants, whether increased via seed or clonal multiplication, also were maintained as genetically heterogeneous populations of heterozygous individuals. Genetic heterogeneity did not provide absolute in- surance against epidemic pest attacks nor against envi- ronmentally caused yield losses. Species-specific genetic uniformity is sometimes sufficient in itself to support epidemics of disease or insect pests. Thus Dutch elm disease devastated the entire species of American elms in the United States in the 1960s. Nevertheless, it is certain that the large amount of heterogeneity in primitive forms in peasant agriculture did and still does greatly reduce the danger of loss to epidemics within the same crops. This advantage ap- pears to be no longer present for those countries that grow highly uniform modern crop varieties, at least not in the same way as is seen in peasant varieties. Varietal Mixtures It is not simply the number of crop species and of individual varieties that made landrace crop fields more diverse. Spatially and temporally crops were intermixed in ways that reduced their vulnerability to pests and diseases and increased their efficiency in using water, nitrogen, and light. Some of these intermixes were intercropping, overcropping, sequential (relay) planting, and adjacent patches or mosaics. Elements of this strategy are found in modem day multiline mixtures and relay planting with distinctly different genotypes (for example the early and late forms of the crop). Similarly mixes of resistant and vulnerable cultivars have dem- onstrated yield stabilities that are many times that of large plantings of single varieties. The mixtures appear to keep insect pest populations from overrunning the susceptible varieties because of their patchiness. Such human designed agricultural systems that mimic more natural ecosystems appear to be more stable. Much of this research has been pioneered by Professor M.S. Wolfe 45 % H . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~..... . . .. % . . . . . . .. .. . .z . . .... at the Cambridge Plant Breeding Institute, United King- dom in the 1970s and 1980s (Wolfe 1985, 1988). There is a continuum from a multiline in which the individual lines differ only by identified resistance genes, to line mixtureswhose individualmembers are segregates from common parents (not isogenic), to optimal variety mixtures that appear to be the best option to supply heterogeneity for disease resistance (Browning 1988; Simmonds 1988; Wolfe 1988). A growing body of literature supports the concept that a field made up of heterogeneous elements yields at least as well as their separate components, be they pyramided disease resistance genes in different multilines (the clean crop multilines of Borlaug 1958) or multilines that stabilize the inroads of the pathogen at low, nondamaging levels without stacked disease resistance genes (dirty crop multilines of Browning and Frey 1969); varietal mixtures (Wolfe 1985); or species mixes such as wheat and oats, wheat and barley, maize, beans, and squash, or barley and peas. In fact, on average the mearn yield of the mixture exceeds the mean yield of the com- ponents when extended over a number of years (Wolfe 1985). Modem agricultural practices traditionally have used large inputs of fertilizers, herbicides, fungicides, and insecticides to optimize the field conditions in an attempt to stabilize yields. General indications are that additional applications of these inputs have levelled off relative to yield (Duvick 1986). They are no longer as cost- effective as they once were, and even more importantly massive applications are becoming socially unacceptable. All of these factors point to the attractiveness of varietal mixtures that can compensate for biotic stress through their dampening effect on the spread of diseases. In addition, varietal mixtures will require no new demands on plant breeding programs because currently available varieties are, in most cases, adequate for mixture devel- opment. Mixtures present an evolutionary dilemma for 46 pathogens but no absolute assurance of durability. As in ~~~~~~~.. .. .. . . ..... S.S ...... . . ........ . . . % ... . monocultures, diversification of varieties in mixtures (diversification of varieties over time in monocultures) and differences in resistance genes among varieties are the most reasonable answers to forestall the genetic vulnerability problem. In industrialized agriculture, monoculture has rapidly become an accepted tradition encouraged by the industrial users of agricultural prod- ucts. The value of heterogeneous crops is often more readily accepted by the farming communitythan by those who serve it. The reason is simple: the advantages of the heterogeneous crops system benefit the farmer, but the changes needed, at least in industrialized agriculture, have to be introduced by the agricultural service industry and society in general. Plant Breeding in Developing and Developed Countries In the developing world, high-yielding varieties (HYV) for the maj or crops have come into dominancejust within the last 15 to 25 years. The immediate effect of these introductions is to make available novel genes for resis- tance to widespread diseases for which landraces may have been susceptible. Simultaneously rapid losses of indigenous diversity are occurring. As a few hybrids from the same breeding programs come to dominate a devel- oping country's production, increased vulnerability is almost certain to emerge. It is very doubtful whether all breeding programs in developing countries can react as rapidly to future epidemics as the U.S. maize breeders did between 1970 and 1972. A classic example of a genetically vulnerable variety without replacement is the wheat 'Sonalika' in South Asia. Farmers will stay with their chosen varieties with remarkable persistence as long as the variety performs well in their hands. This is exactly what has happened with 'Sonalika' where it has been removed from the recommended variety list because of known serious disease susceptibility (Dalrymple 1986), but continues to be planted over vast acreages because it is so early, pos- sesses amber grain, and fits the multicropping schedule. 'Sonalika,' which was released in 1967, has for more than a decade been known to be susceptible to a new race of leaf rust (and therefore needs to be replaced) but has not had the plant breeding attention to generate a suitable replacement. Farmers cannot abandon this susceptible variety (Sonalika accounted for 40 percent of wheat acreage in 1988) because there isn't a satisfactory re- placement. Plant breeding at the local and regional levels continues to fail these farmers. For the past 25 years, leadership in wheat and rice breeding for developing countries has come from CIMMYT and IRRI. There is a strong possibility that with limited funds these institutes will devote a smaller proportion of their efforts to variety development in the future. The national breeding programs in developing countries are, in general, not as well-funded nor as effective as the breeding programs in the IARCs. There are significant policy questions, therefore, as to whether adequate boreeding efforts worldwide will indeed be devoted to boringing on new wheat and rice varieties at frequent intervals for the developing world. This question is of particular significance if the IARCs move away from maintenance breeding in favor of more 'upstream" activi- ties such as biotechnology or politically favored activities such as elite HYVs for marginal environments. Those countries with a high proportion of area sown to the HYVs are on a treadmill. They cannot turn back to cultivation of the indigenous peasant varieties without serious delay due to the time that would be needed for re- education of the farming population and restocking of seed of indigenous varieties. They cannot continue to plant the presently used HYVs indefinitely because new pest races inevitably will appear, causing disastrous epidemics in the aging varieties. This latter problem is particularly acute in the tropical and subtropical areas with no cold season (and often little dry season) to 48 annually set back the pest populations. ... : ., , ... . . . -- . .. ,,--- ;So.................... . . . . gR ...... X . At least 5 to 10 years of targeted efforts are needed to put strong breeding programs into place or to improve ineffective breeding programs. Those countries with undersized or ineffective programs that are now depending upon the IARCs for new HYVs have no time to waste. They must begin as soon as possible to strengthen their national breeding programs up to the size and efficiency that will serve their countries properly. Clearly there is a significant lack of skilled breeders at the national level to achieve these goals. Part of the strategy is for developing countries to work cooperatively with surrounding countries with similar climatic conditions and cultural practices. The pooling of several small programs can, in theory, reach the same goals by providing a sufficient base for bringing out improved varieties at frequent intervals. Unfortunately, histories of planned, government-led intercountry coop- eration have few success stories. Remarkable skills in management and diplomacy are needed to make such cooperation productive. Worldwide cooperation and collaboration in plant breeding are now needed more than ever before. The history of development of successful new varieties shows, without exception, that the varieties could not have been developed if breeders around the world had not freely exchanged germplasm and information. With higher proportions of the crop area being sown to highly bred varieties (and conversely, with smaller and smaller areas being planted to indigenous farmer varieties), the need for an active worldwide plant breeding collaboration is stronger than ever. The developing countries are now a significant factor in genetic vulnerability. Both hybrids and HYVs have come into dominance within the last two decades. The immediate effect of these developments has been in- creased production and rapid losses of local landraces that the elite germplasm has displaced in farmers' fields, and essentially the same genetic backgrounds can now be found in the agricultures of both developed and 49 ',c,''':<':':': R':':' ' . gg: .......................... .. .. g, developing countries. This translates into a wider net to pick up pests and pathogens in the context of varietal invasion, new encounter vulnerability, and pathogen invasion. Because a few elite materials have come to dominate the basic cereals and legume crops world- wide, increasing vulnerability is almost certain to emerge. The burden of genetic vulnerability up until now has been placed primarily on the shoulders of plant breeders because elements exist in the technology of plant breeding that can be designed to minimize the impact of genetic vulnerability. Farmers have always been alert to varia- tions in a crop and have taken advantage of unplanned, fortuitous genetic variants that appeared long before these evolutionary processes had the names of mutation, introgressive hybridization, and polyploidy. Artificial selection is a powerful tool of plant breeders both in controlled crossing or peasant farmer opportunism with landraces. Almost always selection can be employed to overcome the expression of genetic susceptibility once it appears and is recognized. The trouble with genetic vulnerability in modern elite varieties is that uniform monocultures are so widespread and the impact so magnified that it doesn't matter if the plant breeders can correct the problem. That it has happened at all is sufficient to disrupt the world food supply. Collectively, plant breeders have a very important influence on the amount of genetic uniformity to be found in commercial crop varieties, however, de- mands of the industry and limitations of research re- sources ultimately define how diverse these varieties really are. Genetic uniformity in and of itself is not necessarily undesirable (if deployed wisely) and is nec- essary for much of our present agricultural markets (Plucknett and others 1987), however, to disperse a uniform variety very extensively is to spread a wide net for pests and pathogens. The price of the emphasis on maximum productivity 50 is vigilance, based on a thorough knowledge of current ........ .. .. ....... .+. +. . {its<+ ;+ varieties, and the ability to trace any new infection with great care. Ultimately, varieties under development should originate from parentage wider than the varieties they displace. For most breeders, however, there are few incentives to go any further for breeding parents than the very small number of elite materials at the top of the pyramid. Individual breeders watch with vigilance for outbreaks of pests and pathogens, but there is no overall management strategy to quickly bring genes from the lower levels of "gene in reserve" for most of the major crops, and there is no strategy at all for keeping the reserve ranks filled with planned diversity of genotypes for resistance to pest, weather, or soil problems. This strategy must surely be at the very core of genebank policies and management. Nowhere is the impact more serious than the ex- tension of the high-yielding grain belt germplasm into microhabitats and marginal lands of the tropics (CIMMYT 1987). This has cast a larger net to capture pests and pathogens. These habitats are often without the pro- nounced dieback seasons of the major temperate grain belts when severe cold (U.S. corn belt), severe drought (Punjab, wheat), or flooding (South East Asia, rice) push back the pest and pathogen populations. Presently with wheat being harvested over most of the year somewhere in the world, the potential to capture pests and pathogens has increased. Instead of promoting commercial wheat seed in Iran and Turkey forming a continuous carpet from the Punjab into Europe, or commercial seed for Nepal and Thailand forming a continuous carpet from the Punjab to China, there ought to be a mosaic of diversity so there will not be a connecting bridge of genetic uniformity to promote the march of a pest or pathogen. The idea of a genetic "firebreak" using bands of diverse germplasm to slow down the march of a potential vulnerability needs to be in place. Clearly, an integrated, genetic-based, crop man- agement strategy is needed to mitigate unwanted increases in genetic vulnerability of widely grown elite plant ma- terials. 51 One of the best strategies for all vulnerabilities is periodic reassessment. When susceptibility is ig. nored it creates vulnerability, but susceptibility re- assessed generates options. Future Needs and Priorities If plant breeders breed from only adapted elite material as many commercial companies tend to do, and aim only for the short run, narrow germplasm bases result, thus increasing the potential for genetic vulner- ability. A certain amount of long-range breeding aimed at increasing genetic diversity must always be present. However, it is difficult to determine the minimum number of plant breeders necessary to insure this goal of providing sufficient genetic diversity in reserve. If the total number of individuals committed to breeding decreases, the amount of germplasm actively used by breeders also may decrease unless some way is devised of increasing output per breeder, particularly with regard to enhancement or breeding for broadening the useful germplasm base. MIuch potentially useful gernplasm is lost due to lack of personnel. An inordinate amount of public responsibility must be assumed by a few individuals who prepare genebank or landrace materials for further breeding wvork. The degree to which these few can prebreed and efficiently create, evaluate, release, and store valuable germplasm is limited. Furthermore as breeders retire and administrators change their emphasis, valuable breeding pools and germplasm collections often are dis- carded, a tragic loss for the public at large. Changing patterns of susceptibility of maj or crops to diiseases will always result in surprises. The measure of success is how appropriate our preparedness and timely our responses are to the vulnerability. Investing by the IARCs in improved minor crops for the zones between maj or grain belts should be a priority to protect their yield advancement with maize, wheat, and rice. Improvement in the evaluation of germplasm in genebanks and the 52 development of alternate maternal parents (cytoplasm) ,....':..B.:,,''. ....B .....................'B: i ''' B'':'BB'B'B'B, ' ' '' should broaden the backup varieties to current elite materials. To do this we need plant breeders not only at the IARCs, but also at the regional level dealing with local adaptation. Improvement of crop varieties depends on new and diverse gerrnplasm resources, and also the human skill to see promise in segregating populations early on in the breeding process. There exists an obvious need for more plant breeders worldwide to hybridize crops and generate genetic diversity than exist presently. The return to a more diverse genetic mosaic will be an important strategy to reduce genetic vulnerability within and between crops. Agriculture, Germplasm Conservation, and Strategies for the Future The challenge as we approach the next century will be to design agricultural management strategies and institutional arrangements that can successfully ame- liorate the negative environmental effects of agricultural and urban intensification primarily in the grain belts. The uniformity of major crops and the fact that they have largely displaced indigenous landraces is a given of the present agricultural consumer context. Subsistence farmers are a decreasing minority. Most farmers sell some or all of their harvest for cash and are, therefore, yield over cost-of-inputs driven. They need to use the best varieties available to them. Genetic vulnerability is inherent in the use of uniform elite germplasm. New varieties are continually brought up from breeders' plots to replace older ones. However, the lack of a broad-scale management strategy to anticipate expected problems and minimize their impact exhibits tunnel vision about how biological systems organize themselves and an in- sensitivity to the plant breeding establishment's public responsibility. Positive actions to deal with genetic erosion, genetic wipeout, and genetic vulnerability at both national and international levels must involve the following initiatives: 53 :.::.::: :.:..:..: :' :5:...: .: :':'':........................... f.H..v.':.:B:B:BY;.:Bf::.:::B'. * Create new programs to identify genetic erosion and vulnerability-variety surveys and collabo- ration in international efforts to monitor the use and geographic distribution of elite germplasm. * Develop appropriate genepools (especially in the public sector) introgressed by exotic derived germplasm to support the commercial breeder with diverse useful materials. * Increase training in the maintenance of plant ge- netic resources both nationally and internation- ally. Active genebanks should be linked to prebreeding and in moving genes from the col- lections into enhanced materials but not finished varieties. The future will be best protected from genetic vulnerability if worldwide cooperative networks of crop specific prebreeding develop. * Conduct basic and applied research in order to more efficiently and effectively measure genetic distance between varieties and improve evaluation techniques, especially relating to the nature of resistance to pests and pathogens and pest and plant interaction; in conjunction with early de- tection of changes in the virulence of pests and pathogens as well as shifts in the varietal picture. * Educate and inform the food industry, farmers, seed suppliers, plant breeders, and others about the relationship of genetic and spatial diversity with regard to the potential for crop vulnerability so they can develop effective management strat- egies such as parallel breeding, genes in reserve, enhancement breeding for gene pools, wide crosses and biotechnology; and alternative management strategies to monocultures such as crop rotation, tillage practices, crop mixtures, multilines for resistance, pyramided resistance factors, ma- nipulations of pest parasites, pest trap crops, insecticides and fungicides, and better monitor- 54 ing. Summary and Charge for Change The quest for an assured food supply has done more to decrease biodiversity and alter the landscape and environment than any other activity in which we engage. Because food production is population driven, the de- mands for increased productivity mean that the ability of natural systems to conserve genetic resources has been diminished (genetic erosion) to the point that we are now dependent on genebanks to preserve the genetic estate. There is a large amount of germplasm now held in the genebanks of national programs and the IARCs. The wide dispersion of these holdings is desirable but the lack of the ability to fully use the collections has limited the effectiveness of many national programs, both within the country and in exchanges between countries and regions. Better germplasm is the fastest and least expensive input for increasing crop plant productivity. Many national genebanks need to become more active in the search for new useful genetic materials to be used in the breeding process. To compound the problem only a very few crop varieties now occupy the major acreage worldwide and these highly uniform monocultures exhibit considerable potential for genetic vulnerability. More than ever before, international efforts are re- quired to help slow genetic erosion, establish and encour- age internationally responsible and proactive genebanks, and help prevent epidemics in the developing countries where the greatest threats of genetic vulnerability and germplasm erosion now exist. Because oftheirmajorrole in insuring global food stability, the donor community and the IARCs can no longer continue to neglect the issues posed by the decreasing number of plant breeders relative to human population and genetic vulnerability. The current haphazard, uncoordinated, and unsystematic approaches to the problem reflect dangerous and inap- propriate national and global priorities. The new strategies should be more adaptive and employ less preventive maintenance breeding to sustain 55 ........... .. . ..~..... :. ..... . :::: :: : : > : : : : the yields required by an expanding human population. We can no longer afford to ignore the continued loss of biodiversity in agriculture. If society took seriously genetic erosion and genetic vulnerability, we would have better management practices in place than our presently dangerously reckless pattern of casual neglect. Acknowledgements Special thanks to Shauna Lee who has been loyal in helping me make deadlines while keeping a busy de- partment functioning. Also I'd like to thank Allison Aheam, one of my students, for the illustrations, and my thanks to the reviewers who have made this a better manuscript. 56 References Alcorn, Janis B. 1991. .EthIcs, Economics, and Conser- vation." In M. Oldfieldand J.Alcorn, eds., Bfodiwsity: Culture, Conservation and Ecodevelopment. Boulder, Colo.: Westview Press. Altieri, M.A., M.K. Anderson, and L.C. Merrick. 1987. "Peasant Agriculture and the Conservation of Crop andWild Plant Resources." ConservationBiology 1:49- 58. Anderson, E. 1952. Plants, Man and Cie. Boston: Little, Brown. Borlaug, N.E. 1958. iThe Use of Multiline or Composite Varieties to Control Airborne Epidemic Disease of Self-Pollinated Crop Plants." Proceedings First Inter- national Wheat Genetic Symposium. Winnipeg: Uni- versity of Manitoba. Brennan, J.P., and D. Byerlee. 1991. tThe Rate of Crop Varietal Replacement on Farms: Measures and Em- pirical Results for Wheat." Plant Varieties and Seeds 4:99- 106. Brown, A.H.D., O.H. Frankel, D.R Marshall, and J.T. Williams, eds. 1989. The Use of Plant Genetic Re- sources. Cambridge: Cambridge University Press. Browning, J.A. 1988. GCurrent Thinking on the Use of Diversity to Buffer Small Grains against Highly Epi- demic and Variable Foliar Pathogens: Problems and Future Prospects." In N.W. Simmonds, and S. Rajaram, eds., Breeding Strategies for Resistance to the Rust of Wheat. Mexico D.F.: CIMMYT. Browning, J.A., and K.J. Frey. 1969. 'Multiline Cultivars as a Means of Disease Control." Annual Review Phytopathology 7:355-82. 57 ~~~~~~~~~. . . . ... ... .. ... . W..+* CGIAR (Consultative Group on International Agricultural Research). 1990. "Partners in Conservation: Plant Genetic Resources and the CGIAR System." Wash- ington D.C _. 1989. "CGIAR Policy on Plant Genetic Resources." Washington, D.C. Chang, T.T. 1985. "Crop History and Genetic Conserva- tion: Rice-A Case Study." Iowa State Joumal of Research 59: 425-460. CIMMYT (Centro Internacional de Mejoramiento de MaIz y Trigo). 1988. Recent Advances in the Conservation and Utilization of Genetic Resources: Proceedings of the Global Maize Germplasm Workshop. Mexico D.F. __.1987. TheFutureDevelopmentofMaizeandWheat in the Third World. Mexico D.F. Cohen, J., J.T. Williams, D.L. Plucknett, and H. Shands. 1991. 'Ex Situ Conservation of Plant Genetic Re- sources: Global Development and Environmental Concerns." Science 253: 866-872. Collins, W., and R. Phillips. 1991. "Plant BreedingTraining in Public Institutions in the United States: A Survey Conducted by the National Plant Genetic Resources Board." Diversity 7:28-32. Dalrymple, D.G. 1986. Development and Spread of High Yielding Wheat Varieties in Developing Countries. Washington, D.C.: U.S. Agency for International De- velopment. Duvick, D.N. 1986. 'Plant Breeding: Past Achievements and Expectations for the Future." Economic Botany 40:289-297. _. 1984. "Genetic Diversity in Major Farm Crops on the Farm and in Reserve." EconomicBotany 38:161- 58 178. o. 4 4. + h . z F + + N {.+ ................................. 0 0. - *. .., . : . : :. ..... FAO (Food and Agriculture Organization of the United Nations). 1981. Agriculture: Towards 2000. Rome. Frankel, Sir Otto. 1989. "Point of View: Perspectives on Genetic Resources." In Delivering Diversity, CIMMYT 1988 Annual Report. Mexico D.F. . 1974. "Genetic Conservation, Our Evolutionary Responsibility." Genetics 78:53-65. Goodman, M. 1990. "Genetic and Germplasm Stocks Worth Conserving." Journal of Heredity 81: 11-16. Harlan, J.R. 1975. Crops and Man. Madison, Wisc: American Society of Agronomy. Hawkes, J.G. 1983. The Diversity of Crop Plants. Cam- bridge, Mass.: Harvard University Press. Heiser, C.B. 1990. Seed to Civilization: The Story of Food. Cambridge, Mass: Harvard University Press. Hoyt, E. 1988. Conserving the Wild Relatives of Crop Plants. Gland, Switzerland: IBPGR/IUCN/WWF. IUCN (International Union for the Conservation of Nature and Natural Resources). 1980. World Conservation Strategy. Gland, Switzerland. Juma, C. 1989. The Gene Hunters: Biotechnology and the Scramblefor Seeds. Princeton: Princeton University Press. Kato-Y, T.A. 1976. "Cytological Studies of Maize (Zea mays L.) and Teosinte (Zea mexicana [Schradl Kuntze) in Relation to the Origin and Evolution." Massachusetts Agricultural Experiment Station Bulletin 635. Keystone. 1991. Oslo Report: Third Plenary Session Key- stone Intemational Dialogue Series on Plant Genetic Resources. Washington, D.C.: Genetic Resources Communication System. 59 % ,~~..: :. .... :. . .j...H : Kloppenburg, J.R., ed. 1988. Seeds and Sovereignty: The Use and Control of Plant Genetic Resources. Durham, N.C.: Duke University Press. Marshall, D.R. 1990. "Crop Genetic Resources: Currenrt and Emerging Issues." In A.H.D. Brown, M.Clegg, A. Kahler, and B. Weir, eds., Plant Population Genetics, Breeding and Genetic Resources. Sunderland, Mass: Sinauer Associates. NAS (National Academy of Science). 1972. Genetic Vul- nerability of Major Crops. Washington, D.C. Oldfield, M.L., and J.B. Alcorn, eds. 1991. Biodiuersity: Culture, Conservation and Ecodevelopment. Boulder, Colo.: Westview Press. Peeters, J.P., and N.W. Galwey. 1988. "Gerrnplasm Col- lections and Breeding Needs in Europe." Economic Botany 42:503-521. Peeters, J.P., and J.T. Williams. 1984. "Toward Better Use of Genebanks with Special Reference to Infor- mation." Plant Genetic Resources Newsletter60:22-32. Plucknett, D.L., N.J.H. Smith, J.T. Williams, and N.M. Anishetty. 1987. Genebanks and the World's Foocl Princeton: Princeton University Press. Prescott-Allen, R., and C. Prescott-Allen. 1990. 'How Many Plants Feed the World?" Conservation Biology 4:365-374. Simmonds, N.W. 1988. "Synthesis: The Strategy of Rust Resistance Breeding." In N.W. Simmonds, and S. Rajaram, eds., Breeding Strategies for Resistance to the Rust of Wheat. Mexico D.F.: CIMMYT. ___,_ ed. 1976. Evolution of Crop Plants. London: Longman. 60 United Nations. 1989. Department of International Eco- nomic and Social Affairs: World Population Prospects 1988. ST/ESA/SERA 106. New York. Vavilov, N.I. 1935. "The Phytogeographic Basis of Plant Breeding." In The Origin, Variation, Immunity and Breeding of Cultivated Plants: Selected Writings of N.l Vavilov. (Translated from the Russian by K. Stan Chester). Chronica Botanica 13: 1949/50. Waltham, Mass. Wilkes, G. 1991. "In Situ Conservation of Agricultural Systems." In M. Oldfield, and J. Alcorn, eds., Biodiversity: Culture, Conservation and Ecodevelop- ment. Boulder, Colo.: Westview Press. . 1990. "Germplasm Conservation andAgriculture." Paper presented at the Pennsylvania University BioDiversity and Landscapes Conference. October 25, 1990. . 1989. "Germplasm Preservation: Objectives and Needs." In L. Knutson, and A.K. Stoner, eds., Biotic Diversity and Germplasm Preservation: Global Im- peratives. Dordrecht: Kluwer Academic Publishers. . 1988. "Plant Genetic Resources overTenThousand Years: From a Handful of Seed to the Crop-Specific Mega-Genebank." In J. Kloppenburg, ed., Seeds and Sovereignty: The Use and Control of Plant Genetic Resources. Durham, N.C.: Duke University Press. _. 1977. "The World's Crop Plant Germplasm: An Endangered Resource." BulletinAtomic Scientist33:8- 16. _= 1971. 'Too Little Gene Exchange." Science 171:955. Williams, J.T. 1991. "Plant Genetic Resources: Some New Directions." Advances in Agronomy 45:61-91. 61 *~~~~~~~~~~~~~~~~~~~~~~~~~~....... .: . . .....n g .1989. 'Practical Considerations Relevant to Effec- tive Evaluation." InA.H.D. Brown, O.H. Frankel, D.R. Marshall, and J.T. Williams, eds., The Use of Plant GeneticResources. Cambridge: Cambridge University Press. Witt, S.C. 1985. Brief Book: Biotechnology and Genetic Diversity. San Francisco: Center for Science Inforrna- tion. Wolfe, M.S. 1988. 'The Use of Variety Mixtures to Control Diseases and Stabilize Yield." In N.W. Simmonds, and S. Rajaram, eds., Breeding Strategies for Resis- tance to the Rust of Wheat. Mexico D.F.: CIMMYT. . 1985. "The Current Status and Prospects of Multiline Cultivars and Varieties Mixtures for Disease Resistance." Annual Review Phytopathology 23:251 - 73. 62 ® Printed on Recycled Paper