Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
4 Enhanced Weediness: A Major Environmental Issue GENE1lAI PRINCIPLES Perhaps the single most commonly voiced concern about the introduction of genetically modified plants ~ that it might have the potential to inadvertently produce a new weed or increase the aggressiveness of existing weeds (R. K. Colwell et al., 1985; Sieve et al., 1989~. This chapter discusses three aspects of the concern: whether the experience with the introduction of exotic plants into new environ- ments (sometunes with the result that a weed problem is created) is a valid analogy for the introduction of genetically modified plants; the potential for domesticated crops to revert to a wild or weedy state; and the potential for hybridization between domesticated crops and wild relatives that might create or enhance weediness. Evaluation of these issues first requires a careful definition of terms. The term "weed" has been variously defined, depending on the different perspectives of ecologists, agronomists, and the public. in this report we define a weed as an unwanted or undesirable plant in some human environments, that is, a plant that persists In hu- man environments but is neither a crop (used for food, fiber, fuel, pharmaceuticals, or turf) nor an ornamental plant. 37
38 A characteristic of human environments and consequently a strong agent of selection among weeds is frequent disturbance, as occurs in arable fields, roadsides, foot paths, and the margins of reservoirs. Consequently, many plants that have become persistent weeds are species that arose earlier because their phenotypes per- mitted them to colonize special natural environments that exist in frequently disturbed sites. Such plants often display rapid growth, a short life cycle, high seed production, and long-distance dispersal of seeds (Baker, 1974~. Not all colonizers are weeds, however, nor are ad weeds colonizers. Some weed species have also apparently required additional char- acters ~ order to thrive in close association with humans. These ad- vantages include escape from biotic control agents such as predators, pathogens, and competitors (Harper, 1965~. Such an escape is e~ec- tive if a plant is suddenly transported far beyond its native range and therefore the range of one or more of its enern~es. It is not surprising, then, that in most parts of the world, including the United States, the bulk of the weed flora are exotic plants (Holm et al., 1977; Smith, 1985; Mack, 1986), members of a species that enters a range in which that species has not occurred before (Mack, 19853. Perhaps most successful (most widespread, persistent, and abundant) are those weeds that have not only immigrated, but also have a long history of close association with human settlement (Baker, 1974~. Whether a plant becomes a weed depends on the relationship of the plant to its environment, especially with respect to control mechanisms that hold the organism in balance with that environ- ment. A plant can become a weed if it escapes control by migrating to a new environment that lacks the factors that controlled the plant in its original habitat. In addition, a plant reman In its original habitat may effectively escape a particular control factor, such as predation by a specific insect pest, by gaining a trait that imparts to it the ability to overcome the control factor. Any added trait that enhances performance (such as frost resistance or drought tolerance) would also be analogous to providing the plant with an advantage sometimes gained by plants in a new environmental range. Although this description is theoretically valid, it is necessary to keep in mind that there is extensive experience in these kinds of modifications in classical breeding. So far, weediness has not resulted from the addi- tion of the traits of pest or herbicide resistance, nor frost or drought tolerance.
39 THE 1lElATIONSHIP BETWEEN THE INTRODUCTION OF E:XO TIC PLANTS AND GENETICALLY MODIFIED PLANTS The term exotic species, as used here, refers not only to en- t~rely novel species in a new habitat, such as the Asian weed ku~zu (Pueraria lobata) in the United States, but also to any species with an expanded geographic range, even when closely related plants are already present. In addition, exotic species usually refer to plants whose ranges were extended as a result of human intervention. Ecological :[rnplications of ~troducmg Plants with Many New Waits Exotic species may not be strictly analogous to genetically modi- fied organisms because many exotic species differ by many traits Tom any of their neighbors in the new environment. Consequently, the immigrants (such as Agropyron repens, Eicchornia crassipes, Schinus terebinthifolius) will owe their success In spread and eventual nat- ural~zation to a suite of characters (Hohn et al., 1977; Barrett and Richardson, 1986; Morton, 19783. Genetically modified plants that are likely to be introduced In the near future (say, over the next 10 years) will diner by only one or a few traits from cultivated forms already in the same environment (the introduction of glyphosate- tolerant tomatoes). Ku~zu Is a familiar example of a deliberately mtrocluced exotic organism that has proven to have undesirable features. It illustrates the public's worst perceptions of errant organisms and s~multan~ ously exemplifies an exotic organism that ~ not analogous to any hypothetical genetically modified organism. Originally introduced into the United States from China and haps in the late nineteenth century for ornamental purposes, ku~zu was eventually touted as an excellent stabilizer of soil embankments and as a forage crop on unproductive land. Cash incentives were even provided at one time to encourage farmers to plant it on abandoned fields (Miller, 1983~. By the 1950s, however, detrimental aspects of ku~zu were recog- nized, as the vine often grows far beyond the site of its local mtro- duction. It now commonly grows over forest canopies and telephone lines ~ the southeastern United States. Ku~zu's success is based on a combination of features: it readily propagates vegetatively, it can
40 grow on infertile soil and low amounts of soil water (Forseth and Ter- amura, 1987), and it has few (if any) serious parasites or herbivores in its new environmental range. Ku~zu exemplifies how the combined action of many traits mtro- duced into a new environment results in a weediness problem. Our knowledge of invasions and particularly the characteristics that spell success or failure for immigrants is limited (Harper, 1982; S~mberIoff, 1985), despite the attempts to identify the putative characteristics of successful weeds (Baker, ?986; Bazzaz, 1986~. Ecologically l~portant Changes that Result from Smog Genetic Alteratiom Even though exotic species such as ku~zu are not strongly analo- gous to genetically modified plants, circumstantial evidence suggests that a change in only a few characters can sometimes make a plant a successful invader. Within the large grass genus Bromus are several annual species that have become successfully naturalized In different temperate regions. Bromus tectorum spread rapidly In the interior Pacific Northwest in the early part of this century, whereas other members of the genus such as B. motlis and B. iTizGeformis are much less common even though they were introduced earlier (Mack, 1981~. In contrast, B. mollis is much more prominent than B. tec- torum in the Central Valley of California, and B. secalinus can be a serious weed of cereals in northern Europe (Salisbury, 1961~. The differences among these closely related species that explain their var- ious success in new environmental ranges may be related to different tolerances to frost (B. rigidus and B. rubens are less tolerant than B. tectorum) and different flowering times (B. japonicus flowers before the onset of drought) (Hulbert, 1955~. These species are morpholog- ically similar and also share many ecologically important traits, yet they differ in their degree of success in their new ranges. The exotic woody genus Casuarina provides another example in Florida. The two species, C. equisetifolia and C. glauca were de- liberately introduced into southern Florida. The first has become a serious pest, while the second persists only locally. The most appar- ent difference between these closely related species Is the inability of C. glauca to produce seed in the new range (Morton, 1980), which thereby limits its dispersal. Other examples of environmentally important single-trait changes are demonstrated by the spread of Chondritia juncea ire
41 Australia and the role of insect herbivory in influencing the compet- itive ability of barley. Chondrilia juncea (skeletonweed) is a serious weed In the wheat-grow~g regions of southeast Australia. It has three morphological forms In Australia termed A, B. and C that differ most obviously in leaf shape, flower morphology, and fruit char- acter~stics. Before a biological control program wan initiated ~ the early 1970s, form A was much more widespread than the other two. But form A has proven to be much more susceptible than forms B and C to the deliberately released rust fungus, Puccinia chondrillina. As populations of form A have become infected, they have become less competitive than they hack been, and their range has declined. Much of the range vacated by form A has been filled concorn~tantly with forms B and C (Burdon et al., 1981~. A similar reversal of competitive roles has also been documented between two cultivars of barley (HOT]eUm vulgare) that display a difference in their resistance to the aphid Schizaphis gTaminum. Un- der greenhouse conditions the aphid-res~st ant cultivar competes less weD in mixtures of the cultivars. If the aphid is introduced into the mixtures, the competitive advantage of the susceptible cultivar ~ lost (Windle and Franz, 1979a; Winnie and Franz, 1979b). These examples illustrate that small genetic differences be- t~veen closely related plants con produce phenotypes with different ecological properties that can increase or alter a plant's geographic range or enhance its aggressiveness in its normal range. How likely is this phenomenon for genetically mollified crops or other plants berg considered for field testing? Although most ecologically im- portant traits remain unchanged, the interaction among these traits determines whether a species will become naturalized in a region. For example, a species could spread because it tolerates herbivores and parasites and tolerates some aspects of the physics environ- ment (such as salinity) in the new range. Gottlieb (1984) compiled a list of diverse traits in plants that can be governed by one or a few genes. Whether the plant is erect or prostrate, branched or not, an annual or a biennial, or bears its leaves basally or higher on the stem can all be governed by a few genes. The suggestion from this list of traits is that major changes in plant architecture and subsequent performance could be achieved through rather small gene changes or insertions by recombinant techniques. Such changes in architecture would be readily detectable ~ greenhouse and field tests. The likeli- hood that these changes would occur randomly (and be retained) is very small.
42 ~ ~ ~e We do not know to what extent successful naturalization of exotic organisms hinges on their possession of one or a few traits rather than a group of characters. Multiple genes inducing multiple crates Snouiu Increase the probability of assembling an organism that can cause ecological changes if grown on a large scale. Several exotic species (for example, Cytisus scoparius, Ales eu- ropaea, Leucaena [eucocephala) owe much of their successful natural- ization to their ability to fix nitrogen in a new environment that is chronically low in nitrogen (Vitousek, 1986~. Nitrogen fixers, such as Ain?`s spp., characteristically are the first invaders on newly formed volcanic soils. Myrica faya, a smut exotic tree, is rapidly altering the nitrogen balance on volcanic sites in the Hawaiian islands. As the nitrogen content of the volcanic soil has increased, new species have become established on these sites (Vitousek, 1986~. Relatively minor genetic changes can produce plants with altered ecological properties, a phenomenon plant breeders have capitalized on for decades; for example, introducing a single gene in wheat can impart resistance to a specific race of stem rust. Similarly, herbicide-res~stant canola and soybean plants have been produced by minor genetic changes. Such changes have not resulted in increased weediness of these widely used crops. TElE ABII`ITY OF CROPS TO REVERT TO A WILD OR WEEDY CONDITION Crops that have been subjected to long-term breeding (for example, beans, maize, and wheat) are less likely to revert to a wild state than crops that retain many wild characters (artichokes, forage grasses, and grain amaranths). Highly domesticated crops have lost their ability to compete effectively with the wild species in natural environments. Domesticity arises because many characters that would enhance weediness (seed shattering, thorns, seed dor- mancy, and bitterness) have been deliberately elimunated from the crop plant through intensive breeding efforts. The reassuring history for cultivated crops does not completely preclude a genetically modi- fied crop from becoming weedy, but it suggests that the likelihood of that event is small. As new traits are inserted into cultivated crops, they might possibly change the crop in an ecologically significant way, but past experience with classical breeding h" shown this to be a manageable problem. Field trials should identify such possibilities. The descendants of crops may become weeds ~ agricultural
43 fields, and in some circumstances they may move beyond the bound- aries of the field and become weeds ~ sern~natural or even natural communities. More than a decade ago, Harlan (1975) compiled an often-cited list of the wild races of crops that included many row crops such as beets, cabbages, and watermelons. The relevance of these examples depends, in part, on the level of domestication in the crop. Some crops such as artichoke, sugar beets, and some citrus (Gade, 1976; Pickersgill, 1981; Thomsen et al., 1986), seem prone to become weedy. The ability to revert to a weedy condition has never been attributable to traits deliberately retained in the domesticated crops-that is, traits that have been the object of an active breeding program. HYBRIDIZATION BETWEEN CROPS AND TH1:IR WILD RELATIVES Two closely related ecological questions that may be important to the introduction of genetically modified plants are (1) Does hy- bridization between crops and their wild relatives result in transfer of traits from the cultivated form to the wild relative? and (2) Does such gene flow increase the weediness of wild relatives? If the oh portunity exists for the transfer of genetic traits from a genetically modified organism to a wild (and potentially weedy) relative, a po- tential problem exists. The problem poses three relevant questions: (1) Does the genetically modified crop have extant relatives? (2) What is the extent of hybridization between crop and relatives in nature? and (3) What is the current ecological role of the relative in natural ecosystems? Practically all crops have wild relatives at some taxonomy level. The more important question Is whether wild relatives occur in the range in which the genetically modified crop is grown or wiD be grown. The answer varies, as no one region of the world includes the home range of most crops, although arid central Asia and Asia Minor are the centers of origin for many crops (Table t13. Southeast Asia includes the home range of many weeds. Temperate North America, especially the United States, includes the home ranges for very few crops, as U.S. agriculture is based largely on crops of foreign origin. This paucity of crops derived from North American sources means there will be relatively few opportunities for hybridization between crops and ward relatives in the United States, except where both
44 TABLE 4-1 Crops and Their Probable Regions of Origin. tNote That Comparatively Few Crops Are Native to North America) Crop Scientific Name Common Name EUROPE AND TEMPER-ATE ASLA Cereals Avena sating L. Oats A. stri~osa Schreb. Fodder oats Hordeum ~ruI~are L. Barley Secale cereale L. Rye Triticum aestivum L. Bread wheat Pulses Cicer arietinum L. Chick-pea Lens esculenta Moench Lentil Pisum sativum L. Garden pea Vicia faba L. Broadbean Root and Beta vul~aris L. tuber crops Brassica raPa L. Daucus carota L. Raphanus sativus L. Beet, manger, chard Turnip Carrot Radish Oil crops Brassica camPestris L. Rapeseed Carthamus tinctorius L. Safflower Linum usitatissimum L. Flax, linseed Olea eurouea L. Oli Fruit and Ficus carica L. Fig nuts Ju~lans retrial. English walnut Phoenix dactvlifera L. Date palm Prunus amY~dalus Stokes Almond P. armeniaca L. Apricot P. atrium L. Cherry P. domestics L. Plum Pvrus communis L. Pear Vegetables Cucumis melo L. Melon and spices Allium cePa L. Onion A. sativum L. Garlic Brassica oleracea L. Cabbage, cauliflower, Brussels sprouts, kale, kohlrabi, broccoli Cucumis sativus L. Cucumber Lactuca sativa L. Lettuce Forage crops Bromus Dermis Leyss. Smooth bromegrass DactYlis ~lomerata L. Orchardgrass, cocksfoot F`estuca arundinacea Schreb. Tall fescue Medicazo sativa L. Alfalfa Phleum pretense L. Timothy Trifolium spp. The true clockers Drug crops Digitalis Purourea L. Digitalis Papaver somniferum L. Codeine, morphine, opium
45 Table 4-1 (continued) Crop Scientific Name Co~runon Name AFRICA Cereals Orvza ~laberrima Steud. African rice Pennisetum americanum (L.) Pearl millet K. Schum. Sorghum bicolor (L.) Moench Sorghum Pulses Vienna un£uiculata (L.) Walp. Root and Dioscorea cavenensis Lam. tuber crops Cowpea Yam Oil Crops Elaeis zuineensis Jacq. Oil palm Ricinus communis L. Castor oil Fruits and nuts Coloc~rnthis citrullus (L.) Watermelon Fiber plants Goss~rPium herbaceum L. Old world cotton o Forage crops C`rnodon dactvlon (L.) Pers. Dizitaria decumbenn Stent Era~rostis lehmanniana Panicum maximum Jacq. Drug plants Scoffed arabica L. Cereals and Bermuda grass Pangolagrass Lovegrass Guineagrass Coffee CHINA Fazop~rum esculentum Moench Buckwheat pseudocereals Organza sativa L. Rice Panicummiliaceum L. Proso millet broomcorn , millet ,Setaria italica (L.) Beaux. Pulses Glycine max (L.) Merr. Root and Brassica raps L. tuber crops Dioscorea esculenta (Lour.) Italian millet, foxtail millet Soybean Turnip Chinese yam Oil Crops Brassica camuestris L. Rapeseed B. iuncea (L.) Czern. & Coss. Mustard seed oil Vegetables Alium baker) Regel Chinese shallot and spices ,Cinnamomum cassia Blume Spice Cucumis satires L. Cucumber Zin~iber officinale Roscoe Ginger Drug plants Camellia sinensis (L.) Ktze. Cinnamomum camphor (L.) 1 Tea Camphor tree
46 Table 4-1 (continued) Crop Scientific Name Common Name SOUTHEAST ASIA AND PACIFIC ISLANDS Cereals and Orvza sating L. pseudocereals Oil crops Cocos nucifera L. Sesamum indicum L. Fruits and nuts Vegetablen Elettana cardamomum (L.) and spices Maton SvzY~ium aromaticum (L.) Merr. & Perry M~ristica fra~rans Piker ni~rum L. Solanum melonzena L. Citrus aurantiifolia Swingle C. aurantium L. C. limon (L.) Burm. f. C. nobilis Lour. C. paradisi Macfad. C. sinensis (L.) Osb. Musa acuminta Colla M. balbisiana Colla Starch and sugar plants (not roots) Cereals Zea maYs L. Fruits and nuts Anacardium accidentals L. Rice Coconut Sesame Lime Sour orange Lemon Tangerine Grapefruit Sweet orange Banana (A genome) Plantain (B genome) Cardamom Closure Nutmeg Black pepper Eggplant Saccharum officinarum L. Sugarcane MESOAMERICA AND SOUTH AMERICA Corn Cashew Ananas comosus (L.) Merr. Pineapple 13ertholletia excelsa HBK. Brazil nut Papaya Gray papaya Avocado Guava Carica papaya L. Carica candicans A. Persea americana Mill. Psidium zuaiava L. Vegetables and spices Capsicum annuum L. Capsicum baccatum L. Cucurbita maxima L. Cucurbita pepo L . Phaseolus Paris L. LYcopersicon esculentum Mill. Solanum tuberosum L. Vanilla planifolia Andr. Pepper Pepper Squash Squash, pumpkin Bean Tomato Potato Vanilla
47 Table 4-1 (continued) Crop Scientific Name Common Name Fiber plants Gossynium hirsutum L. Drug plants Nicotiana tabacum L. Theobroma cacao L. NORTH AMERICA Oil crops Helianthus annus L. Upland cotton Tobacco Cacao, chocolate Sunflower Fruits and nuts Vitis labrusca L. Fox grape V. rotundifolia Michaux. Muscadine grape Vaccinium macrocarpon Aiton Cranberry Vaccinium (several species) Blueberry Fra~aria several species) Strawberry Rubus idaeus Richardson Red raspberry Rubus (several species) Blackberry Rubus (several species) Dewberry Vegetables Helianthus tuberosus L. Jerusalem artichoke ADAPTED FROM: Harlan, 1975. crop arid wall relatives have immigrated (Table 12~. The incidence of hybridization between genetically modified crops and wild relatives can be expected to be lower here than in Asia Minor, southeast Asia, the Asian subcontinent, and South America, and greater care may be needed in the introduction of genetically modified crops in those regions. If a crop has no relatives within the distance its pollen can travel, no hybrids will develop. Spatial separation Is an obvious barrier to hybridization, but only anecdotal knowledge exists on the actual limits of pollen transport (Elistrand, 1988~. Furthermore, even if relatives are nearby, there Is no assurance that viable hybrids wid be produced, as there often are many formidable barriers to gene flow, such as differences in ploidy level, flowering time, tends breeding systems (Sirnmonds, 1979~. ~ fact, the deliberate introduction of genes from wild relatives into certain crop species by classical breed- ing techniques has been achieved only by manipulating the flowering tune and by repeated hand pollination (as in potatoes). Even if fertilization is accomplished naturally, there is no assurance that
48 T~LE 4-2 Some C~ps Growing SympatricaHy in tbe Onited States witb Congene" or Wild Races with Wbicb Natur~1 Hyb~dization Is Possible. C=p Prim~ Ceue Pool Sor~um bicolor (so~bu=) RsDh~us sst1vus (radisb] SetaHa itallca (~xt~1 ~uet) Br~sica r~s (tumip) Br~slca c~mnestris (rape) Amaranthus c~entus (am~rantb]; ^. c~udatus; ^. bvoochondriscus Bets vul~ds (beet) D~ucus carota (~t] "eliantbus annuns (sunDower) Cucurblts oeno ~qu~b, pumpkin] Bec~le cereale (~e) L~ctucs sativ~ (~ttuce) ~ens satlva (oat) Cvnara scol~mus (~hoke) S. halenense (Jobnson ~-sj R. r~ohanlst~m (wild rsdiab) S. italics ~equently naturahzed ~ ~ weed, m~ not ~st in the United States B. c~mpestHs (~ ~) B. camDest~s (~Hd ~, Deld mustard} 8. hvbrldus; A. oowellIi; A. retronexus - h~e o~ w~dy ~e in Europe D. cs~ta soot c~ota H. annuns (wild ~o~bs); H. bolande~ C. tex~na (w~d ~) S. cereale ~d S. montanum L. se~ols A. ~tus (w~d o~) C. scolvmus (wild types) ~TED P ILY FRO~: N. W. Si~ond~ ed~ 1979, and ~~renc" tbe~in. ~rtber plants w1D be produced. ~r a gene to p~s be~en rel~ tlve and crop ~nd be permanently lncorp grated luto eltber tbe crop or tbe rel~lve] ~trogre~lon (~troducdon ~ a gene ~om one gene co~lex luto ~otber) ~st occur regul~ly (^derson' 1949j. Ibls occurs ~ exceedlugly 1~ hequency ln many crops ~d w~d relative . . . comDm~lons. Evldence kr gene introgress10n by bybrldlzatlon be~en crops
49 and wild relatives has often been only circumstantial. Because plant breeders are usually concerned with the detection and elunination of wild traits in a crop, the low incidence of documented transfer and ~ntrogression that occurs from crops to wild relatives may be an artifact. A complication in reliably identifying such Retrogression continues to be the possibility of convergent evolution between crop and wild relatives. The mechanism by which this could occur is ease fly envisioned: Art agronorn~c practice such an seed sorting by size imposes strong directional selection in a wild relative (or even an unrelated weed) for those phenotypes with the same seed size as the crop (Barrett, 1983~. Seed size, shape, and even color can be re~nark- ably similar between the crop and the weed without hybridizations occurring. Forty years ago, plant breeders In India selected for increased anthocyan~ production in cultivated rice in an attempt to Prove the ability of paddy workers to d~scri~nate between otherwise in- distinguishable seedlings of cultivated and wild rice (Oryza species). Although the cultivated rice seedlings were readily identified at first by their purple leaves, within several plant generations the trait had been transferred to the wild relative, thus rendering the trait use- less from a cultivation standpoint (Parker and Dean, 1976~. Other putative examples of gene flow from crop to wild relative have been reported for crosses between corn and teos~nte, Eastern Carrot" and wild carrots, ~kayseri" alfalfa and weedy relatives, and between durum wheat and wild emmer wheat. The evidence Is mainly mor- pholog~cal and therefore subject to alternative interpretations (for example, convergence after mutation in the wild relative and subse- quent directional selection) (Small, 1984~. Other examples, also based largely on morphological evidence, occur among the cultivated Amaranthine ca?~datus, A. cruentus, and A. hypochondriacus. Each of these species forms hybrids with one or more weedy amaranths ~ California and Mexico. Gene flow to the weedy amaranths is probably more obvious and persistent because of the strong selection by hand-cultivation against the preservation of hybrids with the wild parent's trait of dark seed (Saner, 1967; Tucker and Sauer, 1958~. ~ the Sacramento-San Josquin delta, Tucker and Saner (1958) identified many amaranth hybrids that resulted from crosses between crop add wild relatives. They maintained, without direct evidence, that under cultivation in the light, highly fertile organic soil in the region, hybrids could out-compete their weed parents (A. hybridus, A. powellii, and A. retropexus) because they
so had acquired traits from their crop parents for more robust stature and high fecundity. Gene flow has apparently occurred from crop to wild relative in rye (Secale spp.) in California, where a weedy rye probably derived from a cross between S. cereale and S. montanum has become increas- ingly corklike through introgression with the cultivated S. cereale. This introgression has proceeded to such an extent that farmers have abandoned efforts to grow cultivated rye for human consumption and are deliberately sowing the hybrids for forage (Jain, 1977; Suneson et al., 1969~. ~ each of these examples, the putative transfer of a trait from the crop to the wild relative has resulted In the relax fives' becoming more similar to the crop; ~ the above-cited example with Asian rice, the ~ntrogression resulted only in an enhancement of mimicry of the crop. Evidence ~ restricted to morphological or cytological similari- ties between the crop and the wild relative. However, much of the evidence Is circumstantial rather than exper~rnental; clear demon- stration of introgression depends on molecular analyses of isozymes or other techniques. Recent work with molecular marker loci has refuted several earlier claims of ~ntrogression in Helianthus (Riese- berg et al., 1988) and some reports of introgression between maize and teos~te (Doebley, 1984~. Even with ~sozyme studies there Is the possibility for an alternative interpretation; the crop and the wild rel- ative may share aDeles derived from a common ancestor rather than through more recent introgression. Consequently, the best evidence for recent gene transfer arises in cases in which a wild relative pos- sesses alleles in common with a crop, but only in those populations that have recently come into sexual contact with the crop. Convincing evidence for a transfer of genes from a crop to a wild relative does exist in several crowed complexes: African rice, maize, and Cucurbita. Second (1982) hap shown that African rice, Oryza brev:7igulata, contains more isozym~c variation than cultivated rice, O. glaberrima. His data suggest that this variable weedy rice arose through introgression between the wild form and cultivated rice. Doebley (1984) found evidence for introgression of cultivated maize into Zea diploperennis; one plant possessed two alleles that had not been found previously ~ the wild species but that are common in maize. Because the two loci are tightly linked, there is at least the strong suggestion that the chromosome segment carrying these loci was transferred through hybridization Alto Zea diploperennis. In the southern United States the cultivated Cucurbita pepo (squash) occurs
51 in the same area as the wall species, CUCUTbi~a texana (Texas gourd). Decker and Wilson (1987) found that alleles typical of the cultivar can occur in the wild species. This introgressiorr enhanced weediness in the sense of making the hybrids more difficult to distinguish from the crop (that is, their mimicry of the crop increased), but the hybrid was no more aggressive, nor did it have an enhanced ecological range. Consequently, the products of the inadvertent transfer of crop genes to relatives have been confined to the field in which the plants were grown. From the standpoint of eradicating the weeds, the result of this introgression is at worst undesirable. The hybridization between cultivated sorghum and one or more of its wild relatives is more serious. "Hybrid grain sorghums (SOT_ chum bicolor,} Is produced through the cytoplasmic male-sterility method In which two inbred lines are hybridized. The seed is har- vested from the male-sterile plant. If pollen of one or more weedy sorghums is inadvertently allowed to fertilize the stigmas of the male- sterile plants, the offspring are useless commercially and represent a genetically diverse cluster of races and "off-types called shatter- cane. These plants usually express many traits of the wild parent, such as the perennial habit (inherited from Sorghum halepense), height, or self-sow~g seed (Baker, 1972), a trait inherited from Sorghum sudanense. Hybrids bearing traits of S. halepense (Johnson grass) present potentially serious weed problems because the vegetatively vigor- ous S. halepense Is eradicated only with great difficulty and expense (Holm et al., 1977; Warwick and Black, 1983~. The direct role of introgression with the cultivated sorghums In the enhancement of weediness in S. halepense ~ not clear, but the circumstantial evi- dence at least suggests the production of more persistent plants. De Wet (1966) maintains that S. halepense in its native range in the 01d World has never been an excessively weedy plant and that its weediness was enhanced coincidentally with its introgression with cultivated sorghum in the United States. Acquisition of these traits is unusual in that their advantage to the weedy offspring ~ not con- fined to enhancing the weed's mimicry of the crop. If these Johnson grass populations extend their already major ecological role outside agricultural fields, they will represent the most extreme category of known risk associated with gene flow from crop to weedy relative. Biotypes of S. halepense in the northeastern United States apparently have acquired traits of ecological importance through introgression,
52 including such crop-like features as earlier flowering, greater seed pro- duction, larger individual seed weight, and subsequently more rapidly emerging seedings than other biotypes (Warwick et al., 1984~. The male sterility method produces a similar, although less ~e- rious, weed problem In the cultivation of sugar beets for seeds in northern Europe. If these mal~sterile plants are inadvertently polli- nated by the pollen of Beta v?~Igaris subsp. maritime (wild sea beet), some cultivar x wild Fit hybrid seed eventually wait be produced in the crop field (Pickersgill, 1981~. While hybridization between a crop and its wild relative may not be prevented by morphological, cytological, and developmental barriers, there is little likelihood that domesticated traits will be retained In a wall relative. Much of the emphasis in plant breeding has been toward traits that would reduce adaptation to the wall (for example, enhanced of} content in the seed, or an enlarged fleshy root), especially if enhanced production for these features came at the expense of plant fitness. Important cornmercia] trmts, such as pest resistance, that have the potential to alter the ecology of wall relatives have not been a problem with the possible exception of gene transfer from cultivated sorghum to Johnson grass. SUMMARY POINTS 1. The analogy between the introduction of an exotic species into a new environment and the introduction of a genetically modi- fied crop plant is tenuous because introduced exotic plants that have caused problems bring with them many traits that enhance weedi- ness, whereas genetically modified plants are modified in only a few characteristics. 2. Genetic modifications of only a few genes can produce a modified plant with significant, ecologically important alterations. However, genetically modified crops are not known to have become weedy through the addition of traits such ~ herbicide and pest resistance. 3. Domesticated crops, such ~ soybeans, corn, and wheat, have been genetically modified to such an extent that they can no longer compete effectively with wild species in the natural ecosystem. These crops are unlikely to revert to a weedy condition upon further genetic modification. Some forage grasses are more likely to revert to a weedy condition.
53 4. Most crop plants in the United States are not native, and, unless weedy close relatives have been imported, no close relatives with which the crop knight hybridize are present. However, where cross-hybridiz~ng wild relatives do exist ~ close proximity (such as the sunflower), precautions may be necessary to limit gene flow from the crop to the wild relative. Gene introgression, when demonstrated, has often caused the wild species to become more like the crop, with consequences of enhanced weediness of the wild relative largely confined to agricultural fields.