My struggle with weeds over the years has made me aware of the damage they can inflict in gardens, farms, and native ecosystems. I have learned to be vigilant and untrusting of even the smallest, innocent-looking weed seedling and yank it out upon first sight. I have grown curious as to why many weeds spread invasively whereas most crops and native species do not. Both crops and weeds have pollen that can spread widely, right? So why the difference in invasiveness? To answer this question, we need to consider where weeds originate, why they persist and reproduce, and how domesticated crop plants differ from weeds. These issues have become important in the debate about the potential impact of genetically engineered (GE) crops. Some people worry that the presence of GE pollen in the environment will create a new breed of invasive, uncontrollable weeds that will overrun pristine environments or irrevocably alter the genetic makeup of native species. Foreign “Invaders” The weeds in my garden share similar characteristics. Consider yellow star thistle, a non-native plant with gray-green lizard-shaped woolly leaves that is toxic to horses. If ingested they can become ill with a neurological disorder resembling Parkinson’s disease in humans. Over the last 150 years this weed has spread over twelve million acres in California. Yellow star thistle possesses an important invasive trait: it is able to complete its life cycle quickly. It germinates with the first rains of fall, sending its roots down to depths of six feet or more where it sucks up all the moisture so that there is none left for the slower-germinating native species. Second, yellow star thistle is an alien weed that has evolved adaptations that allow it to survive and spread. Introduced from Europe, it traveled to California from Chile as a stowaway in alfalfa seed, where it was then inadvertently planted with the hay crop. Third, yellow star thistle produces a large amount of seed and has aggressive vegetative structures. In midsummer, one yellow star thistle plant produces seedheads bearing long tan spikes that can yield 100,000 seeds. These examples illustrates the ways in which weeds can “outsmart” their domesticated cousins and create havoc. In contrast to invasive weeds, domesticated crops are tame. Consider a field of sweet corn-tall plants with gaudy tassels perched on top of single stems, and large ears. These traits—large fruit, reduced branching, gigantism, reduced seed dispersal, and a lack of genetic diversity—are all signs of domesticity. In the California Central Valley, this corn is grown on farms located quite close to the inner coast range areas. Because of their proximity, it would seem that the crop plants could escape to these foothills. They have not. The foothills are abundantly covered with weedy oats, bromes, and starthistle, but domesticated crops from Central Valley farms are notably absent. There is no corn, no soybean, no alfalfa, no cotton, no tomatoes, no safflower nor rice growing in these areas. Although these domesticated plants are also aliens—tomatoes and corn from Central and South Americas, cotton from what is now Pakistan, safflower and alfalfa from the Near and Middle East, and rice from China, any residual weediness has been eliminated through many years of breeding. This is one of the reasons that the genetically modified corn and cotton, grown here for 150 years, have not established in the foothills. GE cotton and corn, the primary transgenic crops grown in the Valley, are not likely to survive either. After all, a GE crop is still a crop and crops make lousy weeds. The traits that make these plants good for farmers make it hard for them to survive in the wilderness. The Impact on Pollination What about GE pollen? Will it drift over to the nearby foothills to create a new kind of weed that will pollute native ecosystems? What if transgenes move from a GE crop to a weedy relative? Can transgene pollen flow somehow transform a crop into a weed or change an ordinary weed into a “super” weed? Most experts say that this is unlikely. That is because it takes many genetic changes to become a weed, arising from a combination of gene flow, spontaneous mutations and other factors such as changes in environmental conditions. There can be no gene flow, that is to say, no sex, without two willing partners. And most plants are quite choosy, preferring a close relative rather than someone outside its family. Pollen from crop plants (GE or non-GE crop) can travel around all it wants—in gusts of wind, on the pollen basket of bees, as cargo of flies or in the hands of human plant breeders—but unless the pollen alights upon a compatible mate, there will be no fertilization and therefore no seed. And if there is no progeny to pass genes onto, there can be no gene flow. In the Central Valley, genes from GE crops plants cannot be shared with the native populations nearby, because the GE crops grown here have no sexually compatible relatives in the foothills. This means that the GE species grown in this great valley are trapped. It is as if California were a large, oval-shaped, flat-bottomed platter with steep, slippery sides holding all the GE crop plants at the bottom. Of course, as more crops are genetically engineered, the picture will not always remain so simple. This is because cultivation of other crops could potentially create problems under certain conditions. For example, most ecologists believe that if pollen carries a trait that confers a “fitness” advantage (thereby enhancing viable seed production), and it has wild relatives nearby, it could potentially establish in some environments and become invasive. If the gene confers no fitness advantage it would be lost from the population over time. If it confers an advantage and is passed on to relatives, it would be maintained in the population. For example, if a drought tolerant gene from wheat hybridizes with a related weed called jointed goatgrass and if the hybrid establishes, it could become more of a problem in the western United States. There is certainly evidence for cross-hybridization of crops with wild relatives, but few if any of the resulting hybrids have become invasive. For example, in Quebec, Canada, domesticated GE Brassica napus (canola) is able to hybridize with a weedy relative called wild radish (Brassica rapa). According to Norm Ellstrand, a population geneticist at University of California, Riverside, “canola is as yet the only case known in which engineered genes from a commercial crop have been found in natural populations.” Although the transgenes could be found in hybrids between the two Brassica species, they slowly disappeared over subsequent generations and their presence, therefore, did not alter wild populations. Another study demonstrated that these Brassica hybrids actually decreased competitiveness of the wild radish species, turning this particular weed into a “wimp.” Scientific studies looking at the issues of gene flow between domesticated and wild relatives have shown that crop domestication has not benefited the wild relatives. Despite intensive breeding for stress tolerance in annual crops, “there appear to be no known cases where populations that are substantially more invasive in the wild were generated as a consequence.” Apparently it is quite difficult to turn a docile crop into a promiscuous weed. For example, creeping bentgrass is a perennial weed with extraordinarily light pollen that cross-pollinates with at least twelve other species of grass. It has been cultivated on Oregon golf courses for decades. Unlike crops that have trouble surviving off the farm, creeping bentgrass can easily survive in the wild. Nevertheless, golfers and caretakers of fairways like this weed because it is low-growing and easy to take care of. Now, two companies, Monsanto and Scotts, have genetically engineered this weed for tolerance to the widely used herbicide glyphosate (often sold as Roundup). Other weeds on the golf course will be killed by Roundup but GE creeping bentgrass will survive in the presence of the herbicide. Not surprisingly, GE pollen behaves no differently than its non-GE counterpart. Researchers have found creeping bentgrass transgenes in the progeny of wild populations 14 kilometers away from the source field. The transgene is not expected to spread in the population in the absence of the herbicide and is therefore not a problem for gardens or the wilderness. In other words, the GE weed is still a weed, but survives no better than its non-GE counterpart. Still the case brings up interesting questions as to what should be cultivated or genetically engineered. From the point of view of a gardener who spends several hours a week pulling weeds and adding them to my compost pile, it makes no sense to plant weeds, let alone cultivate them. On the other hand golfers seeking smooth fairways might prefer that I take my weed-magnet of a garden elsewhere. While GE plants currently grown in California do not have an opportunity to interbreed with wild species, some of the California crops are exported widely and could possibly end up in environments where there are sexually compatible species. Mexico, for example, imports several million tons of corn from the United States each year. The Concern for Genetic Diversity In a study of corn landraces (crops selected for their adaptations to specific locations and their culinary characteristics) in Northern Oaxaca, Ignacio Chapela, a professor in the Department of Environmental Science, Policy and Management at UC Berkeley published a paper in Nature providing evidence for the presence of transgenic DNA in these landraces. The published results ignited an explosion of worldwide publicity because transgenic corn had never been approved for cultivation in Mexico, and there was concern that the presence of transgenes might compromise the genetic diversity of these landraces, valued because they carry prized genes for disease resistance and other agronomic or gastronomic characteristics. Although the results presented in the initial publication were widely disputed and then refuted by a larger peer-reviewed study in 2005 by Ortiz-García et al., the paper prompted an important debate over possible biological, economic, and cultural implications of gene flow. These issues are increasingly important because Mexican corn growers want to use GE to improve productivity and poor consumers rely on this staple. Carlos Salazar, president of the Mexican National Confederation of Corn Producers, estimates that more than 90 percent of small and medium growers would use GE seeds if they were available. Recently, Mexico’s corn growers signed an agreement with Monsanto to buy and plant genetically altered seeds. Will a future massive planting of GE corn create a problem for local landraces? As Paul Gepts, a geneticist at UC Davis, points out, because domesticated non-GE modern hybrid varieties are now widely planted in areas of high biodiversity, “modern” genes are already present in local landraces, often introduced by local farmers who wish to generate new varieties. It is unlikely that a single transgene by itself would reduce the genetic diversity of native populations to a greater extent than is already occurring. Fortunately, at the policy level, native landraces have actually benefited from the discussions on GE corn in Mexico. The GE corn debate has led to greater recognition of the value of indigenous landraces and Mexican growers plan to initiate activities to protect these landraces, including setting up a maize germplasm bank. They recognize that cultivation of modern crops, GE or non-GE, needs to be examined carefully in order to safeguard the center of genetic diversity where pollen flow could impact the genetics of local plant populations.