Dangerous Liaisons: When Cultivated Plants Mate with their Wild Relatives

  • Norman C. Ellstrand
Johns Hopkins University Press: 2003. 268 pp. $65, £48

The ecological effects of genetically modified (GM) crops remain controversial, despite evidence of the crops' agricultural benefits, such as reduced pesticide use, fewer human poisonings and increased net incomes for farmers. Their most ardent critics argue that GM crops lead to the rapid evolution of resistance in pests, harm non-target species and soil organisms, and create ‘superweeds’ by introducing transgenes into wild plant populations.

However, the transgenic crops that have been commercialized to date seem largely to have defied these predictions. For example, there has been no detectable increase in insects' resistance to the bacterial toxin produced as a pesticide by transgenic Bt crops. Resistance to the herbicides used with GM crops (mostly glyphosate) is still relatively minor compared to that with other widely used herbicides. The damaging effect of Bt crops on Monarch butterflies proved to be exaggerated, and studies on other non-target species have consistently failed to find any significant direct effects. Studies of the interactions of soil organisms and processes with GM crops have also failed to show any significant detrimental effects.

In Dangerous Liaisons, Norman Ellstrand addresses the risk of GM crops sharing their genes with wild and weedy plants — either uncultivated populations of the crop itself or closely related species. His well-written book is aimed at a wide audience of students, academics and policy-makers. The popular press seems to have a difficult time with stories about pollen flow from GM crops, typically extrapolating from studies on sunflowers, maize, canola or sugar beets to all GM crops grown anywhere. Ellstrand's book will help them identify more specific risks. The book will surely gain wide attention because of its sexy title and emphasis on GM crops, but I was more impressed by Ellstrand's documentation of the broader effects of conventional agriculture in swamping wild populations of rare plant species.

Ellstrand provides an introductory section for readers who are not population geneticists, before detailing hybridization between domesticated plants and their wild relatives, and then presenting his interpretation of these observations.

Despite his effort to provide this broad context for hybridization between crops and wild plants, and its consequences, I suspect that most readers will focus on chapter 7, where Ellstrand contrasts his views with an opening quote from the Israeli plant scientist Jonny Gressel: “Most crops have no interbreeding relatives in most of the world.” Ellstrand reviewed the data for the world's 25 most widely planted crops, summarized their interbreeding with wild plants in a single table, and showed that 22 of them do hybridize with wild relatives somewhere in the world. I suspect that this table will be the most widely referenced in the book, and wish that a few of the distributions were more precisely stated. For example, cotton, beans and potatoes are listed with a “multicontinental” distribution of hybridization; more precisely this refers to Latin America (and, for cotton, some islands in the Caribbean and Pacific).

But what about Gressel's proposition, especially in the context of GM crops? Using statistics from the database of the Food and Agriculture Organization of the United Nations, cited by Ellstrand (http://apps.fao.org), I checked on the dominant GM crops. At least 87% of the world's soybean crop, and 95% of the world's maize and cotton, are grown in countries for which Ellstrand lists no hybridization — and even for those countries with hybridization, such as China for soybeans, wild relatives are found in only some areas. Many of the other crops are complicated to tabulate, but Gressel seems to be correct that most crops are not interbreeding locally with wild relatives.

This still leaves the possibility that serious problems could arise in the few areas of the world where hybridization can occur. This currently seems possible for GM canola in Canada and the United States, and for transgenic maize that is probably growing illegally in Mexico (but has apparently escaped documentation in the refereed literature). But even after reading this book, I haven't seen any evidence of harm to human health or to the environment (including weediness) from such hybridization. Where are the super-weeds that were predicted to occur from the exchange of transgenes with wild relatives?

In contrast to the lack of evidence for deleterious effects of gene flow from GM crops, there is evidence that conventional agriculture has adversely affected wild plants through genetic swamping of their populations, and that wild plants have generated weediness in crop–weed hybrids. As noted by Ellstrand, “problems associated with hybridization between conventional crops and their wild relatives received scant attention until potential gene-flow problems were described for transgenic crops”. For example, hybridization with cultivated rice has been implicated in the near-extinction of an endemic Taiwanese wild rice. Hybridization of maize with its ancestor teosinte may be contributing to the extinction of teosinte populations. Indigenous cotton in the Galapagos Islands could be at risk of extinction or replacement as a result of hybridization with cultivated cotton. Ellstrand cites similar evidence for at least another nine species. He also documents in great detail the history of sugar beets in Europe, where hybrids between cultivated beets and their progenitors, the sea beets, have caused major weed problems.

Everyone interested in the effects of cropping on plant biodioversity, the evolution of weeds and the risks of GM crops should read this book. Critics and supporters of transgenic crops will continue to debate whether the relatively benign environmental and agronomic disadvantages of GM crops have been due to largely to luck or to an adequate regulatory system. Ellstrand reminds us in detail that the reliability of future successes depends on more careful risk assessment of hybridization with wild relatives.