allergenicity concerns. It is important to note that modern biotechnology can also be used to reduce the allergenic risks associated with our current food supply. For example, Matsuda et al. (1998) have published papers showing that they reduced the major allergen in rice by approximately 80% by using antisense rDNA technology.

Other safety issues which have received attention are those involving ecological risks such as the effects of gene flow. Studies have been conducted to assess the potential for gene flow among and within related species (see section 2.7 and section 3.4.1). The ability of transgenic plants to cross-pollinate with their wild relatives and form offspring with enhanced weediness has been investigated when herbicide-tolerant rapeseed plants were back-crossed with a wild relative. The hybrid progeny plants produced an equivalent amount of seed as the wild genotypes and were also herbicide-resistant (Snow et al. 1999). That study indicated that back-crossed generations of hybrids between transgenic and nontransgenic crops can have the same potential to flourish as other plants. In a more controversial study, wild type Arabidopsis thaliana plants were found to be fertilized by pollen from transgenic plants more often than by pollen from nontransgenic plants (Bergelson et al. 1998). In addition to those experiments, the use of models has been explored to assess the invasiveness of engineered organisms, although indications are that these models will require several years worth of data to be validated (Kareiva et al. 1996).

Although some transgenic pest-protected plants have the potential to reduce pesticide use and thus to prevent substantial environmental damage, there is concern that gene products from the plants could harm beneficial insects or birds (nontarget species) that are in direct contact with the plants or that feed on insects that are (see section 2.6 and section 3.1.2). Hillbeck et al. (1998a, b) found that when chrysopid larvae were reared on prey that were fed Bt-producing corn, they had 62% mortality. When they were reared on prey that were fed non-Bt corn, mortality was only 37%.

Another experiment indicated that Bt toxins can bind to humic acids from soil, be protected from biodegradation, and persist in the soil (Crecchio and Stotzky 1998). It is not known whether nontarget organisms would be affected by bound toxin molecules in field situations. Other studies indicate that Bt toxins generally degrade quickly in the soil (Palm et al. 1994; Sims and Sanders 1995; Palm et al 1996).

A well-publicized recent laboratory study indicated that when mon-