Pollen densities on the host leaves were estimated using a double-sided adhesive tape glued
onto microscope slides. Caterpillars were sampled from the plants at the beginning and end
of pollen shed. They were carefully replaced back on the plant after identification.
Conventional bioinsecticides based on Bt protein have been used for several years in the
control of pests, even before the development of Bt transgenic crops. Studies on their
nontarget effects have generally shown no negative impact on predator (parasitoids to the
insects) populations. However, these results cannot be fully extrapolated to plants expressing
the Bt protein. The microbial Bt products contain Bt protoxins, which are activated in the
insect’s midgut by proteases (see Chapter 8). Some of the transgenic plants, on the other
hand, express partially activated Bt proteins, which could have a potentially different
impact on the insect populations. Hence, it can be argued that there is a need to investigate
whether the unique delivery system, and the constant exposure of the protein to the
insects, has an effect on natural enemies.
13.3.2. Statistical Analysis and Relevance for Predicting
Potential Adverse Effects on Butterflies
Field testing requires careful analysis. For the German field trial on caterpillars, we used a
statistical evaluation called the “proof of safety” between Bt maize and the nearisogenic
variety (ISO). Maize pollen density was estimated to be 52–972 pollen grains/cm^2 on
Chenopodium albumand 100–894 pollen grains/cm^2 inSinapis alba. No significant
differences were observed in pollen densities between plant species. Note the wide range
of potential exposures. Of the nine butterfly species recovered from the field, only two—
Plutella xylostellaL. andPieris rapaeL.—were abundant enough to be considered for
statistical analysis. Caterpillars in both of these specieas are considered to be pests on
mustard crops, such as canola, cabbage, and broccoli, but not on maize. Throughout the
study period, the numbers of caterpillars (of bothP. xylostellaandP. rapae) were lower
in plots with insecticide treatment (Fig. 13.2). Pollen density on the plant leaves can be
affected by several factors, including relative humidity, growth stage, and distance from
maize fields, as well as shape and structure of host leaves (e.g., waxy or hairy surfaces).
It was observed that more pollen was shed (as inferred from pollens/cm^2 ) from Bt maize
in comparison to the conventional maize; however, this could be attributed to the better
health of the plants themselves. The Bt plants were observed to be more robust than
their isogenic counterparts because they were not damaged by European corn borer as
were the ISO plants, which would be expected to lead to the production of more pollen.
Hence, no reliable conclusions about the (possible) more adverse effects that they could
have on butterfly species could be deduced. No statistically significant detrimental
effects of the Bt pollen on the larvae were found (Fig. 13.3). The most important reason
for the differences in laboratory results and those from field testing (the latter indicating
low overall risk) is the very low level of Bt protein exposure to caterpillar in the field as
Bt corn pollen is a much rarer food source under realistic environmental conditions. A
less important reason is the temporal overlap between caterpillar development and pollen
shed. By the beginning of pollen shed, caterpillars often develop to the final instar stages
(Fig. 13.4). Susceptibility to Bt protein is known to decline with older caterpillars,
thereby reducing the effect that Bt pollen could have on them. Similar studies were done
on the monarch butterfly to estimate the potential risk under field conditions in the
United States. After considering distribution data of the monarch butterfly and their host
plants, overlap between pollen shed and development of larvae, and exposure of larvae
13.3 AN EXAMPLE RISK ASSESSMENT: THE CASE OF Bt MAIZE 317