FYI- Here is May Berenbaum's (Univ. of IL Entomology Dept.) reply to Jane Rissler (Union of Concerned Scientists) regarding May's research on Bt pollen and swallowtails, which was criticized by ISIS in an article titled "Swallowing the Tale of the Swallowtails: No "Absence of toxicity" of Bt Pollen" posted to the list on 6/23.

Dear Jane,

Thank you for forwarding the article from ISIS News (I'm unfamiliar with the publication -- who publishes it?). I'm not really surprised to see this reaction to our study but I do think the author of the article is not being completely objective in some of his or her remarks. For one thing, there is no single ideal way to design a "proper control" for a study -- the design of the control depends entirely on the nature of the question being asked. In our case, we were asking the question, "Does living near a Bt cornfield cause higher levels of mortality than living farther from a Bt cornfield?" and not "Does living near a Bt cornfield cause higher levels of mortality than living near a non-Bt cornfield?" The control the author suggests is appropriate to address the second question but not the first.

As I mentioned in our phone conversation, we actually considered designing the study in the second way but it proved to be impracticable -- the Bt cornfield under study was so large that we could not have constructed one quickly enough that it would have ended up phenologically and ecologically similar enough to have made a comparative study effective (or even interpretible). As for the rain, rain is what happens in a field experiment. We didn't plan for it to happen, but it's what happens in cornfields in Illinois in the summer and our field study was designed to test impacts of Bt corn pollen under field conditions. It may be true that pollen washes off leaves and microscope slides at different rates but I'm confused by the objection raised here -- if rain is actually effective enough to have "washed away the pollen from the surface of the leaves," then how could Bt pollen cause mortality? In fact, we have no reason to believe that rain works that efficiently on parsnip leaves or that rain doesn't wash pollen off microscope slides coated with Vaseline. Notwithstanding, the objection doesn't seem internally consistent.

As for variable mortality in the field, that, too, is what happens in a field study. The rates of first instar mortality we observed in the field are very comparable to rates reported by Feeny et al. (1985) in upstate New York for this species, long before any Bt pollen was a factor. It's not impossible to draw a conclusion from a study with variability -- the appropriate conclusion to draw is that the postulated explanatory factor does not in fact explain the variability observed. The point that mortality rates should be correlated with larval mass is baffling to me -- I know of no studies in the literature that demonstrate predators of black swallowtails preferentially select large prey and the author of the piece doesn't provide any literature to support the existence of this relationship for any species. In fact, at least some of the predators (e.g., anthocorid bugs) most likely prefer smaller, less active prey, so the postulated correlation isn't even intuitive.

I'm also baffled by the criticism of the lab study -- we tested a series of dosages (which is the standard way to assess toxicity) that bracketed (actually exceeded at the high end) concentrations measured in the field. Moreover, caterpillars were given the opportunity to consume these leaf discs in their entirety over the three days of the bioassay, as they presumably would in the field. Our laboratory study was done in large part as a check on the field study -- a "worst case" scenario to see if Monsanto 810 pollen could cause mortality if the caterpillars are force-fed pollen and if environmental variation is eliminated. The lab study does support our field study -- it's not proof, but it is corroborative. I'm puzzled that the bioassay used by Losey et al. 1999 doesn't cause concern to the author, given that there was no quantification of pollen administered, no confirmation of endotoxin level expressed, and, most importantly, no confirmation that the caterpillars actually ingested the pollen presented (the criticism that is now being leveled at us).

The monarch larvae could well have simply starved to death due to antifeedant effects of Bt pollen (an effect that would likely not be manifested in the field because alternate, pollen-uncontaminated leaf surfaces would have been available). In our case, caterpillars were provided with disks until they were completely consumed (so pollen had to have been ingested) and then they were provided with new, untreated disks.

It's extremely disappointing to see such inflammatory language used (e.g., "abuse of science") in this critique, particularly when it appears that equally strict standards are not being applied to studies with outcomes that are more consistent with a particular world-view. As you undoubtedly know, Losey et al. did not report levels of pollen used in their bioassay, did not provide a pollen control with a similar genetic background to the Bt pollen, did not measure amount of pollen ingested, did not quantify endotoxin content, did not use a species of milkweed that grows near U. S. cornfields, and did not do a field study that could take into account such factors as behavioral avoidance of pollen, UV breakdown of endotoxin, and even rain. The author seems to have great confidence in Obrycki et al., which is not yet published, despite similar problems (e.g., use of Asclepias curassavica, use of monarchs from a longstanding laboratory colony, and use of leaf material brought into the laboratory for bioassay rather than exposure of caterpillars in the field). For whatever it's worth, Obrycki et al. didn't apparently use a "proper control" either in that all of their leaf samples were taken from within or near a Bt cornfield; as far as I can tell from the abstract, there were no samples taken from a nearby "control" non-Bt cornfield.

We know that our field study has limitations and in fact we made a point in several places in our paper of explicitly addressing the limitations of the study (e.g., "This is not to say that monarch butterflies are unaffected by Bt corn pollen"), which is more than Losey et al. 1999 did. This summer, we intend to do another field study, this time with three species of Lepidoptera and with Novartis Max 454, which we know to be toxic in the lab. Because we believe it's the proper design, we will be using the same design that we used this year -- varying distance from cornfield to vary pollen exposure. If in fact we find (as we well might) that Max 454 does cause mortality to lepidopterans under field conditions, will an ISIS writer jump on the study and call our conclusion an "abuse of science"?

Our interest in conducting this study was not to vindicate Bt corn -- in fact, our lab study does indicate that some varieties pose a risk. Our interest was to find out what's really going on. I hope, as a scientist, that is your goal as well. As I indicated on the phone, I believe sincerely that policy decisions should be made on solid scientific evidence. I don't believe that there is enough solid scientific evidence to make a rational decision on Bt corn at this point -- but there will never be if studies that run counter to a particular agenda are rejected out of hand on pretexts that are not applied equally to all studies.

** NOTICE: In accordance with Title 17 U.S.C. Section 107, this material is distributed for research and educational purposes only. **