John Hodgson replies:

An apology is clearly due to Cummings et al. for not having drawn attention in my article to their rebuttal of the critics I cited. In my defense I can only say that my nearly 20 years' experience of science publishing had not prepared me to expect a rebuttal to unpublished comments on an unpublished manuscript.

The criticism of my article and their critics is largely misdirected and their choice of supporting data (other people's) questionable.

They point, for instance, to a range of “potential hazards” of the structural instability of constructs containing CaMV 35S promoter sequences. Their choice of key published data in support of this proposition is the work of Stanley Ewen and Arpad Pusztai on the intestinal effects of potato diets on rats. Whatever its scientific merits (and those have been widely discussed), that paper does not claim that “a significant part of the effects may be due to” the construct or the transformation but merely that “other parts of the GM construct, or the transformation, could have contributed to the overall effects.” In any case, identifying potential hazards is of little use unless the relative magnitude of either the potential or of the hazard is described. Exposure to sunlight after all, even in Milton Keynes or Ontario, will elicit many if not all of the same “potential hazards.” A more proper comparison, perhaps, would be with the risks from the ingestion of plants produced by “conventional breeding” (i.e., through random mutagenesis, crossing, and selection).

The other key point made—that naked transgenic DNA containing CaMV 35S sequence might be harmful to humans—is also not well supported by their choice of reference. The work of Rekvig et al. that they cite concerns the inoculation of rabbits with naked BK virus, a human polyomavirus. How does this evidence of an experimental infection of one mammalian species by an integrating virus from another mammalian species support a thesis that, in essence, seems to be that laboratory-created constructs including promoter regions from a plant virus that does not integrate naturally into host genomes will (1) infect humans (2) integrate into human genomes, and (3) cause “potential hazards.”

The final paragraph of Cummings and Ho's letter linking their arguments to the Biosafety Protocol could be considered revealing. A cynic might posit that the reason their review paper was publicized in December 1999 and January 2000—even though it was not actually published then—was to try to influence decision makers involved in finalizing global rules on the transborder shipment of living modified organisms. However, such a supposition, unsupported as it is by any real data, would require such a convoluted and contorted train of logic as to be utterly unbelievable.