cotton without restrictions. In contrast, EPA placed geographic restrictions on the planting of Bt cotton until additional information could be provided to adequately assess the potential for and consequences of transfer of the Bt gene to related species. EPA was focusing on overall environmental impact, whereas the USDA conclusions were related to plant pest issues.

The agencies indicated that they did not communicate with one another on this issue before making their regulatory determinations. However, USDA issued its determination of nonregulated status in June 1995 and EPA registered Bt cotton four months later in October 1995. It appears that the agencies were reviewing Bt cotton during a similar period, so interagency discussions presumably could have been held. The committee recognizes that science-based decisions can depend on an agencies regulatory perspective and that decisions based on the same information can differ. For example, USDA's FPPA determinations are driven by concern about plant pest risk and crop protection, whereas EPA's FIFRA determinations hinge on the potential for adverse impacts on nontarget species and environmental protection in a general sense. In the case of Bt cotton, differentdeterminations concerning the need for geographic limits appear to have been based on somewhat different regulatory end points and levels of comfort with the available information. This may have resulted in stakeholder confusion and raised questions about the credibility of assessments.

The foregoing example emphasizes the need for agencies to avoid inadvertent duplication or the appearance of inconsistency in decisions by increasing their coordination in developing guidance in subjects of common interest and maintaining communication on data needs that are believed to be mutually exclusive. To enhance coordination the committee recommends that

EPA, USDA, and FDA should develop a memorandum of understanding for transgenic pest- protected plant products that provides guidance to identify the regulatory issues that are the purview of each respective agency (for example, ecological risks and tolerance assessment for EPA, plant pest risks for USDA, and dietary safety of whole foods for FDA); identifies the regulatory issues for which more than one agency has responsibility (for example, gene transfer for EPA and USDA and food allergens for EPA and FDA); and establishes a process to ensure appropriate and timely exchange of information between agencies.

If differences in regulatory findings remain after agency consultations, they should be adequately explained to ensure that regulatory decisions are not in conflict and do not have the appearance of conflict. Agencies should consider using Federal Register notices, EPA pesticide fact