In 1995, ears of a experimental inbred (CG59-2) containing a synthetic Bacillus thuringiensis Cry IA (b) gene driven by PEPC, pith and pollen promoters and artificially infested with Ostrinia nubilalis (Hübner) larvae in small plot studies were free from insect damage, whereas 40-50% of the corresponding non-Bt inbred ears were damaged. Bt inbred ears that were inoculated with Aspergillus flavus Link and Fusarium proliferatum T. Matsushima (Nirenberg) or exposed to natural mold inoculum after infestation with O. nubilalis were free of visible signs of mold, as compared with approximately 30-40% of the non-Bt ears similarly treated. Results in 1996 using the same inbred with a single allele dose of the Bt gene showed similar trends. Mean total fumonisin levels for non-Bt versus Bt inbred ears were not significantly different (2.8 versus 0.8 ppm, respectively) in 1996. In paired hybrid studies run in 0.4-ha (1-acre) fields, an event 176 Bt hybrid had significantly lower amounts of damage and signs of Fusarium spp. mold, but not fumonisin, compared with a corresponding non-Bt hybrid from 1996 to 1998. However, two hybrid pairs that contained either MON810 or Bt11 constructs examined in similar fields at the same site had lower levels of fumonisin in both 1997 (30- to 40-fold) and 1998. High intrafield variability in insect infestation and presence of Helicoverpa zea (Boddie) in Bt hybrids was apparently responsible for fewer significant differences in fumonisin levels in 1998. Similar trends for all three hybrid pairs were noted in small plot trials at another site. Incidence of other ear pests or insect predators varied as much among non-Bt hybrids as they did for Bt/non-Bt hybrid pairs.