EFFECTS OF BACILLUS-THURINGIENSIS ISRAELENSIS (BTI) AND METHOPRENE ONNONTARGET MACROINVERTEBRATES IN MINNESOTA WETLANDS

We studied the effects of the mosquito larvicides methoprene and Bacil lus thuringiensis israelensis (Bti) on the benthic macroinvertebrate c ommunities of 27 wetland ecosystems in Wright County, Minnesota. These larvicides are generally considered safe for nontarget species. After 3 yr of preliminary investigations, including 2 yr of intensive sampl ing, larvicide treatments were applied during 1991-1993. Nine of the w etlands were experimentally treated with methoprene, which disrupts in sect development; an additional set of nine wetlands were treated with Bti, a microbial larvicide; and nine wetlands were left untreated to serve as a control treatment. In general, insecticide treatment had mi nimal effects on nontarget groups during the first treatment year. How ever, during 1992, highly significant reductions due to both methopren e and Bti were observed in several insect groups. Predatory insects we re reduced on methoprene-treated sites but not Bti-treated sites in 19 92. In 1993, treatment with both larvicides resulted in wetland commun ities that were depauperate in most insects. Although effects were obs erved broadly across insect taxa, Diptera, which comprised 79% of the insects, were affected most strongly, especially the dipteran suborder Nematocera, which included 71% of total insects and was dominated by Chironomidae. Minimal effects on noninsect macroinvertebrates were obs erved. Bti- and methoprene-treated sites also showed a reduction in ri chness of insect genera and an increased tendency to be dominated by o ne or a few genera. Pretreatment data, collected under drought conditi ons from the same wetlands, showed that the benthic macroinvertebrate fauna was dominated by mollusks during the drought but became increasi ngly dominated by insects during the wetter years. On the treated site s, insects remained at low density, very similar to the drought condit ions, but the noninsect macroinvertebrates declined on treated sites i n the same pattern as on control sites. Both indirect effects and dire ct toxicity likely contributed to the observed differences. Bti is lik ely to be directly toxic only to nematoceran Diptera; thus effects of Bti on other insect groups may have resulted from disruption of the in vertebrate food web. Methoprene is more broadly toxic; thus observed m ethoprene effects on nonnematoceran groups may have been due to either direct toxicity or food web effects, or both. The 2-3 yr lag time in response of nontarget insects to larvicide treatment demonstrates the need for long-term studies in wetland ecosystems, and the need to reco nsider the conclusions based on previous short-term studies that these larvicides are environmentally safe.