To determine the toxicological effects of complex mixtures of pesticides, we obtained data on 100 pesticide residues in common foods of central Italy. Fifteen pesticides were more regularly detected at higher levels (dithiocarbamates, benomyl/carbendazim, thiabendazole, diphenylamine, chlorthalonil, procymidone, fenarimol, chlorpropham, vinchlozolin, methidathion, chlorpyriphos-ethyl, parathion-methyl, parathion, chlorfenviphos, pirimiphos-ethyl). Using itemized data on daily food consumption in Italy, we calculated that the average exposure for an adult subject was 716 micrograms/day, ranging from 148 micrograms of dithiocarbamates to 1 microgram of pirimiphos-ethyl. We made a mixture of these 15 pesticides at concentrations proportional to the ratio determined in foods and tested it with the Salmonella-microsome assay, with and without metabolic activation with PCB-induced rat liver S9. No mutagenic activity was observed at concentrations up to 500 micrograms/plate. We also tested the same mixture at concentrations ranging from 0.1 to 20 micrograms/ml on human lymphocytes in vitro, and observed a slight but statistically significant increase in sister-chromatid exchanges at 1 microgram/ml. We also administered the mixture in corn oil by gavage to Wistar rats at doses of 1, 10, and 100 micrograms/kg. After 24 hr the ratio between bone marrow polychromatic and normochromatic lymphocytes (a sign of cellular toxicity) was decreased by the exposure, but we did not observe a significant increase in the frequency of micronuclei. We conclude that the pesticide mixture did not have appreciable genotoxic activity in the assays used.