In the laboratory and in cages in the greenhouse, we evaluated the toxicity of two insecticides (lambda-cyhalothrin and spinosad) on the parasitoid, Diadegma insulare (Cresson), and the predator, Coleomegilla maculate (DeGeer), both natural enemies of the diamondback moth, Plutella xylostella (L.). Lambda-cyhalothrin was very toxic to both natural enemies. Spinosad was less toxic to C. maculata adults and larvae, and slightly toxic to D. insulare. Both natural enemies suppressed P. xylostella populations in cages with 80% spinosad-treated and 20% nontreated plants; such suppression was not seen when lambda-cyhalothrin was used. Using broccoli, Brassica oleracea L. variety italica, a common host for P. xylostella, we also studied direct and indirect effects of both natural enemies in the presence and absence of the two insecticides and to different P. xylostella genotypes: resistant to the insecticide, susceptible, or heterozygous. Neither natural enemy could distinguish host genotype if P. xylostella were feeding on nontreated plants. They could also not distinguish between larvae feeding on spinosad-treated plants and nontreated plants, but D. insulare could distinguish between larvae feeding on lambda-cyhalothrin treated and nontreated plants. Our studies suggest that lambda-cyhalothrin has direct toxicity to these two natural enemies, can affect their host foraging and acceptance of P. xylostella and consequently would not be compatible in conserving these natural enemies in a program for suppression of P. xylostella. In contrast, our studies suggest that treatment with spinosad has much less effect on these natural enemies and would allow them to help suppress populations of P. xylostella. These findings are discussed in relation to the evolution of insecticide resistance and suppression of the pest populations.