The effects of very low concentrations of the insecticide carbaryl on a corn (Zea mays) monoculture ecosystem and an old-field ecosystem were studied in Ithaca, New York. Four sprayings at "='3-wk intervals were applied and the effects on above-and belowground arthropod faunas and on rate of litter decomposition were investigated. The specific questions addressed were: ( l) are the effects oflow-concentration insecticide stress different in kind or just quantity from those reported in previous studies conducted with higher concentrations? (2) is the complex old-field ecosystem more resistant to these effects than the simple corn monoculture?The insecticide caused a much greater reduction in total arthropod numbers on plants in the old field than in the corn, and this could be accounted for mainly by the effect on herbivores. Despite continued insecticide application throughout the season, the effect of spray on this herbivore reduction decreased over time in the old field. Spraying significantly increased the (predators + parasites)/ herbivore ratio; thus herbivores were disproportionately more affected than their natural enemies by the insecticide. These results are in direct contrast with the previously reported studies using carbaryl at lO times the concentration. Insecticide reduced the soil arthropod fauna much less than it did the arthropods on plants. There was no significant statistical effect of spray, field type, or date on total arthropod fauna on all groups aggregated. The only important effect of the spray on the soil arthropods was an outbreak of scavenging oribatid mites in the sprayed plots. Consistent with the above finding, rate of litter decomposition slowed significantly, but only temporarily, in the sprayed plots. There were no effects of plot type.Thus, low concentrations of carbaryl that might result from drift into fields adjacent to sprayed areas appear to have important ecological effects unpredictable from a knowledge of the effects of high concentrations. The two most important findings here, that the insects in old fields were more sensitive to the insecticide and that spraying increased the enemies/herbivore ratio, suggest that qualitatively different ecological interactions may be occurring in response to low levels of insecticides. Although pesticide droplet size and density were controlled in the present study, under natural drift conditions these factors might further complicate the already unpredictable effects of pesticide concentration per se. The ecological effects of pesticide drift will be most important in areas dominated by small farmers and in areas where airplanes are used to apply chemicals. Because unsprayed crop and second-growth areas play an important role as refugia for both natural enemies and crop pests, further research is warranted investigating the effect of low pesticide dosages on complex ecological responses.