The objective of this study was to generate a comprehensive data set of chemically induced malformations in Drosophila using a detailed morphological examination of the entire fly (phase one). These data were analyzed, in blind, with the goal of developing a standardized set of criteria which could be used in a new, rapid, and economical Drosophila bioassay useful in the preliminary screening for potential developmental toxicants. After 32 chemicals were tested, formalized criteria were developed to form the basis of a new Drosophila bioassay. These criteria were then applied to the data from the same 32 chemicals (phase two). The data from only 18 of these chemicals met all requirements for evaluation, e.g., statistical significance, minimum fly numbers, sufficient challenge concentration administered, etc. In the new bioassay, rather than the detailed and time-consuming examination of the entire fly for a multitude of morphological defects, only two specific anatomical sites are examined. These sites are the humeral bristle and the wing blade, with focus placed on two structural defects--a bent bristle and a notch in the wing. These defects were the only two external malformations among the multitude of defects observed in flies treated in the first phase with the 32 chemicals which demonstrated the following characteristics: 1) A consistent concentration-response in flies treated with a variety of developmental toxicants; 2) a lack of response with most presumptive non-developmental toxicants; and 3) consistently low-background incidences in control flies. In both phases, developing Drosophila were exposed to the test agents from the egg through three larval stages by incorporating a range of concentrations of each chemical into the culture medium. Emerging adults were examined for an array of defects as part of a detailed morphological examination in the first phase, including bent bristles and wing notches. In the second phase, only bent bristle and wing notch data were evaluated. The incidences of bent humeral bristles and wing notches from flies exposed to each of the 18 chemicals were compared with those of concurrent controls. Of the 18 chemicals that could be evaluated using the new bioassay, 13 were known developmental toxicants while the remaining 5 were presumptive negative agents. Ten of the 13 mammalian developmental toxicants were correctly identified with this test (false negative rate of 23%). Four of five apparent non-developmental toxicants were correctly identified for a false positive rate of 20%.(ABSTRACT TRUNCATED AT 400 WORDS)