Systematic design of replacement chemicals with reduced toxicities will require knowledge of mechanisms of action of parent compounds, especially in species which occupy the environment of most likely exposure. For aquatic systems, the rainbow trout has proven a valuable model for studying mechanisms of carcinogenicity. By comparison, small aquarium species show great potential as in situ field monitors of aquatic contamination by toxic chemicals but are less developed for mechanism studies. Fish species, especially rainbow trout, have also proven useful alternatives to traditional rodent models for comparative studies on mechanisms of action of nonaquatic carcinogens. These kinds of comparative studies form an essential basis for extrapolation of animal studies to man. Carcinogenicity testing of individual compounds and their replacements can provide only limited information on the expected impact of such chemicals on natural populations, since these populations are unavoidably exposed to potent modulators of the carcinogenic response. Hence any program which aims at redesign of commercial chemicals with reduced toxicities must have as a prior aim the full understanding of the mechanisms of joint carcinogen-inhibitor-promotor interactions. Because of their high sensitivity, low cost per individual, and low background tumor incidences, fish models such as the rainbow trout may be the only vertebrate models in which it is economically practical to initiate such complex studies.