A number of assumptions are inherent in the process of using animal toxicity tests to determine water quality characteristics that can be considered safe for aquatic life. These assumptions, expressed as hypotheses, include: (1) the transfer of laboratory acute toxicity tests to a field situation is possible without serious distortion; (2) multispecies or community level responses to toxicity exposure will not be markedly different from single-species responses; and (3) intermittent exposures with short time intervals between events will produce the same ecological responses as a continuous exposure of the same concentration-time integral. These hypotheses were tested in two recycling stream ecosystems at the Monticello Ecological Research Station of EPA's Duluth, Minnesota Environmental Research Laboratory.