Abstract
Laboratory tank experiments indicated that turbidity levels up to 160 Nephelometric Turbidity Units (NTU) did not reduce the feeding rates of adult smelt ( Retropinna retropinna Richardson), or inanga (Galaxias maculatus Jenyns), on Daphnia sp. Mean feeding rates were higher for inanga than for smelt. However, the lower rates for smelt were related to the greater variation in individual feeding rates. Specifically, a few fish fed intensively while most did not. As the feeding ability of both species was unaffected by turbidity levels up to 160 NTU, they were less sensitive to turbidity than juvenile banded kokopu ( Galaxias fasciatus Gray). Juvenile banded kokopu should therefore be used as the benchmark species for setting turbidity guidelines to protect native fish in rivers where this species is present. Both juvenile and adult rainbow trout ( Oncorhynchus mykiss Richardson) appear to be more sensitive to turbidity than adult inanga and smelt, so should be used as benchmark species in all other waters.