Interrelations between zooplanktivorous fish and zooplankton were examined at the Poor Knights Islands 20 km off the east coast of Northland, New Zealand from 1980 to 1983. The pomacentrid Chromis dispilus was the most abundant planktivore at all locations; high densities of other planktivores were also found. The rankings of these species varied considerably among locations. These differences may have been due in part to Caprodon longimanus (Serranidae), Scorpis violaceus (Kyphosidae), and Decapterus koheru (Carangidae) malung forays for food away from the immedate vicinity of rocky reefs. The hypothesis that fish have a localised effect on zooplankton was investigated in detail within a small reef area (-2500 m') on 7 separate occasions. Distribution patterns of planktivorous fish changed according to current direction. Fish were always most abundant on the incurrent side of the reef and within an archway during the day. Large differences in densities of zooplankton were detected along a 200 m transect where samples were taken upcurrent, within, and downcurrent of the archway during the day. Lowest zooplankton densities were usually found in the archway where planktivorous fish were abundant. At night when fish were absent from the water column, there was a trend for highest abundances of plankton within the arch, relative to upcurrent and downcurrent sites. A similar 200 m transect parallel to the arch, but 1 km offshore where planktivorous fish were absent, showed no significant differences in density of plankton along its length during the day or night. Zooplankters that showed greatest reductions in density in the vicinity of feeding fish were most abundant in the guts of C. dispilus sampled from the arch. Estimates of removal rates of zooplankters by fish based on concurrent estimates of fish densities, plankton, feeding rates, diet of fish and current speeds suggest that fish were capable of causing the measured reductions in zooplankton concentrations through the archway. The diets of C. dispilus changed among sampling occasions and corresponded partly to changes in the composition of zooplankton captured in nets. Accordingly, any effect that fish have on zooplankton may change between times. Ecological consequences of predation by fish are discussed in terms of zooplankton communities, fish, and rocky reef environments.