I investigated the ability of predators to influence the patterns of s
pecies richness and abundance of non-piscivorous fishes on small, arti
ficial reefs replenished by natural recruitment. Periodic removal of p
redators effectively reduced the species richness and abundance of pre
dators on removal reefs. The difference between the number of predator
s on control and removal reefs was greatest immediately following the
removal of predators and attenuated between removals. During periods o
f recruitment, species richness and total abundance of recently-recrui
ted, non-piscivorous fishes were generally greater on predator-removal
reefs than on control reefs. Species richness and total abundance of
resident non-piscivorous fishes were not affected by the removal of pr
edators in the first year of the experiment. Both abundance and specie
s richness of residents, however, were greater on the removal reefs du
ring the second year of the experiment. The difference in the response
s of the two age classes to the removal of predators suggests that pre
dators may affect community patterns of older age classes through time
-lagged effects on the survivorship of younger age classes. At the end
of the experiment, species richness was positively related to abundan
ce for recruits and residents. The effects of removing piscivorous fis
hes on the abundance of non-piscivorous fishes were similar for specie
s considered separately. A greater number of species of recruit and re
sident fishes were more abundant on reefs from which predators had bee
n removed. These data suggest that predators can play an important rol
e in structuring communities of fishes on coral reefs.