1. In a southern Californian stream naiads of a lestid damselfly (Arch
ilestes grandis Rambur) were much less abundant, moved less, exhibited
fewer conspicuous behaviours, were more likely to occur in refuge are
as, and had different diets in pools containing versus pools lacking r
ainbow trout (Oncorhynchus mykiss Richardson). To determine if trout w
ere responsible for these patterns, we removed trout from some stream
pools and added them to pools lacking trout, with unmanipulated trout
and troutless pools acting as controls. 2. The abundance and emergence
of A. grandis were drastically reduced, and the proportion of lestid
populations in refuge areas greatly increased, when trout were added t
o pools; however, the removal of trout had less drastic effects on les
tid abundance and distribution. Aeshna walkeri (Kennedy) was also more
abundant in pools lacking than in pools containing trout. 3. Trout ma
nipulations affected lestid behaviour, with swimming being observed on
ly in troutless pools and movement tending to be greater in pools lack
ing rather than containing trout. 4. One week after manipulations star
ted, the number of prey items per lestid gut was higher in troutless c
ontrol than in trout addition pools. Ostracods and chironomids were mo
re abundant, and mayflies were less abundant, in the diets of lestids
from pools lacking versus containing trout. Comparisons of the environ
mental abundances of prey taxa and lestid diet composition indicated t
hat lestid selectivities for Caenis were higher, and those for Paralep
tophlebia, ostracods, and Eubrianax lower, in trout than in troutless
pools. 5. Although similar at the beginning of manipulations, head wid
ths of lestids in troutless control pools were greater than those in t
rout addition pools after 3 weeks.