The importance of chemical cues in predator-prey interactions has rece
ntly received increasing attention from ecologists. The sources of che
micals to which prey species respond often originate as cues released
by the predator (reviewed by Kats and Dill, this issue). Alternatively
, cues may be released by other prey animals when they detect or are a
ttacked by a predator. Such cues, known as chemical alarm signals, are
particularly common in aquatic systems. These signals provide the bas
is of our current review. Short-term behavioural responses of prey ani
mals to alarm signals have received the most attention. Behavioural re
sponses of prey resemble those exhibited to known predators, and are t
herefore likely to make receivers less vulnerable to predation. More r
ecently, studies have shown that benefits to alarm signal receivers ex
tend beyond the immediate behavioural response of nearby conspecifics
over a few minutes. For example, alarm signals are important in mediat
ing the learning of unknown predators and dangerous habitats. Furtherm
ore, alarm signals have been implicated in induced morphological defen
ces, and defences that involve shifts in life history characteristics.
In some systems, predators may have ways to counteract alarm signals
used by their prey. We suggest that researchers must consider alarm si
gnalling and associated learning phenomena when formulating hypotheses
and designing future experiments.