We recorded the responses of individual intertidal crabs, Heterozius rotund
ifrons, to stimuli presented singly and in combinations in the laboratory.
Undisturbed crabs did not respond to the introduction of odour from a crush
ed conspecific but did respond strongly to food odour. Undisturbed crabs re
sponded equally to food odour alone and a combination of food and odour fro
m a crushed conspecific. When tactile stimulation was applied, as when the
crab is grasped by a predator, individual H. rotunidfrons assumed a rigid,
appendage;extended posture for several minutes, rests with predatory fish s
howed that this catatonic posture is a very effective predator-defence mech
anism. The duration of the catatonic state was decreased by the addition of
food odour but increased by the addition of alarm odour (crushed conspecif
ic) or the combination of alarm and food odours. Thus, which chemical stimu
lus was dominant was reversed by tactile input (i.e. dominance was continge
nt upon context). The effect of alarm odour on food odour responses lasted
4 h. Visual input in the form of a shadow passing over the crabs, either be
fore or after tactile induction of the catatonic state, also increased the
duration of that state. However, the duration of the catatonic state follow
ing exposure to both cues associated with danger (shadow + alarm odour) was
similar to that of the control level. The crabs appeared to switch strateg
ies when three cues associated with danger (tactile grasping, alarm odour a
nd shadows) were detected, either simultaneously or over a 4-h period. The
results illustrate the highly contingent nature of the behaviour of these c
rabs. (C) 2000 The Association for the Study of Animal Behaviour.