DEATH-FEIGNING IN TERRESTRIAL AMPHIPODS. 1 95 



food. The behavior of the terrestrial 0. agilis in relation to solid 

 objects is little modified beyond that of the aquatic species. 

 The thigmotaxis of this form is certainly protective in function, 

 not only by enabling the animal to escape detection by lying 

 quiet, but by leading it into situations such as under stones or 

 into crevices which are inaccessible to its enemies. 



The behavior of Orchestia palustns shows an interesting con- 

 nection between that of Talorchcstia and O. agilis. O. palustris 

 is considerably larger than agilis and is not so active in its move- 

 ments. It is often found in marshes some distance from the sea- 

 shore. It usually endeavors to escape by running away and 

 resorts to hopping only under necessity. The tendency to get 

 under or between objects is as strongly developed in this species 

 as in agilis, and contact has apparently a stronger quieting effect 

 upon it. When lying quiet 0. palustris may be poked about 

 more or less without being aroused from its thigmotactic lethargy. 

 Only rarely, however, can it be picked up without its making 

 efforts to escape, although it is much less responsive than 0. 

 agilis. The conduct of this species is intermediate between the 

 thigmotactic response of agilis and the death-feigning of Talor- 

 c/iestia. Some specimens might almost be said to possess a death- 

 feigning instinct. The curling up of the body and the bending 

 of the antennae are, not carried so far as in Talorchestia, but the 

 same actions are performed which, if carried out in a more de- 

 cided manner and persisted in longer, would result in what would 

 commonly be called feigning death. The death-feigning instinct 

 of Talorchcstia is an instinct which, I believe, has its root in the 

 thigmotactic responses common among other amphipods. One 

 may easily conceive that by the selection generation after gener- 

 ation of those individuals of 0. agilis in which the thigmotaxis 

 is most persistent and in which the body is drawn up in the most 

 compact form during the response a mode of behavior like the 

 death-feigning instinct of Talorchcstia might readily be produced. 



It seems not improbable that an instinct having its phyletic 

 root in a simple thigmotactic response may in course of time 

 come to be comparatively independent of contact stimuli. The 

 persistence of death-feigning in Talorchcstia depends far less 

 upon contact than the thigmotactic reactions of the aquatic Am- 



