Memory of information gained during shell investigation by the hermit crab, Pagurus bernhardus

Hermit crabs have the opportunity to assess shells by sight, initial contact and active investigation with their appendages, and it might be expected that information thus gained would be remembered. This possibility was tested by offering crabs shells with partially blocked apertures so that the subsequent rejection could be timed. Crabs appeared unable to distinguish between a previously rejected shell and a novel shell by sight or contact alone but they actively investigated a previously rejected shell for less time than they did a novel shell. The data indicate that it is the shell itself that is recognized, not the location of the shell. The longer the interval between two investigations of the same shell the less was the reduction in the duration of active investigation. These data are discussed with reference to optimal mechanisms for shell acquisition.