A. Agassiz—Instinct ? in Hermit Crabs. 291 
It is of course very difficult to apply to Invertebrates many 
of the laws of natural selection, and thus far we know so little 
We can, therefore, only explain the faculty of performing this 
act as inherited, or else as a simple mechanical act rendered 
necessary by the conditions of the young hermit cra 
latter seems the more probable case from the nature of the test 
stage and the one preceding it is the curling of the abdomen; 
in the anterior pa 1 
seemed to satisfy him as well as a shell, there being several 
empty shells at bi 
taking the animal out and occupying his place; all acts which 
seem to require considerable intell 
able forethought ?. 
Newport, Aug. 23, 1875. 
