THE NAUTILUS. 



41 



however, shows that 



while one, C. plana, grows u])on the inside or concave surface. 'The 

 curvature of the surface upon which the Crepi- 

 dulas grow has naturally produced a similar 

 curvature in their own shells as it would -.other- 

 wise be impossible for them to retain their hold. 

 We therefore find that the three species first 

 mentioned are convex in shape, while C. plana 

 is concave. 



With this fact before us the question naturally 

 arises : why is not C. plana a mere form of C. 

 Crepiduia fornicata. Jomicata produccd from eggs of the latter spe- 

 cies which happened to be deposited on the inside of the host shell? 

 A series of specimens collected with their host 

 young fornicata is sometimes found on the inner 

 or concave surface of Natica and Fnlgur and 

 retains all the characters of the species, being 

 easily separated from the young of plana. The 

 fact, however, that none but very small fornicatas 

 are found in such j^ositions, seems to show that 

 the tendency to a convex form makes it impos- 

 sible for them to retain their hold for any length 

 of time on a concave surface. In other words. Crepiduia plana. 

 the tendency towards convex and concave shells has now become a 

 fixed specific character. 



As to the original ancestry of C. plana there does not seem to 

 me any doubt but that either this species has descended from a shell 

 of the C. Jomicata style or better perhaps that both species have 

 developed from a free snail which formed the habit of attaching 

 itself to other shells and whose characters have been gradually 

 altered to suft its habits. Some of the individuals have thus adapted 

 themselves to the convex and some to the concave surfaces of their 

 hosts, and so eventually produced two distinct species. This theory 

 is strengthened by studying the Crepidulas from other parts of the 

 world, for my friend Mr. Pilsbry, tells me that on nearly every coast 

 where a convex Crepiduia is found there is also a concave species 

 corresponding to our C. plana. 



As regards our three convex Crepidulas, C. glauca and C.convexa 

 are undoubtedly much more closely related to each other than either 



