NOTES. 457 



CHAPTER IX. 



Nate 108, page 284. Zoologists and geologists alike are wont lo re- 

 gard all the land moUusca, or rather their shells, as peculiarly fitted to 

 indicate the affinities and relationship of living and extinct faunas. 

 Now, I do not dispute that they may sometimes be of the greatest 

 utility in this respect, hut I must here express my conviction — a convic- 

 tion derived from years of study of the animals as well as of their 

 shells— that in many cases we have absolutely no right whatever to 

 avail ourselves of the shells of land moUusoa for such comparisons ; and, 

 moreover, that their classification by the shells, which is universally 

 adopted by conchologists and geologists, and which they have accepted 

 as a natural one, is absolutely and totally worthless and unnatural. Thus 

 every argument based on the assumption that the genera and sub- 

 genera as at present distributed are natural divisions, indicating the 

 true affinity of the species they include, must be purely imaginary, a, 

 mere castle in the air (such, for example, as Qeotroclms, Sulimtm, 

 Raehu, Somoms, Hapahis, Nanina, Leueoehroa, &c., &c. ; comp. "Wallace, 

 Geog. Did. Animals, ii. 512 et seq.). 



Nate 109, page 287. The careful investigations which I pursued for 

 years, extending over many hundred species, have brought me more and 

 more to the idea that it may be possible to determine the route of migra- 

 tion followed by many genera of land moUusoa by a diligent examina- 

 tion of their natural affinities. This evidently cannot be done by 

 an examination of the shells exclusively. These, of course, must not be 

 neglected, but their systematic value has hitherto been greatly over- 

 estimated, especially by geologists, and without a close familiarity with 

 the animals themselves we can but very rarely determine the affinities 

 of the species with any certainty. Hence our first task must be to sepa- 

 rate those groups of the land mollusoa whose shells do, in fact, afford a, 

 sure indication of their systematic position from those in which the 

 shell is quite or almost useless for such a, purpose. To what a great 

 degree this is often the case is shown by the Philippine genus Cochlo- 

 styla, of which the shells are so excessively variable — in spite of the 

 similarity of structure in the animals themselves— that no conchologist 

 could possibly describe the genus from the shells. Hitherto we have 

 always had a genus under the name of Vitrina, but species were in- 

 cluded iu it which belong not merely to different genera, but even to 

 different families ; these are so much alike as to the shells that, 

 according to that character alone, it was inevitable that they should 

 get classed together. In my work on the land mollusca I have shown 

 that almost all the shells of the Philippines known as Vitrina belong to 

 the genus Helica/rion and the family Zonitidse, while Vitrina is one of 



