148 BARBOUR: ZOOGEOGRAPHY. 



ported across the open sea fortuitously, or by what has been called flotsam and 

 jetsam dispersal, is much less than was previously supposed. I have discussed 

 this question at considerable length in "Notes on the herpetology of Jamaica" 

 (Bull. M. C. Z., 1910, 52, p. 280-284). To attempt to trace the history of this 

 change of opinion would be a long task, and would have no place in this connec- 

 tion. Many naturalists still adhere to the old views, and there can be no doubt 

 that in certain cases fortuitous distribution does play some part in providing 

 continental islands with faunae. Generally speaking, however, it is a negligible 

 one ; and the tendency is stronger and stronger to conclude that the dry surface 

 of the earth has undergone enormous and fundamental changes in extent since 

 life has been existent. Thus, the Galapagos Islands, considered by Darwin, 

 Wallace, and others as typical oceanic islands, are in all probability truly con- 

 tinental in nature; at least this opinion is gaining ground (c/. G. Baur, Amer. 

 nat., 1891, 35, p. 217-229, 307-326; also Amer. nat., 1897, 31, p. 661-G80). 



So also with the Azores, Madeira, and the Canary Islands. 



The marvellous land-snail fauna of some of the Pacific Islands, especially 

 of the Hawaiian Islands, is no longer considered as consisting on!}' of individuals 

 which reached the island fortuitously; but rather as an enormous group of 

 species, greatly modified and divided up amongst themselves through long 

 isolation upon the different islands, or even in the different valleys of the various 

 islands; and, moreover, these species are shown to be the descendants of ancient 

 types. If individuals in the past occasionally reached the islands by floating, 

 they would still do so ■ once in a while. An influx of continental indi^•iduals 

 would tend, of course, to keep the groups of individuals upon the islands more 

 or less like those upon the shores whence the immigrants came. That this 

 is not the case has been most ably argued by Pilsbry (Proc. Acad. nat. sci. Phila., 

 1900, p. 568-581), who shows that, in the first place, many genera of land-snails 

 reach back to tlie Oligocene unchanged in specific characters; and that the mod- 

 ern family groups of snails diverged far back in Mesozoic time. This shows the 

 difference in the conditions which one finds in dealing with the fauna composed 

 of invertebrate groups of animals alone. A land-mass cut off in early Tertiary 

 times might lack most mammals, and yet present moUuscan genera identical 

 with those upon other land-masses, or upon the continents. This, in brief, is 

 the basis of the argument which Pilsbry employs in commenting upon the antique 

 character of the Polynesian snail fauna. After mentioning the primitive char- 

 acter of the Orthurethra, Heterurethra, and the Sigmurethra (Aulacopoda), 

 he shows their astonishing abundance in Polynesia, along with the absence of 



