310 The American Naturalist. [April, 
the family. The paleontological history of the Testudinide 
reaches back to the Bridger Eocene. From this formation the 
oldest land tortoise, Hadrianus, has been recorded, which is nearly 
related to Manouria, characterized by the double caudal-shield, as 
in the Emydide. Forms very much like those from the Gala- 
pagos we find in the Miocene. We do not go too far to say that it 
is probable that during the Eocene period, and possibly a little later, 
the Galapagos were still in connection with the continent. 
The important question is, Where was this connection? In 
their general characters the fauna and flora of the Galapagos show 
resemblances to the great Mexican and Sonoran province, and 
also to the West Indies. And it may be that the connection was 
with these regions (and it seems more probable than any other), 
but of course it is quite impossible to bring to-day any posi- 
tive proof for this idea. It would appear that the whole west- 
coast of America has undergone subsidence. We find there a 
great number of islands: Prince Wales, Queen Charlotte, Van- 
couver, Santa Barbara, Guadaloupe, and so on. That all these 
islands have been in connection with the continent at a former 
period seems to be certain. They appear as the result of 
subsidence. The Revilla Gigedo Islands are in the line of this 
sunken district. Farther south we find the small island Clipperton, 
and in a southeastern direction the Galapagos. Between Clipper- 
ton and the Galapagos two islands, Duncan and Galego, have been 
: recorded ; but they are of a doubtful nature,—at least they have 
not been seen again in latter times. But could we not imagine 
that they have disappeared in the course of this and the last 
century by subsidence ? 
Near the Mexican coast we have the Tres Marias Islands. These 
are considered as continental even by Wallace; but the mor 
distant Socorro of the Revilla Gigedos are considered by Wallace 
_ as oceanic. Wallace believes, therefore, in subsidence in regard 
to the Tres Marias, in elevation in regard to the Revilla Gigedos, 
simply because there are no mammals on the barren Revita 
Gigedos, and because they are placed within the thousand-fathom 
line. The fauna of the Revilla Gigedos is typical of that of lowèf 
California and the Sonoran province, and I believe also that the =i 
