1908-9. ] THE GEOLOGICAL CONNEXIONS OF THE CARIBEAN REGION. 375 
America while the eocene flora retreated South and East, and is now 
represented by the living indo-australian flora. In dealing with these ques- 
tions these facts must ever be borne in mind ; but whatever may be the 
case as regards the connexion referred to, it in no way invalidates the 
evidence for a more southerly connexion between North Africa and 
South America. The Flora and the Fauna of North America and those 
of the Caribean Region are so distinct and have so few elements in 
common that we cannot suppose any extensive migration between them, 
while the alliances between the Faunas of North Africa and South 
America are clear and unmistakable. 
The contributions of S. P. Woodward to the question should not be 
overlooked. In his “Manual of Mollusca” (1856, page 403), he says 
“the presence of several species of the old-world Genera, Clausilia and 
Streptaxis, both wanting in North America, becomes a significant fact 
when taken in connexion with the affinities of the higher animals of 
South America and Africa. These imply a landway across the Atlantic 
at some very remote period more direct than would be afforded by the 
Continent which is believed to have united the boreal regions at the 
close of the miocene age.” Further he refers (page 404) to the resem- 
blance between certain land shells found in the Eocene of Europe to 
living shells of South America. Subsequent writers have also dealt with 
this subject. 
Bland (Ann. Lyc. N.H., New York, 1869, page 241) notices the 
observation of Julien (Ann. Lyc., N.H., New York, 1866, page 251) that 
Tortoises occur at Sombrero allied to those of the Galapagos. And in 
the same place will be found a remark of Cope’s that the discovery in 
Anguila of a large extinct Rodent indicates that the Caribean Continent 
had not been submerged prior to the close of the post-pliocene, while a 
wide strait separated it from the shores of North America. If our 
present views are correct this Caribean Continent must have been the 
Atlantis ; for the Caribean Sea instead of being a Continent was a Gulf 
of the Pacific, And the inference as to the separation of this Atlantis 
from North America is no doubt correct; but the date of the sub- 
mergence is more doubtful. 
Andrews (Descriptive Catalog of the Tertiary Vertebrata of the 
Fayum, 190%) is inclined to regard the evidence as lending additional 
support to the theory that in mesozoic times Africa and South America 
were connected by land, and that “the final separation of the two 
continents did not take place until eocene times.” 
