l6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 65 



the absence of Peripatus from these islands, may be thus accounted 

 for. 



Grenada lies not upon the ridge supporting Trinidad and Tobago, 

 but upon the ridge supporting St. Vincent, St. Lucia, and the islands 

 beyond. There is no evidence that it ever was connected with Trini- 

 dad or with Tobago. Certain elements in the fauna of Grenada, such 

 as Epiperipatus among the onychophores, and very many types among 

 the other groups of organisms, recall the fauna of Tobago and Trini- 

 dad, and separate Grenada sharply from the islands to the north. I 

 believe that the island of Grenada, including the Grenadines to the 

 northward as far as Bequia, first became separated from St. Vincent 

 by the formation of a deep channel between them, and at a consider- 

 ably later epoch, after the fauna of Grenada had become further 

 modified by additions direct from South America (and not by way of 

 Trinidad and Tobago), it became separated from South America, to 

 which it had been joined in the general region of Margarita Island. 



The fauna of Barbados (including, so far as we know, no ony- 

 chophores) is the fauna of an oceanic island purely, being composed 

 entirely of representatives of the most widely ranging and most easily 

 transported of the organisms of the adjacent islands. Barbados has 

 been entirely submerged since it formed a part of the ancient Antil- 

 lean land. 



No onychophores have ever been found in Cuba, though they have 

 been diligently sought for there by a number of competent naturalists. 

 If any are ever discovered it will be interesting to see whether they 

 will belong to the subgenus Peripatus, like those on the other islands, 

 or to Epiperipatus, like those on Grenada and Tobago, or to both 

 Epiperipatus and Macro peripatus, like those on Trinidad. 



The uniformity of the onychophores throughout the West India 

 archipelago, both in the Greater Antilles and in the Lesser, is of much 

 interest in indicating the original and fundamental unity of the 

 entire group of islands. They do not indicate a zoogeographical 

 division into a Greater and a Lesser Antillean fauna for the reason 

 that their genera are uniformly distributed both in South and Central 

 America, so that the same faunal elements would enter either group 

 of islands in the event of a continental connection. 



The close faunal affinity between the Antilles and the mountain 

 region of western Venezuela, Colombia, Panama, and Costa Rica 

 indicated by the species of the subgenus Peripatus is not a true faunal 

 affinity. It merely shows that in the Antilles and in the mountain 

 region Peripatus has in exactly the same way been protected by bar- 

 riers which have prevented the intrusion of the more efficient com- 



