654 W. TRELEASE FLORA OF CENTRAL AMERICA 



be carried b}^ water. Their seeds are fairly resistant, thin, and easily 

 blown about by the wind ; but there is no reason to think that this insures 

 dissemination to any great distance. Even on the mainland, as in the 

 classic region of Tehuacan, the species are often narrowly limited geo- 

 graphically. In the West Indies this restriction is accentuated, no doubt 

 as a result of water barriers. 



The few Antillean species of Fwarcea are suggestive. One of them 

 {F. cubensis) which occurs in Cuba and Haiti is very closely related to 

 the Yucatecan cahum {F. cahum).^ The commonest and most wide- 

 spread (F. tuberosa), which is found throughout the chain, is a close 

 relative of the Brazilian species, which has been grown so long in Mauri- 

 tius as to have acquired the name Mauritius hemp (F. gigantea). A 

 third species {F. macro fliyllcu) y which seems to be indigenous to the Ba- 

 hamas, is very like a form of northern South America. So far as these 

 facts are indicative, they suggest immigration from both south and west; 

 the former apparently earlier, if extent of distribution bears any relation 

 to time. 



Agave, which is represented in the West Indies by about 50 indigenous 

 and endemic species,^ presents these in 6 distinct types: the Antillanae, 

 of a dozen species, are confined to the Greater Antilles; the Bahamanse, 

 closely allied to the preceding, and with half as many species, are exclu- 

 sively Bahamian; the Caribseae, with 15 species, are confined to the 

 Caribbees, or Lesser Antilles. These plants are aU large, of the "century 

 plant" or "maguey" type. The southernmost islands also possess a re- 

 duced edition of this type, the Yiviparse, with five species, of which one 

 is peculiar to Trinidad and the adjacent coastwise islands, and a sixth 

 species of the group occurs in the coast region of Venezuela. In the 

 northern islands, also, a smaller type occurs, represented on the G-reater 

 Antilles by five species (Antillares) and in the outlying Bahamas by two 

 v^ry xerophytic species (Inaguenses). The Antillanae, Bahamanse, An- 

 tillares, and Inaguenses of the north are clearly differentiated from the 

 equivalent Caribseae and Viviparae of the south. 



There is no evident reason why a species of either group should not 

 range through the entire chain of islands, like the wide-spread species or 

 group of scarcely segregable species of Furcrcea {F. tuberosa), but they 

 do not do so. The Agave of Saint Thomas {A. missionum) is one of the 

 Antillanae; the Agave of Saint Croix, 30 miles or so away {A. Egger- 

 s'iana) , is one of the Caribaeas. Less striking, but even more suggestive, 

 are the facts that the species of either group are severally localized on a 

 single island or on contiguous islands, and that species of any group 



