PEEIPATUS IN SOUTH AMEEICA 347 



members of the Andean group. This discontinuous distribu- 

 tion is a very noteworthy fact, for it cannot be explained by 

 the supposition that some member of this Andean group may 

 still exist and have been overlooked in the intermediate vast 

 tract of country, because many specimens of Peripatus have 

 been discovered in Costa Bica, Nicaragua and Panama, all 

 belonging to the Caribbean group. Accidental dispersal, such 

 as marine currents, cannot be invoked as being responsible for 

 this distribution. It is due, in my opinion, to a former direct 

 land connection between western Mexico or Guatemala and 

 some part of the west coast of Soutli America. That the 

 mountain system of Guatemala suddenly terminates at the 

 edge of the Pacific, and that it formerly had a westward 

 continuation, has been alluded to, and I have mentioned also 

 several cases of discontinuous distribution that I thought 

 were due to the existence of an ancient land, more or less in- 

 dependent of Central America. The newt Spelerpes is one 

 of these. Its headquarters seem to be in Mexico. A few occur 

 in Guatemala, Costa Eica and Chiriqui. Further south we 

 meet with the genus again in Colombia, Ecuador and northern 

 Peru, but nowhere else in South America. The tortoise 

 Chelydra rosignoni occurs in ^Guatemala. It is absent from 

 the rest of Central America, yet in Ecuador we find an isolated 

 colony. Another tortoise, Geoemyda punctularia, inhabits 

 Guatemala and Mexico. Southward it is only known from 

 Ecuador, Colombia, Venezuela and further east. The gecko- 

 like genus Eublepharus is probably an exceedingly ancient 

 one, its range being most peculiar and strikingly western. 

 One species occurs in California, another in Mexico, still 

 another in Panama, and lastly one in Ecuador. All the re- 

 maining species, which show great resemblance to the 

 American ones, are confined to southern Asia. We probably 

 have to deal in this case with a persistent type which through- 

 out many geological periods has retained the same characters 

 and has died out in the still existing land fragments of the 

 ancient Pacific continent, whence it originally spread east 

 and west after its subsidence. There are numerous other 

 examples, particularly among plants, implying that the land 

 which I described as lying westward of Central America 

 once touched the South American continent, probably near 



