BARBOUR AND NOBLE: LIZARDS OF THE GENUS AMEIVA. 419 



p. 562) and Barbour (Mem. M. C. Z., 1914, 54, p. 326) have suggested 

 that the Antillean Ameivas were derived from a centre of dispersal 

 in northeastern South America, and that they have spread thence 

 northward over a continuous land area to the Greater Antilles. Pro- 

 ceeding northward along the islands we find species which show a 

 gradual transition in morphological characters, and there is no obvious 

 break in the series, except where the evidence is wanting, as for ex- 

 ample where the species on Dominica seems very different from that 

 of St. Vincent, we must remember that the form which formerly 

 inhabited Martinique is undescribed zoologically and is probably now 

 extirpated by the mongoose. This gradual transition, as we have 

 said, points to a land migration and not to distribution by flotation. 

 The latter means would not account for the presence of the genus 

 upon so many islands, without presupposing an enormous amount of 

 rafting. Such a constant flotation would have kept new immigrants 

 coming to the islands already populated, as well as to those as yet with- 

 out Ameivas, and would surely have tended to keep the whole Antil- 

 lean group of individuals more homogeneous than they are. There 

 is no real reason for supposing that there was more carriage in the 

 past than at present. Then the derivation would probably have been 

 from several stocks, whereas the Lesser Antillean Ameivas are all 

 derived from the Amciva ameiva stock, the Antillean and mainland 

 races having probably had a common origin from an ancestral wide- 

 spread stock which became differentiated as the stations occupied 

 became separated. The comparatively fixed characters observed 

 among the individuals of the island races stand at sharp contrast to 

 the great variability of the same characters in the mainland races, 

 and this points to a long complete isolation. Interchange of indi- 

 viduals between the islands is unthinkable on any basis, as their 

 physical geographic characters make the setting free of rafts impossible. 

 By the flotsam theory individuals must have reached all islands by 

 rafts directly from mainland rivers. 



Gadow (P. Z. S., 1906, p. 277-375) has shown that the closely 

 related genus Cnemidophorus is composed of species having remark- 

 ably variable characters and that it is necessary to consider the sum 

 of the distinguishing features when comparing two forms. Similarly 

 in Ameiva too great stress cannot be laid upon a single character 

 within a species, especially upon the mainland. This variability may 

 make two species, probably but distantly related, appear closely 

 similar. Some of these curiously close resemblances between widely 

 separated forms may be mentioned, as they are interesting from an 



