TKiri) LIZATtDS OF THE GENUS CXEMIDOI'IIOKUS 



225 



iiiL'iit of the leiuk'iicy to possess u tiiicly retieiilatcd or iiiu'Iy .spoUecl 

 pattern. All of these species, but particularly hacatus and catalin- 

 ensis, are separated from the mainland of Lower California and 

 Sonera by greater distances and depths than are ceralhensls and 

 celer^'ipes^ and, as might be expected, the variation from tessellatus is 

 on the whole more striking. 



The subspecies, canus of Sal Si l*uedes Island and North San 

 Lorenzo Island and martyrls of San Pedro Martir Island, intergrade 

 with tessellatus throngh the whiptail on Smith Island, to the north 

 of them, and with each other through the form on the geographically 

 intermediate South San Lorenzo Island. The islands on which these 

 p()])ulati<)iis occui' are in a dii'cct line with each other and were no 



HLMlyxis 



ceraltensii 



tessellatus 



Ances\r-a\ 



tessellaius 



stock 



. Tnax\mus 



Ancestral S^ock of the Tesse\l3.iu3 Group 



FlGUIiE 31. DiAliUAM UF THE SUPPOSED RELATIO.NSIIIl'S WITUIX THE TESSELLATUS 



GROUP 



doubt connected in the post-Miocene (as suggested by Gadow, 1905?>, 

 p. 236, and others) and probably the union continued through the 

 Pliocene and perhaps even into the Pleistocene. The pattern of 

 maHiji-is and canus ditfei-s from that found in a minority of speci- 

 mens of tessellatus only in the iiner degree of its reticulation. This 

 and the existence of the intergrades, mentioned above, makes it obvi- 

 ous that both martyrls and canus have been directly derived from 

 tessellatus. 



The dwarf sj)ecies, hacatus and catalinensis, are not known to inter- 

 grade with each other or with either martyris or tessellatus. Although 

 hacatus is found on San Pedro Nolasco Island, near the coast of 

 Sonora, and the latter occurs on Santa Catalina Island, near the 

 Lower Californian mainland, the two forms resemble each other 



