218 Annals of the Carnegie Museum. 



less than three geographic races of this species. Average measure- 

 ments are as follows: 



Wing. Tail. Bill. Tarsus. 



Nine males from Mexico 117 41 29 49 



Six males from Central America 115 40 30 51.5 



Ten males from the West Indies 116. 5 40 30 51.5 



Six females from Mexico 133-5 46 31.3 51 



Five females from Central America 131 43.5 32 53.5 



Ten females from the West Indies 132 44 33 55.5 



Sex for sex, Mexican examples are decidedly duller and darker 

 below than those from the West Indies, in which the maroon color of 

 the under parts is much brighter. In the former series the greenish 

 black of the breast merges more gradually into the maroon of the 

 abdomen, which is often overspread with a shade of brown, while in 

 the West Indian birds the transition is more abrupt, and the brown 

 shade lighter or absent. The color-differences are no less marked 

 above, although their character is reversed, for while the upper parts 

 in the Mexican birds are lighter, more rufescent (nearer Hessian brown 

 of Mr. Ridgway's Color Standards and Color Nomenclature), in the 

 West Indian skins they are darker and more purplish (nearer maroon). 

 There is also a decided difference in the size of the frontal lappet in 

 favor of the latter series. 



Taking up now the Central American series, which includes examples 

 from Honduras, Nicaragua, Costa Rica, and Panama, we find them 

 almost exactly intermediate between the Mexican and West Indian 

 birds. With a larger series the slight discrepancy in size between 

 Central American and Mexican birds shown by the above table would 

 doubtless disappear. The frontal lappet, however, certainly averages 

 larger in the former, although not so large as in the West Indian birds. 

 In the color of the upper surface the Central American birds most 

 resemble those from Mexico, while below they are almost as bright 

 as those from the West Indies. In short, if the latter are to be separ- 

 ated at all, as I believe they should be, it will be necessary to recognize 

 three races of this species instead of two. While selected specimens 

 may be very similar, the average collective differences are quite suf- 

 ficient in my judgment to justify subspecific separation, certainly as 

 much so as in some other groups, the Ground Doves for instance. 



Before the question of names for these three forms can be decided 

 it will be necessary to fix the type-locality of Fiilica spinosa Linnseus, 



