BIRDS 275 



It will be seen from the above that there is in the genus a regular 

 progression toward complete melanism, i. e., toward a form in which 

 both sexes and all ages of the young would be entirely black. Hence, 

 it appears to us that the color may be taken as the most important fac- 

 tor in the evolution of the genus. Almost the only other variable 

 character is the beak. Hence, by plotting with the color variation as 

 the ordinate and the bill variation as the abscissa, we can arrive at an 

 approximate scheme of the relationships of the different subgenera, 

 species and subspecies of the genus. The diagram on page 276 is 

 formed in this way. This diagram is made out from the characters of 

 the males only, except in the upper or Cactornis half of Stage VI, for 

 in all the others the females remain in Stage III. 



The diagram shows that the young of Cactospiza during their 

 growth pass through Stages I and II while the adults never get beyond 

 Stage III. The young males of CainarhyiicJnis traverse Stages I, II, 

 III and IV, arriving by maturity at Stage V. This is their ultimate 

 condition. The males and females of Geospiza begin at Stage II and 

 the male goes through all the stages up to Stage VI, which represents 

 the maximum of blackness attained by any of the Geospizcc. The 

 males of Cactornis are the same when adult as the males of Geospiza^ 

 but both males and females in their growth begin at Stage IV (or go 

 very rapidly through Stages II and III). The female in this group 

 reaches Stage V, or a condition similar to it. Hence, the average 

 amount of blackness in the subgenus Cactornis^ considering all the 

 forms, is greater than in Geospiza^ and for this reason we place it 

 higher in the diagram. This scheme brings G. conirostris coni- 

 rostris at the top of the entire Geospiza series. The position of the 

 species in the diagram to the right or to the left of the main vertical line, 

 indicates the relative slenderness or thickness respectively of the bills. 



The position of the species as given in the diagram certainly repre- 

 sents their degrees of resemblance, but xve do not claim that it cer- 

 tainly represents their natural relationships. We have no way of 

 determining to what extent convergent evolution has operated in caus- 

 ing forms to resemble one another. However, in the discussion of 

 the species and subspecies we have followed the order indicated in the 

 diagram, working in each direction away from the main vertical line. 



Subgenus Cactospiza Ridgway. 



Cactospiza Ridgway, Proc. U. S. Nat. Mus., xix, p. 546, 1896 (Type, Cac- 

 tornis pallida Sclater and Salvin). 



No black on the plumage at any stage. Sexes similar, and the 



young resemble the adults. Color, above brown with dusky centers 



