A.I Z Snodgrass, Geosfiza, Cocoinis^ and Ccrf/it'dia. \ o^^ 



7iis. It differs from the latter in about the same way that the 

 Cocornis skull differs structurally from the skull of G. scandcns. 

 That is, the upper mandible is slenderer and less deflexed. The 

 angle of deflexure in C. o. luteola (and the bill does not vary 

 in the genus) is about 35°. This is just as much smaller than 

 the deflexure in Cocornis as the latter is than the deflexure in 

 G. sca7idcns. 



From the foregoing descriptions it is evident that the species of 

 the three genera under consideration can be arranged in a graded 

 series according to the structure of the skull. Such a series 

 would begin with the members of Certhidia, all of which have very 

 slender and gently deflexed bills. Following Certhidia comes 

 Cocornis with a slender but more deflexed bill. Separated from 

 Cocornis by a step no greater than that from Certhidia to Cocornis 

 comes Geospiza scandens. This species, although structurally 

 intermediate between those on each side of it, makes a digression 

 as to size, being much larger than either Cocornis or G. fuliginosa 

 acutirostris which otherwise follows G. scandens in the series. 

 From G. f. acutirostris the series is uninterrupted to such forms 

 as G. stremia and G. tnagnirostris in which the bill is enormously 

 large and the upper mandible greatly deflexed, and in which the 

 skull has an almost unbird-like appearance on account of the curi- 

 ous shape and the great development of the crests. 



All that the writer here intends is simply to call attention to the 

 fact that there is a gradation in the skull characters of these three 

 genera, progressing by almost equal steps from one extreme to 

 the other. If any phylogenic theory can be based on this fact 

 then the classification of the three genera accepted at present can- 

 not be correct, for Certhidia is regarded as a member of the 

 Mniotiltidffi and Geospiza and Cocornis are placed in the Frin- 

 gillidae. The GeospizcE as birds have certainly a most Fringillid 

 appearance. The same, however, cannot be so positively asserted 

 concerning the skull of even the least modified species. 



A study of the bills of nestlings would probably have little 

 phylogenetic value. Three stages in the growth of G. fuliginosa 

 parvula are shown in Plate XVIII, figures 23, 24 and 25. 



