BIRDS 



343 



dividuals of the last, i. e., those living on Bindloe, and the fact that 

 the melanistic tendency in the adult females and young in the Bindloe 

 form of G. scandens approaches nearest in degree to that of G. con- 

 irostris, indicates a natural transition from G. scandens to G. con- 

 irostris. The few facts that might be taken as evidence of a relation- 

 ship between G. c. propinqua and G. seftentrionalis have already 

 been given under the latter species. 



Since, in the general evolution of the groups Cactospiza, Cama- 

 rhynchus, Geospiza and Cactornis, the color of the plumage is seen 

 to be a much more constant character than the size and shape of the 

 bill, and a characteristic, in all other cases, of apparently natural 

 groups, we see no reason why it should not be relied on in the case of 

 G. conirostris, and be made the basis for including this species in the 

 group Cactornis. Taking the color as the more fundamental charac- 

 ter in the classification of all these groups, and the shape of the bill 

 as a secondary one, then G. conirostris conirostris must be regarded 

 as the most specialized of all the Geospizce, and be placed at the top 

 of the series, a position that has by all previous writers in the genus 

 been given to G. magnirostris . 



70c. GEOSPIZA CONIROSTRIS PROPINQUA (Ridgway). 



Geospiza propinqua Ridgway, Proc. U. S. Nat. Mus., xvii, p. 361, 1894 

 (Tower Island), and xiv, p. 543, 1896 ; Bull. U. S. Nat. Mus., 50, Pt. i, 

 p. 499, 1901. 



Geospiza conirostris propingua'RoTnscm'Li) and Hartert, Novit. Zool., vi, p., 



159, 1899. 



Geospiza conirostris subsp. ? Rothschild and Hartert, Novit. Zool., vi, p. 



160, 1899 (Culpepper Island). 



Range. — Tower and Culpepper. «^ 



We have seven adult males from Tower taken in June, one adult 

 male from Culpepper taken in December, and two young males and 

 two young females from Tower. We apparently have no adult females. 



The shape of the bill in this subspecies is very similar to that of G. 

 c. conirostris. Between the bills of some specimens from Tower and 

 Hood there is absolutely no difference, but the bill of the Hood 

 Island variety averages larger. The bill of G, c. propinqua generally 

 has a more curved culmen and less acute tip. 



The collection contains eight adult males. Two of them have not 

 quite reached the purely black phase of Stage V, having a few narrow 

 whitish edgings to the feathers of the lower part of the abdomen and 

 a few brownish feathers on the back. Two specimens have a slight 

 tinge of chestnut on the under tail coverts. Most of these males are 



