BIRDS OF NORTH AND MIDDLE AMERICA. 



515 



on the rump; middle and greater wing-coverts margined terminall_y 

 witti a rather more buffy or light brownish hue, but still not approach- 

 ing tawnj^ or rusty; sides of head, chin, and throat dusky or dull 

 i blackish brown, faintly streaked with dull whitish, more distinctly along 

 ! the median line; rest of under parts dull bufpj^ white, immaculate on 

 ' middle of abdomen, elsewhere broadlj' streaked with dusky, the streaks 

 giving way on sides and flanks to a nearl}' uniform light olive; bill 

 wholly cleai' deep cinnamon, paler on lower and terminal portions of 

 mandible; legs and feet brownish black; length (skin), 96.52; wing, 

 71.12; tail, 41.91; culmen, 17,78; maxilla fx'om nostril, 12.70; gon3^s, 

 10.16; basal width of mandible, 9.40; basaldepthof bill, 10.67; tarsus, 

 22.10; middle toe, 15.75. 



Galapagos Archipelago (Charles Island; Indefati<yable Island?). 

 Cactornis hreviroslrin Eidgway, Proc. U. S. Nat. Mus., xii, no. 767, ¥eh. 5, 1890, 

 108, fig. 4 (Charles I., Galapagos Archipelago; collection U. S. Nat. Mus.). 

 Geospiza bremrostris Rid&way, Proc. U. S. Nat. Mus., xix, 1897, 541, pi. 57, fig. 6 



(monogr. ). 

 Geospiza conirostris brevirostris' 'Rotbscuild and Hartert, Novit. Zool., vi, Aug., 

 1899, 159 (Charles I.; crit.). 



Genus COCORNIS Townsend. 



Cocomis^ Townsend, Bull. Mus. Oomp. Zool., xxvii, no. 3, July, 1895, 123. (Type, 

 C. agasdzi Townsend. ) 



Similar to the more slender-billed forms of Geospiza {^''Cactornis''''), 

 but bill proportionallj^ smaller, more slender, and more decidedly 

 decurved. 



coverts and other features which characterize young birds in their first year. It cer- 

 tainly can not be referred to G. fatigata, which has the bill altogether longer and at 

 the same time much narrower in both its vertical and transverse diameters. The size 

 and shape of the bill agree verj' closely with those of G. brevirostris, though, as might 

 be expected from the difference in age or sex, it is not quite so strong. 



Without having seen the type, Messrs. Rothschild and Hartert refer, quite confi- 

 dently, this form to G. conirostris as a subspecies; but in doing so I feel sure they are 

 in error, having carefully reexamined the type and compared it with the smaller 

 specimens of G. conirostris (my G. media) . G. brevirostris is, in fact, far more like 

 G. frcdereula, and there is quite as good reason for considering G. brevirostris related 

 to G. fratercula as to G. conirostm, as the following measurements will show: 



Locality. 



Minimum measurements of G. 



conirostris, male 



^Type of "G. raedia" 



Type of G. brevirostris 



Maximum measurements of G. 



fratercula, male 



Average measurements of G. fra- 



ter^^ula, male 



Wing. 



77.47 

 78.74 

 68. S8 



67.31 



65.79 



C\ilmen 

 Tail. from; 

 base. 



48.26 

 48.26 

 46 99 



43.18 



41.15 



18.54 

 20.32 

 18.29 



17.27 



17.02 



Depth 

 Of bill 

 at base. 



13.97 

 15.75 

 11.43 



Gonys. 



10.41 

 10.67 

 10.16 



9.65 



9.40 



Width 

 of man- 

 dible at 

 base. 



11.43 

 9.40 



10.16 

 9.65 



Tarsus. 



Middle 

 toe. 



22. 86 16. 51 

 22. 86 16. 51 

 20.83 ; 15.75 



20.57 

 20. 07 



14.99 

 14.48 



' "From Cocos -]- opvn = bird." 



