( 160 ) 



Itidgwav. We can tlieretore only 

 The type is one of the smallest ol 

 ill I onr adnlt males. 



Five adnlt males from the new collection measure as follows : — 



sliajte of the beak so aMy puiiited out by Jlr 

 consider it to be a subspecies of G. conirostris. 



CulmeD. 



22 mm. 



23 „ 

 19 „ 

 22 „ 



7. Geospiza conirostris subsp. ? 



We have three j'onng specimens of a Gcospiza from f'nlpepper Island which 

 we believe to be a form of G. conirostris, from which they differ in not being qnite 

 so dark, and the light margins to the feathers being more bnffy, and the beaks 

 appear to be a little larger, the back paler, the wing-coverts rather broadly margined 

 witli rnsty rnfons. 



We cannot come to any definite decision without examining adnlt specimens. 



8. Geospiza dubia dubia Gonld. 



Geospiza ihihia, J. Gould in Proc. Zod. Soc. Lmd. pt. V. p. 6 (1837)— .?')«/. T'"//. Beagle. III. Birds 

 p. 103 (1841) ; Salyin in Tiwis. Zool Soc. IX. pt. 11, p. 480 (1876) ; Sharpe, Cm. B. Br.l. Mug.. 

 XII. p. 9 (1888) ; Ridgway in Proc. U.S. Nat. Mns. XIX. p. 519 (1890). 



The type of this species came from Chatham Island, and is no longer in exist- 

 ence (cf. Sharpe, I.e.). The Smithsonian Institution received a series from the 

 ^^ All/atross" voj-age, and Messrs. Baur & Adams, as well as the Harris' expedition, 

 collected good series of it. 



This species differs from tlie foregoing ones in its much shorter bill, and from 

 ll.forti.-i in its larger-sized bill. 



Besides forty-seven skins from Chatham, we have five from Barrington iiml 

 fourteen from I)uncan Island, which we cannot separate from G. dttljia. There 

 is, however, no perfectly adult tnale from either Barrington or Duncan Island, and 

 the Barrington specimens are on an average rather small. 



9. Geospiza dubia albemarlei Ridgw. 



Oeo.ynza albemarhi, Ridgway in Pioc. U.S. Nat. Mus. XVII. p. 3G2 (1894), »/. op. cil. XIX. p. 523 

 (189G) (not separated from G.fmiis). 



Mr. Ridgway originally compared his G. albenuirlei with his G. media 

 (= coniro.<itri.s) and G. duhia GId., but afterwards, on comjiaring his Albemarle 

 specimens with G. fortis, came to the conclusion that they were liardly, if at all, 

 different, but admitted that some specimens were nearer G. dubia than G. fortis. 



From the examination of a large series from Albemarle and three from 

 Narborough Island it is evident, however, that this form can be separated, but must 

 be treated as a subspecies of G. duhia, and not of G. fortis, because it is always 

 separable from the latter by its larger bill and wing. The differences from G. dubia 

 are exceedingly slight, but generally the wing is 2 to 4 mm. longer, and the cnlmen 

 a little more arched. 



