'2^8 Skot>grass, F'ood a /id Size q/ Bill m Geosfiza. \j^c\. 



general truth. We cannot show how far it actually holds true of 

 other species on the archipelago. The similarity in the food of 

 these two common forms on Seymour and Barrington Islands is 

 so striking, however, that one is almost forced to the conclusion 

 that all the species of Geospiza eat simply whatever seeds are 

 accessible to them. 



IV. Different species at different localities may feed on the same 

 kinds of seeds. 



The truth of this proposition may be seen by a comparison of the 

 food of Geospiza conirostris conirostris (PI. XI, Fig. 5) on Hood, 

 G. scan dens fatigata on Seymour and Barrington, and G.fuUginosa 

 parvula at Iguana Cove, Albemarle, and on Seymour and Barring- 

 ton. The largest part of the food of G. conirostris conirostris in 

 May consisted of seeds Nos. 14 and 15. Next in numbers were 

 seeds Nos. 10 and 17. All but two of thirteen birds had eaten 

 No. 14, and all but one No. 15, while Nos. 10 and 17 were each 

 represented in six stomachs. Seed No. 5 was found in one bird. 

 No. 23 in one bird, No. 26 in five birds. No. 28 in three birds, No. 

 29 in one bird, and No. 57 in one bird. 



Hence, the food of G. conirostris on Hood Island is in the main 

 the same as that of G. fuliginosa parvula at Iguana cove, Albe- 

 marle, and on Seymour and Barrington Islands, and is also the 

 same as that of G. scandens fatigata on Seymour and Barrington. 

 There are thus three species of Geospiza with very different bills 

 (PI. XI, figs. I, 5 and 6), living at three localities, whose food is 

 almost identical at approximately the same time of the year. 



V. Different species at the same or at different localities 7nay feed 

 on different seeds. 



If the size and shape of the bill is dependent on the character 

 of the food, this proposition should be a general truth. However, 

 the material under consideration affords only a few instances of it. 



Geospiza pachyrhyncha (PI. XI, Fig. 9) is peculiar to Tower Is- 

 land. The stomachs of seven specimens taken in June contained 

 only seed No. 57. We have no data to show what the food of other 

 species on Tower consists of. Vegetation is extremely scant on 

 the island, and all the birds may be forced to eat the same seed. 



A specimen of Geospiza strenua (PL XI, Fig. 8) taken in January 

 on Narborough had only seed No. 22 in its stomach. Five speci- 



