Vol. XIX 



1Q02 



J 'i'So-DGRASS, Food and Size of Bill I'tt Gcospiza. l^C 



From the facts given in the table the following propositions 

 may be pretty well established. 



I. The food of Geospiza as a za/iole differs ffo?n that of Neso- 



MIMUS. 



This is evident from the fact that of the twenty-one seeds found 

 in the Nesomimus stomachs, only eight were found in Geospiza 

 stomachs. Of these seed No. 8, which was the predominant seed 

 in the food of Geospiza fuliginosa pan'ula at Tagus Cove, Albe- 

 marle, was found in two specimens of Nesomiitius from Bindloe 

 Island. Seed No. i8, which occurred in eight A'esofuiviits 

 stomachs, was present in only three Geospiza stomachs. The 

 other seeds common to the diets of the two genera were of rare 

 occurrence in each. 



Nesomimus eats a great many more insects than does Geospiza, 

 being insectivorous and granivorous in about equal proportions. 

 Numerous pieces of large insects were found in nearly all of the 

 Nesomimus stomachs, including grasshoppers, files, beetles, cater- 

 pillars, and also spiders and centipedes. The seed part of the 

 diet differs from the food of Geospiza in consisting of larger 

 seeds, of seeds that the much smaller-mouthed Geospizce could 

 not handle. 



II. The same species at different localities may feed oti different 

 seeds. 



The truth of this statement is best shown by a study of the food 

 of Geospiza fuliginosa parrula (PI. XI, Fig. 6) the most widely 

 spread and most abundant form of Geospiza on the archipelago. 

 In the stomachs of nineteen specimens of this variety from Tagus 

 Cove, Albemarle Island, collected in January, there was found a 

 total of only eleven species of seeds. Of these, seed No. i had 

 been eaten by two birds, No. 2 by one bird, No. 4 by two birds, 

 No. 6 by two birds, No. 8 by nine birds, No. 15 by one bird. 

 No. 22 by five birds. No. 24 by three birds, No. 41 by tw^o birds, 

 No. 58 by one bird, and No. 59 by one bird. This shows that 

 the birds here feed on seed No. 8 more than any other, and that 

 seed No. 22 was second in numbers. Moreover, these two seeds 

 were present in much greater numbers than the others in each 

 stomach in which they occurred. 



From Elizabeth Bay, Albemarle, there are only two specimens 



