3 68 Snodgrass, Food aiid Size of Bill in Geospiza. \o^ 



scandcns rothschildi. G. affinis, G. O'assirostris, G. prosthemelas^ 

 G. heliobates. The specimens were collected from the islands 

 of Albemarle, Narborough, James, Seymour, Duncan, Charles, 

 Hood, Barrington, Tower and Bindloe. The dates run from 

 December till June, inclusive. 



Comparison has been made of the food of individuals of the 

 same species at different places, and of the food of different spe- 

 cies at the same and at different places. The results are some- 

 what conflicting. In any case one would require a great amount 

 of evidence to come to any definite conclusions. Then, too, 

 there is always a doubt created by the fact that the specimens were 

 not taken on the different islands during the same months, and by 

 the fact that the seasons vary considerably at different localities. 

 What might appear to be evidence of a difference in food habit 

 between a species on one island and a different one on another 

 island, might be nothing more than a seasonal change of diet due 

 to different plants being in seed at the two times. However, a few 

 conclusions may be positively deduced, the results being sufficient 

 to warrant the discussion. 



The detailed records of the two hundred and nine stomachs 

 are omitted. The data obtained are given in the following 

 table, and the seeds are illustrated on Plates XH and XHI. The 

 seeds have not been identified, but the names are not necessary. 

 They are drawn to show their relative sizes, and are referred to 

 in the succeeding discussion by their numbers on the plates. 

 Figures i to 44, inclusive, except figure 42, are magnified 6|- 

 times. The others are magnified only half as much. The 

 stomachs of Mockingbirds {JVesomimiis) from eight islands have 

 been examined in the same way. The records of these are given 

 at the end of the table, and the seeds are figured on the plates 

 along with the Geospiza seeds. The stomachs of all contained a 

 total of sixty kinds of seeds. Seeds Nos. 59 and 60 are not 

 figured. 



Following is the table (pp. 369-374) containing the records of 

 the examination. It of course does not show the relative numbers 

 of each seed present. 



