NO. 7 PROTECTIVE ADAPTATIONS McATEE 7 



quoted in the following pages include more or less of the handiwork 

 of practically every prominent American systematic zoologist of the 

 period. Nevertheless everything is not identified, far from it ; expert 

 assistance has not been availa1)le in some cases when needed, too far 

 digested residues sometimes occur, and rarely we find also well-j-jre- 

 served but puzzling objects that indefinitely defy classification. 



These, however, are but minor flaws in the system ; the ground- 

 work of our faith in tlie results of stomach analysis is the law of 

 averages. Given good distribution geographically and seasonally, 

 which necessarily follows from miscellaneous collecting carried on 

 for so many years, the reliability of results varies directly with the 

 number of stomachs. The collection (80,000) here reported upon 

 is believed sufficient to furnish fairly dependal)le data, although addi- 

 tions are made almost daily to the list of animals identified from l)ird 

 stomachs. 



The total number of identifications of animals from these stomachs, 

 counting those of whatever degree, once for each time identified 

 irrespective of the number of individual specimens concerned, is 



237-399-' 



It was impracticable to compute the total numl)er of individual ani- 

 mals concerned for the reason that these were not counted in all 

 cases. Moreover this figure would not have been especially useful 

 in the absence of estimates for comparison of the actual animal popu- 

 lation of significant areas. In casting about for a standard which 

 would aflford some idea of the frequency of occurrence of animals of 

 various groups, the estimated number of species therein proved to be 

 the only one availaljle for the whole range of the animal kingdom. 

 That the number of species in taxonomic groups bears a general re- 

 lation to the number of individuals can not be questioned. It is easy 

 to point out exceptions, but remember we can only deal with this 

 problem in an approximate way, and it goes without saying that on 

 the average a group more numerous in individuals will have devel- 

 oped more species than one less numerous. The correspondence is 

 not exact, but it is sufficient to give a fair working idea of the position 

 of the various groups in the scale of frequency of occurrence, the 



'The tabulation necessary to yield this figure was an immense one (covering 

 nearly a thousand typewritten pages) and has been found, it is not surprising, 

 to contain some errors. These are so small, however, that rectification of them 

 would not cause changes of more than a fraction of one per cent in any part 

 of the results, except in the table for Coleoptera, pp. 65-67. Hence they do not 

 invalidate the figures at all for the purpose here used of showing in a general 

 way the tendencies exhibited by our birds in their choice of animal food. 



