104 THE SCOTTISH NATURALIST 



any other source during the past twenty-three years, and 

 it is the one now employed by practically all workers in 

 the United States. 



The second method has, so far as I am aware, only 

 been employed by Rorig, the objection to it being its 

 impracticability. 



The first method has, with various modifications, been 

 used by a few American workers and practically all British 

 workers, but as McAtee x has so pertinently and ably pointed 

 out, it has assumed no less than five different forms, viz. : 



1. The total number of insects, seeds, etc., are given 

 that are consumed by the whole collection of birds examined. 



2. States the number of times a particular kind of food 

 is taken by the total number of birds. 



3. Enumerates the number of birds eating certain kinds 

 of food, and the number of specimens taken. 



4. " The proportion of the number of times a certain food 

 is taken, to the total number of times all foods are taken, 

 is considered the percentage of that food in the diet." 



5. " Estimates the proportions of food items according 

 to their numerical representation among the total of all 

 specimens eaten." 



It is at once evident that such a method, capable of 

 assuming so many different forms, stands condemned. But 

 there are much graver objections to the numerical method 

 as will be seen later. 



McAtee 2 has dealt with this subject in great detail, but 

 even at the risk of repeating much that he has so ably 

 stated, I think a useful purpose will be served by once again 

 setting forth the chief advantages of the percentage by bulk 

 or volumetric method. 



Forbes 3 referring to this method states that "opportunity 

 is afforded for careful and trustworthy estimates of the 

 ratios each element bears to the other, so that the average 

 significance of the food can be discovered. Practically this 

 is indispensable. Whatever method fails of this, while its 



1 Auk, 191 2, vol. xxix., pp. 449-464. 



- Op. at. 



3 Bull. III. State Lab. Nat. Hist., 1880, vol. i., No. 3. 



