464 McAteb, Contents of Bird Stomachs. [o^ 



portance of elements of the food that have parts very resistant to 

 digestion, a difficulty which is reduced to the minimum when pro- 

 portions are estimated according to the volumes. 



Numerical systems are not sufficiently comprehensive. Finely 

 comminuted, fleshy, or pulpy food, or food occurring in indefinite 

 masses cannot be reckoned by numbers. Under the percentage- 

 by-bulk system, all food can be included in the computations. 

 Intelligible comparison of one part of the diet with another or of 

 the food of one species or group of species with that of another, as 

 well as graphic representation of the proportions of the food, are 

 only possible when the volumetric method is used. This system 

 is the better therefore, as the more complete is always superior to 

 the less. 



On the other hand statements of the frequency of occurrence 

 of food items in bird stomachs may perhaps be taken as rough 

 indices of availability of the food or of relish for it. And statements 

 of the number of individuals in stomachs have an interest as 

 " records," the interest being in direct proportion to the bigness 

 of the number. 



The ideal system from the writer's point of view is one that 

 combines the good points of both the numerical and volumetric 

 methods — a system which, as a matter of record, counts individ- 

 uals as far as possible, or at least in enough instances to assure 

 the inclusion of typical cases, and which further estimates the 

 proportion of all important items by bulk. Such a system has 

 been approved and used by Forbes, Beal, Barrows, Kalmbach, 

 Judd, Sanderson, Dearborn, Weed and the writer, among American 

 investigators. The consistency with which it has been applied 

 is in striking contrast to the vagaries of numerical systems which 

 have scarcely been used alike by any two writers. We have shown 

 that the chief criticisms that have been aimed against the volu- 

 metric system, apply equally well to the numerical methods. 

 The latter have other weak points that do not appear in the per- 

 centage-by-bulk system, and the few good points peculiar to the 

 numerical system can profitably be combined with the volumetric 

 method. This gives us a compromise technique that contains all 

 of the good features and a minimum of the weaknesses of its com- 

 ponents. 



