28 THE COMMON CROW. 



clew to the character of some of the soft food. The utmost care is 

 needed to distinguish fragments of the hard parts of insects, particu- 

 larly jaws, and the heads of weevils, and also small seeds that are 

 likely to be overlooked among the bits of sand and gravel. The finely 

 broken bits of insect shell usually float. Advantage is taken of this 

 fact by throwing water into the vessel in a strong jet so that it 

 may entangle bubbles of air and rise to the top, where they can be 

 skimmed off. 



The fine matter collected in the filter is examined under the hand 

 lens, or the compound microscope if necessary, and by comparison with 

 the coarser material it is possible frequently to determine not simply 

 that it is vegetable or animal dust, but to name the particular animal 

 or plant from which it came. In any case, the animal, vegetable, and 

 mineral components of this fine material are separated and added to 

 the coarser material already isolated. 



After the materials contained in the stomach are recognized the next 

 step is to determine the percentages of animal, vegetable, and mineral 

 matter. For this purpose the smallest heap or mass of identified matter 

 is selected as the unit, and each of the other heaps is divided into 

 portions of like size. Then, by a very simple process, it is possible to 

 determine the proportion of each ingredient of the food, and it has 

 been found most convenient to express this proportion in percentages. 

 For instance, if the heap of sand and gravel is selected as the unit, 

 the animal matter may be 2J times as much and the vegetable matter 

 3 J times, giving a total of 7 parts, each of which represents 14.3 per 

 cent. In this case, therefore, the record would stand: Animal matter, 

 35.7 percent; vegetable, 50 percent; mineral, 14.3 percent. The pro- 

 portions by weight would be very different from those by bulk, and 

 these again would differ from those based on the numbers of individual 

 seeds or insects found in the stomach. 



In the case of a bird which eats insects only it might be possible to 

 use the numerical method with some accuracy ; yet even then much would 

 have to be left to individual judgment in estimating how many small 

 insects were equivalent to one large one, or how many harmful insects 

 would be necessary to offset the consumption of a given number of ben- 

 eficial insects. Moreover, only under the most favorable circumstances 

 would it be possible to determine just how many individuals of each 

 kind were represented in the stomach contents, for, even if swallowed 

 whole, so soon as digestion begins the individual insects become dis- 

 membered, crushed, and broken, and within a short time only the 

 hardest parts, such as the heads, wing covers, legs, and jaws, remain 

 in recognizable condition. 



It has seemed best, therefore, in attempting to determine the propor- 

 tions of the various food substances in Grow stomachs, to depend upon 

 the method of equal masses or bulks, which method is adopted in the 

 present bulletin. In most cases the number of individual seeds, insects, 



