'(l 



' 12 THE RELATION OE SPARKOWS TO AOEICXJLTUBE. 



•liil I of identifying in the parent's beak the insects usually fed to nestling 



birds is exceedingly difficult, yet it yields more satisfactor}^ results 

 than examination of the stomachs of the nestlings, not onlj^ because 

 stomach examination shows nothing as to frequency of meals, but also 

 because the soft insects given by most of the smaller birds to their 

 young are generally unidentifiable in the stomach. 



EXPERIMENTS WITH CAPTIVE BIRDS. 



Ver}^ often birds that are too shy to be watched in the field may be 

 kept in captivity and experimented with. If the experiments are 

 carefull}^ conducted much can be learned as to the amount of food 

 eaten, preferences in food, and questions relating to the dissemina- 

 tion of the seeds of fruit and weeds bj^ birds. Birds selected for 

 experiment should be recentlj^ trapped, because those that have 

 been long in confinement usuall}^ develop unnatural tastes. In test- 

 ing preferences in insect food it is convenient to place the insects on 

 a piece of cork anchored in the center of a bowl of water. This pre- 

 vents them from escaping and makes the conditions almost identical 

 in the case of each kind. The insects should be equal in volume. 

 Thus, a fair experiment would be made if a ladybird {Coccinella 

 9-notata) and a 12-spotted cucumber beetle {Diahrotica 12-pundata) 

 were both put on the cork island at once for the bird to select from, 

 because both insects are of about the same size. The same principles 

 apply to fruits and seeds, though, of course, the use of the cork is not 

 important in their case. The food of nestling birds may be sometimes 

 studied to advantage by removing the young from the nest and plac- 

 ing them in a cage almost out of reach of the parent bird, so that in 

 feeding them the latter will drop a large proportion of the food just 

 inside the cage. 



LABORATORY EXAMINATION OF STOMACHS. 



The contents of the crops (or gullets) and stomachs of wild birds 

 IP'' are examined to find just what elements of food the bird has chosen 



iiBiIll and the proportion each bears to the total amount of food ; and if the 



number of stomachs examined be large enough, the difference due to 

 individual variation is eliminated. The identification of the food 

 giji) I found in a bird's stomach is difficult, for two reasons: (1) Because of 



1,, the great variety of substances that may be found in the stomach, 



j|;;j ; and (2) because of the fact that the semi-digested pieces and fragments 



of insects, fruits, and seeds are often so comminuted that exact iden- 

 tification is well-nigh impossible. The method of procedure employed 

 ' , in the laboratory of the Biological Survey in making stomach exami- 



nations is as follows : The stomach is slit open with a scalpel, and its 

 contents are first washed into a pan by a jet of water from a wash bottle 

 I and then, with the exception of the sand and gravel, poured into a 



