210 THE GYPSY MOTH. 



supplied with an opera glass, by means of which he could 

 determine at a greater distance whether or not the birds were 

 feeding on the caterpillars of the gypsy moth. Few birds 

 were shot and few dissections made as compared with the 

 number of birds reported as feeding on the caterpillars. 

 This might serve to discredit the accuracy of the observa- 

 tions, were they not made at close range, and when the 

 caterpillars were large enough to be readily identified. No 

 birds were shot except where it was absolutely necessary to 

 determine whether it was the gypsy moth caterpillar or 

 some other upon which they were feeding. No observations 

 were accepted as conclusive unless the observer was known 

 to be careful in his work and conservative in his statements, 

 or unless ample corroborative evidence was obtained. While 

 there is some possibility of error in field observations, there 

 is also a possibility of error in stomach examinations. Di- 

 gestion in birds is so rapid that it is impossible to specifi- 

 cally identify some portions of their insect food unless the 

 bird is killed within a very few minutes after the insects are 

 eaten. When birds are feeding on larvoe, a large proportion 

 of the stomach contents is often unrecognizable.* 



Birds do not always swallow hairy caterpillars whole. In 

 many cases they tear them open, eating only some of the 

 internal parts which are unrecognizable upon dissection of 

 the stomach. Other birds appear to kill wantonly many 

 caterpillars and moths. Woodpeckers, jays and chickadees 

 have been seen to snap them up and then drop them to the 

 ground, sometimes uninjured but often mortally hurt. 

 Sparrows and other birds have been seen to kill many moths 

 which they do not eat. 



The Most Useful Birds. 

 The records of the observations made on birds comprise 

 one hundred and fifty typewritten pages. A glance over 

 these pages shows that the greater number of observations 

 have been made on less than a dozen species, and indicates 

 that these are probably the birds most useful in destroying 



* Considering the possibility of error in records made from either observation or 

 dissection, it would seem that one should be used as a supplement to and a check 

 on the other. 



