( xxxvi ) 



able conditions, such as, when the butterfly passes very close 

 to the bird's perch, or when the attention of the insect is dis- 

 tracted during feeding, courting, ovipositing, etc." — these 

 were Mr. Marshall's main suggestions in explanation of the 

 paucity of evidence (Trans. Ent. Soc. Sept. 20, 1909), and he 

 gave over 670 records of actual attacks, many of them multiple, 

 drawn from various sources. His striking illustrations — of 

 English kestrel and garden-warbler, familiar birds not known 

 to eat butterflies, yet now observed preying on them systemati- 

 cally — and, more lately, my own results, seem rather to bear 

 out the first of these suggestions; wings too very often are 

 removed (though also frequently swallowed), and the head 

 is sometimes rejected, and the butterfly by some birds eaten 

 piecemeal — and the truth of Mr. Marshall's last suggestion 

 I can bear out from actual observation. It is a point of 

 great importance. 



Another very important consideration, which I suggested 

 in " The Ibis " for October 1912, is that in localities or seasons 

 in which, as at Chirinda, the butterfly population is very 

 small indeed relatively to that of other orders, we cannot 

 reasonably expect to see anything approaching as many attacks 

 on butterflies as on these other insects. Both this view and 

 Marshall's (that birds probably accept easy chances) are borne 

 out by the fact that when, temporarily and locally, I have 

 removed the disparity in population by definitely releasing a 

 number of butterflies to birds, and when I have given them 

 the easy chances by placing disabled butterflies in their way, 

 I have many times witnessed more or less numerous attacks. 

 One such experiment elicited as many as sixty attacks by 

 eight species of birds in what could have been little over 

 half an hour. " But if it is necessary to disable a butterfly 

 to get it attacked you admit that butterflies are not attacked 

 under natural circumstances." The objector here forgets 

 that butterflies are not always flying zigzag at top speed, 

 far from cover. Kelatively easy opportunities such as 

 Marshall suggests occur very frequently in nature — and I 

 have seen some taken. But the fact that they entail the 

 butterfly's presence in or within a few inches of cover im- 

 mensely lessens our chances of seeing the attacks. During 



