Production of Mimetic RcsemUances among Biittcrfiics. 337 



vestigation has ever been undertaken in those tropical 

 regions where mimicry among butterflies is most prevalent, 

 and where we should therefore expect to find the greatest 

 destruction by birds. Moreover, this line of inquiry is un- 

 fortunately much less satisfactory in the case of butterflies 

 than in most other insects, owing to the fact that birds so 

 frequently snip off" the wings before eating them, thus 

 rendering their recognition almost impossible ; and there 

 can be little doubt that the presence of butterflies in birds' 

 stomachs must have been overlooked through this cause. 

 Thus it is rather to ornithologists that we must look for 

 any material addition to our knowledge of this subject ; 

 and an attempt to carefully identify the food brought to 

 their young by insectivorous birds would probably yield 

 much valuable information. 



Enough has now been said to show that the assumption 

 that birds seldom attack butterflies is certainly premature ; 

 and for my part, I am convinced that when the subject 

 has been more fully and systematically investigated that 

 assumption will prove to be entirely unfounded. 



We may now turn to the list of available records. So 

 far as concerns those already published, I believe it to be 

 fairly complete, though a certain number must inevitably 

 have escaped my attention. For the remainder I have 

 to offer my hearty thanks to the following friends and 

 correspondents, who have very kindly furnished me with 

 a number of valuable unpublished observations : Mr. 

 Parkinson Curtis, Professor C. B. Davenport, Dr. F. A. 

 Dixey, Dr. F. Doflein, Mr. C. H. B. Grant, Colonel N. 

 Manders, Mr. S. A. Neave, Father O'Neil, Professor E. B. 

 Poulton, Mr. R. Shelford, Mr. C. F. M. Swynnerton and 

 Professor August Weismann. 



The number of Observations is unfortunately still too 

 small to justify any broad generalisations from the results 

 shown, but a few salient points may be noted. 



The paucity of records in the case of Lycaenidae and 

 Hesperiidae is probably due to the comparatively small 

 size of these insects, which would thus render them 

 difficult to identify under such conditions. The great 

 majority of the evidence is furnished by the Nymphalinae 

 and Pierinae, and of these the latter subfamily is always 

 easily first. This is in accordance with Bates' experience 

 on the Amazons. For he says: "I could not, from their 

 excessive scarcity, ascertain on the spot that the Leptcdides 



