1912.'] NATURAL SCIENCES OF PHILADELPHIA. 313 



man has fired his last shot ineffectually at a charging tiger or rhinoce- 

 ros; he naturally shows a "marked reaction" by taking to his heels, 

 not because he or his ancestors have had similar experiences, but 

 because he can run. So with the butterflies, skilful efforts to escape 

 do not necessarily indicate previous racial experience of the same 

 nature. 



The principal failing of the selectionists always has been a vast 

 ignorance of what wild birds really eat. They have made very 

 little effort to acquire such knowledge, and their speculations through- 

 out show the lack of it. Practically the only large body of authentic 

 information on the natural food habits of birds is contained in the 

 records of the United States Biological Survey. They comprise 

 detailed identifications of the contents of more than 48,000 bird 

 stomachs representing all families of birds and collected in hundreds 

 of localities in the United States at all seasons. The United States 

 has a goodly representation of butterflies, yet only five of these 

 48,000 stomachs contained remains of Rhopalocera. It is hoped 

 this will be more satisfactory to the selectionists than the " negative 

 evidence" they are accustomed to cite with contempt. 



The extreme artificiality of Pocock's experiments and the inappli- 

 cability of the results to the natural relations of British birds and 

 insects are so evident that it is not worth while to comment on the 

 details. A few comparisons of the results with those of experiments 

 recorded by Poulton are of interest as showing the inconsistency, 

 inter se, of experiments. It has not been possible to collect a large 

 number of such comparisons because Poulton's experiments were 

 chiefly with lizards and few with birds, while the opposite is true of 

 Pocock's. The varying stages in which the insects were presented 

 also tend to limit comparisons. The table includes all possible 

 direct comparisons and only one pair in eight shows real correlation. 



Birds: Poulton. Pocock. 



Vanessa urticai, larva D A 7 R 4 D 1 



" pupa R A 2 R 2 D 1 



Clisiocampa neustria, larva D A 1 



Euchelia jacobaz, ad A A 1 R 4 



Cosmotricha potatoria, larva D A 1 R 4 D 1 



Anthrocera filipendula, ad. A R 4 



Lizards: 



Apis mellifera, worker A R 3 



Pieris rapw, ad A 20 A 2 



Notes on Pocock's experiments, by Prof. E. B. Poulton, are given 

 21 



