218 MR. C. F. M. SWYNNEKTON ON THE 



Nymphalinse and skippers, Catopsilia, various grasshoppers and Noctuids, etc. 

 I think that the only striking exception was in the case o£ our commonest 

 bulbul {Pycnonotiis layardi). Two birds o£ this species that I tried both tended 

 to place Pierinse generally particularly low, and one of them at any rate 

 placed Danainse unusually high. This departure from the more usual order 

 may, of course, have been a case of vitiation due to captivity, but I should 

 rather hesitate to accept that explanation myself. It is more probably 

 a case of specialization — or, if the peculiarity be confined to individuals, it 

 may be analogous to the differential results on the natives who eat them of 

 the larvse of the moth AnapJie panda. 



I would in any case like to make it quite clear that I am not referring to 

 the larger preferences as between Orders. Thus, if any species specializes on 

 Diptera, it is likely to place these insects generally far higher than will a 

 bird that is not to the same extent a Dipteron eater : cf. the contrast in this 

 respect between ray drongos and my rollers. 



3. Newly-captured birds have, in so far as their preferences have been 

 ascertained, confirmed long-caged individuals, with only the occasional and 

 relatively slight transpositions that may be expected even between two wild 

 examples of the same species. In this connection Roller C^s list may 

 usefully be compared with that of Roller B. 



4. My tame ground-hornbills, roaming always unconfined and not fed at 

 the house at all, had much the same preferences, in so far as they were 

 ascertained — as, e.g., my captive rollers. 



5. Such preferences as I have elicited from wild birds — and they have 

 been fairly numerous — also go to bear out the views of my captive birds. 



This is the case for the reUability of my experiments. I will now frankly 

 state those points in which I believe them to be not entirely reliable. 



1. Through relative lack of exercise. I have evidence — not conclusive, for 

 the species were not the same, but certainly I think fairly strong — in favour 

 of the view that unconfined birds, probably with the digestion born of exercise 

 and of general greater enjoyment of life^ may more readily and frequently 

 become hungry enough to eat highly indigestible insects than even non- 

 fretting captive birds. This, as will at once be seen by a study of my 

 experiments on them, is not equivalent to saying that they do not discriminate, 

 or to saying that the order of their preferences may not be approximately 

 the same as those of captive birds of the same species. 



2. Through relative lack of practice. It being obviously impossible, how- 

 ever much: trouble one may go to, to provide the fidl amount of practice 

 enjoyed by a wild individual, mistakes from lack of it are likely to be 

 especially frequent amongst captive animals. 



Mistaken tastings, for instance. Not that for the purpose of these experi- 

 ments anytbing but mistaken eatings matters, for a rejection is as good as, 

 pr better than, a refusal ; and mistaken eatings are rare even amongst 



