'i\82 



MR. C. F. M. SWYNNEETON ON THE 



A little later he accepted and ate a Mycalesis camjnna and a large 5 Saturniid 

 moth with undeveloped wings, which I had found half drowned in a puddle and 

 now gave to "fill up," refused, then tried with disinclination and rejected the next 

 Mycalesis cmnpina, but readily ate, after crushing each, an Eurytela hiarhas, a 

 Byhlia goetzms, a Phrissttra isokani S , and a Papilio angolanus, a fair-sized 

 grasshopper and another Byhlia, four or five small grasshoppers, then refused, 

 afterwards rejecting, a Mycalesis campina, but crushed and ate a Nej^tis agatha. 

 After a few more small grasshoppers he refused to take either a Mycalesis campina 

 or a Neptis agatha^ but accepted a dry Papilio angolanus. It was an old specimen 

 that I took from my collection, having no more fresh ones, in order to ascertain 

 whether the species would now be refused without tasting ; it was, of course, 

 rejected. The roller next refused for a moment, then, after looking closely at it, 

 accepted and ate a Henotesia perspicua, and accepted a dry Papilio demodocus^ but 

 on attempting to crush it threw it awa}^, refused to taste an tJveres cissus, accepted, 

 crushed, and swallowed a PhopalocamjJta libeon, I'efused, then accepted, crushed 

 well, and ended by throwing away an A. schoeneia, showed eagerness for a Char axes 

 candiope, which, however, I reserved, and with equal eagerness accepted and ate a 

 Cyligramma latona. 



[^Order : — • 



Prob. P, angolanus, C 1 . 

 H. perspicua and J 

 P. demodocus. [2. 



G. candiope, C latona, with 



greater eagerness than 

 R. libeon, A. schoeneia. 



3. 



P. angolaiiios, 



Grasshoppers, 



E. hiarhas and perhaps 



S. goetzius and P. isokani c?. 



M. ruejypelli above 

 T. senegalensis. 



A\ agatha 3.' 



4. M. camjnna. 



5. B. mesentina at any rate 

 not beloiv 



6. T. senegalensis, T.hi'igitia, A. 

 johnstoni, & P. echerioides rS . 



Everes cissus.'] 



Conclusions from Experiments on Captive Rollers. 

 Many interesting conclusions and suggestions appear to arise from the above 

 experiments. I will postpone the discussion of most of them until I come to 

 publish m,y far more numerous and equally interesting experiments on very many 

 other .species of captive insectivorous birds. 1 will refer briefly to only one r,esult 

 here. It is that of the fine grading. It would appear that relatively few species 

 of prey are eaten by a given individual enemy to repletion-point. Relatively to 

 enemies generally — even individuals of the same species with their slightly- 

 differing tastes — the number of absolute " Grade 1 " species is 3'et further reduced. 

 There are thus probably comparatively few species that do not at some time 

 require to be distinguished by certain enemies from species acceptable nearer to 

 repletion-point than themselves. Where their own parent-form was higher-grade 

 than themselves they vvill, in order to gain any real advantage from their added 

 nauseousness, have had to be distinguished from that parent-form. Where 

 the present nauseousness has arisen by many steps the oft-repeated necessity 

 for differentiation from successive parent-forms might, theoretically, have in 



