290 MR. C. F. M. SWYNNfiRtON ON THE 



I think that the experiment on the whole justifies the above order, though some 

 placings are naturally lower than others. Nothing was eaten to quite the same 

 point of repletion as in the Coreid bugs. Additionally, M. agathina was probably 

 at any rate not better than M. campioia, or this than iV". agatha, or E. dryope and 

 L. thalassina than P. dardamis <S ; P. dardanus c? was eaten apparently more 

 readily than P. lyceus once, but perhaps in consequence of stimulative influence by 

 the latter ; A. schozneia was probably not better than P. cardui, or yellow G.jlorella 

 than A. schceneia or P. Mppocoon, or P. lyceus than the latter or C. jiorella. 

 P. hippocoon seemed very near or equal to C. florella and A. schceneia. 



(2) Note at the time: The various occupants of grade 7 " were evidently more 

 obviously unpleasant than the pupa of Acrcea serena. This was treated exactly 

 similarly by the kingfisher [Halcyon cyanoleucus). In both cases it was obvious that 

 the birds were not at all sure whether it was fit to eat (at that stage) or not, and it 

 seemed as though they must almost possess some special gauging power at the back 

 of the throat {cf. experiments on swallows, passim). Neptis agatha was eaten after 

 twenty-four grasshoppers later, showing that each of the above insects was decidedly 

 nearer the i^ejected agathina in degree of unpleasantness than to agatha., and the fact 

 that M. agathina was played with — the roller being probably nearly hungry enough 

 to eat it — just before she rejected the Acroea pupa shows the same. I tasted a little 

 of the jvxice exuding from the pupa after its first rejection and it was not markedly 

 unpleasant, as in the case of the larva and imago^a slight poppy flavour, I thought. 

 The ichneumon has great tenacity of life. After its crushings by the birds and an 

 additional sharp crushing of the thorax by myself, it quickly recovered and moved 

 about in most lively fashion. Lygoius swynnertoni (found feeding in numbers on 

 Vernonia living stoniana) is much more easily killed. Its extreme unwelcomeness, 

 in spite of its practical lack of smell, while the large strong-smelling Coreids were 

 eaten with relish in preference to the most palatable butterflies and grasshoppers, 

 shows that the more typical fttigr-smells are in no way an indication of unpleasant- 

 ness to the roller. The Coreids are literally in swarms on the oranges near the 

 forest, sucking their juices and blackening their surfaces with their numbers — a 

 spot visited almost solely by bulbuls. On the trees being shaken they fly or, rather, 

 drift ofi' slowly in crowds, numbers landing on the ground just round about. They 

 are common, too, on tender terminal shoots of cofi"ee, conspicuously exposed on the 

 surface and usually several together. They are very sluggish in their general 

 habits."] 



Exp. 73. — August 22. The bird was very hungry. She readily ate, after crushing 

 each, a Terias^ a Mylothris rueppelli (sex not noted), and a S M. agathina., rejected 

 an Acrcea johnstoni and an Acrcea esebria (fulvous c? ), also a 5 MylothrHs agccthina. 

 She then readily ate a cJ Mylothris agathina, again rejected the $ and once more 

 readily ate a S ', tasted and rejected Lygceus svjynnertoni and the ichneumon (of 

 yesterday's species), readily ate another Mylothris agathina J , tasted and rejected 

 Acrcea serena., and readily ate a Terias. 



I now gave her a smallish grasshopper, after which she readily ate another Terias 

 but crushed and at once rejected a J Mylothris agathina. Then followed two more 

 grasshoppers and another Terias, three more grasshoppers and another Terias, five 



