306 MK. C. F. M. S'W'YNNEKTOX ON THE 



[Summary : — 



(a) 1. C pollux (and H. loahlbergi, S. cacta^ P. ca^xlui, P. cebrene, 

 S. anacardii apparently not preferred to it). 

 2. P. demodoGus, P. angolanus, various favourite grasshoppers. 



G. candiope was also preferred to P. angolanus and the grasshoppers.] 



Exp. 85. — September 10. Fed the roller on various Orthoirtera, &c., till she would 

 eat no more. She then refused but went on to taste and reject Papilio demodocus 

 and Papilio angolaiius, just took Hypolimnas loahlbergi listlessly in the tip of her 

 l)ill and dropped it, also Precis cebrene, and refused without tasting Pyrameis cardui, 

 Precis natalensis f., Salamis anacardii, and Salamis cacta. After an interval she 

 refused without tasting Pajnlio demodocus, Euralia loahlbergi, and Salamis cacta, 

 crushed and readily ate a mimetic Baoris and a Lycsenid (probably Zeritis of 

 outskirts), but then went on to eat also Papilio demodocus and Eurytela dryope, 

 rejecting at once however after she had tasted it Neptis agatlia. I decided not to 

 reofter the two main butterflies of the experiment as the bird was too hungry once 

 more and I had given away such grasshoppers as remained. The Salamis cacta was 

 in any case now dead. 



[Probable that the Baoris was preferred to P. demodocus and, if she really knew 

 ther)i, to H. loahlbergi and ;S'. cacta. The Zeritis, eaten after stimulation, was not 

 necessarily better liked than the E. dryopeJ] 



Exp. 86. — Sejytember 11. Had just been feeding and had returned to her usual 

 perch presuixiably satisfied, leaving a. nuniber of very favourite insects uneaten. 

 She then refused obstinately to touch Mylothris ruep2)elli or a Terias, but readily 

 accepted and ate Neptis agatha, most readily accepted a number of grasshoppers by 

 hand, including those she had abandoned, and finally absolutely refused to touch 

 either another grasshopper, Neptis agatha, Papilio lyceus, or Papilio demodocus. 

 Three minutes later she tasted and rejected Papilio lyoius and Papilio demodocus, 

 looked very suspiciously at both Euxanthe wakefieldi and Hypolimnas wahlbergi, 

 both wingless, the bodies in each case being very reminiscent of Danaince, then took 

 each in a gingerly manner in the tip of her bill and dropped it, I believe quite 

 without tasting, I'eadily accepted and ate Precis natalensis f., Pajnlio jjolicenes, 

 Papilio antheus, and Papilio hip])ocoon (each with one wing), doubtfully refused 

 without tasting Euxanthe ivakefieldi and Hypolhnnas wahlbergi, readily ate Papilio 

 hippocopn and Precis archesia, refused without tasting Papilio demodocus and 

 Papilio lyceus, but readily accepted and ate Pajnlio hippocoon, Precis natalensis 0, 

 and Precis archesia. She then refused Precis natalensis, Papilio hippocoon, and 

 Euxanthe wakefieldi, ci'ushed slightly in the point of its bill and dropped Precis 

 cebrene, accepted and ate Salamis anacardii and Salamis nebulosa, but refused the 

 next as also several of the preceding butterflies reoffered. 



Later she i-eadily ate Papilio angolanus and tliree small grasshoppers, refused 

 without tasting the next Papilio angolanus, also Pajnlio lyceus and Papilio 

 demodocus, tasted and rejected Euxanthe wakefieldi and refused without tasting 

 Precis archesia. It was obvious that tlie bird found the. thorax of the Euxanthe 



