EXPLANATION OF FORM AND COLOURING. 313 



later she readily ate a Papilio angolanus, a Mycalesis campina, and four small 

 grasshoppers, after which she once again refused to touch another; refused, 

 then tasted and rejected Papilio leonidas, and refused without tasting Papilio 

 angolanus. 



Ten minutes later she obstinately ignored both, also grasshoppers. 



[Summary: — 



1. P. lyceus. 



2. P. leonidas and P. angolanus, and grasshoppers. P. leonidas 



not above P. angolanus. 

 Note ability to reduce black and white Cetoniid.] 



I was short of butterflies for the experiment, but in the late afternoon, two 

 Neptis agatha having come in in the meantime, I again gave her by hand a number 

 of grasshoppers, after which she readily ate a Mycalesis campina, four more grass- 

 hoppers, and a Neptis agatha, six moi-e grasshoppers, and a second Neptis agatha. 

 I now, rather expecting a rejection, offered her the same species of Cetoniid that 

 she had already eaten in the morning. A moment afterwards I regretted doing so, 

 for it was eaten after the usual crushing and being a comparatively bulky article of 

 diet carried the bird a stage too far and caused her to now refuse Papilio leonidas 

 and Papilio angolanus without tasting. Four or five times subsequently by lantei-n- 

 light at about ten-minute intervals I offered the two species of Papilio, but the 

 bird had settled down for the night and absolutely refused to touch them. 



The Papilio leonidas was already dead in the above experiment, but having been 

 kept on damp earth was still perfectly fresh and supple. It was, at any rate, not 

 appreciably preferred to the Cetoniid, which was once more reduced. 



Exp. 97. — October 20. Hungry; crushed and rejected a § Aletis monteironis, 

 Acrcea satis S f^nd a spotted Hypsid moth, Argina amanda, accepted most readily 

 and ate a Terias hrigitta, o,nd tasted and rejected an Acrcea asema. 



Later in the day, to test the probable value of the I'esemblaiice between Charaxes 

 hrutus and Papilio echerioides 6 , I carried out the following experiment, at the 

 same time experimenting on Pycnonotus layardi(B) with model and mimic reversed. 

 1 first fed the roller up on va.rious Orthoptera (including migratory locusts, crickets, 

 locustids, and small grasshoppers of various kinds) until she absolutely refused to 

 touch another. 



I then offered her a d Pajnlio echerioides with upper surface of wings fully 

 displayed. She took it by the base of one of the forewings, crushed it there and 

 threw it away, refusing to have anything to do with it on my reofl["ering it, I then 

 offered instead Charaxes hrutus, its upper surface similarly displayed. The roller 

 refused to have anything to do with it. On my closing its wings, however, and 

 reoffering it, this time of course with only the narrowly- barred under surface 

 showing, the bird seized it with actual eagerness, crushed it well, and swallowed it 

 with evident relish. The wings were not removed and as the insect was brought 

 into final position for swallowing, head first, they projected prominently on each 

 side of the bill with their upper surface vividly displayed so that the bird could 

 not have failed to take note of it. That she had done so was proved by the fact 

 that on my now ofiering a second Papilio echerioides 6 with the upper surface once 



