EXPLANATION OF FORM AND COLOURING. 319 



large or too much trouble to eat in the then state of her appetite, edging away 

 from it when I continued to hold it to her ; refused for a time a Myc. camjnna, 

 then tasted it slightly and withdrew from it, and finally knocked it out of my 

 forceps irritably with a sidelong blow of her bill. She readily, however, ate a 

 Neptis agatha, then once more refused the Myc. camjmta. 



[Summary : — H. ervphla, iV". agatha, and, on manner, P. angolanus were preferred 

 to M. cam2nna, and JV. agatha to Phymateus viridipes.] 



Exp. 108. — November 26. Noticing that the two fragments of yesterday's 

 Phymateus (the thorax was still alive) had almost completely lost their charac- 

 teristic smell (it was still just noticeable on opening the wings at the point where 

 they joined the thorax), I decided to once more offer it to the roller. I therefore 

 fed her till she refused to eat any more grasshoppers, and refused, then doubtfully 

 tasted and rejected Pajnlio angolanihs. She then accepted the Phymateus 

 abdomen and subjected it to a very pi'olonged tasting, making once or twice 

 as though to swallow it before at last rejecting it, I then reoffered it twice and 

 it was each time once more accepted, crushed, and dropped. On the third occasion 

 it was refused without tasting. 



Ten minutes later she again refused, but after some persuasion tasted slightly 

 and rejected Pajnlio angolanus, twice tasted and doubtfully rejected a small portion 

 of the Phymateus abdomen, afterwards refusing it without further tasting, and 

 twice tasted and dropped its thorax, afterwards I'efusing it too without tasting. 

 She then accepted an unidentified beetle, pressed it once or twice, but made no real 

 attempt to crush it, and finally dropped it, obviously because too replete to go to 

 the trouble of crushing it. On my reoffering it, she treated it again in the same 

 way. She then tasted slightly and rejected Papilio angolanus, crushed and readily 

 ate Precis artaxia, refused through sheer I'epletion, then accepted and ate a Precis 

 madagascariensis ; refused, then ate a $ Arrugla basitta ; refused, then ate two 

 Precis artaxia and rejected with disgust the Meloid beetle, Lytta designata, 

 exuding from its thorax a small quantity of yellow juice. 



Ten minutes later she readily ate after crushing it the small Cetoniid, Leucocelis 

 p)arallelocollis, expei-iencing no difficulty from its hardness or polish. 



[Summary: — Note: "The long and repeated trials given by the roller to day to 

 the Phymateus when practically replete contrasted strongly with its prompt 

 dismissal of the insect yesterday when considerably hungrier, and seemed to show 

 that the unpleasant taste and odour are chiefly confined to its frothy exudation 

 and disappear with its disappeai'ance." 



P. ataxia (and probably P. madagascariensis, Arrugia hasuta 5 , and Leucocelis 

 parallelocollis) was preferred apparently to P. angolanus, the unidentified beetle 

 (which the roller at that stage of appetite would not trouble to crush), Phymateus 

 viridipes, and even favourite species of grasshoppers.] 



Exp. 109. — November 27. Tasted well and rejected a Zonocerus elegans and 

 refused persistently without tasting a wingless Acrcea esebria, retreating before it 

 when I continued to offer it. It was interesting that after this lapse of time she 

 should continue to remember clearly the Acreeine body and to associate it with a 



