EXPLANATION OF FORM AND COLOUEING. 247 



accepted and eaten. The roller was then twice offered a Papilio lycetis $ and each 

 time tasted it well and rejected it, but tasted and most readily ate a Precis archesia, 

 a Haonanitmida dcedalus, a Pyramieis, a Precis cehrene, and a Sarangesa. 



She then readily ate a Charaxes candiope, a Gharaxes hrutus, a Precis cehrene, and 

 a Byhlia, but crushed and rejected the next Byhlia and the $ Pseudacrcea lucretia. 

 She was, however, probably now replete, as evidenced by the fact that she per- 

 sistently refused without tasting Charaxes brutus, Hamanumida dcedahis, Precis 

 cehrene, and Latnpides bcetica. 



[Summary : — 



(1) After M. agathina was rejected the bird ate 1 Melanitis leda, 1 C. neanthes, 

 4 C. cithmron, 5 C. brutus, 5 C. candiope, 5 Byhlia, 1 E. dryope, 1 E. hiarbas, 

 1 A. phalantha, 2 P. cardui, 2 P. cehrene, 2 P. archesia, 1 P. ceryne, 2 H. dcedalus, 



1 B. mesentina, 1 j -P- isokani, 1 c? and 1 5 Z. thalassina, 1 P. de'modocus, 1 P. lyceus, 

 and 1 Sarangesa [M. agathina rejected 40 places, including 14 large Charaxes, from 

 repletion-point). 



(2) After rejecting a Terias she ate all these less the B. mese^itina and 2 C^. cithoi- 

 ron (Terias rejected 37 places, including 12 large Gharaxes, from repletion-point). 



(3) After rejecting the S and § P. lucretia she ate all in list (1) except the 

 Belenois and the 2 C. cithceron just mentioned and except 2 C brutus, 2 (7. candiope, 

 and the P. isokani thorax (P. lucretia rejected 32 places, including 8 large Gharaxes, 

 from repletion-point). 



(4) After rejecting M. campina she ate list (1) with the exceptions already 

 mentioned and the further exceptions of 2 Byhlia and the P. demodocus ( = 29 places). 



(5) After rejecting the last of the main lot of Byblias she ate 1 M. leda, 1 G. ne- 

 anthes, 2 C. cithceron, 3 G. brutus, 3 G. candiope, 1 Byhlia, 2 P. cardui, 2 P. cehrene, 



2 P. archesia, the P. ceryne, 2 ^. dcedalus, and the Sarangesa ( = 21 places). 



(6) After rejecting ^. perspicua (so waveringly that she may have been nearly 

 hungry enough for it) she ate 1 G. cithceron, 2 C. brutus, 1 G. candiope, 1 P. cardui, 

 2 P. cehrene, 1 P. archesia, 1 if. dcedalus, 1 Byhlia, and the Sarangesa (=11 places). 



(7) After rejecting P. lyceus she ate these less 1 C. hr%ittis and the C. cithceron 

 ( = 9 places including only 2 large Charaxes). 



(8) Terias senegalensis probably not liked less than M. agathina ; P demodocus 

 than ilf. campina ; P. lyceus than Ps. lucretia ; the last Byhlia eaten not less than 

 (7. brutus, M. dcedalus, or P. cehrene. Hesitation over Z. thalassina $ , probably 

 due to its Mylothris-like appearance. 



(9) By disregarding throughout the experiment the possibility of special stimu- 

 lation (such as probably caused Byhlia to be eaten to repletion-point) one might 

 deduce much from this experiment. Taking it into account, all that can fairly 

 safely be said with regard to preferences is that P. archesia was apparently pre- 

 ferred to P. lyceus, and, with G. brtittcs, to II. perspicucc (which on manner may 

 have been preferred to M. cccmpina) ; with G. hrutus and M. leda to " unstimu- 

 lated " Byhlia ; with G. brutus, M. leda, and A . phcdantha to P. hici-etia ; with these 

 three and Z. thalassina <S to M. campina; with these four and P. demodocus to, 

 again, Byhlia and P. lucretia, also to Terias and perhaps P. isokani ; and with these 

 five and B. mesentincc to M. agathina.^ 



