EXPLANATION OF FORM AND COLOURING. 255 



birds as part of their ordinary food, they were usually eaten only when (1) the bird 

 was hungry, or (2) I had first reduced them with a hammer. Under the latter 

 circumstances they have been eaten by various birds up to near repletion-point, and 

 it is probable that their great hardness and size and spiky legs are a very needed 

 as well as a very efficient protection. Their habits are not specially procryptic. 



The very prompt rejection of Neptis saclava, on taste, after a quite unsuspecting 

 acceptance, is highly convincing. Preferred to it were Gharaxes hrutus and perhaps 

 Gharaxes candiope, also Precis cehrene. The last-named butterfly was apjaarently 

 regarded as better worth considering than Atella phalantha.] 



Exp. 36. — Jione 6. Extent of hunger uncertain ; was feeding on grasshoppers 

 and termites when I commenced the experiment. Readily ate Precis c'ibrene, two 

 Gharaxes bruius, a Henotesia perspicua, a third Gharaxes hrutus and a Gharaxes 

 candiope, refused, but immediately afterwards changing her mind, accepted, crushed, 

 and readily swallowed a Neptis saclava and a Mycalesis campina, the latter with 

 the usual chocolate smell, ate four Gharaxes hrutus, and tliereafter persistently 

 refused to touch Heptis goochi. Neptis agatha was, however, accepted, crushed, 

 and eaten with apparent readiness, but Neptis goochi was once more refused 

 without tasting. She then readily ate two Henotesia perspicua, a Gharaxes hrutus, 

 and a Gharaxes candiope, and tasted and rejected an Ypthima. On my reoffering 

 it, it was again properly tasted, and this time flung to the other side of the cage. 

 Precis cehrene, Atella phalantha, and a Mycalesis campina were now eaten without 

 hesitation, but there was some delay between each of these last offerings as I had 

 exhausted my prepared material and had to hunt up each separately in the 

 envelopes and remove the wings. 



[Summary, &c. : — (1) After N. goochi was refused, 1 N. agatha, 2 H. perspicua, 

 1 G. hrutus, 1 G. candiope, and 1 P. cehrene were eaten, the two Gharaxes perhaps 

 in part as the result of stimulation. 



(2) After the rejection of Ypthima, P. cehrene was eaten. Whether Atella phalantha 

 and (especially) Mycalesis campina could be safely added to the lists of " preferred " 

 butterflies would depend entirely on the length of the delay before each was ofiered, 

 quite apart from any appetizing influence that the P. cehrene may have exerted. 



(3) Comment at the time. — " The roller's constant refusal of Nejitis goochi without 

 tasting seems significant — cf. past experiments. The Ypthima had the usual faint 

 smell of lemonade."] 



Exp. 37. — June 7. The bird was hungry. ISTot having many Gharaxes I gave 

 her a number of grasshoppers of many species. She then tasted and rejected 

 Neptis saclava, but readily ate Pi^ecis cehrene and Neptis agatha. After eleven 

 more small gi'asshoppers she readily, to my surprise, ate the Neptis saclava. I then 

 gave five large grasshoppers. The bird had some difiiculty in swallowing the last 

 two, and they could be plainly seen in the back of her throat whenever she opened 

 her bill. A Neptis saclava was now persistently refused without tasting, a Pyrameis 

 was recognized and readily eaten, a Byhlia, smelling slightly of civet (dark hind- 

 wing with pearl spots), was tasted and thrown away with evident dislike, but a 

 § Leuceronia argia was eaten, as also a Gharaxes candiope — the latter slowly, after 



