EXPLANATION OF FOKM AND COLOURING. 243 



Ghrysippus ; ate all the first P. cardui, P. cebrene, C. brutus (3), C. neanfhes, 0. ci- 

 thceron, and C. candiope (2) after refusing Neptis saclava, Neptis agatha, and 

 Eurytela hiarhas; C.hrutus{Z), G. neanthes, G. cithceron (1), and G. candiope (2) 

 after refusing, waveringly, G.florella; and the three largest Gharaxes (1 citJiceron, 

 2 cfMifiiope, and 3 brutus) after refusing :Z^isHas seneg(densis,Glutophrissa saba,Belenois 

 severina, and, possibly from suspicion of its white coloration, Gharaxes zooli7ia. 



From the wavering nature of the Gatopsilias, rejection it seemed not unlikely 

 that it too would have been eaten nearer repletion than either species of Neptis, 

 the Eurytela, the Mylothris, the Nyctemera, or the Danaida. The fact that the 

 last three species were refused no less than 17 places away from repletion (11 of 

 these places being occupied by such bulky insects as Gharaxes brutus, cithceron, and 

 candiope) would seem to indicate that the roller placed them very fairly low. The 

 great contrast in manner might be taken to suggest that Neptis agatha, the 

 Leuceronia, and the earlier Belenois would have been placed higher than they, but 

 all, I think, that can be said with any safety is that the Leuceronia at all events 

 was not placed lower than the Mylothris, etc. It and its companions should have 

 been reoffered after the latter's rejection. 



N. agatha, N. saclava, and E. hiarbas were rejected at 9 places from repletion 

 (the 9 including 6 large Gharaxes), the Catopsilia at 7 (quite unusually low for it). 

 The G. neanthes was at any rate not below the later Eelenois, the G. saba, and the 

 Terias senegalensis, and these three insects were refused at 6 places from repletion. 

 The 3 large species of Gharaxes were at any rate not below P. cardtd, P. cebrene, 

 P. archesia, G. neanthes, or any species of grasshopper that I was able to offer. 

 The latter point is interesting in relation to experiment 25. 



Last part of experiment : P. cardid and P. cebrene above the Byblia, A . encedon, 

 P. angolanus, P. lucretia, N. saclava, and G. saba ; G. neaiithes and P. archesia 

 above the two last-named insects and P. clelia not below them. G. nea^ithes and 

 P. archesia not below P. angolanus or P. lucretia*. 



Exp. 37. — May 25. Had been allowed no food since a good meal, possibly three 

 hours before, and semed distinctly hungry. She readily ate Neptis goochi, refused 

 emphatically and persistently without tasting a Terias senegalensis, then stretched 

 out eagerly for Neptis goochi, crushed and at once swallowed it, readily ate, after 

 crushing it, Herpcenia eriphia and, with some hesitation, the previously-refused 

 Terias (now stripped of wings), ate a Nychitona medusa and Phrissura isokani J 

 (the last with a front wing attached ; all the others, unless otherwise stated, with 

 one hindwing), then tasted and rejected Mylothris yulei with very evident dislike, 

 but readily ate a Rhopalocampta with distinct Amauris-^uieW. The eating of the 

 Nychitona and Terias was probably the result of immediately preceding stimulation. 



I now gave her a Gharaxes brutus natalensis and one Gharaxes candiope, after 

 eating which she tasted and very readily swallowed Neptis goochi and refused very 

 persistently without tasting the wingless body of Terias senegalensis, once leaning 

 forward and half tasting but drawing back without attempting to remove it from 

 the forceps. She crushed, however, and readily ate Phrissura isokani and a 



* Some of the acceptances iu this experiment may have been the result of stimulation: I have 

 made the necessary allowances for this in carrying out these results to the final list. 



