232 ME. C, F. M. SWYNNERTON ON THE 



the bird to the various species of butterflies, his preferences in which I hoped to 

 ascertain later. That the experiments proceeded no further was due to his 

 premature escape. 



C. COBACIAS GARSULUS. B. EaRLY EXPERIMENTS. 



1909. Mar. 7. Another adult European roller was limed and brought to me 

 this afternoon. 



[There is very little to be deduced from the following fifteen experiments — to 



]y[g_y 18 though they mostly afford illustrations of the phenomenon commented 



on for Roller A on March .3. In any case, neither they nor those on A were 

 intended as preference experiments. These may perhaps be regarded as com- 

 mencing, for this bird, on May 20. Those of April 1, 9, and 15 were "mimicry" 

 experiments. 



In connection with those experiments that commence with " Plenty of food 

 uneaten in cage," I may remark that I had probably at that period still to realize 

 that this roller (unlike some of my other birds) frequently allowed herself to 

 o-et moderately empty before going down again to feed, and that abundance of 

 uneaten food however pleasant, did not therefore necessarily indicate approximate 

 repletion.] 



Exp. 6. Ma7\ 9. Crushed and swallowed a wingless Pajnlio lyceus, refused the 



next but accepted it three minutes later, and in this way ate seven in all with a 

 refusal and about a three minutes' interval between each. [She was evidently 

 throughout barely hungry enough for them, a few minutes' digestion after each 

 refusal rendering her just hungry enough for another acceptance.] 



Exp. I.—Mar. 10. Crushed well, and, though she had not appeared to be 

 hungry, ate with apparent relish a Charaxes zoolina, a Charaxes guderiana, and 

 two Charaxes candiope, accepted with evident distinction a Charaxes brutus natal- 

 ensis, crushed it slightly and dropped it. On my reofi"ering it ten minutes later 

 she ate it readily, as also a second and a Charaxes neanthes. 



[In concluding my account of this experiment I stated that I was left with the 

 impression that the Charaxes were more enjoyed than the Papilios of the day 

 before, I must confess, however, that there is nothing in the two experiments 

 as recorded that necessarily indicates this, seeing that in neither case was the state 

 of appetite definitely noted.] 



jjxp. 8. Mar. 23. Had plenty of food in the cage and did not appear in the 



least hungry, but readily ate, after crushing each, five Hypolimnas misippus in 

 rapid succession (a S and four $ ). I now left five or eight minutes to allow for 

 after-efi'ects and ofi"ered two more $ ? , both of which were readily accepted and 

 eaten. Each butterfly in this experiment had one hindwing attached to ensure 

 recognition by the bird. 



[The same applies to all the following experiments, unless otherwise stated. 

 The short interval for after-efi'ects was a habit of mine at that time, plausible 



