368 mr. c. f. m. swynnerton on the 



Experiments on Coeacias oasbuzus. C. 



Exp. 215. — March 7, 1911. A just-captured roller (C. garrulus) was brought to me 

 this morning. By the afternoon he was already so much at ease as to readily 

 accept insects from the forceps. I gave him a Rhahdotis larva and a number of 

 grasshoppers, after which he refused to touch an A. caldarena, ?k D. ckrysippus, an 

 A. esehria or an M. rueppelli, but readily accepted and ate a Padraona zeno, refused 

 a Terias senegalensis, a Lr/o. lemnos, a M^c. cavipina, and a Creiiis boisduvali, these 

 last two offered with upper surface displayed, but accepted readily and ate a 

 Leuceronia thcdassina S , refused a (S Lycus sp., again refused the Mylothris and 

 the Terias, also a Precis natalensis Q f., readily took, crushed, a,nd ate an 

 A. schceneia and after it to my surprise swallowed down the Myc. campina, but 

 seemed annoyed and refused to touch another or the Crenis re-ofiered, even with the 

 underside up, refused a second Leuceronia thalassina and again the Lycus, accepted 

 a little doubtfully but on crushing it readily ate an Andronyomos 'iieander, refused a 

 P. demodocus and a P. lyc^us, and on my persisting in oflering them took each fiom 

 the forceps and without tasting them dropped them on the ground, but readily ate 

 a Gyligramma latona, then took and dropped a Rhahdotis larva and a small grass- 

 hopper, tried and rejected Charaxes candioj^e and Sphingomorpha chlorea. The 

 Charaxes and the S. chlorea were each simply snatched from the forceps, crushed and 

 di'oppedand a grasshopper was similarly treated ; but, judging from the preceding- 

 portion of the experiment, the bird must in any case have been nearly replete. 



{Order : — 



. 7 • (I.e. latona and A. neander. 

 A. schceneia. { 



2. L. thalassina (S . 



Lycus sp. S 1 

 G. boisduvali. 

 Precis natalensis © f . 



3. A. caldarena, A. esebria, D. chrysippus, M. rtieppelli, 

 T. senegalensis, Lycanesthes lemnos, M. cwnijnna (and, 

 (or whatever it was I if not taken for M. campina, C. boisduvali). 



taken for). |^ 



G. latona also over P. demodocus and P. lyceus and quite as high as the grasshopper, 

 G. candiope, S. chlorea, and the Rhabdotis larva. 



P. zeno over at any rate the first four species in Grade 3 above. 



But the great point of the experiment was that a bird captured that very day 

 and without previous experience of them in captivity should have given proof of 

 its experiences in the wild state by refusing to touch a Danaine, two species 

 of Acrcea, a Mylothris, a Terias, a Lyccenesthes, a Mycalesis, and a Lycus, and, when 

 nearly replete, a P. demodocus and a P. lyceus — all of them more or less disliked by 

 my long-captive roller, — while readily accepting and eating species that my long- 

 captive roller would also under those conditions of appetite have readily eaten. 

 The refusal of the Precis natalensis was probably due to its ^c'/'OB«-like appearance, 

 and the jSrst refusal of the Grenis seemed due, possibly, to its likeness to 

 31. campina, but its subsequent refusal when the underside also was shown may 

 have indicated that, if known, it was itself then unacceptalole.] 



