228 MR. C. F, M. 8WYNNERT0N ON THE 



suitable species on which to carry out what was, perhaps, my most important 

 series of butterfly experiments. 



In view oF its relative importance I would like again to emphasize the fact 

 that from the outset I studied the matter of food and spared no pains to 

 provide my captive rollers with a thoroughly varied diet on as natural lines as 

 I could devise. A reference to the list I have already given will show, I 

 think, that I was probably not entirely unsuccessful in this. 



2. The Experiments. All three of my rollers (A, B, and C) were from the 

 outset remarkably tame and fearless, showing relatively little alarm at passing 

 dogs or servants, and accepting food from the forceps almost at once. They 

 were lethargic birds. B, in particular (the bird on which most of my experi- 

 ments were carried out), had a very large cage and plenty o£ perches, but 

 excepting at the northward migration-time (when she sometimes \mis restless 

 at night) she took very little exercise, remaining as a rule quietly perched in 

 one place, excepting when she descended to eat or drink or shifted her 

 position with the sunlight, of which she was very fond. This was, of course, 

 a pity, and it was quite likely, owing to this lack of exercise, that she relatively 

 seldom, even after a fast, developed so ravenous an appetite as to eat Acrceince 

 readily. She ate heartily, nevertheless, and kept in excellent health and 

 condition throughout. 



All three birds expressed their sentiments with regard to the offerings 

 made to them with, usually, a lack of ambiguity that rendered them exceed- 

 ingly satisfactory subjects for the purpose in hand. All three also appeared 

 to be exceedingly cautious tasters and, though often emphatic enough in their 

 refusals, seldom swallowed an insect, however small or however well-known, 

 without first more or less crushing it (often running it through their bill from 

 side to side as they did so). That is to say, very few insects were eaten 

 witliout a [)reliminary tasting, the bird maintaining a highly judicial calmness 

 and precision of demeanour that tended to inspire the onlooker with a certain 

 amount of contidence even while it amused him. 



B. Experiments on Coeacias gabbulvs. A. 



1909. Feb. 23. An adult European roller brought me by a woman who had, 

 she said, just limed it. Accepted food from the forceps. 



Exp. 1. — Feb. 24. Offered the common mud-coloured beetle, with hard elytra, 

 the roller crushed it very slightly with one bite of the point of his bill, and at 

 once swallowed it down, tried and at once rejected a Neptunides polychrous, 

 shaking his head after doing so (the same individual had already been rejected by 

 Centropus burchelli, A), crushed a few times with the point of his bill, and finally 

 rejected a Papilio lyceus with wings, and, on my offering a second without wings, 

 masticated it as before and swallowed it. No sisrn of dislike was shown, but a 



