234 MR. C. F. M. SWYNNERTON ON THE 



cautiously tasted it, promptly rejecting it. She then refused, without tasting, 

 Precis natalensis f . and Pseudacrcea trimeni, but accepted and ate a Catopsilia 

 Aorella with the greatest readiness. All the butterflies in this experiment were 

 offered with all wings attached and upper surface fully displayed. 



[The bird certainly appeared to differentiate between the Acrcea acara and the 

 Pseudacrcea, though less sure, apparently, of the latter and Acrcea areca. It is 

 possible that Precis natalensis escaped through its rough general likeness to 

 Acrcea acara. The two Acrseas were evidently definitely disliked — at any rate, 

 relatively to the only butterfly eaten, G.Jiorella. In this case no complication 

 arises in connection with the presence or absence of wings (as happened in the 

 similar experiment on Roller A). G. Jlorella, the stiffest-winged of all, was 

 swallowed, with its wings, at the very end of the experiment.] 



Exp. 13. — Apr. 4. Plenty of other food in cage. Ate with the greatest readiness 

 and no sign of dislike three Papilio dardanus 5 f- hippocoon, each with one 

 hindwing attached. 



[As the roller's state of appetite was not definitely indicated, nothing useful can 

 be deduced from this experiment.] 



Exp. 14. — Apr. 5. Independent Experiment by my Wife. — Each butterfly, as 

 usual, had one hindwing attached. The bird ate readily six Precis celrene in rapid 

 succession. My wife now waited five minutes for possible after-taste or after- 

 effects to make themselves felt, and offered a seventh. It was recognized with 

 evident pleasure and swallowed with as much apparent relish as its predecessors. 

 The bird now ate in rapid succession six Precis clelia, showing actual and increasing 

 eagerness and craning out for each butterfly as it appeared near the bars. After 

 the usual short interval a seventh was offered. The bird, evidently recognizing it, 

 reached out, snatched it from the forceps, and ate it with evident pleasure. 

 She now readily accepted and ate two Antanartia schoeneia, seven Belenois, and 

 three Scdamis anacardii, all in rapid succession. 



The third Salamis was accidentally dropped from the forceps and the bird 

 appeared quite concerned, leaning well over and gazing down at it. It was quickly 

 picked up and reoffered, and this time eaten rapidly. A fourth Salannis, dead this 

 time but not appreciably dry, was now offered, but the bird dropped it two or 

 three times, each time after crushing it. She ate it, however, readily on its being 

 reoffered fifteen minutes later. After this she readily ate two Eurytela hiarhcis. 



[The bird was probably fairly hungry at the commencement, but, seeing that 

 there was no rejection early in the experiment by which one might have gauged 

 her state of appetite, all that can be deduced is that, when hungry enough to 

 do so, the roller would eat with every sign of relish the butterflies mentioned, 

 including Belenois and Precis, in considei-able numbers. There was nothing to 

 indicate what grade any of the butterflies fall into, or how they stand relatively 

 to one another.] 



Exp. 15. — Apr. 9. A Mimicry Experiment. — The roller tasted and promptly 

 flung away Amawis ochlea and refused it without tasting on its being reoffered, 



