to limitations of the Milllerian Hypothesis of Mimicry. 141 



The rejection of the first specimen of the scarlet-tipped 

 Teracolus achine has likewise no significance, for on several 

 occasions subsequently they ate fourteen similarly coloured 

 specimens of this genus " with undoubted appreciation." 

 Again on ofiferinsr a larva of Bcleiiois mesentina to the 

 female baboon she was clearly afraid of it and tasted it 

 most cautiously, yet on finding it all right she " eagerly 

 devoured over twenty of them " (/. c. p. 885). Yet the 

 male, which was far more timid and suspicious in tempera- 

 ment, "was much too frightened to eat them in spite of 

 the example of the female." Had these larvae been 

 offered to the male only, it is probable that an entirely 

 erroneous conclusion as to their extreme unpalatability 

 would have been deduced from his behaviour. Cases of 

 this kind might be easily multiplied, as, for instance, 

 where unsuitable insects are offered, or where suitable 

 insects are given to over-fed animals, and so forth. 



The fact remains that even when reasonable care is 

 exercised there must be many opportunities for error in 

 estimating degrees of palatability or the reverse. Yet 

 it seems a fair contention that, in the absence of obvious 

 mistakes or inaccuracies, the true value of such ex- 

 perimental evidence can best be assessed by the man 

 who actually carries out the experiments. He will be 

 fully cognisant of all the conditions under which the ex- 

 periments were made ; he will presumably be acquainted 

 with the special idiosyncrasies (often a very important 

 point) of those animals upon which he experiments ; 

 finally, he alone is able to see and compare those subtle 

 differences of manner, facial expression, etc., which con- 

 stitute our only guide in estimating the psychological 

 effect produced upon the animal. If similar experiments 

 are repeated by a number of other observers the chances 

 of error will be much diminished, and we shall thus obtain 

 a solid foundation of fact upon which to build up those 

 broad generalisations after which we are seeking. 



A final point may be noticed. It has been suggested 

 to me that perhaps we may be unable to demonstrate by 

 experiment these incipient degrees of distastefulness which 

 have been postuhited for many genera of NymiphaliniB and 

 Pierinm. But if these unpleasant flavours are so slight 

 that insectivorous animals in captivity are unable to 

 appreciate them, it seems scarcely probable that they 

 will have any real selection value, or that they will be 



