440 NATURAL SCIENCE. June. 



Larvae of C. verbasci were taken feeding exposed on the upper 

 side of leaves of mullein ; they were green, yellow, and black, very 

 conspicuous, and not hairy. One was offered to the jackdaw when the 

 bird was very hungry ; it was refused at first, then tasted once only, and 

 dropped with greater signs of disgust than in the case of D. caeruleocepliala. 

 After this he would not take anything at all from me for a considerable 

 time, and appeared very uncomfortable. 



In the case of Acronycta psi and Bombyx rubi the animals ex- 

 perimented on were three slowworms (Angiiis fmgilis) and one lizard 

 [Zootoca vivipara). These were purposely kept very hungry, but though 

 the larvae were left with them for three days they refused them 

 entirely, never even attempting to taste them. Both these larvae, 

 however, were hairy as well as conspicuous, and, therefore — assuming 

 that conspicuousness means unpalatability — were doubly protected. 

 The fact that an animal may possess more than one kind of protection 

 must of course be taken into account in making these experiments, 

 and for this reason the experiments with D. caei'ulcoccphala and C verbasci 

 were the most satisfactory, because here the unpleasant attitude was 

 almost certainly taste only. Mr. Poulton says [Pyoc. Zool. Soc. Loud., 

 March i, 1887) that Mr. Jenner Weir found the two last-named 

 larvae to be disregarded both by birds and lizards. This certainly 

 supports Professor Wallace's original suggestion of the unpalatability 

 of conspicuous larvae to " some, at least, of their enemies," while, on 

 the other hand, the results obtained by me also favoured Mr. Poulton's 

 suggestion as to hunger putting a limit to this method of defence, 

 since the larvse were tasted, and tasting would be as fatal to them as 

 eating. 



Mr. Beddard (" Animal Colouration," p. 164) found that A . psi 

 was eaten by Lacerta viridis, and at least tasted by a thrush, but he 

 does not mention whether either was hungry on these occasions, a 

 point of some importance. In any case, his results differed from 

 mine in this experiment, but then a different species of lizard was 

 used. L. viridis would seem, from other experiments of Mr. Beddard's, 

 to be less sensitive to unpleasant attributes than other lizards, 

 Z. vivipara, for instance. This difference of behaviour does, as Mr. 

 Beddard points out {loc. cit., p. 155), show that unpalatable animals 

 with warning colours are not always exempt from attack, but this 

 very fact rather tends to support the view held by Mr. Poulton that this 

 means of defence can only safely be adopted by a limited number. On 

 the other hand, it militates strongly against Dr. Eirig's theory, quoted 

 by Mr. Beddard, that brilliant colours (i.e., abundant secretion of pig- 

 ment) are the cause of inedibility. Exceptional cases like that of 

 L. viridis esiting Acronycta psi are difficult to explain on this hypothesis, 

 but they appear more comprehensible on the theory of limitation. 

 A . psi might be regarded as one of the few examples of the failure as 

 a means of defence both of the unpleasant attribute and its adver- 

 tising warning colours, and L. viridis as having had its power of 



