The Bionoiihics of South African Inaccts. 23 



they are not likely to be released. But in the case of 

 inedible species it is different. For if my surmise is true, 

 that insectivorous invertebrates are not capable of appreci- 

 ating warning colours but have to taste all their captives 

 before being able to tell whether they are edible or not 

 (which I think is clear from my experiments), then tenacity 

 of life (as a protective agency) will be as useful an 

 acquisition against invertebrates as warning coloration is 

 against vertebrates, and come into play when the latter is 

 useless. Of course tenacity is of use against the experi- 

 mental tasting of young birds, lizards etc., but this does 

 not seem to me to be a sufficiently cogent factor to 

 develop the power to such a high pitch. For if the 

 insects had only these enemies to contend against, even 

 supposing every specimen experimentally tasted died from 

 its injuries, the protection afforded by the warning colours 

 would still be ample. Indeed I believe that the tough- 

 ness of inedible insects has been primarily developed to 

 counteract the injuries from invertebrate foes (which are 

 incapable of reasoning as to whether an insect is edible or 

 not), and that therein lies its chief utility, though it may 

 prove useful incidentally in other cases. 



A. luxtoni appears to have more juice in proportion 

 than liorta, and I regard it as a more highly-developed 

 species, from a distasteful point of view, in that it can 

 exude juice at will from its thorax, and thus show its 

 nauseous qualities without necessarily having to be injured 

 like horta. When squeezed the juice often oozes from 

 the ends of the antennae and all the nervures of the 

 fore-wings when they are cut. But, as I have pointed 

 out {vide Expt. 21), the results of experiments with it 

 are unreliable. 



The treatment of A. horta b}^ the spiders would almost 

 give some colour to your suggestion that the inedibility of 

 species may be due to unpleasant internal effects rather 

 than the mere taste, for B ate one specimen and A, C, and 

 D two each before they seemed to become aware that it 

 was not good to eat, from which I should conclude either 

 that the unpleasant effects are subsequent to eating or 

 that their sense of taste is not sufficiently acute to recog- 

 nize a nasty flavour at once. But the latter conclusion 

 appears to be invalidated by their prompt rejection of L. 

 chrijsipptts and A. cchcria. Anyway their selection seems 

 to show that there are grades of unpleasantness, and, as I 



