1887.] VALUE OF COLOUR AND MARKINGS IN INSECTS. 193 



enemy, but palatable to another ; and to this extent Wallace does 

 point out a limit to the application of this principle of defence. But 

 the counterbalancing limit which I suggested is of course entirely 

 different, for I argued that a Vertebrate enemy may be forced by 

 stress of hunger to eat an insect although unpalatable to it. 

 Although the latter limit is thus quite distinct, it would certainly in 

 time become identical with the former, as the distaste for the insect 

 gradually disappeared after it had been repeatedly eaten. In fact it 

 will be shown to be probable that many (if not all) of the instances 

 in which an evidently distasteful insect is eaten by certain Vertebrates 

 originally rose in this way. These suggested additions to Wallace's 

 theory of protection by warning colours were capable of being put 

 to the practical test. To achieve this object it was only necessary to 

 ascertain whether an insect-eating Vertebrate could be induced by 

 hunger to eat a gaily coloured and conspicuous larva which it was 

 always known to refuse when other food was present, and which was 

 evidently very much disliked on the few occasions of preliminary 

 " tasting," which would always occur long before the time when the 

 disgusting morsel would be reluctantly swallowed. I shall presently 

 show that my suggestion was in every way confirmed by the test; 

 but before giving an account of my own experience I will allude to 

 all the previous experiments which have been made in support of 

 Wallace's theory. 



I. Brightly Coloured or Conspicuous Larva. 



At a meeting of the Entomological Society of London (see Proc. 

 Ent. Snc. ser. 3, v. p. Ixxx, 1867) Wallace made his important 

 suggestion as to the biological value of conspicuous and gaudy colours 

 in caterpillnrs. It is obvious that the question of the value of such 

 colours in the larval stage is almost the same as in other stages, 

 and it was chiefly from the determination of the use in the latter case 

 (due originally and principally to Bates) that Wallace suggested that a 

 similar solution would be found to apply to the former also. Never- 

 theless there are reasons why such a niethod of defence is especially 

 applicable to the laival stage. I have shown that there is a spe- 

 cial reason in the anatomical construction of larvae which explams 

 why these orsanisms require to be deieuded from slight injuries 

 (see Trans. Ent. Soc. Loud. 188.5, pt. ii. Aug. pp. 321-323). A 

 larva " may be described as a soft-walled cylindrical tube which 

 owes its firmness, and indeed the maintenance of its shape, to the 

 fact that it contains tluid under pressure. The pressure is exerted 

 by the muscular parietes of the body. The advanta'.;e of this 

 coiii-truction is as obvious as iis danger; the larva possesses a motive 

 force which can i)e applied to any movable part of the surface 

 through the medium of thefluid.". . . " Thiscon-truction isextremely 

 dangerous ; for a slight wound entails great loss of blood, while a 

 moderate injury must prove fatal. The larvae of Smerinlhus 

 ocellatus (and many others) nibble off each other's horns, and the 

 wounded larvae (although they do not seem to be aware of the injury) 



