346 Mr. G. A. K. Jklarshall on 



eye-spots, especially upon the under-side of the wings, as 

 directive marks leading an enemy to attack a non-vital 

 part, and they tend to refute Portschin ski's explanation 

 of them as the representation in colour of drops of some 

 specially-protective fluid (see p. 398). 



Butterflies of different groups, Hcsperidx, Picrinx, 

 NyriqjhcdincV, were freely eaten, but the rejection of the 

 abdomen of Pccpilio corinncus by the captive bird which 

 afterwards ate Bylilia and Tcrias, can only be explained 

 on the supposition of unpalatability, and the same was 

 evidently true in a more marked degree of Acrseas and 

 L. chrysiirpns, although parts of these would sometimes be 

 eaten, while on Feb. 15, after starvation for twenty-four 

 hours, many Acra?as were swallowed whole. The behaviour 

 on this occasion renders it certain that, as in the case of 

 Bucorax coffer, L. chrysipijus was far more distasteful to 

 the kestrel than the Acrteas. 



The rejection, after trial, of the evil-smelling Coreid 

 bug A. airvipcs, greedily eaten by baboons, is a good 

 example of the difference in value of the same defence 

 with different enemies. 



The treatment of Coleoptera almost invariably supported 

 the theories which explain the meaning of insect colouring 

 as cryptic, warning, etc. 



The following beetles were eaten by the kestrel : 

 Curculionid^e, with cryptic colouring [Oosomus, sp., and 

 Eremnus, spp.) ; the large, slow-moving, conspicuous, black, 

 earthy Heteromeron Anomaliinis plehcius, when the chitin 

 was broken; the smallish Buprestid Amdlystcrna vitti- 

 jjcnnis (dark metallic green or coppery with white stripe 

 on each elytron); the Heteromeron Dichtha injlcda, dark 

 brown with reddish stripes, conspicuous and slow-moving 

 WkeAnomcdijnis; the medium-sized Scarabeid Onitisalexis 

 with elytra and legs brown, and thorax iridescent green. 



It is probable that most of the defensive fluid had been 

 already discharged in the case of the Carahidai of the 

 genera Piczia, Polyliirma, and Grcvphii^dcrus, of which the 

 acid secretion was seen to be a very positive protection 

 when there was opportunity for its operation on a normal 

 scale. The Longicorn Ceroplcsis fidlax with a Cantharid 

 type of colouring may be synaposematic, as it was only 

 eaten very slowly although the bird had been kept without 

 food. The impression produced by the stridulation is of 

 much interest (see p. 403). 



