204 MR. E. B. POTJLTON ON THE PROTECTIVE [Mar. 1, 



true of gregarious larvae. Thus a group of phytophagous Hymeno- 

 pterous larvfe may remiiiii inconspicuous while uniisturbed, but 

 nevertheless the approach of an enemy determines united moyemeats 

 in the colony which render the whole strikingly conspicuous, and which 

 may be attended later by the emission of au offensive smell from the 

 numerous ventral glands of all tiie individuals simnltaneonsly (e. g. 

 Croesus septentrionnlis). In tiie other larvse which suddenly assume 

 a terrifying attitude " the effects produced approximate somewhat to 

 an intensely exaggerated caricature of a sort of generalized vertebrate 

 appearance, probably of the serpent type (at any rate in C/icerocampa), 

 such as would be most efficacious in the case of birds. It is likely 

 that the terrif\ing appearance of our own larvye in temperate 

 latitudes first arose in tlie tropics, where the imitated cause of alarm 

 to the enemies of the larva is real and obvious. And it is probable 

 that the success of the same method in countries where tiie reptilian 

 fauna cannot be said to constitute a source of alarm is due to 

 inherited memories of a tropical life which live on, as that instinctive 

 fear of anytliing snake-like which is so commonlv exiiibited by the 

 higher land-vei tebrates including ourseKes." (Poulton, Trans. £nt. 

 Soc. Lond. 1886, pt. ii. June, pp. 15(), 157). 



The success of this combination of defensive measures depends on 

 the extraordinary sensitiveness of the larvge, so that the tiansitiou 

 from the one method to the other is instantaneous, and in the c;ise 

 of the suddenly assumed terrifying attitudes, the enemy is additionally 

 alarmed by tlie way in which some dreaded ohject seems, as it were, 

 to spring into exi-tence. It is very unfortunate that so few expe- 

 riments have been made upon this most interesting group of 

 larvge. 



Just as it was sn^gested that insect-eating Vertebrates might, 

 under the influence of hunger, be induced to eat and finally to relish 

 distasteful larvae, so we must expect that the same cause would in the 

 end prevent this elaborate system of intimidation from being success- 

 ful. In this case, however, there is no prejudice against an unplea- 

 sant taste or smell to be overcome, and it is most probable that the 

 larvae would be in great danger as soon as the imposition was detected. 

 It is perhaps on this account that these methods are adopted by an 

 exceedingly small proportion of larvae, but also because a certain size 

 is necessary for any chance of success. Nevertheless this size is 

 less than might be anticipated, for the anterior part of the body with 

 large eye-like marks is generally swollen out into a resemblance to 

 the head of a serpent, while the larval body is partially concealed 

 among the leaves of the food-plant, and, in many positions, what is 

 seen merely serves to suggest a far more extended length than that 

 which actually exists. Wallace has suggested that it is very probable 

 that the " spectacles " of the Cobra are terrifying marks, which warn 

 the enemy against approach, and it is most interesting to note that 

 the Chcerocampa-lair\se mimic the terrifying eye-like marks of a Cobra- 

 like serpent, and not the real eyes of a serpent, which are relatively 

 small. (Table II., pp. 206, 207). 



Having thus tabulated the results of experiments upon undoubtedly 



