244 THE VARIATION OF ANIMALS IN NATURE 



enemies (iguanas, turkeys and peccaries) the warning colours 

 can be of no avail. Cuenot (i 921, p. 512) has further objected 

 that divers noxious forms (toads, vipers, torpedo fish) have 

 a homochromatic coloration. Conversely we find it very 

 difficult to obtain evidence that the striking or brilliant colours 

 of (e.g.) many of our British slugs have any ' warning ' value. 

 Cuenot (I.e. p. 513) makes the suggestion that the conspicuous 

 colours of venomous forms may simply be due to the fact that 

 the owners are otherwise well defended, either by their powers 

 of flight (reef fishes) or by their hidden life (Elaps), and their 

 conspicuous colours are not disadvantageous. 



On the whole we have to admit that the frequency of con- 

 spicuous colours among noxious animals is high enough to 

 require explanation, and that the ' warning ' hypothesis is not 

 to be lightly dismissed. We think, however, that a good deal 

 more exact investigation (e.g. of the frequency of the correla- 

 tion) is needed, and in particular far more knowledge as to 

 whether ' warning ' colours are actually avoided by predators. 



It is very probable that in some of these animals the warning 

 colours have an important function in saving the bearer from 

 unnecessary attack. But there is probably an equal or greater 

 number of examples where one or other feature of the associa- 

 tion is lacking, and there has been a tendency to assume that 

 brightly coloured forms must be protected without any very 

 good evidence as to whether they are actually preserved from 

 attack. We may consider as an example the Heteropterous 

 bugs of the family Pentatomidae. A number of species 

 (e.g. European species of Graphosoma) are brightly coloured 

 and sit about very conspicuously in bright sunshine, often 

 gregariously, so that the group stands out from its sur- 

 roundings. These bugs are protected by a powerful odour, 

 very unpleasant to man and possibly to most insectivorous 

 animals. Yet we find that the same protective odour occurs 

 throughout the family, of which many (perhaps the majority 

 of) species are not brightly but cryptically coloured, and by 

 no means expose themselves in conspicuous positions. It is 

 difficult to obtain satisfactory evidence as to how far the pro- 

 tective devices of warningly coloured animals are efficacious ; 

 this is particularly true where protection is by means of nauseous 

 taste, since human predilections are of little value, and experi- 

 ments on animals in captivity are liable to give very uncertain 



