1887.] VALUE OF COLOUR AND MARKINGS IN INSECTS. 215 



Minchin has called my attention to some experiments made by 

 M. (le Niceville upon tlie images of Acrseinse &o. (Butterflies oV 

 India, Burmali, and Ceylon, vol. i. part ii. p. 318). " M. de Nice- 

 ville has experimented with the carnivorous Mantis on many of the 

 Butterflies believed to be offensive to birds, and he has found that 

 A. violcB is the only Butterfly which all the species of Mantis he has 

 experimented with refuse to eat." 



One other consideration remains to be partially discussed before 

 giving the list of experiments. The meaning of conspicuous colora- 

 tion may be said to be clear and definite in the case of larvae, being 

 only capable of the explanation that it is of value for protective pur- 

 poses. But in the perfect forms the other explanation of colour is 

 always possible, i. e. its use in courtship, and, as a consequence, its 

 growth from small beginnings and its gradual perfection by sexual 

 selection. By what criterion, then, is it possible to judge of the 

 meaning of bright and cons[)icuous colours in any particular instance? 

 In the first place, the brilliant tints due to sexual selection alone 

 rarely usurp the whole surface of an insect, and tliere are certain parts 

 (exposed in the protective attitude of rest) in which we expect to find 

 such brilliant tints replaced by the (usually) sober colours which 

 harmonize with the environment. For instance, this is well known 

 to be the case with the uppersides of the upper wings in most 

 moths, and with the undersides of both wings in nearly all butterflies. 

 If, therefore, such exposed parts are cons()icuously coloured, strong 

 presumptive evidence will be afforded for the explanation of the 

 colour as belonging to the "warning class," Another test is 

 iouud in the mode of flight, which may be expected to be such as 

 will clearly display the colours no less than when the insect is at 

 rest. The question is no doubt complicated by the two kinds of 

 coloration running into each other. Thus Wallace has shown 

 that the shapes and colours due to sexual selection may run riot in 

 localities (certain islands) where enemies are largely" excluded by 

 barriers, and in the same way the brilliant colours of dangerous or 

 nauseous insects may perhaps be sometimes exj.lained by equal 

 immunity, although due to other causes. But a conspicuous appear- 

 ance gained in this way will be always liable to be used for an 

 entirely different object as " warning coloration." But if such a 

 change of use took place, we should expect some change of pattern 

 or some new combination of colours, for this reason, that "warning 

 colours " have one meaning which is not associated with sexual 

 colours, i. e. the production of a conspicuous appearance and the 

 more or less complete subordination of everything to this end. 

 Nevertheless it is almost certain that the appearance of any species, 

 however specialized for other ends, possesses sexual significance, 

 and appeals as an adornment to the modified taste of the individuals 

 concerned ; and we have a rough criterion of the extent of the 

 modification in taste, when we compare such appearances with those 

 which possess a sexual value alone, which are concealed except 

 during flight and which are especially displayed during courtship. 

 Prof. Aleldola has recently drawn my attention to an observation of 



