18S7.] VALUE OF COLOUK AND MARKINGS IN INSECTS. 231 



always) more abundant species which it resembles. In the former 

 classes of resemblance we have groups of two or more conspicuous 

 forms all possessing unpleasant attributes, which become convergent 

 in external appearance, or which maintain an initial nniformity, 

 and in eitlier case are mutually benefited by the process. In the 

 latter class the resemblance would be a source of danger to the 

 mimicked species if the edibility of the mimicking species were 

 discovered ; and the experiments detailed in the present paper show 

 how likely it is that such qualities would be discovered if tlie lutter 

 species became relatively abundant. Nevertheless, until the dis- 

 covery was made, the mimicry would be an advantage to both 

 species, for the reasons already adduced. In the following Table the 

 colours of conspicuous insects are tabulated, /. e., those contained in 

 Tables I. and IV., excluding S.fuciformis and the conspicuous larva 

 of L. pini, the latter being omitted because I have never seen a 

 specimen, and because the appearance differs greatly in the various 

 figures I have been able to consult (see Table A, pp. 232-235). 



I have described the colours of the imagos at rest to correspond 

 with the larvae and pupae ; in flight the following effects are seen : — 

 Imagos of Wasp, Bombus, JSomada marshameUa, E. jacobcece, A. 

 filipendulce, A. grossidariatu, the two Coccinellidse, Telephorus and 

 Chrifsome/a, would show much the same colours as at rest, although 

 in E. jacobcECB and A. JUipenduIce the red would be in larger amount 

 because of the under wings, and in the Coccinellidse, Telephorus and 

 Chrysomela, the black would be in far larger amount because of the 

 body. The imagos of S. menthastri, S. lubricipeda, and P. aurijlua 

 would hardly show the black spots in flight, but would appear 

 whitish, yellowish, and white respectively. 



This comparison is exceedingly interesting if it is remembered that 

 the colours which are repeated again and again are those which are 

 known to produce the greatest eff'ect. Thus the greatest possible con- 

 trast is afforded by black and white, and next to this by black (or 

 some very dark colour) and yellow, orange, or red, the brightest 

 colours in the spectrum, which possess a far higher illuminating- 

 power than any of the others. Hence we find that the colours of all 

 the conspicuous insects which have been tested are in all except five 

 cases included in the short list given above. And these five only 

 differ in the inclusion of blue in one case, and of green in the other 

 four cases. Hence we probably see that in addition to the advantage 

 gained by convergence which has been alluded to above, benefits have 

 been derived from the colours which have been employed ; and as the 

 choice of the most conspicuous colours is limited, it is seen that a 

 certain amount of similarity must follow incidentally from the 

 number of forms of life among which the few combinations are 

 divided. Hence convergence has been aided and perhaps given its 

 starting-point by the action of another principle of coloration also 

 favoured by natural selection, and leading in the same directiori as 

 convergence itself (see Table B, pp. 236, 237). 



Just as similarity in colours was favoured by the limited number 

 of suitable combinations, so there are a few eminently conspicuous 



