Colouring in Birds and Insects. 561 



light. This, however, follows, if we consider that the con- 

 dition for such a negative phase carve is that the transition 

 layer have an index of refraction greater than that of the 

 second medium ; and as the refractive index for magenta is 

 low at the blue end of the spectrum and high at the red end, 

 the inversion of sign is strictly in accord with the theory, of 

 which indeed it furnishes a striking confirmation. 



On applying the simpler general tests of metallic reflexion 

 to the case of iridescent plumage of birds, scales of butterflies, 

 and wing-cases of beetles, one is at once struck with the 

 close resemblance these bear to the aniline colours, in every 

 particular : for 



1st. The intensity of the reflected light is much greater 

 than for the " non-metallic "" plumage, &c, in some cases 

 approaching the value of the reflexion factor of the metals 

 themselves. 



2nd. The reflected light is always coloured, showing either 

 a rapid change of index of refraction, or of coefficient of 

 absorption with the wave-length or colour ; and, indeed, it 

 may perhaps be objected that these colours are far more 

 vivid than any of the reflexion hues of the aniline dyes, or 

 of any other case of "surface colour" hitherto observed. 



3rd. In the cases which could be investigated for this 

 relation (unfortunately rather few) the transmitted light is 

 approximately complementary to that which is reflected. 



4th. The change of co'our with changing incidence strictly 

 follows the law already mentioned — the colour always 

 changing towards the blue end of the spectrum as the 

 incidence increases. 



This remarkable agreement has been pointed out by Dr. 

 B. Walter in an admirable essay, " Die Oberflachen- oder 

 Schiller-Farben/' and it is shown that none of the other 

 causes of colour phenomena (in particular interference and 

 diffraction) can be effective ; the laws which govern these 

 last being totally different. It is, therefore, somewhat sur- 

 prising to find that the contrary view is still held by many 

 eminent naturalists, and it is hoped that the further evidence 

 here presented may serve to emphasize the distinction between 

 "metallic''' or "surface''' colour and the remaining classes 

 of colour (due to pigments, interference, and diffraction). 



In attempting to apply the more rigorous optical test of 

 the measurement of the phase-difference and amplitude-ratios, 

 one is met at the outset with the serious difficulty of the 

 absence of true " optical " surface. In fact, the material- we 



