566 On Metallic Colouring in Birds and Insects. 



On immersing the specimen in oil or other liquid little or 

 no change is observed, except in those specimens in which a 

 small communicating aperture exists in the neck (point of 

 support) of the scale. The oil can be seen to gradually fill 

 the interior, and simultaneously all trace of colour vanishes*. 



It appears, then, that the colour in this case is due to fine 

 striations on the interior surface of the scale. 



Plustiotis resplendens. 



This is a beetle whose whole covering appears as if coated 

 with an electrolytic deposit of metal, with a lustre resembling 

 brass. Indeed, it would be difficult for even an experienced 

 observer to distinguish between the metal and the specimen. 



On examination with the Babinet compensator it was found 

 that the reflected light was circularly polarized even at normal 

 incidence, whether the incident light was polarized or natural. 

 The proportion of circularly polarized light is greatest in the 

 blue, diminishing gradually in the yellow portion of ths 

 spectrum and vanishing in the orange-yellow — for which 

 colour the light appears to be completely depolarized. On 

 progressing towards the red end of the spectrum traces of 

 circular polarization in the opposite sense appear, the proportion 

 increasing until the circular polarization is nearly complete 

 in the extreme red. 



It was at first suspected that the phase difference (not 

 always as great as one quarter, but varying between *15 and 

 •25) was due to linear structure, as in the case of Alorplio 

 alga ; but on rotating the specimen about the normal no 

 change resulted. The effect must therefore be due to a 

 "screw structure" of ultra microscopic, probably of mole- 

 cular dimensions. Such a structure would cause a separation 

 of natural incident light into two circularly polarized pencils 

 travelling with different speeds, and having different 

 coefficients of absorption. 



Such cases have been observed in some absorbing crystals; 

 but whereas in these the difference in absorption between the 

 two circularly polarized pencils is quite small compared with 

 the total absorption— here one of the two is almost totally 

 reflected, while there is scarcely a trace of the other. 



If this hypothesis be correct, however, the selective ab- 

 sorption (or reflexion) being reversed at the other end of the 



* Sometimes a faint indication of colour remains (usually greenish) 

 which shows the characteristics of surface colour. It is probable that 

 this surface colour acts conjointly with the eft'ect of diffraction, and 

 indeed the character of the spectrum indicates an excess of green which 

 may be thus accounted for. 



