ON COLOURS IN METAL GLASSES, ETC. 279 



then purple. As p. still further diminishes, the relative absorption of red continues to 

 become small, so that the film becomes pink. Finally, at p. = 0, the absorption of 

 green is less than that of blue, and the colour has changed from pink to red.* It is 

 further seen, from fig. 12, that the reflecting power E, has, as p, began to diminish, 

 become more yellow. At p, = '60, when the colour of T is still purple or pink, the 

 colour of R has become green ; and thenceforward R remains green as p. diminishes 

 from '6 to zero. 



Similar consideration of figs. 9 and 11 shows that in the case of silver, as p. 

 begins to diminish from unity, the colour by transmitted light becomes at first bluer, 

 then changes to purple in the neighbourhood of p, = '8, and thence, through pink, to 

 red or " amber " as p, further diminishes to /A = 0. In fact, it appears from the four 

 colours for which calculations have been made, that there is, for any value of p., a 

 well-defined absorption band at some position in the (visible or invisible) spectrum, 

 and that, as p. diminishes, the position of this absorption band moves from the 

 infra-red through the visible spectrum towards the shorter wave-lengths, being 

 at X = -630 for p, = about '83, at X = "589 for p, = about '80, at X = '500 for 

 p, = about '69, and at X = '450 for p. = about "55. Fig. 13 shows that the colour of 

 the reflected light becomes distinctly blue at about p = '75, and remains blue down 

 to p. = 0. 



With a view to determining what may be the explanation of the colours and 

 changes of colour exhibited by gold and silver films, we have now to compare the 

 latter colours with those which we have found above to be consequent upon a mere 

 isotropic change in density. 



BEiLBYf has prepared gold films by using paints in which " the metal had been 

 brought into solution in an essential oil." Having smoothly coated a plate of glass 

 or mica with the paint, he heated it to a temperature of about 400, thereby driving 

 off the oil and other volatile constituents. A film of pure gold with full metallic 

 reflection, and transmitting green light, is left adhering to the glass. 



When these films are kept at a high temperature for some time, they change colour. 

 By transmitted light, the original olive-green colour becomes at first bluer, then 

 changes to purple, in which, as the annealing process is still continued, the red 

 predominates more and more over the blue, until finally the purple has given place 

 to pink. The reflecting power of the film has, meanwhile, diminished. But the 

 colour of the light reflected from the blue films remains yellow, while the pink films 

 reflect a green colour. I have before me a gold film prepared in this way and 

 subjected to lengthy annealing. By transmitted light it appears striated with pink 

 and blue bands. By reflected light the blue striae become golden, but the pink 

 striae green. Under the microscope the film appears continuous, and is quite thick. 



These colour changes, both with transmitted and with reflected light, are just 



* Or yellow, if the colour is faint. See the second footnote on p. 243 above. 

 - t Loc, cit., p. 40. 



