ON COLOURS IN METAL GLASSES, ETC. 251 



paper. We have already* seen that if gold glass when first made in the furnace be 

 rapidly cooled, the glass remains colourless. In order to obtain ruby glass, the molten 

 glass must be left in the annealing oven and maintained at a high temperature for 

 about three days. If the glass is too violently heated or is kept too long at a high 

 temperature, it becomes turbid, reflects brown light, and develops first an amethystine 

 and then a blue tint by transmitted light. But it now appears that the gold cannot, 

 as previouslyt suggested, be in solution in the colourless glass when first heated ; for 

 if metallic gold were in true solution in the glass it would have the vapour colour 

 indicated by the dotted curve in fig. 1. The gold must therefore be gradually reduced 

 during the annealing process. So long as the glass remains hot enough to admit of 

 molecular movement, the molecules of gold go together to form spheres, and these 

 small spheres tend to coagulate into crystallites.} If the glass cools before the 

 coagulation of the small spheres, a gold ruby glass is obtained. If, however, some of 

 the small spheres have coagulated into crystallites, the density of which exceeds '6 

 of that of normal gold, these crystallites will reflect light which is predominantly 

 yellow or red.|| The glass will thus reflect brownish light; and since the more 

 refrangible rays are less reflected than those of longer wave-length, the red end of the 

 absorption curve will, owing to the crystallites, be raised relative to the blue. The 

 glass will thus appear bluer than when no coagulation has occurred. Further, as 

 these crystallites may be of dimensions comparable with a wave-length of light, they 

 destroy the optical continuity of the medium and produce turbidity. Now the blue 

 colour of a gold glass is always associated with turbidity and a brown appearance by 

 reflected light, so that the formation of crystallites of gold in the glass accounts for 

 the irregular blue and purple colours which gold glass sometimes exhibits. 1 



Diffusions of gold particles in water the so-called "colloidal solutions" of gold- 

 have been prepared by FARADAY**, ZsiGMONDY,tt and STOEKL and VANINO^, who 



* Loc. cit., p. 392. 



t Loc. cit., p. 392. 



J In the case of copper ruby glasses the process continues until actual crystals of copper are formed, 

 but I have not seen gold crystals in a glass, although it is probable that they are occasionally formed. 



KIRCHNER and ZSIGMONDY ('Ann. der Phys.,' XIII., 1904, p. 591) estimate that a clump of gold 

 particles in a blue gold-gelatine preparation contains at least 50 per cent, of gold. See below, p. 254. 



|| See fig. 12, below. 



U The blue and violet [purple] colours of the glasses D and E in SIEDENTOPF and ZSIGMONDY'S table 

 (see ' Phil. Trans.,' A, 1904, p. 397), as well as the red, yellow and brown colours of the cone of light 

 emitted by them, are thus explained. STOKES (Royal Institution Lecture, 1864, 'Collected Works,' 

 vol. IV., p. 244), without entering into the question why gold glass ordinarily transmits pink light, says 

 that, it being the property of gold to transmit bluish light, the metallic gold in suspension causes the blue 

 appearance. 



** FARADAY, Bakerian Lecture for 1857, printed in 'Phil. Trans.' for 1857, and reprinted in his 

 'Researches in Chemistry and Physics.' References will be made to the pages of the reprint. 



ft ZSIGMONDY, 'LiEB. Ann.,' vol. 301 (1898), p. 29, and 'Zeitschr. f. Electrochem.,' vol. IV., p. 546. 



II STOEKL and VANIXO, 'Zeitschr. f. Phys. Chem.,' XXX. (1899), p. 98. 



2 K 2 



