ON COLOURS IN METAL GLASSES, ETC. 267 



Sir WILLIAM RAMSAY, when I first called his attention to the explanation of the 

 coloration of glass by radium which is afforded by supposing the radiation to reduce 

 metal in the glass, suggested that the reduction might be effected by the discharge 

 of free ions of the metal. Since that time the further evidence that has accumulated 

 seems to 'favour the truth of this theory. Thus, as all the colour-changes from pink 

 to blue exhibited by gold glass can be imitated with suspensions of gold in water, 

 the glass appears to behave as a liquid, although a very viscous one ; and it seems, 

 therefore, reasonable to suppose that the salt of a metal which will dissociate in 

 water will dissociate also in glass. As an alternative hypothesis, we might suppose 

 the compound molecules broken up by the rays. But, were this the case, the a-rays 

 would be far more efficient than the /3 in producing the colour. And this is not 

 true ; for the coloration produced in the splinters of gold and silver glass, as well as 

 in soda and potash glasses, are not, apparently, stronger on the sides of the glass, 

 but seem to be of uniform strength throughout. From this it appears that the /8-rays 

 are alone capable of producing the colour. This is in accord with the former 

 hypothesis. For the ions of the metal in the glass would be positively charged, and 

 their discharge by the negatively-charged /3 particles (or cathode rays) would change 

 them into molecules just as the sodium ions in the electrolysis of common salt are 

 discharged at the cathode, and thus are transformed into molecules of sodium, 

 imparting a violet colour to the water and capable of forming caustic soda. 



It appears, therefore, possible that all glasses contain free ions of metal, and that 

 it is by the discharge of these ions, and consequent reduction of the metal, that 

 cathode and Becquerel rays are able to produce coloration in them. 



10. Numerical Values of the Optical Constants of Media containing Large Volume 



Proportions of Certain Metals. 



The preceding sections of this paper have treated only of the optical properties of 

 those media for which the volume proportion, //., of metal is very small. The 

 consideration of media in which p. may have any value up to unity will now, however, 

 be resumed, in order to discover what may be the physical explanation of those 

 colours and changes of colour which FARADAY,* BEILBY,! and others have found to 

 be exhibited by thin metallic films. In 11 of the former communication J the 

 question whether films built up of small spheres of silver or of gold would, for any 

 given volume proportion of metal, transmit red or yellow light more easily, was dis- 

 cussed, and the conclusions reached were compared with the results of Mr. BEILBY'S 

 experimentsf on the effect of heat on thin films of metal. The present section 

 extends the scope of that enquiry. 



* Bakerian Lecture for 1857, 'Phil. Trans.,' A, 1857. Reprinted in FARADAY'S 'Researches in 

 Chemistry and Physics,' pp. 391 et seq. (Reference will be made to the pages of the reprint.) 

 t 'Roy. Soc. Proc.,' vol. 72, 1903, p. 226.. 

 t 'Phil. Trans.,' A, 1904, p. 415. 



2 M 2 



