ON COLOURS IN METAL GLASSES, ETC. 283 



transparent and tending to shrink during the process.* Thus a silver leaf which 

 before heating was opaque, or only able to transmit deep blue light, and that very 

 feebly, was so altered by heating that the light of a candle could be seen through 

 forty thicknesses.! But in every case the original colour of the leaf, whether of gold 

 or of silver, returns when the leaf is subjected to pressure. 



The differences between the effect of heat on chemically prepared films and on 

 beaten leaf correspond to differences between the laminatedf structure of the leaf and 

 " the closer and more horn-like texture of the films deposited by chemical agents.''^ 



The optical properties of a laminated metal leaf may be estimated and compared 

 with the corresponding properties of an amorphous or granular film of the same metal, 

 if the optical constants of a plate built up of a number of flat spheroids with their 

 polar axes normal to the plate can be calculated. The general problem of the 

 transmission of electromagnetic waves by a medium composed of a number of minute 

 similar and similarly situated ellipsoids, distributed at random many to a wave-length, 

 has now been solved, and it is hoped that the discussion of the optical properties of 

 gold and silver leaf, of the change in those properties which is produced by heat, and 

 of the relations of metal films (spheroidal, granular, and amorphous) to polarised light, 

 may form the subject of a future memoir. 



With these exceptions, namely, the properties peculiar to beaten leaf and the 

 relations of metal films to polarised light, all the experimental relations of gold (and 

 other metals) to light, which FARADAY described in his Bakerian Lecture have now 

 been discussed, and we are led to the conclusion that the phenomena exhibited 

 whether by chemically or electrically deposited films, or by particles of gold diffused 

 in glass, jelly, or water are due to different groupings of the metal molecules and to 

 variations in the mean distance between adjacent molecules, and in no case are they 

 due to allotropic modifications of the molecules themselves. 



13. CAREY LEA'S "Allotropic" Silver. 



In the former communication] | it was suggested that CAREY LEA'S "allotropic" 

 silver was in reality only finely divided silver, the division being sufficiently fine to 

 admit of the films being optically continuous. 11 He advances** two principal arguments 



* Loc. cit., p. 396. 



t FARADAY, loc. cit., p. 399. 



t BEILBY. loc. cit., p. 43. The difference in structure is shown by the fact that while mercury will 

 diffuse slowly and uniformly in the compact film, in the leaf thin streams of mercury may be seen shooting 

 rapidly in all directions. 



BEILBY (loc. cit., pp. 48 et seq.) has shown that a layer of exceedingly flat spheroids is generally found 

 on the surface of a metal. 



|| 'Phil. Trans.,' A, 1904, p. 419. 



U It is not necessary to suppose the microstructure of the finely divided silver to be granular, as was 

 done in the former paper. It may be in part granular and in part amorphous. 



** Fide ' Amer. Journal of Science,' 1889, and 'British Journal of Photography,' March, 1901. Also 

 'Phil. Mag.,' vols. 31, 32 (1891). 



2 O 2 



