2go MR. J. C. MAXWELL GARNETT 



those which have been shown above to be consequent upon a continuous diminution 

 in the density of a gold film, which throughout remains either amorphous or 

 "granular" (i.e., possessing a microstructure of small spheres). The view that the 

 film is initially amorphous or granular, and that heating diminishes its density, is 

 supported, as has already been pointed out,* by the fact that the curves in figs. 8 

 and 10 show that the absorption of light increases rapidly as /* begins to diminish 

 from unity, while BEILBY'S films exhibited just such an increase of absorptive power 

 when first heated. This view is also in accordance with the loosening of structure 

 which is suggested by the great decrease in electric conductivity which accompanies 

 heating. But direct evidence of the correctness of the view that heating produces 

 decrease in density is not wanting, for BEILBY! has estimated the thickness of a film 

 which had been annealed to the purple stage. He found, by weighing the gold from 

 a given area, that, had the density of the gold been then normal (p = 1), the film 

 would have been lGO/x/.t thick, whereas, under the microscope, the thickness seemed 

 to be much greater than this. The density of the gold in the purple film thus 

 appeared to be less than in the normal green films. 



We conclude, therefore, that (a) the films, as first prepared, are amorphous or 

 granular in structure; and (/*) heating diminishes the density of the film, while 

 pressure is able to increase the density ar/ain.^ 



Further, BEILBY found t that, when the heating of a film was continued after it had 

 reached the purple stage, " the film assumes a frosted appearance by reflected light 

 and becomes paler by transmitted light." The frosted surface appeared, under the 

 microscope, to consist of granules at least lOO/x^a in diameter. This phenomenon 

 suggests that in the earlier stages of annealing, smaller granules were formed, which, 

 as annealing proceeded, ran together to form larger granules : and the formation of 

 such minute granules, while, according to our analysis, it does not aft'ect the optical 

 continuity of the film, will explain the diminution in density which occurs on heating. 

 It is, therefore, most probable that (c) tl\e diminution in density produced by heating 

 is effected by the passage of metal from the amorphous to the granular phase and the 

 growth of the larger granules at the expense of the smaller; and the increase in 

 density produced by pressure may be accompanied by the passage of metal from the 

 granular to the amorphous phase. \ 



The optical properties of the films of gold whicli FARADAY produced by reducing 

 that metal from its solution by means of phosphorus, tend to show that these films are 

 composed of amorphous or granular gold of density less than the normal. The films 

 appeared to consist of pure gold ;|| when first prepared the films appeared of a grey 



* 'Phil, Trans.,' A, 1904, p. 415. 



t IMC. cit., p. 41. 



I Of. the effect of pressure on FARADAY'S " phosphorous " films after heating see next page. 



[Note added 31st August, 1905. Subsequent analysis has, however, shown that a sufficient flattening 

 of the granules would cause the colours of the standard metal (//. = 1) to be exhibited.] 



|| FARADAY, loc. cit., p. 408. 



