ON COLOURS IN METAL GLASSES, ETC. 263 



It appears that while the graphs of R' and R widely differ, the positions of the 

 respective maxima falling near X = '360 and X = '436 respectively, the graph of R 

 closely resembles that of R 0) * the maxima of R and of R' occurring at almost the 

 same value of X. We conclude that the presence of small spheres of silver throughout 

 the stained region of the glass will account for the blue reflection ; and we thus 

 confirm the view, to which absorption phenomena led us, that silver glass consists of 

 a suspension of small spheres of silver in a colourless glass. 



Before leaving the consideration of the blue reflection from silver glass, it may be 

 noticed that the light is not reflected as from a plane interface between glass and 

 silver glass. Thus when the source of light is an electric arc, the blue colour is 

 clearly discernible by an observer whose eye is not in the straight line determined by 

 the ordinary law of reflection. This effect is due to the irregularity of the interface, 

 the silver not having penetrated the glass to a uniform depth. AH alternative 

 explanation, however, suggests itself, the blue colour might be due to independent 

 radiation from discrete spheres (or molecules) of silver so far apart as not to form an 

 optically homogeneous medium. The intensity of the emitted light would then be 

 proportional to (a 2 +4/3 2 )/X 4 (in the case of spheres, or (a'- + 4/3'-)/X l in the case of 

 molecules). Further, the blue colour would be equally visible if the light illuminating 

 the discrete spheres (or molecules) entered the silver glass from the air side or the 

 clear glass side ; and this is not the case. 



It is of interest to notice that while each individual sphere in glass radiates out 

 light of an intensity proportional to (a*+ 4j6 2 )/X 4 , a surface separating a glass, containing 

 many of the spheres to a wave-length of light, from a region of the same glass in 

 which no spheres are present, reflects light with an intensity proportional to or + 4/3". 

 This is due to the fact that the number of spheres (on the reflecting surface), the 

 phase of the forced vibrations of which lies at any instant between given limits, is 

 proportional to X 2 ; so that the intensity of the reflected light is proportional to X' 

 times the intensity of the light emitted by a single sphere. 



8. Diffusions of Copper. The Nature and Form of the Suspended Particle*. 



We proceed to discuss the colours produced by diffused particles of copper in order 

 to discover the cause of colour of copper ruby glass. The values of the expression 

 nV//iX for v = 1-56, v = 1'5, and v = I'O given in Table III., are plotted in fig. 6, the 

 maxima being determined as in the case of fig. 1 (cf. 5 above). As in the case of 

 silver, these curves should fairly accurately represent the absorptions produced by 

 copper spheres in glass v = T56, in glass v = 1'5, and by copper spheres or molecules 



* The fact that E increases from yellow to red, while E diminishes in the same range, would be 

 accounted for if the black paper with which Sir WILLIAM ABNEY backed the stained face of the glass 

 reflected 2 per cent, of the light incident on it. Further experiments are to be made on this. 



