336 Prof. Wood and Mr. Kimura on Scattering and 



on the original negative the density of "/" is certainly double 

 that of, "6." 



As subsequent experiments showed that the dense vapour 

 reflects polarized light in much the same way as a film of 

 metal, it was of some interest to see whether the reflecting 

 power of the quartz surface in contact with the vapour 

 passed through a minimum before beginning to increase with 

 increasing vapour pressure. In the case of metallic deposits 

 on glass, the reflecting power of the glass is considerably 

 diminished by very thin layers of metal, when the reflexion 

 is from the glass side ; in fact it is reduced nearly to zero if 

 the thickness of the metallic film is just right. If a number 

 of strips are silvered cathodically on the face of a prism of 

 small angle (five to ten degrees) with exposures to the 

 discharge of from say one minute to ten minutes, one or 

 more strips will be found which appear quite black in 

 reflected light when viewed through the glass, though all 

 reflect much more powerfully than the glass when viewed 

 from the silver side. 



The mercury vapour was examined for a similar phe- 

 nomenon in the following way: — The convergent 2536 beam 

 from the monochromator was reflected from the prismatic 

 plate of the bulb, and the two reflected beams received on a 

 plate of uranium glass. The temperature of the bulb was 

 then gradually raised and the fluorescent images on the 

 uranium plate watched. If the vapour behaved exactly as a 

 metal film of increasing thickness, the image formed by 

 reflexion from the inner surface ought to fade away gradually 

 and then rapidly brighten. No trace of such a phenomenon 

 was observed. The two images remained of the same in- 

 tensity until a temperature of 250° was reached ; above this 

 point the image due to the inner reflexion rapidly brightened 

 as the temperature rose, reaching its maximum brilliancy in 

 the neighbourhood of 300°, at which temperature it appeared 

 to be four or five times as bright as the other image. 



This specular or metallic reflexion of the light by the 

 vapour occurs only when there is exact synchronism between 

 bhe luminous vibration and the free period of the system 

 which causes the 2536 line. This fact is emphasized because 

 there is another type of selective reflexion which occurs 

 when the synchronism is not exact, and which is the result 

 of the refractive index of the vapour. This will be discussed 

 presently, after the polarization experiments have been 

 treated. 



