ABSORPTION OF LIGHT, ETC. 51 



take light reflected from glass at a certain angle, which is 

 called the angle of polarisation, the reflected light is polarised, 

 and capable of being extinguished by an analyser such as a 

 Nicol's prism. Light reflected from a metal is not polarised at 

 any angle of incidence, though it is partially polarised at an 

 oblique angle. I say partially polarised, but I will leave the 

 explanation of that to my friend Mr. Spottiswoode, who will 

 give a lecture on that subject. Bronzy crystals of perman- 

 ganate of potash agree in that respect, to a certain extent at 

 least, with the metals ; if you examine the light by reflection 

 you find that it is not capable of extinction by analysing 

 under any conditions. If you examine it at such an angle 

 of incidence that a vitreous substance would give you light 

 capable of extinction, the light becomes weaker and of purer 

 green. I have analysed the' light reflected by a crystal under 

 these conditions, by a combination of a prism and a Mcol's 

 prism, so as to extinguish what light would have been reflected 

 from glass under similar conditions, and this curious result came 

 out. I must premise that crystals of permanganate of potash 

 are too intensely opaque to allow you to examine them by 

 transmission, but you can make a solution of them and ex- 

 amine that, and it shows these bands of absorption which are 

 shown on the diagram [referred to]. Now on examining, in 

 the manner I have mentioned, the green light reflected from 

 the crystals at an angle similar to that at which light reflected 

 from glass would have been quenched by a Nicol's prism, this 

 curious result was obtained ; the spectrum was seen to consist 

 of four bright bands, and perhaps a trace of another, the rest 

 of the spectrum being wanting. Now what were the positions 

 of those four bands ? When the positions were observed, by 

 referring them to the standard fixed lines of the spectrum, 

 which were seen at the same time, they were in the positions 

 represented in the under part of that figure ; they agreed in 

 position with the first four of the five dark bands seen in the 

 transmitted light. The spectrum begins to get comparatively 

 faint in the region of the fifth band of absorption, and there 

 was hardly a chance of seeing the fifth bright band if it had 

 been there ; but you see that whereas, as regards transmitted 

 li.^ht, the crystals pass alternately through maxima and 

 minima of transparency alternately from the . condition of a 

 vitreous substance to the condition of a metal, as to the avidity 

 with which they absorb the light corresponding to these 



E 2 



