1883.] Absorption of Ultra-Violet Rays by various Substances. 73 



Q (3285) on one side, and to s (3045) on the other. In the higher 

 region there is a second absorption sensible about wave-length 2600, 

 and increasing in intensity up to about wave-length 2580, beyond 

 which point it is complete. 



Chlorine peroxide gives a succession of nine shaded bands, at nearly 

 equal intervals, between M and S, with faint indications of others 

 beyond. In the highest region this gas seems quite transparent. 



A slice of chrome-alum a quarter of an inch thick, is transparent 

 between wave-lengths 3270 and 2830, its absorption gradually 

 increases on both sides of those limits, but rather more rapidly on 

 the more refrangible side than on the other, and becomes complete 

 below about wave-length 3360 and above wave-length 2730. 



A very thin plate of mica shows absorption beginning about S 

 (3100), rapidly increasing above U (2947), and complete above wave- 

 length 2840. 



A thin film of silver precipitated chemically on a plate of quartz 

 transmits well a band of light between wave-length about 3350 and 

 3070, but is quite opaque beyond those limits on both sides.* 



A thin film of gold similarly precipitated merely produces a slight 

 general absorption all along the spectrum. 



The difference between the limits of transparency of Iceland spar 

 for the ordinary and extraordinary rays, inferred from theory by 

 Lommel, we find to be very small, and hardly to be detected without 

 using a considerable thickness, three inches or more, of the spar. 



We had expected to be able to apply the well-known photo- 

 metric method by means of polarised light to the comparison of 

 intensities of ultra-violet rays. Ordinary Nicol's prisms are not 

 applicable to ultra-violet rays on account of the opacity of the Canada 

 balsam, with which they are cemented, but through the kindness of 

 the President of the Society, we obtained from him the loan of a pair 

 of Foucault's prisms. Upon taking photographs of the spectrum of 

 the iron spark through this pair of prisms at various inclinations 

 between the planes of polarisation of the two prisms, we found that for 

 the whole range between the position of parallelism and the inclina- 

 tion of 80° there was no sensible difference of effect upon the photo- 

 graphic plate, though the length of exposure was in all cases the 

 same. For inclinations between 80° and 90° there was a sensible 

 and increasing diminution in the photographic effect as the planes of 

 polarisation of the polariser and analyser were more nearly at right 



* Corcra has noticed (" Spectre Normal du Soleil," p. 23, note) that such a 

 film of silver is transparent for certain ultra-violet rays, but he places them about 

 wave-length 270, which does not agree with our observations. Chardonnet (" Comp. 

 rend.," February, 1883) says that the transparent band extends from O to S. 

 W. A. Miller ("Phil. Trans.," 1863) noticed that for a certain distance in the 

 ultra-violet a silver reflector did not reflect the incident rays. 



