WOOD. — NITROSO-DIMETHYL-ANILINE. 61 



tion spectrum was photographed by pressing out a drop of the fluid 

 substance between a quartz lens and plate, the whole being kept warm 

 by a current of hot air. In this way a film of variable thickness was 

 obtained, which when brought close to the slit of the quartz spectrograph, 

 enabled a record to be made of the relative intensities, positions, and 

 general form of the absorption bands. One of these photographs is 

 shown in Figure 4, Plate 3. 



The nitroso begins to vaporize at a temperature only a little above 

 its melting point, and the absorption of the vapor is interesting, for 

 unlike most absorbing vapors and gases, this substance shows a broad 

 absorption band, similar to the liquid, only shifted well down into the 

 ultra-violet. The nitroso was vaporized in a glass tube, the ends of 

 which were closed with quartz plates. The tube was previously ex- 

 hausted, and was heated by a water bath through which it passed. The 

 absorption spectrum was photographed at different temperatures, ranging 

 from 85°, the melting point, to 100°. The absorption does not begin 

 as a narrow line, as is usually the case with gases, but with a broad 

 band, which increases in intensity as the density of the vapor increases. 

 The centre of this band is at wave-length 34, while the centre of the 

 band in the case of the liquid is at 43. 



I made numerous experiments to see whether the density of the vapor 

 could be sufficiently increased to cause a shift iu the position of the band 

 towards the red, all of which failed, owing to the fact that the nitroso 

 decomposes at temperatures above 150. The substance was heated in 

 strong sealed bulbs in an air bath, but decomposition always resulted 

 before a density sufficient to shift the absorption band in a measurable 

 degree had been obtained. 



On the other hand the position of the absorption band can be shifted 

 by increasing the density of the medium in which the nitroso vapor is 

 present. If a solution of nitroso in ether is heated above its critical 

 temperature in sealed glass tubes, the centre of the absorption band can 

 be given almost any position between that of the vapor baud and the 

 solution band, by varying the amount of ether in the tube, or in other 

 words by varying the density of the vapor. 



I made a number of attempts to prove that the double overlapping 

 band shown by the solid was connected with the double refraction of 

 the nitroso crystals, by placing a Nicol before the slit of the spectro- 

 graph and photographing the absorption spectrum of a thin crystalline 

 film of the substance between two quartz plates. A number of spectra 

 were taken with the Nicol in different positions, a comparison of which 



