60 PROCEEDINGS OF THE AMERICAN ACADEMY. 



ground and polished, and one face of it brought in contact with a 

 small drop of nitroso fused on a strip of blackened brass. One end 

 of the strip was heated in a flame, the nitroso remaining fluid as a 

 result of the heat conduction along the strip. The object of using the 

 blackened brass was to get rid of reflection from the back surface of 

 the nitroso. The light reflected from the front surface of the quartz 

 was thrown to one side, owing to the inclination, so that everything was 

 eliminated except the reflection from the quartz-nitroso surface. The 

 spectrum of this selectively reflected light is shown in Figure 2, Plate 3. 

 The minimum at wave-length 36 will be seen to have disappeared, and a 

 new one will be found at 29 just about where we should expect it. This 

 minimum would doubtless be more pronounced were it not for the fact 

 that the absorption coefficient has at this point a not inconsiderable 

 value, which will cause the bounding surface in question to have a higher 

 reflecting power than if both media were perfectly transparent. 



Absorption of Solid, Liquid, and Gaseous Nitroso. 



No determinations of the dispersion of the solid nitroso have been 

 made, owing to the difficulty of getting suitable prisms: when the 

 fluid prisms cool off the nitroso crystallizes in aboresceut forms, and no 

 longer transmits regularly. If the prism is held close to the eye, and a 

 lamp flame viewed through it, enormously deviated spectra are seen, due 

 to the formation of crystals having a much larger angle than the prism. 

 The substance in the solid condition is double refracting, the two spectra 

 which every pristn furnishes being extinguished in turn by a revolviug 

 Nicol prism. The absorption baud of the solid nitroso was studied by 

 photography. A drop of fluid nitroso was pressed in a clamp between 

 two hot plates of quartz, which were allowed to cool under pressure. 

 In this way a very thin film of the solid substance was obtained. On 

 photographing a spectrum through this screen it was found that the prin- 

 cipal absorption band had broken up into two, placed symmetrically with 

 respect to the band shown by the fluid. One of these doubtless belongs 

 to the ordinary, the other to the extraordinary ray. The centres of the 

 bands are at wave-lengths 36 and 46, while the centre of the single band 

 possessed by the liquid is at A = 43. 



In Figure 3, Plate 3 we have two photographs of this double band. 



The absorption of the liquid nitroso does not differ very materially 

 from that of its solution in glycerine, which I have described in a previ- 

 ous paper on screens transparent only to ultra-violet light. The absorp- 



