WOOD. — NITROSO-DIMKTHYL-ANILINE. 65 



ciirred in the extreme visible red. It was not sulRcieutly intense to 

 interfere with measurements, and whether it was due to the presence 

 of a weak, diffused baud, siinihir to the one at A = 25, or to a stroujf 

 band of metallic absorption, I was unable to say, the observations being 

 confined to its extreme edge. I expect to investigate this point, as soon 

 as our apparatus for the study of the infra-red region of the spectrum is 

 in working order. I hope to be able, in the near future, to make a 

 determination of the dielectric constant of the substance. With these 

 points thoroughly investigated, a more rigorous application of the dis- 

 persion formula to the results will be possible. I feel less hopeful of 

 getting satisfactory data regarding the absorption, but shall at all events 

 make the attempt. 



The Dispersion of Toluine. 



In examining the ultra-violet absorption of some organic solvents, I 

 found that toluine has a fairly strong and narrow absorption band at 

 wave-length .00027, in addition to the band of metallic absorption 

 below .00020, which is the one which chiefly influences the dispersion 

 of the substance. It occurred to me that it would be interesting to 

 determine the dispersion of the substance through this region, and see 

 to what extent the curve was modified by the weaker band. So far as I 

 was able to find, the ultra-violet dispersion of toluine has never been 

 determined, which is not very surprising, since only very acute prisms 

 transmit anything below the band above mentioned. The same method 

 was employed as in the case of nitroso-dimethyl-aniline, namely the 

 crossed prism method adapted to the quartz spectrograph. 



The dispersion was first measured in the visible spectrum with a 

 hollow prism of 60^ angle. Small prisms were then made of quartz 

 plates with the toluine between them, the angles varying from two to 

 six degrees. The angles were determined from the photographs of the 

 deviated spectra obtained with the quartz spectrograph, using the data oli- 

 tained with the spectrometer itf the blue region of the spectrum. Figure 6, 

 Plate 3, shows one of the deviated spectra, the ultra-violet being to tlie 

 left, and illustrates well the general form of the dispersion curve in this 

 region. Figures 7 and 8 show the deviated and undeviated spectra, the 

 one above the other, obtained with prisms of two different angles. It is 

 at once apparent that the absorption band at wave-length .000271 modi- 

 fies the curve to no small degree. On the original negatives a very 

 slight trace of oppositely curved branches, such as are always present at 

 the edges of strong absorption bands, is discernible. The baud is quite 



VOL. XXXIX. — 5 



