WOOD. NITROSO-DJMETHYL-ANILINE. 53 



V 



an adjustable slit inserted in its place. The spectrum can be made to 

 pass across this slit by turning the micrometer screw, aud by noting the 

 readings when known lines in the spectrum fall on the slit, the instru- 

 ment can be calibrated. This is easily and quickly accomplished by 

 means of an ingeniously arranged microscope which can be thrown into 

 and out of position as desired. A small lens carried in a tube in front 

 of the slit focuses the monochromatic lisdit which issues from it on the 

 slit of the spectrometer. 



The dispersion of the nitroso was measured in the following manner. 

 The prism having been set at minimum deviatiou, and the usual adjust- 

 ments made, the deviated image of the slit, illuminated in monochromatic 

 light, was brought into the field of the telescope. The gas flame was 

 then moved out of position and the current of hot air stopped. As the 

 fluid nitroso cooled the deviation increased, the reading being taken just 

 at the point of solidification, when the image disappeared gradually owing 

 to the crystallization of the medium. The warm air was then turned 

 on again and a second reading taken in the same way. A number of 

 prisms were used, the angles varying from one to ten degrees, those 

 of small angle being necessary when working with the bluish green, 

 owing to the absorption. In this way the dispersion was measured 

 between the extreme red aud wave-length .00050, below which point 

 the prisms refused to transmit sufficient light to make readings possible. 

 The values obtained in this way are given in the following table. 



The results are shown graphically in Plate 1, together with the dis- 

 persion curve for* bisulphide of carbon for the same region of the 



