66 PROCEEDINGS OP THE AMERICAN ACADEMY. 



narrow and fairly strong, yet it does not seem to modify the curve to 

 any great degree on the red side, — at least the curve has nothing like 

 the steepness which it has in the remote ultra-violet. On the other side 

 of the band, however, the curve is seen to be depressed to a considerable 

 distance. The influence of the band seems, therefore, to be uusymmetri- 

 cal. This is to be ascribed, I believe, to the fact that the absorption on 

 the ultra-violet side of this band is quite strong, as will be seen from 

 Figure 8. 



The plates were measured in the same manner as the nitroso plates 

 and a table of the refractive indices is given below. 



The curve is shown graphically on Plate 1. 



Applying the dispersion formula to the results we find, neglecting the 

 band at .00027 for the centre of the band of metallic absorption in the 

 ultra-violet, the value .000182, which is not inconsistent with the photo- 

 graphic records. For the other constants we find m = 1.37 and m' = .77. 



The sum of these two constants should give us the dielectric constant 

 of the substance if no other bands of absorption are present. The sum 

 m + m' = 2.14, while the dielectric constant is 2.36 (determination by 

 Palaz), from which it seems probable that there is an absorption band in 

 the infra-red. 



The formula expresses the dispersion fairly well except for a narrow 

 range immediately on the more refrangible side of the band at wave- 

 length .00027. Adding another term to the formula will not help 

 matters much in this region, in my opinion, owing to the indefinite nature 

 of the absorption. 



The substance seems to be worthy of a more complete investigation. 



