1910-11.] Absorption of Light by Inorganic Salts. No. IV. 555 
can, therefore, only be due to something vibrating in the cobalt atom. The 
height of the bands varies somewhat in each case, this being probably due 
to the acid radicles with which the metal is associated. Again, in the 
nickel salts we have in each case the same three maxima at A = -40 5/ul, 
A = *690yoc, and A = l’210 / a. These bands can, therefore, be due only to some- 
thing in the nickel atom. Of course in cobalt iodide the band at A = ’510 / u 
and in nickel iodide the band at A = '405//, are swamped by the bands due 
to iodine. 
If now we look for an effect due to the acid, we can trace none in the 
case of the sulphates. In the sulphates there is no band we can attribute 
to the S0 4 radicle, though it probably influences the height of the bands 
due to the metal. The chlorides and fluorides each show indications of a 
band in the extreme ultra-violet just off our range ; we can trace the 
presence of this band in the rapid increase of the value of A. This increase 
is more marked in the fluoride than in the chloride. In the chloride there 
is also a weak band at A = '330/x, which has already been noticed by Jones 
and Anderson.* Both the bromides indicate a band off our range and a 
decidedly interesting one at A = ’28 5/ul, which attains the same height in each 
case. It has apparently not been observed before. 
The case of the nitrates and iodides is, however, by far the most 
interesting. Owing to the large values of A, the parts of the curves in 
the ultra-violet could not be shown well on the same scale as the rest. 
They are, therefore, plotted separately in the two following diagrams, the 
values for the cobalt salts being shown by o s and for the nickel salts by x s. 
The two nitrates are almost identical. The band at '300/x has, of course, 
been observed by Hartley and by Jones and Anderson, but no measurements 
of A have previously been made on it. The iodides both show three bands, 
situated respectively at the end of the range, at A = '292 ju and at A = ”3 55/x. 
The heights for nickel iodide are about five times the values for cobalt 
iodide. It should be stated that, owing to their unstable nature, the values 
of A for the iodides are not so reliable as for the other salts, especially at 
the low concentrations employed in this part of the spectrum. The values 
given are, it is hoped, free from serious error, as they are each the result 
of three or four careful independent determinations. For purposes of 
comparison the values of A for a saturated aqueous solution of iodine are 
shown by means of n’s on the same diagram. In determining the absolute 
value of A for iodine some trouble was experienced owing to the latter 
vaporising from the solution. To show the effect on A of the decomposi- 
tion of cobalt iodide a dotted curve is shown which represents the values 
* Jones and Anderson, Publication of the Carnegie Institution, No. 110. 
