1910-11.] Absorption of Light by Inorganic Salts. No. III. 539 
posed, giving the chocolate-brown precipitate referred to. The fluoride 
was a light green powder sparingly soluble in cold water. It was taken 
as NiF 2 3H 2 0. 
The following table gives the results for the visible spectrum : — 
Values of A for Nickel Salts in the Visible Spectrum. 
A. 
d = 2 cm. 
d=l cm. 
c= *140. 
c = 2‘27 ; 
first five and 
last two pts. 
jr strength. 
c=l-40 ; 
first four 
and last pt. 
| strength. 
First ten 
pts. c = '488, 
next three 
strength, 
last four 
jV strength. 
c = l-56 ; 
first five 
and last pt. 
^ strength. 
c = 1-398 ; 
first three 
and last pt. 
strength. 
Fluoride. 
Chloride. 
Bromide. 
Iodide. 
Nitrate. 
Sulphate. 
717mm 
1*83 
2*07 
2-06 
217 
1-87 
2-03 
687 
1-59 
1-88 
1-91 
1-99 
1*65 
2-04 
653 
1-57 
1-78 
1*76 
2-00 
1-57 
1-93 
625 
1*24 
1-20 
1*22 
1-66 
1*16 
1-27 
602 
•71 
•68 
•72 
1*41 
•64 
•70 
582 
•39 
•353 
•40 
1-48 
•363 
•41 
563 
•27 
•204 
•237 
1-80 
•206 
•21 
547 
•30 
137 
•172 
2*10 
•139 
•15 
529 
•22 
•100 
•141 
2-58 
•089 
•12 
514 
•24 
•064 
•108 
3*10 
•063 
•07 
499 
•19 
•049 
•082 
4‘5 
•046 
•057 
486 
•26 
•108 
•120 
7-0 
•083 
•102 
475 
•32 
•222 
•237 
11*6 
•174 
•174 
463 
•45 
•390 
•39 
13-2 
•299 
•341 
453 
•61 
•580 
•52 
18-8 
•427 
•460 
444 
•68 
•62 
•73 
29-0 
•58 
•63 
434 
M2 
M2 
337 
1-04 
•88 
In every case except the sulphate the values in the table are the mean 
of two independent determinations. For the sulphate only one determi- 
nation was made. The values are not for the same value of c through- 
out the spectrum. Towards the ends where the absorption is greater 
the solutions were diluted. It would have been less open to objection 
to keep c constant and alter d, but this would have entailed more trouble. 
The variation of A with c for these points was investigated after- 
wards. Owing to the solutions being so weak, the values for the 
fluoride did not agree nearly so well with one another as in the case 
of the other salts. The values of A for adjacent wave-lengths are, of 
course, fully independent; the successive strips of spectrum taken do 
not overlap. 
Two of these solutions have been investigated with a spectrophotometer 
