1911-12.] Absorption of Light by Inorganic Salts. No. VI. 59 
The above calculation would seem to show that in the red phase each 
cobalt chloride molecule is associated with approximately fifteen molecules 
of water. This is the phase existing in dilute aqueous solution, and I shall 
meantime term it the polyhydrate. 
The absorption spectra of three definite and distinct phases of cobalt 
chloride, viz. the anhydrous salt, the hexahydrate, and the polyhydrate, 
have now been investigated ; and in the cases of the first and the last, values 
of A have been determined. No traces have so far been found of any other 
hydrate. With these data at my disposal I have returned to the investiga- 
tion of the saturated aqueous solution. 
In the first place, its spectrum though not simply that of the polyhydrate 
bears a great resemblance to it. The maximum of the band in the green 
lies practically at the same wave-length, although the values of A are in- 
creased and the shape differs slightly. There is a small band with maximum 
at X = 680 miu, probably indicating the presence of anhydrous salt. Obviously 
A 
the concentration of this phase in solution cannot be more than ^ or 
ZoZ 
about j-J-o of the total concentration, and this would not account for the 
change in the band in the green. 
W. N. Hartley in his investigation of absorption spectra found that 
5 c.c. of water dissolved 8 gms. of the hexahydrate at 16°, the room 
temperature of this research, in forming a saturated aqueous solution. 
This gives the proportion of 143 molecules of water to 1 molecule of cobalt 
chloride, and would enable most, if not all, of the salt to enter the polyhydrate 
phase. This fact and the similarity of the bands in the green for the dilute 
and saturated solutions have induced me to assume meantime that at least 
nine-tenths of the salt does go into the polyhydrate phase. 
In fig. 3 I have plotted A as a function of X for dilute and saturated 
solutions. The values of A for the dilute solution are taken from Paper II. 
of this series, and the values for the saturated solution are from this paper. 
These are connected by curves X and Y respectively. I then deducted 
nine-tenths of the ordinates of X all over the curve from the corresponding 
ordinates of Y. The remainders obtained are shown in the heavy line, and 
are a rough indication of the absorption of the remaining tenth of the 
chloride in solution. The curve shows clearly a band with maximum at 
X = 550 wul, and less clearly a band about X = 490 /x/x. It is to be noted that 
the more distinct band of the hexahydrate crystal, viz. that at X = 550 /x/x, is 
the more distinct here. 
We thus have evidence of the presence to a slight extent of the 
hexahydrate in the solution. This conclusion has the support of those 
