52 
Proceedings of the Royal Society of Edinburgh. [Sess. 
temperature, and that the temperature had reached a constant value some 
time before the absorption reached a corresponding constant value. Probably 
the lag exists in falling from a high temperature as well as in rising to it. 
Hence what we obtained by the former method as the value of A when 
the temperature was at 80° may be the constant value of A for a 
temperature of 86° or 88°. The values of A given in my table are, with 
the exception of the last three, constant values for a temperature of about 
83° or 84° ; the last three were got as the temperature was rising, it being 
impossible to get values of A for higher temperatures with the same 
thickness of solution for these wave-lengths on account of the great increase 
in the value of A. 
As this effect of lag tended to make my results less reliable, and 
evaporation from the heated solution, though reduced greatly and to a 
certain extent eliminated, had not been quite prevented, I gave up the hope 
of obtaining measurements of the progress of the change at different 
temperatures, accurate enough for the determination of the law which it 
followed. 
It is a well-known fact that solutions of cobalt chloride in ethyl alcohol 
are blue, and that as water is added the colour gradually changes through 
various shades till it reaches the colour of an ordinary aqueous solution. 
This change seems to be of exactly the same nature as that produced in 
the aqueous solution by heating. It also allows the easy and accurate 
measurement of the intermediate stages. I therefore turned my attention 
to investigating it. 
In the first place, I made a mother solution, almost saturated, from 
which all the others were derived. Some hexahydrate chloride was care- 
fully heated in a crucible till it lost all water of crystallization, passing- 
through the dark-blue dihydrate state to the light-blue anhydrous state. I 
had then 123 grammes of the anhydrous salt which I dissolved in 90 c.c. of 
absolute alcohol of specific gravity ‘794, obtaining 92 c.c. of solution. The 
specific gravity of the solution was '91, and the concentration of the chloride 
in gramme-molecules per litre was T028. This is solution I, and from it I 
made solutions II, III, IV, and V by adding further quantities of absolute 
alcohol. Thereafter, by adding water to the alcoholic solutions, I made 
from I solutions la, 16, Ic, Id, le, I/, I g, Ih ; from II solutions Ila, 116, II c, 
II d, lie, II/; and from III solutions Ilia, III6, IIIc, III d, the water being- 
added in such proportions that the concentration of the cobalt chloride in 
solution, measured in grm.-mols. per litre, decreased regularly in each series 
by one-twentietli or multiples of one-twentieth of its value in the purely 
alcoholic solution of that series. 
