1910-11.] Absorption of Light by Inorganic Salts. No. II. 535 
One advantage of this method is its simplicity and the very steady 
temperatures obtained by it. The bottoms of the moulded cells were, 
however, not plane enough for photometry, and these cells had to be given 
up. Square glass boxes were then tried, one inside the other, and some 
results were obtained. These boxes were, however, not very satisfactory, 
and the figures recorded in this paper were obtained by the simple method 
used in the infra-red. The apparatus illustrated in the diagram is being 
altered at present, and will be employed in a more accurate investigation of 
the effect of temperature and density which Mr Brown is at present 
undertaking. 
In the simple method, then, which was chiefly employed here, the lower 
slit was kept free. The cell with the solution was placed before the 
upper slit and the difference of the nicol readings in two adjacent quadrants 
taken (2d). The cell was then heated, placed again before the upper slit, 
and the difference of the nicol readings in two adjacent quadrants again 
taken (2d')- Let A' denote the molecular extinction coefficient at the 
higher temperature. Assume that 2d, 2d' are the angles less than 90°. Let 
s be the fraction of the incident light transmitted by the glass cell when 
filled with water only, and let cr be the ratio of the intensity of the light 
falling upon the upper slit to the intensity of the light falling upon the 
lower slit when both slits are free. Then 
sorlO~ Acd = tan 2 # ; scrlO~ A ' cd = tan 2 0'. 
Hence 
i qccha'-a) _ ta . n 2 0 
tan 2 #'’ 
/ 2 
or A' - A = — (log tan 0 - log tan O'). 
As c and d are known, the increase in A is thus easily determined. 
In order to eliminate irreversible changes, the mean of repeated heatings 
and coolings was taken. The average temperature of the hot solution was 
