199 
so as to obtain uniform surfaces. These surfaces were then 
dried, the colorimeters employed were the same as in the 
last experiments. I looked at Wb through a column 3 c.m. 
long. I held the other cylinder over Wa and endeavoured 
to get the same tint. The mean of two columns, one 
probably a little too long and the other probably a little too 
short, gave 8 ’2 as the proper length. So I made the column 
of this length, I thought that the tints were the same. Now 
if the law hold with regard to turbid solutions, if we increase 
both columns by the same length, the tints will again 
correspond. I took 4 c.m. as the common increment. The 
lengths of the columns were made 7 and 12*2. I thought the 
resulting tints equal. The lengths of the columns were 
made 11 and 16*2. I thought they were about equal, possibly 
Wa slightly lighter. My impression varied a little with 
the illumination of the surfaces. The lengths of the 
columns were now made 15 and 20*2, the tints seemed 
about equal, possibly Wa slightly lighter. The difference 
if any must be small, as is shown by the following experi- 
ment. Taking Wb seen through a column 15 c.m. long, as 
the standard tint, I endeavoured to get the same tint with 
Wa seen through the other cylinder. For the lower limit I 
got 19 '9 c.m., and for the upper limit 21-0 c.m., the mean of 
these is 20*45, not far removed from 20*2. Afterwards I 
thought a column of this length satisfied. I now made the 
lengths of the columns 19 and 24*2. I thought that the 
tints again corresponded. 
Some experiments of a similar nature were made with a 
stronger diffusion, it contained 3*747 cub. c. in 500 cub. c. 
The tints of grey used were the same as in the last experi- 
ments. The standard tint at the commencement was Wb 
seen through a column 4 c.m. long. To get a similar tint 
with Wa one determination for the upper limit gave 6*5, 
and one determination for the lower limit gave 6*1. The 
mean of these is 6*3. A column of this length seemed to 
