138 
a result considerably remote from tlie correct value. Also 
when the disc was placed inside the solution at the depth 
given by theory, it seemed too dark. Next, a comparison 
was made by looking through the cylinders at external 
white surfaces. Standard solution 4000 in 500 cub. c. of 
water, length of column 8'4. 
ABC 
6000 5-1 6588 
Thus the result is considerably different from the real value. 
Also both experiments concur in giving too high a value. 
The theoretical depths moreover were tried with external 
surfaces, and seemed slightly too great. My next experi- 
ments were made with solutions containing 2400 and 1600 
of sulphate of copper. The discs being inside the results 
were as follows : standard solution 1600 in 500 cub. c., depth 
of disc 8*3. 
ABC 
2400 5-6 2371 
The number under B was the result of seven trials. The 
value under C is a fair approximation to the real value. I 
also tried external surfaces with these solutions. Standard 
solution 1600 in 500 cub. c., length of column 8*4. 
ABC 
2400 6*2 2161 
somewhat further from the correct value than we might 
expect. I also tried the theoretical length of column ; with 
external discs it appeared a little too light. The results 
obtained with the last standard solution are inconsistent 
with the results obtained with the first. In the latter case 
the results are too low, while previously they were too 
high. Errors of observation arising from imperfect percep- 
tion of colour, from imperfection of instruments and 
unfavourable conditions of light (for many of the experi- 
ments were made during the winter months, sometimes on 
gloomy unfavourable days) would no doubt contribute to 
