OF LIGHT BY TURBID SOLUTIONS. 195 



Exp. XXIII. — Comparison-diffusion contains 28 8 cnb. 

 c. in 500 cub. c. 



A B C D 



i-o i"i I "05 °'^^ 



The tint given by the theoretical column I thought slightly 



lighter. 



Exp. XXIV.— Comparison-diffusion contains 33-6 cub.c. 



in 500 cub. c. 



A B C D 



0-88 0-9 0-89 076 



Tlie tint given by the theoretical column I thought slightly 



lighter. 



Exp. XXV.— Comparison-diffusion contains 38-4 cub. c. 



in 500 cub. c, 



A B D 



073 0-83 078 0-66 



The tint given by the theoretical column I thought slightly 

 lighter 3 they were^ however^ very nearly the same. 



This second series of experiments confirms the result of 

 the first series, that there is a slight departure from the 

 rule when the strength of one diffusion is several times a 

 multiple of the other. Both series show a tendency to 

 make the column in the comparison-cylinder slightly too 

 long. We must not hastily conclude that in these cases 

 the theory is not applicable. It seems to me that we should 

 expect such a result ; for the medium is not perfectly trans- 

 parent, and in one cylinder we have a column of this medium 

 (water) several times a multiple of the length of the column 

 in the other. This would require some slight compensation. 

 In these experiments I have taken the lower level of the 

 meniscus as the proper reading. In some of the stronger 



o2 



