some Copper Salts in Aqueous Solution. 331 



The final value of (1— p) is calculated from two photo- 

 metric measurements of the same kind; one, the correction 

 for the apparatus, the other the measurement of the light 

 which passes through the solution. These will both be sub- 

 ject to an error of the magnitude given in column Y. The 

 probable error of the final result will therefore be double this 

 amount. The numbers in column Z represent this final error, 

 calculated as percentages of (1 —p). The large error in the 

 violet is due to the small intensity of the light used in that 

 part of the spectrum, and also to the fact that the eye is less 

 sensitive to differences in the intensity of light in the violet 

 than in the middle of the spectrum. At the red end of the 

 spectrum a special error makes itself felt. The measurements 

 are only theoretically accurate, when the part of the curve 

 representing the intensity of the light in the part of the spec- 

 trum in which they are made may be considered as a hori- 

 zontal line. When the absorption-curve of the substance 

 under examination rises or falls rapidly this is no longer the 

 case, and for this reason the measurements on the copper 

 salts in the red end of the spectrum are less accurate than in 

 other parts. 



My first measurements on the absorptions by copper sul- 

 phate showed a lack of agreement between the numbers ob- 

 tained with the same solution when different lengths of layer 

 were used. This was found to be due to the light not being 

 parallel, as the differences disappeared when parallel light 

 was used. Something similar appears to have been observed 

 by other investigators. Vierordt* found (with a lamp 15 

 centim. from the slit, and distilled water in the trough) the 

 ratio of the intensity of the light which had passed through 

 the empty half of the trough to that which had passed through 

 the water to be 08 ; the calculated number is 0*936. Kriiss f 

 found the ratio in a Schulz cell 0*905, that calculated being 

 0*993. I also found the ratio in a Schulz cell 0*91, when the 

 light was not parallel. 



The cause of these differences may be found in the fact that 

 placing a layer of a medium with a greater refractive index 

 than air in the path of the light has the same effect as bring- 

 ing the light nearer to the slit through a distance which is 



given by the formula ll J, where I is the thickness of the 



layer of the medium, and (jl its refractive index. It must also 

 be remembered that, so long as the collimator is filled with 

 light, an alteration in the distance of the lamp makes no dif- 

 ference in the intensity of the light in the spectrum. The 

 * Anwendmiy &c. pp. 5 and 21. f Kohrimelric, p. 153. 



