12 



Studies on Solution. 



TABLE 1. Nickel 

 Chloride in Water. 



Wave-length 434jt;*. 



A very important quantitative study of the light-absorption of solu- 

 tions has been carried out by Houstoun and his co-workers 1 (the refer- 

 ence is to the last of a series of eleven papers) . Many phases of the gen- 

 eral problem were considered and frequent reference will be made here 

 to the separate papers. His work is unique hi that it is the only record 

 we have of the determination of A for solutions for wave-lengths in the 

 infra-red. Even this work, although of a more complete character 

 than any of the researches hitherto attempted, did little more than 

 touch upon the relation between A and c. The absorption curves 

 were determined for the region of the spectrum from 645juju to l,270jujt 

 for the chloride, bromide, iodide, nitrate, and sulphate of cobalt. 2 This 

 was done for a strong and for a dilute aqueous solution of each salt, 

 in all cases the values of A for the more concentrated solution were found 

 to be greater than the corresponding values for the dilute solution. 



Houstoun also made a further study of the chloride and bromide of 

 cobalt, nickel, iron, and copper. 3 Solutions of each salt were prepared 

 varying in concentration from saturation to moderate dilution. A was 

 determined for a single wave-length lying on the edge 

 of an absorption band. The results for nickel chlo- 

 ride as an example are given in table 1. The values 

 of A are seen to decrease with dilution reaching a 

 minimum value, and then to remain fairly constant. 

 Table 1 and other similar tables show that A increased 

 again for the more dilute solutions. This increase was 

 considered either as within the error of experiment or 

 due to the chemical change taking place in the solution. 



In all of Houstoun's work A was determined by 

 comparing a cell containing the solution with a cell of 

 the same thickness containing the pure solvent. This 

 method is open to criticism, but the difference be- 

 tween the A thus determined and the true value was probably less 

 than the errors in the values of A due to other experimental causes. 



APPARATUS. 



The apparatus used for determining the coefficient of light-absorp- 

 tion has been developed by previous workers in the Johns Hopkins 

 laboratory. The quantitative work was begun by Guy, 4 who built a 

 sensitive radiomicrometer and used this in connection with a glass- 

 prism spectrograph. The apparatus was greatly improved by Shaeffer 6 

 during the following year, and the apparatus used in the present inves- 

 tigation and described in this paper is the same in all respects, except 

 for minor details, as that used by Shaeffer and his co-workers. 



'Proc. Roy. Soc. Edinburgh, 33, 156 (1912-13). *Ibid.,31, 521 (1910-11). *Ibid.. 33, 147(1912-13). 

 'Carnegie Inst. Wash. Pub. No. 190, 29 (1913). 6 /Wd. t 230, 44 (1915). 



