338 Dr. T. Ewan on the Absorption-Spectra of 



or nearly so^ and his experiments are therefore comparable 

 with mine. 



The changes in the absorption-spectra of the solutions which 

 occur when the concentration is changed are very clearly 

 visible in the curves which represent the absorption-ratios. 

 These are drawn for convenience with the divisions of the 

 tangent-screw as abscissae (instead of wave-lengths), and the 

 values of the absorption-ratios as ordinates. Curves 1 and 2 

 belong to the strong and dilute solutions of copper sulphate, 

 3 and 4 represent the absorption-ratios of the copper-chloride 

 solutions containing about 4 and 2 equivalents in 1 litre, and 

 5 belongs to Cu(X0 3 ) 2 . The curves of the most dilute 

 solutions of the three salts lie very close together and are 

 almost identical with No. 2. The great similarity of the 

 absorptions of these three solutions is best seen, not by com- 

 paring their absorption-ratios, but by comparing the fractions 

 of the light transmitted by them, calculated for solutions 

 containing the same amount of copper, and for unit length of 

 layer. The comparison is carried out in Table XII. Com- 

 paring the absorptions in this way, it is possible to see 

 whether the differences between them are greater than the 

 experimental error or not. This is very difficult when the 

 numbers are calculated as absorption-ratios, as the errors in 

 the measurements are much exaggerated in parts of the spec- 

 trum where the absorption is small, and vice versa. 



Table XII.* 



Region in 

 Spectrum. 



6277— 599-4 

 599-4—574-2 

 574-2—551-6 

 551-6— 5341 

 534-1-518-2 

 518-2—502-3 

 502-3— 488-3 

 488-3—474-6 



0uS0 4 . 



CuCl 2 . 



0-2775 



0-3018 



0-5089 



0-5403 



0-7130 



0-7157 



0-8333 



0-8398 



0-9008 



09087 



0-9447 



0-9474 



0-9725 



09709 



0-9840 



09831 



CuS0 4 

 -CuCL 



-00243 

 -00314 

 -0-0027 

 -0-0065 

 -0-0079 

 -0-0027 

 +0-0016 

 4-00009 



^uNO„) 2 . 



0-3096 

 0-5430 

 0-6854 

 08254 

 0-9069 

 9513 

 09759 



CaCl 2 

 -Cu(N0 3 ) 2 



-0-0078 

 -0-0027 

 +0-0303 

 +0-0144 

 + 0-0018 

 -00039 

 -00050 



The numbers for copper chloride and copper nitrate are 

 very similar, the differences are scarcely larger than the 

 experimental error and vary in sign. The differences between 

 copper sulphate and copper chloride, on the other hand, are 

 nearly all in the same direction and sometimes considerably 

 exceed the error of experiment. 



* The numbers are calculated for solutions containing 1 equivalent in 

 1 litre. 



