some Copper Salts in Aqueous Solution. 321 



taining the concentrated solution. The diagram, fig. 1, will 

 make the arrangement sufficiently clear. 



Rff. 1. 



A B C is a trough having a movable water-tight partition of 

 glass at B. DE was the compensating-tube, full of distilled 

 water : F G the tube destined to contain the concentrated solu- 

 tion. A B and D E were each 888 millim. long, and B C and 

 F G each 102 millim. The light coming from a large Argand 

 burner was rendered approximately parallel by means of 

 a lens. The two bundles of light, after traversing the 

 solutions, pass through the glass cube H, by which means 

 the two spectra are seen, the one above the other, and only 

 separated by a narrow black band due to the edge of the 

 cube, which was not quite sharp. The experiments were 

 carried out as follows : — The tubes B C and F G were filled 

 with a strong solution of, say, copper sulphate, and the tubes 

 AB and DE with distilled water. The volumes put into 

 A B and B C were proportional to the lengths of the two 

 parts. The lamp was then arranged so that the two spectra 

 appeared equally bright, and as nearly as possible the same in 

 every respect ; measurements of both were then made. The 

 movable partition at B was then removed, the solution in B C 

 mixed with the water in AB_, the partition replaced, and 

 measurements made again of the spectrum. By this 

 means any change due to dilution which occurred would be 

 detected. The absorption of the copper salts examined 

 consists of a strong general absorption in the red, and a 

 greater or smaller absorption in the violet and blue. The 

 readings were made by means of a vertical fibre of silk in the 

 eyepiece of the observing-telescope. The positions of the 

 extreme ends of the luminous spectrum were read, and also 

 the positions in which the silk fibre was just visible against 

 the luminous background of the spectrum. The spaces 

 between these readings were taken as the penumbra. The 

 measurements made on the divided arc of the spectroscopy 

 were reduced to wave-lengths by interpolation in a curve 

 constructed from measurements of various known lines, which 

 were repeated from time to time. 



