ON THE HALOGEN HYDRIDES AS CONDUCTING SOLVENTS. 



141 



and it was only after months of failure that an apparatus was designed, by means of 

 which sufficiently accurate analyses were obtained. 



The apparatus consisted of two parts, the transferrer and the absorber. 



The transferrer (T, fig. 5, see p. 144) consisted of a wide H-shaped tube, with a 

 capillary tube a passing through three of its branches, the two tubes being sealed 

 together at b and &', as shown in the figure. By filling the space between the tubes 

 with a mixture of carbon dioxide and ether, the capillary between 6 and I' could be 

 cooled to 81. In fig. 5 the transferrer is shown when placed in the electrolytic cell. 



The absorber (fig. 2) consisted of a stoppered tube A, connected by C with the 



-e 



Fig. 2. 



Fig. 3. 



bubbler B ; this bubbler was so constructed that it was impossible for water to be 

 either ejected from the apparatus or sucked back into A. A smaller vessel b of the 

 same type was contained within the apparatus. The calcium chloride tube d, which 

 was provided with a stop-cock, was attached to B by a ground joint. 



The method of using the apparatus was as follows : A quantity of moist garnets 

 were first placed in the tube A, and a quantity of glass beads in the section c of the 

 absorber. The requisite amount of distilled water was then placed in the bubblers B 

 and b, and the tube d and stop-cock f were replaced. The apparatus was then 

 weighed, a glass counterpoise of approximately the same size, shape, and weight being 

 used. The tube A was next immersed in a mixture of solid carbon dioxide and ether, 

 and after the cap k, fig. 3, had been placed in position by means of the rubber cork m, 



