ON THE BOILING POINTS OF AQUEOUS SOLUTIONS OF ELECTROLYTES. 211 



using distilled water, which was redistilled from a flask containing a little potassium 

 bichromate and sulphuric acid into one containing a small quantity of barium hydrate, 

 thence to a condenser with Jena glass worm to a Jena glass flask for its reception. 

 The water was kept in a Jena glass bottle. By the method above described for water 

 purification, a water was obtained of specific conductivity of "23 x 10" 6 , Kohlrausch 

 units cm. -1 ohm"'. During the time the experiments were being made, the results of 

 which are given, a register of water purity was kept, its electrical resistance being 

 frequently measured as a safeguard. 



In conductivity work a Wheatstone bridge, after the Kohlrausch pattern for 

 electrolytes, with telephone as current indicator, was used. The bridge wire was 

 calibrated by the Strouhal and Barus method,* and the electrodes of the resistance 

 tubes were platinised.t 



Observations were made by the method already given (page 196), to determine the 

 amount of vapour present in the tube during an experiment, and the change in con- 

 centration allowed for. 



It was assumed that in finding values of — for the boiling point, the ratio of 



the dilutions at 100° of two solutions might be taken equal to the corresponding ratio 

 at 15°. Experiments were made to determine the error involved, which was found 

 to be no more than - 03 per cent, even for a sevenfold normal solution. 



When making up solutions, a table of solubilities was consulted, to ascertain the 

 composition of a solution saturated at 15° Centigrade. Five hundred or a thousand 

 cubic centimetres of solution were made up of this concentration, which was done 

 by weighing so many gramme equivalents of salt according to the concentration desired, 

 placing it in a flask, dissolving in water, and bringing the solution to the standard volume 

 500 or 1000 cubic centimetres at 15° Centigrade. The various dilutions were obtained 

 from this by the addition of solvent to a measured number of cubic centimetres of 

 solution. For this 20, 25, and 50 cubic-centimetre pipettes were used, and 100, 200, 

 300, and 500 cubic- centimetre flasks, all of which had been tested at ' Charlottenburg. ' 

 In addition I tested the pipettes myself. One of these dilutions, usually the tenth 

 normal, was tested quantitatively as a safeguard against error. 



It will be noted that I have found conductivity values at very high concentrations, 

 indeed in some instances beyond the solubility at 15° Centigrade. These were obtained 

 for a reason which will appear in Part IV. of this paper, as follows : I desired the conduc- 

 tivity at a certain concentration beyond the solubility at 15°, but within the solubility 

 at 100° Centigrade. In each case I calculated the amount of salt required to the 

 amount of solvent to be used in the experiment, and placed it in the boiling resistance 

 tube, adding the solvent afterwards. This was then heated to the desired temperature, 

 and the reading taken. Several such experiments were made for various salts ; the 

 results so obtained are marked with an asterisk. 



* Wied. Annal., 10, 326 (1880). t Zeit. phys. Chemie, 21, 297. 



