Section jj. Preparation of the Solutions. 73 







outer cylinder with water, which was heated by a coil traversed by high- 

 pressure steam, the free steam thus produced surrounding the sides of the 

 bath, and returning through an outside condenser to the bottom. The 

 bomb was immersed in the liquid xylene, which was continually stirred by 

 a propeller. The temperature was always constant to 0.01 and about 0.1 

 below the boiling-point of water at the prevailing pressure. 



Vapor baths were used at the other temperatures. Brombenzene was 

 used as a boiling substance for 156, and naphthalene for 218. The 

 heaters were of the form described in section 3, Part II. Bromnaphtha- 

 lene and benzophenone were used as boiling substances for the tempera- 

 tures of 281 and 306. The heaters used at these higher temperatures 

 were of the rotating type described in section 29, Part III. An inverted 

 mica cone was used to prevent the condensed vapor from dripping onto 

 the top of the bomb. 



The procedure in the conductivity measurements was essentially the 

 same as that pursued in the previous investigation. The bomb was ordi- 

 narily rinsed thoroughly with the solution about to be used ; but when a 

 specific-volume measurement was to be combined with the conductivity 

 measurements, the bomb was rinsed twice with alcohol and then with 

 ether, and finally a weighed quantity of solution added. The solution was 

 not in any case boiled to expel the air; but, before heating, the air pressure 

 was reduced to 3 to 5 cm. mercury. In the first measurements with cell 

 11, this was not done until after the measurement at 18, in order to pre- 

 vent the formation of air bubbles on the electrode, which could only be 

 dislodged by shaking. The use of the rotating carriage in the later 

 experiments allowed the pressure to be reduced before the measurement at 

 18 was made. 



33. PREPARATION OF THE SUBSTANCES AND SOLUTIONS. 



The sodium and potassium chlorides were purified by precipitation with 

 hydrochloric acid, the latter being recrystallized from hot water. The salt 

 was finally dried and ignited. 



The potassium sulphate was purified by repeated crystallization from 

 hot water ; it was then dried and ignited. 



The silver nitrate was recrystallized once from warm water and care- 

 fully dried in a current of dry air at 140. The salt was always kept over 

 calcium chloride in the dark. Portions of the salt were dissolved in water, 

 the silver was precipitated as silver chloride, and the precipitate was 

 washed and dried. The calculated percentages of silver nitrate for the 

 first lot of salt analyzed in February, 1904, were 99.86, 100.01, and 99.88 ; 

 mean 99.92. The second lot analyzed in April, 1904, gave 99.78, 99.82 

 and 99.89, and in July, 1905, gave 99.88 and 99.84, the mean being 99.86. 



