184 PROCEEDINGS OF THE AMERICAN ACADEMY. 



stant, which is indicated by the constancy of the conductivity, the bomb 

 is removed from the bath, shaken, and returned as quickly as possible. 

 To shake it wliile hot, a pieoe of asbestos cloth, with a piece of woollen 

 cloth outside, is used. If the shaking is omitted, the measured conduc- 

 tivity may be too high by as much as 0.5 per cent. This was found to 

 be due to the following facts: At 140° there is still a considerable 

 vapor space left in the bomb, the entire cover being above the liquid 

 surface. During the first part of the heating the xylene vapor is con- 

 densed so rapidly by the bomb that it extends up only for a little 

 distance above the bottom of the bomb, leaving the upper part com- 

 pletely out of it. This causes an evaporation of pure water and a con- 

 densation of it all over the colder cover, leaving the solution too 

 concentrated. If the bomb is shaken after reaching the temperature of 

 the bath and quickly returned, the same action does not repeat itself, 

 since the top is now as hot as the bottom. The bridge readings are 

 now continued (usually for about thirty minutes) till one perfectly 

 constant for ten to fifteen minutes is obtained. 



The bomb is then transferred to the naphthalene bath. Shaking was 

 found to have no effect at this temperature, owing, doubtless, to the fact 

 that the liquid level has then risen almost to the cover, so that large 

 drops cannot adhere to the latter, and to the fact that the surface tension 

 has diminished, so that less water is held clinging to the walls of the 

 narrow chamber in the cover. 



The bomb is next brought into the 281° bath. It is now necessary to 

 keep constant watch of the conductivity between the upper auxiliary elec- 

 trode and the walls of the bomb, so as to be sure that too much solution 

 has not been put into the bomb. The reasons for putting in solution 

 enough at the start to so nearly fill the bomb at the highest temperature 

 are first, to reduce the va2')or space at all the temperatures as much as 

 possible, since a correction has to be m:;de for the amount of solvent in 

 this space; and secondly, to see that the bomb is absolutely tight even at 

 the highest temperature, when the solution is in contact with the upper 

 electrode. This latter is important, since a leak, if it took place above 

 the liquid level, would cause a loss of pure solvent and a consequent 

 increase in the concentration of the solution. 



After completing the measurements at 281°, the bomb is returned to 

 the 218° bath, then to the 140° bath, and finally it is brought back to 

 26°. In going from a higher temperature to a lower much time is saved 

 by cooling the bomb, in front of a fan outside the bath, to a temperature 

 which is at least as low as that next desired ; for while heating in a 



