NOYES AND COOLIDGE. — ELECTRICAL CONDUCTIVITY. 187 



constant decrease in the conductivity of the upper chamber), the conduc- ' 

 tivities between the walls of the bomb and both the upper and lower 

 electrodes are measured carefully, and the temperature of the bath is 

 observed. The experiment is then complete. 



The ratio of these two conductivities is calculated, and the corre- 

 sponding volume is obtained by interpolation from a plot obtained as de- 

 scribed in Section VI! This volume, corrected for the expansion of the 

 bomb and then divided by the weight of solution in the bomb, gives the 

 specific volume of that solution at the temperature in question, and this 

 quantity divided by the specific volume of the solution at 4° gives the 

 volume of that quantity of the solution that would at 4° occupy one 

 cubic centimeter, this quantity being most convenient in subsequent 

 computations. 



VI. Standardization of the Apparatus. 



TJiermometers. — These were calibrated by the method recommended by 

 Crafts,* first for irregularities of bore and then at the fixed points 0°, 

 100°, 218°, and 306°. For the last two temperatures the vapors of boil- 

 ing naphthalene and of benzophenone were used. These substances 

 were obtained from Kahlbaum, and were purified in the manner sug- 

 gested by Crafts, until their melting points came within his limits. The 

 form and dimensions of the vapor bath used in establishing the 218° and 

 306° points were essentially those given by him. To reduce the temper- 

 atures lying between the fixed points to the gas scale, his table of correc- 

 tions for French glass was also used, our thermometers being of the same 

 make as those used by him. At the beginning every temperature meas- 

 urement was followed by a lag ice-reading ; but this was found to be un- 

 necessary, since the thermometers showed no lag. The ice-reading was, 

 however, taken frequently, to make sure that the zero did not change from 

 another cause, — evaporation of mercury from the bulb below and con- 

 densation in the chamber above. This effect was not apparent even after 

 long use at temperatures up to 280°; but above this the zero would fall 

 perhaps 0.1° from two or three days' use. When in use at 306°, instead 

 of taking an ice-reading the thermometer was first placed in the benzo- 

 phenone heater and then in the calibrating apparatus containing perfectly 

 pure benzophenone. The difference in reading (usually amounting to 

 0.1-0.2°) was deducted from the true boiling-point of benzophenone at 

 the observed barometric pressure. 



* Am. Chem. Jour., 5, 307-338 (1883-84). 



