240 Conductivity of Aqueous Solutions. Part VIII. 



bomb (No. 3) with a similar electrode was used, the quartz cup in which 

 was 1.40 cm. in height and 1.37 cm. in diameter. This will be called 

 Cell in. 



CONDUCTIVITY MEASURING APPARATUS AND INDUCTOR. 



A slide-wire bridge of the roller type, described by Kohlrausch and 

 Holborn, and made by Hartmann and Braun, was used to measure the 

 conductance. The coils were of manganine and of 1, 10, 100, 1,000, and 

 10,000 ohms resistance. Each coil was compared directly in the Insti- 

 tute's testing laboratory with manganine standards having the Reichs- 

 anstalt seal and certificate. The slide-wire was calibrated twice by the 

 method of Strouhal and Barus. The corrections both to the coils and slide 

 wire agreed within the experimental error with the results of Kato 

 obtained a few months earlier. 



An ordinary interrupter was used. The minimum sound in the tele- 

 phone was very good except for the most dilute and most concentrated 

 solutions, and fairly good for them. 



HEATERS. 



For the work up to and including 156, a liquid bath of pseudocumene, 

 heated electrically by an inside and outside coil, and well stirred, was 

 used. Cooling was effected by running tap water through a copper coil 

 immersed in the bath. The temperature was regulated by the observer, by 

 varying the current through the coils. It could be held at a desired tem- 

 perature within the negligible variations of 0.02 at 18 and 0.1 at 156. 



For the temperatures of 218, 260, and 306 vapor baths of boiling 

 naphthalene, isoamyl benzoate, and benzophenone, respectively, were used. 



THERMOMETERS. 



Up to and including 100, mercury thermometers graduated in tenths 

 of a degree were used. Since stem exposure could not always be avoided, 

 they were calibrated in position as used by comparison with a standard 

 Baudin thermometer, having a Bureau of Standards' certificate. The ice 

 and steam readings remained substantially constant throughout the work. 

 The error in the bath temperature could hardly have exceeded 0.02 at 

 the 18, 25, 50, and 100 points, but at 75, owing to the necessity for 

 applying a large stem-exposure correction to the standard Baudin, the 

 error may have been as much as 0.05. The temperatures above 100 

 were probably determined with an accuracy of 0.2 - 0.3. At these tem- 

 peratures a 360 Alvergniat thermometer, graduated in degrees, was used. 

 The ice, steam, naphthalene, and benzophenone points were directly deter- 

 mined. Intermediate corrections were computed for 128 and 156 from 

 the bore calibration, allowing for deviations of the mercury from the 



