ip6 Conductivity of Aqueous Solutions. Part VII. 



to the known boiling point of naphthalene tinder the corrected atmos- 

 pheric pressure as determined by a mercurial barometer. The 306 point 

 on the Alvergniat thermometer was frequently determined by heating it 

 in a calibrating bath containing pure benzophenone, prepared by crystal- 

 lizing a Kahlbaum preparation from absolute alcohol. For the boiling- 

 points of both naphthalene and benzophenone the values on the hydrogen 

 thermometer determined by Jaquerod and Wassmer* were employed. 



77. INSTRUMENTAL ERRORS AND THEIR CORRECTIONS. 



There was no appreciable inaccuracy in the temperature measurement 

 at 18 ; and at 218 and 306 the measurement certainly gave the true 

 temperature of the bomb within 0.2. An uncertainty of 0.1 in tempera- 

 ture at 218 corresponds in the worst case to less than 0.1 per cent in the 

 conductance, as the temperature-coefficient at this point is always less than 

 1 per cent. At 306, 0.1 corresponds at the maximum to about 0.3 per 

 cent. No variation was, however, noticeable in the conductance after it 

 had reached its final value, so that the error, if any, is probably all in the 

 temperature value. 



The slide-wire was calibrated three times by the method of Strouhal 

 and Barus: once by division into ten parts, and twice by division into 

 twenty parts. The results agreed within 0.1 mm., and the correction was 

 at no point greater than 0.2 mm. The 1, 10, and 100 ohm coils of the 

 rheostat were compared, on a Carey-Foster bridge, with the Reichsanstalt 

 standards in the Electrical Department of this Institute of Technology. 

 The 1,000 and 10,000 ohm coils were tested by making up a Wheatstone 

 system, using two standards as ratio arms and a third as known resist- 

 ance, adjustment being made on the slide-wire of the Carey-Foster bridge. 

 The maximum error found was 0.15 per cent, in the 1-ohm coil. 



The measured resistance includes the resistance of the leads from the 

 bridge to the bomb. This was measured by the drop-of-potential method, 

 the bomb being placed in position as usual, except that the lower electrode 

 tag was wired tightly against a polished spot on the bomb itself. To the 

 resistance thus measured must be added, first the increase due to the heat- 

 ing of the leads inside of the bath, which was calculated from the size 

 and temperature-coefficient of the copper wire, and second, the resistance 

 of the stem of the electrode. The latter is, however, only 0.002 ohm. The 

 maximum lead resistance was 0.034 ohm, at 306, while the lowest total 

 resistance measured was 19 ohms. The only possibility of variation in the 

 lead resistance was at the removable contacts between electrode tag and 

 electrode, bomb and carriage, carriage and supports, and the outside flexi- 



*J. chim. phys., 2, 52 (1904). 



