Section 77. Instrumental Errors and Corrections. 19? 



ble leads and the main leads. All of these surfaces were polished with 

 fine sandpaper before each heating. Special experiments showed that the 

 brass contact surfaces are almost unaffected by tarnishing, but that the 

 steel surface resistances are increased appreciably by a film of oxide ; also 

 that the variation in resistance at the sliding contact of the carriage on its 

 supports is inappreciable. 



The current used in the measurements was made as small as possible, 

 so as to avoid the ejection from the electrode of adsorbed material. This 

 was accomplished by using the smallest possible voltage on the coil, after 

 weakening its spring by filing partly through it. Any error from polari- 

 zation caused by asymmetry of the coil, was eliminated by commutating 

 the current and taking the mean of the two readings. 



The excess of pressure due to air in the bomb was only a small fraction 

 of the total pressure ; for instance, if the air is evacuated before the heat- 

 ing down to a pressure of 2 cm. of mercury, and the vapor-space at 218 

 is 2 c.cm., then the air pressure at 218 is 0.5 atmosphere, while the vapor- 

 pressure is about 22 atmospheres. With 2 c.cm. vapor-space at 306, the 

 air pressure is about 1 atmosphere, while the vapor-pressure is about 97 

 atmospheres.* Hence the variation in conductivity due to the residual 

 air pressure is probably negligible. 



Down to the lowest level ordinarily used, namely with the bomb three- 

 quarters full, the height of the solution in the bomb has no effect on the 

 conductance-capacity. In the experiments for determining the vaporiza- 

 tion-correction at 306, however, the bomb was only half full at 18; the 

 effect on the conductance-capacity was determined by filling the bomb only 

 to this level with a standard potassium chloride solution. The results are 

 given in section 81. The correction for the variation in conductance- 

 capacity with the temperature was made as described in section 36, 

 Part IV. 



78. PREPARATION OF THE SUBSTANCES AND SOLUTIONS. 



The weights used in weighing out the solid substances and solutions 

 were all standardized in terms of the one-gram weight as standard. All 

 weights were reduced to weights in a vacuum before being used in cal- 

 culations. The atomic weights used were those reported by the Interna- 

 tional Committee in 1904, referred to oxygen as 16.00. 



All solutions both strong and dilute, except those used for determining 

 the conductance-capacities of the apparatus, were made up, analyzed, or 

 titrated wholly by weight; the results are therefore independent of tem- 

 perature, and are expressed in terms of milli-equivalents per kilogram of 



*Batelli, Landolt-Bornstein-Meyerhoffer Tabellen, p. 122 (1905). 



