NOYES AND COOLIDGE. — ELECTRICAL CONDUCTIVITY. 189 



The lead wires are now bolted on and the bomb is placed in the 

 liquid xylene bath, serving ordinarily for the 26° measurements, and 

 the temperature of the latter is raised by means of the heating coil. The 

 li(piid level in the bomb is at tlie start about 3 mm. below the point of 

 the auxiliary electrode, so that the resistance of the upper cell is shown 

 by the conductivity apparatus to be infinite ; but upon heating, the level 

 rises and finally touches the electrode, whereupon the resistance sud- 

 denly sinks to perhaps 1000 ohms. The temperature of the bath (per- 

 haps about 130°) is now held constant until the solution in the bomb 

 has also attained it, as will be indicated by the resistance of the lower 

 and, far more sensitively, by that of the upper cell becoming constant. 

 Both these resistances are then noted, and the temperature is measured. 



The temperature is now raised by steps of three or four degrees until 

 that ratio of the cell constants is reached which corresponds to the bomb 

 being almost completely full. This limiting ratio can be determined 

 cold at any time by measuring the resistance of the lower cell and then 

 inverting the bomb and measuring that of the upper cell. Finally, the 

 cell-constant ratios are plotted as abscissae and the corresponding vol- 

 umes as ordinates, whereby a straight line is obtained. 



The computation of the volumes is made with the help of the follow- 

 ing data. Zepernick and Tammann * have found that equal volumes 

 of a 0.52 normal potassium chloride solution and of water at 0° upon 

 heating from that temperature to 140° become different from each other 

 by only 0,1 per cent. It is therefore perfectly safe to assume that the 

 expansion of the 0.02 normal potassium chloride solution used by us is 

 the same as that of pure water. From Hirn's f results the specific 

 volume of water at the temperature in question, but under a pressure 

 of 14.8 atmospheres, may be obtained. At 135°, the mean temperature 

 of the calibration experiments, the vapor pressure is 3.1 atmospheres. 

 Hirn's result should then be reduced to this pressure. The coefficient 

 of compressibility of water has been investigated by Pagliani and Vicen- 

 tini $ up to 100°. Plotting values and extrapolating gives 0.000048 for 

 the coefficient at 135°, or for the fractional decrease in volume per at- 

 mosphere pressure. Hirn's value should then be increased by 0.000048 

 X (14.8 — 3.1) X 100 = 0.056 per cent. Multiplying the value so 



* Ztschr. phys. Chem., 16, 665 (1895). 



t G. A. Hirn, Ann. chim. pliys. (4), 10, 32 (1867). His series of observations 

 covers tlie range of temperature up to 180°. Between 110° and 143° ins values 

 differ from those found by Zepernick and Tammann by only 0.02 per cent. 



t Landolt und Bornstein, Tabelien, 96 (1894). 



