BRTDGMAN. — THERMODYNAMIC PROPERTIES OF LIQUIDS. 21 



used here is determined. It is unfortunate that these values did not 

 appear in time for use in the present computations. They are quoted, 

 however, so as to afford a comparison with the results of Amagat, 

 which are here used as the standard for low pressures. It makes 

 no difference with the essential conclusions of the paper, however, 

 which set of data is accepted as most probably correct. The only 

 effect would be to change the initial values of some of the thermody- 

 namic properties; their magnitudes at high pressures will not be 

 altered. 



The precise steps in combining into a final result the changes of vol- 

 ume computed from the two sets of readings at high and low pressures 

 were as follows. From the high pressure readings, the changes of 

 volume at 40° (AV, cm.^/cm.^) were plotted against the displacement 

 of the slider of the resistance bridge, the zero of pressure being at 

 10 cm. displacement. The scale of the diagram was large; 2 cm. for 

 0.1 inch piston displacement, and 55 cm. slider displacement for the 

 maximum pressure. From the low pressure readings AV in cm.^/cm.* 

 was found at 20'^, and was plotted against slider displacement, the 

 zero being at 5 cm. A smooth curve was drawn through these points. 

 From this curve, knowing the constants of the manganin coil, AV 

 was found at 500, 1000, 15^00, and 2000 kgm. and 500, 1000, 1500, and 

 2000 atmos. AV in terms of atmos. was now corrected so as to be 

 reckoned from 1000 atmos. as zero. The values of Amagat for AV at 

 20° were next computed with 1000 atmos. as the zero. These two 

 sets of data were compared, and the new values for the lowest interval 

 adjusted so as to be in agreement with Amagat between 1 and 50Q 

 kgm. It will be noticed that this preliminary comparison with 

 Amagat does not enter the final result; it was an orienting comparison 

 for obtaining some idea of the accuracy at low pressures. Using 

 Amagat 's value for the lower interval, the changes of volume at 

 pressures corresponding to 2.5, 5.0, 7.5, and 10.0 cm. were next deter- 

 mined, reckoned from zero pressure. These values were now corrected 

 to 40° with the data obtained for the thermal dilatation and were 

 recomputed with 10 cm. as the zero of pressure. The changes of 

 volume at 40° obtained from both the low and high pressure sets of 

 measurements were now plotted together on one diagram, and a smooth 

 curve drawn through the points. From this curve the changes of 

 volume with the kilogram as the unit of pressure were read off at 500 

 kgm. intervals, starting from an origin at 500 kgm. The changes 

 of volume so found Avere now smoothed to give regular differences. 

 The smoothing was performed in a manner somewhat different from 



