BRIDGMAN. — THERMODYNAMIC PROPERTIES OF LIQUIDS. 19 



of the slider of the bridge wire on which the resistance of the manganin 

 coil was measured. Because of hysteresis effects it was necessary 

 to make two sets of readings, one with increasing and the other with 

 decreasing pressure. Care was taken that any two corresponding 

 readings should be at as nearly as possible the same pressure, so that 

 it should be allowable to take the average of two corresponding 

 displacements as the best value of the displacement at the average 

 of the two corresponding pressures. The next step was to draw a 

 smooth curve through the average points, and from the curve to 

 tabulate the displacement at regular intervals of pressures (5 cm. on 

 the bridge wire, or about 1100 kgm.). The second series of measure- 

 ments on the same liquid was then treated in the same way. These 

 two series of measurements differed somewhat as to the quantities of 

 materials used, that is, the hquid under investigation, the kerosene, the 

 mercury, and the steel. But the amounts w^ere so nearly the same 

 that it was permissible to take the average displacement of the two 

 series as the displacement that would have been found for the mean 

 between the quantities of material used in the two series. In only a 

 few cases did the amounts of material in the two series differ so much 

 that it was not permissible to do this. In these cases the displace- 

 ment for the mean quantity had to be determined in a way which need 

 not be described in detail. The average of the displacements obtained 

 in this way were now corrected for the effect of the kerosene, the 

 mercury, the steel, and the distortion of the steel containing vessel. 

 In applying the corrections, the mean of the five auxiliary experiments 

 in which the liquid was replaced by Bessemer steel was used. The 

 correction was applied in essentially the same way as for water. 

 The corrected result gave the motion of the piston due to the com- 

 pression of only the liquid under investigation. This, with the known 

 cross section of the piston and the weight of the liquid, determined the 

 change of volume at any pressure of 1 gm. of the liquid. Finally, by 

 using the values for the density at 0° deduced from the recent Tables 

 of Kaye and Laby, the results were reduced to the change of volume in 

 c. c. of a quantity of liquid which at 0° C. and atmospheric pressure 

 occupies 1 c. c. This is the unit quantity which is here adopted 

 throughout, and seems to have been most usually used in work of this 

 kind. In particular it is the unit quantity of Amagat. 



The computation just described applies only to the measurements at 

 the higher pressures. The results are tabulated as changes of volume 

 from 2000 kgm. as the zero of pressure. If the pressure during the 

 high pressure measurements went lower than this, as it usually did. 



