BRIDGMAN. — THERMODYNAMIC PROPERTIES OF LIQUIDS. 15 



measurements are due to the assistance of Mr. S. L. Gokhale. During 

 the entire time there was no change in the internal diameter of the 

 cyHnder of so much as 0.0001 inch. To ensure further that there was 

 no progressive change in the cyhnder during the measurements, the 

 comparison measurements with the Hquid replaced by Bessemer steel 

 were made at five intervals during this time. The maximum diver- 

 gence of any of these readings from the mean was only 0.5%, better 

 perhaps than would at first be expected from the method. 



The original data are so numerous that it seemed undesirable to 

 give them here in full. Every point recorded involves six readings, 

 two of pressure and four of piston displacement. On the average 

 each liquid involved 140 points, 75 for compressibility and 65 for 

 thermal expansion. This makes a grand total of 12500 readings in 

 the original data. It was thought to be sufficient to give a few 

 sample curves of complete data, and to state for each liquid the 

 average departure from the mean of the two series of compressibility 

 and dilation measurements. The average departure from the mean 

 of the two series of compressibility measurements for the twelve 

 liquids was 0.15% of the maximum, and the departure from the mean 

 of the thermal expansion measurements was 2%. The changes of 

 volume due to changes of pressure are much greater than those due 

 to temperature, so that the compressibility measurements determine 

 the final accuracy of the volume. 



In regard to the purity of the liquids, it was not thought necessary 

 to take special precautions, because the properties studied here are not 

 much influenced by the presence of impurities. The compressibility, 

 for example, of a mixture of two liquids of small concentrations of the 

 one in the other is an additive function of the compressibility of the 

 two components. An example of this fact occurred in the prelimi- 

 nary work on ether. The first measurements were made with the 

 ether enclosed in a glass bulb. This bulb broke on the application of 

 pressure because of the great viscosity of the kerosene. The com- 

 pressibility measurements of this preliminary run were measurements, 

 therefore, on a mixture of ether and kerosene. In spite of the fact 

 that this ether was very much contaminated the result showed that 

 the compressibility of this mixture was only different by 4% from what 

 it was when the measurements were made on the same quantities 

 of ether and kerosene prevented from mixing. 



The liquids used were obtained from Eimer and Amend. They 

 were either the purest manufactured by them, or else Kahlbaum's 

 purest. It was to be expected therefore, that only slight impurities 



