BRIDGMAN. — THERMODYNAMIC PROPERTIES OF LIQUIDS. 29 



for the dilatation. In order to save space, these tables are not given 

 here. The thermal dilatation enters into many of the other ther- 

 modynamic quantities listed in this paper. In computing these 

 quantities the values of the dilatation given in the tables have been 

 used, not the values obtained from the curves given later. The 

 same is true for all the other thermodynamic properties listed in the 

 paper; tables were first computed for all of them before curves had been 

 drawn for any. In this way any progressive error was a^•oided which 

 might have been introduced by the use of diagrams. Although each 

 diagram shows the property in question with as great an accuracy 

 as is justified experimentally, it might be that if a computation in- 

 volved the transference of points from one diagram to another several 

 times, the error so introduced might finally mount up to more than the 

 experimental error. 



Isothermal Compressibility. — The compressibility, or the 



fdv\ . rr^ ■ 



quantity { ^ j , was the next to be determmed. This was found by 



a method somewhat analogous to that for the dilatation. Evidently 

 the compressibility does not vary greatly with temperature. If the 

 compressibility can be found as a function of pressure for one constant 

 temperature, then the compressibility at other temperatures can be 

 found by applying a small correction. The temperature chosen for 

 the direct determination of the compressibility w^as 40°, since this 

 was the temperature at which the change of volume with pressure 

 had been found. The compressibility was determined graphically 

 from a large scale drawing of the change of volume against pressure. 

 An alternative method would have been to calculate mathematically 

 the slope from the approximate formula for the change of ^•olume, 

 and then to correct this by the graphically determined slope of the 

 difference curve. But this method would fail at the lowest pressure, 

 500 kgm., and at the higher pressures it did not pro^'e necessary, be- 

 cause the simpler direct graphical method was sufficiently accurate. 



It was now possible to correct the compressibility at 40° to the 

 other temperatures of the tables by the use of difference curves. Let 

 us suppose that it was desired to find the compressibility at 60°. 

 Pigure 6 represents graphically the operation which was actually 

 performed by a computation. The curve of volume at 60° against 

 pressure was displaced downwards (shown in the dotted line) so as 

 to have the same origin as the curve for 40°. The difference between 

 the curves was plotted on a large scale against pressure, and the 

 graphically determined slope of the difference curve used as a cor- 



