BRIDGMAN. — THERMODYNAMIC PROPERTIES OF LIQUIDS. 13 



'The pressure was then increased at 80° to the next higher vahie, about 

 3500 kgm., and the corresponding set of measurements made at 20° 

 intervals, this time with decreasing temperature. The other four 

 sets of readings were made in a similar way. The highest pressure 

 reached, on the last reading at 80°, was about 12500 kgm. From here 

 the temperature was decreased, so that the final reading at 20° was 

 made at a pressure of about 11500 kgm. Here is an incidental ad- 

 vantage of the modified method of making the expansion measure- 

 ments, because mercury solidifies at 20° at less than 12000 kgm., 

 so that if the former method had been used a smaller value of the 

 mean maximum pressure would have been necessary for the last read- 

 ing at 20°. Of course the modified method necessitates a slight 

 change in the method of making the computations, which will be 

 described later. 



All the measurements on each of the liquids were repeated to ensure 

 greater accuracy. The measurements were made in two series; all 

 twelve liquids were first measured, and, except for the preliminary 

 experiments, no repetitions made until at least one set of readings 

 had been made for each liquid. This has the advantage of separating 

 the measurements on any one liquid by a considerable interval of 

 time, and thereby eliminating any possible effect of temporary varia- 

 tions in the apparatus. During the first extended series of readings 

 the two sets of readings for each liquid at high and low pressures were 

 made with two separate fillings of the apparatus. At the lower 

 pressures the viscosity effects do not play any part, so that it was not 

 necessary to leave so large a space between the bulb and the walls of 

 the cylinder as was necessary for the higher pressures. In this way 

 a larger bulb could be used at the lower than at the higher pressures, 

 so that somewhat greater sensitiveness could be obtained at the lower 

 pressures because of the greater quantity of liquid. But this made 

 it necessary to take the apparatus apart after the set of readings at 

 low pressures, and replace the larger bulb by the smaller one for the 

 run at high pressures. This made considerable trouble, which did 

 not seem adequately compensated by the somewhat greater accuracy, 

 so that in the repetition of the experiment the readings at the high 

 and low pressures were all made with the same filling of the apparatus. 

 The readings at low pressures were made before the readings at high 

 pressures, and so before the piston had been upset. 



The final detailed procedure was as follows. The readings at low 

 pressures were made first. It is not necessary to describe them in 

 detail as they were similar to the readings at the higher pressures, the 



