PAPERS ON PHYSICS 



293 



After allowing the system to reach a constant tempera- 

 ture for a time, the deflection of the gauge was taken and 

 the total pressure read from the calibration curves. The 

 mercury vapor was then driven into the sample container 

 and held there by liquid air, after which the residual gas 

 pressure was measured as above. The difference of the 

 two readings gave the vapor pressure of mercury at that 

 temperature. Four separate sets of readings were taken, 

 using nineteen temperatures ranging between — 7- and 

 34,9 = C. All of the values are within about 6% of a mean 

 curve, which is estimated as the accuracy that is attainable 

 by this method. (See Fig. 2.) 



5' 



20' 



25° 



30' 



35'C 



FIG 2. 



The data obtained agrees fairly well with that of Morley, 

 van der Plaatz, and with the extrapolated values from 

 higher temperatures by Ramsey and Young. The percent 

 variation in the actual readings is much less in the present 

 method. Measurements by Knudsen, in 1909, give results 

 only 1 3 to 1 2 of those of the three methods mentioned 

 above. His method, however, would be expected to give 

 values too low, and it is not probable that either of the 

 above methods would be in error as much as Knudsen's re- 

 sults indicate. A detailed study of all of the methods seems 



