﻿314 Mr. Baly and Dr. Ramsay on the Relations of Pressure, 



after ascent and descent of the mercury sometimes differed 

 by as much as four millimetres. 



It is also to be noted that the mercury must be raised very 

 slowly, or gas will be trapped between the mercury and the 

 glass. We also found it necessary to make the gauges of 

 English lead-glass, as we found this to be the only glass that 

 would stand the high temperature jackets without cracking. 



Experiments to test the Accuracy of the 

 McLeod Gauge. 

 In order to determine once and for all the value of the 

 McLeod gauge as a means of measuring vacua, some experi- 

 ments were made with air. Many sets of readings were 

 obtained, and they show the absolute failure of the gauge to 

 measure vacua. It fails more and more as the vacuum 

 improves. If the gauge had given trustworthy readings, the 

 values of pv obtained at all reading-points should have been 

 constant. We give here a set of values of pv measured at a 

 moderate vacuum (p here refers to pressure in the gauge ; 

 not to the vacuum in the apparatus) : — 



P- 

 millim. pv. 



4-1 100 



4-1 81-83 



4-1 69-921 



4-3 67-25 



4-6 59-57 



4-6 41-78 



4-9 33-47 



5-0 23-09 



6-3 15-09 



8-0 9-427 



The other sets of values of pv obtained during this exhaus- 

 tion varied from 100-95, 100-74, 100-65, 100-38, 100-18, 

 100-14. In each of these sets the first and highest pv is 

 raised to 100, and all other values in that set raised in pro- 

 portion. 



The set of values of pv given above, decreasing from 100 to 

 9 was plotted against pressures and the curve (No. I.) drawn 

 which gives the value of pv constant for minimum pressure 

 somewhere about 270. This gives a vacuum of 9 millionths 

 of an atmosphere, instead of *3 millionth. 



But the curve, as thus drawn, is not correct, for the 

 decrease in the values of pv is due to two causes. First, to 

 the condensation of the gas on the glass as the pressure rises ; 

 and secondly, to the fact that Boyle's law does not hold good 



