1902.] 



On the Measurement of Temperature. 



487 



of air or oxygen is passed into the liquid it ceases to boil, and may 

 become superheated to the extent of more than one degree. Further, 

 it is very difficult to obtain a sufficient quantity of pure oxygen for such 

 a measurement. 



In our experiments a bulb in which a small quantity of pure oxygen 

 could be liquefied was immersed together with the thermometer bulb 

 in a vacuum-vessel containing liquid air or oxygen, through which a 

 fairly rapid current of air was passed. The bulb containing the pure 

 oxygen communicated with the lower chamber of a barometer, so that 

 the measurements of the vapour pressure were quite independent of 

 the atmospheric pressure, with a mercury pump, and with an apparatus 

 for generating pure oxygen from potassium permanganate. Simul- 

 taneous readings of this barometer and of the thermometer, which 

 contained hydrogen or helium, gave the vapour pressures of pure 

 oxygen at temperatures which could be varied between 80° and 90° 

 Abs., according as the vacuum vessel surrounding the thermometer 

 bulb, &c, contained freshly made liquid air or nearly pure oxygen. 



Four thermometers were employed in these measurements, the 

 capacities of the bulbs being approximately 90 c.c, 12 c.c, 26 c.c, 

 and 27 c.c. Only one single set of measurements on the hydrogen 

 scale were made with the large thermometer, which was the same 

 instrument as was employed in the determination of the pressure 

 coefficients of the gases, for though it was possible to observe a steady 

 temperature by means of it with a degree of accuracy approaching 

 1 part in 20,000, it was found impossible to maintain so large a 

 thermometer bulb at a constant and uniform temperature, without 

 employing very large masses of liquid oxygen. The results obtained 

 by means of this thermometer differ only by 0*1 ° from the mean of 

 those obtained by means of the three smaller thermometers. The 

 temperatures observed by means of the three smaller thermometers 

 rarely differ by more than 0*Q3° from the temperature, corresponding 

 to the same pressures, taken from the smoothed vapour pressure curve. 



The form of the thermometers employed in this research, and in 

 the measurements of the vapour pressures of liquid hydrogen, was 

 practically the same as that described in the previous abstract ; full 

 details are given in the paper. The pressure on the gas in the 

 thermometer was directly observed by means of a manometer attached 

 to the apparatus, and was independent of the atmospheric pressure. 

 The temperature of the dead space and mercury column were determined 

 by means of mercury thermometers, and the mean temperature of the 

 vertical portion of the stem immediately above the bulb was measured 

 by means of an auxiliary gas thermometer, with a cylindrical bulb of 

 the same length as that portion of the stem of which the temperature 

 is uncertain. 



The coefficient of expansion of the glass of which the thermometer 

 VOL. LXX. 2 L 



