Distillation of Liquid Air. 525 



Olszewski's. On the other hand, if the points at intermediate 

 pressures be given some value, a very fair curve can be drawn 

 which very nearly gives the same boiling-point as found on 

 the Callendar thermometer as above, viz. 90°"96 abs., the 

 actual point being 90°' 90. Which curve to adopt as best 

 expressing his results I do not know, for no data are given 

 of the 25 observations whose mean is given as one point. 

 The difference between the boiling-point I have given and 

 that found by Olszewski is o, 36. 



Of the boiling-point of nitrogen, Olszewski* gives a 

 measurement in his series of five vapour-pressures, which are 

 as follows : — 



Pressure. Temperature (aba.). 



o 



35 atmospheres 127*0 



3L „ 124-8 



17 „ 116-5 



1 „ 78-6 



60 mm. (melting-point) 59*0 



The straight line of Ramsay and Young's ratios was drawn 

 for these points. The last two by no means agree, and were 

 therefore given equal weight. The vapour-pressure curve 

 was drawn, and the boiling-point of nitrogen was given as 

 77°-38. The difference, therefore, between the boiling-point 

 I have given and that obtained by smoothing Olszewski's 

 numbers is 0°*16. Olszewski's values for the boiling-points 

 for the two gases are thus both lower than the ones 1 have 

 found. 



In order to test Brown's suggestion that the ratio of the 

 partial vapour-pressures of the two components in the liquid 

 directly determines the composition of the vapour, it was 

 necessary to make measurements of the vapour-pressures of 

 oxygen and nitrogen with the temperatures referred to the 

 Callendar thermometer. Since these vapour-pressures were 

 only required between the temperatures 77° and 91° absolute, 

 two points on each curve were sufficient, and the complete 

 curve was drawn by the Ramsay and Young method. Of 

 these two points, the pressure of 760 mm. at the boiling-point 

 was one, and for the other it was necessary to determine the 

 vapour-pressure of each gas at the boiling-point of the other. 

 This was carried out first for nitrogen by immersing a small 

 bulb containing atmospheric nitrogen into pure liquid oxygen, 

 and then measuring the pressure under which it liquefied. 

 The mean of four readings was />=2805'4 mm., T = 9U°'96. 



* Co nipt. Rend. xcix. p. loo. 



