528 Dr. F. A. Lindemann and Dr. F. W. Aston on the 



mercury in gas-holder A and opening the connecting stop- 

 cock between it and a. After another interval the stopcock 

 was turned, the mercury raised in A and the gas forced 

 into bulb b. The mercury was next lowered in both A and 

 B, the former receiving the second fraction from a while the 

 latter withdrew the first fraction of the gas now in b. The 

 fundamental assumption on which this arrangement was 

 made was that at this stage, if the vapour-pressures of the 

 gases are nearly the same, the gag in A. would have the same 

 composition as that left in b, and that they therefore might 

 be mixed. This was done by raising the mercury, which not 

 onlv drove the gas from A into b but also the lightest 

 fraction from B into c, where it again fractionated, each 

 process driving the lower boiling gas forward and keeping 

 the higher back. 



The apparatus may contain any number of units, the 

 whole system being made cyclical and continuous by joining 

 the charcoal bulb at one end with the gas-holder at the 

 other. Four such units were actually employed, and after 

 four operations the liquid air was removed from a and the 

 residue it contained was pumped off completely with an 

 Antropoff pump as the first contribution to the heaviest 

 fraction, in the same way that in D was also pumped off as 

 that of the lightest. The bulb a was then immersed again 

 in liquid air and the process continued. 



After about two-thirds of the gas had been collected in 

 this way as light and heavy fractions, that remaining was 

 all pumped out as the middle fraction. The process was 

 next repeated with the light and heavy fractions in turn, the 

 intermediate ones being combined by a definite rule. 



By this arrangement, which does many operations at once,, 

 the small quantity of helium contained in the original gas 

 was removed in a remarkably short time, after which the 

 neon was subjected to continual fractionation for three 

 weeks. The gas had now been through about 3000 fraction- 

 ations and was divided into seven main fractions ; the 

 density of these was determined by the quartz micro-balance 

 (F. W. Aston, Roy. Soc. Proc. A, vol. lxxxix. 1914), the 

 figures for the pressures giving the same zero as oxygen at 

 76*35 were as follows : — 



(1) (2) (3) (4) (5) (6) (7) 



121-05 120-95 121-05 120-90 121-00 121-05 121-05 



; The mean of these, 121*05, gives a molecular weight of 

 20*19, which is identical within experimental error with the 

 accepted one of 20*200 determined by Watson (J. C. S. 1910). 



