52 
PKOFESSOE ^r. EAMSAY AlsD DR M. W. TEAYEES 
under a pressure of over an additional atmosphere, in order to facilitate liquefaction. 
This device was particularly useful in the separation of neon from argon, and in the 
liquefaction of argon, mixed with the heavier gases, which could then he accom¬ 
plished without the necessity of causing the air to hoil at reduced pressure. 
By enclosing the vacuum-vessel shown in the diagram in a wider outer vessel, filled 
with liquid air, and Ijy reducing pressure in the inner vessel, the temjjerature could be 
reduced as low as 205° and maintained with hardly any labour, owing to the very 
slow ingress of heat. This device was particularly useful in the separation of neon 
from argon ; hut it was unnecessary in the separation of argon from the heavier gases. 
Determination of Density .—As the densities of the gases obtained the jmocess 
of fractional distillation gave us the surest indication of the manner in which the 
fractionation was progressing, very many determinations were made during the 
course of the investigation. Except in the case of argon itself, the quantities of 
gas at our disposal were exceedingly small; and we were consequently obliged to 
reduce the dimensions of the weighing globes to the smallest capacity compatible 
with the degree of accuracy which we desired to attain, and to design a combined 
gasholder and manometer for introducing as much of the gas as possible into the globe. 
Some four years ago, in conjunction with Dr. Noemax Collie, in a paj^er read 
before the Chemical Society, we pointed to the possibility of determining the density 
of helium to within 1 per cent, above or below its actual value by weighing the gas 
in a globe of about 30 cub. centims. capacity. The results obtained for the much 
heavier gas, argon, naturally showed much closer concordance ; and we have found 
that the accuracy of such experiments is conditioned only the sensibility of the 
balance. Since neon, the lightest of the new gases, is five times as heavy as helium, 
we resolved to employ a globe of 7 culj. centims. capacity. Determinations of the 
density of the same sample of gas, carried out by each of us independently, showed 
that it was possilfie to obtain results of sufficient accuracy and concordance, even 
when the pressure on the gas did not exceed 250 millims. of mercury; this state¬ 
ment is substantiated l^y the figures in the following talde 
The gas chosen is neon containing helium ; the capacity of the bulb is 7‘152 cub. 
centims. 
tVeight of gas. 
Pressm-e. 
Temperature. 
Density. 
tv. E. 
grammes. 
0-00531 
millims. 
788-6 
16° 20 
8-4S 
^I. tv. T. . . 
0-00.339 
507-1 
16-95 
8-44 
tv. E. 
0-00331 
422-5 
18-40 
9-94 
tL tv. T. . . 
0-00320 
417-2 
20-15 ’ 
9-80 
In these and all other determinations, the tlensity has been referred to 0 = 16. 
The Weujliing Globes .—Three globes were employed of capacities corresponding to 
