ON ARGON AND ITS COMPANIONS. 
53 
163‘170 cub. centims., 32MSS cub. ceiitiius., and 7'152 cub. ceiitims. ; the largest was 
used only to determine tlie density of pure argon. Sealed glass globes of the same 
external volume as the weighing globes were used as counterpoises. 
The Weights .—The values of the weights were carefully referred to that of a 
centigramme rider. It was not necessary to know their absolute values, for since tlie 
volumes of the globes were obtained by weighing them filled with water at a known 
temperature with the same weights as were used in weighing the gases, the weiglits 
of the gases were referred through the weight of an equal volume of water to tlie 
weight of ail equal volume of oxygen. The weight of a litre of oxygen was taken 
as 1’4296 gram. (‘ Phil. Trans.,’ A, 1895, Part 1, p. 201)"^'. 
The Balance .—The globes were weighed in a long-lieam Oertling balance to O'Ol 
milligram. They were usually allowed to hang for aljout half-an-hour in the balance- 
case, before the final weighing was made. 
The PrejKiration of the G'a.scs.—The gases were sparked witli oxygen in jiresence 
of caustic soda to remove nitrogen, &c. ; the oxygen was subse(|uently removed by 
means of phosphorus, tlie water-vapour being at the same time absorbed by tlie pro¬ 
ducts of its combustion. As tliis jirocess differs somewhat from any previously descrilied, 
it may be advisable to enter into it in some detail. After the gas had been sjiarked, 
it was withdrawn into a burette, and sepaiuted from tbe soda. A tube was then 
completely filled with mercury, and a piece of plios})horus, as large as a small pea. 
was introduced, so that it floated iq) aliove the mercury. The top of the tube was 
then warmed, so as to melt the })lios})horus, and the gas ivas introduced from the 
gas-burette, through the mercury, bubble by bubble; as each liublde entered it 
caused the phosphorus to infianie. Wlien all the gas liad lieen introduced, the top 
of the tube was again warmed, and a rain of phosj)borus fell through the tulie. 
After cooling, the gas could be withdrawn if desired ; liut it was often ke})t in the 
same tulie until recpiired. It aj)j)ears to be necessary to have a sufficiency of oxygen 
in the gas which is admitted to the phosphorus ; for if the plios})l)orus does not 
inflame, we have found that the spectium shows lines belonging to jihospliorus ; if, 
however, the phosphorus actually catches fire, there is no sign of any volatile 
compound being present in the gas. 
The apparatus enq)loyed for introducing tlie gas into the globe is shown in fig. 3. 
The gas can be introduced through the siphon into the Imrette A, which can be placed 
in communication with the globe E. The side-tulie B, whicli is of tlie same diameter 
as the upper part of the burette, facilitates tbe measurement of the pressure on the 
gas, after it has been allowed to fill tlie globe, thus eliminating any correction for 
capillarity. The pressure was read off' on a silvered glass scale, to winch the burette 
was fixed. The tulie c, which was about 2‘5 milllms. in diameter, contained a roll of 
* In previous determinations with the same bulbs different weights were used, and the capacity is given 
as a slightly different mimlier. As these were also referred to water, no correction as regards the 
densities previously determined is necessary. 
