286 THE KINETIC THEORY OF GASES. 



was accordingly treated with water, and the resulting ammonia ab- 

 sorbed by means of weak sulphuric acid. There was merely a trace of 

 gas which refused to be absorbed, and on examination it turned out to 

 be the familar hydrogen, which was formed by the action of the water 

 on some metallic magnesium which had escaped combination with 

 nitrogen. This experiment was interesting, inasmuch as it proved 

 that magnesium refuses to combine with even the smallest trace of 

 argon. The ammonia resulting from this treatment, it is true, might 

 have conceivably contained a compound of the new gas, but a similar 

 sample had previously been decomposed, so as to obtain from it its 

 nitrogen, and that sample of nitrogen had been found by Lord Bay- 

 leigh to possess the same density as a sample of nitrogen of which the 

 source could not be traced to the atmosphere. Lastly, it was conceiv- 

 able that the hydroxide of magnesium might have contained some 

 compound of the new element. It was therefore treated with water, 

 and the soluble portion separated from the insoluble. The soluble 

 portion, on examination, proved to contain nothing but the carbonate 

 of magnesium. The insoluble portion was not further dealt with, but 

 was kept in reserve. 



The argon of the atmosphere was next examined. A large quantity 

 having been prepared, it was purified, and by passing it into a vessel 

 immersed in liquid air, made to boil at even lower temperature than 

 usual by pumping away the air-gases as they boiled off, the argon, too, 

 was completely changed into liquid. Liquid argon is clear and color- 

 less, whereas liquid air has a faint blue tint, owing to the blue color 

 of the oxygen it contains. The argon was next made to boil, by allow- 

 ing the temperature of the liquid air to rise a few degrees, and the 

 first portions of argon-gas were collected separately, the remainder 

 going back into the gas-holder in which it had originally been stored. 

 The gas thus obtained was lighter than argon and more difficult to 

 liquefy; this was shown by the necessity of compressing it into the 

 bulb in which liquefaction took place. The most volatile portions of 

 this liquid were next collected separately, and the gas proved to be 

 still less dense than the former sample. It was not possible to liquefy 

 more than a small fraction of this last specimen of gas, to however low 

 a point the temperature of the boiling air was reduced; and after 

 another repetition of the same process the gas appeared to be as light 

 as the process could make it. Its density was 9.75 times that of hydro- 

 gen, and making allowance for a small quantity of argon, which it 

 must necessarily have contained, this number becomes reduced to 9.6. 1 

 The weight of a molecule, compared with the weight of an atom of 

 hydrogen, as previously explained, must therefore be 19.2; and 19.2 

 lies between the atomic weights of fluorine, 19, and of sodium, 23, fall- 



1 This gas has since been found to contain a trace of helium, the presence of which 

 would lower the above density. The actual density will, therefore, he somewhat 

 higher than 9.6, but it will probably not exceed 10. It has not yet been determined. 



