1898.] on Liquid Air as an Analytic Agent. 819 



Ramsay, the geologist, estimated the mineral ingredients obtained 

 from this source in one year would equal a square column 9 feet in 

 diameter and 140 feet high. Roscoe detected by spectroscopic ex- 

 amination the presence of lithium, strontium and copper. The sample 

 of Bath gas examined by Rayleigh contained scarcely any oxygen and 

 but little carbonic acid. The weight in a given globe of the N from 

 the Bath gas (2 • 30522) is about half-way between that of chemical 

 nitrogen (2 '299) and "atmospheric" nitrogen (2*3101), suggesting 

 that the proportion of argon is less than in air, instead of greater, as 

 had been expected. Later experiments by Rayleigh proved that 

 this nitrogen contained helium as well as argon. 



The sample of gas from the Bath Spring was treated exactly in 

 the same way as the hydrogen mixtures before referred to. During 

 liquefaction there was a marked difference in the appearance of the 

 liquefied gases, for while the hydrogen and air mixtures on condensa- 

 tion gave clear transparent liquids, the product from the Bath gas 

 was turbid, and a precipitate gradually formed which by transmitted 

 light looked yellowish-brown. The yellowish-brown precipitate is a 

 hydro-carbon, probably of the petroleum series, having a marked 

 aromatic smell, and is liquid at the ordinary temperatures. It was 

 probably this gas which Professor Williamson gave as marsh gas in his 

 analysis. Further research will be made on this substance. Another 

 peculiarity of the liquid nitrogen obtained from Bath gas is that, on 

 examining it with a spectroscope, even through a thickness of two 

 inches of liquid, no trace of the characteristic oxygen absorption spec- 

 trum could be obtained. In all other attempts to make nitrogen for 

 liquefaction on the large scale, oxygen could always be detected in 

 the liquid by means of its absorption spectrum. Another phenomenon 

 was that the gas from the King's Well could not be entirely condensed 

 by refrigeration with liquid air boiling in vacuo. After the cooling 

 had continued for some time, the gas ceased to flow into the condens- 

 ing vessel, and the upper part of the vessel was occupied by a gas 

 that would not undergo liquefaction at the temperature together with 

 substantially liquid nitrogen saturated with the said gas. 



About 70 litres of the Bath gas were condensed, certainly the 

 largest quantity of this gas ever subjected to chemical examination. 

 This was boiled off, and a^ by accident too much nitrogen had vola- 

 tilised along with the gas, oxygen was added, and the mixture sparked 

 over alkali, to get rid of the excess of nitrogen. The samjile of gas 

 directly collected from the liquid nitrogen contained about 50 per 

 cent, of helium. During the sparking the helium lines were well 

 marked (along with others, the origin of which must be settled later), 

 and a vacuum tube filled with the product of the sparking gave a 

 splendid spectrum of the gas. The recorded unknown lines in the 

 Bath helium were subsequently detected along with helium in the 

 more volatile portion of liquid air.* Eight months after my paper 

 to the Chemical Society, and some two months after this address was 



♦ See 'Nature,' vol. Iviii. p. 570, Letter of Sir William Crookes, Oct. 11, 1898. 



