1876.] -^^^ [Sadtler. 



members of the marsli-gas series of hydrocarbons. Thus the gas from 

 Pioneer Run he found to have essentially the composition of propyl hydride 

 (C., Hg), with small quantities of carbonic acid and of nitrogen ; the Fre- 

 donia gas appeared to be a mixture of marsh-gas (CH4) and ethyl hydride 

 (Cj Hg), with a small quantity of carbonic acid and 1.55 per cent, of nitro- 

 gen ; the Rogers' Gulch gas was CH^ almost exclusively, with 15.86 per 

 cent, of carbonic acid, and a small quantity of nitrogen ; the Burning 

 Springs gas almost pure CH^ with a little CO^ ; the Petrolia gas a mixture 

 of marsh-gas (CH4) and ethyl hydride (Cj Hg), with a small amount of 

 carbonic acid. 



However, in some cases the composition as given above was only appa- 

 rent, as in the case of the Pioneer Run gas, for on passing the gas through 

 alcohol a part was absorbed, which was afterwards shown to be butyl 

 hydride (C^ H,o), while the part unabsorbed showed nearly the composition 

 of marsh gas (CHJ. It was evident, therefore, that what appeared to be 

 propyl hydride (C3 Hg) was in reality a mixture of marsh-gas (CH J and 

 butyl hydride (C^ Hio). 



In 1870, Prof. Henry Wurtz made an analysis of the gas from a well in 

 West Bloomfield, N. Y. The results of t]iis analysis are found in Silli- 

 man's Journal (2) XLIX., p. 336. He found : 



Marsh-gas 82.41 



Carbonic acid 10. 11 



Nitrogen 4. 31 



Oxygen 0. 23 



Illuminating hydrocarbons 2.94 



100.00 



The specific gravitj^ of the gas was .693. 



The methods of analysis were not the usual ones of gas analysis, but 

 some new absorption methods devised by himself in conjunction with 

 Prof. B. Silliman. This, I believe, is the extent of the published informa- 

 tion on the subject. 



The gases which I collected and analyzed were : 



First, the gas of the Burns Well in Butler County; secondly, that of the 

 Harvey Well in the same county; thirdly, that from the Leechburg Well 

 aci'oss the Kiskeminitis river from Leechburg, in Westmoreland County 

 and fourthly, the gas bubbling from a spring at Cherrj^ Tree in Indiana 

 County. As the results of Fonqu6 just given had rendered the presence of 

 higher members of the marsh-gas series than CH^ probable, at two of 

 these wells, I made, in addition, absorption tests with absolute alcohol and 

 with bromine, hoping to prove qualitatively the presence or absence of 

 these higher hydrocarbons as well as those of the ethylene series. For 

 while marsh gas itself is insoluble in absolute alcohol, ethyl hydride 

 (Cj Hg) is dissolved to the extent of 1^ volumes, propyl hydride 

 (C3 Hg) to the extent of 6 volumes, and butyl hydride (C4 Hjo) 

 to the extent of 18 volumes in one volume of the alcohol. A pas- 



