122 PROCEEDINGS OF THE AMERICAN ACADEMY. 



This work was uudertakeu with the expectation that Pennsylvania oil 

 would prove to consist in its higher portions of a series C„Ho„, and it is 

 only after a great amount of laborious study, and. a vast accumulation of 

 analytical data, presenting unquestionable evidence that, for the hydro- 

 carbons distilling within these limits of temperature, except those col- 

 lecting at 196° and 216° from Canadian petroleum, this series must be 

 discarded, that I accept this conclusion. Concerning the composition of 

 Ohio and Canadian petroleum, in the beginning of this work, I had no 

 preconceived ideas. 



The separation of constituents with higher boiling points presents 

 greater difficulties than those in the portions distilling below 150°. In 

 the lower portions there is no danger of decompositions during distillation, 

 both on account of greater stability of the hydrocarbons and because the 

 unstable bodies, such as the nitrogen, oxygen, and sulphur compounds, as 

 well as the more complex hydrocarbons, distil for the most part at higher 

 temperatures. Then distillation in air has little if any effect on the lower 

 constituents, while the portions distilling at higher temperatures cannot 

 be volatilized in the presence of air without more or less decomposition. 

 As has been shown, distillation in vacuo prevents changes due to inherent 

 instability and it excludes air, but it increases very much the labor of the 

 separations. It has been found to be especially serviceable in separating 

 the hydrocarbons under consideration, since these bodies are contained 

 for the most part in portions of the first distillate boiling above the 

 limits of cracking, and are consequently contaminated by the products of 

 decomposition when distilled in air, and the decomposition products are 

 difficult to remove. Evidently any experiments on a laboratory scale 

 may fail to reveal the presence of bodies that are present in proportion- 

 ately minute quantities. It would be interesting, and from a commercial 

 point of view doubtless profitable, to establish a more extended investi- 

 gation involving the manipulation of at least a hundred barrels of crude 

 oil, continuing the fractional separations of all portions until the individual 

 hydrocarbons were as perfectly isolated as is done with smaller quantities 

 in the laboratory. Such an investigation could only be undertaken at 

 large expense, and a long time would be required to reach desirable 

 results. 



In undertaking this subject, the course to be followed was plain. 

 There is but one method for the separation of these hydrocarbons, and it 

 yields satisfactory results only after long and tedious application. Con- 

 cerning the question as to whether fractional distillation can be relied on 

 for the separation of hydrocarbons with boiling points not far removed, 



