594 PROCEEDINGS OP THE AMERICAN ACADEMY. 



tained but not yet published indicate that the same series of hydrocar- 

 bons constitute the less volatile portions of Canadian petroleum, and 

 probably also of Ohio petroleum. Results already published show that 

 the less volatile distillates from California and Texas petroleum are 

 composed of the same series and other series still poorer in hydrogen. 



The unexpected appearance of the series C n H 2n _ 2 in Pennsylvania 

 petroleum suggests a closer relationship between this petroleum and the 

 heavier oils from other fields, such as those in Texas and California, 

 than was suspected. To gain further information concerning the heav- 

 ier portions of Pennsylvania oil, we allowed three kilos of the sample 

 from which the hydrocarbons described in this paper were prepared, to 

 evaporate spontaneously in the air in a strong draught, but with no appli- 

 cation of heat. At the end of thirty days there remained one kilo that 

 gave as its specific gravity 0.8620, practically the same as that of C 28 H 54 , 

 0.8G94. A combustion gave percentages of carbon and hydrogen re- 

 quired for the series C n H 2n _ 2 . A distillation showed that 65 per cent of 

 the residual oil was composed of hydrocarbons above C 15 H 30 . The com- 

 position of these hydrocarbons will be ascertained iu connection with the 

 study of natural and commercial paraffine, which is now in progress. 



This interesting relation, and others between the numerous petroleums 

 from different oil fields that have been examined in this laboratory will 

 be presented more at length in a later paper. 



As a general summary of the results described in this paper, hydrocar- 

 bons have been identified as shown in table on opposite page. 



It appears that the liquid hydrocarbon C23H54 iu fraction 310°-312° 

 has the same number of carbon atoms as the solid hydrocarbon octocosane 

 in the same fraction. In this respect the hydrocarbons in this fraction 

 differ from those in the two preceding fractions, in each of which the 

 liquid hydrocarbon is one carbon higher than the solid constituent. De- 

 terminations of the molecular weights of hydrocarbons with high carbon 

 content can only be made by the boiling-point method; and even with 

 the greatest care in manipulation, this method is somewhat uncertain for 

 high molecular weights of solid hydrocarbons, for the reason that the 

 rise in boiling points diminishes with the increase in molecular weight. 

 With oils there is less difficulty. For instance, the molecular weight, 

 370.2, of the liquid hydrocarbon C 2 7ll 5 2» given on page 592, is one of 

 five closely concordant determinations by different persons. As an illus- 

 tration of the care necessary in details, heating with a lamp supplied with 

 gas from the laboratory mains is so irregular on account of variation in 

 gas pressure that the gas must be supplied from a tank under water 



