118 GENESIS OF NATURAL GAS AND PETROLEUM. [Feb. 5, 



in form of a carbide richer in carbon than is producible in the 

 electric furnace, and therefore resembling the carbides of aluminum 

 as regards its action upon water. In such case a gas somewhat 

 similar to natural gas might result. Published analysis of meteoric 

 iron and of iron found in plutonic rocks do not tend to show, how- 

 ever, that the iron found in nature ever contains carbon in such 

 quantity as to lead to the belief that a definite carbide of this metal 

 exists comparable to the carbides of aluminum, akaline earth, and 

 alkali metals. 



If aluminum carbide and the carbides of related metals are to be 

 regarded as the source of natural gas, we must look for the occur- 

 rence of the lighter metals at depths at which the hypothesis of 

 Mendeleeff would require us to suppose that the heavy metals pre- 

 dominate. It seems, therefore, probable that a few only of the 

 metals in form of carbides could have been concerned in the pro- 

 duction of natural gas, and these are the very metals which on 

 account of their lightness are supposed by this hypothesis to give 

 place to those of high specific gravity in regions where the chemical 

 changes in question have occurred. 



On account of its stronger affinity for oxygen, aluminum may be 

 supposed under all conditions tending towards oxidation to assume 

 the form of an oxide more readily than iron, and where aluminum 

 occurs in presence of the heavier metals it will probably precede 

 these in the order of time in uniting with oxygen. But its oxidation 

 would remove it from the sphere of action leading to the produc- 

 tion of hydrocarbons. 



The conclusion seems justified that where aluminum occurs in a 

 metallic state, or as a carbide, below or in the earth's crust, the 

 heavy metals will also abound and notably iron. 



If the chemical composition of natural gas is such as to warrant 

 the belief that its production was due to the action of steam upon 

 iron carbide, the hypothesis of Mendeleeff would at once receive 

 strong support. If, on the other hand, chemical considerations 

 show that iron cannot have been concerned in the process, the 

 question then arises. Why has iron carbide been suppressed in the 

 subterranean reactions giving origin to natural gas ? 



The term iron carbide has here been used to signify iron contain- 

 ing a little carbon, such as cast iron, but not implying a real com- 

 pound containing iron and carbon in atomic proportions. 



Analytical data concerning natural gas drawn from deep-lying 



