CARBON AND THE HYDROCARBONS 877 



mountain chains, 57 because water with iron carbide ought to give iron 

 oxide and hydrocarbons. 58 Direct experiment proves that the so-called 

 spiegeleisen (manganiferous iron, rich in chemically combined carbon) 

 when treated with acids gives liquid hydrocarbons 59 which in com- 



s7 During the upheaval of mountain ranges crevasses would be formed at the peaks 

 with openings upwards, and at the foot of the mountains with openings downwards. 

 These cracks in course of time fill up, but the younger the mountains the fresher the 

 cracks (the Alleghany mountains are, without doubt, more ancient than the Caucasian, 

 which were formed during the tertiary epoch) ; through them water must gain access 

 deep into the recesses of -the earth to an extent that could not occur on the level (on 

 plains). The situation of naphtha at the foot of mountain chains is the principal 

 argument in my hypothesis. 



Another fundamental reason is the consideration of the mean density of the earth. 

 Cavendish, Airy, Cornu, Boys, and many others who have investigated the subject by 

 various methods, found that, taking water = 1, the mean density of the earth is nearly 

 6'5. As at the surface water and all rocks (sand, clay, limestone, granite, &c.) have a 

 density less than 8, it is evident (as solid substances are but slightly compressible even 

 under the greatest pressure) that inside the earth there are substances of a greater density 

 indeed, not less than 7 or 8. What conclusion, then, can be arrived at ? Anything 

 heavy contained in the bosom of the earth must be distributed not only on its surface, 

 but throughout the whole solar system, for everything tends to show that the sun and 

 planets are formed from the same material, and according to the hypothesis of Laplace 

 and Kant it is most probable, and indeed must necessarily be held, that the earth and 

 planets are but fragments of the solar atmosphere, which have had time to cool con- 

 siderably and become masses semi-liquid inside and solid outside, forming both planets 

 and satellites. The sun amongst other heavy elements contains a great deal of iron, 

 as shown by spectrum analysis. There is also much of it in an oxidised condition on 

 the surface of the earth. Meteoric stones, carried as fragmentary planets in the solar 

 system and sometimes falling upon the earth, consisting of siliceous rocks similar to ter- 

 restrial ones, often contain either dense masses of iron (for example, the Pallosovo iron 

 preserved in the St. Petersburg Academy of Sciences) or granular musses (for instance, 

 the Okhansk meteorite of 1886). It is therefore possible that the interior of the earth 

 contains much iron in a metallic state. This might be anticipated from the hypothesis 

 of Laplace, for the iron must have been compressed into a liquid at that period when 

 the other component parts of the earth were still strengly heated, and oxides of iron 

 could not then have been formed. The iron was covered with slags (mixtures of silicates 

 like glass fused with rocky matter) which did not allow it to burn at the expense of the 

 oxygen of the atmosphere or of water, just at that time when the temperature of the 

 earth was very high. Carbon was in the same state; its oxides were also capable of 

 dissociation (Deville) ; it is also but slightly volatile, and has an affinity for iron, and iron 

 carbide is found in meteoric stones (as well as carbon and even the diamond). Thus the 

 supposition of the existence of iron carbides in the interior of the earth was derived by 

 me from many indications, which are to some extent confirmed by the fact that granular 

 pieces of iron have been found in some basalts (ancient lava) as well as in meteoric stones. 

 The occurrence of iron in contact with carbon during the formation of the earth is all 

 the more probable because those elements predominate in nature which have small 

 atomic weights, and among them the most widely diffused, the most difficultly fusible, 

 and therefore the most easily condensed (Chapter XV.) are carbon and iron. They 

 passed into the liquid state when all compounds were at a temperature of dissociation. 



68 The following is the typical equation for this formation : 



8Fe m C,, + 4 m HoO=wFe 5 O 4 (magnetic oxide) + C^Hgrn (see Chapter XVII., Note 38). 



59 Cloez investigated the hydrocarbons formed when cast-iron is dissolved in. hydro- 

 chloric acid, and found C,,H 2 ,! and others. I treated crystalline manganiferous cast-iron 

 with the same acid, and obtained a liquid mixture of hydrocarbons exactly similar to 

 natural naphtha in taste, smell, and reaction. 



