PETROLEUM 715 the bulk of petroleum produced issues from rocks older than the Carboniferous, while the formations yielding bitumen, in by far the greater number of localities, are of Eocene age. In the great "oil-region" of the United States petroleum occurs in crevices to a very limited extent. In Canada and West Virginia it occurs beneath the crowns of anticlinals, and in Pennsylvania it saturates the porous portions of formations that lie far beneath the influence of superficial erosion, like sand-bars in a flowing stream or detritus on a beach. These strata are not of any particular geological age, but run through a vast accumulation of sediments embraced in all the forma tions between the Lower Devonian and Upper Carbonifer ous. They lie conformably with the enclosing rocks, and slope gently to the south-west. The Bradford field in particular resembles a sheet of coarse-grained sandstone 100 square miles in extent, by from 20 to 80 feet in thickness, lying with its south-western edge lowest and submerged in salt water, and its north-eastern edge highest and filled with gas under an extremely high pressure. In Galicia the sandstones holding the oil are very much dis turbed, while in the Caucasus the deposits of sand are erratic both in regard to position and extent, and lenticular in outline, being enclosed in a formation consisting of stiff blue clay. Chemistry. The first chemical research upon petroleum was conducted by Vauquelin in 1817 upon the naphtha of Amiano. Prior to the discovery of petroleum in commer cial quantities, a number of European chemists had made determination of the atomic constitution of several different varieties, and it had become generally understood that the oil consisted of an equal number of atoms of carbon and hydrogen. It has since been determined that some varieties of petroleum contain nitrogen and others contain sulphur and oxygen. These last-named elements are, however, to be properly considered as components of impurities. The proximate principles of petroleum have been determined and examined chiefly by Schorlemmer in England, Pelouze and Cahours in France, and C. M. Warren and S. P. Sadtler in the United States. Many other chemists have contri buted valuable assistance to the work. These researches have established the fact that Pennsylvania petroleum con sists chiefly of two homologous series of isomeric compounds having the general formula CnH-zn+2, at one extremity of which marsh gas is found and solid paraffin at the other (see PARAFFIN). This oil also contains a smaller propor tion of the olefine series, having the formula C,iH-2n, with traces in the Bradford oil of the benzole series. Rangoon petroleum contains a larger proportion of both the olefine and the benzole series than Pennsylvania oil. It has been thown that Caucasian petroleum contains the additive compounds of the benzole group which have the same per centage composition as the defines and furnish an illumin ating oil containing more carbon than Pennsylvania oils of the same specific gravity. The residues from the manu facture of petroleum have been shown to contain very dense solids and liquids of high specific gravity, having a large proportion of carbon and possessed of remarkable fluorescent properties. Some petroleums are easily oxidized into asphaltum and kindred products. Colourless illumin ating oils under the action of light absorb oxygen, which is converted into ozone, and they become yellow and viscid and of greatly impaired quality when the action is prolonged. Origin. The origin of petroleum has been a subject of speculation among scientific men during the last half cen tury. It is a subject involved in much greater obscurity than the origin of coal, for, unlike coal, it has no organic structure ; hence it can only be inferred upon circumstantial evidence that it is of organic origin ; yet such evidence is so strong that few competent judges have ventured to decide otherwise. The arguments in favour of a chemical origin have been advanced almost wholly by a school of French chemists during the last twenty years. They are based upon the results of a class of experiments first inaugurated by Berthelot, in which powerful deoxidizing agents like the alkali metals or iron at a white heat are caused to react with steam and carbonic acid. The hydrogen of the water and the carbon of the carbonic acid, having been deprived of their oxygen, unite in the nascent state to form a mix ture of oily fluids closely resembling petroleum. Sufficient quantities of these oils have been prepared to prove their identity with each other and with crude petroleum. Be fore concluding from this circumstance that petroleum is the product of similar reactions, it is necessary to assume a condition of the earth s interior concerning which we know nothing ; and, while the theoretical chemistry of the earth, based upon the nebular hypothesis, does not forbid such possibilities, there are other considerations relating to the origin of petroleum based upon the known rather than the possible that render the assumption that petroleum is of mineral origin forced and unnecessary. It is found that, when shale, coal, peat, wood, or animal matter, in fact any recent or fossil organic matter, is subjected to destructive distillation at low temperatures, there is ob tained among other products an oily fluid which chemistry shows to consist chiefly of the same compounds of carbon and hydrogen as are found in Pennsylvania petroleum. There are other petroleums, however, occurring in Canada, Tennessee, and other localities somewhat different in com position, which are often found under conditions that make it extremely difficult to account for their origin upon any hypothesis that does not regard them as a product of the decomposition of animal remains. They fill the cavities of fossil corals and orthoceratites in Canada and of geodes in Tennessee, in all of which the oil appears to be hermetically sealed until the rock-mass is broken. The formation in which these oils occur consists of thickly-bedded Silurian limestones that were probably deposited in a deep sea at a somewhat high temperature, in which vast quantities of sea-animals perished and became buried. It is therefore most strictly in accordance with observed facts to assume that these oils, in whatever manner they may have been produced from the original animal remains, are indigenous to the rocks in which they are found. These indigenous oils do not occur locally in considerable quantity, although the aggregate amount scattered through any formation in which they occur can easily be shown to be large. In those localities, notably north-western Pennsylvania and eastern Ohio, where petroleum occurs in large quan tity, it occurs quite uniformly, saturating heavy beds of uncemented sandstone. This sandstone is overlaid with an impervious shell of slate, containing much silica, that holds down both the oil and gas within the sandstone under great pressure, not locally in cavities but over wide areas. The sandstone is also, so far as can be ascertained, underlaid with a vast formation of shale more than 1000 feet in thickness, containing large numbers of fossil animals and such a quan tity of fossil sea-weeds that Dr J. S. Newberry has suggested that the Silurian ocean here contained a veritable sargasso sea. This shale, so filled with the remains of fucoids, has been several times submitted to destructive distilla tion, and has yielded as high as 50 gallons to the ton of distillate oil that was in many respects scarcely to be dis tinguished from crude petroleum. During the present century the French chemical geologists have held that all forms of bitumen are the product of metamorphism. Prominent among these may be mentioned Daubree, who in his Observations sur le Metamorphisme has shown the strict correspondence between his laboratory experiments, in which all forms of bitumen were produced, and the opera-
