and at Chicago, Dr. Sterry Hunt concludes that the oil is indigenous to the Niagara, Corniferous, and Trenton limestones. Lesley asserts the same as regards the sandstones and conglomerates of Kentucky; and Orton considers that this is the case with the oil of North-Western Ohio. Similar observations are made as regards the asphalt of Trinidad (Wall), the oil of Shropshire in England (Höfer), that of Upper and Lower Alsace (Andreae), the bitumen of Seyssel (Davies), the asphalt of Limmer (Credner), the bitumen of the Lias clay and of the Wealden formation in Northern Germany (Eck), and the oil of the Punjab (Medlicott), and of Khátan in Baluchistan (Townsend).

The Pennsylvanian 1st oil-sand yields the heavy brownish-green oil of Franklin, Venango (31° to 33° B.), the greenish-black oil of Warren County (46°), the dark amber oil of the district between Titusville and Pleasantville (47.5°), and the light amber oil of Washington County (52°).

The formation of large deposits of oil depends as much upon the presence of suitable strata to receive and retain it as upon an adequate source of supply. "So common is the occurrence of petroleum in stratified rocks, that wherever a close-grained shale occurs, there is almost always, at least a small accumulation of oil directly underneath it. The same thing occurs when an impervious stratum of any other composition than shale occurs in the series" (Orton).

The principal deposits which provide the necessary porosity for the accumulation of the oil are sandstones, conglomerates, and limestones. Shaly-sandstones and slaty-shales also serve as reservoir rocks in a lesser degree. In the case of limestones, a natural porosity, as in the case of the "crinoidal" varieties, or a certain amount of change resulting in the formation of interspaces capable of receiving the oil, appears to be necessary for the formation of a true "reservoir-rock." Such change is usually dolomitic, and consists in the conversion of the calcium carbonate forming the limestone, into the double carbonate of calcium and magnesium, known as "dolomite," which occupies less space than the unaltered limestone. It is, therefore, characterised by the production of such numerous spaces between the dolomite crystals, that the rock becomes capable of retaining a large volume of oil.

This dolomitic change appears to be incapable of occurring, save in the purer limestones. The Trenton limestone, for instance, is thus modified only where almost free from silica, the changed parts showing about 54 per cent. of calcium carbonate and 44 per cent. of magnesium carbonate. The major portions of the Trenton limestone are too impure to permit of the change, and are found destitute of oil or gas.

Even in rich oil-fields, the dolomite has only been formed in a small proportion of the stratum. When followed northward in Ohio and Indiana, the Trenton limestone is found to have become dolomitic through a small thickness only of its upper beds. The changed and unaltered portions occur at short intervals, but only the former contain the oil and gas. The change usually affects from 10 to 50, and, in some cases, 100 feet of the stratum, and has occurred along a line passing into Indiana through the principal oil and gas fields of Ohio.

In addition to possessing a porous structure, the reservoir-rock must be entirely covered by an impervious layer, the commonest and most perfect cover being a fine-grained shale, whose imperviousness and freedom from fracture exert the most important influence on the capability of the reservoir-rock to retain its liquid or gaseous contents. Continuity of the impervious cover is essential, and the broken nature of the deposits of central and eastern Pennsylvania readily accounts for the absence of oil and gas in those districts.

In oil-bearing territory, the occurrence of a porous rock beneath such a cover usually results in the formation of an oil field. "Almost every important mass of shale in the Ohio series has been proved to be somewhere the cover of accumulated petroleum, as, for example, the Cuyahoga and Berea shales covering the Berea grit, the Ohio shale covering the Corniferous limestone, the Niagara shale covering the Clinton group, and, finally, the Utica shale covering the Trenton limestone" (Orton).

If not charged with oil, porous beds are found to contain water, which may be fresh when near the surface, but is salt and often sulphurous at lower depths. As the oil is removed from a stratum, it becomes replaced by water. In the case of gas wells, however, the flooding by water does not invariably occur, many wells having been pumped for years to such an extent that a vacuum of as much as 12 lbs. per square inch is registered without any appearance of water. In gently-dipping strata of sufficient inclination, such as the Bradford sand, the gas, oil, and water form distinct horizons, the gas being uppermost and the water lowest.

As regards the capacity of the various oil strata to serve as an oil reservoir, experiments performed on the rock itself have shown that an oil-bearing pebble rock may contain one-tenth, or even one-eighth, of its bulk of oil, and the pores of the rock would permit of the ready removal of the largest supplies ever obtained, without the necessity for the channels which were at one time supposed to exist.1 The dolomitised portions of the Trenton limestone have been found to possess about the same capacity. Dr. Orton2 has pointed out that if a stratum "a few hundred feet in thickness carries but one-tenth per cent. of petroleum, every square mile of such territory would contain more oil than has ever been removed from a like area of the most productive field; and (loc. cit., p. 86) taking Dr. Sterry Hunt's calculation of the oil contents of the petroliferous dolomite of Chicago as a basis, he has determined the probable contents of the almost-universally petroliferous Waterlime stratum of Ohio. "Estimating its petroleum content at one-tenth of one per cent. and the thickness of the stratum at 500 feet, both of which figures are probably within the limits, we find the petroleum contained in it to be more than 2,500,000 barrels to the square mile. The total production of the great oil-field of Pennsylvania and New York to January, 1885, is 26,000,000 barrels. It would require only three ordinary townships, or a little more than 1 See Carll, Second Penn. Survey, i. 251.

First Ann. Rep. of the Geol. Survey of Ohio, 1890, 58.

100 square miles, to duplicate this enormous stock from the Waterlime alone." The oil-sand of the region of Baku is estimated to contain about one-fifth of its volume of oil.1 These estimates would account for a yield far exceeding the amount that has actually been obtained, although, as pointed out by Ashburner, small areas of the best fields of Pennsylvania have yielded as much as 900,000 barrels per square mile. As regards the capacity of gas strata, Mr. Carl12 has made the following interesting calculations:

66

Although large quantities of gas had been allowed to go to waste in the Murraysville field, comparatively little had been piped from it prior to November 1st, 1884. If we take this date for the commencement of the general pipe-line deliveries, the field has now (January 1st, 1889) entered upon the fifth year of its exhaustion. It is said by those who are qualified to judge of the matter, that an average of at least 300,000,000 cubic feet of gas per day has been drawn from the pool during the last four years. To get some conception of the enormous storage capacity of a rock capable of making such an output, let us look at the following simple calculations and comparisons:

"A standard oil barrel of 42 gallons [American] contains (42 × 231) 9,702 cubic inches, or 5.6146 cubic feet. Under a rock pressure of 900 pounds or 60 atmospheres to the square inch, every cubic foot of storage room in a gas sand should contain 60 cubic feet of gas. Hence a rock chamber capable of holding a barrel of oil would be able to hold sixty times that amount of gas or (5-6146 × 60) 336-8760 cubic feet; consequently, a gas well flowing 3,368,760 cubic feet of gas per day would, theoretically, relieve as large a space in the rock-reservoir as an oil well flowing 10,000 barrels per day, and a gas well of 33,687,600 cubic feet (some wells have been estimated as high as that), would equal an oil well of 100,000 barrels.

"Multiplying the daily average by the number of days in four years we have (1,460 × 300,000,000) 438,000,000,000 cubic feet as the total output of gas. Reduced under 60 atmospheres, this would occupy a storage space of 7,300,000,000 cubic feet, a chamber which would contain 1,300,181,669 barrels of oil, almost four times as much as has been produced in the whole oil regions of Pennsylvania and New York since the discovery of oil in 1859. A reservoir capable of holding this enormous quantity of fluid would require to be a little over 13 miles long, 2 miles wide, and 10 feet deep."

Where the oil and gas strata are comparatively undisturbed, as in the principal districts of the United States, it usually happens that each well draws its supply from a considerable area; indeed, so much is this the case that the owners of wells are usually compelled to continue to raise the oil without regard to the condition of the market, to prevent its being obtained by the neighbouring leaseholders. In North-western Ohio, for instance, wells separated by intervals of over half a mile have been found to affect each other, and a single well has been known to draw the gas from an area of several square miles.

On the other hand, certain rich wells, notably the "Manifold" well 1 Mendeléeff, Russian Zeitsch. Technik, 1886, No. 109. 2 Seventh Rep., 1890, 18.

of the Washington field, which appeared to be totally unaffected by its numerous neighbours, and the great "Hutson" gas well struck early in 1890 in the Allen Township of Hancock Co., Ohio, have been wholly or comparatively free from such interference.

The wells of Baku, which receive their oil from an extremely loose and friable sandstone, are, to a great extent, independent of each other, so that less necessity for bringing the oil to the surface exists in that district, and a considerable number of very productive wells belonging to the firm of Nobel and to others are kept capped until the oil is required for use. The great "Droojba" well before-described, showed no appreciable action on other wells in its vicinity while it was flowing; although, when it was closed, a great disturbance occurred at Nobel's No. 14 well.

At Bibi Eibat, near the village of Strikhoff, four wells have been producing within a few yards of each other and all from different depths. In the same district a well was sunk within a few yards of an old well, 70 feet in depth, which had furnished oil for centuries. The new prospectors, however, failed to obtain any oil until they reached a depth of 420 feet. Meerzoeff obtained no oil from one of his wells at Surakhany until a depth of 700 feet had been penetrated, although other wells in close proximity were yielding at 100 feet.

There is, however, an evident connection between many of the wells of Baku. Three wells, for instance, belonging to Messrs. Awakoff, Palashkowski, and Nobel, were found to invariably spout periodically at about the same time. Direct communication has also been found to exist between the "Ararat" and "Sun" wells in Group V. of the Balakhany plateau.

The practical independence of the Baku wells, and the enormous yield of many of them, have led to a widely-spread opinion that this oil field is divided into a number of oil-containing cavities or, as they were described by Mr. Marvin, "subterranean lakes," yielding only to those wells which happened to directly pierce them. They may, however, be attributed with greater reason to a disturbance of the oil strata, resulting in its division into a number of almost independent sections. This view is supported by the records which have been kept by the Brothers Nobel of the appearance and geological character of the strata pierced by their wells. The hypothesis that cavities exist is quite gratuitous, as the natural porosity of the strata is sufficient, as already noted, to account for a far greater yield than is practically obtained.

Professor Lesley suggested that the gas areas, which are evidently small, and are irregularly scattered and hemmed in by water and oil areas, shift their position slightly, partly on account of the pressure of the oil and water and partly from seismic changes. He considers that this movement tends towards the working wells, and will become increasingly rapid as the gas is removed.

The Association of Salt with Petroleum.-The peculiar relations between salt and petroleum and natural gas were noticed at an early date.

The petroleum industry of the United States originated in

the drilling of wells for brine, and the observation that gas and oil were usually found with it; and throughout the globe the association of gas and petroleum with salt, either in solution or in the solid state, is almost universal.

The occurrence of salt in the oil districts of America, Russia, Galicia, and India is too well known to need remark; while in Japan and China it is equally noticeable, many of the salt workings of the latter country having been lighted by natural gas from time immemorial. The salt mines of Szalino in Hungary are similarly illuminated.

An early article by Dr. S. P. Hildreth1 on the saliferous rock formation in the valley of the Ohio, is of interest in this connection :"All salt wells afford more or less of this interesting gas-an agent intimately concerned in the free rise of the water, and universally present where salt water is found. Indeed, so strong is the evidence afforded by the rising of this gas to the surface, of the existence of the salt rock below, that many wells are sunk on this evidence alone. It is, without doubt, a product of the saliferous formation, as it rises in many wells without any appearance of petroleum. In many wells, salt water and inflammable gas rise in company, with a steady flow. In others the gas rises at intervals of ten or twelve hours, or perhaps as many days, in vast quantity, and with overwhelming force, throwing the water from the well to the height of 50 or 100 feet in the air. A well on the Muskingum, 10 miles above M'Connelsville, at 600 feet in depth, afforded such an immense quantity of gas, and in such a constant stream that, while they were digging, it several times took fire from the friction of the iron on the poles against the sides of the well, or from the scintillations from the auger, driving the workmen away and communicating the flame to the shed which covered the works."

The district of Baku was formerly highly esteemed for its rock salt,2 and a large deposit of rock salt extends along the foot of the outer slope of the Carpathians through Galicia and Bukowina into Moldavia. On the other hand, large saline deposits frequently occur without the appearance of any noticeable quantity of oil or gas. This is particularly observable in the great salt deposits of England and of Salzberg. The earlier attempts of chemists to explain the origin of petroleum frequently included salt as one of the factors engaged in the process, but modern theories usually ignore it, the brine which is found in most porous strata being considered as merely a natural constituent of waters which have penetrated into deep-seated strata.

The Association of Mud Volcanoes with Petroleum. - In Russia and India, the relation between petroleum and mud volcanoes is very noticeable. In the former country the presence of these volcanoes is usually considered a favourable indication of the presence of petroleum, as has been asserted by Professor Mendeléeff and others. They are found in the neighbourhood of the Balakhany and Benegadi fields of the Apsheron Peninsula, and are frequent in the Crimea and Am. Journ. Sci., 1833 (1), xxiv. 61. 2 Enc. Brit., 3rd ed., 1797, article "Baku."