bly, the oil absorption was only slightly different, and there was scarcely any change in the chemical analysis. After compression to 70 pounds per square inch, however, the gas changed markedly.

Regarding the results for plant 1, the chemical analysis, the specific gravity determination, and the claroline oil absorption show the gas represented to be a rich one. It will be seen that little difference existed between the composition of the crude gas and the same gas after it had been compressed to a pressure of 50 pounds per square inch. Only after the compression to a pressure of 250 pounds per square inch and cooling, did the composition of the gas mixture change appreciably.

The results obtained with the gas from plant 2, a small plant near McDonald, Pa., show that the crude gas was not very "wet." The absorption by claroline oil was rather low, and the gas was probably near the lower limit of a gas adapted for the production of gasoline. The composition of the gas was not changed to a very marked degree at any stage in the plant operation.

The results obtained with the gas from plant 3, 1 mile from plant 2, and the results with the gas from the latter plant show in a marked degree how natural-gas samples may vary in the same field. After the sample from plant 3 had been compressed to 20 pounds per square inch and cooled, a considerable change in the composition took place. The composition also changed considerably after compression to 80 pounds per square inch and cooling. Nearly 50 per cent of the gas delivered to the compressor consisted of air and nitrogen.

The results for plant 4, in the Glenn pool district, Okla., cover a gas probably about as poor as can be handled commercially. The condensate produced is extremely volatile. Other samples of gas from this same field are rich in gasoline vapors.

The results for plant 5 represent a gas that was derived from wells in the immediate vicinity of those connected with plant 1. The gas is of the same general character. The greatest change occurs after compression to 225 pounds per square inch and cooling.

Regarding plant 6, near Sistersville, W. Va., compression to a few ounces per square inch and cooling produced no gasoline, but after compression to 70 pounds per square inch and cooling, the gas underwent a considerable change in composition. A heavy condensation of gases and vapors occurred.

As regards the results for plant 8, near Steubenville, Ohio, most of the condensate was produced after the gas had been compressed to 90 pounds per square inch and cooled. Increasing the pressure to 160 pounds per square inch and cooling gave little additional yield, as is shown by the test results.

In plant 9 one-stage compression and cooling was practiced. High-grade plant equipment and excellent methods of handling the product were in use.

In plant 10 the composition of the gas, after compression to 125 pounds per square inch and cooling, changed markedly. Compression to any higher pressures would probably be useless, in view of the composition of the waste gas.

GENERAL CONCLUSIONS.

In general the preceding table shows that under existing methods of plant operation condensate is extracted from natural gas that ranges in specific gravity from as low as 0.83 to as high as 1.59 (air = 1) and that the solubilities of the gas in claroline oil ranges from 36.9 (air free) to 85.7 per cent, according to the well from which it comes. The authors hesitate to recommend the installation of a plant to handle natural gas that shows results as poor as the minimum values given in the table. Such gas might produce gasoline in paying quantities and might not. Probably the safest extremes would be a specific gravity of 0.95 (air-1), and a claroline-oil absorption of 40 per cent. The natural gas supplied to Pittsburgh, Pa., with which the authors are most familiar, contains little of the gaseous hydrocarbons, has a specific gravity of 0.64 (air = 1), and has a claroline-oil absorption of about 16 per cent. It is a dry gas and is unsuitable for gasoline production.

It will be seen that in many of the plants the gas delivered to the compressor contained 20 to 40 per cent of air due to inleakage either through the rock strata or through defective connections in the gas lines. Such leakage is obviously undesirable and every effort should be directed toward its prevention.

The waste gases from some plants seem to contain considerable quantities of gasoline-making constituents, as may be observed from the analytical results, including the specific gravity determinations, and the percentages of gas that was absorbed by claroline oil. The waste gases escape directly from the accumulator tank containing the last-stage compression products. These products at the existing temperatures have high vapor pressures, and consequently the escaping gases carry with them some of the vapors of the liquid in the tank.

SPECIFIC GRAVITIES AND ABSORPTION NUMBERS OF NATURAL GASES USED FOR CONDENSATION OF NATURAL GAS.

The authors have compiled the following table to show at a glance the specific gravities and absorption numbers of natural gases used for the condensation of natural gas. The table is compiled from the results shown in the large table preceding. The compilation will be useful for reference in predicting the results that may be obtained from other samples of natural gas.

Specific gravities and absorption numbers of samples of natural gas used for making gasoline.

FIRE HAZARD AT GASOLINE PLANTS.

Serious accidents have occurred in the manufacture of gasoline from natural gas and in the transportation of the gasoline. Several plants have been completely wrecked through the explosions of gasair mixtures that have resulted from the escape and ignition of natural gas in the engine and compressor rooms. Storage tanks have also been demolished when they have burst owing to excessive pressures put on them. Safety valves must be in good working order, and great care should be exercised in installing electric wiring systems and magnetos. All electric wiring should be properly placed, well insulated and made secure on supports. Fuses or electric switches should never be placed inside of buildings where gas might escape. Lighting dynamos and magnetos should be installed in small buildings outside the main building. Electric-light bulbs should be protected by wire guards, because they may ignite explosive mixtures if they are accidentally smashed."

The following table shows the small percentages of gases and vapors occurring in natural gas that are required to form explosive mixtures with air:

Low explosive limits for paraffin gases and vapors.b

According to the above table, even if a natural gas consisted almost entirely of methane, as some natural gases do, an explosion would follow an ignition of a mixture of air and natural gas containing cent of methane. The natural gas of Pittsburgh contains an

per

• Clark, H. H., and Ilsley, L. C., Ignition of mine gases by the filaments of incandescent lamps: Bull. 52, Bureau of Mines, 1913, 31 pp.

Burgess, M. J., and Wheeler, R. V., The lower limit of the inflammability of mixtures of the paraffin hydrocarbons with air: Trans. Chem. Soc., vol. 99, 1911, pp. 2013, 2030.

appreciable quantity of ethane and propane. The authors have determined its explosive limits to be about 5 per cent gas, low limit, and about 11.6 per cent gas, high limit. The "wet" gases, or those from which gasoline is condensed, contain a much higher proportion of the higher paraffin hydrocarbons than does the natural gas of Pittsburgh, and consequently have narrower explosive limits. Their low explosive limits are much lower. The authors have made no determination of the explosive limits of the "wet" natural gases from which gasoline is condensed, but they can closely estimate for many "wet" gases that such limits will range from about 3.5 per cent gas, low limit, to about 9.5 per cent gas, high limit. That is, when some "wet" natural gases are present in the mixture of the gas and air, an explosion will occur between these limits, although the limits for different natural gases will vary somewhat from the figures mentioned.

It will be observed then that only a small proportion of the gas need be mixed with air to form an explosive mixture, and that great care should be exercised to guard against leaks of the gas in plant buildings.

EFFECT OF LEAKAGE OF AIR INTO PIPES AND CONNECTIONS.

As mentioned before, at many oil wells that are connected to gasoline plants the gas is drawn at reduced pressures as low as 25 inches of mercury. Manifestly, under such low pressure, pipe and other connections must be very tight in order to avoid inflow of air. The authors have found as much as 40 per cent of air in the mixture drawn into some compressors. In some instances the air is drawn into the rock strata through old and abandoned wells that have not been properly plugged.

Introduction of air of course cuts down gasoline production and may lead the operator into believing that the quality of his gas as regards gasoline content is low. Residual gas may contain so much air as to be objectionable for further use. A determination of the oxygen in the gas as it comes to the plant would give complete information as to the air content. The authors know of few plants where such determinations are made. Oxygen is not a constituent of natural gas as it occurs in the earth. Hence, its presence in a pipe line shows inflow of air. Pure dry air contains 20.93 per cent of oxygen. Consequently, if the oxygen content of a gas is known, the percentage of air can readily be calculated.

CHARACTER AND USES OF RESIDUAL GAS.

In some gasoline plants gas issuing after some of the condensate has been removed is sent to mains and used for consumption in ordinary ways or returned to the leases to be used for pumping

purposes. In some plants it is wasted. As far as the quality of most of the gas is concerned, unless greatly contaminated with air, it is superior to that ordinarily used in cities and towns. Originally, as mentioned heretofore, it contains the permanent gases methane and ethane that are also contained in the natural gases ordinarily used for supplying cities. In addition there are contained considerable quantities of the more easily liquefied gases propane and butane, and some gasoline vapors. As the gas passes through the plant the gasoline vapors are to a large extent removed. Some of the propane and butane is liquefied and some of the methane and ethane goes into solution. The proportions of propane, butane, methane, and ethane that disappear may be small. The residual gas as it leaves the plant will contain all of the above-named constituents, because a complete removal of even the gasoline vapors is not accomplished in plant practice. The results of analyses on page 61 show the composition of raw and natural gases before and after passing through the plant. The high heating value of the residual gas is apparent, being in one instance almost twice as high as the heating value of the natural gas of Pittsburgh which has a gross heating value of 1,179 British thermal units per cubic foot at 0° C. and 760 mm. pressure.

Frequently the quantity of gasoline vapors in casing-head gas is so high as to cause trouble in gas-engine cylinders from premature ignition. The fact that casing-head gases vary so widely in composition can not of course be anticipated by engine builders. It is also true that casing-head gases may be unsatisfactory for lighting if mantles are used because the standard burner can not be adjusted for complete combustion of the gas. The result is sooty mantles and imperfect light. Removal of the gasoline vapors sometimes overcomes this objection.

SOLUTION OF GAS IN CONDENSATES.

As previously stated, one of the physical changes occurring in the operation of a gasoline plant has to do with the solution of gas in the condensate, that is, when the residual gas is in contact with the condensate in the storage tank. The following experiment and calculation by the authors will serve to show how small and insignificant this change may be.

A residual gas from an operating plant was shaken with refinery naphtha. The naphtha had a specific gravity of 61° B. The solution was effected at a temperature of 20° C. (68° F.) and atmospheric pressure. The naphtha was shaken with the gas supply until no more gas would go into solution. It was found that 1 liter of the naphtha dissolved 1.760 liters of the gas; or 500 gallons of the naphtha would have dissolved 3,331.7 liters of the gas. If the