TABLE 2.-Results of analysis of samples of natural gas-Continued.

of Tillamook,

sec. 10, T. 2 S.,

R. 9 W.

NOTES.

Sample 1871.-The gas sampled comes from the fourth sand, which is 800 feet deep. The rock pressure in the well was 300 pounds per square inch. The capacity of the well was 3,000 cubic feet of gas per day.

Sample 1889.-The sample came from the third sand. Five gas wells were tapped by the pipe line from which the gas sample was drawn. The rock pressure at the wells was 62 pounds per square inch.

Sample 1872.-Collected in Clarion County, Pa., near the town of Mill Creek. The sample came from the Bradford or Little Bradford sand, which is 2,100 feet deep. The rock pressure was 850 pounds per square inch and the capacity of the well 3,000,000 cubic feet a day. This gas contains a higher percentage of the heavier paraffin hydrocarbons than does the gas represented by sample 1871 that comes from the shallower sand.

Sample 1848.-The sample was collected at well No. 14 of the A. W. Starr farm. The gas is from the Speechley sand.

Sample 1838.-The sample was collected from the No. 1 well on the J. S. Somerville farm. The gas comes from the fourth sand.

Sample 3177.—The sample was taken from the Collinsville pipe line in the pumping station of the Kansas Natural Gas Co. at the north end of the Hogshooter field. This line tapped about 40 gas wells. The initial rock pressure of the wells was 490 pounds per square inch. At the date of collection of the sample the pressure had dropped to 120 pounds. This was one of the few samples of natural gas analyzed by the bureau that contained only methane as the combustible constituent when the sample came from a sand in close proximity to sands producing oil. Most gas from the oil regions contains other paraffin hydrocarbons besides methane, and gas coming from the same sand as the oil invariably contains other paraffin hydrocarbons. It is difficult to explain the presence of an enormous quantity of methane gas, such as is contained or was contained in the Hogshooter pool, in close proximity to oil sands. If gas and oil were of common origin in the particular region, one would suppose that the Hogshooter gas should contain higher members of the paraffin hydrocarbons than methane, as do the gases from neighboring oil wells. The phenomenon might be connected in some way with the movement of the gas through strata whereby ethane and still higher paraffin hydrocarbons were separated.

Sample 2121.-The sample represented a casing-head gas from the Glenn sand. It was collected about 2,000 feet east of the railroad station at Kiefer, Okla.

Sample 2445.-The sample was collected at the Cully well on the Ellis farm. The well was 340 feet deep and was drilled in 1903. A large amount of hydrogen sulphide was present in the sample. The determination of the hydrogen sulphide was made by absorbing the H2S in a standard iodine solution and titrating with a standard sodium thiosulphate solution. Other sulphur compounds may have been present; hence the assumption that 2.9 per cent was all HS may be wrong. The gas possessed the odor of H2S.

Sample 2444.-The sample was obtained in the same locality as sample 2445, but from a different well. The well was 280 feet deep. Both wells were drilled in 1903. The hydrogen sulphide content of this sample is much smaller than that of sample 2445.

Sample 1031.-The sample is classified as a marsh gas in distinction from those gases that are found in the oil field and contain appreciable quantities of the higher paraffin hydrocarbons.

Sample 1033.-The sample was collected at a near-by seepage in the same locality. The gas was simply bubbling up through marshy ground.

Sample 1063.-The sample was collected from a slough on the Brown farm. It also represented a marsh gas.

Sample 1893.-The sample was collected from 15 oil wells, Nos. 1 to 15, of the Atlantic Refining Co. (South Penn Oil Co.).

Sample 1066.-The striking feature of this sample is the high percentage of nitrogen. A well had been drilled and the gas was issuing from it in considerable quantity. For present purposes to which natural gas is put, the gas would of course be worthless. Sample 1378.-The sample was collected from a slough about 15 feet in diameter. It represented what was seemingly another marsh gas.

The last eight samples given in the table were taken from natural gas supplied the city in which the sample was obtained.

COMMENTS ON THE ANALYSES.

It will be observed that the samples ranged in heating value from 724 to 1,657 British thermal units per cubic foot at 0° C. and 760 mm. normal pressure, except one sample which had the abnormally low heating value of 21 British thermal units per cubic foot.

The analytical results show only approximately the quantity of the individual hydrocarbons, although the percentages of total paraffin hydrocarbons are correct. The heating values of the samples, as calculated from the analyses, are also correct. A discussion of natural-gas analyses is found in Bulletin 42 of the Bureau of Mines." The causes of erroneous results that are frequently reported are there explained.

METHOD OF ANALYSIS USED.

The ascertaining of the exact proportions of the different hydrocarbons that may be found in natural gas has long been a stumbling block in gas analysis. The ordinary eudiometric method of analysis offers little in the way of a complete separation of a natural gas into its various constituents. Determination of the total paraffin-hydrocarbon content, with an approximate determination of the individual paraffins present, has been the only end attained. The Bureau of Mines in working on this problem succeeded in separating a natural gas into its individual paraffins by means of fractional distillation at low temperatures. Natural gas was first liquefied by means of liquid air, and then separated into its constituents by fractionation in vacuum at different temperatures.

The results of a complete analysis, including the quantity of each paraffin hydrocarbon found by the above method follow. For comparison, the results of an eudiometric analysis of the natural gas of Pittsburgh are also included.

a Burrell, G. A., and Seibert, F. M., The sampling and examination of mine gases and natural gas, 1913, 116 pp.

Included in the nitrogen content of the above analyses is 0.03 per cent of carbon dioxide present in the natural gas. The natural gas supplied to Pittsburgh can not be used for the commercial production of gasoline, although it contains sufficient of the higher paraffin hydrocarbons, the butanes, pentanes, and hexanes, to produce some condensation (drip) in the pipe lines in the winter time. These hydrocarbons are present in small quantity, as shown by fractionation experiments conducted by the bureau. It is only because of the immense volume of gas passing through the lines that appreciable condensation of vapor occurs.

A "wet" natural gas from which gasoline is obtained commercially was also subjected to fractionation." The results follow:

9.5

Butanes, pentanes, hexanes, etc.............

Total, including 1.00 per cent nitrogen and 0.03 per cent carbon
dioxide.......

PROPERTIES OF THE PARAFFIN HYDROCARBONS.

100. 0

The properties of those hydrocarbons of the paraffin series that concern the gasoline industry are given below:

a A technical paper covering in detail this method of separating gases and the results of experiments by the bureau is being prepared.

Holleman, A. F., Organic chemistry, edited by A. J. Walker, 1910, p. 41.

e Landolt and Börnstein, Physikalisch-chemische Tabellen, 3d ed., 1905, pp. 416, 425 (J. Thomsen). d Gas at ordinary temperature.

e Wright, L. T., Illuminating power of methane; Jour. Chem. Soc., vol. 47, 1885, p. 200. Landolt and Börnstein, Physikalisch-chemische Tabellen, 3d ed., 1905, pp. 185 (Dewar). Landolt and Börnstein, Physikalisch-chemische Tabellen, 3d ed., 1905, p. 185 (Olszewski).

A Frankland, P., Illuminating power of methane: Jour. Chem. Soc., vol. 47, 1885, p. 235.

í Landolt and Börnstein, Physikalisch-chemische Tabellen, 3d ed., 1905, p. 182 (Dewar). * Liquid at ordinary temperature.