Pitot Tube determinations have also been used, especially on natural gas mains. The method is well described by W. B. Gregory, of New Orleans, before the American Society of Mechanical Engineers (see volume XXV of the Transactions.) F. H. Oliphant also describes it in his report on natural gas above noted. The formula of the Pitot tube as applied to gas measurement has been very carefully determined by Prof. S. W. Robinson, whose report is published in the Geological Survey of Ohio, volume VI, page 548. The final formula, calculated for gas of .06 specific gravity and a temperature of 50° F., is

Q = 1690 d2 √h (1+ P3)

All the above formulae were derived from Dr. Pole's formula and are more or less successful attempts to add corrections for density and coefficients of friction due to compression. more complicated formulae are simply attempts at greater refinement.

It has come to be pretty generally understood that each size pipe requires a change in the formula to give the particular results obtained. The coefficient of friction does not seem to vary with different densities of gas, but varies with each size pipe. Unwin says that the coefficient of friction varies with the roughness of pipe, diameter and velocity, and I would add possibly density.

I believe that the formulae recognizing a different coefficient for each size are the most accurate. I do not believe the full accuracy of any formula is yet established. Accurate results from actual tests should be kept by some one, and tabulated until a true formula can be secured. Until such time, practical working results can be had from the best natural gas practice as embodied in Prof. Robinson's formulae, and Cox's computer, also, which give safe results.

I have noted three records, the accuracy of measurement of which I do not know. At Newton, Mass., a 6-inch pipe line 9,600 feet long, 10 lbs. per square inch initial gauge pressure, 9.28 lbs. final gauge pressure, delivered 18,200 feet per hour. At Chester, Pa., a six inch pipe line 7,514 feet long, 27 lbs.

initial gauge pressure, 25.5 lbs. final gauge pressure, delivered 20, 100 cubic feet per hour. (Not satisfied with accuracy of measurement.)

Mr. Oliphant says he checked his formula on delivering natural gas 100 miles into a gas holder through 8 inch pipe. Taking the Newton conditions and using the several formulae we obtained the following results :

I am indebted to Dr. H. B. Harrop for considerable assistance in obtaining several of the above formulae.

4. Pumping Gas. Only a few limiting conditions will be given, with a classification according to the services.

FOUL GAS.

With suction pressures in tenths of an inch water column and discharge pressure amounting to only a few inches, this is purely a rotary positive blower proposition. The vital points here are close governing to protect the hydraulic main seals, and machine construction such that there can be no fouling from tar. Higher speeds by fifty per cent. than heretofore used would insure better governing, while allowing for greater variation of output.

The cooler the gas is handled the less capacity will be necessary in the machine. Where the meter measurements are used to figure the required capacity of the exhauster, the volume at the pressure and temperature at the machine should

be calculated with allowance for shrinkage due to condensable matter dropped, and 10 per cent. for actual slip in the machine. Sufficiently close results may be obtained by using the formula :

TRANSFER OF GAS FROM HOLDER TO HOLDER.

For pressures not exceeding eight or ten inches water column, a centrifugal blower may be used; for pressures over that the rotary positive blower will be more satisfactory. The constant speed machine is the correct application here, and close to maximum speed may be used unless a margin is specially desired. In large stations three machines, any two being of sufficient size for the work, should be used. This is especially true of gas engine outfits, for the economy of large and small units differing but little, sub-division in the above proportions is wise and not too costly.

TRUNK LINE WORK.

The conditions usually vary from 2 to 6 pounds; and high speed, high pressure rotary positive blowers are most suitable. Either constant or variable speed machines are permissible, the constant pressure machine seldom being necessary except as a protection against excessive pressure, due to stoppages of any kind. Here an automatic by-pass is ample protection; though a diaphragm regulator that can be changed readily is an advantage under certain cirumstances. Efficiencies vary from 75% to 92%, according to size and pressure.

TRANSMISSION LINES, IO POUNDS AND UP.

The real problems of pumping gas begin here. Up to ten or twelve pounds the positive rotary gas pump should be considered. Simplicity, flexibility and first cost may be balanced against increased cost of attendance, valve complications and adiabatic efficiency. Above twelve pounds the field belongs to the reciprocating compressor, unless some unusual condition. would justify compounding the rotary blower.

The balance of the problem is a question of fuel against cost of pipe; small pipes and high fuel account, large pipes and low fuel.

THE MECHANICAL ASPECT OF GAS DISTRIBUTION.

In the following revision of the material contributed to the Question Box of the Ohio Gas Light Association, "The Mechanical Aspect of Distribution" has been treated as viewed by the man who was either actually in charge or performing the work. In handling the subject the information gathered in the Ohio Question Box has been added to by information gathered from other sources wherever it seemed necessary to do so, care being taken to give the proper credit.

In case of a difference of opinion, where there was no preponderance of evidence in favor of either side, we have endeavored to present both sides, together with the argument, leaving the reader to decide.

The subject naturally divides itself into the consideration of mains, services, meters, and inside piping. The subject of mains may be further divided into low pressure and high pressure. In the consideration of the subject of low pressure mains, the question of the material used for the main is of the first importance, as the preparation for the work to be done must be made to conform to the best use of the material selected. Therefore we might further divide the subject of low pressure mains into cast iron mains, wrought iron or steel mains, and mains of miscellaneous materials.

5. Cast Iron Pipe. Up to within the last few years, almost all of the low pressure gas mains were laid of the ordinary type of bell and spigot cast iron pipe. "This pipe has the advantage of not being easily corroded, and, barring electrolysis, it will last for at least fifty years in ordinary soil."-(D. McDonald, Western G. A.-1900.)

It might be thought that it would be advisable to order cast iron pipe for gas mains to be coated with either coal tar or asphaltum in the same manner as cast iron water pipe, and thus still further prevent corrosion. This is not the case, however, as the coating forms a thin film between the material.

used for the joint and the pipe itself. In the course of time this film is dissolved by the action of the gas and the joint becomes leaky. In many cases, however, when it is suspected that there may be acid in the soil or the main subject to electrolysis, it would be advisable to coat the pipe as laid. This will be discussed in another section.

Fittings for Cast-Iron Pipe. In ordering fittings for the laying of a cast-iron line, it is advisable in nearly all cases to order the fittings with all hub ends, as this allows the short spigot pieces to be used which are cut from full lengths, and will also permit of uniform calking, as spigot fittings do not permit sufficient hammer room when the bell faces the fitting. It is also well to specify that the pipe and fittings shall be of the standard adopted by the American Gas Light Association, of which the table in the Appendix gives sizes and dimensions.

In case it is necessary these dimensions, also location of hub ends, spigot ends or flange ends, will be changed by the makers. For instance, if it were necessary a cross with one spigot end, one flange end and two hub ends could be procured. When ordering fittings, and especially crosses and tees or fittings having a special arrangement of openings, it is well to send a sketch of the fitting. In designating sizes, the run of the fitting is given first, and then the size of the branches. If necessary, the length of the run and branches should next be given, followed by the style of openings, thus :

In ordering flange fittings the diameter of the flange should be given, also bolt circle and the size and number of boltsalso note if the bolt holes are straddle or on center lines. By center lines is meant a line drawn across the face of a flange through the center in the same plane as the axial line of the side branch if there be any.

In case of side branch tees and crosses, the center line will