The following extracts from a letter to Mr. Dean of the Central Foundry Co. from Mr. James C. Bayles regarding the waste of water and gas through leakage in distribution will be of interest as showing the very great need of further efforts toward the improvement of pipe construction and jointing.

Leakage Of Pipes And Joints 557

The vertical lines of division in this diagram represent the hours from midnight to midnight. The parallelogram in solid block represents the continuous leakage of water which goes on day and night in practically unvarying volume.

American practice is not, on the average, better than British practice in water works engineering.

Fig. 519.

Relation of Use to Leakage in Water Supplies.

To put the average leakage of water mains and services in this country at fifty per cent of the intake, would be far within the truth. Mr. William Hope, C. E., an eminent British water works engineer, stated the truth briefly in his paper before the Institution of Civil Engineers as follows:

"Even now a majority of the water undertakings of this country lose by leakage more than one-half the total quantity of water supplied from the source. The proportion is often higher, and rises in many instances to three-fourths or more, while in comparatively new countries, such as America, Australia and New Zealand, the proportion so lost is still greater."

Nearly all the water distributed in cities and towns is carried by cast iron mains laid with hub and spigot joints.

Gas Leakage.

For American cities I am of the opinion that from 12 to 15 per cent of total output would be a fair average for leakage loss. In some instances it is very much greater. Losses of 25 to 30 per cent are not exceptional, and I have known a company with all wrought iron mains put together with screw joints to have a leakage loss of 60 per cent. In large cities losses materially exceeding an average of half a million cubic feet per mile of main per annum do not surprise the well informed gas engineer.

Expansion and contraction are constant. Pipes in the ground are never at rest, but lengthen and shorten with each variation of temperature. This movement is slight, rarely exceeding two inches in a thousand feet through the range of a year, but it is irresistible. If not accommodated it will accommodate itself. The forces acting upon a pipe are so much stronger than any iron ever made that the rigid pipe is inevitably broken. I attribute 80 per cent, of normal main leakage to the destructive influence of expansion and contraction.

To this may also be attributed the great leakage found at service taps. To expect that two lines of pipe laid at right angles one to the other, rigidly connected and each expanding and contracting longitudinally, will remain in gas or water tight connection is to expect a miracle, and modern engineering does not deal with occult phenomena.

Unequal settlement results from many causes too well known to need explanation. It opens joints or causes fractures according to circumstances. Pipes which cannot bend are readily broken by vertical or lateral displacement.

Disintegration by oxidation and electrolysis need no discussion.

Jar, shock and vibration affect pipe lines in proportion to their rigidity. Lines of pipe so laid as to be flexible and expandable, even with narrow limits, are very little affected by these disturbing influences.

The screw joint of wrought iron pipe is as rigid as the pipe itself. For this reason the threads, being less strong than the pipe wall, are liable to strip under the action of contraction and expansion. I have seen a line of pipe taken apart by hand which it required four men at each joint to put together. The hub and spigot joint is a crude stuffing box placed with an inelastic material. I have seen a line of hub and spigot pipe laid in one day and tested bottle-tight at 6 p. m. leak at 7 a. m. the next morning from the shrinkage due to taking the temperature of the trench over night. The chief advantage of the hub and spigot joint is that when packed with lead it will permit a certain amount of slip and thus to some extent relieve the strains tending to fracture. That it is always a leaky joint is a fact too well known to need the support of argument. When packed with cement, as in much of the modern gas practice, leakage at the joints is minimized, but fractures are much increased. The difference as affecting net leakage is not material, but a great many engineers prefer to deal with occasional great leaks due to breakage rather than with innumerable small ones at joints.

4. Main Leakage and the Public Interest.

If the consequences of main leakage which are of public concern were limited to the waste of what is lost thereby, it would be difficult to make the matter appear important.

In the case of water it would be argued that it "cost nothing" originally and should be "as free as air" to the user. If some loss in distribution was admitted, it would be contended that measures of waste prevention would cost more than waste replacement, and that it is better public economy to lose two gallons in distributing three than to conserve two and a half and lose only one-half. This might be plausible, but it would be essentially untrue. Few municipalities are so situated that their sources of supply of potable water are equal to the triple demand of leakage, waste and use. A normal increase of use may be counted on; waste will increase in more than arithmetical ratio unless checked by metering and of leakage we may be sure it will keep well ahead of both use and waste together. As the height of buildings is increased more pressure is needed for fire purposes and circulation. The escape of water from a given defect in a conduit varies under different pressures as the square roots of the pressures compared. For example: A defect in a pipe which under a pressure of forty-five pounds will leak 12,960 gallons in twenty-four hours (and a round hole one-fourth inch in diameter, or its equivalent, will do this) at sixty pounds will leak 14,431 gallons, and at ninety pounds will leak 18,600 gallons. This is theoretical. In practice, as pressures are increased old leaks grow larger and new ones are developed. With a defective distributing system, the relief of increased pressure is usually very transient. It does not always, if often, increase the available supply. The resulting evils, besides an increased and ultimately burdensome public expenditure, are scarcity of water for such publie uses as street sprinkling and washing, sewer flushing, etc., a low pressure service which must be supplemented by house pumping and an inadequate fire protection. New York is now "threatened with a water famine." Without main leakage it would have in its present supply all the water needed for three times its present population, and might safely postpone plans of water works extension until near the close of the present century.