Mr. J. W. Hill. under this pressure it was to be subjected to a thorough hammertest, and any pipe which showed leaking or sweating at the end of this time was to be rejected. After the pipe was made up in the usual lengths of 12 feet in the trench, the contract required that it should be subjected to a further pressure of 200 lbs. per square inch for a period of not less than 5 hours, this length of time being fixed to enable the inspectors to examine carefully each joint and length of pipe, and if there should be spongy places in the pipebarrel which had escaped inspection at the place of manufacture, this time was deemed sufficient to bring the water to the surface and enable such spots to be marked and the proper remedy applied. The test in the pipe-line also permitted of a trial of such forms of special castings as could not be conveniently tested at the foundry. All special castings in this contract, excepting curves and man-hole castings, were 48 inches to 60 inches in diameter, and were made up in the valve-chambers with flanged joints. Tees and crosses were flanged castings. Forty-eight-inch to 60-inch taper-pipes were constructed with a flange on the small end and a socket on the large end to join with the spigot of a 60-inch pipe; and all straight castings were tested at the foundry under the same conditions as applied to the cast-iron pipe. At the time this contract was made there was only one hydrostatic pipe-press in the United States of sufficient capacity to apply the required test, and the two other foundries which were employed in the manufacture of the pipe were required to build presses of magnitude adapted to this size of pipe and to the proof-test provided for. At one time (1903) the rate of progress in the manufacture of the pipe was so slow that the city was on the eve of going abroad to buy part of the work from the pipe-foundries of Great Britain and Belgium; and it was a pleasure to state that the English pipe-foundries which were called upon in this matter were entirely willing to undertake to meet the conditions of the contract, which were more exacting than any previously laid down in the United States in work of this kind. In addition to the hydrostatic test at the foundry and in the pipe-trench, it was required that there should be cast from the ladle for each piece of pipe at least one tensile test-bar and two cross-breaking bars. The tensile bars were 11 inches long, and approximately 1 inch square, corrections being made for exact dimensions upon breaking the bar in the testing-machine. The cross-breaking bars were 26 inches long, and approximately 2 inches wide and 1 inch deep, and were broken by placing them flatwise on supports 24 inches apart, and applying the load at the centre. The conditions with reference to metal were that

the tensile strength should be not less than 22,500 lbs. (10 tons) per Mr. J. W. Hill. square inch, and that the cross-breaking bars of standard dimensions and span should take a load of 2,280 lbs., equivalent to a maximum fibre-stress of 41,000 lbs. (18.3 tons) per square inch of section, and show a centre deflection of not less than 0.34 inch before breaking. When tenders were received for this contract it was found that the difference between the cost of constructing the work with castiron pipe and with riveted steel pipe amounted to about $186,000.00 (£38,130), an amount which in his opinion was too small to justify the risk incidental to laying in city streets these large lines of parallel pipes of material of less known durability than cast iron; and it was gratifying to note that the tests of the pipe at the foundry and after laying in the trench had never developed a fracture. Out of more than 2,500 pieces of this pipe manufactured to the present time, not one had broken in the proving-press under pressures sometimes as high as 350 lbs. per square inch; and naturally none had broken in the trench under the lower pressure of 200 lbs. per square inch, although the latter might happen should the pipe be injured in transportation. In centering these pipes, instead of following the usual practice of centering from the outside, after the first few lengths were laid this was done entirely from the inside; an inspector inside the pipe indicating by taps of a hammer on the barrel the direction in which the spigot was to be wedged in the socket to bring about concentricity of the spigot end of one pipe and the socket of the next. In the placing of the spigot of one pipe in the socket of the next, an allowance of inch was made for longitudinal motion, partly to compensate for such temperature-changes as might take place in the length of the line before the trench was filled, and partly to allow for inequality in settlement of the pipe after filling, which would not be so well provided for if the spigot of one pipe was brought solidly home against the shoulder of the hub of the next pipe. Valve-chambers were provided at intervals of about 4,000 feet, so arranged with line- and branch-pipe valves as to permit of cutting out of service temporarily any line of pipe which might be injured, and confining the supply during the period of repair to the remaining lines of the system. The worst condition that could arise in service would be the temporary loss of the use of one line of pipe between two valve-chambers, the remaining lines between the two chambers still being serviceable; while of course all the three lines or four lines when completedbeyond the chambers nearest to the point of rupture would be unaffected by the break. Where the pipe entered a valve-chamber

Mr. J. W. Hill. it was reduced from 60 inches to 48 inches in diameter, partly because the manufacture of 60-inch stop-valves had not been as extensive, nor the use of valves of such size as well established, as those of 48-inch stop-valves, and partly because the height of such stop-valves would have materially increased the depth of the trench in which the pipe was laid. The number of chambers, and therefore of reductions in size from 60 inches to 48 inches in each line of pipe between the pumping-station and the terminal point at Torresdale and Kensington Avenues, was six, and the estimated loss of head by reason of the use of reducers from 60 inches to 48 inches, of 48-inch tees and crosses in the chambers, and of 48-inch stop-valves at each chamber, as computed by the formula given by Weisbach,' was less than 0.35 foot for a velocity of 7.39 feet per second through the 48-inch stop-valves and special castings, corresponding with a velocity through 60-inch pipe of 4.73 feet per second: an amount altogether too small to raise any doubt of the wisdom of the reduction of the diameter of pipe from 60 inches to 48 inches while passing through the valve-chambers, in view of the very large reduction in the cost of trenching by the use of 48-inch stop-valves. All valve-chambers were constructed with concrete floors and side and end walls. Upon the walls were placed rolled-steel beams, with arches sprung from the lower flanges of the beams, to support the roadway and paving which was carried over the chamber. Access to the chambers was provided through two manholes at corners diagonally opposite. The special castings in the valve-chambers, tees, crosses and taper-pipes, were made of open-hearth cast steel, having a tensile strength of approximately 60,000 lbs. (26.8 tons) per square inch of section, and the bodies of the 48-inch stop-valves, which were subjected to twisting and oblique stresses during test in the line and in service, were also steel castings, to reduce the chances of rupture. While the lines of pipe were tested to 200 lbs. pressure per square inch upon completion in the trenches, the working-pressure was limited to 125 lbs. per square inch by a series of 4-inch Crosby reliefvalves, one of which was placed on each line of pipe where it entered a valve-chamber. In addition to the relief-valves on the pipe-line, each of the pumping-engines (of which there were twelve of 20,000,000 U.S. gallons daily capacity, or a total of 240,000,000 gallons) delivering into these lines of pipe was provided with a 12-inch relief-valve set to lift at 120 lbs. per square inch. Thus the probability of ever producing a water-ram sufficient to injure

1 J. Weisbach, "A Manual of the Mechanics of Engineering, etc." vol. i. p. 872. London, 1877.

the pipe in service was extremely remote; but anticipating that Mr. J. W. Hill. rupture might occur under unaccountable conditions, and that with the large volume of water flowing through these lines of pipe, great damage might be done to the bottom of the trench, and probably to the parallel lines of pipe, the pipes were everywhere laid upon a concrete floor 9 inches thick, extending from side to side of the trench. It was thought that, if a rupture should occur in service, the washing out of the filling over a pipe would simply open the street and cut the banks at the side, but could not by any chance shift from their positions and injure the adjoining lines of pipe. In constructing the floor, which was laid after the pipe-lines were blocked in position and tested, a concrete cheek, formed as a monolith with the floor and extending up on the curve of the pipe for a vertical height of about 9 inches above the floor, was provided on each side, thus forming a cradle to guard against the lateral movement of the other pipes should one of the lines be fractured in service. In laying this pipe the contractors were required to construct a standard-gauge railway upon the side of the street, and to place thereon a travelling crane with a boom having a reach of 20 feet at right angles to the line of trench, and a capacity, with this reach, of about 10 tons. Each piece of castiron pipe weighed approximately 14,000 lbs. (6 tons), and each of the 48-inch stop-valves weighed about 22,000 lbs. (9.8 tons), these representing the heaviest castings used in the work. The contract provided that cutting pipes for closures should not be done in the usual manner with a hammer and chisel, but with a machine which would cut the pipe at right angles to its axis, and leave the end of the casting true and smooth, and which would accomplish the work without the possibility of fracture of either section of the cut pipe. In planning this system of pipes it was thought that in time there would be considerable tuberculation of the interior from the iron and lime salts found in the water of the Delaware River, which would be the source of supply for Lardner's Point pumping-station; and manholes were provided in each of the 60-inch by 48-inch taper-pipes through which access to the lines of pipe between chambers could readily be obtained, for the purpose of removing incrustations, and scraping and recoating the pipe. The removal, subsequent to preparing the contract, of a line of 30-inch pipe which interfered with the placing of the first four lines along Robbins Street, indicated that 27 years' use had produced very little deposit and no tuberculation in the pipe, and the manholes which were intended to be placed midway between the valvechambers were accordingly omitted, although those in the

Mr. J. W. Hill. taper-pipes were retained. Should occasion arise to examine the interior of these lines of pipe and remove obstructions, this could be accomplished by taking out of service one section of the pipelines at a time, between any two consecutive valve-chambers. These pipe-lines, with a calculated capacity of 240,000,000 U.S. gallons per day of 24 hours, were pumping-mains intended to convey filtered Delaware River water to several districts of the City of Philadelphia, which at the present time contained a population of about 1,100,000. The filtration of the Delaware water would be accomplished by sixty-five covered plain sandfilters each acre in area, constructed on the banks of the Delaware River, about 3 miles up-stream from the Lardner's Point pumping-station. In the valve-chambers each line of pipe was provided with a 24-inch relief-valve to be blown by hand. As the vertical undulations in the lines of pipe were not very great between their origin at Lardner's Point pumping-station and the terminus at Kensington Avenue, it was thought preferable to have these air-valves blown from day to day by an inspector, rather than to trust to automatic valves which might fail to act and probably subject the pipe-line to undesirable shocks by reason of accumulated air. At two points it was necessary to carry the pipes over creeks or large sewers, which necessitated the use of flanged pipe and the building of concrete chambers in the line of the sewer or watercourse, to confine the channels to definite widths. Two lengths of flanged pipe were bolted together in the centre of the span lengths, and acted as tubular horizontal girders of 16 feet clear span, of 60 inches internal diameter and 63.4 inches external diameter, calculated with a factor of safety of about 20, for a centre dead load of pipe, water, earth filling and pavement, of 47,256 lbs. (22 tons), and a maximum centre live load of 40,000 lbs. (17.8 tons). An interesting feature of this contract had been the necessity of moving under pressure a line of 48-inch cast-iron pipe about 1,194-45 feet long, which was one of the two pumping-mains from the original pumpingstation (Lardner's Point No. 1). To place this pipe in its new position to accommodate the four lines of 60-inch pipe in Tacony Street required that it should be changed 11.5 feet in alignment, and 13.17 feet in level at its lowest point. In carrying out the work the deflection from the old elevation to the new was made on the first 200 feet of the line at each end, the remaining portion being run on a tangent. As about 30,000,000 U.S. gallons of water were being delivered daily through the 48-inch main, it being the principal pumping-main, it was impossible to take