III.—THE PIPE-LINE.

The points on which the Commission of English engineers were asked to advise were, as regards pipes and main generally—

(a) "Whether the pipes should be laid in a trench and covered in, or left exposed to view."

(b) "Whether it would be safe to rivet up the whole line of pipes, or whether joints, to allow for contraction and expansion, are necessary; the kind of joint most suitable, should they be necessary."

(c) "Material and method of manufacture of pipes, whether welded or riveted, and whether welding and riveting shall be square or spiral.” The use of cast iron being prohibited by the cost and the difficulty of freight both by sea and by land, the Commissioners were not to take it into account.

(d) "The diameter and thickness of pipes, and method of protecting."

As regards (a), the Commissioners were informed that there were possibly deleterious salts in the soil of a large part of the district through which the aqueduct would pass; and, for this reason, and also in order to avoid pressure on the empty pipes, to save the expense of trenching, and especially to facilitate detection and suppression of leakage, they recommended that the pipes should be laid above ground, uncovered, with expansion-joints.

The Commission recommended that the pipe should be of steel throughout, supported on bolsters, and riveted up in lengths of about 110 feet, with expansion-joints at these intervals, and anchor joints midway, fixed to masses of concrete or piles, in order to prevent the pipes from creeping. The minimum thickness was fixed at inch; and the pipes were to be longitudinally riveted where the pressure was such that the thickness of shell for riveted

pipe was not required to be greater than inch, and welded for all higher pressures; with a minimum thickness of inch.

Tenders for the pipes were accordingly invited from Australia, Europe and America, the quantities specified for being as shown in the Table on p. 70. Tenderers were at the same time invited to submit alternative prices for any other kind of pipe which they desired to put forward. The lowest of the tenders received were as follows, the prices being for delivery in the Colony at a point 22 miles inland :—

The locking-bar pipe, for which alternative tenders were received, had been considered by the Commission and favourably commented on, but was not recommended for so large a scheme, because proof of its successful manufacture and use on any considerable scale was not then available. Subsequently, however, and before receipt of the tenders, 10 miles of main, 25 inches in diameter, had been laid in South Australia. It had been found that pipes made from 1-inch plate and fresh from the closingmachine would withstand a pressure of 400 lbs. per square inch-or nearly twice what would be allowed continuously in practice on pipes of this thickness of plate-without a weep; and, moreover, all pipes which did not stand the test could be passed back to the closing-machine to be reclosed, instead of being subjected to the usual caulking-processes so injurious generally to the plates and jointings. Practical use on a fair length of main also showed that the jointing could be successfully accomplished, thus leaving only questions of comparative cost and comparative usefulness to be considered in deciding whether the new pipe should or should not be used in place of welded and riveted pipes.

Taking first the Australian prices for locking-bar pipes and contrasting them with those for welded pipes, the saving is seen to be very marked, being within a few pounds of 50 per cent. Moreover, the price of locking-bar pipes was but little more than

that of riveted pipes. The lowest tenderers were therefore asked to consider the matter again, and they quoted prices for the locking-bar pipes which contrasted as follows with those received for the riveted pipes :

Making a deduction of inch from the thickness of the plate to allow for corrosion and contingencies, and assuming a safe working-pressure of 7 tons per square inch of net section of metal, the safe head of water on pipes of these thicknesses, and 30 inches in diameter, is shown by the following Table:—

The locking-bar pipe being as strong as welded pipe, it would be possible to effect considerable economy by using-inch and 4-inch locking-bar pipes, in place, respectively, of the 1-inch and

-inch riveted pipes which had been specified originally; but it was recognized in the State, when pipes of so small a thickness as inch were included in those to be tendered for, that great care would be required in handling them, in order to prevent damage; and one result of the favourably low tenders was that a minimum thickness of inch was provided throughout, thus greatly increasing the probable life of the main in the very portions where the soil is worst, and the variations in temperature greatest. Moreover, by having one thickness and one diameter throughout, the contractors were induced to make a further reduction of 58. per pipe, so that the whole length of main was laid with pipes 30 inches in diameter, thus effecting some saving in the capital cost of the pumps, as well as in the cost of pumping.

Summing up the position, the results of adopting lockingbar pipes and a uniform diameter throughout are these:-The section of the ground traversed by the pipe-line is such that, considered purely from the point of view of obtaining minimum pressures on the main throughout, it would be advisable to vary the diameters and thus use up superfluous head; but the variation of pressure with a uniform diameter could not be large if the pumping-stations were suitably located, and this slight disadvantage was considered to be more than counterbalanced by the reduction in the cost of the pipes and the other advantages attending a uniform, and to some extent larger, main. Moreover, the substitution of locking-bar for welded pipe, effected a saving of no less than 50 per cent. of the cost of the latter; and, although, as compared with the riveted pipes tendered for, the locking-bar pipes eventually provided cost 11 per cent. more, on the other hand, the latter were considered superior in several ways. Their frictional resistance, according to older accepted formulas, was less in the ratio of 2.5:3.1, a difference of 25 per cent.; and the probable damage in handling 4-inch in lieu of -inch plate pipes would be less, and the probable life of the pipes would be much longer: for the actual thickness required for safe working being about as 2 of lockingbar to 3 of riveted pipe the substitution of 1-inch plate lockingbar pipe for -inch riveted pipe meant a provision of inch of plate in place of for corrosion and damage; and the substitution of 1-inch locking-bar pipe for 1-inch riveted meant a provision of inch for corrosion in place of inch in the case of the riveted pipe, a difference, therefore, of 133 to 233 per cent.

6

As the adoption of locking-bar pipes obviated the serious and continuous loss of water which was to be anticipated from a pipe having multitudinous rivet-holes, the question was considered whether the soils in which the pipe would have to be laid would tend to shorten its life, and if so, to what extent. As already mentioned, the natural water obtainable on the goldfields is highly mineralized; moreover, it often contains free acids. Therefore thin unprotected pipes in contact with this water could not have any lengthy life-a conclusion which experience has confirmed; but careful analysis of the soils along the pipe-track (Table XI.) showed that, where mining-operations did not entail distribution of such water on to the soil in which the pipes might be buried, this soil has been so much leached as to have lost many of its harmful properties, except, of course, in the salt-impregnated beds of the so-called "lakes." It was decided, therefore, that in the latter situation the pipes should be laid on trestles above ground,

but covered with a low roof of galvanized iron; and that in the remainder of the section they should be buried, thus obviating any necessity for expansion-joints and permitting, in fact, the use of ordinary lead jointing.

The Coating. In determining the composition of the coating to be used, wide extremes of temperature had to be allowed for. The fierce and continuous heat of the goldfields summer, when the temperature in the sun attains 150° to 170° F., is sufficient to render even block asphalt plastic. On the other hand, the frosts of winter would injuriously affect too hard a coating; and, moreover, as experiments showed, the extreme dryness of the soil at certain seasons, together with the heat, would very likely cause some loss of essential oils. As the result of a large number of tests of mixtures, made both at the pipe-works and at the head office, the coating used consisted of one part of asphalt and one part of coal-tar applied as described later, and freely sprinkled with sand while still hot and soft, to reduce the risk of the coating running when exposed in hot weather. No doubt the latter object could have been attained by more boiling, but the harder coating-mixture would have been brittle and more liable to flake off the pipes. Even the coating used ran to some extent when exposed for many days to the hot sun; but all exposure of metal, owing to this and other damage, was systematically made good just before the pipes were buried. The inside of the pipes was similarly coated-except, of course, that no sand was applied; but, as the water passing through is soft, although containing 20 grains of solids per gallon, and as vegetable acids are absent, much corrosion of the interior surface is not anticipated: and where the pipes have been emptied and opened 12 months after water started to pass continuously through them, the interior has appeared to be as clean and good as when they were first laid.

Joints. A simple sleeve joint (Figs. 12) was adopted, the ring being 8 inches wide, and inch larger internally than the pipe externally, to allow for slight variations in the ring, and to permit of the use of lead filling throughout. For working-heads of 320 feet and less, the section of ring used was as shown in Fig. 13, the weight being 126 lbs. ; but for heads of more than 320 feet a stronger form was used, as shown in Fig. 14, the weight per ring being 160 lbs. The finished jointing has proved very effective, the loss through leakage being small. From the pipes alone, on Sections 1-5, it was found to be 343 gallons per mile per diem. From the whole length of 295 miles between the storage-reservoir at Mundaring and the last pumping-station it was found to be 480 gallons