confined by a mound across the river-bed constructed of rubble faced with concrete.

The excavations for the foundations were begun in May, 1898, and on their completion in January, 1900, the building of the wall was started, and was carried on both day and night until completion in June, 1902, an electric-lighting plant and eight arc-lamps placed at points of vantage affording ample light for operations by night.

Cement and Concrete.-In the construction of the weir-wall 76,418 casks of cement were used, and a further 1,090 in the spill-water basin and other subsidiary works, or a total of 77,508 casks, of which 19,767 casks were of German manufacture and the balance British. The German cement was chiefly used in filling the deep excavations made in sinking on the fault in the bed of the river.

The length of the average passage by steamer from London to Fremantle was more than 6 weeks, and by sailing vessel 90 to 100 days; and as on arrival in the State the cement was received into storage-sheds where it lay at least 1 month, but generally for a longer period, during which time tests were made preliminary to its despatch for use, the cement had some chance of losing any "freshness" which it might have had when first placed in casks, and needed comparatively little slaking. A cement which did not demand much slaking before use was especially necessary in connection with the smaller scattered works of the scheme, distributed as they were over 350 miles, and mostly in country whose dry atmosphere would not tend to satisfactory, or, at any rate, speedy, slaking. In these smaller works, the cement, having passed the necessary tests, was used direct from the casks, because to slake and then repack it would have entailed incommensurate expenditure; but at the weir provision was made for slaking fully all cement requiring it.

The tests, which were of an exhaustive character,1 were directed not only towards determining the qualities of each batch of cement, but also to so ascertaining those qualities that after slight treatment in the State parcels which seemed at first to be doubtful might be used without hesitation. Situated, as the works were, at such a distance from the source of supply, this was essen

1 In the year 1902 alone over 9,000 briquettes were made, not only for immediate use, but also to be broken for comparison, 3 months and 6 months and 1, 2, and 3 years after making.

[THE INST. C.E. VOL. CLXII.]

F

tial. Taken altogether, the cements received were very satisfactory, and as the long-date tests become due, and the samples are examined and the briquettes broken, the results confirm the good qualities adjudged after the shorter tests. In all, ninety-two complete analyses were made, of which those in Table IX. (p. 101) may be accepted as average results. The specific gravity varied between 3.05 and 3.13. On the whole, the cement used was exceptionally well ground, that received towards the completion of the works being even finer than the earlier consignments. After the works were begun, a special set of bulk tests was carried out. Several casks of the different brands were sampled, and 25 lbs. of each brand was carefully passed through sieves with a mesh of 14,400 holes per square inch. The residues

The tests of tensile strength ranged from the usual 3-day and 7-day hot and cold-bath to 28-day cold-bath tests, a reserve of briquettes being often retained from the various batches for longdate tests. As a rule the results were very good, even when, fresh from the cask, the cement was subjected to the hot-water treatment. The hot bath was utilized to determine the necessity for slaking, it being found that a cement which showed a falling-off from the cold-bath results, when treated for a similar period in hot water, generally headed the cold-bath records after being fairly slaked and there are numerous series of tests showing satisfactory increase in tensile strength at various dates up to 12 months. Cements, however, which showed a tendency to fall off in strength in hot water had to give undeniably good results after the requisite periods of slaking, before being despatched to the works.

One feature in the relative rise in temperature on hydration of slaked and unslaked cement is worthy of notice. On several occasions samples taken straight from the casks showed a comparatively small rise in temperature, yet the same cement, after exposure to the air, registered a considerable rise. For the purposes of closer examination a long series of experiments was made with the same shipment of cement air-slaked under three different conditions, namely:

(1) Under the corrugated-iron roof of a shed.

(2) Under the wooden floor of the same building, spread in thin layers on a large tarpaulin and turned over daily; and

(3) Under the same floor, but placed in a box covered with wet bags, passages for currents of air being left between the cement and

The maximum

the bags, which were wetted and turned daily. rise in temperature is shown in Fig. 10, Plate 2. It is difficult to assign reasons for these results, but Table X. in the Appendix, obtained from long-date tests of this same consignment, shows satisfactory increase of strength.

The effect of slaking on subsequent expansion was also the subject of a long series of tests. Ordinary glass test-tubes, 6 inches by 1 inch, were filled with stiff grout, and placed in cold-water baths after the cement had set. The tubes usually cracked after 3 days or more when filled with fresh cement, showing a high rise of

temperature on hydration; but as slaking proceeded, so did the energy of the cement decrease. Further, there was apparently renewed activity after months of quiescence, tubes being cracked although the cement core itself remained hard and sound and, so far as the eye could detect, absolutely unharmed.

The receiving- and slaking-shed for the cement (Fig. 11) was built at the temporary end of the tram-line, that is, adjacent to its crossing of the weir-site, the upper floor of the shed being at the same level as that of the railway-wagons. The storagecapacity of the shed was 2,000 casks, and one-half of this quantity was taken by the slaking-tables, arranged on a falling gradient

so that one tipped on to the next and so on, until the cement arrived at a shoot leading on to the trucks, which conveyed it to the mixing-shed. A large portion of the cement, however, did not require special air-slaking. This, after having been put through a fine-meshed sieve, passed down a shoot to the concretemixer without being placed on the slaking-tables. Below the cement-shed, on the same level as the quarry-road, the stonecrushing plant was erected. It consisted of a No. 4 Gates crusher capable of dealing with 20 cubic yards of granite per hour, and driven by a 25-HP. Robey engine. In the same shed, but below the crusher, was situated the concrete-mixer, portable and selfcontained, of the rotating-barrel type, mechanically fed with cement, stone and sand in the proper proportions, and capable of mixing 20 cubic yards of concrete per hour. All this machinery, from the cement-shed to the delivery end of the concrete-mixer barrel, was designed so that every operation which could be effected with advantage, or could be helped, by gravity, was so arranged; and the whole proved very satisfactory in working.

Except about 1,000 cubic yards of sharp, coarse-grained sand obtained from the river about 1 mile below the weir-site, the whole of the sand was brought from either Lion Mill or Bayswater, distant 8 miles and 22 miles respectively, by railway. That from the former was of quartz, and very fine-grained, yielding even and good results in the testing-room. The quarry, however, required heavy stripping of mould, and the sand itself required screening and thorough washing, to cleanse it from vegetable and earthy matters. This entailed the erection of a sand-washing plant. The sand from Bayswater was of a much better class, and required but light washing to free it from all earthy material.

About 30,000 cubic yards of spalls, for crushing to concrete size, were selected from the material obtained in the excavation of the foundations. For the plums and the balance of the spalls required, a quarry was opened on the north bank of the stream, below the weir, and about 70 feet above river-bed.

The weir and all accessories were built of concrete, but in the former, large rough granite blocks, just as quarried, were introduced into the concrete. It was originally intended to build the wall with 50 per cent. of these large blocks; but without proper plant, which was not available, handling would have been very expensive. The concrete consisted of 5 parts by measurement of granite crushed to 24-inch gauge, 2 parts of cleaned sand and 1 part of Portland cement. So long as the wall remained below the

level of the mixing-house, the mixture was conveyed to the work on an endless conveyor working in a trough, with travelling boards secured by ropes and spaced 2 feet apart, thus ensuring that the heavier aggregate was not separated from the matrix on the way. Later, the concrete was conveyed on a trolley-line in skips, to a large derrick-crane, which lifted the skips on to temporary tram-lines on the growing wall: they were then pushed by hand to a travelling steam-crane which lifted each skip in a bridle, overturned, emptied, and returned it to its carriage. The concrete was spread and rammed by hand, the various layers being broken up so far as the width of wall would allow, in order to break bond in both beds and joints; and in addition, the large rough blocks, up to 2 cubic yards in volume, were deposited and thoroughly bedded and grouted, in order to key the bedding-planes together.

For the first 10 feet the batter was lined and the concrete retained by rubble masonry, but above this level wooden framing was substituted. This framing was of Oregon pine, and consisted of uprights 9 inches by 3 inches, and 15 feet long, cut to the sweep of the wall section, spaced 3 feet apart and closely lined on the wall face with 12-inch by 4-inch Oregon boards. For the first few feet upwards, the studs were held in position by shores, but later they were bored for 1-inch bolts about 18 inches long, at vertical intervals of 3 feet. Each bolt was fitted with an 8-inch by 3-inch by 3-inch iron screwed washer-plate, which was built into the concrete, and remained there after the bolts were withdrawn and the holes grouted. Each vertical stud was lapjointed and bolted to the succeeding one, the lap being sufficient to allow of two bolt-holds in the concrete before the lower boards were removed. No cross stays or ties were used across the wall, and the front and back framings were independent. The uprights were aligned throughout with a theodolite, the heads of each section being cut off to the required level and fixed to the width of wall corresponding with that level, with an allowance for outward pressure of the wet concrete. Rendering of the face was not desired or found necessary, as great care was taken, when depositing the wet concrete in contact with the moulding-boards, to keep all stone well back with straight spades, and a good finished face resulted on stripping. The valve-tower and viaduct, which were carried up with the main wall, were similarly built between moulding-boards, the frames inside and out being formed of upright studs, cross silled and lagged with 4-inch by 13-inch tongued and grooved Oregon boards fixed vertically.