wrought iron breaks suddenly a crystalline fracture is the invariable result. A truer test is a slowly applied breaking weight, which should cause a fibrous fracture. Bad iron is never fibrous. Specification for Plates and Bars. "Every plate and bar must be sound, straight, and free from all flaws, and any piece which shows signs of lamination or other defect will be rejected. The edges of all plates are to be planed so that they may bear truly at their joints. All joggles are to be thoroughly well and neatly formed. The butting ends of all ties, angles, and bars are to bear fairly and firmly throughout, and all corners and edges to be neatly finished off. Every piece is to be of the full thickness specified, to be tested by gauging, weighing, or otherwise." Working Strength.-The following table gives the amount of stress. generally permissible, in tons, per square inch of sectional area :— These figures are based on a factor of safety of 4. The Board of Trade has fixed the limit of stress for bridges of wrought iron at 5 tons per square inch, and of steel bridges at 61⁄2 tons. The strength of steel depends on the precise nature of its composition, and the values given above are merely approximate and general. Tests. Cast iron is usually specified to be tested as follows:-. -A sample bar is cast, 3 feet 6 inches long, 2 inches deep, and 1 inch wide. It is supported on bearings 3 feet apart, and loaded at the centre with a weight variously stated at from 25 to 30 cwts., which it is required to sustain without fracture and without exhibiting a deflection greater than inch. Test bars should, if possible, be cut from the casting, but in any case should be cast under exactly the same conditions. A tensile test is rarely required. 5 Wrought iron is generally required to stand a minimum tensile stress before breaking, the contraction of area at fracture not being less than a * These values only apply in the case of short struts. When the length is considerable, failure is more likely to take place through flexure, and special calculations are necessary for determining the nature and extent of the stress. The problem is dealt with in Chapter ix. certain amount. According to the quality desired the following figures are given: In addition to this, certain forge tests are required. Thus, 1-inch plates for the Admiralty are to be capable of bending without fracture while hot from 90° to 125° along the grain and from 60° to 90° across the grain, and while cold, 10° to 15° along the grain and 5° across the grain. For 4-inch plates the cold tests are 55° to 70° and 20° to 30° respectively. Steel, according to Admiralty requirements, must have an ultimate tensile strength of between 26 and 30 tons per square inch, combined with an elongation of 20 per cent. in a length of 8 inches. Lloyd's specification raises the limits to between 27 and 31 tons with the same elongation. Both tests apply, indifferently, along or across the grain. As regards temper, strips cut from a plate heated to a low cherry-red and cooled in water at 82° F. must stand bending round a curve of which the diameter is 1 or 3 times the thickness of the plate, according as the authority is Lloyd's or the Admiralty. Rivets, if of wrought iron, should be capable of being bent double, cold, without sign of fracture. When hot they should stand being hammered down to less than inch in thickness without cracking at the edge. If of steel they should have an elongation of 25 per cent., with 26 to 28 tons per square inch tensile strength, in test pieces of ten diameters, and should be capable of bending double after the same tempering as that applied to steel plates. Weight of Iron and Steel.-Plates of metal, 12 inches square and 1 inch in thickness, weigh 371, 40, and 403 lbs. respectively for cast iron, wrought iron, and steel. Corrosion of Iron and Steel. It is to be regretted that on a point of such vital importance to the dock engineer as the durability of metal structures exposed to atmospheric and aqueous agencies, the evidence is so scanty as to be inconsiderable, so incomplete as to be inconclusive, and so conflicting as to be actually perplexing. This state of things arises from a variety of causes. In the first place, it is only within the last fifty years. that iron has begun to usurp the pre-eminence hitherto enjoyed by wood and stone in maritime construction, and steel is an intrusion of still later date. Consequently there has hardly yet been sufficient time in which to acquire data for the determination of the actual life of metallic structures CORROSION OF IRON AND STEEL. 141 under such conditions, even if systematic experiments had been carried out from the earliest possible moment, which has not been the case. Again the variation in atmospheric conditions is extremely great, the seasons being marked by enormous fluctuations in sunshine, rainfall, and temperature not only for different seasons in the same year, but for the same season in consecutive years. The question is still further complicated by the factor of locality. Then, as regards aqueous influences, there is no definite standard of comparison whatever. The salinity, acidity, density, and temperature differ in almost every unit volume of sea-water, so that it is never precisely the same at any two ports. Rivers, sewers, and ocean currents all contribute to differentiate its composition. It would, perhaps, be a comparatively easy solution of the difficulty to lay down one's individual experience as a dogma for general acceptance, but the wiser and more judicious course will be to set forth such information on the subject as is available, and leave the reader to draw his own conclusions. The following coefficients, given by Thwaite and quoted by Molesworth,* represent the amount of corrosion in lbs. per square foot of surface during twelve months' exposure: If the metal be painted once a year the coefficient to be divided by 2; if once in two years, by 18; and if once in three years, by 1·6. Trautwine states, in apparent contradiction of the above, that while "fresh-water corrodes wrought iron more rapidly than cast, the reverse appears to be the case with sea-water," and that "the corrosion of iron or steel by sea-water increases with the carbon." He admits, however, that * Pocket-book of Engineering Formula, 25th edition, p. 33. + Civil Engineers' Pocket-book, 17th edition, p. 218. wrought iron is affected very quickly, so that thick flakes may be detached from it with ease. The following instances are cited :-"Cast-iron cannons from a vessel which had been sunk in the fresh-water of the Delaware river for more than 40 years, were perfectly free from rust." The cast-iron work of the "Royal George" and the "Edgar," sunk in the sea for 62 years and 133 years respectively, when examined by Gen. Pasley had become quite soft and resembled plumbago. The wrought iron was not so much injured, except when in contact with copper, or brass gun-metal. Two other experimentalists-Rennie and Mallet-adopt antithetical opinions as to the relative corrosion of cast and wrought iron in salt-water. The former maintains a higher rate for cast iron; the latter, for wrought iron. The following table extracted from a paper on the corrosion of iron and steel, by Mr. David Phillips,* relates to a series of experiments made by him with five sets of iron and steel plates, 4 inches square by inch thick, exposed to various corrosive agencies. "To avoid even a suspicion that galvanic action had any influence in these cases, all the plates were suspended on glass rods, and each plate was separated from its neighbour by glass ferrules." It is important to note that Mr. Phillips attributed the generally greater corrosion during the first period of trial to the fact that the weather in the summer of 1879 was much more changeable than that in 1880. In the discussion which followed the reading of the paper, much emphasis was laid by Dr. Siemens, Mr. Barnaby, Mr. Farquharson, and others, on the importance of removing the magnetic oxide scale from the *Phillips on "The Comparative Endurance of Iron and Steel when Exposed to Corrosive Influences," Min. Proc. Inst. C.E., vol. lxv. CORROSION OF IRON AND STEEL. 143 surface of steel, and this received the confirmation of Sir W. H. White, at a later meeting of the institution, when he declared that "as regards the relative corrosion of iron and steel when immersed in sea-water, the experience of the Admiralty during the last six years (1876-1882) showed that if the manufacturers' scale (black oxide) was thoroughly removed, and equal care taken in protecting the surfaces by paint or composition, iron and steel had about the same average rate of corrosion, the steel wearing somewhat more uniformly than the iron."* The question of corrosion principally concerns the dock engineer in regard to the duration and maintenance of metal gates and fittings. Decay mainly takes place below the water-line, where inspection and repairs are alike difficult. In this connection the following data taken from a report by Messrs. Brandt and Hotopp to the Ninth International Navigation Congress possess much interest : "I. In the case of the floodgates at Glückstadt, erected in 1874 and to be renewed this year (1902), the first isolated rust spots on the outer skin are to be found at 4 inches below ordinary low water level; the spots increase in number at 6 inches below low water, and are thickly distributed all over the metal at a depth of 10 inches. The greatest depth to which decay has penetrated in the strip comprised between this line and another, lying about 3 feet 3 inches below low water, is about 4-inch; below this level the metal skin is covered with a layer of short-stalked moss, mixed with shells, the thickness of which increases downwards, and below which the depth and extent of decay grows gradually less and less (to about 1-inch deep near the sill), so that the plates near the sill are almost sound. A few of the rivet heads, starting at a depth of 14 inches below low water, begin to show signs of decay and are furrowed; the decay gradually increases with the depth, so that when the rows of rivets, situated between 18 and 22 inches below low water, are reached, not only have all their heads been completely eaten off, but their shanks have also been already attacked in isolated cases. The decay in this case also becomes less and less with increased depth. The water of the River Elbe, at Glückstadt, is only on exceptional occasions somewhat brackish, but in the outer harbour there is a great deal of deposit, and several drains full of water from the moors empty into it. 3 10 "II. The gates, and more especially the floodgates, in the harbour at Geestemünde, erected in 1861, show a furrow, the rust in places penetrating as deep as inch into the outer metal skin, just above the cover strips lying close below low-water line, and it may be assumed that similar rusty places exist also above the cover strips in lower situations, the upper portions of the outside rivet heads lying close under low water mark have also rusted away. The cause to which this damage is ascribed is the layer * Min. Proc. Inst. C. E., vol. lxix., p. 35. + Brandt and Hotopp on "Iron, Steel, and Wooden Gates," Int. Nav. Cong., Düsseldorf, 1902. |