most gigantic arches ought to be far easier to build in this material. The one element of uncertainty is the contraction and expansion of the metal from heat; but there seems little cause to fear it. When we first made railroads we allowed a quarter of an inch free space between each bar, and took every precaution for freedom of expansion and contraction till one man, bolder than the rest, proposed to butt them one against the other and join them with fish-plates. This has now been done, so that the rail from London to Aberdeen is one continuous unbroken bar; it neither expands nor contracts, but submits, and so probably would a bridge, provided the abutments were sufficiently firm, or if it did expand, it would probably be marked only by a slight elevation at the crown of the arches.

Before, however, engineers had proceeded far in the application of iron to bridges, they perceived that though the metal possessed the quality of resisting compression to ten times the extent of the materials they had usually been employing, it was even more remarkable for its tenacity; nor were they long in finding out how best to avail themselves of this peculiarity. By suspending the roadway from a chain hanging from the summits of two tall towers, they in the first place got wholly rid of the bugbear of expansion or contraction, and were also able to span a greater space with an infinitely smaller quantity of metal than was required for a bridge in compression. So great, indeed, was the economy of weight, that there seemed no practical limit to the extent of the span, while all other structures were liable to be broken by their own weight when extended beyond certain moderate dimensions.

Unfortunately these good qualities were accompanied by others which disappointed the sanguine hopes that were at one time entertained of this mode of construction. Its very lightness rendered it liable to undulation, always unpleasant and sometimes dangerous; and its weight was frequently not even sufficient to resist the action of the wind, which ruined at one time the chain pier at Brighton, and seriously damaged the bridge over the Menai Straits, as well as that at Montrose. Notwithstanding this, Telford's great work has answered its purpose perfectly for the last thirty-seven years, and now that it has been strengthened, may still span the Straits for the next three centuries; while, considering the time when it was erected, it is one of the boldest as well as one of the most graceful works of modern engineering skill.

On the Continent, where scientific knowledge is generally in advance of practical skill, they have carried this principle to excess, by using wire, which is iron in its most perfect form for

tenacity.

tenacity. This has reduced the weight of the bridge so much relatively to the load, as to render the undulation excessive, and frequently to lead to the most frightful accidents. Still the bridge over the Sarine at Friburg has stood for thirty years, with very slight repairs, though its span is 870 feet, while that of the Menai Strait is only 570, and the bridge which recently crossed the Thames at Hungerford Market, which was our largest and typical example of the class in England, was only 6761.

The boldest and grandest application of this principle is the bridge constructed for railway traffic by Mr. Roebling, just below the Falls of Niagara. So rapid has been the progress of engineering science, that if any one had proposed twenty years ago to throw a railway bridge over a chasm 800 feet wide and 245 above such a foaming torrent as that of the Niagara, he would have been looked on as a madman. Yet this has now

been accomplished, and by very simple means. The bridge consists of a rectangular tube 20 feet deep by 26 feet wide, or rather two floors 18 feet apart-the upper carrying the railway, the lower the roadway for ordinary traffic. These are connected together by a series of wooden posts, braced together by diagonal iron tie-rods. By bracketing out from the rocks, the free length of the tube is reduced to 700 feet, and it is then suspended from towers 821 feet apart from centre to centre by four wire cables of 10 inches section, and each containing 3640 separate wires. These are further assisted by numerous braces radiating from the towers, and a multitude of ingenious minor contrivances.

When a train weighing more than 300 tons passes over the bridge, the deflection is said to be only 10 inches; and certain it is that so far it has answered all the purposes for which it was intended, but nevertheless it seems too frail and fairy-like a structure for the rough usage of railway traffic; and trains are not allowed to move across it at a higher velocity than a man can walk. With great care and continuous repairs it may do its work for years to come, but it may any day deposit its load in the boiling flood beneath, and so again separate the provinces it has so boldly united. Indeed, taking it altogether, there can be little doubt that the tubular girder proposed by Robert Stephenson for the same purpose would have been a better piece of engineering. It would have cost more in the first instance, for if the published accounts are to be believed, the suspension bridge cost only 1007. per foot forward; but the durability of the tube would have been practically unlimited, its safety undoubted, and an occasional coat of paint all the repair it would have required.

Fortunately for the engineers it is their privilege to be allowed to think. They are not like the architects, first forced to inquire whether

whether or not a thing was done in the fifth century before, or the thirteenth century after Christ, before they are allowed to act, and the progress of improvement in iron bridge building has consequently been rapid. For certain purposes a cast-iron bridge, wholly in compression, was no doubt a very perfect thing, so also was a wrought-iron bridge wholly in tension; but it was easy to predict that the most perfect result would be attained by a structure which should combine these two properties, so as to take the greatest possible advantage of both.

The best method of effecting this was fully investigated, and practically settled, by the very complete and exhaustive set of experiments which were undertaken by Robert Stephenson and his associates before commencing his great work, the Britannia Bridge. The conclusions then arrived at were so sound and satisfactory that it is scarcely probable any extensive railway structures will in future be carried out on any other principle, though for local traffic simple compression or tension structures may still be used.

Although the principles then evolved are now thoroughly understood by every engineer, they are so novel and so little appreciated by the general reader, that it may be worth while to try to explain what they are before proceeding further.

1

In the above diagram the left-hand side represents the usual form of a cast-iron bridge, supported by abutments, in the same manner as stone bridges are; and its stability of course depends wholly on their immovability. Instead of this, let us suppose that the ends of the arch rested on iron shoes, as at A, and that these were tied together by a chain or bar of iron B: it is evident that by this expedient the arch would be prevented from spreading as well as by the abutments. It will also have this further advantage, that, as the tie expands equally with the arch, and the structure is one homogeneous whole, with only a per

pendicular

pendicular bearing in its supports, you have a better bridge than before.

It is remarkable that the Italian architects in the Middle Ages tried this principle in all their Gothic structures; but an iron tie to a stone arch is both mechanically and artistically a mistake. The expansion and contraction of the metal is always working when the stone is at rest, and the flimsiness of the one compared with the mass of the other always produces an effect so disagreeable that the true Gothic architects on this side of the Alps never adopted it. They always applied a stone abutment to a stone arch, which was as essentially the proper and legitimate mode of construction as the iron tie to the iron arch is now seen to be.

This principle of construction, once seized, was used in fifty different forms. One of the most obvious was to frame the arch and the tie strongly together, as shown in the right-hand side of the diagram, making what is called the bow-and-string bridge, and to run the roadway along the tie, in which form it has been extensively employed in railway structures. At the High Level Bridge at Newcastle the spandrils are filled up level (as on the left of the diagram), and the railway runs along the top, the roadway along the string. At Saltash and Chepstow, Brunel substituted a bent wrought-iron tube for the cast-iron arch, and tied the ends together by a chain drooping in the centre, and suspended his roadway from both. At Mayence, Dr. Pauli improved on this by substituting a wrought-iron T girder for the tube, and proportioning all the other parts more scientifically together, so as to produce what is theoretically perhaps the most perfect truss yet executed. The three spans of the German bridge are only 333 feet each, while the span of the two at Saltash is exactly 100 feet more; but the proportion of the parts is so perfect, that the principle admits of extension up to the limit at which a girder would tend to break itself by its own weight.

The defect of these bridges is that they are a little too clever. If the load were always evenly disposed over their whole surface, and at rest, no doubt every cubic inch of iron would always be doing all its work; but a railway train weighing 400 or 500 tons, and rushing at a speed of forty miles an hour, is a sad disturber of equilibriums: every part that ought to be in tension is at times thrown into compression, and every strut at times becomes a tie, so that engineers generally have agreed to adopt a plain straight girder instead of those with these beautifully calculated curves. The same thing occurred with rails in the infancy of the system. Every mechanic saw, and every mathematician calculated, that a fish-bellied rail must be stronger than a straight one; but the practical result is, that all rails are now made with

parallel

parallel sides, and there is not one of the other class in existence on any locomotive railway in Europe. It will probably be the same with bridges when the true conditions of the problem come to be more perfectly appreciated, except, perhaps, in structures of such magnitude that the weight of the load bears a very small proportion to the weight of the girder, and where the saving of every ton of iron becomes of importance lest the weight of the bridge should itself become a source of weakness.

Barring such exceptional cases, engineers are generally agreed in making the top and bottom flanges of their girder bridges practically parallel to one another, and when these are of wroughtiron, in putting the same quantity of metal into both. According to strict calculation, the proportions between the top and bottom ought to be as six to five; but as the lower or tension part depends wholly on its rivets, and the top or compression piece might almost be stuck together with glue, the same amount of metal is practically required for both, and the form in which it is disposed is mechanically immaterial. The cellular system has some convenience, but it does not seem to give any strength proportioned to the additional cost and difficulty of construction.

One of the most obvious ways of applying these principles is by means of what is called the Warren girder. This consists of a series of straight cast-iron tubes above, butting one against the other, and a chain of wrought-iron links below, and then connecting these two systems by struts and ties placed diagonally where wanted. Theoretically nothing can be more perfect than this arrangement: it is simple, but almost too simple; if one thing goes wrong all goes wrong; and more margin is wanted. for the violent irregularities of railway traffic.

Perhaps, after all, there is nothing better than the simple tubular girder, which was evolved out of the first experiments, and used with such success in carrying the Holyhead Railway across the Menai Straits. The first and most obvious proposal for this bridge was one of cast-iron in compression, which would have been the cheapest and most architectural mode of effecting the object, but the Admiralty interfered, and insisted that a clear headway of 100 feet above high water should be maintained throughout. To meet this difficulty a tube suspended by chains was then suggested, nearly similar in principle to the one recently erected at Niagara; but as the investigation proceeded, it was found that the chains might be dispensed with, if a tube of sufficient rigidity could be constructed to carry any railway train across the greatest opening, which here was 460 feet clear. So complete were the investigations, and so careful the execution of the whole work, that subsequent experience has added little to the knowledge then attained; and, besides being the first, it is, considering