III
Wepass on to some important topics that worry a writer because they cannot be arranged in a neat scheme. Some of them are technical, but everybody will be able to understand their bearing on the main subject. We have seen that fords gave place to bridges very slowly, even in some neighbourhoods where the Church was exceedingly active, as at Rochester.[105]Can you explain why? There were a good many reasons, and among them is the fact that it was a long time before bridges won a good reputation among the people. Wood being abundant everywhere, they were timber bridges at first, and rudely built; many of them were carried away by storms, as Matthew Paris related in the thirteenth century. So people set their hearts on the greater safety of stone bridges; but money was difficult to collect, and stonework cost a great deal more than timber; and no bridge could be built until permission had been gained from the King, often after tedious negotiations. Further, the lands through which rivers flowed were owned at times by rival noblemen,who put a veto on the project, either in a spirit of perverse antagonism or because a stone bridge might benefit one landlord more than another. And it was easy for the stronger man to explain his antagonism in a reasonable manner, for he could say that the cofferdams used in grounding piers diverted rivers from their channels, causing inundations. This objection seems to have been raised pretty often, as many piers were grounded in a very primitive fashion, just by throwing down stones and cement till a bed of masonry rose above water-level.
In the Ballad of Abingdon Bridge, written by Richard Fannande Iremonger in the thirty-sixth year of HenryVI(1458), we find most of the difficulties that attended mediæval bridge-building. Till the fourth year of HenryV(1417) the townsfolk of Abingdon and Culham had nothing but a ford, which could not be passed after a storm of rain or after a thaw. Yet Abingdon lived under the shadow of a great monastery, and roads were constructed from her streets to the ancient or Roman highways. Not even a timber bridge preceded the charming stone one that charity built in 1417, the very year in which HenryVsailed from England with 16,000 men and ravaged Normandy. But in the Middle Ages most people regarded bridges as we in our ignorance regard hospitals, as useful and necessary things to be supported by charitable doles, and not by district rates. To beg is a degradation, no matter what the cause may be, and many a small town could have built for itself a bridge but for the ruling customthat taught it to be a mendicant. Culham and Abingdon waited a very long time before almsgiving got rid of their dangerous ford. The Abbot gave his aid, and Geoffrey Barber paid a thousand marks to the workmen, and Sir Peris Besillis, Knight, provided the stone, and “the gode lorde of Abendon left of his londe, for the breed [breadth] of the bridge, twenty-four fotelarge”:—
It was a greet socour of erthe and of sonde,And yet he abated the rent of the barge.An C. Pownde, and xvˡⁱ was truly payedBy the hondes of John Huchyns and Banbery also,For the waye and the barge, thus it must be sayed.
It was a greet socour of erthe and of sonde,And yet he abated the rent of the barge.An C. Pownde, and xvˡⁱ was truly payedBy the hondes of John Huchyns and Banbery also,For the waye and the barge, thus it must be sayed.
It was a greet socour of erthe and of sonde,
And yet he abated the rent of the barge.
An C. Pownde, and xvˡⁱ was truly payed
By the hondes of John Huchyns and Banbery also,
For the waye and the barge, thus it must be sayed.
But I am happy to add that “the Commons of Abendon” had to do something for themselves. It was “set all in one assent that all the brekynges of the brige the town bere schulde.” In other words, charity had produced a free town bridge, leaving the inhabitants to pay for its upkeep.
During the building of this pretty structure an unsuccessful attempt was made to ground the piers while eleven men baled water from the river. Then a dam was built, and trenches were dug to prevent the water from overflowing the dam. This I gather from the ballad, but the wording is not at all graphic in any technical matter.[106]We are not told why cofferdams[107]were not tried. In the Middle Ages cofferdams were known as brandryths or brandereths; by this name they are mentioned in the Contract Deed for the building of Catterick Bridge over the Swale,A.D.1421; and they were large enough to obstruct most rivers, for they had to surround enormous piers, and the thickness of their sides was never less than from four to six feet. It is interesting to note, in this connection, that during the construction of Old London Bridge, between 1176 and 1209, the Thames “was turned another way about by a trench,” which, according to Stow, began east near “Rotherhithe, as is supposed, and ended in the west about Patricksey, now termed Battersea.” In those days no embankments controlled the Thames at London; wide shores, littered with the odds and ends of a waterside life, were playgrounds for the ebb and flow of the tidal waters; and the main purpose of the “trench” or canal was to lessen the risk of floods while the huge piers were being founded. Stow’s words give us to understand thatallthewater in the Thames “was turned another way about”; a very important feat of civil engineering. Perhaps the purpose of the canal was not so thorough; perhaps it drew from the river sufficient water to lower its normal level by several feet and to diminish the force of the tidal current. In any case, however, Stow’s evidence has great interest.
One of Brangwyn’s animated drawings, the Pont des Consuls at Montauban, comes in here to illustrate the many troubles of mediæval bridge-building. In 1144, when Montauban passed from an unknown village into a known town, its patron or founder, Alphonse Jourdain, Count of Toulouse, commanded that a bridge should be made at once, and that the little township should keep it in repair; but, somehow, for many generations, nothing was done. Sometimes poverty was pleaded as an excuse, and sometimes the Albigeois wars were blamed; but at last, in 1264, the good men of Montauban ventured on a little action. Indeed, they stretched themselves yawningly, and said that a bridge over the Tarn would be a boon indeed. Their ferry was a slow nuisance, we may presume, and their trade ought to be increased by better communications. For twenty-seven years they repeated these truisms; then, in 1291, they bought the island of Castillons or of Pissotte to serve as a foundation for several piers. Tired by this unwonted exertion, Montauban wished to take a long holiday, but Philip the Fair came forward and asserted himself as a king. A bridge over the Tarn must be built! It should have three fortified towers, one at each end, the other inthe middle; and these towers were to be garrisoned by royal troops, so that no harm should happen to the king’s authority. In order to collect money for the bridge-building a tax was to be levied on all visitors to Montauban, and two consuls were to overseer the work. His Majesty chose Mathieu de Verdun, a citizen, and Étienne de Ferrières, who was keeper of the town. They seemed to be honest men, but funds collected for the bridge were used for other purposes, and I know not if this action was justified. It was in 1304 that Philip the Fair gave his instructions, and the bridge was not finished till 1335. Still, the dilatory township had achieved a very fine work of art, noble in design and very well constructed.
It is a brick bridge, 250 m.[108]50 cm. in length. The bricks are excellent in quality, and measure 50 centimetres in thickness, 40 centimetres in length, and 28 centimetres in width. The roadway is nearly flat, and its height above the level of the Tarn is 18 metres. There are seven pointed arches with an average span of 22 metres; and the six piers armed with cutwaters at both sides are 8 m. 55 cm. in thickness. Note how the spandrils are pierced with high arched bays to facilitate the passage of water during floods. These relief arches were copied from Roman models. As for the defensive towers they exist no longer, but the strongest one kept watch and ward over the entrance across the river; it was square in shape, and its summit was a crenellated platform fringed with machicolations. The other end tower—theone on the town side—was also square in form, while the central defence was triangular. It stood on the middle buttress on the side looking downstream, and the lower part of it was used as a chapel dedicated toSt.Catherine. A flight of winding steps went down to a postern, cut through the buttress a little above water-level; and at the other side of the pier, just below the arched bay, was an instrument of torture, a see-saw that carried an iron cage in which blasphemers were ducked in the river.
The Pont des Consuls has one quality that Englishmen ought to study with the greatest care; it is in scale with a great river. To build a vast bridge for a little township was in part a just tribute to the beauty of a noble site, and in part a prophetic compliment paid to the future history of Montauban. How differently we have acted in our London bridges! We have disgraced the Thames with the Railway Viaduct from Charing Cross, for instance, and neither Waterloo nor London Bridge does justice to the size of our Nation-City. There are three or four good bridges on the Thames, notably those at Maidenhead and Richmond, but they are nothing more than delicate works of refined engineering. Not one is inspired by awe, the only feeling that can bring home to our minds the wondrous grey antiquity of the Thames and the immensity of London. So we have feared to be great in the historic symbolism of bridge-building, unlike the citizens of Montauban, who were lifted far above their indolence by a brave inspiration as ample as was the Tarn after a flood.
LE PONT DES CONSULSLE PONT DES CONSULS OVER THE TARN AT MONTAUBAN IN FRANCE, FOURTEENTH CENTURY
LE PONT DES CONSULS OVER THE TARN AT MONTAUBAN IN FRANCE, FOURTEENTH CENTURY
In 1823-4, when George Rennie designed New London Bridge, London was probably two hundred times as big as was Montauban in the fourteenth century; and certainly the Thames was not inferior to the Tarn as a historic inspiration. Yet Rennie failed to understand the importance of being large in scale. In less than fifty years his work was “insufficiently wide for the traffic”;[109]and since then, on a good many occasions, we have been asked to disfigure London Bridge with overhanging footpaths. “London can well afford to pay for new bridges, but can by no means afford to part with a single object of real beauty.”[109]For Rennie’s bridge, despite all errors of scale, has points of charming interest. Her roadway has a graceful curvature that delights the eye, her arches have an excellent shape, and the variation in their size could not well be bettered.[110]Later we shall see (p.325) that much money was ill-spent on hammer-dressing the whole external face of the masonry; but an engineer with a very weak feeling for scale was afraid to use either scabbled stone or stone with a rough-axed facing. Rennie learnt all that he could learn by studying fine models of style, such as the Roman bridge at Rimini, but his own equipment as an artist was terrene.
Would that we had in England an old bridge equal tothe Pont des Consuls! Would that old London Bridge had been delivered down to our sixpences and shillings! Yet I suppose we must consider ourselves lucky in the fact that historic bridges in Great Britain, though much inferior to those on the Continent, are fairly numerous in districts where there has been but little increase of traffic. We possess three bridges with defensive gateways (Stirling, Warkworth, and the Monnow Bridge at Monmouth); five with chapels, or with relics of chapels (St.Ives in Huntingdonshire, Derby, Bradford-on-Avon in Wiltshire, Wakefield, Rotherham); and many good specimens exist of bridges with angular recesses built out from parapets and forming part of the piers.[111]These recesses were designed not only as shelter places for wayfarers, but because they lessened the cost of production, inasmuch as they gave width to narrow footways; and so their value in an old bridge is very similar to that of bay-windows in cottage rooms.
Very often the modern engineer has misunderstood their origin, and, regarding them as decorations, he has used safety recesses to ornament his wide bridges, just as he has put battlements on iron parapets and stuck machicolations on defenceless gateways. Brangwyn has drawn for us three or four big Gothic bridges with safety recesses. Among them is a fine structure over the Main at Würzburg,in Bavaria; there are eight arches, and the length is 650ft.This bridge dates from the year 1474, but his adornment with statues of saints belongs to later times. Indeed, the architecture and decoration take us from the end of the Middle Ages to the year 1607, when the spirit of the Renaissance was active and generative.
BRIDGE OVER THE MAINTHE BRIDGE OVER THE MAIN AT WÜRZBURG IN BAVARIA (1474-1607)
THE BRIDGE OVER THE MAIN AT WÜRZBURG IN BAVARIA (1474-1607)
Here is an old defensive bridge that does not resemble a common man-at-arms: in him there is a fine courtesy, as of a knight long used to the etiquette of tournaments; but yet the technical inspiration is rather inferior to that in his greatrival, the Moselle Bridge at Coblentz, built in 1344, by the Elector Baudouin, and charmed with a mellow grace that imparts a rare distinction to the vigour of fourteen bold arches. The Moselle Bridge is 1100ft.long, or ninety-five longer than London Bridge. There is but one fault, and this one fault belongs to the Middle Ages: the ten piers obstruct the river too much, and two or three of them might have been omitted without harm to any strategic consideration.
In the Middle Ages almost everything was looked at from the standpoints of attack and defence. Bridges as well as soldiers needed armour, so their gateways and towers were built in a military fashion, and at times curious traps were devised along the footways. For example, consult the “Pacata Hibernia,” and you will find an engraving of Askeaton Bridge,[112]with a sort of hangman’s trapdoor at each end of the footway. In 1586, or thereabouts, Askeaton Bridge had another peculiarity: a castle stood close to it on an island in the river; and between the castle and the bridge was a fortified platform with two gateways.
It happened often, in mediæval times, that one arch was a drawbridge. Take Old London Bridge as an example. One of her twenty arches—the thirteenth from the City end—was a toll-gate for merchant shipping, and a drawbridge to gap off enemies from the town. It served this latterpurpose in 1553, when Sir Thomas Wyatt and his insurgents tried to enter London. Everybody knew which was the movable arch, because it was connected in all popular talk with the tower that rose beside it, a terrible and gruesome tower, for on its summit executioners displayed the heads of decapitated persons, who ranged from common bandits to the great Sir Thomas More.
Some defensive bridges in Old England had an important look as late as the reign of GeorgeIII. This applies to the Welsh Bridge at Shrewsbury, which had a noble tower at the entrance that looked towards Wales. Perhaps it belonged to the reign of EdwardI, as a statue of Llewellyn was placed over one of the arches. At the present time our fortified bridges are minor specimens. The “auld brig” over the Forth at Stirling, once “the key of the Highlands,” is the most interesting architecturally. He still retains a defensive gateway at each end, and his four arches, now closed to traffic, have a bold and pleasant rhythm. They date from the last years of the fourteenth century. From this century also Warkworth Bridge comes to us; it is a smaller structure, with a triangular recess at each side, projecting from the parapet into the central pier. The gate-tower is at some little distance from the abutment; it has a low and narrow archway under which carters swear unhopefully, believing that their wagons will stick fast. A person who was present on the occasion told M. J. J. Jusserand that a gipsy’s caravan, not long ago, was stopped at the tower on Warkworth Bridge, and waitedthere while the pavement was being hollowed out to make the passage deep enough for a safe journey.
The pier midstream is triangular, and almost as sharp as an arrow-head. This shape is very common in mediæval cutwaters, but it belongs to a technical routine which cannot be regarded as practical. Floods cannot eddy around the flat surfaces of a triangle; they are cut into waves that soon break with an increasing force against the piers and spandrils. On the other hand, when a cutwater is shaped like a Gothic drop arch, or like a tierce-point arch, it meets the current with a much bolder wedge of stone, whose curved sides are better playgrounds for water in spate. Cutwaters of this improved sort are uncommon in mediæval bridges, but some are to be found in French work of the Limousin.
Viollet-le-Duc was the first critic who called attention to this technical matter, and no pontist should fail to note how cutwaters are designed. For example, in a bird’s-eye view of the bridge at Avignon the buttressed piers jut out on each side beyond the narrow footway, looking like boats that support a long line of planks; and I have no doubt that Saint Bénézet had in mind this figure of boats when he planned his roadway over “the arrowy Rhône.” It is far from my wish to compare the little Warkworth Bridge with this French masterpiece, but let us note in its cutwaters a similar character.
Again, when you remember that Warkworth Bridge belongs to the fourteenth century, do you not expect to findin it the pointed vault, whose lighter grace is among the most beautiful things both in Eastern and in mediæval architecture? Yet the two ribbed arches are segments of circles. For many a generation Northern England has been famed for three things—a long-headed thrift, a discontent that is said to be a Radical in politics, and a stubborn hatred for any new knowledge that attacks the dull mimicry of customs. It is to Lancashire, for instance, that you must go if you wish to study in old packhorse bridges the retention of Romanesque forms. A considerable number are described popularly as Roman bridges, probably because they are found on the old pilgrim ways, which, after the Reformation, were scorned as Roman Catholic.
For some reason or other Northern England welcomed in bridges the bluff economy of ribbed arches, while neglecting the more gracious thrift of Early English or pointed vaults. These are easier to build because they need lighter centres or arched scaffolds, and their thrust being less powerful than that of round-headed arches, they require less bulk in the piers. Some writers say that pointed arches interfere with sailing-boats, but this depends on the size of their spans. At Montauban there is room enough for ordinary boat traffic under the Pont des Consuls.
The Pont Valentré at Cahors has ogivale arches, and in one fine drawing Brangwyn studies the technique of their construction. For instance, the embattled piers are triangular, and each of them is pierced transversely by a bay or passage, which is put on a level with the springing of everyarch. Below this bay are three holes; and another line of holes runs across the under surface of the arch beneath the springing.[113]Now, these holes and the bays have a great technical interest, they remind us how the Pont Valentré was built in the thirteenth century. With their help simple scaffolds were erected. The first step was to thrust fir saplings through the holes in a pier till they jutted out on each side; then they were covered with planks and used as footbridges by the workmen, and also as resting-places for barrow-loads of dressed stone, which were lifted up by movable cranes. The service of the masons was effected through the bay in a pier, and the centring of every arch was fixed in those other holes which Brangwyn has represented in his vivacious water-colour.
Not more than two arches were built at the same time. At any moment, in those rude, warfaring periods, work might be interrupted by strife, and its progress was so very slow that it took from ten to thirty years to bring a bridge to completion, usually after a continuous fight against money troubles. Many a hint on economy was borrowed from the Romans, whose enterprise was far in advance of their current cash. Piers that look marvels of solid masonry may be nothing more than shells filled with beaten earth and gravel; and those passages through the piers at Cahors have one thing in common with the relief arches that pierce the spandrils of the Pont des Consuls at Montauban: they enabled the builders to be thrifty.
PONT VALENTRÉPONT VALENTRÉ AT CAHORS-SUR-LOT A FORTIFIED BRIDGE, THIRTEENTH CENTURY
PONT VALENTRÉ AT CAHORS-SUR-LOT A FORTIFIED BRIDGE, THIRTEENTH CENTURY
In a Persian bridge (on the way between Resht, on the Caspian Sea, and Teheran, the capital) thrift hollowed the spandrils into chambers, some of which were used by travellers. This bridge carried a rough highway over the Kâredj River, which runs down from the Elburz Mountains between Kasvîn and Teheran, and disappears in a gravelly plain. In 1874 the Kâredj Bridge was studied in measured drawings by J. Romilly Allen, and eighteen years later (November 19, 1892) the drawings were published in “The Builder,” with a most valuable description. Let us linger for a few minutes over Romilly Allen’s research, as the technique of old Persian bridge-builders has points in common both with Gothic methods and with modern practice also. Some mediæval spandrils are hollow, for example; and a very noted French architect of the eighteenth century, Perronet, not only left empty spaces behind the haunches[114]of an arch, but made tunnels in piers, after the manner used by Pope SextusIVin the Ponte Sisto. And the bridge of Glasgow over the Clyde has tunnelled piers, so this technical detail has a long and entertaining history.
In the Kâredj Bridge, then, the builders had to solve three or four difficulties that strained the usual penury of Persian finance. The river itself must have been a constant trouble while the bridge was being constructed. A rapidmountain torrent with precipitous rocky banks, it pours through a gorge of rock, and at one spot only it forms a good foundation for a wide pier; but this spot has a situation that divides the bridge into inharmonious parts, making symmetry impossible. Allen’s drawing shows both arches, one with a span of 23ft., the other with a span of 72ft.9in.; and between them is a vast pier not less than 31ft.9in.wide. Forty-six feet separate the highest point of the parapet from water-level; and from water-level to the peak of the big pointed vault is thirty-seven feet. In width the bridge measures thirty feet across the outside of its parapets, and twenty-six feet across the roadway, so there is room for a great deal more wheeled traffic than Persia has yet developed along her dusty trade routes.
From this description it is evident that the builders had a stiff job. Timber for centring has ever been scarce in Persia[115]; so in Persian bridge-building the usual plan is to set up a light scaffold just strong enough to bear its own weight and a few rings of brickwork. After a single rib of bricks has been made, other bricks are dabbed against the first set, more being added at the abutment ends than in the centre of an arch; and so, as the work goes on, the arch grows to be self-supporting, like a cantilever bridge. When the middle part of the span has been covered over, the remaining courses at each side are completed with bricks set at right angles to the others. In looking upward at the under surface of a Persian vault a pontist sees that the courses of brick go in two directions, one parallel to thecentral axis of the bridge and another at right angles. Such is the Persian method of building a brick arch; its main object is to evade, without too much risk, the cost of heavy centring, timber being so difficult to get and so expensive to carryabout.[115]
In the four bridges that Romilly Allen studied, between Resht and Teheran, the building was brickwork, and the bricks were rather like Roman tiles; they measured 10in.by 10in.by 2½in.At Kâredj the mortar joints were about ¾in.thick, so that twenty-four courses of bricks with their mortar joints built a wall about 6ft.2in.high. At the thinnest part of the big arch there were only three bricks, giving a thickness of 2ft.6in.; further on there were five bricks, and two more were added at the abutments, where the walls were 7ft.6in.thick. Here is much economy, for thick joints of mortar are not praiseworthy (p.175); and thrift is very noticeable in other details of the workmanship. Beneath the roadway were two chambers with pointed barrel vaults, which were built partly to relieve the haunches of the big arch, and partly to save materials. On one side of the arch the chamber was about 12ft.high; its length, varying with the curve of the voussoirs, and extending across the abutment, ranged from 27ft.to 49ft.On the pier side of the big arch the chamber was not so long, but its height was 12ft.; and the pier itself was chambered in its upper part and pierced below with a Tudor-like arch about 14ft.wide and 11ft.6in.high. The chamber abovehad a cone-shaped roof, and at each side of it were three square-headed windows that measured 3ft.6in.wide by 6ft.high. I am speaking in the past tense, for I know not whether this bridge is still in use; but now we will return to the present tense in a short quotation from RomillyAllen:—
“This chamber appears to be ... a temporary living-room for travellers. It ... communicates with the cells above the haunches of the arch by an opening 4ft.6in.high. The inner room is probably intended to afford sleeping accommodation. The living-room is approached by a staircase in the thickness of the wall leading up from the top of the pier. The Persian name for an upper chamber of this kind is ‘bala-khana,’ literally ‘a house upabove.’”[116]
“This chamber appears to be ... a temporary living-room for travellers. It ... communicates with the cells above the haunches of the arch by an opening 4ft.6in.high. The inner room is probably intended to afford sleeping accommodation. The living-room is approached by a staircase in the thickness of the wall leading up from the top of the pier. The Persian name for an upper chamber of this kind is ‘bala-khana,’ literally ‘a house upabove.’”[116]
Perhaps the finest bridge in Persia is the far-famed Ali Verdi Khan at Isfahan.
Ali Verdi Khan was the general of Shah Abbas, and his bridge, if not the greatest in the world, has no rival that excels it in stateliness. As Lord Curzon has said, it alone is worth a visit to Isfahan to see. I know it in photographs only, and in written descriptions, but I feel its beauty and magnificence. In many respects it resembles the Pul-i-Khaju (p.213), but it is a great deal longer, and no pavilions rise above its tiers of arches. To my mind the pavilions of the Pul-i-Khaju have an architectural value that cannot be rated at too high a level. So I miss their grace in the AliVerdi Khan, though this noble structure ought not to be criticised—except in an ashamed whisper.
BRIDGE OF ALI VERDITHE BRIDGE OF ALI VERDI KHAN OVER THE ZENDEH RUD AT ISFAHAN, PERSIA
THE BRIDGE OF ALI VERDI KHAN OVER THE ZENDEH RUD AT ISFAHAN, PERSIA
There is a gateway at the north end, so we must place the Ali Verdi Khan among the minor defensive bridges. A paved ramp or causeway leads from a great avenue to the gateway; and then a visitor has 388 yards to walk before he reaches the far end. The main road is paved, and its width is thirty feet. Upon each side is a gallery, or covered arcade, two and a half feet wide, which is pierced through the outer wall of the bridge from one end to the other; it communicates with the main road by frequentarches, it opens by similar arches—over ninety in number—on to the river view, and here and there it expands into large chambers, as we see in Brangwyn’s pen-drawing. The chambers used to be decorated with “not too proper paintings,” done in the time of AbbasII. At both entrances the Ali Verdi Khan is flanked by round towers, and staircases in the towers go up to a fine platform which in earlier times was a favourite promenade; but now it is disfigured by telegraph poles, the modern spirit everywhere having an unrivalled vulgarity.[117]“Similar staircases, cut in the basements of the towers, and also at regular intervals in the main piers, conduct from the road level to a lower storey, where, but little elevated above the bed of the river, a vaulted passage runs along the entire length of the bridge, through arches pierced in the central piers, crossing the channel of the river by huge stepping-stones planted in its bed. Colonel Johnson gives the dimensions of these transverse arches as ten feet span and nine feet high; and of the main arches (thirty-three in number) which they bisect, as twenty feet span and fifteen feet high, separated by piers eleven feet thick. There is thus a triple promenade on this remarkable bridge—the vaulted passage below, the roadway and lateral galleries above, and the open footpath at the top of all. I should add that the upper part of the bridge is of brick, the piers and towers are ofstone.”[118]
There is no European structure akin to this, but for a long time Rothenburg on the Tauber has been famous for a two-storeyed bridge; also we know that some modern commercial bridges have an upper road and a lower one, like the High Level Bridge at Newcastle. In every case the idea was suggested by the Roman practice of building aqueducts in tiers.