Building Castles

The following is Chapter 4 of Fry's excellent book, Castles of Britain and Ireland, ISBN 0 7153 0242 6 p David and Charles.

Purchase Details USA

Purchase Details UK

Building Castles

In Chapter 2 we saw that motte castles could be erected in a few weeks. Stone castle construction, however, was a very different matter and generally took years, despite the fact that most stone castles were wanted in a hurry. D.F. Renn estimated that the annual building rate on a great tower was about 3 metres (10 ft) of elevation, allowing for weather, stopping work in winter or if funds ran out. Scarborough, some 27.4 metres (90 ft) tall, took ten years (1158-68). This 3-metre average applies to great towers. In the same period, work on other parts of the castle must have been done simultaneously, such as constructing the curtain with towers (if any), raising the gateway or gatehouse and so forth. At Orford, for example, it is thought that the whole castle, great tower and surrounding enclosure with its smaller (rectangular?) towers and gate went up in seven or eight years, the tower itself in two years. The time may seem lengthy, but the medieval builder had no mechanical aids except the pulley-wheel crane and the wheelbarrow, and stone cutting and dressing had to be done by hand.

Every stone castle posed a veritable catalogue of problems, some of which had to be dealt with in advance and which could not be solved merely by riding roughshod over the feelings of local people. Early in the Norman occupation, the conqueror-lords steamrollered their way through towns and countryside to clear sites for castles and pressed Anglo-Saxons into forced labour. But in the time of Henry 1, who introduced a new spirit of co-operation between Norman and Anglo-Saxon (setting an example by marrying the daughter of an Anglo-Saxon princess), proper formalities were observed. Some sites were paid for, some were exchanged, and there are even

instances of compensation being paid for intrusion upon neighbouring land, such as at Gloucester. It is interesting to look at the various problems, for it is still a matter of some wonder how well the medieval builder coped with them, particularly in the construction of the great tower type of castle.

Some stone great tower castles were superimposed upon motte castles because the site had proved advantageous. Others were built on new sites, generally selected for their strategic position vis-a-vis the lands of neighbours as well as suitability for controlling their own lands, and for the nearness to water sources. There were other important questions for the builders to consider. Could stone be quarried on the site, or nearby, and was it suitable or too soft? If not, where was the nearest quarry? Did it belong to the king or the Church? Most quarries were royal or ecclesiastical property. How much would the stone cost? Where was the nearest waterway for transporting the stone and other materials? Water transport was much cheaper than land cartage, and it was quicker. Many of the best quarries, such as Barnack in Northamptonshire, were close to navigable rivers. How easy would it be to find enough masons and assistants to extract stone (by hammering iron wedges into the layers in the stone beds) and to cut by saw or split the lumps into appropriate shapes and sizes, and to dress (smooth) them? What sort of rubble was available nearby? Were there any Roman remains, such as tiles and brick segments? At Colchester great tower, both brick and dressed stones were taken from the extensive ruins of the old Roman town at Camulodunum and used in the masonry, and can be seen in many parts of the tower, notably in some of the steps of the great stairs. The owner and builder of a castle might also have considered whether he wanted to obtain a supply of the cream-yellow limestone quarried at Caen in Normandy, generally reckoned the best of its kind, to make the ashlar quoins and other features such as arch heads, keystones, window mullions and lintels, loopholes and battlement tops.

While masons were attending to the stone supply problems, carpenters and joiners were worrying about timber availability. Was there a source close to the site and were the trees big enough to produce the lengths of beam and plank? In some cases owners contracted to buy the wood from a supplier; in others they would buy a stretch of forest with the appropriate trees. Oak in one or other of its many species was preferred. Wood was also needed for burning to make charcoal for blacksmiths.

There were other materials wanted in quantity: lead for the tower roofs and water pipes; iron for a multitude of requirements; sand and lime for mortar (cement). All had to be acquired and transported to the site. Details of this kind today would be skilfully estimated down to the last nail, and the operations supervised by one or other of a building team of architect, quantity surveyor, consulting engineer and services engineer, but in those times there were no such disciplines. Estimating had to be guesswork.

Once the material supply problem was satisfactorily in hand, the building gangs were recruited, or pressed into service, generally from the immediate neighbourhood. However, in the case of the castles of Edward 1 in Wales, the authorities found they had to recruit from English counties, from Norfolk to Devonshire, because they could not depend upon loyal service from the Welsh whom they had just overcome. In early Norman days men were not generally paid, but as Norman and Anglo-Saxon began to edge towards a better relationship, castle owners started to pay wages. Pay was not standard, except in so far as it was low everywhere. In the late thirteenth century we read of diggers, carpenters and masons being offered bonuses for good work - and docked wages for absenteeism - but these were the country men recruited for work on the Edwardian castles in Wales. A site supervisor, or custos operationum, at Builth could earn 12d a day in 1277 and a master mason 7 1/2d a day in 1278. Piecework was also done: at Flint, in 1280, masons were paid 1d or 1 1/4d per stone for cutting and dressing. Diggers were paid 3d a day at Rhuddlan. At Deal in 1539, a number of labourers on the site of the coastal fort being built at the order of Henry VIII went on strike because of low wages: they were only paid 6d - and some even 5d - a day. The differentials of the thirteenth century seem much the same as today: thus the head of a building team on a large project would generally expect to earn about 1 to 4 times the basic wage (without overtime) of the builder's labourer.

Work on a castle site began with a careful scrutiny of the ground. If it was an existing motte site, the owner would have known about its earth content and whether there was rock or not. If it was a new site, this would have to be discovered. The presence of rock meant that possibly the great tower could be built on rock foundations. If so, the rock was flattened by gangs of men chipping it away with iron cold chisels, and the plinth of the tower was laid down. If the ground was soft, deep trenches were cut, a yard or two wider than the intended thickness of wall, and filled with an assortment of rubble, stone and timber - a kind of hardcore - which was rammed down. The plinth was laid on these foundations. In many great towers, plinths were battered, that is, they sloped outwards and downwards for structural strength and for ricocheting missiles; though some led straight downwards, such as at Richmond. Then the first levels of wall were put together above the plinth.

In some castles the walls consisted of an inner 'skin' of an aggregate of rubble, old brick, pebbles and almost anything else hard, held together by a tough mortar of sand and lime which set as hard as rock, sometimes prepared by creating a trough of wooden planks between which the aggregate was poured. The method may have been learned by the Normans from the Saracen castle builders in Spain of the tenth century who used tapia, a mix of pebbles and cement poured between boards and left to dry in the sun. A fine example of a castle built of this material is at Banos de la Encina. The aggregate skin was generally clad, in England, with an outer skin of dressed stone, and the corners of the towers were set with quoins of dressed stone which tied in with both skins. In some castles, such as Rochester, the walls were all rubble (Rochester was Kentish ragstone) with corners of ashlar. At Middleham great tower you can see stretches along the wall faces that are stripped of the outer skin, and part of the corners where the quoins have disappeared, leaving the inner rubble skin exposed. At Baconsthorpe, the quoins in the corners of the great flint-built rectangular gatehouse were deliberately taken out to provide for later building nearby.

As the first courses of wall went up, scaffolding became necessary to continue to work. Long poles held together with rope were erected, anchored here and there with horizontal poles, called put-logs, let into the masonry already built. These put-logs slotted into put-log holes, and many great towers (such as Hedingham) today bear the patterns of rows of put-log holes. In most cases they were horizontal lines. A more sophisticated system was used during the building of some of the Edwardian Welsh castles of the late thirteenth century, where helicoidal, or inclined scaffold paths, about 35-40 degrees from the horizontal, supported ramps for hauling or winching materials, notably at Harlech gatehouse. This system appears to have been a speciality of Master James of St George, the Savoyard master-mason. Whether the ramp for a wheelbarrow was more efficient than the pulley wheel and basket is a moot point.

Great tower walls rose in a vertical straight line (although at Oxford the walls tapered inwards with offsets

spaced out in the height), either solid or hollow with passages here and there, and in the greatest of the towers, chambers and staircases until they reached the desired height whereupon they were topped with the wall-walk and parapet and given the outer protective battlements. Many towers had corners which were in effect corner turrets, and these rose higher than the four walls with their battlements. Corner turrets contained the spiral staircases and also rooms, and they acted as buttresses. Some great towers had one corner turret of slightly greater dimensions than the other three.

When the walls were completed they might be coated with plaster and whitewashed, or whitewashed directly on the stonework. Occasionally, the plastering was made to look like coursing of large stone blocks (as may be seen on many Georgian or nineteenth-century town buildings), by means of thin lines etched into the wet plasterwork. The great tower at London is called the White Tower because of the whitewashing it received in the thirteenth century. Whitewashing was not simply decorative; indeed, with the smoke and dirt given out by fires and from slops thrown out of windows and loops, it would not have remained clean for long. It was a preservative for the stone, and it is also held that medieval builders thought it helped to fireproof the castle.

The building operations we have outlined applied to great towers. They were much the same for gatehouses and gatehouse-towers, chapels and halls. Put-log holes can be seen in many smaller curtain wall towers and in gatehouses, such as at Tonbridge. In some of the Edwardian castles, the put-log holes are also in helicoidal pattern on the smaller towers and gatehouses. At Goodrich the three early fourteenth-century cylindrical towers in the inner quadrangle enclosure stand on a square base with spurs up the tower sides, an alternative to a battered plinth. This type of support was particularly appropriate for structures on sheer cliff faces, like the Constable's Gate at Dover.

Since castles were residences as well as fortresses, we should not be surprised to find them having many domestic features. Internally, many had fireplaces, some of considerable decorative attractiveness and elaboration, from the simple thirteenth-century sloping ashlar hood on corbels and columns at Tretower great tower, to the massive 6-metre (20-ft) wide pillar-supported hood at the great hall at Linlithgow. Windows in great towers and halls are endlessly fascinating in variety. Loopholes (arrow slits) are likewise varied. We have mentioned the chapels in the fore-building at Dover. At Castle Rising the chapel is next to the great chamber, at its east end. At Conisbrough the chapel, hexagonal in plan, projects into one of the wedge-shaped buttresses. At Colchester and the White Tower of London, the chapels are in the apsidal ends in the east wall. Numerous great towers had kitchens in the wall thicknesses, or in the basement, or at the top - perhaps a better place, so that the cooking smells could get out without affecting the occupants. At Orford there were two kitchens, one at ground- and one at first-floor level.

What did all this castle building cost? A great deal of work has been done on the costing of building works on castles in the Middle Ages, particularly those in royal hands, and the costs to private owners other than kings must have been much the same. Expenditure for many royal castles is well documented, and sums looked at in relation to the total income of the kings in a year are startling. We may take the position over the period c. 1155-1215, namely, the reigns of Henry 11 and his sons Richard 1 and John. It is reckoned that the king's annual income from taxes and rents was not much more than £10,000. It is also estimated that none of his lords was worth more than about one-twelfth of that £10,000, and that the average knight could indeed live comfortably on £20 a year.

If the king's annual income was about £10,000, then from the evidence in the Pipe Rolls of the Exchequer, he seems to have spent a significant percentage of it on castle works, from new structures like Orford, to new parts at Dover, Newcastle and so forth, to repairs and upkeep on many others. The History of the King's Works considers the outlay on castles to have been the biggest single item of expenditure in all three reigns. Orford cost about £1400 in seven years, Dover had nearly £7000 spent on it in nine, Newcastle cost £1000 in ten years and Bowes £600 in seventeen years. King's Works estimates that about £780 a year was spent on castles by the Crown throughout the period 1155-1215, or a total of over £46,000 - about 7-8 per cent of the total income the Crown received over the 60-year period. Interestingly, while Henry 11 spent some £21,000 on about 90 castles, the great majority of the expenses were on less than 30 castles. In John's reign, £17,000 was spent in 15 years, the lion's share being on only 10 castles: over £500 on Hanley, Horston, Lancaster and Norham; over £ 1000 on Corfe, Dover, Kenilworth, Knaresborough and Odiham (his polygonal great tower castle); and over £2000 on Scarborough. A graph of his expenditure over the 15 years would show an acceleration towards the latter end, in both new works and repairs, and this has been taken as an indication of the growing tension between Henry 11 and his feudal lords at home, and the growing danger of invasion from abroad (Dover, Southampton and other castles in south-cast England are recorded to have received the main attention).

If we turn now to the expenditure of the three Edwards (1, 11, 111) in Wales from 1277 to 1330, the figures for which many writers on castles quote with relish, the sums involved are centred on ten new castles (Aberystwyth, Beaumaris, Builth, Caernarfon, Conwy, Flint, Harlech, Hope, Rhuddlan and Ruthin), with renovations to three Welsh-originated castles (Castell y Bere, Criccieth and Dolwyddelan). Aberystwyth (£3885), Beaumaris (£14,444), Builth (£1666), Caernarfon (£19,892), Conwy (£14,248), Flint (£8951), Harlech (£6224), and Rhuddlan (£9292) give an idea of the sort of money these kings were laying out. These sums are the minimum, as they do not include absolutely everything, but they are near enough for us to stand back in some wonder, more so when it is learned that some of the castles actually began to decay almost before they were finished, that Caernarfon took more than a quarter of a century to complete entirely, and that Beaumaris was never completed nor was it ever involved in any warlike event.