Contents:
Edward I and Scotland
Architectural development of Scottish castles
Siege engines
Sieges
Later medieval castles

Edward I and Scotland
The main castles to he held by the English in Scotland were: south-east, Berwick, Jedburgh and Roxburgh. In
Lothian, Dirleton, and Edinburgh. Bothwell in Central Scotland, Bothwell and Stirling.  South-west, Caerlaverock,
Dumfries and Lochmaben. North, Kildrummy, in Aberdeenshire. There were many others such Castle Urquhart
on Loch Ness, St Andrews on the east coast, etc. There was therefore no need for Edward I to launch the sort of
building programme that there had been in Wales.

In the course of the Great Cause, the hearing of the disputed Scottish succession in 1291-2, Edward I took
charge of all the Scottish royal castles, 23 in all. They were, of course, handed to John Balliol once judgement
went in his favour. The war began in 1296. Edward obtained the surrender of the Scottish castles, and of the
Scottish king, with little difficulty. The English attempted no work of fortification apart from the rebuilding at
Berwick, where Edward planned a new town there, much like the new towns in Wales. The castle at Berwick,
however, had not been demolished, and there is no indication that work was planned on it.

In 1297 the work of conquest was undone in the rising under William Wallace and Andrew Moray. Most of the
castles fell into Scottish hands –  Edinburgh, Jedburgh and Berwick did not, and the status of some others, such
as Roxburgh, is uncertain. Stirling was taken by the Scots soon after their triumphant victory in the autumn at
Stirling Bridge. 1298 saw Edward’s revenge at the battle of Falkirk, and the recovery of Stirling, as well as of
Lochmaben in the south-west, and Jedburgh in the south-east. The English-held castles fell into three main
groups. South-east: Berwick, Roxburgh and Jedburgh. Southern side of the Forth: Edinburgh, Linlithgow and
Stirling. South-west: Lochmaben and Dumfries controlling the route into Galloway from Carlisle. There were
other castles, of course, in English hands, such as Ayr on the south-west coast. But there was no significant
presence north of the Forth.

Building works were very limited. Lochmaben: Robert Clifford put up a ditch and palisade. In 1300 similar
expansion of the defences of Dumfries was ordered. All the work was in wood and earth. That year saw the
capture of Caerlaverock from the Scots, in a celebrated siege. This was a relatively small, but modern, castle in
the south-west. Its garrison had been a thorn in the flesh of nearby Dumfries and Lochmaben. In 1301 Bothwell
in central Scotland was taken after a brief siege. Edward remained in the north over the winter, staying in
Linlithgow. Additional defences were constructed, again in earth and wood, but early in 1302 it was decided to
build in stone. Master James of St George was in charge. A twin-towered gatehouse, a stone curtain, with
towers at each end, would command the peninsula stretching into the loch. But this was abandoned; the work
was to be done in earth and wood. Other work was done at Selkirk and Kirkintilloch, but again there was nothing
very elaborate. In 1303-4 Edward again wintered in Scotland, this time in Dunfermline in Fife. All that was done
by way of defensive work was the creation of a ditch and bank. It was only after the surrender of the most of
the Scots, and the subsequent capture of Stirling in 1304 that significant castle-building was planned. Three new
castles were planned, at Tullibody and Inverkeithing on  the north shore of the Forth, and one at Polmaise about
three miles east of Stirling. Edward had been quite clear in letters that control of the Forth was vital – he saw
the crossing of it (which he called the Scottish Sea) as essential for conquest. The plan was to pay for them out
of the fines paid by the Scots who had come to Edward’s peace – the money, according to the king,was to be
used   ‘for the work of the castles that we are having built in the said land of Scotland for the security of the
said land and keeping the peace, or to be put to other use, as we see should be done.’ Little work was in fact
done. Under Edward II there was a steady process of the Scots gaining English-held castles, and in very many
cases slighting them, so that they would be unusable in the future. Berwick fell, through treachery, in 1318, and
the English were unable to recapture it despite a major campaign in the following year. The Scottish wars were
turned around under Edward III. Berwick was regained in 1333 after the battle of Halidon Hill, but while some
work of repair was undertaken by the English at Edinburgh, Stirling and elsewhere, there were no substantial
works done.

Scottish architectural historians have wanted to see much more that is Edwardian in the castles of the late
thirteenth and early fourteenth centuries. Excavation of Bothwell shows  that it was intended to build a
twin-towered gatehouse – or even that such a gate was built and subsequently demolished. The enceinte would
then have been considerably bigger than what stands today. Such a gatehouse, for some at least, is indisputable
‘Edwardian’. Another example is Kildrummy in Aberdeenshire. Here again, a twin-towered gatehouse of
Edwardian proportions, but now sadly ruined. Caerlaverock in Dumfriesshire presents another example. This is,
unusually, a triangular castle in plan. A twin-towered gatehouse is at one angle, round towers at the others.
There has been dispute over the date, but it is widely suggested that it is indeed English build, of the 1290s.

The Bothwell gate is fairly simple – there is no question of it extending back into the courtyard on the Harlech
scale. The castle was begun in the 1240s, and there is no reason why the gate should not date from that period.
Kildrummy was built by the bishop of Caithness between  1223 and 1245, and again there is no reason to suppose
that the gatehouse was not a part of the original structure. To prove Edwardian work, you would have to show
that there were the distinctive elements of Savoyard style present, as in the Welsh castles. That simply is not the
case. As for Caerlaverock, who could have built it in the 1290s? The castle was granted to Robert Clifford in
1298, and it is hard to believe that he had time to build it, and for the Scots to take it over, by the siege in
1300.

Military value of castles. Stirling held out for over a year in 1299. Turnberry and Ayr held out in 1301, and
Linlithgow successfully survived siege in early 1303. The main role of the castles was as bases for raids into
enemy territory. In 1298 a raid was organised, with 40 cavalry drawn from the Roxburgh garrison, 30 each from
Berwick and Edinburgh, 10 each from Selkirk and Jedburgh. The garrison of Norham was asked for 20 men. In
1301 John de St John, commanding in the south-west, was asked to have a force of 120 men-at-arms constantly
ready to make forays in the winter from Lochmaben and Dumfries.

The castles were administratively important; most of the constables were sheriffs for the areas under their
control. Little revenue was received – the constables were preoccupied with military issues. The role of the
castles, though, was surely vital to Edward’s efforts in Scotland. The summer campaigns were often surprisingly
ineffective, if the Scots were unwilling to give battle. If the country was to be held, then it needed the
permanent presence of English troops in English-held castles. The Scots were clearly well aware of this – the
policy of slighting castles when they were taken shows this.

Architectural development of Scottish castles.
A brief summary. Scottish castles do not fit into quite the same pattern as do English ones. Certainly there
were motte and baileys, especially in the south-west – Scotland was subject to Norman infiltration if not
invasion. North of the Forth, however, mottes are rather less common. Stone castles did not develop as in
England, for the great square keep was not a feature of Scottish castles – apart from Carlisle keep, built by
David I. Scottish stone castles of the twelfth century have left little trace. There were some stone fortifications
in the Western Isles, notably Castle Sween, a quadrangular enclosure probably of late twelfth or early
thirteenth-century date. Rothesay is probably early thirteenth-century in date; it consists of an oval curtain wall,
with four equidistant round towers. Other castles in the west, such as Mingarry, are a simple curtain wall, of
irregular shape. Inverlochy (Invernesshire) is a classic thirteenth-century castle, a rectangular curtain with four
round corner towers, one larger than the rest. More in the mainstream is Bothwell, with a remarkable fine
thirteenth century donjon, which suggests a strong French influence. It resembles the very grand round donjon
at Coucy, now destroyed. Caerlaverock, however, in the south-west, is a remarkable triangular castle, with two
towers at one angle forming the gatehouse. Kildrummy is a little like Caerlaverock, in being essentially
triangular, but it has a donjon, and two of the sides are curved outwards, with D shaped towers half-way along.
Similarities of detail have been noticed between Kildrummy and Edward I’s Welsh castles – the ‘cusped
shouldered arch of the twin-light windows’ which can also be found at Caernarfon, while a fireplace jamb
apparently had parallels to Conwy. There is, however, no evidence of English building taking place at these sites.
It may be that the Scots employed some English masons – but this is not work of the Edwardian occupation of
Scotland. It may be that some English detail at Bothwell dates from the English tenure of the castle in 1331-7,
but there is no reason to suppose that the Scots could not build up-to-date castles. The triangular plans, the
presence of donjons, the absence of the rear stair-turrets on the gatehouses – these all mark them out as
distinctively different from the English castles. As for the greatest Scottish castles, Edinburgh and Stirling, we
know little of what they were like in about 1300 – so much of what stands today is of far more recent
construction. But Dirleton, built on a rocky outcrop, perhaps gives some idea, with round towers of good quality
masonry and a good curtain wall.

The tradition of great castles came to an end in the fourteenth century. The end was brought all the quicker by
the Bruce policy of razing castles which were likely to be used by the English as occupation bases. Bothwell, for
example, had its great donjon half cast-down in 1337. Tantallon, of mid-fourteenth-century date, is
exceptional. Rather like Dunstanburgh, it consists of a curtain wall across the neck of a promontory, with towers
on each end, and a rectangular gatehouse in the middle. Doune was late fourteenth-century in date, and was a
small courtyard castle, with well integrated accommodation. It does not have the conventional twin-towered
gate, but rather a single circular tower controlling the entry. Behind it lies a substantial block, which extends
into a range containing what is conventionally regarded as the lord’s hall, and the common hall. It was probably
intended that the castle should be completed with a full range of accommodation all round. Doune demonstrates
the need to incorporate domestic accommodation into the defensive envelope.

The main Scottish tradition that developed in the later medieval period was that of the tower house. Threave in
Galloway, built by Archibald the Grim, third earl of Douglas, in the late fourteenth century. It is rectangular, 60
feet by 40, and 70 feet high. Its walls are 8 feet thick. The building has five floors, and the wall-head was
defended with a timber gallery. It resembles an Anglo-Norman keep in a superficial sense, but has little of the
complexity of mural chambers that a keep of this size would have had.

Craigmillar, also late fourteenth-century in date, was a simple tower, but with an important element in its plan –
a projecting wing making it L shaped, and providing it with smaller chambers in addition to the main rooms.
This was to become a dominant plan of Scottish tower-houses. At Borthwick, licensed in 1430, there are two
such wings enabling much a more complex internal structure. This was a well-fitted building internally, with fine
fireplace in the hall and a good deal of decorative stonework. There was an unusual garderobe, which was fitted
with a system of removable containers rather than the normal unhygienic chute. The building is vaulted, an
additional sophisticated element.  This was an exceptional building. Most towers were fairly simple structures,
and of course a great many were built outside our period, in the sixteenth century. The tower houses combined
defence with accommodation in very much the same way as an Anglo-Norman keep. There is no easy
explanation as to why this type came to dominate as against the courtyard castle – it offered no very obvious
advantages, and fashion must have been an important element. The defensive elements of the tower houses
should not be overplayed. They often had ground floor entrances, and the machicolations and bartizans at the
wall head were at least as much decorative as defensive. Gunloops and arrow-loops do not feature on any scale –
it is the outer defences of Scottish castles which begin to see artillery defences on an increasing scale from the
mid-fifteenth century.

 Siege Engines
There is a reasonable amount of agreement in the secondary sources about this topic – you will find fairly
standard descriptions of siege towers, mangonels, trebuchets and the like. Much that is written is based on
nineteenth constructs, rather than reliable contemporary evidence.

The movable siege tower or belfry has a long history. They were wooden structures, normally wheeled, with
several platforms or storeys. They had to be taller than the castle they were attacking. At Lisbon on the second
crusade one English one was 83 feet high, another 95.  There are two particularly well-known examples of siege
towers from the Anglo-Scottish wars. Bothwell castle was captured by Edward I’s forces by the use of a great
belfry, brought on carts from Glasgow. A special roadway had to be constructed to bring it up to the castle. In
1315 the Scots attempted traditional siege warfare at Carlisle, and built a siege tower. But Carlisle is on low
ground, and the machine sank into the mud and could not be moved. Moving these machines must have been a
major problem. In the siege of Lisbon in the mid-twelfth century a belfry was moved forward a mere ninety feet
on the first day. Fire was another problem. At the siege of Acre on the third crusade the French built two
machines, a tower and a testudo, a strongly armoured low movable shelter. The defenders of the city threw
brushwood at the two machines, and then threw Greek fire at them with a machine. The brushwood caught,
and both French machines were burned.

The simplest assault weapon was the scaling ladder. Early in Edward II’s reign at Roxburgh, the castle ‘was taken
through the exertions of James Douglas, who was on the side of the Scots. This James came secretly to the fort
by night, brought up ladders stealthily and placed them against the wall; and by this means he climbed up the
wall, and leading his companions upon the sleeping or heedless guards, attacked those he found and took the
castle.’  When the English found the Scottish ladders after a failed assault on Berwick, they were highly
surprised, and very impressed by the ingenuity of the design.

Throwing weapons. The usual division is that between machines  worked by tension, as with a bow, those which
used torsion (twisted ropes), and those which used a counterweight to provide motive power. A large crossbow
could be mounted on a frame, with a mechanism to cock it – this was an arbalest. The next stage up was the
Roman-style ballista. Here, the arms of the bow were fixed in twisted ropes or sinews, set in a frame. This
enabled a more powerful engine to be built, with less danger of a bow snapping. There are not many illustrations
of these, but we have, for example, payments at Newcastle in 1297-8 for an engine made to wind up a screw
crossbow. This shot what are described as ‘shafts’, not quarrels, and this may have been a ballista.

The other type of torsion engine is that conventionally known as a mangonel, though it is not clear that is
correct. The concept was of a single throwing arm, with a cup on the end. The base end would be fixed to
twisted ropes. When it was pulled back, the torsion was increased. When released, it would hit a padded
cross-beam, loosing the stone that had been loaded into the cup. Such machines are illustrated in an
eighteen-century edition of the classical author Polybius. The Onager was similar, but had a sling on the end of
the beam, which gave greater leverage as it hurled its stone. It is described by Ammianus Marcellinus, writing in
the late fourth century, but we have no medieval descriptions or illustrations. There is just one medieval
illustration from the Walter Milemete manuscript. It is a sketched drawing, not a fully painted illustration, and it
shows quite the beam with a cupped end, held in a twisted skein at one end. It is mounted on a wheeled frame,
and the drawing shows a man-at-arms twisting the skein tight with a four-armed wheel or capstan. The drawing
may represent an attempt to reproduce some classical drawing, rather than being from life.

The main type of throwing machine used the principle of a sling combined with that of a lever. A lever, with the
fulcrum placed towards one end, offers a way of changing the nature of force, transforming a short, relatively
slow movement into a quick, long one.  A hand sling consisted of a stick with a cloth or leather pouch slung at
one end. Such a sling could cast a stone a long way, for it lengthened and speeded up the throwing arm. A larger
beam could be put in a frame so that it could swivel. Then, several men could pull the beam down – it would be
fitted with ropes for them to pull.

The origins of this type of machine lie, almost certainly, in China. The Slavs used what seems to have been a
traction machine (i.e. a hand-pulled one) in 597 at Thessalonika, but details of how machines in the west
worked are not known. A treatise written for Saladin in the late twelfth century shows that he was familiar with
lever machines. There is also a description of some kind of counterweight machine. All of the early lever
machines were hand-operated, in some cases by very large gangs of men, all pulling simultaneously. Then, by the
turn of the twelfth century, it was realised that if you put a heavy counterweight on the end of the beam,
energy would be stored in the counterweight. No pull was needed to release it, only to cock it. This type of
machine is normally called a trebuchet, and appeared in the west in the early thirteenth century. The next
change was to abandon the fixed counterweight on the end of the beam, and in its place have a swinging box.
There were various advantages to this. If you have a fixed counterweight, it is not easy to change its weight. So
the machine will always shoot at the same speed, and each projectile of the same weight will land in the same
place. The way to alter range would have been to use lighter or heavier ammunition. With a swinging box, you
could load as much or as little weighting material (stones or earth) as you wanted, and change the range by
adding or subtracting to this ballast.

There were various variants possible. A mixed counterweight and pulled machine is illustrated in a
thirteenth-century French bible, and a similar machine is depicted in a carving of the machine which killed Simon
de Montfort the elder at Toulouse in 1218. In the fifteenth century a machine called a Cuillard  had a single
central mount, but the beam was split, as it were, so that it carried two counterweights which dropped either
side of the central mount. But this is just a variant on the counterweight trebuchet.

These were potent machines, capable of hurling a 250 lb projectile a couple of hundred yards, with very
considerable force. The trajectory was a high one, so it could lob stones over a castle wall with little difficulty.

There is much evidence for siege engines in England, but it is usually impossible to say much about what type
they were. There is little about them in the eleventh or early twelfth centuries. Throwing engines of some kind
were referred to at the siege of Faringdon in 1145, but most of the engines in this period were belfries or other
movable shelters. At the siege of Lisbon the English had two throwing machines, operated by teams of a hundred
men working in rotation. This implies that they were ‘traction trebuchets’, hand-powered machines pulled by
teams of men. There are not many references to siege engines in the rebellion of 1173-4, but carpenters were
hired to make engines when Henry II came to Huntingdon in 1174, and a mangonel costing one mark was bought
for Dover. The Scots had a couple of machines at Wark, but when the first was used, the first stone hardly came
out of its sling, and landed on an unfortunate Scottish knight. In 1194 the pipe rolls refer to both mangonels and
petraries at the sieges of Marlborough and Nottingham, and under John there are a many references to Turkish
mangonels.

Prince Louis had a huge machine in his siege train, called the Malvoisin, surely a counterweight trebuchet. At
Bedford in 1224 there were both mangonels and petraria. The engines of this period were substantial – it took
ten days to get the engines for the siege of Bytham from London in a dozen carts, at about ten miles a day. By
the early fourteenth century Edward I was in a position to deploy some thirteenth engines at Stirling. Lead was
taken from church roofs for counterweights – this suggests that some at least must have been
fixed-counterweight engines. The greatest of the machines was the Warwolf . Trebuchets continued to be built
and used through the fourteenth and early fifteenth centuries; it was not until mid-century that it was quite
clear that guns had replaced them. We have just one surviving plan for a trebuchet, drawn by the thirteenth
century Villard de Honnecourt. It gives the measurements of the box as 12 x 8 x 12 – this was a massive
machine, capable of taking a counterweight of twelve tons.

At the siege of Mortagne in the early 1340s the town was pestered by a fine engine, which hurled stones by day
and night. An engineer within the town built a small machine, ‘well-made and tempered, and so well pointed,
that it was used only three time; the first stone fell within twelve paces of the engine of the Valenciennois; the
second was nearer to the box, and the third was so well aimed that it struck the machine on the shaft, and split
it in two.’ (Froissart)

One of the more puzzling machines is the springald. There are lots of references to springalds, which were
commonly used within castles, by the defenders. It seems that the motive power derived from a single huge
bow, which was set vertically in a socket in the base of the machine. This could then be pulled back and cocked
by ropes and a windlass. For projectiles, it used bolts or shafts. These were placed in a holder, which could be
elevated and aimed. When the bow was released, it struck the bolt. To give an example of the type of defensive
armament: at Southampton in 1353 there were 12 crossbows with one foot, and 300 quarrels for them. There
were 12 bows with 300 arrows. There were 12 springalds, with 300 bolts (attiliis), two great engines called
‘magnels’ and a small engine called a Tripoget.

The final engine to be discussed is the cannon. Gunpowder was known to Edward I, who had it at the siege of
Stirling in 1304. Guns first appeared in the 1320s, and the earliest illustrations appear in English manuscripts of
1327, presented to the young Edward III. This cannon shot a bolt, like an arrow, not a ball. Written references
also begin at the same period; they were used at the siege of Tournai in 1340, and on the battlefield on Crécy in
1346. By the 1370s, guns were getting larger, and the early fifteenth century was the age of really huge
bombards, such as Mons Meg at Edinburgh.

Guns could be made of bronze, which would have been cast. Iron, however, was much more plentiful and cheap.
The technology, however, was not available to cast large pieces in iron, though small cast iron guns were
made.The alternative to casting was forging. Strips of iron could be welded together to form a tube. Rings could
be placed round the tube to strengthen it, and a strong piece of iron welded onto the back. There would
normally be a removable breech chamber  – this was known as early as 1342. If you had such a breech, then you
could use a gun inside a castle, without needing to move it every time you wanted to load it. Stone projectiles
were used for the largest guns, metal for the smaller ones.

The men responsible for the siege machinery were, of course, professionals. They were expert engineers and
carpenters, not earls or knights. One, was the Gascon Master Bertram, who served both Henry III and Edward I.
At Caerlaverock, the main engineer was a man called Brother Robert of Ulm, who at 9d a day seems to have
somewhat underpaid, for a master engineer. The main carpenter was Adam of Glasham. Reginald the Engineer
was employed in 1300 to make machines for Berwick; another employed there, intriguingly, was a Durham
monk, Thomas of Bamburgh. Their skills were presumably practical not theoretical; it is, of course, possible to
work out the parameters of a medieval trebuchet mathematically, but it was surely in practice done by eye and
by experience.

Sieges
1. Sieges ending in negotiation. There were in legal terms two ways of taking a castle. It could be stormed,
either by surprise or after the walls had been breached. Or it could be surrendered by arrangement, after
negotiation. In formal terms, the first step was for the commander of the besieging forces to demand the
castle’s surrender. The constables of Nottingham and Tickhill surrendered in 1191 to Prince John without a fight.
Their lord, Roger de Lacy, ordered their arrest – they fled, and were regarded as traitors to their lord. In 1203,
during the French conquest of Normandy, Philip obtained the surrender of Vaudreuil before he even brought his
siege machines into action. Saer de Quency and Robert FitzWalter, the commanders, were treated with
contempt for this – Philip II kept them in chains, and demanded a huge ransom of 5,000 marks for them. John,
however, to maintain his honour, put it about that they had surrendered on his instructions. To surrender
without a siege was therefore dreadful. In 1453 it was said in the indictment of Somerset that ‘It hath been
seen in many realms and lordships that for the loss of towns or castles without siege, the captains that have lost
them have been dead and beheaded, and their goods lost.’

The formal start of the siege came not with the demand for surrender, but with the firing or loosing of the first
weapon. Once that was done, then honour was more easily satisfied should the garrison surrender. The longer
resistance carried on, the more honourable in one sense the constable and the garrison were. But there was a
downside to carrying on resistance. If the defenders refused offers of negotiation, then the laws of war were
severe. All the goods of the garrison or townsmen were regarded as forfeit, and rape and slaughter were the
order of the day. Instances of widespread slaughter are relatively rare. The capture of Berwick in 1296 by
Edward I was certainly the occasion for a bloodbath. Caen in 1346 also witnessed horrific scenes. The sack of
Limoges by the Black Prince in 1371 is another example.

Between the alternatives of a swift surrender which would be seen as treasonable, and holding out to the bitter
end, which could result in savagery by the besiegers, lay negotiation. Particularly common were agreements
whereby a truce was agreed for a specific period. At the end of it, if no relieving force had appeared, the castle
would be surrendered – if relief did come, the siege would continue. Surrender would be on honourable terms –
the garrison would be spared life and limb, and might even be allowed to exit with all their possessions. The
besiegers might be seen to be chivalrous and generous opponents, and the defenders to have behaved
honourably.

The most famous example is from Stephen’s reign. Stephen was besieging John FitzGilbert, the Marshal, in
Newbury. A deal was struck; John would not bring any men or supplies into the castle during the period of truce.
His son William was handed over as a hostage to ensure that the bargain was kept. But John slipped some
supplies into the castle, and was caught. The young William should have been hanged. John claimed not to care:
‘Have I not the hammer and the anvil to make new sons?’. But Stephen was charmed by the young boy and
spared him. In 1173 Roger de Stuteville, ‘who never liked treason or to serve the devil’, constable at Wark,
realised that he could not prevail against the Scots. He decided to ask for forty days truce, so that he could go
overseas, or send letters there, to get aid from his lord, Henry II. William the Lion gladly agreed. However,
Stuteville sent out messengers, and went to England himself. He brought back a sufficient army within the terms
of the truce to deal with the Scots, and Wark was saved. Stirling in 1314. The Scots under Edward Bruce agreed
terms with the English garrison under Philip Mowbray in early summer 1313 – if they were not relieved before
Midsummer 1314, the castle would be surrendered. That, at least, is the story given by Barbour, and accepted
by other writers. At all events, an agreement was reached, the castle was to be relieved by Midsummer 1314.
The outcome, of course, was that Bruce and the Scots encountered the English as they approached Stirling, and
defeated Edward II’s huge army at Bannockburn.

Another example of an agreement from the Scottish wars is that which led to the battle of Halidon Hill in 1333.
Here, the town and castle of Berwick were besieged by the English. The defenders negotiated a fifteen day
truce, surrendering a dozen hostages as guarantees. If they had not been relieved by 11 July they would
surrender. On the morning of the 11th, the Scots slipped across the Tweed, burned Tweedmouth, and got
victuals and reinforcements across the river into Berwick. The return of the hostages was demanded. Edward
pointed out that the terms of the agreement had not been met. Berwick had not been relieved from the
Scottish side, but the English. The first hostage, Thomas Seton, was duly hanged. So as to save the lives of the
remaining hostages, a new agreement was reached. Complex clauses defined the relief of the town. If the Scots
forced their way over the Tweed by Berwick Stream to the west; if the Scots were victorious in battle on
Scottish ground between the Tweed and the sea; if they got 200 men-at-arms into Berwick with a loss of no
more than thirty. In the event, Douglas chose to risk a set-piece battle rather than lose Berwick. The result was
the disastrous defeat of Halidon Hill, and the taking of Berwick by the English.

2. Major sieges. Bedford in 1224 – William de Bréauté, brother of Falkes, kidnapped the royal justice Henry of
Braybrooke, a challenge which the crown had to meet. The king sent messengers to the castle, asking that he be
allowed to enter, and demanding Braybrooke’s return. William de Bréauté refused. The king ordered the
investment of the castle by his troops. The defenders cleared the walls and ramparts, making ready their
defence. Siege engines arrived. These kept up a fierce attack – the defenders countered with missiles. Many
were killed. The king’s workmen made a wooden castle, tall and symmetrical, in which his crossbowmen were
placed. They could see into the whole castle; no one within dared take off their armour. As the siege developed,
there was a petraria and two mangonels on the west, bombarding the keep. Two mangonels on the east
attacked the ‘old tower’. One to the south, and one to the north, attacked the curtain wall. Then there were
two wooden towers, for crossbowmen and observers. There was also a Cat, providing protection for miners
working away to bring down the walls. It took four assaults to take the castle. In the first, the barbican was
captured, and four or five foreigners killed there. In the second, the outer bailey was taken, and with it horses
with their harness, breastplates and crossbows, oxen, live pigs and bacon carcases. The buildings in the outer
bailey, containing hay and grain, were burned. In the third attack the miners brought down the wall next to the
old tower, or keep, and the breach enabled the men to get into the inner bailey. Finally, the miners fired the
keep, so that it was cracked, and smoke poured into the room where the defenders were. They then
surrendered.

Kenilworth. The siege began in November 1265, but in the early stages it looks as if the crown was unable to
bring together the forces needed. In March 1266 a royal messenger was sent to talk to the rebels, who sent him
back with his hand cut off. The royal army was divided into four divisions, and nine trebuchets were set up to
bombard the castle. Hurdles, to protect the besieging troops, came from Oxfordshire, Worcestershire and
Northamptonshire. Two belfries were built, but both were destroyed by the castle’s engines before they reached
the walls. In July negotiations were attempted; the archbishop of Canterbury and the papal legate, Ottobuono,
offered terms, and excommunicated the garrison when they were rejected. In retaliation, the garrison dressed
one of their number up like a papal legate, and he solemnly excommunicate the king, the legate and all their
men. An assault by barges across the water-defences failed. The garrison, by the autumn, were running low on
food. In mid-December, the garrison marched out before the term was up. Supplies had run out.

Caerlaverock, 1300. The one reason why it is notable is that a remarkable work was written about it, a heraldic
poem. Now the main purpose of the poem was to provide a record of the coats of arms of all those present.
Each of the bannerets present got a little character sketch as well. ‘Guy earl of Warwick, who of all that are
mentioned in my rhyme, had no one who was a better neighbour than he was, bore a red banner with a fess of
gold and crusilly.’ John de Beauchamp bore handsomely, in a graceful manner, and with inspiring ardour, a
banner vair.’ Some got a very brief mention; Robert Clifford has about twenty lines, concluding ‘If I were a
young maiden, I would give him my heart and body, so great is his fame.’ The castle is described – it was ‘shield
shaped, with a tower on each angle, but one of them was a double on, so high, so long, and so large that under
it was the gate with a draw-bridge’. It had a moat filled with water. It could not be approached from the south,
for there were dangerous defiles of woods, and marshes and ditches, so it was approached from the east, down
a slope. The troops set up their tents and cut down trees to make huts. Then the navy arrived, with the engines
and provisions. Then the infantry began to attack the castle, but so effective was the shooting from within the
castle that in one short hour many of the attackers were wounded or killed.  When the men-at-arms saw that
the infantry had sustained such losses, ‘many went there running, some jumping, and rushed in such haste that
they had no time to speak to anyone. Stones were thrown – I think by both sides ‘for to kill and wound was the
game at which they played. Great shouts arose among them, when they perceived that any mischief occurred.’
‘They threw up many a stone, and suffered many a heavy blow’. The poet clearly thought all this was a bit too
impetuous. ‘Willoughby in his advance got a stone in the middle of his chest, which he should have had protected
with his shield.’ The gate came under attack, but ‘There were showered upon them such huge stones, quarrels
and arrows, that they were so hurt with wounds and bruises that it was with very great difficulty that they were
able to retire.’ The attacks continued. ‘Those within bent their bows and crossbows, and aimed their springalds,
and were quite ready to throw and to hurl.’The defence was heroic, ‘But their courage was considerably
depressed during the attack by brother Robert, who sent numerous stones from the robinet, without ceasing
from dawn till dusk. Moreover on the other side he was erecting three other engines, very large, of great power
and very destructive, which will confound the castle.’ The garrison could not endure the bombardment, and one
of them put out a pennon as a symbol of surrender, but an archer shot him through the hand into his face. He
begged that they do no more, and at the moment the marshal and constable ordered the assault to cease. The
castle then surrendered.

The financial accounts of the royal wardrobe contains details such as the purchase of two great leather scales to
weigh the trebuchet ammunition. Brother Robert of Ulm is mentioned, as a master engineer. Master Adam of
Glasham, a carpenter, was retained at the king’s wages to take a siege engine from Lochmaben to
Caerlaverock, which took him 10 days. The Grace Dieu of Ross carried 5 tuns of wine from Kirkcudbright to
Caerlaverock, a ten day voyage. Another ship took a fortnight to bring flour to Caerlaverock from Whitehaven.

Stirling in 1304. Edward I’s great siege.. Langtoft states that the king set up very large engines there; the
garrison was no more than two knights, twenty men-at-arms, a friar, a monk, and 13 women. They had an
engine themselves, but its beam broke. Edward’s engines began to send stones through the walls and
crenellations, crushing the houses within. Then he ordered his carpenters to make a horrible engine, called
Ludgar, and when it hurled, it cast down the whole wall. Acording to Rishanger, when the besieged saw the
moat being filled with earth, and scaling ladders being prepared, they surrendered. Guisborough is similar. The
English had wooden machines that threw stones of 100, 200 or even 300 lbs. They brought down the
battlements, but the defenders held out, and killed many with their own machines. Then the king ordered the
moat to be filled, but the defenders managed to set light to the wood used for this. Then he ordered machines
by which the walls could be climbed, and filled the moat with stones and earth. On this, the defenders gave up.
Another chronicle refers to the use of a battering ram. Documentary sources show that the king had thirteen
throwing engines. A double team of carpenters worked for most of the time of the siege constructing the
Warwolf. A letter reveals what happened. Once the machine was ready, the garrison offered surrender. Edward
was, however, determined to try his new toy; no one was to come in or go out of the castle for 24 hours. The
Warwolf was tried out; the defeated garrison were allowed out of the castle. There is a reference to an oriel
window or gallery constructed in the town, so that the ladies of the court could watch the progress of the siege.
The rock on which the castle is built made mining impossible, and access was not easy for a belfry – that
required flat ground. Just about everything else that could be tried, was tried. What the Warwolf really was,
goodness knows – but a huge trebuchet does seem the most likely.

Bamburgh, besieged by the Yorkists in 1464. There, Ralph Grey held out. He and the garrison were threatened.
Because King Edward IV did not want to see such an important castle damaged by gunfire, Grey would lose his
head if it proved necessary to bring the bombards into action, and every shot fired would mean that one more
head would roll. The threat did not have its desired effect, and the guns were brought into action. Dijon
apparently sent its projectiles through Grey’s chamber; he was wounded by falling masonry. Eventually the castle
surrendered, and Grey was duly beheaded.

3. What happened to the prisoners at the end of a siege? At Bedford the king took the extreme measure, and
hanged the lot. He had the support of Langton and de Burgh in this. The king was in the right in legal terms.
Caerlaverock is different. The Song has it that the defeated garrison was treated honourably, granted life and
limb by the king, and given a clean shirt each. The Lanercost chronicle, on the other hand, has it that many
were hanged. Another source explains that the garrison did offer surrender, but only on condition that they were
granted life and limb, and their possessions. This angered the king, who promptly ordered the assault. Record
evidence shows that the constable, Walter Benchafe, and twenty-one of the garrison were sent to prisons in
England. As for Stirling, it seems that Edward made dire threats of dismemberment and hanging, but in fact
none of the garrison was hanged, apart from one man who had previously betrayed the castle to the Scots. The
constable, William Olifant, was sent to the Tower, and the others to different prisons in England.

Cases of the execution of defeated garrisons are not common. In 1138 Stephen executed all the members of the
Shrewsbury garrison, including Arnold of Hesdin, their commander, in a wholly untypical act of brutality. In 1153
the future Henry II executed just the archers in the garrison of Crowmarsh Gifford, pardoning the knights.
Richard I had all but one of the garrison of Chaluz-Chabrol executed in 1199 – they had refused to agree
surrender terms, and the castle had to be stormed. At Kenilworth in 1266 the garrison were allowed the terms
of the previously agreed Dictum of Kenilworth, under which men were allowed to buy their way back into favour
by paying fines assessed on the degree of their involvement in the rebellion, paying so many times the annual
value of their land. It marked a real change when the garrison of Leeds castle in Kent was executed in 1322 by
Edward II.

Far more common are cases where a garrison was reasonably well treated. When David I of Scotland took Wark
in 1138, he even gave the garrison fresh horses to replace those which they had eaten. In 1136 King Stephen
swore that he would kill the garrison of Exeter when he took the castle. In the event, he held back from such
extreme action. He has sometimes been accused of weak-mindedness in this; in practice, he was taking the
normal line adopted by monarchs. It probably did Stephen more good to pardon the defenders of Exeter than it
did to execute those of Shrewsbury. Edward I’s treatment of the Scots is interesting. I have discussed
Caerlaverock and Stirling; all the other examples of the capture of castles up to the Bruce rebellion show the king
treating defeated garrisons with clemency. After Bruce rebelled, however, it was a different story. A bloodbath
followed, and when castles fell, such as Kildrummy, execution was the order of the day.

How important were the laws of war? Was common sense the order of the day? John had wanted to execute the
prisoners taken in the siege of Rochester; but Savari de Mauleon protested. ‘My lord, our war is not yet over.
Therefore you ought to carefully consider how the fortunes of war may turn; for if you now order us to hand
these men, the barons, our enemies, will perhaps in a similar event take me or other nobles of your army, and
following your example, hang us; therefore do not let this happen, for in such a case no one will fight in your
cause.’ The clemency showed by Henry II in the rebellion of 1173-4, or that shown by the victors in the civil wars
of the 1260, and the siege of Kenilworth in particular, was not weakness, but good common sense and even
statesmanship.

4. The Scottish revolution: surprise attack and escalade. In 1307 or so, James Douglas disguised his men as
merchants, and made them ride past the English held Douglas castle. Some of the garrison, seeing what they
thought was easy booty, rushed out. The Scots put off their disguise, and killed them. Following this success,
Douglas took the castle. At Forfar, a local forester scaled the castle with ladders, and took the castle because
the guards were unprepared. At Perth, in 1313, there was a proper siege. After six weeks the Scots packed up
and departed. But Bruce had scaling ladders made, and a week later returned by night. He and his men waded
the moat, up to their necks in water in January, got over the walls, and took the place totally by surprise. At
Linlithgow a peasant called William Bunnock regularly took hay to the castle. One day he put eight men, well
armed, under the hay, and himself walked nonchalantly beside the cart. The cart’s driver carried a hatchet, and
once the gate was opened, and the cart blocking it, he cut the traces. The men jumped out from under the
hay, and started their work of slaughter. Some of the garrison had gone out to harvest grain – they were soon
killed by other Scots waiting in ambush. Linlithgow was then demolished on Bruce’s instructions. At Roxburgh
Douglas had rope ladders made, with wooden steps tied into them, and a strong square iron hook that would fit
into an embrasure. He and his men came up at night, wearing dark cloaks over their armour, crawling on hands
and feet and pretending that they were oxen. They set up the ladders, killed a guard, and then slaughtered most
of those in the castle, when they found them carousing in the hall. The constable managed to escape into the
keep where he and a few men held out; their honourable surrender was soon negotiated. At Edinburgh the Scots
found a local man, William Francis, who had been in the garrison as a youth. He had then found a secret way
down the castle rock so that he could go and visit his girlfriend – he now offered to show the Scots this route. He
led a group of about thirty up. They carried a short scaling ladder which they needed for the low wall at the top
of the rock, and so made their way into the castle. Piers Lubaud, the Gascon commander of the garrison, then
changed sides and joined the Scottish cause.

Such techniques of surprise attack, escalade and trickery were not, of course, entirely novel. There was the case
of the very small Welshman who had entered Cardiff castle by surprise in the twelfth century. There were the
sexual wiles used by Robert FitzHildebrand in Stephen’s reign – he won Portchester castle by seducing the
constable’s wife. Robert FitzHubert in the same period used leather scaling ladders at Devizes. But such
consistent use of surprise tactics by the Scots was novel.

Later Medieval Castles – Architecture

The courtyard castle. In this, accommodation and defensive walls would be fully integrated – no longer any
question of lean-to’s built up against the walls. Goodrich is an early example from the later thirteenth century,
presaging future development. William of Caverswall had a licence to crenellate in 1275, and he built a
quadrangular castle with octagonal corner towers, with a pair forming a gate, at Caverswell (Staffs). Maxstoke
in Northamptonshire is a good example from the mid fourteenth century  – unusually, its towers are multangular
not round. The classic late medieval castle is Bodiam, another of this type. It has a wide moat, and the full
paraphernalia of towers and gatehouse. Wingfield, built by Michael de la Pole in the 1380s, is another
quadrangular moated site, with an impressive gatehouse. From the fifteenth century there is Herstmonceux in
Sussex, again rectangular, rising from a moat, with gatehouse and corner towers, as well as additional turrets or
towers along the walls. This contained not one, but two courtyards, for within it there is a cross-range. Kirby
Muxloe in Leicestershire was unfinished at the death of Lord Hastings, its builder, in 1483. It is built of brick,
not stone, and has the usual quadrilateral plan. The corner towers are square; the gatehouse towers
multangular. There are examples in the North as well. Bolton in Wensleydale is a massively formidable
rectangular courtyard castle, its square corner towers not projecting far from the curtain walls. Lumley Castle is
another. Smaller, but again of similar plan is Danby in Cleveland. Sheriff Hutton in North Yorkshire was built by
the Nevilles in the late fourteenth century – again, it was a quadrilateral courtyard type castle.

The domestic requirement for these castles were, clearly, first of all great hall, with a passage separating it
from the pantry, buttery and other offices. There would be a kitchen, and a chapel. Store rooms were also
needed. All of this could be provided in an orthodox, old fashioned hall and solar type plan, but what the
courtyard castles also provided was extensive ranges of what Thompson calls ‘cellular lodgings’, bed-sitting rooms
or small suites of rooms. It might also be possible to double-up some rooms, and have, for example, two halls.
The courtyard plan made good sense in that it allowed a high level of accommodation to be provided in an
integrated way within a relatively small overall envelope. Caister was a courtyard castle built by Sir John Fastolf
in the fifteenth century. It had one striking tall round tower. An inventory enables the rooms to be identified.
There were summer and winter halls, with chambers above each. Kitchen, buttery, wardrobe, larder and cellar
were among what might be called the working rooms. There was a chapel. Then there were lots of chambers,
with names such as Inglose’s chamber, Porter’s chamber, Chamber of Dame Millecent Fastolf, Chamber of John
Boteler. In all, fifty rooms can be identified. The courtyard plan made good sense as a way of providing rooms
on this scale.

The great tower. In the early fourteenth century Edward II built a keep of sorts at Knaresborough. At Dudley Sir
John Somery built a rectangular two storied keep, which has round towers on the corners. Later in the century
Sir John de la Mare built a rather similar structure at Nunney in Somerset, a sort of rectangular keep with four
round towers. At the shorter sides the towers almost meet. At Durham bishop Hatfield made use of the Norman
motte, and built a keep on it which I suspect may have been in emulation of that at Windsor (which is probably
twelfth century in origin). From the fifteenth century we have Ralph Cromwell’s extraordinary brick tower at
Tatershall in Lincolnshire. Five stories in height, this was a medieval skyscraper. It was not, however, a
self-contained building like a twelfth-century keep  –  it was an integrated part of a castle the rest of which has
been lost. Ashby-de-la-Zouche in Leicestershire is a four-storied stone tower, with a section to one side having
seven stories, so as to provide more chambers. Unlike Tatershall, this was intended as a self-contained building,
even having its own well. It is fashionable to denigrate the military role of these fifteenth-century castles, but
Raglan in the Welsh marches has an extraordinary powerful octagonal keep, separate from the rest of the castle
which provided all the chambers etc. This ‘Yellow Tower of Gwent’ was a massively powerful military structure.
In the north there is Warkworth. This is best described as a square, with subsidiary squares placed half way along
each side. The design thus forms a sort of cruciform shape. It is unique, and provides a full set of self-contained
accommodation. Store rooms in the basement, grand staircase, hall, kitchen, chapel on first floor, chambers on
the second. It stands on an old motte; there is a sense in which the site demanded some grand treatment.

Additions to existing castles At Kenilworth John of Gaunt built a splendid hall with associated chambers etc. At
Warwick, the earls created a splendid grand front. A barbican thrust forward from the line of the curtain walls,
and two unique towers stood at either end of the curtain. Caesar’s Tower was a five storied lobed affair, with a
machicolated gallery – it was a sort of clover-leaf design. Guy’s Tower, at the other end of the curtain, was a
tall multangular effort, again five stories high, all vaulted. At Windsor, Edward III spent lavishly, largely on
complex suites of chambers occupying the bailey of the original twelfth-century castle.

French influence. Most of these castles were built by men who had long experience of the French war. The
builders of the castles of the late fourteenth century, men like de la Mare and Dalyngrigge, builder of Bodiam,
were all experienced soldiers. Fastolf in the next century was a highly experienced commander. The Scropes
(Bolton) were a legal family, and Lord Cromwell was the king’s treasurer – not a soldier. But the majority of the
builders were presumably familiar with the castles of Brittany, Normandy, and South-West France, and may well
have wanted to build in French style when they returned home. Caesar’s Tower at Warwick is pure French in
conception. The twelfth-century donjon at Etampes was four-lobed in design; the chemin-de-ronde, or
machicolated gallery around the tower finds many French, but no English, parallels. The Yellow Tower of Gwent
has a resemblance to Largoët-en-Elven in Brittany, though it does not match it in height. Nunney again has a
very French feel to it.

Just how military were these castles? Charles Coulson argues, in a series of dense and combative papers, that
some of them were really just for show, and not intended for serious conflict any more than, say, the
nineteenth-century pastiche castles of Scottish Baronial style. This is not altogether new. Hearne in the
eighteenth century commented on Wingfield: ‘Upon the whole, this Fabrick seems to have been formed rather
to inspire the idea of dignity, and to oppose a popular tumult, than to resist an enemy’. Bodiam, built in Richard
II’s reign, was Coulson’s prime example. Here we have the classic picture-postcard castle. It has, at least
superficially, all the right elements – wet moat, gatehouse, towers, battlements, arrow-loops and, befitting its
late fourteenth-century date, gunloops. The towers provide flanking fire along the curtain walls. The castle was
licensed to Sir Edmund Dalyngrigge, and it states that the castle was built in defence of the country.

The licence suggests that the castle was built near the sea, for the defence of the adjacent country, and for
resisting the king’s enemies. This will have been Dallingrigge’s explanation when he sought the licence, and does
not represent royal command.

The standard windows at Bodiam were barred and had shutters, but were clearly windows, not defensive
arrowloops. The chapel windows, and the windows of the lord’s chamber, are large. There was no attempt to
place the windows so as to make flanking fire possible. There is a postern gate, with no indication of any
massive beam that could be placed across to close it in time of need. The evidence suggests that doors and
shutters were of domestic quality, not military, fastened with ordinary sliding bolts. The main gate is strong
enough, with three portcullises, strong outer doors, etc., but Coulson argues that ‘There is no “military” logic in
trebling the main closures while leaving a short and direct approach to a weak back door’.

Bodiam is crenellated, and its gates are in addition machicolated in up-to-date fashion. There is little problem
here – these are effective military elements. Certainly it would have been better to extend the machicolation
further, but all Coulson can say is that ‘It all “doth protest too much” with splendid mendacity’. The merlons are
a little low, and the design is adjusted so that they fit the length of the wall. Coulson suggests that it was done
so that Dalyngrigge could man the walls, to impress visitors. The gunloops are typical of the period. They would
not have been easy to use – very little deflection was possible, so aiming was very hard if not impossible. The
design of the loops, says Coulson, ‘show that a display of aggression, not real vindictiveness, was in the
designer’s mind’. And he suggests a ‘saluting base whence to greet honoured guests with his own personal
Crakkys of Warre firing ‘blank’.

The moat was emptied in 1970, by about a dozen workmen in a day. So, the moat could easily have been
drained in the medieval period. The site, argues Coulson, was skilfully chosen – there was probably a natural pool
there. The castle was ingeniously constructed so that the walls would appear to rise directly out of the moat –
there is a sloping plinth, but it lies underwater. The purpose of the moat, argues Coulson, was to create an
impressive entrance to the castle, going sideways along a long wooden bridge to an octagonal stone island,
before turning into the castle itself. ‘At the outer barriers, having negotiated the circuitous water-girt
roadways, a guest of the greatest honour would be received by the Lord; conducted past the specious challenge
of Dallingrigge tenantry and deferential domestics, via the Octagon and Barbican, each point doubtless the
occasion of ceremony.’ So, the purpose of the moat was ceremonial, not military.

Coulson goes on to suggest that at Bodiam, ‘Its covert contradictions are so many coded disclaimers of undue
social pretension.’ In other words, Dalyngrigge was anxious not to offend his social superiors, and so inserted in
his castles signs of weakness. At the same time, he wished to display his wealth and status with a grand building.

Maxstoke was built in the 1340s by William Clinton, earl of Huntingdon. It lies within a moat, and looks
impressive. But it has weak features. Unlike Bodiam, it does not rise direct from the water, but has a ‘berm’,
ground outside the walls but within the moat – assailants could land and scale the walls. The gatehouse is tall and
slender, with a couple of fine windows in the central portion. Coulson has an odd discussion of the crenellations –
the merlon has an exaggerated coping, but there is none to the crenel. Grooves were fitted for shutters to the
crenels. This is ‘suspiciously crude and over-scale as though purposely conspicuous from outside’. Coulson does
not mention the very fine chapel window, which conspicuously weakens the curtain wall. The entrance was well
defended, with ‘murder holes’, double-leaved door and portcullis. Again, there is the suggestion that the
parvenu was deliberately not building something too grand, and that Clinton was not wishing to be seen to be
challenging the might of nearby Warwick.

It cannot be the case that the designers of Maxstoke and Bodiam simply did not know their job, that they were
too unintelligent to realise that to pierce the curtain wall with large chapel windows was to weaken the fabric.
Were the military elements at these castles nothing more than stylish features? They do look more than that –
the gunports at Bodiam have genuine beds behind them on which guns could be mounted. The gatehouse
defences at both castles, gates, portcullises and murder-holes are real. Men such as Clinton and Dalyngrigge,
experienced soldiers, would surely have been uncomfortable if such features had just been fakes. Was it perhaps
that they realised that full-scale defences were simply not needed in peaceable England?

Bolton in Wensleydale is a courtyard castle, but very different in aspect from Bodiam. Licenced in 1379, it was of
very much the same date. No moat, towers barely projecting from the walls, gate consisting of a passage next
to one of the towers – this is from a very different tradition. Externally it has something of the look of such a
keep. But, of course, no keep would have a courtyard within it. It is a particularly stark building externally. As
usual, chapel windows break the solidity of the walls, but otherwise fenestration was very limited, and only at
the upper levels. It is also limited on the courtyard side, and at ground-floor level the different sides of the
courtyard are not interconnected. The building is massive – the towers are five stories high. There was no means
of keeping enemies at a distance, for there is no indication of an outer bailey. There were no fashionable
gunloops, and the treatment of the wall-head is not clear.

Tatershall, built by Ralph lord Cromwell, is hard to assess, since all we have is the great tower, and not the rest
of the castle of which it was an integrated part. There was a castle on the site, licenced in 1231, probably an
irregular polygon with corner towers. It was demolished by Cromwell in the 1430s, and he built his new castle in
brick, not stone. The great tower stands five stories high. It has octagonal corner turrets, battlements and
impressive machicolation. At the same time, it features ground-floor entry, and has large windows. The
machicolations and windows are in stone. This was not an independent building. Immediately in front of the
tower was a hall, and the ground-floor entry doors were in fact interior doors, leading from the hall into the
tower, which was in effect a very grand chamber-block. Thompson talks in terms of ‘fantasy and
self-dramatisation’. This is surely a building in castellated style, in which the military elements are not serious at
all. It has no gun-loops or arrow-loops, save for single loops at the top of the corner-turrets. The parapets to the
turrets, intrerestingly, have a sort of fake machicolation.

Caister, John Fastolf’s castle built in the first half of the fifteenth century, is a classic courtyard castle, with
elaborate fully integrated domestic accommodation. It stands within a wet moat, rising directly from the water
like Bodiam. There is one remarkable tall tower, containing five chambers, each with window, fireplace and
latrine. The walls are thin. The castle has no substantial flanking towers, and the gate is relatively simple.
Fenestration for the hall is generous. There are gunports, however, and a curious isolated flat-topped tower may
have been designed to take artillery. It does not look, from what remains, particularly formidable. Yet Caister
underwent a siege for a month in 1469. It did not in the end hold out ‘Your brother and his fellowship stand in
great jeopardy at Caister, and lack of victuals, loss of men failed in gunpowder and arrows, and the place sore
broked with guns of the other party’, wrote Margaret Paston. Nevertheless, it did well, and makes one think
that these apparently not very military castles were a good deal more formidable than some commentators
assume.

Raglan, built in the second half of the fifteenth century, is puzzling. The main part of the castle is martial
enough in appearance with heavy machicolated towers, gunloops and the rest. But it is also lavishly fitted with
windows. This is a courtyard house, with inner and outer courtyards. Then there is the great Yellow Tower of
Gwent. It is hexagonal, and stood four stories high. It stood proud of the castle, defending it with gunports at
ground level, firing in every direction. The walls are massive, eleven feet thick, but have wide and elaborate
window openings. Despite this, this was surely a military building par excellence. One suggestion (Thompson) is
that it was a kind of bastion or bastille, built outside the main castle as a means of protecting it from attack.

Kirby Muxloe and Ashby deserve attention, as the very last castles of our period. Ashby is a massive stone tower,
with effective machicolation, the potential for independent defence with its own well, and an elaborate provision
of chambers and rooms. Its entrance, however, is on the ground floor – but it has provision for a portcullis.
Windows made it comfortable, but would have been a problem in a siege situation. The building appears
something of a schizophrenic mess, but in practice it combined ample domestic accommodation with all the
military strength that Hastings required. Kirby Muxloe was a complete contrast – a brick courtyard castle, begun
in 1480. This has the right elements – wet moat from which the castle walls rise directly, twin-towered gate,
flanking corner towers (square, though, not round). It is well-equipped with gunports. Some of them are directly
aimed at other parts of the castle, so providing flanking fire along the unfinished curtain walls. The walls were
only about four feet thick, not sufficient to deal with a serious siege. Windows in the gatehouse and the one
tower to have been built again belie the martial image. This is more a country-house in castellated style than a
true castle.