Clause 28 was in all important respects unchanged from number 18 of the Articles of the Barons, and in both documents it appeared as the first in a small group of clauses concerned to regulate the activities of royal officials, with specific reference being made in two of them to the constables who controlled the king’s castles. Those castles were fundamental to the governance of England under the Angevin kings, for all of whom they were, in William of Newburgh’s striking expression, `the bones of the kingdom’,1 but they took on a particular importance under King John. In the years around 1215 John held about seventy castles in England alone (their number fluctuated somewhat, since forfeitures or the deaths of bishops or tenants-in-chief might temporarily add to them, as happened, for instance, after the death in April 1208 of Philip of Poitou, bishop of Durham, which gave John control of Norham Castle, a major fortress on Northumberland’s border with Scotland, for several years).2 For reasons both military and administrative castles bulked large in his style of government, as instruments of control and administrative bases, as royal residences, and in some cases also as major depositories for stores of cash. The official who received custody of one might have to give hostages to the distrustful king,3 who possessed silver-gilt basins decorated with castles.4 The maintenance of both their buildings and their occupants was the king’s responsibility, and as Clause 28 shows, it had come to be achieved, at least in part, through the exercise of the king’s right of purveyance.
Traditionally, purveyance enabled the king to maintain his own peripatetic household by requisitioning the supplies it needed from the communities among which it passed, and making payment later for what it took. It was a specifically royal privilege, and not one to which any baron was entitled.5 The grant which John FitzRobert (one of the twenty-five barons appointed to enforce Magna Carta) made to the burgesses of Corbridge, probably some time after 1215, in which he forbade any sergeant residing there to take any purveyance (prisam) from the townspeople against their will, almost certainly reflected the particular conditions in a town which had been a royal manor until King John granted it at feefarm to John FitzRobert’s father in 1205.6 In all likelihood either John FitzRobert or his officers had attempted to continue to exploit the king’s right of purveyance, and thereby helped to generate the controversia which the grant was intended to allay. But although it was tolerated because it seemed unavoidable, purveyance was always deeply unpopular, as Clause 28 and John FitzRobert’s charter alike demonstrate. On later evidence, it was a ready source of corruption – goods might be taken when there was in fact no need for them, or bribes extracted for having goods returned, or simply for being left unmolested - while payments were all too often made slowly if they were made at all.
Henry I’s attempt to control the exploitation of purveyance, in regulations probably issued in 1108,7 does not seem to have been more than partly successful, and indeed its use, as a means of providing government with subsistence while deferring payment for it, appears to have expanded as government did, being employed by sheriffs and other royal servants as well as by the king’s household. Misuse of purveyance may have been one of the malpractices which lay behind Article 10 of the inquest of sheriffs of 1170, which ordered inquiry `whether the sheriffs or any of their bailiffs, or the lords of the vills or their bailiffs, have restored any of the things which they have taken, or have made any peace with their men, since they have heard of the coming of the lord king, in order to prevent any complaint thereof reaching the lord king or his justices.’8
Magna Carta was silent about purveyance as it was customarily taken for the benefit of the king’s household, but its being exploited, and abused, by royal officials as well was a different matter. Clause 28 does not in fact represent the first attempt to prevent this, though the fact that the same noun – prisa – was, tellingly, used to define both purveyance as such, and also extortion in general, means that it is sometimes difficult to tell whether it is the king’s privilege or the misdeeds of his officials which is being referred to, as it is, for instance, in proceedings involving Ranulf de Glanville in 1176/7, when acting de propria prisa, and abetted by the prisa of his six sergeants, he was found to have exploited his control of Westmorland to take valuables, livestock, grain and timber worth £1571. 1s. 1d.9 There is similar uncertainty concerning one of the articles of the eyres of 1194/5, which originally instructed royal justices to inquire `de prisis et tenseriis [protection money] of all the king’s bailiffs, both justices and sheriffs, constables, foresters and their servants’ since 1189.10 Archbishop Hubert Walter, in his capacity as justiciar, put a stop to this inquiry, and it remains unclear what its target was, but there can be no doubt that the practice of taking of goods to stock castles had become a regular one, for on a few occasions John felt it necessary to issue charters exempting their beneficiaries from it. In 1199 his confirmation charters for the knights of the Temple and the Hospital ordered that neither `their own grain nor that of their men is to be taken to provision castles’,11 and in 1205 he granted the same privilege to his own foundation of Beaulieu Abbey12 (another charter containing this clause, on behalf of the Cistercians, was cancelled, probably as a result of the recent death of Archbishop Hubert).13
The abuse of purveyance, and a response to it, can be no less clearly seen in an addition which was made to the articles of the eyre in 1208, concerning `Purveyances made (de prisis factis) by sheriffs or constable(s) or by any bailiff(s), against the will of those whose chattels were taken’. Moreover this was almost certainly accompanied by another new article, `Concerning those who take bribes so that corn and other goods are not taken for provisioning castles’ - payment to avoid purveyance was the inevitable corollary of its usage.14 Ambiguities remained, but it may not have been coincidence that in the same year that the new articles were introduced, Walter Clifford, the recently-dismissed sheriff of Herefordshire, proffered 1000 marks (£666. 13s. 4d.) to recover the king’s good will and to avoid any inquest being made de prisis suis in that county. He was pardoned 100 marks, but paid the rest within a year.15
John spent over £17,000 on castles during his reign.16 He built two new ones, at Hanley and Odiham, but most of his outlay involved repairs and extensions to existing fortresses. Only once in the first eight years of his reign did recorded expenditure exceed £1000 (in 1204/5, no doubt reflecting the danger of invasion after the loss of Normandy), but thereafter his outlay increased sharply, with a notable emphasis on the fortresses of the north of England. Exact figures are unavailable, but the surviving records show that he spent at least £1616. 4s. 11½d. on castles in 1209/10, £2892. 18s. 3½d. in 1210/11, and £2201. 1s. 3d. in 1211/12. The fall to £632. 8s. 5d. in 1213/14 may reflect financial difficulties, or perhaps just different priorities, as John devoted ever more of his resources to his forthcoming campaign in France, but it is also possible that the English castles were now regarded as strong enough to serve their various purposes without much more money being spent on them.
Those castles would have been useless, however, without garrisons inside them. These seldom appear to have been large, except in times of imminent danger, and even then they might be smaller than expected. Following the death of Richard I, which saw a serious outbreak of disorder, and no doubt a reinforcement of garrisons, Tickhill Castle in Yorkshire was recorded as having been occupied by just five knights and ten sergeants, along with two smiths and six carpenters,17 while in 1216 the newly-built but decidedly minor castle of Odiham, Hampshire, famously held out for a full week against the French forces besieging it, even though it was manned by only three knights and ten sergeants.18 Two years earlier the same castle was recorded as being held by four knights and eighteen sergeants, still a modest number, while the far more important castle of Windsor was garrisoned by, at the most, ten knights and twenty sergeants, and even the Tower of London was manned at first by only ten knights, twenty sergeants and four crossbowmen, numbers later increased to fifteen, twenty-four and seven respectively, forty-six soldiers in all.19
But although the number of fighting men in a castle may not always have been very great, soldiers will seldom have been its only occupants. There must always have been resident watchmen, doorkeepers and others responsible for upkeep and security, and these men will have been more numerous in the larger fortresses, and especially those which were also in intermittent use as royal palaces, like Corfe or Nottingham. The arrival of the king and his retinue would then have augmented further the number of mouths to be fed. In a county town, even one which the king did not visit in person, if it held a castle which was also the head-quarters of the sheriff, then the latter’s officers and staff would likewise have needed accommodation and upkeep. Prisoners in a county gaol were expected to live off their own money and goods, but John’s suspicious regime was also apt to bring hostages into castles for safe-keeping, and these could not have been maintained thus. In March 1213 Robert de Vieuxpont was ordered to hand over Salisbury Castle to Arnold of Auckland `with the prisoners and hostages which are in it’,20 and the release of hostages commanded by Clause 49 of Magna Carta must have eased the pressure on space and supplies in many English castles – Hereford, for instance, where six hostages were ordered to be released in August 1215.21 Finally there were non-human residents in the form of horses. Supplies of oats, especially for the better animals, and also of hay must always have been in demand to feed the horses ridden by messengers and servants as well as by knights, and though many castles probably had demesne meadows which could usually meet the demand for hay, it is likely that the oats recorded as purchased for garrisons – 300 quarters bought for Lancaster Castle in 1215, for instance22 - were required less for the men than for their mounts.
Supplying a castle was thus likely to be an expensive business even in peacetime, and more so in times of war or disturbance. In 1199, probably in response to the disorder which followed the death of Richard I, Launceston Castle in Cornwall was held for thirteen days by a force made up of five knights, twelve mounted serjeants and twenty serjeants serving on foot, thirty-seven men with, presumably, at least seventeen horses. To supply men and mounts for less than a fortnight, a total of £18. 15s. were spent on eighteen sides of bacon, ten horseloads of corn, eighteen of rye, fifty of oats (most of these were probably intended for the horses), four quarters of salt, a barrel of wine, unspecified quantities of firewood, coal, hay and ale, and also four doublets (purpuinz).23 It was probably usual for stocks of victuals to be built up in times of emergency, then reduced once the crisis was over. How large they could be in a major fortress is suggested by the sales of supplies from Nottingham Castle, altogether worth at least £145, recorded in 1194 – they must have been gathered ahead of the siege of the castle earlier in the year, but once the castle had fallen to King Richard most of them were no longer needed and were therefore sold off. The foodstuffs disposed of included 229 quarters of corn, 297 quarters of rye, 1180 sides of bacon and 400 cheeses.24
Since Nottingham Castle was held against the king by supporters of his brother, it seems likely that its garrison simply took provisions as and where they could be found, without any pretence that they were being commandeered as a lawful act of purveyance. But even when legitimately carried out, purveyance seems to have been potentially problematic. Sheriffs and constables (often the same, especially when the sheriff had his office in a castle situated in a county town) who had taken goods against a promise of future repayment could offset their subsequent disbursements against the revenues for which they accounted at the exchequer. In this there were similarities to the way in which the financing of work on the king’s buildings was organised,25 but also one important difference. In both cases the sheriff or constable needed to show a royal writ authorising his action, but whereas for buildings he also had to have the supporting testimony of two or three `law-worthy men’ who were required to accompany him to the exchequer and there swear that the money had indeed been spent as the king ordered and the sheriff claimed, after which the latter was allowed the sum in question on his account, for purveyances the authority of a royal writ was usually all that was necessary. There were a few exceptions. In 1193, for instance, the sheriffs of London successfully claimed allowance of £52. 14s. 10d. spent on stocking the Tower with wheat, oats, cheese and wine, `by the king’s writ and by view of John son of Erlecon and Alexander son of Sperling ...’,26 and in 1215 the sheriff of Northamptonshire spent nearly £50 on grain and ham for Northampton Castle `by royal writ and by view of Philip Malebranche and Robert of Bugbrooke’.27 But far more often it was only works which were guaranteed per visum, while supplies were said to have been paid for on the sole authority of a royal writ, suggesting that purveyance was usually carried out under little or no external supervision, and so with a greater danger of corruption and abuse.
In discussing purveyance, it is seldom possible to be certain when it was used to provision castles. Exchequer records which sometimes list the goods being paid for as well as their value hardly ever state categorically that the expenditure being accounted for was disbursed in payment for previously-requisitioned goods, although it is very likely that this was indeed often the case. Clause 28 itself, along with earlier measures like the 1208 addition to the articles of the eyre, constitute the strongest evidence that purveyance was frequently resorted to, but supporting detail is largely absent. The issue is further complicated by the fact that there were other ways in which castles could be provided with food, drink and other necessary supplies, most of them recorded in the last years of John’s reign, though it would be rash to assume that none was ever employed earlier. One such was simply to sell existing stocks of food in a castle, presumably because they were in process of decay, and use the proceeds to buy new ones. The sale of meat, grain and other goods at Knaresborough Castle in 1214, including grain that may have been three years old, raised £51. 3s. 3d., more than enough for the oxen, horses, beans and herrings which were then purchased.28 In May 1215 Norwich Castle was to be provisioned with meat and grain from the lands of the bishopric, which was vacant at the time,29 while a year later it was corn from a rebel’s land which was to be stored in Corfe Castle.30 At Winchester cash owed from the city’s fee-farm was to be used in 1215 to finance the stocking of the castle,31 as it was also at Gloucester in the following year, where the sheriff was also ordered to bring in `the corn of all the knights of your county whose lands are within ten miles’ distance of Gloucester ...’.32 In 1216 the royal manor of Kingsthorpe, Northamptonshire, was entrusted to the sheriff, so that its issues could be applied to the keeping of Northampton Castle,33 and in the same year Hubert de Burgh was ordered to let Cecily d’Avranches have all her son’s land `except the land which he finds it necessary to have for the stocking [ad garnisturam] of Dover Castle.’34 By then Rockingham Castle was being maintained from a castlery (castellaria) made up of four royal manors and a number of wardships.35
Cash could be paid where it was available, or when its use seemed especially appropriate. On 25 July 1215, for instance, Sherborne Castle in Dorset was entrusted to John Marshal, whose keeper was instructed to hand it over `with all the provisions which you have had bought with our money, and if any money is left over from that assigned to the purchase of provisions, it is to be handed over to the same [John] ...’.36 But at a time when the supply of coin was limited, not least because John’s ever-increasing demands on his subjects caused large quantities to be taken out of circulation, either stored in his castle treasuries or shipped overseas for payment to his continental allies, purveyance must often have been the only recourse available to sheriffs or constables who needed to man and equip the castles in their keeping. By 1215, as one might expect, the scanty evidence indicates that the garrisons of royal castles, and therefore the quantities of provisions needed to maintain them, were growing considerably. According to the Barnwell chronicle, John responded to the baronial occupation of London on 17 May by strengthening his strongpoints (praesidia) and castles with men and victuals.37 Figures for the men are scarce, though it is probably indicative of the mounting crisis that when John appointed a new constable for Scarborough Castle in March 1215, he ordered the two existing ones to come to him, and instructed them to leave the castle manned by sixty sergeants and ten crossbowmen,38 while soon afterwards it was reinforced by ten knights as well as by thirteen more sergeants and another three crossbowmen;39 it seems safe to say that this was a considerably larger body of troops than had held the castle a year or two previously.
There is better evidence for the victuals, though John seems to have started adding to them somewhat earlier than the chronicler reported. In April 1215 John had ordered Henry of Braybrooke, sheriff of Northamptonshire, to supplement the provisions already in Northampton Castle, so that the forty quarters of wheat already in store there were increased to 200, and the number of sides of bacon brought up from eighty to 300.40 The sheriff obeyed his instructions, and indeed, he did rather more than that, for at a cost of £49. 11s. 1d. he bought 220 sides of bacon, 103 quarters of wheat and 136 quarters of oats. As a result he had considerably more than was needed, or perhaps more than the castle granary could hold, for he soon had to sell grain worth £21. 18s. 2d.; he had been ordered to pay 2s. per quarter, suggesting that he sold all but about twenty quarters of the 240 he had only just purchased.41 Other castles were restocked at the same time. At Lancaster the sheriff spent a total of £162. 10s. 10d. on wheat, malt, sides of bacon, oats, herrings and salt, along with eighty cows and 130 sheep, all of them `on the hoof’, and he also laid out £5 on 10,000 crossbow bolts,42 suggesting that he was taking precautions against a possible siege, though none took place. At Wallingford, which stood in greater danger, the sheriff spent only £10 on 100 horseloads (summis) of wheat and £18. 18s. on `186 pigs bought to stock Wallingford Castle’,43 but that castle may well have been better maintained than Lancaster (the latter also needed repairs costing £33. 18s. 4½d). Expenditure on works on castles was rarely significant during this year, but £50. 15s. 11d. were spent on Knaresborough, and no less than £402. 2s. on Kenilworth.44
According to the well-informed `Anonymous of Béthune’, in 1216 the vital fortress of Dover was manned by 140 knights and a very large number (moult grant plenté) of sergeants.45 But apart from the enlarged force which manned Scarborough Castle in 1215, and the thirteen who held Odiham, there is otherwise only meagre evidence for the size of the garrisons which held King John’s castles during the civil war at the end of his reign. A little more can be said about the castles which were held against him. These must often been small – the Barnwell Chronicle referred to rebels building `new fortlets’ (novas munitiunculas) in the summer of 121546 – but some were substantial. When Belvoir Castle in Leicestershire, a powerful fortress belonging to William d’Aubigné, surrendered to John in December 1215, thirteen knights and twenty-eight serjeants, forty-one men in all, were described as having come into the king’s allegiance.47 William himself was then a prisoner in the king’s hands, following his capture shortly before in Rochester Castle, after a prolonged siege in which he had led a baronial force which according to Roger of Wendover consisted of 140 knights `cum sequela sua’48 - the retinues of so many knights are likely to have amounted to a considerable force, and presumably contained the crossbowmen who were said to have killed many of the besiegers. Colchester Castle, a royal fortress which had recently been strengthened by John but later fell into baronial hands, to be held by a combination of English and French troops, contained at least 117 of the latter when it surrendered to the king in March 1216.49 It seems reasonable to suppose that the major royal castles, at least, were similarly well-manned when they faced an imminent siege, and they may have been better stocked with provisions than Rochester was. D’Aubigné and his followers were said by Wendover to have found the castle `destitute not only of weapons and food, but also of every kind of goods, except those which they had brought with them’, and to have been able to victual themselves only with what they could seize in the city. In a telling comment on the way castles under threat might have to be supplied, they did not have time to plunder the countryside round – praedam in provincia agere – before they were besieged. They would, in fact, have resorted to precisely those methods which Clause 28 condemned in order to maintain themselves when under attack. As it was, shortage of victuals eventually forced them to surrender.
Once civil war had broken out it was inevitable that both the king’s castellans and their rebel counterparts should have resorted to purveyance to stock their fortresses. When Engelard de Cigogné, who became constable of Windsor Castle in 1215, was sued by Alan the vintner of Reading in 1224 for wine he had allegedly taken `in time of peace, before the war’, he denied all misdoing in peacetime, but admitted that `in time of war, when he had Windsor Castle in his keeping, it may well be that he took some of his wine on the king’s behalf, to stock the castle, as he did with others ...’.50 John’s order of 23 July 1215, addressed to all the men of Yorkshire, ordered them to give up not only the lands and tenements, but also the castles and supplies (munitiones), which they had taken tempore guerre,51 suggests that as soon as hostilities began not only castles but also the supplies needed to provision them were seized, and since the order did not distinguish between the friends and foes of the king, that the men of both sides acted in the same way.
Magna Carta was drawn up before civil war broke out in earnest. No doubt Clause 28 was principally intended to remedy a burdensome abuse, but its likely repercussions upon a regime which was so dependent on castles for the maintenance and exercise of its power, and which also relied heavily on the services of mercenary soldiers, can hardly have escaped the barons who included it in the Charter, and in that respect it supplemented other clauses which were more overtly concerned to reduce John’s military strength, notably 50 and 51. Taking into account the assault made in the Charter on John’s methods of raising money, and the reduction in his revenues which resulted from it, it should have been much more difficult after 1215 for any king to try to dominate the country through garrisons maintained in his castles. Forcing him and his officers to pay cash down for supplies needed by soldiers manning his strongholds, while at the same time reducing the amount of cash available to them, should have been a simple way of clipping royal wings. There was, however, a potential flaw in Clause 28 which must have threatened to undermine its essential purpose, in its concession that payment for goods taken in purveyance might be deferred `with the consent of the seller’. In this half-hearted way the barons accepted that it was not only the king’s household which might have to be supplied through purveyance. And in doing so they left the way open for would-be purveyors to compel the necessary consent, by bringing to bear upon those whose goods they wished to requisition the force which they had at their disposal, that same force, indeed, which the proposed purveyance was intended to maintain. Unsurprisingly, therefore, misuse of purveyance remained a grievance for a long time to come.
William of Newburgh, Historia rerum Anglicarum, in R. Howlett (ed.), Chronicles of the reigns of Stephen, Henry II, and Richard I, 4 vols. (Rolls Series, 1884-9), i, 331.
Details from R.A. Brown, H.M. Colvin, A.J. Taylor, The history of the king’s works i: the middle ages, 2 vols. (1963), supplemented by R.A. Brown, `Royal castle-building in England, 1154-1216’, EHR 70 (1955), 353-98. Welsh castles are excluded from the figures given here.
T.D. Hardy (ed.), Rotuli litterarum patentium 1201-1216 (Record Commission, 1835 – hereafter Rot.Lit.Pat.), 34.
R.A. Brown (ed.), The memoranda roll for the tenth year of the reign of King John (1207-8), Pipe Roll Society new series 31 (1957), 125.
W.S. McKechnie, Magna Carta (2nd edn., Glasgow, 1914), 329-33.
M.T. Martin (ed.), The Percy cartulary, Surtees Society 117 (1911 for 1909), 283-5, H.H.E. Craster, Northumberland County History x: Corbridge (Newcastle, 1914), 59-61; the interpretation of this grant offered here differs somewhat from that of J.C. Holt, Magna Carta (2nd edn., Cambridge, 1992), 58-9. John FitzRobert’s grant also appears to refer to Clause 20 of Magna Carta.
J.A. Green, The government of England under Henry I (Cambridge, 1986), 21.
D.C. Douglas and G.W. Greenaway (eds.), English Historical Documents ii: 1042-1189 (2nd edn., 1981), 472.
PR 23 Henry II (1177), 81-2.
W. Stubbs (ed.), Chronica Rogeri de Houedene, 4 vols. (Rolls Series, 1868-71), iii, 267.
T.D. Hardy (ed.), Rotuli chartarum, 1199-1216 (Record Commission, 1837), 1-2, 15.
L. Landon (ed.), The cartae antiquae, rolls 1-10, Pipe Roll Society new series 17 (1938), no. 222 (pp. 109-111).
Rotuli chartarum, 154.
H.T. Riley (ed.), Munimenta Gildhallae Londoniensis, 3 vols. (Rolls Series, 1849-62), i, 118; C.A.F. Meekings and D. Crook (eds.), The 1235 Surrey eyre, 3 vols., Surrey Record Society 31-2, 37 (1979-2002), i, 92-3.
PR 10 John (1208), 191; PR 11 John (1209), 61.
Brown, art.cit. note 2 above, 356; the figures relating to expenditure are also taken from this article.
PR 8 John (1206), 78.
H.G. Hewlett (ed.), Rogeri de Wendover liber qui dicitur flores historiarum, 3 vols. (Rolls Series, 1886-9), ii, 182-3.
PR 16 John (1214), 124-5.
PR 17 John (1215), 56.
PR 1 John (1199), 242.
PR 6 Richard I (1194), 87.
E. Amt and S.D. Church (eds.), Dialogus de Scaccario (Oxford Medieval Texts, Oxford, 2007), 134-7.
PR 5 Richard I (1193), 158.
PR 17 John (1215), 49-50.
PR 16 John (1214), 66-7.
T.D. Hardy (ed.), Rotuli litterarum clausarum i, 1204-1224 (Record Commission, 1833 – hereafter Rot.Lit.Claus.i), 198.
Ib., 276, 288.
W. Stubbs (ed.), Memoriale fratris Walteri de Coventria, 2 vols. (Rolls Series, 1872-3), ii, 221.
PR 17 John (1215), 49-50, 51.
Ib., 13 (Knaresborough), 28 (Kenilworth).
F. Michel (ed.), Histoire des ducs de Normandie (Société de l’histoire de France, Paris, 1840), 170.
Memoriale Walteri de Coventria ii, 222.
Rogeri de Wendover ... flores historiarum, ii, 146.
Curia Regis Rolls xi, 1223-4 (1955), no. 2368 (pp. 469-70).
Clause 29 (The 1215 Magna Carta)
Clause 30 (The 1215 Magna Carta)
Clause 30 (The 1215 Magna Carta)
Please note: commentaries are presently available only for clauses marked with *; more commentary to be added in due course.