The Medieval Town Defences of Glamorgan

STUDIA CELTICA, XXXV (2001), 161–212

The Medieval Town Defences of Glamorgan

C. J. SPURGEON Royal Commission on the Ancient and Historical Monuments of Wales

Introduction

Nine boroughs were established within the bounds of the historic county of Glamorgan in the Middle Ages, two in the lordship of Gower (Loughor and Swansea), and seven in the lordship of Glamorgan (Aberafan, Caerphilly, Cardiff, Cowbridge, Kenfig, Llantrisant and Neath). This paper considers four boroughs which were certainly fortified, and one, Neath, for which there is sufficient evidence to suggest the possible line of its vanished defences. Vestiges of the town walls and a gate survive at Cardiff and Cowbridge, and at Swansea recent development has exposed a fragment of the town wall with traces of a possible tower. At these three towns the outline and characteristics of the masonry defences are well attested by early town plans and records, and largely confirmed by excavations. Kenfig was never provided with stone walls, but it still retains substantial traces of its defensive rampart and ditch. There is no documentary or phys- ical evidence for fortifications at the other four boroughs. Extensive excavations at Loughor found no trace of the suggested medieval refortification of the defences of the Roman fort.1 The nine medieval boroughs, with one exception, were founded by the chief lords of Gower or Glamorgan. The exception was the latest, Aberafan, established c.1304 by Leision ap Morgan Fychan, the Welsh lord of Afan. Eight boroughs were urban centres which developed beside castles. Cowbridge, however, was unique in Glamorgan and rare in Wales in not conforming to this pattern.2 The walled town was set on low ground half a kilometre from the pre-existing Llanblethian Castle, and its defences were probably provided at its foundation to compensate for this unusual lack of protection. In castle-towns which were fortified the defences were almost invariably integrated with those of the castle, as at Cardiff, Kenfig, Neath and Swansea. Cardiff was the earliest town in the lordship of Glamorgan. Cardiff Castle was raised by William the Conqueror in 1081, when he is said to have founded the town. Coins of the Conqueror minted at Cardiff suggest urban development beside the castle. Land granted to Tewkesbury Abbey by Robert Fitzhamon, c.1102, was said to lie beside the

1 H. S. Owen-John, Annual Report, 1983–84 2 Beresford, New Towns, 183, 527–8. (Glamorgan–Gwent Archaeological Trust, 1985), 66–115. 162 C. J. SPURGEON borough (iuxta burgum) of Cardiff.3 Earl Robert (c.1113–47) granted privileges to the burgesses of Cardiff, though this is only attested in an undated recital of the liberties and customs conferred by both Earl Robert and his successor, Earl William.4 It is possible that some of the privileges stemmed from a grant by Fitzhamon (ob.1107), and that the burgus recorded c.1102 was a settlement of significant size. Kenfig and Neath were the next boroughs to be established in Glamorgan, following Earl Robert’s appropriation of the western lowlands in the 1120s. Robert’s borough of Kenfig is first noticed in a charter dating to the last years of his life (1140–7).5 The land beside the town included a burgage in the west street ‘outside the gate of the town’. Already fortified, and developing a suburb for its burgesses, the town must have been established soon after the castle was founded c.1120. Neath town was established at the same time beside the castle on the east bank of the Nedd, facing the castle and lordship on the west bank in lands granted by Earl Robert to Richard de Granville. When Granville ceded these lands in 1130 to found Neath Abbey there is the first reference to the new town (nova villa) at Neath in the abbey’s founding charter.6 Neath was given formal recognition as a borough in the charter of Earl William (1147–83), which endowed the burgesses with the liberties and customs of Cardiff.7 Swansea was the first of the two boroughs established in Gower. Its castle and contiguous settlement were founded c.1107 by Henry de Beaumont, earl of Warwick, to serve as the administrative centre of his lordship. There is no mention of a town when Swansea Castle was attacked by the Welsh in 1116, but urban development had occurred before 1137–41, when a mint established there produced silver pennies bearing the name of the castle and town.8 Swansea attained full borough status with the grant of extensive liberties and customs by Earl William de Beaumont (1153–84).9 Earl Richard de Clare chose to locate his borough of Cowbridge on the old Roman road, the Portway, some distance from the existing castle of the St Quintins in Llanblethian, and conveniently placed mid-way between his boroughs at Cardiff and Kenfig. The rapidly expanding town was granted its formal charter by Earl Richard in 1254, and it was soon rivalled only by Cardiff, and was among the largest boroughs in Wales. Cardiff, Cowbridge and Swansea were the most successful foundations. They enjoyed their most flourishing periods of growth and prosperity in the thirteenth century. In 1300 Cardiff was by far the largest borough in Wales, with 421 burgages occupying its fortified enclosure of about 17 hectares (about 42 acres). Cowbridge expanded considerably from its foundation in 1254, with 276 burgages recorded in 1306, placing it sixth behind Cardiff in Wales.10 The defended area at Cowbridge was 5.5 hectares (13.5 acres),

3 Clark, Cartae, I (xxxiv), 37 (confirmed by Henry Cartae, IV (mlxxv), 1418; A. Ballard, British Borough I, 1106: Cartae, I (xxxvi), 39). Charters (1042–1216) (Cambridge, 1913), pp. xxx, 24. 4 Ibid., I (xciii), 94–7 and note. 8 Boon, Welsh Hoards, 49–50. 5 Patterson, Glouc. Charters, no. 68, pp. 73–4; Clark, 9 Clark, Cartae, I (cxxxviii), 136–8; Ballard, British Cartae, I (ci), 103; Epis. Acts, ii (L105), 637–8. See also Borough Charters, pp. xxxi, 19, 38, 47, 51, 54, 58, 63, Glam. Co. Hist., iii, 338. 65, 83, 89, 118, 137, 140, 145, 153, 212, 216, 238 and 6 Clark, Cartae, I (lxvii), 74–6; Phillips, Vale of Neath, 249; Glam. Co. Hist., iii, 362–4. 556–8. 10 Beresford, New Towns, 255–6, table IX.2. 7 Patterson, Glouc. Charters, no. 159, p. 148; Clark THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 163 but by 1300 its eastern and western suburbs had more than tripled the area occupied by its burgages. Boroughs experienced a steady decline throughout the fourteenth and fifteenth cen- turies, largely as a result of the general fourteenth-century depression resulting from periodic famines and plague, but also due to the advent of more settled times. Boroughs like Caerphilly and Llantrisant, founded as castle-towns in less fertile or hospitable areas in the interior, were left with increasingly obsolete and under-manned fortresses. The same was probably the case for Neath and Loughor, which certainly experienced gradual decay in the late Middle Ages. At Aberafan and Kenfig it was the natural phenom- ena of advancing sand dunes which gradually overcame them; by the 1530s, when they were noticed by Leland, they were only little villages ‘devoured by the sands’.11 Cowbridge has the most extensive vestiges of its defences extant, comprising sections of the walls defining its south-west angle, where the base of a salient tower remains, and the South Gate. Cardiff retains only its largely rebuilt West Gate and the stub of the North Wall at the opposite side of the castle. At Swansea there remains only an apparent fragment of the East Wall, seemingly incorporating part of a tower. Neath displays a stump of the town wall which projects from the castle gatehouse. At Kenfig, where walls were never built, there is a considerable angled portion of the town rampart, with one section of its external ditch, which have been spared by the encroaching sand dunes. The considerable industrial and commercial development at Cardiff, Neath and Swansea explains the almost complete disappearance of their medieval town walls. Ironically, it was the serious destruction caused by enemy bombing in 1941 which indi- rectly led to a further understanding of the medieval defences at Swansea. The centre of the city was obliterated, including the medieval core, and during subsequent reconstruction many excavations were carried out that located and recorded traces of the vanished defences. Less extensive excavations have traced and recorded a few sections of the former walls at Cardiff.

Cardiff

The medieval borough Cardiff Borough developed on a glacial terrace to the south of the castle beside the navigable Taff, and about 2 kilometres above its conﬂuence with the Bristol Channel, where it was crossed by the Roman road, the medieval Portway. The foundation of the town of Cardiff has been attributed to William the Conqueror, who raised the large motte there in 1081.12 The site was taken over by Robert Fitzhamon in 1093 to serve as the caput of his lordship of Glamorgan. Fitzhamon’s grant of land to Tewkesbury Abbey,

11 For the development of the boroughs: Glam. Co. Cowbridge: The Archaeology and Topography of a Small Hist., iii, 333–77; M. Beresford, New Towns of the Market Town in the Vale of Glamorgan (Swansea, 1980). Middle Ages (Gloucester, 1988); R. A. Grifﬁths (ed.), 12 Ann. Margam, 4; Breviate Ann., printed in Arch. Boroughs of Medieval Wales (Cardiff, 1978), 1–7, 102–28 Camb. (1862), 273; Brut, B S, 83 (s.a. 1080). (Cardiff), and 262–86 (Swansea); D. M. Robinson, 164 C. J. SPURGEON

FIGURE 1. Cardiff: John Speed’s map of 1610. c.1102, some of which lay beside the Taff, juxta burgum, suggests that he had granted the adjacent settlement the privileges of a borough.13 Earl Robert of Gloucester (c.1113–47) certainly granted ‘liberties and free customs’ to the burgesses of Cardiff. The portreeve (praepositus) of Cardiff is first recorded in Earl Robert’s agreement with Bishop Urban, drawn up in 1126.14 A grant made by Earl William of Gloucester c.1175 implies that he had extended the area of the borough. In this grant William endowed the church of St Mary and St Thomas the Martyr with 10s. of the rents derived from the new borough he had founded where his garden had been outside the town of Cardiff.15 Various suggestions have been made regarding the location of William’s ‘new borough’. Clark equated it with the eastern suburb of Crokerton, based on the dubious suggestion that the corrupted form ‘Crokherbtown’, only introduced as late as 1766,16 reflected folk memory of the earl’s garden of ‘crock-herbs’ or vegetables. An equally improbable suggestion located the ‘new borough’ on the flanks of the road skirting the east side of the castle beyond the North Gate. More plausible is the interpretation offered by D. G. Walker, based on two features

13 Clark, Cartae, I (xxxiv), 37. 15 Patterson, Glouc. Charters, no. 49, p. 63; Clark, 14 Ibid., I (l), 54; Epis. Acts, ii (L45), 620; Corbett, Cartae, I (ciii), 104. Glamorgan Lordship, 47. 16 D. R. Paterson, Arch. Camb. (1920), 36. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 165 of the topography of medieval Cardiff. He observed that St Mary’s church lay at the southern-most extremity of the town, while Womanby Street, Quay Street, Church Street and St John’s Street outlined the part of the town closest to the south flank of the castle, possibly tracing the line of a former entrenched outer bailey, or the defences of the original urban settlement. On Speed’s plan of 1610 (Figure 1) these streets appear to proj- ect the lines of the east and west ditches of the castle, and possibly followed the ditch of such an outer enclosure. Rees also deduced a two-phased development from the layout of the streets and suggested that traces of a wall noted at certain points along the line of Quay Street and Church Street may have marked the south side of the early Norman town.17 He cited no authority for the wall, and no record of it has been traced; it may be discounted as evidence for the primary borough, which was certainly not enclosed within stone walls. Walker adds the tentative suggestion that the suspected enclosure on the south side of the castle may have been the earl’s garden taken over by his ‘new town’. However, its estimated area of almost 3 hectares (9.4 acres) is too large for this, and the location of William’s extension remains uncertain. The palisaded defences of the early town are first attested during the Welsh revolt which followed the death of Earl William in 1183, when the town of Cardiff was burnt.18 The Pipe Roll for 1184–5 records expenditure at Cardiff on the castle, houses, and mills, and on the repairs made on the town gates and palisade (portas et paliciam ville de Card[iff]).19 The town appears to have recovered and prospered during the thirteenth century; with 423 burgages in 1307 the town attained its highest recorded total.20 By then it had outgrown the defended area, with suburbs at Crokerton, outside the East Gate, and at Soudrey hamlet along Shipman Street beyond the South Gate. The town walls and the five or six gates probably replaced the old palisaded defences in the latter half of the thirteenth century. The walls were certainly in place before 1314, when Earl Gilbert de Clare III’s post-mortem described Cardiff as ‘a certain market town enclosed by a wall’. In the town there were then 380 burgages;21 it is possible that the burgages no longer returned had been destroyed by the insurgents in Hawyse’s War, perhaps in the suburbs beyond the walls, though such war damage is unnoticed at Cardiff. The accounts of John Giffard of Brimpsfield, custodian of the lordship, rendered to 29 September 1316 record widespread damage during the revolt of Llywelyn Bren, including that at Cardiff. Repairs were undertaken at the castle, and the issues of the borough, its mills, and the profits of its court were all much reduced.22 The reduced yield from burgage rents may reflect reduced commerce rather than an actual attack on the town in 1316; severe damage was certainly inflicted in the vicinity, but the town wall at Cardiff had been repaired in the previous year,23 which may have deterred the insurgents.

17 Rees, Cardiff, 23–4 and n. 35. ing PRO, Chancery 133/130). 18 Ann. Margam (s.a. 1185), 18; Hist. King’s Works, ii, 21 Cardiff Records, i, 278 (I P M, Gilbert III de 650–1; Glam. Co. Hist., iii, 37. Clare). 19 Pipe Rolls, 3 Henry II, 1184–5, p. 6; Clark, 22 Clark, Cartae, III (dccxliii), 813–15 (misdated Cartae, I (clxxi), 171. 1281); Cardiff Records, i, 137–9. 20 Cal. I P M, iv (Edward I), no. 435, pp. 322–4; 23 Rees, Cardiff, 20, n. 25. Cardiff Records, i, 268; Beresford, New Towns, 553 (cit- 166 C. J. SPURGEON In May 1321 the town was sacked, along with the castle, during the baronial uprising, but it seems to have recovered within a few years. The inquisition following the death of Hugh Despenser in 1349 did not register any perceptible effects attributable to the Black Death, which raged in Gwent in that year; Cardiff Castle, with the vill 1 ‘enclosed by a stone wall’, was still valued at £52. 19s. 5 /2d., but revenues were considerably reduced following the subsequent outbreaks of the pestilence in 1361 and 1369.24 The rebellion of Owain Glyndw^r posed a serious threat to Cardiff after Owain led a successful campaign in south Wales in August 1403. The fragmentary account of William Rye attests strenuous efforts to repulse the insurgents. He released cash to the mayor of the borough to pay the wages of 138 men and 435 archers, and the garrison also included 24 men-at-arms. Arrows, guns, powder and cannon balls were accounted for, as well as expenditure on the defences of the castle and town.25 In October 1404 Henry IV dispatched a force of over 450 men to relieve the siege of Cardiff, but despite this it appears that the town was sacked, along with the castle, before the close of the year.26 There is only one authority for the capture of Cardiff, the chronicler claiming that only the Franciscan Friary in Crokerton was spared when the town was burnt.27 Damage inflicted at Cardiff during the Glyndw^r revolt is attested long afterwards in the records. In 1451 Richard Nevill granted a charter to the burgesses in which he confirmed their charter from Isabel Beauchamp (1423) and rewarded them for restoring the town defences. They had ‘very laudably fortified, defended and furnished our afore- said vill, as well in the walls, towers, gates and trenches as otherwise, not without their very great labour, costs and expenses’. In recognition of such ‘great fidelity and close affection’, he granted them immunity from answering charges for any felonies, tres- passes or other misdemeanours other than before the constable or bailiffs of Cardiff.28 Proof that these comprehensive reparations were occasioned by the attack of 1404 is provided by a later record. In 1492 when Glamorgan was held by Jasper Tudor, duke of Bedford, his coroner, Ralph Bampton, accounted for many lost rents from vacant burgages ‘destroyed, laid waste and burned by the rebel Welsh many years ago’; he also made an annual payment due to the bailiff as murage for the repair of the walls ‘which were very greatly damaged in the time of the rebellion of Wales’.29

Early modern Cardiff Despite the vicissitudes of the late medieval period Cardiff remained one of the largest towns in Wales. Leland (c.1536) found Cardiff to be well waullid, and is by estimation a mile in cumpace. In the waulle be 5 gates. First Portllongey [South Gate], in English the Ship Gate, ﬂat south. Then Porte Dour, in English the Water Gate, by southe weste. Then Port Miskin by north weste [West Gate],

24 Cardiff Records, iii, 288–9; NLW, Floyd Papers, IV (mxciv), 1454; J. E. Lloyd, Owain Glendower 3748D, I P M, Hugh, Lord Despenser; William Rees, (Oxford, 1931), 54, 74, 76, 89–90, 152; Glam. Co. Hist., in R. W. Southern (ed.), Essays in Medieval History iii, 183–5. (London, 1968), 182, and Rees, Cardiff, 57. 27 Eulogium Historiarum, ed. F. S. Haydon, Rolls 25 PRO, E. 101/44/5, Exchequer, Accounts Various, Series, iii, 401. 5–6 Henry IV (30 Sept. 1403–29 Sept. 1405). 28 Cardiff Records, i, 38–41. 26 Cal. P R, Henry IV, ii, 296 and 439; Clark, Cartae, 29 Ibid., 174–5, 180. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 167 2. Cardiff: the probable line of town defences. IGURE F 168 C. J. SPURGEON so caullid bycause it ledith the way into the lordship of Miskin. Then Porte Singhenith flat north [North Gate], so caulled bycause that menne passe by it into Singhenith. Then Porte Crokerton flat est [East Gate], so cuallid of the suberbe that joynith hard to it . . . Ther be 2 paroche chirchis in the towne, wherof the principale lying sumwhat by est is one, the other our Lady is by southe on the water side. Leland also noticed the chapel of St Perine in Shoemaker Street and a chapel just within the West Gate; Crokerton suburb and the lesser suburb outside the South Gate; and the former houses of the Dominican and Franciscan friars.30 Four of Leland’s five gates are identifiable with the names given in parenthesis above and marked on Figure 2. He locates the ‘Porte Dour’ or Water Gate to the south-west, near the river and St Mary’s church (where Speed marks a gate in 1610), but he omits the ‘Blunch Gate’ and the suspected ‘Golgate’, both on the west side facing the river. A further rental by the royal bailiff in 1550 is less detailed, but notices the suburbs of Crokerton and Soudrey, Cock’s Tower on the eastern defences, and Wharton Street, St John’s Street, the Hayes, Duke Street, Womanby Street, and St Mary Street.31 Merrick was fulsome in his praise when he described Cardiff c.1580. As the ‘chiefest town of Glamorgan’, he claimed it had been founded before the Norman Conquest, but nothing in quantity or beauty to that it is at present. The river Taff runneth near the town walls in the west part of the town, and washeth the wall, but somewhat too hard, for part of it is thereby overturned, and the sea floweth to the walls, where, at the west angle, is a fair quay . . .32 As early as 1552 the residents of Cardiff had expressed grave concern at the tidal erosion, which had undermined a section of the west wall of the town and threatened its quay, three bridges, and the church of St Mary. These fears were expressed in a dep- osition to the Court of Augmentations, and largely accepted.33 The proceedings appear to explain why the river had breached the town defences and was threatening the entire settlement. The town authorities were long accustomed to rely largely on charges levied annually on the parishes of St Mary and St John, derived from their well-endowed chantries, to repair and maintain the bridges, walls, gates, quay, sea-walls and weirs. Such charges had varied, as need required, between £10 and £40. With the suppres- sion of the chantries in 1548 the lower amount only had been assigned annually by the crown commissioners; whether or not this payment had been regularly made is unclear, but the burden of the cost of repairs was alarming the townsmen. The repair, building and amending of between 650 metres and 740 metres of sea-walls and the quay were considered an urgent necessity. The quay was built of ‘great stone and timber’ and constituted a defence for one section of the town walls; it had already been rebuilt three times in the previous twenty years. In addition, six weirs, maintained at a cost of 200 marks or more in recent years, were again decayed; in stilling the waters of the Taff, the weirs constituted

30 Leland, Itin. Wales, 34–5. 32 Merrick, Morg. Arch., 86. 31 Cardiff Records, i, 255–8. 33 Cardiff Records, iii, 43–9. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 169 the only safeguard and maintenance . . . of the said Town and bridges as also of the parish church called Saint Mary’s . . . the lack of the maintenance and repairing of the said weirs shall be the utter destruction of the said town. Regarding the town walls, it was stated that sections on the south-west were in ruin and would be entirely destroyed without urgent repairs and the restoration of the protecting weirs. It was also noted that there were six gates entering the town, one more than Leland had noticed. Two of these were gatehouses, and all had been maintained by the bailiffs; two had been repaired for £15 recently; the other four needed repairs costing 40 marks. Merrick’s notice of the ‘overturned’ section of the western wall attests the validity of the fears expressed in 1552.34 According to Merrick, Cardiff was environed with a fair high wall, garreted [that is provided with watch-towers] and place thereon to walk, saving where the river Taff and the tide, undermining it, overturned part thereof . . .’ Upon this town [wall] are four fair gates [besides the castle gate and other posterns] in the four quarters thereof . . . the East Gate; . . . the West or Miskin Gate; . . . the North or Senghennydd Gate; . . . the South Gate, near unto the quay. On the town wall was builded a tower, called Cock’s Tower, to defend the town against the danger of the sea. John Speed’s pictorial plan of Cardiff made in 1610 (Figure 1) furnishes the earliest and most informative view of the town at a time when it still retained its essential medieval characteristics. The plan almost entirely confirms the descriptions of Leland and Merrick, and the information in the submission to the Court of Augmentations. The depiction of the town defences is credible. They are shown running clockwise and unbroken from the West Gate, via the North, East and South Gates, to what may be Leland’s Water Gate close to the river on the south side of St Mary’s church. The West Gate is correctly placed in relation to the tall Beauchamp Tower of the castle. The North, East and South Gates are suitably located, as is also Cock’s Tower at the angle formed by the town wall to the east. Speed shows none of Merrick’s garrets or watch-towers on the town wall, only the probably larger and more significant Cock’s Tower which he had mentioned separately. The western defences between the West Gate and the Water Gate had entirely vanished according to Speed’s depiction. The encroachment of the river had greatly extended the breach in the west wall noticed by Merrick; the river now flowed within its line for over 400 metres between the quay at the bottom of Quay Street to the Water Gate. To the north of the quay, beyond the loop of the Taff, it is not clear why Speed indi- cates no trace of the northern section of the west wall that extended to the West Gate. A drawing by Francis Place made in 1678 shows a substantial wall enclosing this front to the north of the quay, and the wall is similarly portrayed on the panoramic view of Cardiff from the west by the Bucks in 1748. Neither of the views shows crenellations on the wall, but it is so substantial that it may have been the medieval town wall.35 Speed’s

34 Merrick, Morg. Arch., 87–8. Francis Place’s drawing); Cardiff Records, ii, pl. 12, 35 D. Moore, The Earliest Views of Glamorgan between 68 and 69 (for the Buck panorama). (Glamorgan Archive Service, Cardiff, 1978), 8–9 (for 170 C. J. SPURGEON pictorial record furnishes no information regarding the Blunch Gate leading on the quay, or the suspected Golyate or Golgate further to the south. The Blunch Gate, also known as Blanch Gate, Blunts Gate and Blounts Gate, appears to have been set at the lower end of Quay Street. It is not mentioned before 1542, when ‘the gate called Blounts Yate’ served as a means to locate burgages in a rental. The name was applied to the adjacent quay by 1666, and in 1785 the bailiff ordered that ‘the gate near the quay called the Blunch or Blunt Gate be taken out’ (that is demolished).36 The Golgate, a suspected gate or postern to a subsidiary quay, has been postulated from the Golgate, a narrow lane sloping down westwards from St Mary Street to Westgate Street. Early forms of the name may imply a derivation from a gate opening towards the river in the west wall: Gall Gate (1748), Goll Gate, Gollygate (1750), and Golly Gate (1786).37 The presumed Golgate would have been swept away by the Taff before 1610. Leland’s Water Gate or ‘Porte Dour’ (Welsh = Porth Dwr) would have vanished before 1678, when the river carried away all but the east wall of St Mary’s church.38 The Blunch Gate appears to have escaped a similar fate as the river advanced to the western edge of St Mary Street by 1849–51, when it was diverted to its present course in ‘the New Cut’.39 The minutes of the town council show that as late as 1759 the East Gate was repaired and rebuilt, and that ‘centry houses’ (porter’s chambers or guard-rooms) over the North Gate and by the West Gate had been leased out since the previous century.40 Later minutes show the abandonment of this customary responsibility for the maintenance of the town defences.41 A scheme to remove the gates was suggested to ease growing traffic congestion in 1767, but they were still let ‘at farm’ and facilitated the collection of tolls. In 1781 the decision was made to open up the main east-west road by pulling down the East and West Gates. In the event the bailiffs directing the work failed to demolish totally the West Gate; vestiges remained sufficient for credible restoration in 1921. The North Gate was demolished in 1786, the bailiffs being instructed to carry out the work and make good the breached side walls. The South Gate survived until 1802, when the inhabitants of St Mary’s parish offered to take it down at their own expense to provide a more convenient roadway into the town, making good the flanking ends of the old wall; the offer was accepted by the council. As late as 1828 the town plan of Cardiff had hardly altered since that depicted by Speed.42 Its population of around 5,000 was still confined to the old enclosed area and the modest south and east suburbs that had existed since the late twelfth century. The Glamorganshire Canal, from Abercynon to the sea at Cardiff, was built in 1792–4. It skirted the town on the east side, and from the north-east angle of the defences to the South Gate its course had been cut along the line of the old town moat.43 The town wall, now mostly facing the canal, remained a closely guarded asset to the corporation well into the nineteenth century. Like the town gates, the town walls and ditch had been

36 Ibid., i, 218; iv, 337; v, 341–2; Rees, Cardiff, 21. the OS for the corporation shows the ‘New Cut’ and 37 Cardiff Records, ii, 392, and v, 371; Rees, Cardiff, the old river bed before Westgate Street was built over 21–2. it. 38 Ibid., 103–4. 40 Cardiff Records, iv, 272, 296. 39 Ibid., 270, and pls. XXX and XXXI for town 41 Ibid., 333, 338, 358; Rees, Cardiff, 222–3. plans depicting the ultimate stages in the advance of 42 Ibid., pl. XXX facing p. 261. the river in 1828 and c.1840. A 1851 plan made by 43 Ibid., 232–5. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 171

FIGURE 3. Cardiff: the West Gate. leased out with adjacent plots of land to provide an appreciable source of income to the borough. The town ditch had been leased as early as 1671 for ninety-nine years at an annual rent of 3s. 4d., and by 1798 another lessee held both wall and ditch from the North Gate to the East Gate, which the marquess of Bute wished to obtain.44 Cock’s Tower at the angle of the eastern defences, which posed problems to barges negotiat- ing the turn, was leased at an annual rent of 6d. for twenty-one years in 1771; in 1811 it was leased along with a section of the town wall in the Hayes for forty-two years at 5s. annually, but another tenant surrendered the lease in 1821 when the corporation obtained estimates for repairing the tower for public use.45 The latter scheme was not implemented; in 1835 Charles Williams leased Cock’s Tower and an unspecified length of wall for 999 years at 1s. a foot (0.3 metres). The council’s minutes reflect conflicting actions taken in regard to the maintenance of the wall, which gradually diminished as the century advanced. On the one hand, there are punitive measures taken for encroachment: James Parry was charged 1s. rent for erecting a chimney on the wall in St Mary’s parish (1816); publicans of the Unicorn and the Masons Arms, Crokerton, both had to pay 1 guinea a year for encroachment on the town wall (1839); and an order was made against Robert Thomas for removing a wall on land leased to him (1843).46 Conversely, other actions of the council facilitated the gradual destruction of the wall: in 1834 three unspecified sections were sold at

44 Cardiff Records, iv, 296, 355–6. 46 Ibid., 368, 424, 429–30. 45 Ibid., 365, 382–3, 407. 172 C. J. SPURGEON

FIGURE 4. Cardiff: the West Gate. public auction to defray the cost of the new market hall; in 1835 78 feet (23.80 metres) of the wall were sold to the marquess of Bute at 1s. a foot; in 1843 two encroachments were permitted to improve a school playground and accommodate a weighing ofﬁce.47 Periodic attempts by the corporation to recover control over the town wall, as in 1838–9,48 did not halt the widespread, piecemeal attrition of the wall as townsfolk quarried it for building stone. Sections of the wall survived until the late nineteenth century. Photographs show its internal face still rising about 2.5 metres to the east of

47 Ibid., 405, 408, 430. 48 Ibid., 415, 424. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 173 Working Street in 1888; its ivy-clad, external face fell directly into the canal. Cock’s Tower was also photographed at that period, when its squared outer part rose almost to the eaves of flanking buildings and projected slightly beyond them into the canal.49 Such vestiges were erased or encased in modern buildings by the twentieth century, occasional fragments being sighted from time to time during redevelopment. The canal was finally in-filled in 1956–8, including the tunnels which had carried it beneath Kingsway from the central part of the eastern castle moat and the tunnel under Queen Street immediately in front of the site of the East Gate. The exact course followed by almost the entire circuit of the town defences, as marked in solid lines on the plan (Figure 2), can be plotted from Speed’s view and the various other records. Only to the west, in the sector subject to erosion by the Taff, has it been necessary to indicate an approximate line. The West Gate (Figures 3 and 4) faced north-west and was set at the outer end of a short section of the town wall perpendicular to the west wall of the castle and 54 metres north of its south-west angle. The perpendicular wall ran out about 16 metres to the rear of the gate, from which the vanished western town wall began its southward course overlooked by the castle wall and embracing the former West Street leading to the gate. This arrangement was devised to enable the more elevated castle battlements to con- tribute to the defence of the gate and contiguous sections of the town wall. The role of the castle in protecting the West Gate probably inspired Richard Beauchamp, earl of Warwick (1423–39), to site his great machicolated Beauchamp Tower a few metres beyond the northern cross-wall and directly commanding the gate; it is traditionally held that Glyndw^r’s successful assault on Cardiff in 1404 broke into the town at the West Gate, and the new tower of c.1430 may have reflected that disaster. The medieval road from the West Gate led north-west across the flood plain to a timber bridge set about 73 metres upstream of the present bridge over the Taff. The bridge, depicted by Speed (Figure 1) near the house of the Black Friars, was regularly restored and maintained until the late eighteenth century. It was only after severe flood damage in 1792 that a new stone bridge was commenced on the present site and in line with the main east- west thoroughfare skirting the south side of the castle. On the completion of the new bridge in 1796 the short West Street and the obsolete road to the abandoned crossing were by-passed.50 Two of the town mills were sited just outside the west wall to the south of the West Gate. To power these mills the approaching leat was channelled down the western moat of the castle and across the front of the West Gate, greatly enhancing the defences. The order to pull down the West Gate in 1781 was not fully implemented. In 1862 G. T. Clark described part of the gate which retained its iron gudgeons (door-pins). The standing masonry seen by Clark survived in 1921 when the area was cleared by John P. Grant, estate architect to the marquess of Bute.51 Grant confirmed Clark’s observations; a surviving section of one side of the gate-passage was 1.83 metres long, over 1.00 metre thick and 2.44 metres high. The fragment represented the outer end of the

49 Ibid., ii, pls. after 336; Rees, Cardiff, pl. X, after 51 Clark, Arch. Camb. (1862), 225; J. P. Grant, Cardiff p. 60. Castle: Its History and Architecture (Cardiff, 1923), 33–7. 50 Ibid., 21, 26, 132–4. 174 C. J. SPURGEON passage and retained a projecting jamb of the gate flush with the outer face of the gateway, behind which there were the three iron gudgeons noticed by Clark. The masonry consisted of blocks of Lias limestone, squared and coursed, set in good lime mortar. There were also traces of the associated foundations; clearance revealed a cambered roadway of cobbles and substantial remains of the stone bridge crossing the stone-revetted leats in front of the gate. The vestiges were the only surviving remains of any of the town gates, and to enhance this area below the residential quarters of the restored castle Grant rebuilt the gate in 1921. He based the reconstruction on eighteenth-century drawings made before the partial demolition in 1781.52 The views seem to depict a simple gate without an upper chamber, which was probably a fifteenth-century rebuilding after damage inflicted in the 1404 assault. The reconstructed gate is a simple vaulted passage flanked by thick side-walls, the whole in advance of the converging town walls and slanted north-west towards the medieval river crossing. A tall pointed arch of restored voussoirs, matching that in the eighteenth-century views, leads into the passage. There is a reconstructed, four-centred vault above the passage. The projecting jambs of the double-leafed doors were set to the front, flush with the outer face of the gate. The south jamb is original, surviving up to the springing of the arch and retaining two of the three iron gudgeons recorded by Clark. The medieval fabric is constructed of Lias limestone and extends a little within the south jamb; the restored masonry is of glacial pebbles with sharply squared and dressed Lias limestone quoins. Above the passage is a simple, open, crenellated roof platform, matching that shown in the early drawings. A short length of reconstructed town wall runs back to link with the west curtain of the castle about 2.50 metres south of the Beauchamp Tower. A flight of modern steps runs up westwards against the inner face of the wall to reach its wall-walk and the platform over the gate. At the inner east end the wall is pierced by a tall postern gate with a pointed arch with segmental rear arch. The postern is modern, providing access to the wide berm within the leat flowing down the castle moat. To the south of the gateway the foundations of a mortared wall possibly represent the town wall running south; the whole of the south-west external face of the gateway is reconstructed and shows no trace of the abutment of this wall. Within the gate the cambered cobbled roadway is largely ancient and differentiated from restored cobbling by red tiles. The stone bridge spanning the watercourse in front of the gate was also cleared and restored by Grant. Sandby shows a broad mill-pond backing against the bridge, which incorporated two arched culverts at its inner east end to convey water to separate stone- lined channels running south to the mills. Sluices would have controlled the flow of the water to activate the machinery of the mills; an outer watercourse was not revetted, and represented the normal passage of surplus water from the mill-pond. Grant believed that the outer watercourse had been spanned by timber beams set in a chase in flanking responds; one such beam was recovered from the mud. Foundations were found on the site of the mills further south, which were standing when the early views were drawn.

52 For Paul Sandby’s view of 1775, see RCAHMW, anonymous view, see RCAHMW, Inventory of Inventory of Glamorgan, iii/1a, fig. 119, p. 180, and Glamorgan, iii/1a, fig. 121, p. 182, and Cardiff Records, Cardiff Records, iii, facing 188; for a contemporary v, facing 241. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 175 Grant’s reconstructed bridge followed the presumed fifteenth-century lines, though he asserted that the substructure was earlier work. His finds from the bed of the water- courses included part of a water-wheel and pottery dating back to c.1400. The North Gate lay close to the east side of the castle, the south side of which bounded the north side of the town for about 183 metres. The eastern section of the north wall of the town sprang from the castle about 14 metres to the north of the reconstructed Roman south-east angle bastion; the bastion and the Roman wall were buried in the great earthen bank of the bailey in medieval times. This section of the town wall extended about 113 metres to the east, where it turned south-east just west of the pedestri- anized street known as the Friary. A section of the town wall 7.90 metres long survives, running east from a point 3 metres beyond the reconstructed castle wall, where it originally abutted the bailey bank formerly masking the Roman wall. The town wall rises to 1.70 metres above the present level of the lawn surrounding the castle. It is almost 1 metre thick and constructed of coursed Lias limestone blocks pierced with putlog holes, but only the lower 0.71 metres is original, the upper part having been rebuilt; red tiles distinguish the original from the reconstructed masonry. In the 1950s a section of the castle ditch survived against the outer face of the wall, which was carried down to more than twice its present height above the ditch fill. At the bottom of the ditch was an arched doorway with recessed blocking; it may have constituted a culvert permitting water from the wet moat and leat to the north to pass around the part of the castle facing the town. The North Gate was on the projected line of the wall on the castle lawn and about 26 metres east of the reconstructed castle wall. Its site lies beneath the Kingsway, the main thoroughfare leading north to the civic centre. Its character is only known from a distant view by Paul Sandby looking towards it from the north in 1775, but wrongly designated as the West Gate.53 The view does not correspond with the record regarding the lease in c.1715 of the ‘centry house’ over the North Gate. The depiction shows a gate like the West Gate, with a similar, open, crenellated roof platform in place of an upper chamber. The section of the wall linking with the castle is shown standing as high as the gate and with an uncrenellated parapet pierced by ten loops. The town wall running eastwards from the gate appears to have retained several more normal merlons in 1775. Several excavations explored the eastern section of the northern town defences in 1974–5 during the widening of the Kingsway.54 Cuttings on both sides of the length of surviving wall linking with the castle located medieval occupation levels within the wall, which were 1.50 metres below the present surface; externally, a number of fillings of the ditch were identified. East of the Kingsway and north of the projected line of the wall the inner scarp of the in-filled town ditch was detected beyond a berm about 2.70 metres wide. The ditch was traced northwards from this inner lip for 5.50 metres, where it reached a depth of 3.05 metres below the present level. It was estimated that the full width of the ditch would have been 9–12 metres. During building operations on the

53 The Sandby print is reproduced ibid., iii, facing 12 (1974), 30; B. Vyner, Arch. in Wales, 15 (1975), 57; 98, and Rees, Cardiff, pl. XXIV, facing p. 149. South Wales Echo (Monday, 12 January 1976). 54 J. and P. Webster and J. M. Lewis, Arch. in Wales, 176 C. J. SPURGEON Rose and Crown site in January 1976 substantial foundations were found on the line of the wall. The town walls described a right angle on the north-east, where there was probably a tower or at least one of the ‘garrets’ noted by Merrick. To the east the wall followed an irregular course in three straight sections to the South Gate. The northern section ran south-east as far as Cock’s Tower, broken near its northern end by the East Gate. The East Gate was set about 68 metres south-east of the north-east angle of the town. It straddled the main east-west road, the former Crokerton Street and now Queen Street. It is not known from any early drawing except for its clear uninformative depiction by Speed. The Glamorganshire Canal was carried along this side in the ditch from the north-east angle of the defences, passing beneath the road fronting the East Gate in a tunnel which continued some distance to the south of the road. Vestiges of this section of the town wall were photographed in 1880 behind Working Street; of random Lias limestone blocks, it stood about 2.50 metres high but displayed no architectural features.55 No visible section of the long northern line of the eastern wall is known to survive, and most of it has certainly vanished entirely with the redevelopment which followed the in-filling of the canal. Its line is marked on the floor of St David’s Way, a shopping arcade in the modern St David’s Shopping Centre. Cock’s Tower was set where the eastern circuit of the defences changed direction to head south-south-west. Photographed across the canal in the late nineteenth century, it was squared with an outer face rising directly from the canal. Its foundations were exposed in 1962, projecting about 2.50 metres beyond the wall and with an outer face measuring about 5.50 metres. The siting of the tower at an angle posed problems to the bargees as they manœuvred their long barges around its projection.56 The tower is the only tower known to have flanked the town walls, and it was differentiated from the watch-towers on their line for which Merrick is the only authority. The town wall crossed the lower end of the Hayes just east of its junction with Caroline Street and then continued along the west side of Mill Lane, turning to make its final short run to the South Gate at the lower end of St Mary Street. Here the former canal in the town ditch veered south towards the sea having followed the entire length of the eastern defences. The South Gate, like the East Gate, is known to have been depicted only on Speed’s plan. Since its demolition in 1802 no reports of any sighting of its foundations have been noticed. The town wall to the west of the South Gate followed a line corresponding with the northern limit of the premises of the Great Western Hotel to cross the present Great Western Lane about 12 metres north of its junction with Station Approach and a similar distance south of the multi-storey car park on the west side of that lane. A short distance north, at about ST 1831 7598, the site of St Mary’s church lies beneath the south-eastern part of the car park. The town wall continued west for about 40 metres to where it curved north-west. At this curve near ST 1826 7591, and a short distance south of the south-west corner of the multi-storey car park, stood the twin-towered gate seemingly depicted on Speed’s plan (Figure 1). Although Speed marks no actual opening for the gateway, the close spacing of the two towers probably equates with the Water Gate mentioned by Leland. The gate was close to the brink of the

55 Cardiff Records, ii, pl. after 336. 56 Rees, Cardiff, pl. X, after p. 66 and p. 256. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 177 encroaching Taff in 1610. At that date the river would have run south to the sea across ground now occupied by the eastern parts of the bus station and the Central Railway Station. The west wall of the town followed a course running north-north-west from the south- west angle for a distance of about 554 metres to the lower end of Quay Street. This long section had been undermined by the river by 1610. The main plan (Figure 2) marks the approximate line followed by the western defences before the Taff began to encroach within them. It is claimed that a section of the wall was exposed in 1885 near the Golgate, but its precise location is not recorded; its foundations were also thought to have been discovered in 1849–53 during the building of a new Town Hall on the west side of St Mary Street, when the lower part of a newel stair was identified.57 More recently, two substantial sections of walling close to the suggested line were encountered during the building of the multi-storey car park on the site of the former Westgate Street Fire Station, in the angle made with the south side of Quay Street. The sections of wall lay parallel to the north side of Westgate Street and 8 metres within its line. Overall they defined an 11 metre length of substantial wall, 1.20 metres thick and up to 3.57 metres high, its top only slightly below the present ground level. These fragments were interpreted as part of the nineteenth-century quay made obsolete when the Taff was diverted to its present course in 1849–53, and their position is marked in brick on the ground floor of the car park. However, it is possible that a section of the town wall had been exposed; only 3.50 metres south of Quay Street, the remains would have been close to the Blanch Gate. Further south two walls ran back at right angles to the substantial wall described above. Even if these walls were correctly interpreted as a dock enclosure or slipway of late date, this need not conflict with the view that the wall parallel with Westgate Street represented part of the medieval town wall.58 The old OS large-scale maps mark the site of the Blanch Gate at the bottom of Quay Street and plot the faceted line of the west wall running north from that gate to the West Gate. This line in shown on the main plan, although no evidence has been found to justify its precision.

Cowbridge

Development of the borough The walled town of Cowbridge was founded on 13 March 1254 by Earl Richard de Clare as a demesne borough, and it prospered and expanded so rapidly that by 1307 it had become one of the most populous boroughs of Wales.59 It was located on the Portway, the old Roman road, which provided it with the good communications essential for its growth as a trading centre. In addition, it lay at the centre of the Vale proper, the most

57 Ibid., 20, 22, 310. (Cardiff, 1922); J. Richards, The Cowbridge Story 58 P. V. Webster, ‘Excavations in Quay Street, (Bridgend, 1956); B. L. James and D. J. Francis, Cardiff 1973–74’, Arch. Camb., 126 (1977), 88–90, Cowbridge and Llanblethian Past and Present (Barry, 96–8. 1979); D. M. Robinson, Cowbridge: The Archaeology and 59 Beresford, New Towns, table IX.2, 255–6. The Topography of a Small Market Town in the Vale of principal accounts of Cowbridge are: L. J. Hopkin- Glamorgan (Swansea, 1980). James, Old Cowbridge: Borough, Church and School 178 C. J. SPURGEON

FIGURE 5. Cowbridge: the medieval walled town and its suburbs. fertile area of the Vale of Glamorgan, and was well situated to serve as the main market for the largest and richest demesne manor, administered from Boverton by Llantwit Major, near the coast about 6.5 kilometres to the south, but extending north to Llysworney, where it abutted Llanblethian manor a short distance west of Cowbridge. It was similarly convenient as a market serving the lord’s knightly tenants of the shire- fee. Cardiff lies about 19 kilometres to the east, across the Ely and Taff rivers, and Kenﬁg is 20.5 kilometres to the west. The foundation of Cowbridge can be seen as part of more determined policies in the exercise of lordship begun by Earl Richard and developed by his successor, Earl Gilbert. Cowbridge was not given its own castle. Llanblethian Castle was only about 0.5 kilometres south-west of the town and it continued to be maintained; as late as the sixteenth century it served as the prison of the lordship, and its constable acted as the mayor of Cowbridge. Robinson has made a detailed analysis of the growth of medieval Cowbridge based on a careful reconstruction of the medieval burgage plots of the town, using the carto- graphic evidence of the 1841 Tithe Survey and the ﬁrst 25-inch Ordnance Survey map of 1878.60 By this means he was able to plot 285 characteristically long, narrow burgage plots, all but eighteen of them set end-on to each side of the old road, the exceptions being those lining the outer end of the medieval Rood Street (now Church Street) in

60 Robinson, Cowbridge, 39–43, and ﬁgs. 9, 10 (n. 2). THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 179 the south of the walled area (Figure 5). The plots were disposed in three distinct but contiguous zones: at the centre, about a hundred lay within the walls; outside the walls, about eighty-six were ranged along the road in the western suburb (Westgate); and about ninety-nine formed the eastern suburb (Eastgate). The burgages were generally uniform, at an average of about 67 metres long to the north of the road and about 55 metres to the south side, but there was some variation in their widths, suggesting that some of the nineteenth-century boundaries represented the amalgamation or subdivision of original plots. The total of 285 plotted, however, is close to the maximum recorded medieval total of 320. The increase in burgages in the time of Gilbert de Clare II would have generated a spread of settlement well beyond the primary walled area, where Robinson estimated that a maximum of eighty to ninety burgages could have been accommodated, assum- ing that some of the 1841 plots represented subdivided burgages. His suggestion that this total had been reached by at least 1275 is mistaken, based on the 135 burgages pur- portedly recorded in 1281, but in fact relating to the year 1316.61 It is probable that extra-mural settlement would have become necessary as early as 1269, when the average annual increase of 5.4 burgages under Gilbert II would have brought the total to ninety and exhausted available space in the walled area. The suggested early development of extra-mural suburbs from about 1269 would strengthen the view that the town walls were built when the borough was founded by Earl Richard. The claim that the walls date to the early fourteenth century is not demonstrated by any documentary or architectural evidence, and no dating evidence was produced when two sections were cut across the line of the north wall. The excavations also failed to discern any evidence that the walls replaced a primary rampart. The internal stone-revetted bank behind the surviving section of the wall running north from the south-west angle-tower is an eighteenth-century feature in the large garden behind Old Hall. It is possible, but improbable, that Gilbert II inspired the Cowbridge defences; given the estimated rate of growth this could only have been before 1269, by when it is estimated that the enclosed area could accommodate no further burgages. With its church and market-place, it seems more probable that Gilbert had inherited a borough already enclosed within strong walls and with only enough vacant land for another twenty or thirty burgages. Earl Richard’s preference for the low-lying site on the Portway, rather than the hill-top beside Llanblethian Castle, would explain the immediate need for defences to protect his new town.62 Cowbridge continued to expand under Gilbert II’s widow, Joan.63 After her death in 1307 a period of stagnation set in, probably exacerbated by warfare and the Black Death.64 The second half of the fourteenth century witnessed a full recovery and further expansion at Cowbridge, attested by the greatly increased number of burgesses in 1425.65 The second ﬂourishing of the borough seems to represent the climax of its for-

61 Clark, Cartae, III (dccxlvi), 828 (Gifford’s survey Cartae, III (dccxlvi), 828 (misdated 1281, and usually of 1361, misdated 1281). misquoted as evidence for only 135 burgages); Cal. C 62 Beresford, New Towns, 183, 527–8. R, Edward II (1318–23), 541–2; Beresford, New 63 Cal. I P M, iv (35 Edward I), no. 435, 322–4. Towns, 554 (citing PRO, Chancery, 135/105). 64 Beresford, New Towns, 554 (citing PRO, 65 Glam. Co. Hist., iii, 355. Chancery 134/33); Glam. Co. Hist., iii, 340; Clark, 180 C. J. SPURGEON

FIGURE 6. Cowbridge: plan of the town defences. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 181 tunes. Robinson estimates a population of around 1,200 in the early fourteenth century and a minimum of about 1,400 inhabitants in 1425. Thereafter there began a long period of decline, shared with the other boroughs in Glamorgan. As the fifteenth century progressed many burgages were abandoned, and revenue from market tolls and brewing declined.66 The borough showed evidence of further decay in 1514; houses lay empty and derelict, and revenues had fallen sharply. Long before 1514 it is probable that the burgesses had ceased to yield any profits to the lord.67 Cowbridge, however, still sounded quite pros- perous in Leland’s account: As much of Cowbridge as is enclosid with the waull stondith on the est ripe and the bridge of ston there. The great suberbe of Coubridge is cis pontem. The waulle of Cowbridge is a 3. quarters of a mile aboute. There be 3. gates in the waulle, the est, the west and Porte Meline by south.68 He noticed the church as a dependency of Llanblethian parish church, failing only to record the small North Gate of the town. The location of the town on the main route though the region probably saved it from the serious decline sustained at the more remote boroughs at Llantrisant and Caerphilly. The Acts of Union of 1536–43 enhanced the status of Cowbridge as one of the contributing boroughs electing a member of Parliament for the new county of Glamorgan, where Cowbridge ranked third after Cardiff.69 Cowbridge maintained its status as an important market town serving a rich agricultural hinterland. Merrick, c.1580, records its two markets each week and its two annual fairs; he also noticed its school and town hall.70 A grammar school was founded in the early seventeenth century.71 In 1781 the population was 759, and the same number of inhabitants was returned in the first census of 1801; the number increased to 1,107 by 1821; by 1971 there were 1,224 within the old borough boundaries, excluding large new suburban housing estates extending south towards Llanblethian.72 When a northern by-pass was opened in 1965 the town was relieved of the congestion of heavy traffic along its long narrow street. After this it retained its old market function, but it has also become a desirable residential commuter town.

Description The Town Walls (Figures 5 and 6) outline a pentagonal enclosure embracing 5.50 hectares (13.50 acres), and measuring about 335 metres east-west by 217 metres north- south at its widest central point. The enclosed area is bisected by the medieval Portway, which formed High Street between the East and West Gates and served as the market area of the town. The road passed through the East Gate about 45 metres beyond its

66 Ibid., 355–7, with graph and detailed tabulated 70 Merrick, Morg. Arch., 37, 81, 136, 139. revenues for 1425–58. 71 I. Davies, A Certaine School: A History of the 67 Ibid., 358. Grammar School at Cowbridge (Cowbridge, 1967); 68 Leland, Itin. Wales, 32. Hopkin-James, Old Cowbridge, 207–77. 69 Robinson, Cowbridge, 64. 72 Robinson, Cowbridge, 65, 76. 182 C. J. SPURGEON crossing of the Thaw. Besides the main gates demarcating each end of High Street, there was a South, or Mill, Gate on the line of the south-west curtain, and a gate to the north. To the north of High Street the parallel North Wall and squared northern angles outlined a long rectangular area. A less regular outline was imposed on the walls enclosing the larger area south of High Street by the south-western course followed by the Thaw from a point just below the river crossing. The West Curtain and the shorter East Curtain constricted by the river maintained the square form to the north. To the south, however, long unequal south-west and south-east facets produced an irregular pentagonal enclosure. Only the South Gate survives; the West Gate was demolished in 1754, the East Gate c.1770. The street pattern within the enclosed area was a simple T-plan, the short Rood Street (now Church Street) running south from the mid-point of High Street to the South Gate. A short alley ran north from a point further west along High Street to link with the North Gate, here named ‘North Lane’, since it was designated only as a ‘footway’ in a survey of 1630.73 The two main streets of the town defined three zones which col- lectively accommodated about 100 burgages and the church. Burgages were mainly con- centrated to the north, flanking High Street, with only about seventeen ranged beside the lower end of Rood Street near the South Gate (Figures 5 and 6). Divided only by the narrow ‘North Lane’, contiguous burgages entirely filled the enclosed area to the north of High Street, facing those ranged along the south side of the street and broken only by Rood Street. ‘The Tower of Cowbridge’ recorded in 1487 has long vanished, but its location within the walled town may be deduced from a deed endorsed ‘carta le towre de Cowbryge’.74 The deed, dated 26 September 1487, granted by the feoffees of William Prior, burgess of Cowbridge, gave to John Thomas, son of John ab Ieuan ap Thomas of Llanfihangel Place, certain properties within and without the walls. The lands granted by William Prior’s feoffees included ‘a tower with all curtilages and walls’ within the walls of the town. The curtilage with the tower can be identified as that occupied by the town house of the family of Thomas of Llanfihangel Palace, now the Bear Inn, which incorporates fabric attesting a three-unit, lateral-chimney hall-house of late medieval date. The Dunraven estate, successor to the Thomas family, still held the later Bear Inn property in 1782.75 It is possible that the ‘tower’ of 1487 was set on the line of the North Wall of the town. Alternatively, it may have designated a detached, late medieval, first-floor hall represented by the barrel-vaulted undercroft now supporting an eighteenth- century assembly room to the rear of the Bear Inn. William Prior, the grantor and suspected founder of the chantry at the town church,76 was not ‘William the Prior’, as some have suggested, but William Prior, burgess of Cowbridge; there is no record of a priory at Cowbridge. The walls survive virtually intact only on the south-west, with substantial sections of the West and South-West Curtain extending from the extant South-West Angle-Tower of the town. The South-West Curtain continues, rebuilt on the old foundations, as far

73 Hopkin-James, Old Cowbridge, 48. 163, and plan between 160–1. 74 Clark, Cartae, v (mccli), 1735–6. 76 James and Francis, Cowbridge, 43–4. 75 Clark, Limbus, 272; Glamorgan Historian, 9 (1973), THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 183 as the South Gate, which also survives. Recorded sightings of the North and East Curtains and the topography of the remaining south-east flank permit the delineation of the entire circuit of the defences. They enclosed 5.50 hectares (13.50 acres), far less than the 33 acres first cited in 1833 by Lewis and closely followed by most subsequent authorities.77 Robinson suggested that the initial defences were of earth and timber and that the primary area enclosed extended as far as the Thaw to embrace 8.50 hectares (21 acres).78 However, no evidence was found during excavations to suggest that sections of the North Curtain had been inserted in a primary rampart. A survey of 1630 suggests that the North Gate, unnoticed by Leland a century earlier, was the first gate to fall out of use, the ‘common footway’ leading to it from High Street having been blocked by a stable erected by Richard Says.79 The main gates continued to control traffic along High Street until the mid-eighteenth century. By then, however, the medieval walls were crumbling and the main gates were perceived as a hindrance to traffic through the town. For this reason, Thomas Edmonds of Old Hall was permitted to take down the West Gate in 1754, provided he carried away the stones at his own expense. The borough rolls mention the East Gate in 1763 and 1768, but by 1775 it had gone, and an adjacent property was described as ‘near where the Eastern gate lately stood’.80 The ‘Town Ditch’ which fronted the walls is no longer visible. It was maintained in 1610, when it was decreed that ‘noe maner of pson shall caste noe duste, dounge, nor noe other filthe in the streates nor in the Town ditches’ or ‘drive anie beastes into . . . the same’. The Town Ditch was noticed again in a survey of 1630, and later records suggest that its circuit remained open and continuous until the 1790s. The south ditch is mentioned in 1672, 1776, 1786, and 1787, and that on the west in 1748 and 1791. In 1794, however, Thomas Thomas was presented for ‘making level’ a section of the old ditch on the south-east side, and in 1808 it seems that the west ditch in Eagle Lane had been ‘fenced and stopped’. The ditch on the east side seems to have remained open in 1813, and in 1832 it was ‘cleansed’ at the expense of the corporation. The ditch along the northern side of the town was open and apparently insalubrious in 1853, when it was blamed for a serious fever epidemic in the town; to correct this nuisance a covered drain was inserted and the ditch was filled in.81 The borough records suggest that the Town Ditch held water. A presentment of 1748 instructed Thomas Wyndham to raise the wall near the West Gate and ‘stone’ ( revet?) and keep clean the watercourse opposite John Long’s pool. The watercourse was either the ditch itself or a leat designed to feed it. The ditch on the south side is termed the ‘main gutter’ in 1776, and in 1853 the northern ditch clearly provided inadequate drainage eastwards to the river, being ‘an offensive open gutter . . . containing a quantity of putrid matter’.82 The northern ditch was partly revealed in excavations in 1981–2, which also disclosed

77 Lewis, Top. Dict., i, s.n., Cowbridge; Beresford, 79 Hopkin-James, Old Cowbridge, 71; Richards, New Towns, 554; H. L. Turner, Town Defences in Cowbridge Story, 144. England and Wales (London, 1971), 213; I. Soulsby, 80 Hopkin-James, Old Cowbridge, 71–2. The Towns of Medieval Wales (Chichester, 1983), 115; 81 Ibid., 32, 35, 48, 72–5. James and Francis, Cowbridge, 17, 35. 82 James and Francis, Cowbridge, 102. 78 Robinson, Cowbridge, 44–5. 184 C. J. SPURGEON

Figure 7. Cowbridge: the South Gate. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 185 the North Gate of the town. Only the inner lip of the ditch and part of its scarp was traced, with a massive masonry bridge abutment fronting the gate. From the inclination of the scarp of the ditch, it was estimated that the outer lip probably lies on the far side of North Road, giving it an overall width of about 15 metres. A narrow berm, 1.50 metres wide, separated it from the town wall flanking the gate to the east.83 The South Gate (Figure 7) survives in good condition after recent restoration. Its masonry is of random rubble, much patched, with dressings of Sutton stone. This simple structure has been dated to the early fourteenth century, but no architectural feature attests such a date. The gateway consists of a square block, comprising a passage flanked by two thick walls and set astride the south-west section of the town wall at the lower end of Rood Street. The town wall survives on the west, joining the middle side of the gatehouse, but on the east side the adjacent wall is gone and its site is occupied by eighteenth-century outbuildings. The outer part of the gateway, facing south-west, possibly projected outwards into the ditch. The passage is spanned at each end by a rough arch of Lias limestone voussoirs, both springing from the flanking walls. The outer arch is elliptical and possibly rebuilt on original Sutton stone quoins at the outer angles of the passage. Outward subsidence towards the ditch may have necessitated the buttresses of late date that were added to the outer faces flanking the entry. The inner arch towards Rood Street is segmentally pointed, its roughly dressed Lias voussoirs springing from the side-walls. The gate-passage, excluding the modern external buttresses, is 5.70 metres long and unvaulted. The doors were set 1.80 metres within the passage towards its outer end, their position marked by the projecting jambs bearing another segmentally pointed arch, all framed with plain, broad-chamfered Sutton stone ashlar. A narrow, rectangular murder slot protects the front of the door, and there was provision for a portcullis, but very recent patching has masked the surviving portcullis groove on the west side recorded in an unpublished Royal Commission survey of c.1940. The passage within the recessed doorway is open above, but was formerly covered by a timber floor set on transverse beams supported by offsets to the flanks. Immediately behind the doors the east wall displays a blocked medieval doorway with a pointed reliev- ing arch of rough voussoirs. The doorway probably gave access to the stairs leading to the upper floor. The upper floor and original roof level are gone, and the present low parapet is probably of post-medieval date. A water-colour by Paul Sandby painted in 1777 shows the gate from outside the walls; since that date its appearance has hardly altered.84 In 1777 a wall ran along the outer lip of the ditch, broken in front of the gate for a five-bar gate opening to the unmade track heading to the town mill and Llanblethian. The South Gate was repaired by the corporation in 1805 and 1862. The South-West Wall extends about 115 metres north-westwards from the South Gate, and about 48 metres to the south-east. At both ends it terminated at projecting angle-towers or bastions. To the west of the gate the first 30 metres of the line are marked by a modern boundary wall, with a cottage and garden occupying the site of the ditch. From its junction with the garden boundary the town wall survives as a rebuilding on old foundations for about 18 metres.

83 J. Parkhouse, Annual Report 1981–82, Glam- 84 Moore, Earliest Views, 33. organ–Gwent Archaeological Trust, pp. 12–16. 186 C. J. SPURGEON Then there is a length of original walling for about 22 metres, standing 3.60 metres high with a thickness of 1.80 metres and pierced by a small modern doorway leading into the grounds of the Old Hall (occupied by the Health Centre in 1995); the wall here retains a strong external batter. After a break in the masonry, a further section extending about 12.20 metres is again rebuilt on old foundations. The final section extending 27.50 metres to the South-West Angle-Tower is original. White chalky mortar charac- terizes the extensive patching throughout and the area of rebuilding. The original masonry is of random Lias limestone rubble with yellow clay bonding instead of mortar. The wall extending the south-west line eastwards from the modern building abutting the South Gate is comparatively modern, but it ends by outlining another boldly projecting, round angle-tower and may be accepted as a rebuilding on the old line. Within this section of the line the old grammar school grounds are set at a higher level. The South-West Angle Tower is circular and about 5.60 metres in diameter. It stands 3.60 metres high; the lower part, rising on a strong batter, is original. The upper portion was rebuilt in the eighteenth century as a summer-house, with rendered faceted sides dashed with calcite gravel, and the interior filled in to first-floor level. At that level there is a small fireplace of the same period, but the roof has gone. The West Wall contiguous with the South-West Angle-Tower is substantially intact for about 40 metres, its ditch occupied by the cattle market. Here, as it heads towards High Street and the West Gate, it is 1.80 metres thick and survives to a height of 3.40 metres with a pronounced external batter. The southern section of the West Wall is masked internally by an eighteenth-century earth terrace, contained by a stone revetment wall and outlined to the front with simple mock crenellations. These late features, along with the summer-house, represent the landscaping of the grounds of Old Hall. Beyond the original section the line is preserved by a modern boundary wall as far as the Old Masons Arms on High Street next to the former West Gate. The Old Masons Arms incorporates a 1.80 metres-thick section of the town wall in its west gable. The West Gate, demolished in 1754, was probably a more impressive structure than the South Gate, but no evidence can be cited to confirm this. The projected line of the West Wall beyond High Street is accurately marked by the boundary wall running along the east side of Eagle Lane. Immediately behind this wall, in the grounds of Woodstock House, excavations in 1981 distinguished a band of yellow sandy mortar thought to have represented the foundation trench of the West Wall.85 The footings of the North Wall were exposed at its east end during the building of an extension at the north end of the Institute.86 The footings were 2.13 metres wide, cemented with yellow mortar largely consisting of clay, and with a marked external batter. This siting of the foundations was not precisely located in relation to North Road, and for long it was assumed that the North Wall must have followed a line approxi- mating to the south side of North Street. In 1979 excavations at the rear of the Bear Inn disclosed that the north ditch extended back well within the south side of North Street and that one small section of the town wall within its line was built into the walls of an outbuilding.87 The wall was 3 metres high and 2.20 metres wide and had a marked

85 Parkhouse, Annual Report, 16. 87 A. Davidson, Arch. in Wales, 19 (1979), 40; 86 W. F. Evans, Arch. Camb. (1914), 304. Robinson, Cowbridge, 44–5, and fig. 11. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 187 external batter. Further sections of the North Wall were identified at two points in 1982. Behind the Midland Bank on the property immediately east of the Bear Inn was a fragment 2.30 metres high, 3.60 metres long and 1.80 metres wide, with the characteristic external batter and yellow mortar of the medieval fabric. A squared termination marked an opening at the west end of this wall section; excavation located a bridge abutment of solid masonry fronting this opening and the inner scarp of the town ditch on which the abutment rested; the remains were part of the site of the North Gate. Within the wall a small trench disclosed two medieval footings, but it was not possible to determine whether these represented part of a gatehouse.88 Behind the Midland Bank the North Wall had been set about 5 metres within the south side of North Street. The second sighting in 1982 confirmed the withdrawn alignment to the rear of Taynton Cottage, towards the east end of the wall, where excavations exposed the stump of the North Wall beneath an outbuilding and set back 5.20 metres from the road.89 Following the filling of the North Ditch in 1853 its area was encroached upon at the rear of properties previously defined by the line of the wall. It is very probable that many more sections of the wall were demolished during this encroachment. The plan (Figure 6) tentatively marks angle-towers at the northern angles on analogy with those known at the south-west of the enceinte. The East Wall is the shortest and the least well-established. It is generally maintained that it ran north-south just to the east of the Town Hall. An excavation in 1980 disclosed a ‘section of wall possibly related to the medieval town wall near the East Gate, and the lip of the ditch’ at SS 9956 7461. No further details were reported, but the location is close to the presumed line as it nears another suggested angle-tower which probably marked the south-east corner of the town. The South-East Wall followed a slightly re-entrant course in two unequal facets from the presumed but vanished south-east tower to the probable tower marked by the rounded revetment projecting at the south angle. From there the shorter facet of the South- East Wall appears to be marked by the modern revetment supporting higher ground within the grammar school grounds and the contiguous churchyard. At the south-east angle of the churchyard the facet ends with a rounded section of masonry which appears to incorporate medieval fabric at the base of its battered face. The feature is too close to the site of the South Tower to represent another such structure and only effects a slight set-back at the start of the remaining and longer facet on this flank. The boundary walls of properties marked the latter facet as far as the south-east angle until open ground within the line was taken over as part of a large garden centre around 1970. The facet was followed by a re-entrant of the borough boundary (Figure 6), which seems to confirm it as the line of the vanished town wall; only the boundary wall to the south- west end joining the corner of the churchyard remains. The general course of the South- East Wall was dictated by the south-western trend of the River Thaw along this flank.

88 Parkhouse, Annual Report, 12–16. 89 Ibid., 17–18. 188 C. J. SPURGEON Kenﬁg

Kenfig lies on the east bank of the Afon Cynffig. The river now reaches the sea 2 kilometres to the west, but before its deflection by the encroaching sands it formed a broad estuary running south from the medieval site, now marked by the large Kenfig Pool and lesser ponds.90 Castle and town lie on the north side of the extensive area of sand dunes known as Kenfig Burrows, immediately south of the dismantled sidings beside the main-line railway, and about 350 metres west of where they are crossed by an elevated section of the M4. The Roman road between Cardiff and Loughor, the medieval Via Regalis, Via Maritima, or Portway, crossed the Afon Cynffig near the site, probably about 560 metres to the east at Pont-felin-newydd, where its alignment may be marked by Water Street to the north of the river and Heol-y-sheet to the south.91 In this regard, it should be noted that the castle keep incorporates fragments of Roman bricks and tiles, and that two Roman coin hoards and Roman sherds have been found in the vicinity, as well as an Early Christian brooch.92 It is possible that the castle and town occupy a Roman military site on this navigable estuary, emulating the re-use of other forts on the line of the Roman road at Cardiff, Loughor and Neath.93 The persistent tradition that Kenfig was successively fortified by Iestyn ap Gwrgant (1080) and Robert Fitzhamon (1092) is a fabrication of the forged Gwentian Brut.94 However, Merrick’s citation of the lost cartulary of Neath Abbey to assert that Robert, earl of Gloucester, founded Kenfig is credible.95 His borough was probably established soon after he had raised the castle in the 1120s, and it was soon provided with the defences first alluded to in a charter he issued between 1140 and 1147. In this charter Robert granted to Ewenny Priory 21 acres of arable land beside the town, with a burgage in the west street outside the gate of the town.96 Earl William (1147–83) founded the church of St James at Kenfig before 1154, to be held for life by his clerk, Henry Tusard, with reversion to Tewkesbury Abbey.97 St James, which lay outside the town, and the chapel of St Thomas within its walls were confirmed to the abbey by Bishop Nicholas (1173–83).98 Before 1166 Earl William also granted 20 acres of land and a burgage at Kenfig to Gregory Fitzrobert for the use of his clerk, Elias.99 Earl William appears to have granted to Kenfig’s burgesses the liberties and customs of Cardiff.100 Their vulnerable location to the west of the lordship was demonstrated in 1167, when the town was burned by the Welsh for the first time.101 Following the death of Earl William in 1183 there was a serious Welsh uprising. In

90 See area map, RCAHMW, Inventory of Glamorgan, Cartae, I (ci), 103; Epis. Acts, ii (L105), 637–8. See also iii/1a, fig. 232, p. 317. Glam. Co. Hist., iii, 338. 91 Gray, Kenfig, 30–2, fig. facing p. 318; RCAHMW, 97 Patterson, Glouc. Charters, no. 271, p. 175; Clark, Inventory of Glamorgan, iii/1a. Cartae, I (cx), 111; Glam. Co. Hist., iii, 98, 338. 92 Glam. Co. Hist., ii, 287, 331–4, 444b; BBCS, 3 98 Llandaff Epis. Acta, no. 31, pp. 28–31; Clark, (1926), 76, and 14 (1950), 87; Arch. Camb., 83 (1928), Cartae, I (cxxxvi), 134; Epis. Acts, ii (L 170), 659. 200–2. 99 Patterson, Glouc. Charters, no. 182, p. 163; Clark, 93 RCAHMW, iii/1a, p. 32 and fig. 9, p. 33. Cartae, I (cv), 107. 94 J. E. Lloyd, Proc. of the British Academy, 14 (1928), 100 Glam. Co. Hist., iii, 338; Beresford, New Towns, 376–7. 199, 555. 95 Merrick, Morg. Arch., 42, 101. 101 Ann. Margam, 16. 96 Patterson, Glouc. Charters, no. 68, pp. 73–4; Clark, THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 189 the expectation of an attack timber for palisading was conveyed to Kenfig from the woods near Chepstow, to strengthen the defences of the town and castle. From Bristol three brattices (bretescarum) were dispatched, along with 200 picks, suggesting a major effort to deepen ditches. Despite the preparations, a Welsh assault reduced the town and its mill to ashes; the castle, though damaged, was held by its royal custodian. The revolt ended in July 1184, and in the following year twenty-four ships were commis- sioned to convey more timber to Kenfig to repair the gates and palisades of the castle and town. On account of damages suffered the surviving burgesses were excused their rents. The Pipe Rolls disclose no hint of building in stone.102 Earl Robert’s stone keep within the palisade of Kenfig Castle remained the only masonry structure. Kenfig recovered during the lordship of John of Mortain (king, 1199–1216). In his time the church of St James and its chapels were granted to Margam Abbey at farm for an annual rent of 10 marks (£6. 13s. 4d.) to Tewkesbury Abbey.103 John also confirmed to Margam lands and a burgage it held at Kenfig by the gift of Earls Robert and William and the burgesses there.104 Kenfig was burned by Hywel ap Maredudd of Meisgyn in 1228.105 In 1232 Morgan Gam destroyed the town, excepting the church; the castle keep (turris) held out, although it was still only protected by palisaded and ditched defences (fossa et sepe).106 In 1243 Hywel ap Maredudd, for the second time, led an attack on Kenfig and reduced it to ashes.107 By the end of the thirteenth century the palisaded rampart encircling the Norman keep of the castle had been replaced by a polygonal curtain-wall with a gatehouse facing south into the town. The strengthening of the castle was probably carried out towards the end of the tenure of Gilbert de Clare II (1263–95), but there is no evidence to suggest that the earth and timber defences of the town were similarly replaced in stone. History records no hostilities at Kenfig between 1243 and 1295, but the latter year witnessed its sixth destruction by the Welsh, shortly before the death of Earl Gilbert II. On this occasion both castle and town were destroyed by Morgan ap Maredudd with a force from the Blaenau. Gilbert’s inquisition post mortem records Kenfig Castle as ‘recently burned’, and burgage rents were reduced to only 13s.108 The town made a strong recovery at the end of the thirteenth century. By 1307 it sheltered 142 burgages and enjoyed the profits of 142 acres of land, a windmill and a water-mill, and markets and a fair on the feast of St James (25 July). The same number of burgages was registered in 1314.109 Kenfig was assaulted in 1316 by Llywelyn Bren. On this occasion the castle was successfully held by Leision de Avene, lord of Afan,

102 Pipe Rolls, 30 Henry II (1183–4) and 31 Henry 107 Clark, Cartae, III (dccclviii), 859; ‘Cardiff II (1184–5); Ann. Margam, 17–28 (misdated to 1185); Chronicle’ (s.a. 1244); Breviate Ann. (Arch. Camb., 1862), Clark, Cartae, I (clxxi), 170–6; Glam. Co. Hist., iii, 37–8; 279; Glam. Co. Hist., iii, 50, 338. Lloyd, Hist. Wales, ii, 571–2; Hist. King’s Works, ii, 108 Cal. I P M, 24 Edward I, vol. iii, 24 Edward I, 650–1. p. 244; Corbett, Glamorgan Lordship, 141–2; A. L. 103 Llandaff Epis. Acta, no. 45, p. 43; Epis. Acts, ii (L Evans, Margam Abbey (1958), 76. 319), 701. 109 Cal. I P M, 35 Edward I (1307), 8 Edward II 104 Cardiff Records, iii, 7; Arch. Camb. (1867), 318. (1314). 105 Ann. Margam, 35–6; Glam. Co. Hist., iii, 46–7. 106 Ann. Margam, 39; Breviate Ann. (Arch. Camb., 1862), 287. 190 C. J. SPURGEON but the town was taken and ravaged.110 Shortly after the attack John Giffard, royal keeper of Glamorgan, recorded that ‘the greater part of the town was destroyed in war’; forty-two burgages rendered nothing, having been reduced to ashes.111 Kenfig was again destroyed in the barons’ rising of 1321.112 In 1324 Edward II granted Hugh Despenser a charter confirming to the burgesses of his boroughs, including Kenfig, the freedom from toll throughout the kingdom, except for the dues respecting wools, hides, woodfells and wines.113 Kenfig began to prosper again, and by 1349 there were 144 burgages rendering £7. 4s. 0d. at the standard rent of 1s.114 The recovery of Kenfig as a mercantile centre was acknowledged in 1359, when it was listed with other Despenser boroughs in a confirmation by Edward III of his father’s charter of 1324, and again returned 144 burgages in an inquisition.115 Kenfig’s earliest known detailed charter of privileges was granted in 1360 by Edward Despenser.116 The charter regulated the election of officers and laid down guidelines for administering justice; rules were set for the conduct of trading, and merchant guilds were permitted. The transaction of business in the borough was by then already regulated by the ordinances of Kenfig, ostensibly drawn up in 1330.117 These ordinances are known only from a later version. The ordinances included strict controls to regulate the activities of various tradesmen, including butchers, bakers, brewers, corn-merchants, grocers, fishmongers, taverners, and tanners. Affairs were administered by the portreeve, aldermen and burgesses, assisted by officers. A hayward regulated grazing on the common land. A sergeant policed the town and kept the prison, and presumably superin- tended the rules governing the use of the borough ducking-stool. The cross-keeper regulated trading conducted at the town cross, which probably stood in High Street, the only street named in the ordinances (though Monk Street occurs c.1283 and a West Street is noticed outside the enclosed town in the 1140s). Town gates, walls and ditches are mentioned, though only in regard to regulations enforced within them and the prohibition of dumping ‘dust, dung or other filth’ within 50 feet of them. No mention of gates and walls in the ordinances (or other records) suggests that they were of masonry. Within the defences the borough had a guildhall with a prison for offending burgesses, which was set over a ground-floor ‘lower prison’ reserved for strangers. It is not clear whether the court mentioned sat in the guildhall or was accommodated in a separate building. There was a shambles, where meat was sold, which like the other trading centre at the cross and the guildhall may have stood in High Street. The detailed impression of Kenfig conveyed by these ordinances can be augmented by other records. The extra-mural settlement first noticed in the 1140s probably survived into the fourteenth century, for it contained the parish of St James, which was

110 Cal. C R, 1313–18, 162, 406; Clark, Cartae, III 113 Cal. Chart. R., iii, 461; Cardiff Records, i, 15–16. (dccclxvii), 1024; Glam. Co. Hist., iii, 73, 79, 344. 114 Cal. I P M, 23 Edward III (1349). 111 Clark, Cartae, III (dccxlvii), 835–6 (misdated to 115 Cal. Chart. R., v, 164; Clark, Cartae, IV (mxxii), 1281 by Clark, an error followed by Beresford, New 1303–5; Cardiff Records, i, 29–30; Cal. I P M, 33 Edw. Towns, 356, 555). III, vol. x, p. 416. 112 Cal. C R, 1318–23, 541–2; Cal. F R, 1319–27, 116 Clark, Cartae, IV (mlxxiv), 1411–17; Arch. Camb. 100, 189; J. Conway Davies, ‘The Despenser War in (1871), 176–85; M. Weinbaum, British Borough Glamorgan’, Trans. Royal Hist. Soc., 3rd series, 9 Charters, 1307–1660 (Cambridge, 1943), 150. (1915), 21–64; Glam. Co. Hist., iii, 170–1; Arch. Camb. 117 G. T. Clark, Arch. Camb. (1871), 243–56. (1927), 174. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 191 held at farm by Margam from Tewkesbury Abbey well into the fifteenth century, along with the chapel of St Thomas within the town. Also within the suburb was a hospital or leper house (maladaria), recorded there c.1216.118 John Giffard’s account of 1316 reports that the town gallows for hanging thieves had been newly constructed for 6d.119 In 1376 the inquisition post-mortem at the death of Edward Despenser recorded a ‘Bothale’ (booth hall or market hall?).120 A series of charters issued to Kenfig during the fifteenth century gives a deceptive impression of untroubled prosperity.121 Natural forces alone caused the abandonment of Kenfig. There is no reason to credit the claim that its decline was advanced by its destruction by Owain Glyndw^r or his followers around 1404.122 No contemporary record can be cited for this destruction, nor for the associated claim that Kenfig Castle had been revictualled in 1403 by order of Henry IV; the castle of ‘Kenflyc’ provisioned by that order was Cefnllys in Radnorshire.123 On 6 April 1405 Kenfig Castle and borough were among the Glamorgan lands granted to Joan, queen of England, with no suggestion that they had suffered damage.124 A detailed study of the evolution of the invasive sand dunes which afflicted the coast of south Wales from Pembrokeshire to Glamorgan in the fourteenth and fifteenth cen- turies has been made by L. S. Higgins.125 His study suggests that by the fourteenth century there was a critical coincidence of a serious degree of marine erosion suffered by protective coastal flats and the onset of a period of exceptionally wet and stormy weath- er and abnormal tides. In combination these factors caused the exposure of previously stable, ancient littoral accumulations of sand, activating a renewed invasive advance of the long dormant dunes from the coast. At Kenfig the first indication of a renewed advance of the sands appears in 1316, when Margam Abbey complained that a large area of its pasture there had been rendered worthless by sand, a complaint that was renewed in 1336.126 In 1365 there is the first mention of Kenfig Pool (Kenefeg is Poil), indicating that the estuary, which had sheltered twenty-four ships in the crisis of 1184, had been impeded by blown sand to take its present form as an impounded lagoon.127 The sands do not appear to have reached the town, which had recently attained its peak of 144 burgages and, with its extra-mural suburb, sheltered an estimated population of 800. That the authorities were oblivious to the impending disaster is suggested by a charter of 1397 in which the bishop of Llandaf ordered that the abbot of Tewkesbury Abbey should rebuild the parish church of St James and that the vicar should ever after keep it in repair.128

118 Clark, Cartae, II (ccclv), 355–6; Arch. Camb. 123 Cal. P R, Henry IV, vol. ii (1401–5), 180–1. (1898), 147. 124 Ibid., Henry IV, vol. iii (1405–8), 4; Cal. Anc. 119 Clark, Cartae, III (dccxlvii), 836. Petitions, 382; Gray, Kenfig, 66; Arch. Camb. (1927), 120 Cal. I P M, xiv, no. 209. 180–1. 121 Henry IV (1401): Clark, Cartae, IV (mlxxxvi), 125 L. S. Higgins, Arch. Camb., 88 (1933), 26–67 1443–5; Cardiff Records, i, 32–3; Richard Beauchamp (summarized in regard to Kenfig by F. J. North, (1421) and Countess Isabel (1423): Clark, Cartae, IV Sunken Cities (Cardiff, 1957), 114–18). (mxcviii and mcxxvii), 1490–1 and 1508–9; Arch. 126 1316: Gray, Kenfig, 21–3; Glam. Co. Hist., iii, 354; Camb. (1871), 187–90; Henry VI (1453): Clark, Cartae, 1336: Clark, Cartae, IV (dcccclxii), 1199. V (mcxcv), 1633–4; Cardiff Records, i, 43–4; Edward 127 Ibid., IV (mxxvi), 1313; Gray, Kenfig, 182; D. H. IV (1465): Clark, Cartae, V (mccxii), 1660–3; Cardiff Williams, The Welsh Cistercians, ii (Tenby, 1984), 323. Records, i, 43–4; Henry VII (1497): Cal. C R, 128 Clark, Cartae, IV (mlxxvii), 1431–3; Arch. Camb. 1494–1509, 106; Cardiff Records, iii, 23. (1898), 149; Gray, Kenfig, 86. 122 Gray, Kenfig, 65–6; Arch. Camb. (1909), 498–9. 192 C. J. SPURGEON After 1400 it is clear that the advance of the dunes relentlessly engulfed the town and its lands, as documented in the last quarter of the century when the process was all but complete. By 1480 the Via Regalis had been abandoned and Kenfig was by-passed by a new main highway constructed further north through Pyle and Cwrt-y-defaid near Margam Abbey (the present A48).129 At Pyle on the new road, 2.4 kilometres east of Kenfig, a second church of St James was newly dedicated by 12 August 1485 to replace the old parish church at the town.130 There is an undocumented but not improbable assertion that the old church of St James had been abandoned by 1428.131 Between the latter year and the order of 1485 the citizens of Kenfig could have resorted to the chapel of St Mary Magdalene, established c.1250 on the ridge only 800 metres south-east of the town, where some burgesses were certainly resettled. St Mary’s still serves the community of Mawdlam, and it is credible that its Norman font and two altar slabs found in its yard were brought from old St James; a floriated tomb-slab retrieved from that site is now at Margam Abbey, and other dressed stone fragments have been recorded at Kenfig Farm.132 By 1500 pasture and demesne lands were overblown, as well as most burgage plots; a windmill was abandoned, and in 1487 a corn-mill and the tolls of the piepowder court had been farmed out. Notwithstanding resettlement at Mawdlam and Pyle, a revised rental yielded greatly reduced returns from burgage rents in 1493, and by 1516–17 total income was one-fifth of that yielded in 1314–15.133 The full extent of the desolation is conveyed by Leland in the late 1530s: ‘There is a little village on the est side of Kenfik [river] and a castel, booth in ruine and almost shokid [choked] and devourid with the sandes that the Severn Se there castith up.’134 The desolate scene was similarly noticed later in the century: Kenfig, a borough town sometime of good account but long since decayed by over- flowing of the sand . . . an old church and castle overblown . . . so that little of the castle is now to be seen. (Merrick, c.1580); Castle and the towne were drowned by the sand of the sea . . . there remayneth but out curtledges bering the name of the Towne of Kenfigg which hath the whole liberties yet remayeninge . . . saving the weekely markettes and annual faires. (Lewis, c.1597).135 The fifty-second clause added to the Kenfig ordinances of 1330 when they were repro- mulgated in 1572 noted that only three of the burgages in the town had not been engulfed by sand and arranged for eight burgesses to enclose, ditch and allot twenty- nine new burgages on Cefn Cribwr Common to the east of Pyle.136 The 1572 clause also

129 North, Sunken Cities, 116, and Higgins, Arch. 134 Leland, Itin. Wales, 29. Camb., 88 (1933), 34, 64. 135 Merrick, Morg. Arch., 101; Rice Lewis, A Breviat 130 Clark, Cartae, V (mccclxix), 1916–17; Gray, of Glamorgan, 1596–1600, ed. W. Rees (Newport, Kenﬁg, 94–5. 1954), 105. 131 Higgins, Arch. Camb., 88 (1933), 64. 136 Clark, Arch. Camb. (1871), 243, 245, 252–4; 132 Gray, Kenﬁg, 83–4, 90–3; Arch. Camb. (1898), Higgins, Arch. Camb., 88 (1933), 34–5. 132–6. 133 Glam. Co. Hist., iii, 354; Clark, Cartae, V (mccl), 1733–4. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 193

FIGURE 8. Kenfig: aerial view of the castle and borough earthworks. emphasized that with great difficulty the full rent due to the lord had always been paid, confirming Lewis’s statement that the resettled burgesses in their ‘out curtledges’ beyond the abandoned town continued to maintain the name and corporate liberties of Kenfig. In 1665 a survey of the dispersed borough noted that only one cottage beside the site of the old castle remained in the old town, and by the eighteenth century corporate affairs were conducted in a guildhall on the upper floor of the Prince of Wales Inn, Mawdlam, which still had custody of the borough’s silver mace and copper ale-taster’s pint measure in 1898, following the abolition of the corporation of Kenfig in 1886 under the Municipal Corporation Act of 1883.137 The borough occupied a large enclosure of roughly quadrangular form extending to the south of the castle, which was set in its north angle. Two unbroken sections of the town rampart are visible (Figures 8 and 9). A straight section runs for about 76 metres along the south-west flank and formed a square angle with a further section running parallel with the river for 65 metres, recorded in 1950 but no longer visible. The east-

137 A. L. Evans, The Story of Kenﬁg (Port Talbot, 1960), 39–40; R. W. Llewellyn, Arch. Camb. (1898), 136–8, illustrating the mace. 194 C. J. SPURGEON

FIGURE 9. Kenfig: plan of the castle and borough. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 195 ern section of rampart may be traced for 140 metres as it defines a gently curved angle. Only on the south-west is there any trace of the external ditch. An extensive and lofty dune separates the surviving lines of rampart, and lesser dunes have masked all traces of the sections linking with the castle. During his excavations of 1924–32 A. J. Richards cut a section through the better preserved south-west defences. The rampart was com- posed of gravel and glacial or river pebbles derived from the ditch, which was 14.70 metres wide from lip to lip and 4.30 metres deep. No trace of the palisade mentioned in the records of 1183–4 was discerned, and there was no evidence to suggest that it had ever been replaced by a stone wall. Richards also ascertained that the ditch continued along the north-west flank, although the rampart on that side was set close to the parallel course of the river and would have joined the castle moat, which may have originally tapped the river to the north side.138 The enclosure surmised by projecting visible and recorded sections of the rampart would have enclosed 3.40 hectares (8.25 acres); this exceptional size precludes acceptance of the enclosure as the castle bailey.139 St James’s church is located at SS 8006 8240, 267 metres south of the castle and 85 metres south of the calculated position of the southern angle of the town rampart, now lying beneath the large dune. Carlisle observed ruins of the church here in 1811 and ‘great quantities of human bones’ cast up in the drifting sand. By 1898 Llewellyn could only recover fragments of dressed stone, which included a column base of Sutton stone of early character, and he also noticed many human bones in the sand.140 The only other recorded evidence relating to the medieval suburb was furnished by Lethbridge and David. In 1928 they reported traces of medieval dwellings exposed in the sands 175 metres south-west of the south-west rampart at SS 7984 8247. Finds from the spot included green-glazed and reddish-brown medieval sherds, iron nails, bones, scraps of bronze, and iron arrow-heads. There were also hearths and Romano-British sherds, and subsequently a sixth-century penannular brooch and a Roman coin were recovered.141 Within the town defences ‘old house floors’ are said to have been observed in the shifting sands.142 The footings of at least three rectangular buildings have been recorded from the air as parch marks centred at SS 8002 8258 to the rear of the south-west rampart (Figure 9). No visible or recorded evidence reveals the location of the town gates. The causeway crossing the ditch at the west angle is probably modern; the site was occupied by military installations during the 1939–45 war. Traces of masonry cut- ting across the line of the north-east ditch at the termination of the surviving line of rampart were initially interpreted as the traces of an entrance, but subsequent explo- ration disclosed a domestic range of four units raised in stages. This was probably a sixteenth-century farm house, and may be identified with the only dwelling remaining at the borough ‘on the site of the old castle’ in 1665.143

138 A. J. Richards, BBCS, 2 (1924), 264, and 6 Camb. (1898), 135–6. (1931), 98; Arch. Camb. (1927), 165 and fig. 4, 168. 141 Arch. Camb. (1928), 200–1. 139 For example, Gray, Kenfig, 58–9 (and plan 142 Ibid., 378. between); I. Soulsby, The Towns of Medieval Wales 143 A. J. Richards, BBCS 6 (1931–3), 98, 294–5; D. (Chichester, 1983), 151. Jones and I. Soulsby, Arch. Camb., 126 (1977), 145–7. 140 Carlisle, Top. Dict., s.n. Kenfig; Llewellyn, Arch. 196 C. J. SPURGEON Neath

The ambiguous mention of a new town (nova villa) at Neath in the foundation charter of Neath Abbey (1130) is probably the earliest reference to the settlement south of Earl Robert’s castle on the east bank of the Nedd. The charter notes that the town was protected by a ditch, which implies that it also had an internal palisaded rampart. No traces remain of the town ditch, rampart and gates, and only one possible medieval documentary reference is known for any town defences later than those implied by the charter of 1130. Some authorities have been reluctant to accept that Neath was ever a fortified borough.144 However, given its remote location on the western boundary of the lordship of Glamorgan and its turbulent history, earlier writers were probably correct in regarding Neath as a defended town.145 Although no specific documentary record asso- ciates Earl Robert with the castle and ‘new town’, his appropriation of the land on the east side of the river is attested at that period by his grant of adjacent land at Briton Ferry to Neath Abbey.146 Neath had attained the status of a borough under Earl William (1147–83), who granted its burgesses a charter which endowed them with all the liberties and customs enjoyed by his burgesses of Cardiff.147 On 29 June 1231 Morgan Gam joined Llywelyn ab Iorwerth in an attack on Neath. The castle was destroyed and, by some accounts, the town was burnt and its inhabitants exterminated.148 The castle and town were presumably restored by July 1244, when the Neath garrison expelled a party of Welsh raiders which was plundering the lordship. In 1246 the bailiff of Neath was instructed to ensure the compliance of named Welshmen with terms agreed with Margam Abbey in restitution for damages inflicted on the property of that house.149 In September 1258 Llywelyn ap Gruffudd attacked Neath; the castle managed to resist the attack, but the town was not able to withstand it, and its mill and 150 houses were destroyed.150 The town recovered under Earl Gilbert de Clare II, and at the time of his widow’s death in 1307 there were 128 occupied burgages at Neath.151 Before the end of July 1314, however, a Welsh attack on Neath had reduced eighty burgages to ashes.152 Payn Turberville of Coity became royal custodian of Neath in 1315. His account for the period to 20 April 1316 was drawn up by his widow, Gwenllian.153 A further revolt led by Llywelyn Bren disturbed Glamorgan early in 1316, but it is not possible to discern in Gwenllian’s account whether further damage was inflicted in Neath town. She records reduced receipts from tolls, the prise of ale, and adds that certain rents from extra-

144 For example, I. Soulsby, Towns, 190; A. Corbett, Glamorgan Lordship, 68, 133. Hopkins, Medieval Neath, Ministers’ Accounts, 149 ‘Cardiff Chronicle’ (s.a. 1244); Clark, Cartae III 1262–1316 (Pontypool, 1988), 13–14. (dxxvi and dxxvii), 532–6; Epis. Acts, ii (L 437–8), 145 Lewis, Top. Dict., s.n. Neath; Clark, Cartae, I, 76, 726–8. and Land of Morgan, 133. 150 Ann. Theokes., 197; Glam. Co. Hist., iii, 337; Arch. 146 Clark, Cartae, II (cccxviii), 315; IV (dcccclxiii), Camb. (1887), 101; Clark, Land of Morgan, 116; 1201; and V (mccxx), 1681; Patterson, Glouc. Charters, Corbett, Glamorgan Lordship, 135. no. 242, 172 (citing Dugdale, Monasticon, v, 259; Rot. 151 Cal. I P M, 35 Edward I (1307), no. 455, Chart, 174). pp. 322–4; Beresford, New Towns, 67, 556. 147 Clark, Cartae, IV (mlxxv), 1418; Patterson, Glouc. 152 Hopkins, Medieval Neath, 33–43 (the first Charters, no. 159, p. 148. account) and 43–66 (the second account). 148 Bruts, B S, Pen 20, R B H; Cronica de Wallia 153 Ibid., 65–6, 73–82; Ministers’ Accounts, PRO, SC (BBCS, 12 (1946), 27–44); Clark, Land of Morgan, 94; 6/1202/8, m.3. FIGURE 10. Neath: possible course of the town defences. 198 C. J. SPURGEON mural tenants had not been received. Besides works at the castle, she also accounted for repairs to the ‘great gate of the bailiwick’, which may be the only other medieval record of the town defences. Neath was plundered in the ‘Despenser War’ of 1321,154 but the damage was presumably repaired to a great extent by 1326, when the restored Neath Castle sheltered Despenser and Edward II, 5–10 November.155 By 1331–2 total income derived from the town matched that produced in 1314, and most of the burgages destroyed in that year had been restored.156 By 1349 the total of occupied burgages had risen to 150.157 Neath Castle was garrisoned in the years 1402–4 during Glyndw^r’s rebellion, but there is no evidence that hostilities were suffered at that period. For Neath, as for the other Glamorgan boroughs, the fifteenth century was a period of declining prosperity. By 1491–2 the town hall (‘Bottehall’) was derelict, the sea had inundated the borough’s meadows and pastures, and fee-farm rents were respited due to poverty. In 1514 market and fair tolls were no longer collected, and the 1521 ordi- nance preventing Welshmen from the uplands becoming burgesses was a sign of des- peration.158 Leland noticed the ‘litle toun and castelle’ of Neath in the 1530s, where ‘there cummith up shippelettes almost unto the toun . . . to the very bridge of tymbre that is sumwhat lower on the water then the toun’.159 In contrast to the besanded and abandoned port town of Kenfig, Neath maintained its maritime trade, not always legally; in 1572–3 the portreeve was prosecuted for importing wine and other merchandise without paying duty.160 Neath came to Sir William Herbert in 1551, and his successors held it until 1751.161 A survey of 1666 notices coal mines belonging to the burgesses, some of them recently opened. Coal had been exploited at Neath since medieval times, but these new mines heralded the subsequent industrial development of the eighteenth century and the rise of the modern town. Kenfig Borough provides a model for how the early defence of the new town at Neath might have been disposed. At Kenfig a roughly quadrilateral town was enclosed by a rampart with a deep external ditch, the castle being set in one corner. Similar defences probably enclosed Neath Borough from the outset until the mid-thirteenth century, when its palisade was replaced by stone walls, a conversion which was never carried out at Kenfig. Early plans of Neath depict the town as a compact grid of streets with very little expansion beyond a pentagonal core to the south of the castle. This ancient nucleus was bounded by Angel Street and New Street to the south-west, Wind Street to the south-east, Mackworth Lane and Gold Street to the east, the castle to the north, and James Street to the north-west. The plan (Figure 10) illustrates the ancient lay out before it was greatly altered during redevelopment on the north-west side in the vicinity of James Street

154 Davies, ‘The Despenser War’, 45–6, 54–5; Rees, 158 Glam. Co. Hist., iii, 353–4. Caerphilly Castle, 56–60; Cal. C R, Edward II, 1318–23 159 Leland, Itin. Wales, 30. (London, 1895), 541–2. 160 Exchequer Proceedings (Equity) Concerning Wales, 155 Phillips, Vale of Neath, 55; Rees, Caerphilly Castle, Henry VIII–Elizabeth, ed. E. G. Jones (Cardiff, 1939), 79–80. 253. 156 Glam. Co. Hist., iii, 353 (citing PRO, SC 161 Merrick, Morg. Arch., 76, 106, 138, records the 6/1202/5). governance of the borough by the portreeve (1580). 157 Ibid., 643, n. 65; Beresford, New Towns, 66, 556. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 199 and the clearance of houses formerly clustered around the castle to the north. Before the latter clearance the houses backing onto the castle defined a lunate line concentric with the castle along the line of Castle Buildings and Castle Street. It is probable that this curvilinear line marked the area of a castle bailey. Within the gridded nucleus of the town, the medieval parish church of St Thomas the Martyr occupies a large rectangular plot bounded by Church Place. The broad Old Market Street, parallel with the north-east side of Church Place, was the main commercial thoroughfare. As late as 1811 a Gnoll estate plan marked the market hall in the centre of this street, close to its junction with High Street.162 At their outer ends on the periphery of the old core Old Market Street and High Street respectively joined Wind Street to the south-east and Gold Street to the east. Along with New Street, on the south- west periphery, High Street is mentioned in a sale of chantry lands of Neath church in 1557, as are probably Wind Street and Gold Street if they may be identified with ‘Wyn Srete’ and ‘Gilden Strete’.163 Limited linear development beyond the old nucleus on the lines of High Street, both ends of Wind Street, and of Old Market Street had extended further by 1832, when Samuel Hosgood produced a plan of the town.164 The most significant of these devel- opments fringed Water Street, which projected the line of Old Market Street to the south-east. A stream formerly running through the town, presumably on the line of Old Market Street and long culverted, is said to have given Water Street its name.165 The town walls at Neath, like the first masonry defences of the castle, were probably raised in the time of Earl Richard de Clare, between 1243 and 1258. Excavations in front of the castle gatehouse in 1962–3 disclosed two distinct phases of structural development at that entrance and for the town wall which joined its south tower. Before the excavations much of the lower part of the town wall was buried in the garden behind a derelict house, and only a short section was visible, rising as tusking to its wall-walk and parapet. In the first phase a flight of stone steps incorporating a small postern gate descended the scarp from the castle, enclosed between the outer face of the town wall and a single salient north-west tower which later became the north tower of the gatehouse. The castle was severely damaged by the baronial opposition in 1321. Between 1321 and 1326 the castle gatehouse was created over the truncated remains of the salient tower and the postern gate, with an entirely new flanking south tower raised upon the inner part of the town wall flanking the postern steps and the footings of the plain curtain with which it had formerly linked. The postern gate was blocked and the steps above it were buried. The lower steps beyond the gate were retained as a drawbridge pit in front of the new gateway and closed off by a cross-wall inserted near the foot of the steps. Beyond the postern gate, which was set about mid-way down the steps, the outer face of the town wall displayed a blocked door which must have opened into the bottom of the ditch surrounding the inner ward and given access from there to the suspected castle bailey facing the town.

162 The plan in Glamorgan Record Office is repro- 164 Trans. Neath Antiquarian Soc. (1980–1), pl. 41, duced in Glamorgan Historian, 9 (1973), between 141. 160–1. 165 Soulsby, Towns, 191. 163 Clark, Cartae, V (mccccxviii), 2013–17. 200 C. J. SPURGEON In eighteenth-century depictions of the castle the town wall is clearly portrayed projecting from the south tower of the gatehouse. In 1741 the Buck brothers showed it descending from the flank of the gate and curving south-westwards to face the river, which then ran much closer to the castle and town (Figure 10). Anthony Denis (c.1770) also shows the town wall and the more adjacent course then followed by the river. A view by J. Warwick Smith (c.1790) shows the former proximity of the river, and another by George Bradshaw (1807) depicts only a short stub of the town wall remaining above ground.166 G. T. Clark accepted that Neath had been a walled town on the evidence of the fragment attached to the castle gatehouse.167 An observation by W. Weston Young in the early 1830s seems to provide the only other record of the town wall and a gate on the south-east side: The town gate, through which this road (Wind St. to Briton Ferry) passed, was standing about sixty years ago; it stretched from the corner house of the street . . . across to some small cottages, then standing where the court in front of the great house now is; the foundation on the south-east side, together with a part of the old town wall, is still remaining.168 This claim suggests that there was a gate standing somewhere along the line of Wind Street around 1770 and that foundations and a contiguous section of wall were visible in the early 1830s. The alignment of Wind Street, however, is south-west–north-east, form- ing the suggested south-east flank of the old nucleus of Neath and perpendicular to the obvious main road, which was directed towards Briton Ferry on Hosgood’s map of the town in 1832. The main road was Water Street, which projected the line of Old Market Street beyond the old core; between them these aligned streets crossed the central part of Wind Street. It is possible that the gate stood at this crossing, facing south-east towards Briton Ferry. Such a gate would have constituted the main entry to the town and was possibly the ‘great gate of the bailiwick’ recorded in 1316, its approach road from the south providing the vital link with the main highway, the Portway, at Briton Ferry. From the castle gatehouse a road probably led west along the riverside to a bridge about 250 metres downstream. On the west side of the bridge, which was of timber in Leland’s time (c.1539), were probably the tenements held by the Neath burgesses in the early thirteenth century. Nearby, at SS 7497 9787, was the chapel of St Giles, which survived as an agricultural building until 1863, when it was demolished during the con- struction of the railway station, though Sutton stone dressings were retrieved and set in a boundary wall to mark the site.169

Swansea

Swansea Castle and the town which soon sprang up beside it were sited to take advan- tage of Swansea’s natural harbour, occupying an elevation with a strong in-curving

166 Both reproduced in Glamorgan Historian, 1 168 Phillips, Vale of Neath, 53. (1963), between 48–9. 169 Arch. Camb. (1861), 344; Phillips, Vale of Neath, 167 Clark, Cartae, I, 76, and Land of Morgan, 133. 86–8. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 201

FIGURE 11. Swansea: plan of the town defences. 202 C. J. SPURGEON natural scarp to the east, facing the River Tawe. The castle was sited at the highest and steepest central part of the escarpment; set against the eastern scarp, the position fell gently away on other sides, where the enclosed town eventually developed. The curved natural escarpment towards the river imposed an irregular shape on the borough, which was compounded to the south-east by the long-vanished tidal creek, the Pill, which defined and protected that flank of the settlement (Figure 11). Industrial and commercial development in the nineteenth century and devastating bomb damage in 1941 almost entirely erased visible traces of the medieval town. However, it is possible to deduce the extent, characteristics and layout from early descriptions, plans and drawings, which were to a large extent confirmed and often enhanced in a series of excavations undertaken in advance of the extensive post-war reconstruction. Although Swansea did not receive its first formal charter of privileges until after 1166, it is probable that the town was well-established and had its burgesses during the time of Henry de Beaumont (c.1107–19). In 1116 Gruffudd ap Rhys pillaged the settlement and destroyed the bailey of the castle.170 Henry de Gower (Gouher) had his caput at the restored Swansea, where he minted coins around 1140.171 Earl William (c.1166–84) was the first to give a charter of liberties to Swansea’s burgesses.172 There is no mention of any borough defences, but given the attention paid to military matters in the charter it is likely that Swansea was enclosed at an early date by a palisaded rampart and ditch like those at Kenfig. In 1192 Swansea Castle was besieged for ten weeks by Rhys ap Gruffudd. The siege was only broken by the dispatch of a royal force and supplies from Bristol.173 The Welsh chronicler cited states that it was the town which was besieged and that the townspeo- ple (oppidanos) were only saved from starvation by dissension within the Welsh forces, which led them to abandon the siege. If this account is true, it is the first incontestable evidence for town defences. In 1212 the Margam chronicler records the burning of Swansea by Rhys Gryg.174 The town recovered and was granted an important charter by King John in 1215.175 Before the year was out, however, Rhys Ieuanc advanced on Swansea, where the garrison is said to have burned the town for fear of him – presumably to deny the attackers shel- ter and sustenance and to assist their more effective defence of the castle. The following forty years were relatively untroubled and the town began to prosper. However, in 1257 Rhys Fychan ravaged Gower and burned the town of Swansea (Sweynesey).176 In 1287 Rhys ap Maredudd also burned it.177 An inquiry respecting alienated property in the lordship undertaken in 1319 shows that William de Braose II (1241–90) had granted out both gatehouses of the outer bai-

170 Bruts, B S, Pen. 20, and R B H; Glam. Co. Hist., 93, 113–14, 122, 148, 172; B. Morris, Swansea Castle iii, 361. (Swansea, 1992), 14–17. 171 Boon, Welsh Hoards, 49–50, 53–4. 174 Ann. Margam, 32; Glam. Co. Hist., iii, 220. 172 Clark, Cartae, I (cxxxviii), 136–8; A. Ballard 175 Ballard, British Borough Charters, xxxi, cxlv, 216. (ed.), British Borough Charters (1042–1216) 176 ‘Cardiff Chronicle’, (s.a. 1256). (Cambridge, 1913), xxxi, cxlv, 19, 38, 47, 51, 54, 58, 177 Annales Cambriae, 109; Clark, Cartae, III (dcclvi- 63, 65, 83, 89, 118, 137, 140, 145, 283, 249. ii), 860; Arch. Camb. (1862), 281; R. A. Griffiths, ‘The 173 Annales Cambriae, ed. J. Williams ab Ithel (Rolls Revolt of Rhys ap Meredudd, 1287–88’, Welsh History Series, 1860), 58; Pipe Rolls, 5 Richard I (1192), 78, Review, 3 (1966–7), 129; Glam. Co. Hist., iii, 229–30. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 203 ley to serve as burgages. The northern gate of the bailey, granted to Walter le Pecar, had two flanking towers.178 It is possible that the primary palisaded defences, which probably enclosed the town, had been given masonry walls and gates by the close of the thirteenth century. This would explain why de Braose was confident enough in the secu- rity of his castle to grant out the two gates of his bailey. By 1306, when William de Braose III granted two charters, one to his tenants in Gower, the other to the burgesses of Swansea, it may assumed that the town had fully recovered from the Welsh attack of 1287.179 The confidence of the urban community was reflected in 1317 by a grant of murage and pavage from Edward II. With this grant the burgesses were empowered to levy tolls on merchandise entering the town, the revenue from which was to be used in the upkeep of the town walls and the maintenance of its roadways.180 The period from the baronial rising against Despenser to the abdication of Edward II was an unsettled one, but with the succession of John, Lord Mowbray (1331–54), to the lordship in 1331 some measure of stability was restored to Swansea. In 1332 Henry, bishop of St David’s, founded the Hospital of the Blessed David within the town. The foundation charter provides the earliest evidence for many features of the medieval town.181 The hospital was situated against the town wall, and it was endowed with properties within and without the town, including several burgages against the town wall, two-and-a-half burgages outside Harold’s Gate, and burgages in Fisher Street. The Cross Keys public house incorporates the only surviving fabric of one of the medieval hospital buildings; Harold’s Gate was probably one of the gates of the castle bailey; Fisher Street survived until the blitz and now lies beneath the Quadrant and other modern redevelopment at the south end of the town. In 1338 the burgesses obtained a further grant of murage and pavage from Edward III, with permission to levy tolls for the maintenance of the town’s walls and streets for five years.182 Despite famines and the Black Death, Swansea appears to have prospered for much of the fourteenth century. In July 1400 King Henry IV assigned a third part of Gower to Countess Elizabeth in dower. A valuation of the lordship was made for the purpose of assigning dower and included a detailed list of the lands and rents, including the borough of Swansea. The receiver’s account compared to one for 1376 showed reduced receipts from the corn- mill, tolls and the hundred court.183 Tradition claims that Swansea was taken and ravaged by the Welsh under Glyndw^r, but there is no record of such an attack. The first half of the fifteenth century saw reductions in long-established revenues, but there is reason to believe that they were compensated for by an expanding and lucrative trade in cloth produced in four of five fulling mills, two of them newly built.184 The later

178 Clark, Cartae, III (dcccxciii), 1071. Petitions, nos. 2062 and 5249, pp. 57–8, 163–4. 179 Ibid. (dcccli), 990–9; A. Ballard and J. Tait, 182 Cal. P R, 1338–40, p. 6. British Borough Charters, 1216–1307 (Cambridge, 183 Jones, Swansea and Gower, ii, 28–30; Glam. Co. 1923), pp. xxxiii, c (see index for treatment of its Hist., iii, 371–4. many clauses). 184 Clark, Cartae, IV (mcli), 1552–3; Jones, Swansea 180 H. Turner, Town Defences in England and Wales and Gower, ii, 56–7; W. R. B. Robinson, BBCS, 22 (London, 1971), 218; Cal. P R, 1317–21, 59. (1967), 169–98; Glam. Co. Hist., iii, 374–5. 181 Clark, Cartae, IV (dccccliv), 1180–3; Cal. Anc. 204 C. J. SPURGEON records for the town down to the Acts of Union (1536–43) suggest variable prosperity and some degree of administrative corruption.185 After the Acts of Union Swansea continued to function as an important market town for the western part of the new county, and with its port it was able to exploit its sea trading and, particularly, its export of coal, which increasingly supplemented its tradi- tional exports of butter, cheese, leather and wool. The coal trade made Swansea the busiest port in Wales by the seventeenth century.186 With the exception of the buildings of its south-western quarter, all standing fabric of Swansea Castle had been cleared away or encased in later work by the early nineteenth century. By that period the town walls seem to have suffered a similar fate. Subsequently, periodic sightings of isolated dispersed fragments of the castle and town defences occurred, usually during structural development or earth-moving operations. Such sightings and early maps of the town permitted antiquaries like Colonel W. Ll. Morgan to suggest the probable layout of the medieval castle and town. To a considerable degree their ideas have been vindicated by excavations and observations during the great rebuilding of the 1970s and 1980s. The full extent of the enclosed town has been established, and the line of the town wall and external ditch is known for almost the entire circuit. Its precise course remains to be demonstrated only to the south-east, as it curved eastwards to join the South Gate at the bottom of Wind Street.

The medieval town The irregular plan of the medieval town (Figure 11) was imposed by the incurving eastern scarp towards the river. After the building of the castle on the highest point of the scarp, c.1107, the primary settlement developed along roads to the north (High Street) and the south-east (Wind Street), extending from the castle gates and reflecting the curve of the scarp. To the south-east of the castle the ground sloped down, and the natural eastern scarp decreased and faded out near the Pill. This long-vanished creek, also known as the Cadle, provided a natural line of defence when the town was enclosed. To the west and north of the castle no such natural feature dictated the lines taken by the northern and western walls, and a more regular trace was followed. To the north the natural scarp continued as a prominent feature well beyond the castle, and the north section of the town wall was set square to that scarp with a central North Gate closing High Street. From a squared north-west angle the town wall ran south-south-west in a straight line to the West Gate, near but beyond the north-west corner of the castle bailey. Another longer straight line of wall ran south from the West Gate to complete the western line. Two shorter facets of the town wall flanking the Wassail Gate then began the curving southern line completed along the north flank of the Pill, with the South Gate set at the lower end of Wind Street, close to the angle made with the eastern wall of the town, which here ran north-west along the crest of the natural scarp to the castle and beyond.

185 Jones, Swansea and Gower, ii, 64–6, 70–2; W. R. 1864–70. B. Robinson, ‘The Welsh estates of Charles, Earl of 186 Glanmor Williams (ed.), Swansea: An Illustrated Worcester in 1520’, BBCS, 24 (1971), 384–410; Glam. History (Swansea, 1990), 12–17. Co. Hist., iii, 280–3; Clark, Cartae, V (mcccxliv), THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 205 The circuit outlined above differs in one essential from that proposed by Colonel W. Ll. Morgan. He believed the town defences to the north were formed by the north wall of the castle bailey, projected westward for the short distance to the established line near the West Gate and excluding the northern area flanking High Street.187 While he erred in excluding the northern sector, Morgan’s proposed line for the remainder of the circuit is largely confirmed. From the late thirteenth century land within the large rectilinear bailey of the castle was being leased out to burgesses, some of whom held its towers and gatehouses along with their tenements. The most significant holders of such tenements were the Hortons, who acquired the south-west quarter of the enclosure and whose successor, Sir Mathew Cradock, raised Place House there in the late fifteenth century. Originally, High Street and Wind Street were not linked as a main broad thoroughfare, as now, by Castle Street and Castle Bailey Street within the precincts of the castle. This central link was only made possible after the unrecorded demolition of the north and south gates of the castle. Before then general traffic was probably diverted around the west side of the castle defences. The medieval street pattern was considerably altered during the redevelopment necessitated by the 1941 bombing. The plan (Figure 11) marks some streets which no longer exist, or have been replaced by wider and divergent new streets. Princess Way, for example, runs south from the subway circle, set over the site of the West Gate, on a line west of the former Frog Street. Pre-Second World-War streets reflecting the medieval lay out are shown; of these, High Street, St Mary Street and Fisher Street were recorded in 1400, and Gaer Street, Frog Street, Cross Street and Wind Street were certainly early. The present St Mary’s church is a rebuild of the Victorian church bombed in 1941, but it occupies the medieval site. The Cross Keys public house is much restored, but it incorporates medieval fabric, including early fourteenth-century lancet windows, and it is accepted that it represents the only surviving fabric of the Hospital of the Blessed Mary. Excavations on the north and north-east sides of the Cross Keys in 1976–7 identified what seems to have constituted the ‘priests’ garth’ or garden.188 The town was probably protected by defences of earth and timber from an early date, but no documentary or archaeological evidence is available for them. If the primary defences did not extend to the full circuit of the later walls, the alignment of Cross Street, Fisher Street and Little Wind Street appears to be a possible line for the original defensive works. It has also been suggested that the castle bailey constituted the first town enclosure, but the overall area of Swansea Castle was insufficient to accommodate the early town as well as the official houses of the knights in the bailey. The circuit of walls appears to have existed before 1317, on the assumption that the murage grant from Edward II in that year would probably have been applied to repairing existing walls. The walls are first recorded in the founding charter of St David’s Hospital in 1332, and a further murage grant was made by Edward III in 1338. The Welsh attack of 1287 may have prompted the building of the town walls. The location of post-Second-World-War sightings of the wall are marked on the plan

187 Morgan, East Gower, 92–3 and plan, p. 86. 188 S. H. Sell, Arch. in Wales, 33 (1993), 1–16. 206 C. J. SPURGEON

FIGURE 12. Swansea: tower on the East Wall.

(Figure 11), with crosses and the date of each sighting. The map also shows the former course of the River Tawe, which formed a bold loop which was by-passed by a New Cut dug in 1842–52, leaving the old channel to serve as the North Dock until its closure in 1928. The East Wall, where it ran north-west from the Mount, was brieﬂy disclosed as massive stone walls during rebuilding at the rear of the Post Ofﬁce in 1963.189 Turning to a more northerly course beyond this point, the wall probably followed the crest of the scarp to join the castle, which protected the central sector of the line between Castle Lane and Welcome Lane. Beyond Welcome Lane the town wall continued the line of the east castle wall in two straight facets along the scarp to Morris Lane (formerly King’s Lane). The facets respect the in-curving northerly course of the natural scarp, chang-

189 B. Morris, in R. A. Griffiths (ed.), The City of Swansea; Challenges and Change (Stroud, 1990), 147. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 207 ing direction at the centre of this section at SS 6567 9330, where there are the vestiges of a projecting tower which rises almost 6 metres from the lower level behind the Strand (Figure 12). The tower is integrated with a section of the northern end of the southern facet of the town wall. Only the south flank of the tower survives, damaged and patched but retaining putlog holes and one side of a splayed loop about 3 metres above the base, which was probably solid. The outer south angle of the tower is rounded below, but faceted at the level of the east-facing loop. The tower was first recorded in 1925 as ‘half a round tower, apparently circular at the bottom and polygonal at the top’.190 The North Wall followed a line along the southern sides of Morris Lane and King Street, with the North Gate set on High Street at the centre. The north-east angle of the defences and the south side of Morris Lane are defined by a modern boundary wall which rises on the lower courses of the medieval town wall. When noticed in 1925 the lower, medieval, fabric of the wall in Morris Lane was said to have incorporated arched recesses.191 No longer apparent, the recesses may have been latrine or drainage outfalls. To the west of the site of the North Gate the town ditch was disclosed during the rebuilding of 27–9 High Street, running back along the north side of King Street.192 The ditch ran east-west beneath King Street, with its outer lip set 6 metres within the line of the frontage of the new Bejam store on the north side of the street. The centre of the ditch was traced for about 20 metres in the sinking of the front wall of that store, and the outer lip in a separate excavation. The ditch was V-shaped, 3 metres deep below the present level and about 10.5 metres wide. It represented a secondary deepening and widening of an earlier ditch centred about 2.5 metres further south. The town wall was not discerned, but the projected angle of the inner scarp at the bottom of the ditch suggests that its line would have followed that of the present south frontage of King Street. The West Wall ran south-south-west from a right-angled turn at the west end of King Street, following a straight course as far as the West Gate. The latter gate stood at the east end of the Kingsway, at a point occupied by the western slope of the large, post- Second-World-War circular subway.193 The wall followed a course along Orchard Street, widened and realigned since the blitz, before turning south to run down the east side of the former Waterloo Street, once called ‘Old Walls’ and entirely built over since the Second World War. Its line crossed the present Oxford Street close to Marks and Spencer, and further south it was revealed during excavations in 1978–9 on a site now occupied by the west end of Littlewoods store to the east of Whitewalls, a purely post- war street.194 The excavations disclosed a wall 1.9 metres thick and surviving to a height of about 1 metre. The wall consisted of an outer part, built with local sandstone set in clay and 1.4 metres wide, and an internal sleeve, 0.5 metres thick, constructed of flat angular stones. The internal part of the wall may have been secondary, added like that at Castell Coch to provide or widen the wall-walk behind the crenellations. One crenel- lation survived here on a fallen block of masonry found in the ditch. The ditch fronting the wall at this point was found to be about 6 metres wide and 2 metres deep. It showed

190 W. H. Jones in RISW, Report (1925–6), 31–3. 193 Morris in Griffiths (ed.), City of Swansea, 147. 191 Ibid. 194 K. Lightfoot, Gower, 30 (1979), 76–9; 192 B. Morris, Gower, 26 (1975), 11–15, and Arch. in Morgannwg, 23 (1979), 89–90; Glamorgan–Gwent Wales, 15 (1975), 60. Archaeological Trust, Annual Report (1978–9), 39–43. 208 C. J. SPURGEON evidence of periodic cleaning. A late medieval building set against the rear of the wall sealed occupation debris which included green-glazed and unglazed sherds of late thirteenth- and early fourteenth-century types. After the excavation further sightings of the wall and ditch were noted to the north when foundations for the new store were sunk. The town wall turned south-east near the north-west corner of St Mary’s churchyard, where a sewer trench cut in 1973 revealed a section of its mortared rubble core. The south-east sector commences with the part of the wall which skirted the south- east boundary of St Mary’s churchyard. Immediately west of this, in 1972, B. Morris recorded a strong wall of sandstone set in lime mortar, which had been exposed in pre- liminary works for a new sewer.195 The wall survived to a depth of 1.80 metres or more below the present level, and it incorporated an arched outfall for a drain or latrine. Its line corresponded with the former west boundary of the churchyard, moved slightly eastwards c.1900. In 1973 the sewer trench disclosed further sections of this wall, including the mortared rubble core exposed near the lych gate at the north-west corner of the churchyard. The town ditch fronting the wall was revealed during the sinking of the foundation trenches for the outer wall and colonnade of the C and A Modes store in 1974. The outer lip of the ditch extended along almost the entire length of that colonnade. A partial section was drawn linking the outer lip with the wall and showing that it had been about 10 metres wide and well over 3 metres deep. The Wassail Gate, which stood at the bottom of Frog Street (now St Mary’s Square), was maintained by the corporation until the eighteenth century. At this gate the con- tinuation of the town wall turned to follow a curve to the east-south-east on a line along the north side of Rutland Street, which extends along the in-filled town ditch. In 1980 the outer lip of the ditch was recorded at two points on the south side of the street (SS 6562 9282 and SS 6570 9279), and the inner lip on the other side mid-way between those points.196 The line of the wall beyond the central part of Rutland Street is uncon- firmed, but its approximate direction is dictated by the curved course of the Pill, long- vanished but marked on the plan (Figure 11). Excavations were undertaken in 1976 on the south side of Little Wind Street and between that street and the Pill (now Harbour Road) in advance of redevelopment.197 The excavations discounted the possibility that the town defences had followed a line to the south-east side of Little Wind Street and close against the Pill. More probably the wall took a course just within the line of that street as it closed with the South Gate. Traces of the South Gate are reported to have been revealed early in the twentieth century in Mount Street, just below the junction of Little Wind Street with Wind Street.198 The site of Mount Street now straddles Quay Parade, the major dual carriageway entering Swansea from the east. The name is derived from the Mount, a post-medieval structure, which in its first phase lay directly beside the South Gate. Square in form, it must have occupied the easternmost angle of the walled town, its north-east flank set against the southernmost end of the east wall. Traces of the town wall were thought to have

195 B. Morris, Gower, 26 (1975), 11–15. Wales, 16 (1976), 41; Glamorgan–Gwent Archaeo- 196 S. H. Sell and J. Parkhouse, Arch. in Wales, 20 logical Trust, Annual Report (1976–7), 17. (1980), 64. 198 S. H. Sell, Gower, 32 (1981), 77; Morris in 197 S. H. Sell, Gower, 32 (1981), 71–83; Arch. in Griffiths, City of Swansea, map, 146. FIGURE 13. Swansea: the Buck brothers’ view of Swansea, 1748. 210 C. J. SPURGEON been identified in 1926 in Mount Street, where they were considered as part of the short line joining the South Gate with the angle of the town.199 The Mount, interpreted as a ‘moated mound’ at the lower end of Wind Street, has been identified as the original Swansea Castle raised by Henry de Beaumont c.1107.200 Morris has shown that the Mount was a defensive structure added to the town defences beside the South Gate in the Tudor period and reconstituted in 1760 in an advanced position straddling the Pill.201 Early drawings show that it constituted a rectangular crenellated structure rising an estimated 4 metres above the surrounding level, its walls revetting a gravel mound. Its function was to provide additional protection for the town and the bridge over the Pill and to act as an emplacement for artillery protecting the harbour and its approaches. It was well-established by 1613, when an alderman was appointed as ‘overseer of the Kaye and Mounte’, and may have paralleled such Elizabethan artillery positions as the Saluting Platform which protected the entrance to Portsmouth Harbour.202 The original Mount beside the South Gate appeared on the print of Swansea by the Buck brothers in 1748, which clearly portrayed its crenellated revetment walls and a tall mast rising within it to a ‘crow’s nest’, from which Swansea Bay could be surveyed (Figure 13). At that date the Mount rose from the edge of the Pill, which was spanned by a bridge leading to the gate. By 1762, however, a plan portrayed the Mount clearly in advance of its former position, straddling the Pill and along- side the bridge, and with a sluice to control the tidal flow in the creek. Later plans and drawings confirmed this arrangement, which seemed to cast doubt on the veracity of the Buck print. Scrutiny of corporation records by Morris proved that the Bucks were not mistaken. The records showed that in 1760 the original Mount was taken down and replaced further forward over the Pill, the old structure being perceived as an obstruc- tion to traffic entering the town. The corporation records also indicate that the Mount, the Wassail Gate and other parts of the town defences were maintained well into the eighteenth century. The north wall of the new Mount was rebuilt in 1782, but in 1804 the structure was removed to make way for the Oystermouth Railway, and by 1809 the Pill had been culverted.

Acknowledgements

The work for this paper was done originally for the Royal Commission’s Inventory, Glamorgan: Later Castles (Aberystwyth, 2000). Howard Thomas and Dylan Roberts assisted the author in the ﬁeld. David Browne and Professor Ralph Grifﬁths have edited the text for publication. Illustrations and layout are by Charles Green and John Johnston. The text was word-processed by Christine Norrington-Davies.

199 W. H. Jones in RISW, Report (1925–6), 33. 201 B. Morris, ‘The Mount at Swansea: moving a 200 For example, Armitage, Norman Castles, landmark in 1760’, Gower, 42 (1991), 24–33. 297–8. 202 R. Fox, Fortress, 11 (1991), 30–2. THE MEDIEVAL TOWN DEFENCES OF GLAMORGAN 211 Abbreviations

Ann. Margam Annales de Margam (printed in Annales Monastici, ed. H. R. Luard, Rolls Series, i, 1864). Ann. Theokes. Annales de Theokesberia (printed in Annales Monastici, ed. H. R. Luard, Rolls Series, i, 1864). Arch. Camb. Archaeologia Cambrensis, The Cambrian Archaeological Association, Cardiff. Arch. in Wales Archaeology in Wales, Newsletter of the Council for British Archaeology: Wales. Armitage, Norman Castles E. S. Armitage, The Early Norman Castles of the British Isles (London, 1912). BBCS Bulletin of the Board of Celtic Studies, University of Wales, Cardiff. Beresford, New Towns M. Beresford, New Towns of the Middle Ages (London, 1967). Boon, Welsh Hoards G. C. Boon, Welsh Hoards 1979–1981 (Cardiff, 1986). Breviate Ann. Annals from AD 600–1298 inserted in Breviate of Domesday (PRO, E. 164/1) and printed in Arch. Camb. (1862), 272–83. Brut, BS Brenhinedd y Saesson or The Kings of the Saxons, ed. and tr. T. Jones (Cardiff, 1971). Brut, Pen 20 Brut y Tywysogyon or The Chronicle of the Princes (Peniarth MS 20 version), ed. and tr. T. Jones (Cardiff, 1952). Brut, RBH Brut y Tywysogyon or the Chronicle of the Princes (Red Book of Hergest version), ed. and tr. T. Jones (Cardiff, 1955). Cal. Anc. Petitions Calendar of Ancient Petitions Relating to Wales (thirteenth–sixteenth century), ed. W. Rees (Cardiff, 1975). Cal. CR Calendar of the Close Rolls, Edward I–Henry VII (London, 1892–1975) and, though not calendared, Close Rolls of the Reign of Henry III, 15 vols. (London, 1902–38, 1975). Cal. FR Calendar of Fine Rolls, 1272–1471 (London, 1911–49). Cal. Chart. R Calendar of Charter Rolls, 1266–1516 (London, 1903–27). Cal. IPM Calendar of Inquisitions Post Mortem and other Analogous Documents, Henry III–Edward III (London, 1904–74). Cal. PR Calendar of Patent Rolls, Henry III–Elizabeth I (London, 1902–82). ‘Cardiff Chronicle’ British Library, MS Royal 6 B xi. Cardiff Records J. H. Matthews (ed.), Cardiff Records: Materials for a History of the County Borough, 6 vols. (Cardiff, 1898–1911). Carlisle, Top. Dict. N. Carlisle, A Topographical Dictionary of Wales (London, 1811). Clark, Cartae G. T. Clark, Cartae et alia munimenta quae ad dominium de Glamorgancia pertinent, 2nd edn, 6 vols. (Cardiff, 1910). Clark, Land of Morgan G. T. Clark, The Land of Morgan: Being a Contribution towards the History of the Lordship of Glamorgan (London, 1883). Clark, Limbus G. T. Clark, Limbus Patrum Morganiae et Glamorganiae, being the Genealogies of the Older Families of the Lordships of Morgan and Glamorgan (London, 1886). Corbett, Glamorgan Lordship J. S. Corbett, Glamorgan: Papers and Notes on the Lordship and its Members, ed. D. R. Paterson (Cardiff, 1925). Crouch, Llandaff Epis. Acta D. Crouch (ed.), Llandaff Episcopal Acta 1140–1287 (Cardiff, 1988). 212 C. J. SPURGEON

DWB The Dictionary of Welsh Biography down to 1940, ed. Sir John Lloyd et al. (London, 1959). Epis. Acts Episcopal Acts and Cognate Documents relating to Welsh Dioceses, 1066–1272, ed. J. Conway Davies, 2 vols. (Cardiff, 1946, 1948). GGAT The Glamorgan–Gwent Archaeological Trust. Glam. Co. Hist Glamorgan County History, iii, The Middle Ages, ed. T. B. Pugh (Cardiff, 1971). Series now completed with vols. i, ii, iv, v and vi (Cardiff, 1936, 1984, 1974, 1980, 1988). Gray, Kenfig T. Gray, The Buried City of Kenfig (London, 1909). Hist. King’s Works The History of the King’s Works, ed. R. Allan Brown, H. M. Colvin and A. J. Taylor, i and ii, The Middle Ages (London, 1963). Jones, Swansea and Gower W. H. Jones, The History of Swansea and the Lordship of Gower, 2 vols. (Swansea, 1920, 1992). Leland, Itin. Wales The Itinerary of John Leland in or about the years 1536–1539, iii, Wales, ed. L. Toulmin Smith (London, 1906, repr. 1964). Lewis, Top. Dict. S. Lewis, A Topographical Dictionary of Wales, 2 vols. (London, 1833). Lloyd, Hist. Wales J. E. Lloyd, A History of Wales from the Earliest Times to the Edwardian Conquest, 2 vols. (London, 1911; repr. 1912, 1939, 1948). Merrick, Morg. Arch. Rice Merrick, Morganiae Archaiographia: A Book of the Antiquities of Glamorganshire, ed. B. L. James (Barry Island, 1983). Morgan, East Gower W. Ll. Morgan, Antiquarian Survey of East Gower (London, 1899). Morgannwg: Morgannwg, Transactions of the Glamorgan Local History Society (from 1971, The Journal of Glamorgan History). NLW National Library of Wales, Aberystwyth. Patterson, Glouc. Charters Earldom of Gloucester Charters: The Charters and Scribes of the Earls and Countesses of Gloucester to A.D. 1217, ed. R. B. Patterson (Oxford, 1973). Phillips, Vale of Neath D. R. Phillips, History of the Vale of Neath (Swansea, 1925). PRO Public Record Office, London RCAHMW Royal Commission on the Ancient and Historical Monuments of Wales. Rees, Caerphilly Castle W. Rees, Caerphilly and its Place in the Annals of Glamorgan (Caerphilly, 1974). Rees, Cardiff W. Rees, Cardiff: A History of the City, 2nd edn (Cardiff, 1969). RISW The Royal Institution of South Wales, Swansea.