Held at:

Headland Archaeology Ltd [Archaeological Investigations Ltd]




Original publication


Ewyas Harold Castle: Archaeological Surveys

Place name:

Ewyas Harold




Guest Contribution: Introduction

This archaeological survey is reproduced with the kind permission of Headland Archaeology Ltd [for Archaeological Investigations Ltd], to whom we are most grateful.

Ewyas Lacy Study Group




Unit 1, Premier Business Park,

Westfields Trading Estate,

Faraday Road , Hereford


Tel: (01432) 364901 - Fax: (01432) 364900

Email: andyb@ailtd.freeserve.co.uk


Hereford City and County Archaeological Trust Ltd


Ewyas Harold Castle








Brief background to the site


Aims and objectives


Statement of methods employed


5.1 Documentary research


5.2 Earthwork survey


5.3 Geophysical survey






7.1 Some documentary considerations


7.2 Survey of the extant remains


7.3 Geophysical surveys




8.1 The early castle


8.2 The priory


8.3 The later castle


8.4 Mills and fishponds


Recommendations and potential








Appendix 1: Sources Consulted


Appendix 2: Report on stone carving at Dulas


Figures and plates





Archaeological Investigations Ltd

Limited company: 3356931:

VAT 692 1750 23






Hereford City and County Archaeological Trust Ltd

Registered Charity No. 1060840 Limited Company No. 3283805

Ewyas Harold Castle




1.         Summary


Archaeological Investigations Ltd was commissioned by the Ewyas Harold History and Archaeology Society to undertake a survey of Ewyas Harold Castle and its immediate environs, following their successful bid for a Local Heritage Initiative Grant.


The site is located at SO 384288 and straddles the 80m contour. It is a Scheduled Ancient Monument (30079). The work involved a documentary study, earthwork survey, and resistivity, magnetic and radar surveys. The core of the site can be divided into three components from north to south: the motte, inner bailey and outer bailey (the latter being divided by a ditch called King Street).


As a result of the surveys it was established that stone structures survive both on top of the motte and in the inner bailey. There appears to be a large building adjacent to the motte which may be a main residence. In the outer bailey, traditionally the site of an early medieval borough, there was some indication that two “streets” may once have occupied this area. One of these could form an approach road on the south side of the inner bailey.


Comparison of two late 19th -century maps of the site suggest that a quarry just outside the northeast corner of the inner bailey began to operate between 1876 and 1888 and the current access to the castle from the Dulas Road was built for this. The survey work suggests three principal entrances for the castle. The two early ones comprise the gap in a bank on the north side of the inner bailey (now the access to a bungalow there), and a possible entrance in line with one of the streets mentioned above. It is probable that the entrance at the south-east corner was created at a later date. Part of one of the major circuit walls for the inner bailey was observed in section in the southwest side of the motte.


The southeast corner of the outer bailey reputedly contained the remains of a Benedictine priory. Two possible sites of buildings were also identified here. However, consideration of the grants held in the cartulary of the priory might suggest that the priory lay in the field on the other side of the Dulas Road and also adjacent to the church on the left (east) bank of the Dulas Brook. However, the information in the cartulary is difficult to apply to the modern-day landscape and a degree of conjecture has had to be used.


The work also considered the possibility of occupation on the ridge behind the castle but no evidence was forthcoming in this respect.


The project has identified a number of avenues for further work including a study of the fields around the Dulas Brook and near the church, as well as the potential to enhance the site through providing better access to the top of the motte, and more comprehensive interpretation on the site.



2.         Introduction


Archaeological Investigations Ltd was commissioned by the Ewyas Harold History and Archaeology Society to undertake a survey of Ewyas Harold Castle and its immediate environs, following their successful bid for a Local Heritage Initiative Grant.


The site is located at SO 384288 and straddles the 80m contour (Fig. 1a). It was Scheduled in 1928 with amendments to its scheduling in 1974 and 1998 (when it was expanded to include the outer bailey) and it was the existing scheduled area that formed the core study area in this instance (Fig. 1b).


All the methods to be employed were non-destructive in nature and included a topographic survey, geophysical surveys using resistivity, gradiometry, magnetic susceptibility and radar, as well as a documentary study of the site.


The site lies on raglan mudstone of the Old Red Sandstone with alluvium covering its lower reaches (Fig. 2). It is bounded to the east by the Dulas Road (with the Dulas Brook a little further east than this) and to the west and south by Prill Brook and Prill Lane. Its northern extent is that of the castle defences.


The project fieldwork was undertaken between April and September 2007.



3.         Brief background to the site


A rapid consultation of the SMR reveals no known sites of pre-10th century origin within 1km of the castle. However, the neighbouring settlements contain evidence for early religious occupation at Dulas and Prehistoric activity at Walterstone. The castle itself continues a theme of rare early settlement with the castle being one of the few pre-conquest Norman structures of its type in the country. It is thought that preceding the establishment of the first timber castle on the site by Osbern Pentecost in 1048 there was a need to raise defences against Danish pirates in Dulas Valley. William Fitz Osbern remodelled Pentecost’s castle after the Norman conquest. In 1120 the Dulas monastery moved into the lower bailey. By 1530 the castle was in ruins and in 1645 the castle was recorded as having gone.



4.         Aims and objectives


The aims of the project are:

·                “to understand more about the significance of the site in the county and its regional setting (Boucher 2006)


·                to involve local people in local history and archaeology giving them ownership of the project and its results


The identified objectives are:

·                To carry out a documentary study from 1048 – 1645

·                To undertake an earthwork survey of the site and its surrounds

·                To undertake geophysical survey on the site and in its surroundings

·                To produce a report, exhibition and leaflet based on the results of the above

·                To hold local meetings to present the results and assist in the provision of a web site

·                To provide training of and support for local people so that they can organise their own aspects of the project.


Further aims relating to the local community

A key member of the project team was the local community. The original proposal aimed to pass ownership of the project, its results, its methods as well as its overall structure to the local community. In this respect the methods and approaches described below have varied to accommodate the enthusiasm, support and skills that could be provided by the local community.



5.         Statement of methods employed


Each element is discussed in turn.


5.1       Documentary research



The following repositories were consulted

·                                                                                        Sites and monuments record

·                                                                                        Local research library/British Library (for the map book of 1718)

·                                                                                        Local/national records offices

·                                                                                        National monuments record – air photos held in Swindon.

The following sources were searched

·                                                                                        Cartographic evidence

·                                                                                        Aerial photographs including those held by NMR

·                                                                                        Geology and soils maps

·                                                                                        Secondary sources

·                                                                                        Reports on previous excavations.


5.2       Earthwork survey


The survey involved the collection of data relating to the terrain as well as laying out markers for other survey methods to tie in to.

The core survey work was undertaken using a total station. Enough data was collected to generate a contour plot for the site.

A digital control plan was produced for the site and later filled in with detailed observations using semi-permanent markers left in the field. Notes relating to control measurements, point codes and point ID’s that relate to the data stored in the total station’s memory were made alongside clarity of break of slope and steepness/regularity of slope and vegetation cover. The survey instrument was also used to map the top and bottom of break of slope for all visible features and edges of structures. Spot heights were also recorded in areas without features.

5.3       Geophysical Survey


Four methods were employed.


Magnetic susceptibility

Magnetic susceptibility survey was undertaken through both the collection of soil samples and loop measurement on a 4m or 5m grid. Samples were weighed and measured from within the bailey site. The aim was to help interpret features identified using other survey methods.


Gradiometer survey

A fluxgate gradiometer survey was undertaken within the survey area using Bartington 1m fluxgate tubes and the Bartlett Clarke Consultancy logging system (for which we are indebted to them). Survey data collection was on 1m spaced traverses with readings taken every 0.5m or less. The method is most appropriate for locating cut features and burnt features

Resistivity survey

An RM15 (Geoscan) resistivity meter was used with a mobile probe array. To ensure that different areas of survey were comparable the remote electrode spacing was kept wide. This also allowed for an assessment of the physical properties of buried remains which is not possible using variable remote spacing.

The method is most appropriate for identifying stone/brick built foundations/walls although cut features also show. Readings were taken at 0.5m intervals on traverses spaced 1m apart.

Ground Penetrating Radar

This was used across the top of the motte and the entire bailey. Single traverses were also taken around the moat and down the side of the motte and across the moat.

The method is good at locating buried voids, features cut through solid surfaces or bed rock and also for looking at cross sections through buried structures.



6. Liaison


·          Scheduled Monument Section 42 licence

     Both a licence and Scheduled Monument Consent were applied for on behalf of the project.


·          Liaison

     Liaison was undertaken with the following:

     Ewyas Harold History and Archaeology Society

The land owner(s) – particularly with regard to modus operandi

Herefordshire Archaeology

Inspector for ancient monuments, English Heritage.


7.         Results


The text below lays out results from each individual method employed with some cross referencing between methods.


7.1       Some documentary considerations


7.1.1 The name

Bannister (1902) describes at length his concern over there being a complete lack of substantial evidence on which to base a true derivation of the name Ewyas. Sprackling (1988) makes the following proposition based on the entry in the Oxford Dictionary of English Place Names. Here Ewyas is taken to derive from the “Welsh – ewig = doe, from the latin – ovis meaning sheep. The -as ending meaning Sheep District”. Others see it meaning place of Yew trees Yw-ys, in the Yew district Ystrad Yu (Bull 1869) although there was by no means a unanimous acceptance of this at the meeting where his paper was presented. Sprackling also points out that some scholars believe it to be pre-Celtic and cites the most likely explanation as being that of “McClure … who derives it from an ancient British ‘Gewissi’ = land of the settlers” referring to the Saxon invaders rather than the native Welsh. The Harold part is discussed below; however, to summarise, it would appear to derive from a reference to Harold’s son Robert who was responsible for relocating the priory, founding Dore Abbey and possibly resiting the village. The name in question was apparently used by Leland in c. 1530 and was that of Mapelart, Maperalt or Mapheralt. The “Map” part being the Welsh equivalent of “ap” or “son of” the other being Harold.


7.1.2 The early history

With respect to the castle site and occupation at Ewyas Harold we join Bannister (1902) in the 5th -6th century AD. The scene is set with the River Severn dividing England (or more precisely the kingdom of Mercia) and Wales. Clydawg is recorded as King in Ewyas – he being a holy man (presumably Christian).


Bannister (p6) points to Ewyas being “beyond the dyke” (some three miles to its west) and finally being occupied by the English in the 10th century. Archenfield (an independent kingdom prior to Offa) was annexed to Mercia in the late 8th century.


In 915 we hear of the ravages of the “Black Pagans” or Danish Pirates harrying Ewyas and Archenfield and making off with the Bishop of Llandaff. He suggests that it was at this time “the great mound, on which afterwards was built the castle of Ewyas Harold, was raised” (Bannister 1902, 6).


After Edward the Elder annexed Mercia, the old smaller kingdoms within the newly united England were ruled by Earls. In 1042, on the accession of Edward the Confessor, Herefordshire and its loosely annexed dependencies, Archenfield and Ewyas, were severed from Lady Godiva’s husband’s (Leofric) earldom of Mercia – at the instigation of Godwin (Father of the later King Harold). The balance of power lay with Godwin who out of the four earldoms held Wessex while at the same time his son Harold ruled East Anglia. The newly forged earldom based on Hereford he gave to his other son Sweyn. Sweyn was banished in 1046 after killing a kinsman and seducing and attempting to marry the Abbess of Leominster. Hereford was then given to Earl Ralph (father of the Harold who in the late 11th -early 12th century became the Earl of Ewyas).


Earl Ralph was the nephew of Edward the Confessor, the latter half Norman, the former a Norman kinsman and son of Edward’s sister Goda. However, Ralph was only earl for the four years of Sweyn’s exile but these four years are critical in the history of Ewyas Harold. Ralph gave Richard (son of Scrob) land in the north of the county (Richards Castle), while Osbern  Pentecost was given Ewyas. Bannister (1902, 8) states that Osbern already had the beginnings of the castle in the mound and earthworks raised a century earlier and on these built a keep and surrounded the lower ward with a wall.


The presence of private fortifications such as these was the cause of great unrest and oppression. This was illustrated when Godwin and Sweyn’s attempt to have the Norman overlords removed was thwarted by King Edward and the Earls of Mercia and Northumbria resulting in their exile. Whilst Ralph was engaged in preventing the exiles returning at one end of the country, the Welsh invaded Herefordshire and were victorious at the battle of Leominster where they defeated both English and French (the latter presumably from Ewyas and Richards Castle).


The castle at Ewyas Harold was dismantled in 1052 at the end of a sequence of events that started with Godwin’s struggle for supremacy. Godwin and his son Harold (with a fleet from Ireland) sailed unopposed on London (as by this time Ralph’s fleet had disbanded, and Ralph was in disgrace). The French contingent (Edward, Ralph, Osbern and Richard) rode to head him off but fled abroad or back to their lands when faced with the judgment of the Witan that was likely to result in at least some of their executions. Only Ralph was allowed to keep his position in Herefordshire – Osbern Pentecost and his company were outlawed, although Osbern himself escaped on the mercy of Leofric to serve under Macbeth in Scotland. Following the dismantling of the castle Osbern’s (Pentecost) nephew, Alured of Marborough, was given Ewyas and still held it 20 years after the conquest at the time of the Domesday Survey.


Ewyas was next involved in conflict in 1055 when the exiles Alfgar earl of East Angles combined his Irish force with Gruffydd (Prince of North Wales) and marched through Ewyas, into Archenfield (occupied by Welshmen of English alliance) and on into Herefordshire. Following Ralph’s defeat King Edward sent Harold (Godwin’s Son now Earl of the West Saxons) to Hereford where he drove the Welsh back into South Wales. He also surrounded the city of Hereford with a wall and Hereford was added to the Earldom of the West Saxons.


William Fitz-Osbern is the next significant character to influence developments at Ewyas Harold and its Castle. He was most favoured by William the Conqueror and his establishment as Earl of Hereford (1067-71) illustrates the significance of the county at the time. His role was not only to subdue Wales but also to keep an eye on Mercia (not yet under the rule of King William who was only king in the south at the time). Fitz-Osbern subdued the borders through castle building which included those at Ewyas and Richards Castle. When Edric (“the Savage”) from north of Herefordshire and the Welsh attacked in 1067 (laying waste to the districts of Ewyas and Archenfield) it is perhaps notable that the three sites that held out against the attack were Hereford City (recently fortified by Harold), and the castles at Ewyas Harold and Richards Castle - the sites of the three pre-Conquest castles.


In ruling the land from the mouth of the Severn up to Richards Castle Fitz-Osbern attracted Knights and Burgessess through a system of lavish pay and privilege. For example burgessess of French birth in Hereford could not be fined more than 12 pence for any offence (except for three reserved pleas – i.e. where a defendant refuses to state whether they are either guilty or innocent of given charges three times in succession). He established the great status enjoyed by the Lord Marchers.


Bannister summarises the development of the castle at Ewyas Harold as follows:

·                915 – the tongue of high ground rising abruptly in the angle where two streams meet, had already been cut off by a ditch, and by the addition of a mound, and an outer circle of earthworks turned into a burgh.

·                1046-1051 Osbern  Pentecost built a Norman castle there.

·                1052-1053 Osbern  Pentecost’s castle was demolished.

·                1067 William Fitz-Osbern and Alured of Marlborough built a castle on the lines of the earlier castle. Salter (1989) and Coplestone-Crow (1992) are of the view that William Fitz-Osbern completed this work on his own and then passed it to Alured (or Alfred). Either way the works were completed in time to withstand the Welsh attack of this year.


In describing the later castle Bannister points to several aspects:

·                He identifies the more weakly defended side as being up ridge and that this was defended by a couple of towers.

·                The keep was c. 30 feet in diameter and built of Old Red Sandstone and probably many sided.

·                Access to the keep appeared to be by a flight of steps from the lower ward.

·                The crescent-shaped lower ward was strengthened by an earth bank and enclosed by a curtain wall.

·                The lower ward also contained a number of buildings and a chapel dedicated to St Nicholas.

·                Considerable remains of the castle were still standing in Leland’s day (c. 1530 ).

·                By the civil war a century later local farmers had grubbed up everything, including the foundations, to build barns and lime kilns (according to a Richard Symonds). One of these kilns “hard by the mound” is all that can be seen of the masonry of the castle.


The entry from the Domesday Survey as translated in Bannister (1902, 16) is as follows (original text italic; Bannister’s notes in round brackets in italics; author’s notes in square brackets, no italics):


Alured of Merleberge [Marlborough] holds the castle of Ewias from King William. For the king himself granted to him the lands which Earl William (Fitz-Osbern), who rebuilt the Castle, had given him, that is five carucates [one carucate = about 120 acres] of land in Ewias, and another five carucates at Munstone. The lands pertaining to the Castle, which had belonged to Ralph of Bernai, the King granted to him (i.e. to Alured). There he has in demesne two carucates, and nine Welshmen with six carucates, paying seven sextaries of honey (each year), and twelve bordarers working (for their lord) one day a week. There are four ox-herds, and one man paying six pence (each year). Five Knights, Richard, Gilbert, William and William and Hernold have five carucates in the demesne and twelve bordarers, and three fisheries, and twenty two acres of pasture. Two other William and Ralph, hold two carucates of land. Turstan holds lands bringing in nineteen pence (each year) and Warner lands bringing in five shillings. These have five bordarers. This castelry is worth ten pounds.”


We also know that the church held land in Ewyas Harold at the time of the Conquest. In referring to Bishop Walter’s holdings of 10 hides relating to Didley and Stane, 9 of these were waste “one part is in Alfred’s castlery of Ewyas (Harold), the other part in the King’s Enclosure ” (Morris 1983, 181). Morris adds as a note to this that the Royal enclosure/forest might refer to an extension of Treville Wood.


The following entry relates to Roger de Lacy’s lands in Ewyas,


In the castlery of EWYAS [here Morris states the entry as relating to Ewyas Harold – the Domesday Book is not so specific but this view is also supported by Benn (1941)] Earl William gave 4 carucates of waste land to Walter de lacy. Roger de lacy his son holds them and William and Osbern from him. They have 2 ploughs in lordship: 4 Welshmen who pay 2 sesters of honey; they have 1 plough. They have 3 slaves; 2 smallholders Value of this land 20s.” (Morris 1983, 184);


with further entry for the land of Henry de Ferrers as follows,


“In Cutsthorn Hundred.

In the castlery of Ewyas (Harold) Roger holds from Henry 3 churches, a priest and 32 acres of land; they pay 2 sesters of honey. In the castle he has 2 dwellings.” (Morris 1983, 185).


Bannister (1902,18) refers to the Welshmen mentioned and states that these would have enjoyed the same liberties as their kinsmen in Archenfield, i.e. they would lead the English into battle against the Welsh and retreat at the rear. They paid their dues in honey. From the Domesday record this “Welsh tax” (honey), along with the predominant use of hides, clearly distinguishes the areas of Archenfield and Ewyas from the rest of the county (Jackson 1954). Ewyas is also one of the largest settlements in the county at the time of the survey (although it did not have a market) and this might well reflect the fact that it formed the seat of the baronies of Alfred de Marlborough and later Harold.


Coplestone-Crow (1992, 8) provides a clear statement about the manner in which knights fees worked. It was part of a Baron’s responsibility to provide a number of knights for an agreed period of service. This period was 60 days in the host (i.e. as part of a mobile army – at home or abroad) and 40 day’s castle guard (presumably wherever this was needed). Any days served over and above this were remunerated at a rate of 6d per day (in 1135). Alfred had to provide 20 knights’ fees. The Barons could chose how they retained the services of requisite knights. They could keep them in their own household (stipendiary knights), hire them in when needed or give them lands in exchange for services (called enfeofing). To satisfy his obligation Alfred kept nine knights in Herefordshire and another 16 elsewhere giving him 25 knights, holding 56% of his barony to satisfy the 20 knights fee. The knights would also have to perform service to the lord (i.e. Alfred). However, as the castle at Ewyas was a strategically significant castle, service here was treated as both for the Lord and the King.


The Harold from whom Ewyas takes its name is believed to be the son of Earl Ralph, nephew of Edward the Confessor (Bannister 1902), and not as Leland suggests that “the fame is that the castle of Mapherald was builded of Harold afore he was king, and when he overcame the Welshmen Harold gave this castle to his bastard [also Harold]”. Robinson (1886, 58) states there is no evidence to support this assertion. Harold (earl Ralph’s son that is) in fact might have had a claim to the throne on the death of Edward in 1066. However, it is thought his mother (Gytha or Gueth) was dead and thus he was in the wardship of Queen Edith (Edward’s widow and daughter of Earl Godwin). It is thought that Harold was prevented from taking up his claim to the throne through the influence of Godwin’s son (also) Harold, his godfather. Following the conquest it appears that Harold (of Ewyas) held land in the shires of Gloucester, Worcester, Warwick and Middlesex. How he came to the land at Ewyas is not known nor why he took his title from here (after all Lord Sudeley was apparently a more significant title). What we do know is that Harold reached his majority and came into his other estates around 1070 living to be at least seventy, Harold of Ewyas witnessed a gift to St Johns Abbey in the Colchester Cartulary in 1120 (Bannister 1902, 21).


Despite losing a number of the estates entreated to Alfred (Bernard de Neufmarché and William de Braose took holdings which had provided seven knights to the former estate), Harold was still expected to provide the full 20 knights fee. To tackle this Harold created two honorial-baronies. Here he would give two parts of his Estate-in-Chief, one to Reginald the other to Erkenbald. Each had to provide four knights. The other 12 were provided by knights doing service for their lands or by substitute. Copelston-Crow (1992, 9-10) draws attention to two charters relating to Ewyas Harold and its knights. These date from about 1110 and 1148 and from these it becomes apparent that the knights serving Ewyas Harold had to provide 90 days free service at the castle there. It appears that to encourage this additional service a number of further favours were provided by Harold. These included being provided with houses in the borough to stay in with their families whilst doing service, free firewood and free hunting in the demesnes. As this was the only service required from these knights then it was likely that there would be at least five knights serving the castle at any one time. The charters also demonstrate the use of scutage (i.e. buying out of service) as early as the first part of the 12th century. In this instance it might be possible to buy men-at-arms or foot soldiers for 13s and 4d (1 mark) in 1110 or 6s and 8d (½ mark) in the 1145 charter. So by the beginning of the 12th century knights were being replaced in the manning of castles by non-feudal armed men.


There is some debate over who was the elder of Harold’s sons, Robert and John (his three younger sons being Roger, Alexander and William). Robert inherited the Lordship of Ewyas and John those of Sudeley and Toddington. Bannister (1902, 22) suggests Robert to be the eldest. Little is known of Robert, apart from the fact that he led a force against the Welsh in South Wales and refortified a castle there. He also issued five grants to the monks of the priory at Ewyas, as well as founding the Cistercian Abbey at Dore and being benefactor to the alien Priory of Craswall. His mother Sibilla was made a monk of Dore (by all accounts on her death bed), the monks there seeking the additional patronage she could provide. Despite any grants given she was still buried in Ewyas church yard, unlike her son whose tomb is in Dore Church. Robinson (1868) attributes the construction of St Michael’s Church to Robert (after a paper by Rev. W.C. Fowle in Cambr: Arch: Soc. Journal 1868). This would suggest that it was a church other than that of St Michael of Ewyas that stood at the foundation of the priory in 1100.


Robert Fitz-Robert was the next lord of Ewyas and by his wife Petronilla had an heiress Sibilla. Sibilla married Robert de Tregoz in 1194. Robert de Tregoz followed Richard I to Normandy in 1196/7 and did not return. Sibilla had two further marriages, first to William de Newmarch (d. 1211) and then to Roger de Clifford – becoming ancestress to the Earls of Cumberland. Her son (also Robert de Tregoz) inherited the estate of Ewyas on her death c. 1235. William, her youngest by Robert (senior), became rector of Kentchurch.


Robert married the sister of Thomas Cantilupe (Bishop of Hereford) – Juliana. Robert de Tregoz II followed Simon de Montfort and died in the battle of Evesham in 1265. On inheriting the estate his son, having obtained pardon from the king for his father’s treason, began to expand the holdings of the estate. This included Holme Lacy which was granted on the basis of one Knight’s fee through his uncle, Bishop Cantilupe. He married Mabilia (daughter of Fulk Fitz-Warine) in 1257 or 1258. In the 1277-8 campaign against the Welsh John Tregoz (Robert’s son) provided only three lances but several horses from Ewyas. In 1282 John brought 10 or 12 lances to the King’s muster at Rhuddlan against Llywelyn. The king was likely to have been happy with the provision of horse in the earlier campaign as his force was otherwise weak in this respect.


John Tregoz remained loyal to the King throughout his struggle with the Marcher lords despite giving evidence in support of the earls of Gloucester and Hereford.


To clarify the distinction between lesser and greater barons, Edward I formalised the latter through writs onto the Council (Parliament). In 1295 John Tregoz does not appear amongst the 53 names listed (out of a total of 250 barons) but by his actions had attained that status two years later and was summoned to parliament by writ in 1297 and 1299. The king entrusted the security of the country into the hands of John and a number of others after a dispute between the king and the Earls of Norfolk and Hereford. This arose because the earls refused to be bound by feudal tenure to foreign service. The king left for Flanders without them and civil war was only averted by an uprising in Scotland that had to be dealt with. John died on 21st August 1300 leaving two daughters. His estate reverted to the crown by escheat. On January 25th 1301 a writ was issued directing an “Inquisition Post Mortem ” and the division of the estate between John la Warr (son of John Tregoz’s eldest daughter Clarissa) and Sibilla (his younger daughter, wife of William de Grandison). This is a substantial document describing in great detail the possessions of the Barony of Ewyas. As a result the estate was divided on 31st December 1301 with John la Warr taking the Manor and Castle of Ewyas and becoming lord of the same. His first act was to release his tenants from the ancient service of repairing the mill dam of the castle mill. Duncomb (n.d.) provides an interesting insight into the siting of this mill, stating in 1307 “John la Ware Lord of Ewas Harold released the tenants of the Manor of Ewias Harold from the Custom paid to his ancestors, of reparing the stank of the Mill called Castle-mill, situated near the Gate of the Priory of Ewias ”. Neither John la Warr or William de Grandison managed to keep their tenants happy and constantly had to seize stock etc… in lieu of defaulters. One such issue shows that William of Leykesworth held a manor in the barony in exchange for providing two esquires for the ward of the Castle of Ewyas, which he had failed to do since John Tregoz’s death!


In 1347 John la Warr was succeeded on his death by his 18 year old grandson Roger (his eldest son John already having died). It was Roger in 1358 who made the agreement with the Abbot of Gloucester and Bishop of St David’s to suppress the Priory. According to Strong (1848) John la Warr inherited Ewyas Harold in 1297 (three years earlier than other accounts would have us believe). He adds that during his grandson’s tenure the castle was not used as his residence, Roger preferring Oxenhall near Newent. Robinson (1868, 60) raises the point that the manor and castle of Ewyas were held by Roger in chief (for the king), by the service of keeping the King’s harriers whenever the King came into the area. As such it would seem that this should have at least served as a temporary residence.


Bannister states (1902, 29-30) that neither John nor Roger seem to have spent much time at Ewyas as in 1331 John la Warr enfeoffed (i.e. gave permission to run the estate and castle in his absence) John de Claydon, with the Castle and Manor being transferred to Warine Latimer nine years later. Roger la Warr died in 1370 leaving the Ewyas estate to his son John.


From fragmentary remnants the latter part of the 14th century can be pieced together. It appears that Edward la Despenser held Ewyas Harold on his death in 1375 (in feof from John la Warr II?). In 1390 Sir John Montacute holds some of the Ewyas lands in Herefordshire, his widow holding them five years later. In 1399 John la Warr held the Castle and Manor again. However, it is as a result of what happened in this year that the Castle eventually passed into the care of its longest serving owner/occupier that of Abergavenny and more particularly the Neville family. According to Bannister (1902) it was in 1399 that William Beachamp was enfeofed at the King’s request, having previously inherited Bergavenny, the Castle and Manor of Ewyas being added to the Bergavenny estate in the first year of Henry IV. Other lands in Ewyas remained with the Montacutes (Earls of Salsbury at that time). Robinson (1868, 60) states that in this year Henry IV granted the lands and castle of Ewyas Harold to Sir Philip le Vache (Elizabeth Clifford his wife and his son-in-law Richard, Lord Grey of Wilton already had connections in Herefordshire). The estate passed to William Beachamp in 1403.


Owen Glyn Dwr’s first assault on Herefordshire was in 1402 with Rhys ap Gethin leading a force into the county following the battle of Pilleth where 1100 English were killed. Other efforts to suppress the Welsh failed and in 1403 William Beachamp, Lord Bergavenny, was empowered to fortify Ewyas Harold Castle by the King. Both the castle here and that in Abergavenny withheld four years of onslaught and siege, being relieved by the King’s army on occasion (Bannister 1902, 78). From the letter to the King written by Archdeacon Kingston on 10th June 1403, it would appear that Lord Bergavenny was in dire straights and likely that both castles at Ewyas Harold and Abergavenny were under siege, as Rys ap Gethin had managed to invade Archenfield again and the Archdeacon states that William Beachamp “is on the very point of destruction”. We know that Owen Glyn Dwr still maintained some hostilities even after 1410. In 1413 he turned his back on the pardon offered by the new king Henry V and continued raids on Marcher’s lands. However, by the time of his death in 1416 it appears Wales was relatively under control and Bannister claims this was the beginning of the end of the Castle at Ewyas Harold.


After Williams death in 1411 we know that his widow Johanna continued to hold the Castle and Demesne at Ewyas at least until 1436. As Richard his only son died in 1422 it was Richard’s daughter Elizabeth that, through marriage, brought the Barony of Bergavenny to Edward Neville (son of the Earl of Westmorland), and on her grandmother’s (Johanna’s) death the Castle, ville and lordship passed to him too (Bannister 1902, 30).


From the muniments of Kemes, a manuscript dated 1587, it is implied that the reason for the decay of castles in the marches up to that date stemmed from the fact that the English Lords had only ever used them as defensive houses and once Wales was subject to the Crown they no longer bothered to maintain them as there were better and more profitable places to live. At the time of Leland considerable remains of the castle and chapel of St Nicholas survived which had all gone by the civil war in 1645.


In this case the castle itself does not appear to have been occupied after the early part of the 15th century (Bannister 1902, 88), partly based on Leland’s description that the castle was already falling into ruin. However, in the civil war the Village was garrisoned by Royalists. It appears to have served as part of a supply route through Radnorshire to the north of the county. One other reference relating to the village in the civil war is a field called Blood Field (SO 3930 2775) which is reported as being the site of a battle on 13th November 1642 (HSMR 33548).


7.1.3 The Marcher Lordships

This is a good early example of political manoeuvring that went wrong. In the first instance the idea of maintaining the turbulent Welsh border by giving Norman Lords any land they could conquer probably served its purpose in both keeping the Welsh busy and also stopping the lords turning their attention to the King and trying to wrest overall control from him. The consequence, however, was a series of almost independent states along the Welsh border over which the King had few powers and which developed in themselves a number of powerful customs:


  • Any land annexed from the Welsh was the property of the Lord Marcher. Only if he then lost it and required the King’s assistance to regain it would the lands revert to the Crown. So the Lords had the right to govern their land with no reference to the King.

  • They had the right to war with each other rather than referring their disputes to the King.


Edward I looked to break the power of the Marcher Lords. His first move was to create a professional (paid) army so that he did not have to rely on feudal obligations. The Marcher Lords were not oblivious to this scheme and insisted on providing free men at arms (even in excess of their quotas) and the King eventually had to bribe the Earls of Gloucester and Hereford to provide a paid force. In the end he failed to remove their powers which were accepted in 1297 whilst the King was absent and confirmed by him in Ghent.


7.1.4 Ewyas: the historical topography of the Lordship

It is considered here how landownership, land form and local economics interweaved within the Lordship or Castlery of Ewyas. Other items such as the location of the Priory and establishment of the Church are tackled under the sections on the Priory and Church respectively.


Our earliest reference is 1086 but this gives an indication of what was held before that date. Alured’s (or Alfred’s) castlery included the Castle, inner bailey and borough below it. Based on details in the Domesday book this had a population of about 500 (Coplestone-Crow 1992). There were also nine knights and two priests. These latter, it is thought, served the churches at Llancillo and Kentchurch (Coplestone-Crow 1992, 7). Alfred’s income from the estates was around £300. It also included the parishes of Kenderchurch, Rowlestone, Llancillo, Walterstone, Dulas and Abbey Dore (Coplestone-Crow 1993, 18).


According to Coplestone-Crow (1992, 7) the current civil parish of Ewyas Harold was divided into two lost English settlements, Mulstonestone (where Osbern  Pentecost built his castle), and Manitone (possibly where the modern village of Ewyas Harold now lie)- or, in his 1986 article, suggested as lying between Ewyas Harold and Pontrilas. His further consideration of these (1993) is important in understanding developments in landownership within Ewyas Harold. The crux of his argument is that the pre-Conquest form of land tax was based on Hides. Where such taxation was already in place then the Domesday survey reflected this. However, any new or fairly recent settlements at that time would not have fallen under the old taxation system. So, in these cases, when the Normans assessed them for taxation they were assessed using carucates, (i.e. land in Wales beyond the limits of any relatively long-term English settlement). Both Mulstonestone and Manitone were assessed at five carucates each. Morningtin Straddle (an alternative suggestion for the site of Manitone ) was assessed as worth five hides. What is the evidence these were in Ewyas Harold? In Robert’s Carta of 1166 he refers to one knight’s fee held by Godfrey de Manistone de feodo meæ castellariæ de Ewias (i.e. within his castlery of Ewyas). In the inquisition of 1300 there is reference to Reginald de Mulstonton who held ½ a knight’s fee and the post mortem of the inquisition specifically states that this was within Ewyas Harold. Furthermore in 1352 the sheriff was demanded to deliver “multeston with the rest of the town of Harald Ewyas in Wales” into the hands of two keepers (ibid , 20 ). Manitune or Manitone was therefore an estate or manor, under the lordship of Ewyas Harold, which appears to have been chosen as the site for the new borough, but also contained another settlement called Mulstonestone . This latter may mean “settlement with a mill-stone”). From the above it can be seen that their precise location and status is still the subject of debate.


The lands in Ewyas were all acquired by Harold after 1086. From the Red book of the Exchequer in 1166 Robert held 19 Knights-fees in Herefordshire. These fees lay in and around the “ville” of Ewyas Harold, at Mornington Stradel in the Golden Valley (Copleston-Crow states Ashe Ingen), Eton Tregoz, and Pencombe, as well as other manors outside of Herefordshire. The income of the portion of Alfred’s estates passed to Harold was therefore £200+. Harold’s lordship at Ewyas Harold was smaller than that of Alfred because Roger de Lacy had obtained Walterstone, Llancillo and Rowlestone following Alfred’s death. Also Kenderchurch passed back to the Bishop of Hereford who had held it before 1066 (Coplestone-Crow 1992). Figure 3 illustrates the local parishes that once made up Ewias.


According to Coplestone-Crow (1992) the priory was moved by Robert (Harold’s son) and relocated in the former Borough, the latter having already been resited by him to where the present village is.


Those hearings relating to the title of Ewyas in Herefordshire following the death of John Tregoz in 1300 were held at Hereford (June 20th ), Ewyas Harold (Oct 18th ) and Eton Tregoz (Oct 20th ). It is, however, the detailed inventory relating to the lands in Ewyas Harold itself that interests us here.


Table 1: Lands held in Herefordshire




The “ville” of Ewyas Harold


Mulstonton or Mulstoneston


Puston in Straddel

Poston near Peterchurch

Pencumbe with Caldwell near Bromyard

15 hides here

Monnington Straddel

In the Golden valley



Table 2: Land in the Demesne or home farm




The castle

Recorded as being in Wales


Assessed to be worth 6s 8d

2 Gardens

  • One worth 2s the other 12d per year
  • These were the Castle garden and the garden of the Prior

2 water mills

Worth 26s 8d and 13s 4d


4 caracates “everyone of which contained 100 acres by the short hundred”

Therefore 400 acres @ 3d = 100s


20 acres @ 18d = 30s


6 acres @ 1d = 6d


14 acres @ 6d = 7s


Total value of Demesne = £9 7s 2d


There were nine tenants in chief holding between them six Knights-fees.


In the village of Ewyas Harold there were thirty seven freeholders (each holding a house and land varying in size from a few acres to a caracate), and 75 custumarii (customary tenants) rendering aids and other services, their rents totalling £4 19s 1½d with the value of their aids being 26s 1d (Bannister 1902, 37, 117-120).


Table 3 : Land holdings of free tenants in 1300




Richard de Enwyas

1 messuage, 1 carucate of land

Henry de Eylston

1 messuage, 1 carucate of land

Madok ap Jevered

1 messuage, 9½ acres

Susan ap Seysol

1 messuage, 1 carucate of land

William de la Grave

1 messuage, 20 acres

Jeverard ap Adam

1 messuage, 9 acres

John the clerk

1 messuage

Richard de Madeleye

1 messuage

Jevan ap Thomas

1 messuage

William ap Adam

1 burgage and a half

Philip Tok

8 acres

Henry de Jay

1 messuage, 12 acres

Matilda Lergey

1 burgage

Michael de Pattershull

A certain piece of meadow (he owes 1d for it)

Philip Mile

1 burgage

Jeverard ap Griffiths

1 messuage, 1 carucate of land

Ralph de Enwyas

8 acres

John de Ipre

1 messuage, 1 carucate of land

Hugh Drueyn

1 messuage, 12 acres

Matilda de la Helme

One piece of land owing 3d for it

Julia the Weaver

Half a burgage

John la Hunte

1 burgage

Walter Elyot

11 acres

Hugh le Tournir

Half a burgage

Hugh de Bredwardin

1 burgage, 8 acres

Adam son of Stephen

1 messuage, 2 crofts

Joan Brounesdauthor

1 burgage, 10 acres

Joan the Weaver

1 burgage

Julia Drue

1 burgage

Joan Gilbert

1 burgage

Agnes Nutegale

1 burgage

Joan Pyard

Half a burgage

Margery Michel

1 burgage

Alice Alewy

1 burgage

Joan la Taylor

A certain piece of land for 8d

Hugh Gilbert

1 messuage, 1 carucate of land

Richard Cantlin

Half a burgage

William Marsh

Half a burgage




67s 9¼d


7.1.5 13th Century Ewyas (From Bannister [1902, 72] unless otherwise stated) – Fig. 4

The castle stood on the “Castle Tump” with the chapel of St Nicholas and other buildings standing around the bailey (basse cour ), the curtain wall surrounding the platform of the lower ward. The main entrance was on the south (or King Street) side, a postern on the north side. Below the Postern was the “Lords Garden” through which ran a path from the Postern Gate to the village with a little stream beside it. Near the village the path divided, one branch skirting the wall of the Priory and crossing the Dulas Stream to the churchyard, the other passing between the curtilage of the monks and the curtilage which William Croc once held and then joined Frog Lane or Frog Street. Richard le Norman, Prior in the second half of the 13th century, wishing to enlarge the Priory obtained from John de Tregoz a grant of that branch of the path that led to the churchyard, closed it to the village, and built over it. On the other side of the Dulas Brook was the Lord’s mill, which had a monopoly of grinding for the tenants; it was supplied by an artificial channel (which could still be seen in 1902). Near the mill was the Priory barn. On the other side of the castle from the village ran King Street leading to Langua. The road to Dore ran round Cae-flwyn and past “Potters Field”. The high ground between Ewyas Harold and Abbey Dore was where the Lord’s wood stood. Below the village at the weir was the new mill of the Lord. A mile lower down stream near where Dulas Brook joins the Dore was Heliston mill (Pontrilas now).


To the north of Ewyas was the forest of Treville containing 2014 acres and extending from Kingston to the River Dore (in 1198 the Abbot of Dore tricked King Richard out of Treville forest and cut it down to sell the timber in Hereford for building). To the west was Mascoyt (Maes-y-coed) with a third forest near Grosmont containing a Hermitage called Lanneir.


Further commentary on some of these elements is provided by Sprackling (1988, 21, Fig. 5). With respect to King Street, he states that the existing road by the farm of the same name was a later insertion to join two old roads, the historic branch being picked up in the fields leading towards and along the south ditch of the inner bailey, and then continued to link this to the Lords Mill, church and village. He also states that it probably fell into disuse after the village had been moved, and the priory took over part of the road “to augment their court ”, and that Bannister’s equation of Virogis street with Frog Lane may be due to ignorance on his part regarding the existence of this other road (although Bannister does mention it in his book). However, that argument aside it is still quite possible that the road name derives from via Regis and could be referring to King Street rather than Frog Lane. So referring back to the discussion two paragraphs above, according to Sprackling, in the late-13th century the Priory closed the branch of the lane (most probably King Street) which ran across the fields to the Church leaving Frog Lane (the current road to the village). The former of these is described as “skirting the wall of the priory”, which also has a frontage onto Frog Lane.


7.1.6 The Priory

The priory at Ewyas was the first to be attached to the newly revived Benedictine Abbey at Gloucester (dedicated in summer 1100) and was soon joined by Hereford, Bromfield, Stanley, Kilpeck and Ogmore (Ewenny). The Bishop of Hereford was the first to grant lands to the Abbey followed by Harold of Ewyas. Harold’s endowment to the monks at Gloucester included all the tithes from his demesne and the lands and other possessions of St Michael’s Church (Ewyas). In exchange for this the monks served the church of St Michael’s and provided a chaplain for the chapel of St Nicholas in the castle (Bannister 1902). It appears from the documentary evidence that the Priory was not large to start with. The cartulary (para. 98) refers to Harold granting the monks who serve at the church of St Michael the tithe of his table and that of all his guests, the tithe of all animals killed on his estates and of all hunting, as well as the tithes of various other churches and estates. It also refers to St Michael’s of Ewyas. The commentary attributes a date of between 1115-1125 to this entry but notes that the original gift was dated 1100 (Walker 1976).


The Priory cell founded by Harold is believed to have been within his Lordship at Dulas, and by the church of St Michael there. However, after Harold’s death Payn fitz John seized Dulas and annexed it to the lordship of Ewyas Lacy and it has been long believed that Robert relocated the Priory to the area formerly occupied by the borough in the outer bailey at Ewyas Harold (Coplestone-Crow 1992, 8). This may have been in the 1130’s-40’s which is when Coplestone-Crow believes Ewyas Harold lost control of Ewyas Lacy (pers com ), and incidentally when there was a grant to the priory of wood to build houses with.


What is the evidence for this? Coplestone-Crow (1993) presents a number of useful arguments. It appears to have been long accepted that the priory was originally located at Dulas. However, it was from a confirmation of Harold’s original grant of land called Leghe to the priory by Sibil de Lacy (with her husband’s permission) sometime between 1130 and 1137 that the main clue arises. She couldn’t have been confirming lands within the present parish of Ewyas Harold as the de Lacy’s didn’t hold them (op sit ). So they must have come from one of the parishes originally held by Harold and later acquired by the de Lacys (see history of the topography). Her charter describes Leghe as lying near the church of St Michael and by elimination this was most probably St Michael’s, Dulas. She also gave the monks further rights in her forest of Maescoed. Evidence in the names of fields and woods around Dulas could well represent remnants of the tract called Leghe given to the priory and which, to all intents and purposes, would appear to have been the Dulas valley (Figs 6 and 7).


However, since his article in 1992 Coplestone-Crow has revised some of his thoughts on these matters (pers comi ). The most specific variation is that he now considers the site of the refounded priory to lie somewhere other than the site of the borough. His reason, that Walker’s (1976) transcription replaced horreum with hortus and the description should in fact read a “barn” enclosed by a ditch and not “garden” as per Coplestone-Crow’s previous article. However, Bannister (1902) in his translation referred to the new site of the priory correctly, and most of the descriptions in various historic documents still point to the site next to the “fish ponds” as being that of the Priory.


Why build the priory at Dulas? Coplestone-Crow (1993, 19) guesses that it may have been the site of an earlier Clas type cell and Harold knew of it. This is apparently not uncommon – St Guthlacs at Hereford for example. Archaeologically, there are monumental remains in the wall of a garden on the north side of the Dulas Brook. The RCHME (1931) recorded the following information about this:


Of this building [the old St Michael’s Church] nothing now remains except a reconstructed arch forming an entrance to a garden 150 yards N. of Dulas Court. This arch is of early 12th century date; it is of one moulded order and of semi-circular form and springs from attached shafts with weathered bases and capital carved with crude scrolls and large scallops; the N. capital has a carved face in addition; the arch is of about 5 ft span.


The carvings probably date from two decades either side of 1100 (Malcolm Thurlby pers com ).


Bannister points out that it was not until Robert was lord of Ewyas that the monks had any buildings in Ewyas. He proposes (based on the cartulary) that shortly after 1120 a site was granted for them to build their church (dedicated to St James and St Bartholomew) where previously all his father’s and his own barns had once stood. He also gave them free use of his mill. From Stephen’s rule onwards the Priory seems to have remained practically deserted and as a consequence Robert the younger refused to confirm the gifts of his father. This was finally resolved in 1196 when the grant was confirmed on condition that the Abbot appointed at least a Prior and one monk to Ewyas and the revenue from the churches at Eton Foy, Lydiard and Burnham be set aside for the next three years to go towards the rebuilding of the Priory.


The late 12th century saw a dispute between St Nicholas’ Chapel and the Priory. The monks originally had the right to “hire or fire” the chaplains of St Nicholas but Robert Fitz-Robert had bought this right in exchange for a tithe of 24s a year. Following this the chaplains overstepped their authority – encroaching on the priory’s rights in St Michael’s Church. This dispute had to be sorted out by the archdeacon of Brecon who found in favour of the monks.


It would appear that the Priory was enlarged in 1196 when they acquired a “cutilagium ” (curtilage = a single plot of land) belonging to Walter de Welyston “ad augmentum curiæ suce ” (to increase their land), and again some time before 1300 by which time it was flourishing.


John de Tregoz’s Inquisition post mortem of 1300 indicated that whenever the priory was vacant the Lords of Ewyas would place one man there and the Abbot of Gloucester another and that when the Abbot appointed a new Prior he had to be presented to the Lords of the manor (Bannister 1902, 113).


After 1301 as a means of improving the governance of the Priories the injunctions of Robert (Archbishop of Canterbury) provide some insight into their running. It would appear that monks were sent to the priories, from the Abbey of Gloucester in this case, either as punishment or for their health, and that they could not stay in any one Priory for more than a year. The Priories were also to provide for the brethren, something which by 1317 they appear to have neglected to do at Ewyas, and so the Priors were ordered to pay each brother half a mark of silver at Christmas and another at Easter. However, at Ewyas in the first half of the 14th century the revenues decreased to the point that there was not sufficient to provide even for the Prior and one monk, so they had to be fed and clothed by the Abbot of Gloucester.


There was an indenture between Roger La War of Ewyas and Thomas Horton Abbot of Gloucester (7th May 1358) where, in view of the financial difficulties of the Priory (the Priory could not afford to continue to entertain visitors as was the requirement of the time), Roger agreed to the monks being recalled in exchange for the provision of a perpetual vicar to serve in the church of St Michael and the Chapel of St Nicholas in the castle.


Another reason for the suppression of the Priory was apparently that the character of the neighbouring people was described as changed so that religious zeal is “not only diminished but absolutely destroyed”, and returning monks brought back and spread very bad habits. It was suppressed on 1st February 1358.


7.1.7 St Michael’s Church

Administratively both Ewyas and Archenfield fell under Llandaff and the Bishop there until the end of the 11th century when in Bishop Herwalds later years Archenfield was annexed to Hereford and Ewyas to St Davids. Thus the parish of Ewyas lay in the diocese of St Davids, the Archdeanery of Brecon and Rural Deanery of Ewyas.


Although it is known that a church of this name was passed to the Abbott and Monks of Gloucester in 1100 (and therefore existed before that date), there is according to Bannister nothing remaining of this structure (above ground at least). The nave, which would otherwise have provided the best evidence for the date of origin of the church in Ewyas Harold, was ostensibly rebuilt in the mid-19th century. In fact it is suggested that the church of St Michael referred to in the Gloucester cartulary was that at Dulas and not Ewyas Harold. Coplestone-Crow (1993, 19) makes a compelling case for the date of construction of St Michael’s, Ewyas Harold falling in the latter part of the 12th century. The evidence for this comes from the charter of the priory (Walker 1976, 47 [133]). This is the charter of Basilia – the then widow of Walter of Ewyas, confirming his grants:


one [acre] lay in the north part of St Michael’s, with the curtilage which Walter son of Rydderch (Righered) held, above the Dulas over against Maes-coed hill (Mescoit); Walter of Ewyas gave this acre to the monks with her consent and at her petition to endow the church of St Michael in its dedication” .


The above charter is dated as early-13th century by Walker. Other charters attributed to Walter: numbers 129 (= 1206), 112 (= 1192-1215 ) and 113 (= 1205) are late 12th -early 13th century in date, and William son of Walter witnessed charter 105 dated 1195.


On the 1866 Dulas Estate map there is also a field called Priors Field on the right hand side of the road leading from Dulas to Ewyas Harold and Priors Wood on top of the hill (Figure 7, fields 144 & 149). Considering this latter field it is worth then referring to the charter numbered 107 (Walker 1976, 41). This confirms a grant referring to land between St Michael’s Church and the Castle, and Dulas Brook and the “top of the hill”. The original grant was to make asserts. This description does not entirely make sense based on our current understanding and the modern-day topography of Ewyas Harold. On the one hand the description of land between St Michael’s Church and the Castle, and up hill of the Dulas Brook could refer to Well Orchard on the Tithe map for Ewyas. However, the small size of this land (c. 2 acres?) would seem to be at odds with granting woodland clearings – it would also have to have been wooded. On the other hand the grant refers to the “top of the hill” which in terms of the above interpretation would be the castle. There is a strong possibility that the grant is for clearings in the woodland along the side of the hill between the castle and Dulas, St Michael’s therefore referring to the church that once stood at Dulas Court. In support of this the strip of wood stretching along the upper part of the valley from level with Dulas Court and nearly to the north side of the Castle Lands is called Priors Wood on the 1866 estate map (Fig.7). The date of the confirmation is before 1195. If one accepts the latter argument then the church at Ewyas Harold village would have to post date this charter. Despite the fact that the original charter would be earlier in date, the insertion of the phrase “any other charter purporting to relate to this grant is false and not issued with his consent” implies this was more than just a simple confirmation. It is likely that some clarification was needed in relation to an earlier, and perhaps more vague grant. This being the case any topographical references are likely to still be contemporary with the confirmation of the original grant, and if the above assertion as to the locations of topographic features is true then St Michael’s may not have been built at Ewyas Harold village before 1195. The question as to how the old St Michael’s Church at Dulas would have been serviced once the monks were serving a new church at Ewyas Harold might be resolved with further and wider research.


With respect to the church at Ewyas Harold from the account by the RCHM (1931, 62) the earliest dateable architectural details come from the west tower (mid 13th century), the chancel (two 13th century lancet windows in the south wall) and trussed rafters from the nave and chancel (which, though restored, are of 13th /14th -century type).


On the basis of charter 134 there was a chapel of The Holy Cross in the cemetery by 1357.


The current Vicarage was only built in 1845 although Bannister states there was a vicarage in 1772 (1902, 87).


Going back to the church at Dulas, this was replaced in 1865 by Robert Fielden the then owner of Dulas Court. The original church was demolished (probably because it was in his front garden). One of the wall monuments in the new church is to Parry of King Street. The Parry family owned the original Dulas Court.


7.1.8 The Castle

To assist in later discussion of the castle the various earlier accounts have been arranged in chronological order below.


1533-9 - Leland

Decribes that a “great parte of Mapheralde Castell [from fitz-Harold – son of] yet standinge and a chapel of Seint Nicholas in it. Ther was sumetyme a parke by the castell. The castell stondythe on a mene hill, and on the right bank of Dules broke hard in the boton by it. There is a village by the castell called Ewis Haralde..


1645 - Richard Symonds

Could not see even the foundations of the castle or church they were “ruinous and gone ”.


c. 1800 Powell

Recorded “From the brook or river I ascended a hill covered with bushes or trees and came on a large area [the bailey] of unequal ground, overgrown with fern and woods among which are some ancient oaks. Beyond, infront, appears a high mound which was the keep, also covered with woods. Not a vestige of wall appears above ground, though I was advised that they had been dug up at times and were three yards in breadth in some parts….. South of the castle is a stream which no doubt fed the moat” (Robinson 1868, 61).


Early 19th century - Duncumb?

The following account is taken from a hand written account in the Hereford Library, at least part of which is attributed to Rev John Duncumb.

  • The house called King Street was built from stone taken from the castle.
  • As a consequence only part of the foundation survived – but not enough to interpret the types of structures that once stood there.
  • He noted the survival of the “mound of earth” (motte) and the fact that “much of the foundations appear on the top”.
  • A reference was made to “the outer slope” which “included a space of about two acres, and seems to have been of an oblong shape, which possibly (like Snodhill in Peterchurch parish) may have been that of the Castle”. [It is assumed from this that he is referring to the inner bailey ward – as this is the only part of the site that matches these descriptions.]
  • Plough shares and swords were recovered from the site during the removal of the ruins [probably the foundations].
  • The castle stood in a park but he does not provide any further evidence in support of this.
  • The Priory was founded by Harold in “Dewlas ” (Dulas) in 1100 and moved to a site “In the village”. He stated that this was united with the Abbey of Gloucester in 1300.
  • He stated that the tower of St Michael’s Church was Saxon in style with “Walls many feet thick ” and also that “in an arch on the North wall was a female effigy in alto Relcino on a yellowish stone: the hands clasped on the breast probably intended to represent one of the Tregoz or De la Warres” (Duncomb nd, 239).


It would appear from the above description that he had not looked in detail at the Cartulary of the Priory as some of the fine detail differs to that in the Cartulary itself, and the latter must be considered to represent a more accurate picture.


1887 Clarke (Fig. 16)

The majority of Clarke’s account relates to the history of the castle, its lineage and the church. There are, however, a few paragraphs of description and analysis that are at least in part revealing:

·                He interpreted the motte as being made by truncating the end of the natural ridge and raising its height by c. 40 feet (to 70 feet). The ditch formed by this continued down the slope towards the stream to the north (Dulas Brook). On the south side it ended abruptly in a “natural bank and slope. Here, however, it was in some sort resumed at a lower level and ended in a shallow ditch at the southern, or principal entrance to the castle.”

·                From the above it is evident that he viewed the southern entrance as being the main one accessed by a track following the foot of the bailey slope.

·                The keep was a shell keep, its foundations having been grubbed up to form a trench. This stood slightly towards the eastern side of an oval platform allowing room for towers or a gate house on the west side.

·                There was no trace of a way up the mound or of a well.

·                The keep had been built from a “hard schistose bed of the old red sandstone”.

·                A group of “excavations” (here meaning hollows) indicated that the bailey contained a number of domestic buildings.

·                There was a notch in the line of the bank on the north side of the motte which he interprets as a postern.

·                He identified the limekiln to the south “no doubt built of the materials of the castle; and a sort of house, now a shed, between it and the brook; but the material shows no mark of the tool, and no old mortar.”

·                There were mounds between the castle and the brook, probably referring to the holloway and bank that describes the north edge of the former ponds. He also stated there was no sign of any features on top of the ridge on the northwest side of the motte.


1902 Bannister (Fig. 17)

In the early 20th century we find one of the most detailed accounts of the castle and its environs (Bannister 1902). With respect to the early historical and archaeological discussion Bannister elaborates on theories that prevailed at the time. It is interesting how on the one hand he is willing to dismiss discussion on etmology (place names) as mere speculation, whilst at the same time happily convey the skin colour and language of the local Iron Age population. Further to this his account is very comprehensively researched and much of the information presented close to its primary form allowing the reader to draw his own conclusions from the evidence in may cases.


a)          Described the platform of the lower ward (Bailey) as being uneven and that this was probably due to the robbing of foundations by excavation in the late 16th – early 17th century.

b)          In the NE corner at point marked E (Fig. 17) the ground was most irregular and he thought the large holes must have been dug out. One such hollow SE of his postern gate measured 60 ft long, 22 ft wide and averaged 7 feet in depth.

c)          Fragments of stone with mortar attached could be dug out of these holes.

d)        He viewed the NE side as containing the chief buildings of the bailey.

e)          He described a trench running round the edge of the bailey where the walls had been robbed up and a raised dyke on the outer edge of the east side of the platform.

f)            On the northeast and southwest sides there was no trace of a ditch (it is not clear whether Bannister is referring to the motte or the bailey, however, he goes on to discuss the motte and this description perhaps better fits that feature).

g)          There was a 36 foot deep, 60-70 foot wide ditch, cutting the mound from the high ground behind on its north side. There was no sign of a dam at the NE end of this feature.

h)          At the SW end of the motte ditch there was a watercourse which continued to run round the base of the Castle Hill and joined the Dulas Brook in the village.

i)              The main entrance was interpreted as lying on the south side of the bailey (D on Bannister’s figure; Fig. 17).

j)              The quarry where Castle Lands bungalow now stands was disused by the time of Bannister’s survey.

k)          The mound was overgrown at the time but Bannister makes the following observations. It stood 70ft high. On its summit there was a trench on the southeast side facing the village 3-4ft deep, 12ft wide and c. 60ft long. He stated this trench could be traced at intervals around the top of the mound and concluded that it marked the line of the [curtain] wall and had resulted from the digging up of the foundations. On the NW side he described five projecting ridges, four to six feet wide on top, extending outwards through/from this “wall line”. The northern most extended 21ft from the outer rim of the ditch (which was 14ft wide at this point). He thought the next two projections were about the same length but couldn’t get in to measure them.

l)              He calculated the size of the keep from inside edge to inside edge of his proposed foundation robber trench as being E-W 72ft, N-S 70ft.


1908 Victoria County History

This states that Mr J.H. Round “has shown the identity of the castle referred to in Domesday [i.e. ‘castellum Ewias ’] with  Pentecost’s mentioned in the Saxon Chronicle under the date 1052. ” (Page 1908)


1931 - Royal Commission on Historic Monuments

States that “The castle is a remarkable example of a motte and bailey earthwork ”. Apart from a general description it does make the assertion that the main entrance was on the north side. It also clearly points out that there was no ditch between the bailey and the motte (and must predate any current eyewitness accounts to the contrary). It classes the condition of the castle at that time as being fairly good.


1989 - Salter

Salter makes some general statements about castle building in Herefordshire. Of interest to us here are the statements that there is no evidence that castles in the county were made of stone prior to Stephen’s reign, as well as a list of other shell keeps at Wigmore, Llancillo and Kilpeck.


Of the castle at Ewyas Harold he states “there was no evidence that any of the owners went to the expense of walling the bailey in stone, but the chapel dedicated to St Nicholas which lay within it may have been a stone building. ”.


1988 - Sprackling

Observations on the Welsh Water trench identified a blackened area taken as evidence for occupation in the outer bailey a few years before.


1989 - Stirling-Brown

He states that the work is predominantly based on the RCHME Survey. However, the following points were raised:

·                It had a shell keep with wall foundations 9’ thick

·                Signs of a tower(s) on shell wall near entrance

·                Buried foundations of barbican and stair to the keep

·                Buried foundations of curtain and other buildings in the bailey

·                Possible wet moated hornwork, or west bailey upslope of the motte ditch

·                Probable ditch between motte and bailey – now filled in

·                Clifford, Wigmore and Walterstone provide good comparisons by the same builder (fitz Osbern).


1996 - Shoesmith

States that Harold was probably responsible for moving the Domesday Borough from out of the castle Bailey to the east side of the Dulas Brook (the site of the modern village). He states “it has been suggested that there was another enclosure on the spur to the NW” of the motte, that there was probably a shell keep on the motte with a stair adjoining the curtain wall on its NE side leading from an entry into the bailey from the north.


He also reports that there is an account of excavations in the outer bailey identifying the site of the village and main entrance to the north. This refers to the RCHM (1931) report, but there is no reference here – so it may be in the original survey notes. There was a watching brief in the outer bailey reported in 1983 (Wills).


7.1.9 Cartographic and aerial photographic evidence

No detailed maps of the castle itself prior to Benjamin Fallows’ survey of 1718 were uncovered during the research. Maps earlier than this date only tend to identify the location of the settlement itself. Saxton’s 1577 map (Fig. 8a) identifies the settlement as “Harlewas ”. Fortunately John Speede’s spelling is a little better in 1610 (Fig. 8b) but in this case he labels the castle and not the settlement, “Harldewas cast ”. Interestingly both these maps identify “The old court Dewlas ” on the Escley Brook. With respect to the Royal Commission’s account of the structure surveyed inside the 1866 rebuild of Dulas Court (c. 1600) and Robinson’s manorial history of 1872 which does not identify any earlier lineage than John Parry during the reign of Elizabeth I, it is difficult to surmise when the estate at Dulas Court was established. Both maps indicate the current site of Dulas Court as “Llanihangel “ (1577) or “Llannihangell” (1610) – my best reading of this being St Michael and All Angels church. There is some debate (Smith 2004, 20) as to whether on Saxton’s earlier map the labelling should be read as “Llanihangel Dewlas” or that Dulas was part of “The old court Dulas ” as on Speede’s map. At the very least this provides early evidence for the continuity of the dedication of a church at Dulas.


In 1718 a survey of the lands around the castle at Ewyas Harold was undertaken for the Abergavenny estate (Fig. 9). This is also very similar to the plan of 1805 (Fig. 10) which probably copied it. It shows the lands owned by the estate but of possible equal significance are those not identified on this plan. The castle and its ward are clearly marked (Castle Yard), as is King Street which on this map is shown as continuing as far as the line of the brook but no further eastwards. It also shows a branch running north-westwards along the southeast side of the bailey to the entrance that lies on that side of the castle, although it is not detailed enough to show the entrance itself. The earthwork ditch is shown and labelled around the southeast and northeast sides of the bailey and a gate is marked into this area on its east side. The road to Dulas is not shown at this point but does appear further north. The land on the east side of the ditch was owned by John Prosser, fields to the east of this and adjacent to the Dulas owned by Philip Jackson and the field by the mill by William Parry (probably associated with the Parrys of Dulas). William Parry also owned the land either side of Lords acre with Hannah Parry owning that plot herself (Plot 102 on Fig. 11). Hannah Parry also owned the “village field”, or the half of it to the west of the boundary marker stones. The fields in which the priory is believed to have stood are not included.


All those fields on the ridge northwest of the castle have castle field names and extend to the boundary with Dulas and the end of Priors Wood. The larger field near to the castle is called Park Piece, Castle Yard and Ditch - one thought is that the boundary here is that of the park pale. To the northeast of this are Little and Great Castle Field. The small piece identified as “taken out of Lords Wood Common” is probably no. 257 on the tithe map (Fig. 11).


Of the surrounding topography the northeast edge of the map ends along the line of what is now Ewyas Harold Common and was labelled Lords Wood. In 1805 this is labelled Lords Wood Common. Those fields on this part of the map relate to the operation of the lord’s mill which is also shown. Otherwise all the fields on either side of the mill leat are included, but not Weir Piece (no. 240 on the Dulas Estate map of 1866) on the right bank of the Dulas where the leat joins the brook. A house is shown on the other side of the brook to the mill.


By the time of the Tithe Map (1844, Fig. 11) the Lords Mill has disappeared, although a house and garden are still marked in a similar location to the ones shown on the earlier maps. On this basis the mill (or part of it) may have occupied the plot labelled “plantation” (no. 254). The other point to note from this map is the ownership of King Street Farm (Fig. 11, black outline). King Street Farm may predominantly derive from that land owned alongside the castle. It is quite certain that other plots may have been sold off or added in the intervening years. Those fields south of the Castle such as the village field and priory field all appear to be under different ownership by this time. We know that castle field fell under the original estate from the earlier maps mentioned above. However, it is interesting to note the division between the Castle Field and Castle Orchard – this being along the line marked as stones on the 1888 Ordnance Survey first edition, with Castle Orchard lying outside of the estate on the 1718 map over a century prior to the tithe survey. This needs to be considered alongside other evidence when considering the land which once belonged to the priory.


Clarke’s plan of the castle (1876) implies that the Dulas road ran along the top of the outer bank rather than through the base of the castle bailey ditch as it does now. However, even a cursory comparison of the 1844 tithe map and 1888 first edition indicates that the road has not moved between these dates and on the First Edition it is clearly in the base of the ditch (Fig. 12). There is no evidence to suggest it was ever different to this.


On the tithe map there is a plot of land known as Sling (Fig. 11, field 164). On the 1st edition map (Fig. 12) the part of this plot adjoining the castle moat has been partly quarried away against the hill side, and a pond is shown towards the top of the ridge in the corner of the plot. However, Clarke makes no mention of this quarry in his description of the site in 1876. His description is exhaustive of the whole circumference of the castle and bailey, but on this side he simply refers to the ditch being continued down the slope and a “notch” in the bank being that of a postern gate. Assuming the “notch” in the bank is the gap through which the modern bungalow is accessed, then on the basis of his drawing and description it would appear that the quarry was formed between 1876 and 1888. The quarry may well have been active at the time of the First Edition OS map. On the map it is referred to as “Quarry”, whilst other hollows further up the fields on the ridge are called “old quarry” implying disuse. Also the tracks are all shown connected to the quarry and a building near where the dairy now stands. Clarke also omits any mention of, and does not show on his map, the modern day access up the north rampart bank to the bailey. However, this again appears on the First Edition OS map. Unfortunately it is difficult to rely entirely on Clarke because there are elements of his plan that do not fit other evidence, such as the position of the Dulas road with respect to the ramparts and the configuration of the southern bailey rampart, which does not match the layout of features 10 years later. However, it does not appear that he was attempting to produce a reconstruction of the layout of the site either, so his observations on the north side of the site may well be valid and the quarry and access up the north bank of the bailey are therefore features that appeared in the decade following his survey.


It is necessary to consider the layout of the plot that the quarry lies within as this must predate the quarry itself. The long linear stretch of the “sling” along the side of the hill was known as “the king’s gallop”, the area towards the ridge contains a pond but does not appear to have been quarried to its bounds. In this case the later quarrying appears to follow what was an earlier plot using an existing entrance with a track being added up the side of the rampart to access it. In actual fact, on the tithe map (Fig. 11), the boundary used to run straight through the quarry workings in a similar fashion to the northwest side for the continuation of Clarke’s ditch. The house below exists on the tithe map and part of its plot is now cut from the side of the hill. Based on these latter two points it would appear that Clarke’s interpretation can be relied on here and the quarry began to operate between 1876 and 1888 (Fig. 12).


By the time of the finance map in 1904 (Fig. 13) the features around the quarry have not changed since the First Edition survey, implying that the quarry had probably been worked out by then, and the pond is still shown in the top field above the quarry. This map also shows the distribution of land ownership following the turn of the century. The land held by Marquis of Abergavenny is similar to the holding of King St Farm on the tithe map. It is also worthy of note that the “village field” (Fig.13; 281) is split between the Marquis of Abergavenny (no.155) and the Rev. J. Jones (no. 35), the boundary between these two segments still being marked by a line of boundary stones (which is what is being referred to on the First Edition OS map in this field).


There is little change by the time of the Third Edition OS map in 1920, although on this map a small triangle at the west end of the “village field” is enclosed as an orchard (Fig. 15).


From air photo evidence a track that runs across the bailey at the foot of the motte is clearly visible as late as 1971 (OS 71398 - 100) and although less clearly in use can be made out on an oblique photo of 1990 (Plate 1a). This latter photograph also shows a number of buildings in the bailey, and the line of the pipeline in the village field – apparently as parch marks.


A more recent photograph of the site (Plate 1b) clearly shows the original route of the Dulas Brook (no doubt the section inside being what was once referred to as Island Piece). There is also a hint of some linear features in the upper field, although the configuration of these is reminiscent of drainage rather than something more ancient. The site is shown in its modern context in Plate 2.


7.1.10 Archaeological evidence

Very few archaeological observations have been undertaken in and around Ewyas Harold. On the castle itself there was a watching brief on work to the steps leading up to the inner bailey. This revealed little except for pottery dating from the 16th /17th century near the top of the bailey bank (Dawkes 1999).


In the outer bailey or “village field” a watching brief was carried out during the laying of a water pipe across the field. It would appear, from the record associated with this, that at the west end of the pipe line where it ran parallel with the southern boundary of the field occupation deposits survived (SMR event no 21498, Wills 1983). They extended for 5m eastwards before fading out and comprised a 0.3m thick deposit of clayey silt containing charcoal, crushed and burnt stone, bone and medieval pottery. This deposit was sealed by 0.7m of clean clayey silt which the excavator believes eroded down the slope to fill up the gap behind the bank here. There was also medieval pottery and bone recovered from the surface of the field during fieldwalking in 1988. The field is referred to as having been ploughed flat (Buteux 1996, 4), thus eradicating any evidence for buildings which were observed as late as 1952 in the area of the priory (SMR event 21497).


Buteux (1996) also mentions a further watching brief to the north of the church where a bank was being altered during development. In this case she states that no dateable finds were recovered. However, the SMR entry (5321) indicates that pottery was observed in the exposed clay ends of the feature. Also this feature has been attributed a medieval date. However, there does not appear to be a reference to the type or date of pottery recovered. Other features in the vicinity include a cross slab of c 1254 found in the wall of the castle inn (SMR 19255) and the base of a 14th or 15th century churchyard cross standing 15m south of the church (perhaps associated with the Chapel of the Holy Cross recorded here?).


In the notes for the Historic Marches Towns Survey it was also mentioned that the gable end of the church contained a bread oven and that the new school was built on the site of the medieval vicarage (built after the suppression of the priory). The survey also identifies a plot to the north of the bank near the church as being the site of a moat (SMR 10655). However, the supporting evidence (an air photo) would appear to relate to the layout of playing fields rather than anything more substantive. One further reference in the survey notes is to the monk’s graveyard which is reputedly located at SO 3880 2879 (SMR 19273, based on Sprackling 1988).


7.2       Survey of the extant remains


In this case there are two reasons why the importance of historic surveys of the castle cannot be underestimated. The first is that, over time, features become more obscure with the development of soil and hill slope erosion. The second is that the introduction of new features into the landscape can either partially or fully remove evidence that once existed.


The map evidence has been covered above. However, as an introduction to this section there are three earlier surveys of note, one by Clarke in 1876 (Fig. 16), a second by Bannister 1902 (Fig. 17), the last by Kay in 1952 (Fig. 18). Setting aside the differences between these they show evidence for structures on the top of the motte, evidence for structures around the outer edge of the bailey (Bannister refers to these in his text), and an entrance into the bailey from the south. Kay’s map also shows the full extent of the ponds by the priory field, and Bannister shows part of the old ox-bow of the Dulas Brook.


7.2.1    General description

The motte measures 74m by 67m at its base and 36m by 39m across the top. In the centre of the top of the motte there is a level area measuring 20m by 22m. The top of the motte is 17m above the highest point of the bailey. The bailey itself measures between 120m and 90m in length and 95m to 60m in width. It slopes down from the northwest to the southeast.


7.2.2              The keep (Fig. 19)

From the hollow marked as A1 onwards there are a series of hollows around the top of the mound which are described anti-clockwise from the point where the present path up the motte ends.


A1 and A2 were both the same length 7m, width c. 4m, and depth (c. 2m), and they were both orientated north-south. Their northern ends were open at the break of slope for the top of the mound. The same was true for the other hollows on this side and as far round as F. The boundary between A2 and B was on a slightly different alignment to that dividing A1 and A2 indicating a change in orientation (possibly at right angles to the tangent of the original motte top). B was c. 7m long by c. 6m wide. A small circular hollow lay in the central courtyard about 7m from the edge of B. This has been interpreted as a possible well measuring c. 2m in diameter. Hollow C was slightly more irregular in form with a hint of a dividing wall one third of the way along it. The angle of its two ends made it trapezoidal in plan, its width from the edge being c. 8m, length along the outer circumference of the motte 15m, and inside length (along the side of the central courtyard) 12.5m. D was a rectangular shaped hollow 3m by 3m with another slightly circular hollow (E, c. 3m diameter) next to it. From this point there was a c. 20m long, level-based depression (F) – 7m wide at its greatest followed by a long, c. 5m wide, curving hollow, in this case with a counter scarp along the top of the edge of the motte (G). Between G and A1 was an area of fairly irregular earthworks about 20m in length by 8m in width. It appeared to comprise two slight hollows either side of a low, c. 3m wide mound.


Referring to earlier plans it is possible that C and B are what Clarke (1876) interpreted as two towers forming an entrance on this side of the motte. Kay (Fig. 18) identified A1 – C although he appears to place D where H is and on the lines of D – G interprets a single curtain wall.


This latter point raises the question of the robbing of the curtain wall. All previous authors refer to the hollows on the top of the mound as representing the resultant quarry holes from the robbing out of the curtain wall. From the shape and layout of hollows A1 – D, however, then there is little doubt that these must represent the footprints of the bases of structures. The greater depth of these in relation to the central court would imply a cellared range. Kay does not appear to suggest that these should be considered otherwise in his plan of the site. However, in the case of F and G he and others suggest these were trenches for the robbing of walls. There are two pieces of evidence from the top of the motte that might suggest F was the remains of a footprint of a structure rather than robbed out curtain wall. The first is that the ends of the spaces defined by A1 – D are all open (as is F), the second is the presence of a counter scarp for G and the fact that this changes direction part way along its length without an internal segment for a wall (as is the case in all the other instances on the circuit). The indication is that the curtain wall most probably closed off the open ends of A1 – D + F, and that the hollow at G defines the line of two (12.5m long) segments of the curtain wall with the structures at each end projecting forward from this line.


It seems probable that the keep would have been accessed from the bailey side and therefore the best candidate for a gate house is probably located at H. This is because it is unlikely that the entrance would have been across the top of one of the cellared rooms around the north side of the courtyard and the indication of two shallow hollows on either side of a slightly raised area between best matches a guard room either side of the position of what might be a collapsed arch. E may well be the site of a tower. Further investigation would naturally be required to even begin to resolve any of the above assumptions.


7.2.3    The Bailey defences (Fig. 20)

The majority of the bailey was defined by a steep scarp. Whilst it might be said that there is an associated ditch with this, the only places where this is substantial are locations where roads or routes pass around the foot of the defences. Given the steepness and lack of vegetative cover on many on the slopes forming the bailey rampart it is most likely that it has been formed from the natural strata. If it were artificial then a much greater level of erosion might be expected. With respect to the presence of a ditch then we have to rely on antiquarian accounts and in particular Clarke (1876).


On Clarke’s plan of the site the quarry has yet to start being worked so he was able to observe the presence of a ditch on the northwest angle of the bailey defences, carrying the motte ditch down to the road. He describes it as such. This was not a substantial feature, but there were indications at the time of his survey that it carried on around the foot of the bailey (e.g. the hint of a ditch near to the southern entrance to the bailey).


Along the lines of features such as King Street and the road to Dulas the ditch may well have been deepened through its use as route ways. Had the feature been originally fully excavated then a ditch of similar proportions might be expected on all three sides of the bailey – this not being the case lends support to the argument that the track ways are in someway responsible for the depth of the hollow. In this case then the end of the hill may have been cut back more steeply and the material spread down slope. The man-made features that formed part of the bailey defences appear to be the banks marked I on Figure 20, and the remains of another bank at J. During the survey it was established that I contained a solid mass of stone as it was not possible to insert any sort of the grid marker in it. At J, in the side of the motte, there was an exposed face of roughly coursed masonry with chamfered ends. A possible interpretation of this is that it was a similar feature to that seen at I, but had been cut back at the time the dairy was built levelling the area by spreading its constituents into the ditch. A concrete platform now stands there. There is also evidence of a slight counter scarp, mainly on the other side of the fence outside of the area accessible for surveying, which could be the foot of the wall for the bailey.


7.2.3              Entrances to the bailey (Fig. 20)

On initial inspection of the site there are five candidates for entrances into the bailey. Starting at the north and working clockwise there is the gap within the bank (Fig. 20; I) where the drive leads to the bungalow, the drive leading up to the cattle grid (Fig. 20; S), a slight hollow where the footpath enters the bailey (Fig. 20; Q), the offset section of rampart on the southwest side of the bailey (Fig. 20; T), and the gap on the south side of the motte (Fig. 20 J). It has been established above that S is a recent entrance dating from between 1876 and 1888 and similarly that there was a complete rampart sealing the access at J in 1876. An access at Q would require a bridge, and certainly fits the traditional view of a castle entrance i.e. a draw bridge across a moat. It also coincides with a long standing boundary across the field to the south of the bailey ditch (Fig. 20, U). There is no natural topographical explanation for the position of this boundary within the plot it divides - its orientation does not match any of the other sides of the field, and its location creates irregularly shaped, different sized parcels. One rather attractive proposition is that this was originally the road way leading to the castle. Such a configuration may also help to explain some of the descriptions in the cartulary (Walker 1976).


There is a reasonable consensus amongst authors for the two other good candidates for original entrances. The hollow that bisects I would appear to be original, as using Clarke’s survey (1876) it clearly predates the quarry (the only other contender as a reason for its construction). Whilst this entrance leads straight into the continuation of the motte ditch (according to Clarke) it is also notable that there was a slight terrace at the top edge of the steeper scarp on this side of the hill (Fig 20 , V). This was called the “Kings gallop” (Margaret Woodward pers com ) and as such may be a more ancient route to Dulas (and the original site of the foundation of the priory). The final candidate for an entrance is on the southwest side of the bailey at T. This entrance is clearly of some antiquity as it has at one time or another been connected to the line of King Street (Fig. 20, W) and a hollow way southwest of the castle (Fig. 20, X). Also considering the layout of features inside the bailey it would appear that this entrance was probably in use at least for part of the castle’s occupation because there is a low lying area (Fig. 20, N) at the point where it enters the castle and no sign that this was dug out at a later date.


There appears to have been another track that leads up west of J next to the dairy but this most probably relates to the operation of the lime kiln in the bank near J. Whether this kiln was once a feature of the castle is difficult to determine. Clarke (1876) thinks not, because it had modern mortar in its construction, however, as it is likely that any original feature would have been pointed using lime mortar it is not surprising that it was repointed before the end of its life as a lime kiln.


7.2.4    Bailey internal features (Fig. 20)

Within the bailey the first most noticeable feature is the variation in levels. These must by necessity relate to the natural topography of the original hill slope with a gradient from the motte down towards the village (northwest-southeast). This hillslope has been roughly crafted into terraces; the uppermost around the base of the motte (predominantly around K) which it joins, a middle level to the south of the modern track at its west end, the lowest following the top edge of the bailey rampart on its southeast edge. There is little doubt that the purpose of these terraces was for the location of structures within the bailey.


It is difficult to be sure if the platform at K was once the site of a building on the basis of the topography alone. Local knowledge suggests that this part of the site was used for storing tarmac at one point, so any levelling here might relate to that more recent phase of the castle’s history.


Plate 1 clearly indicates the presence of structures next to S, at L/M and also at O-R (Fig. 20). The topography at L/M indicates the possibility of structures at different levels and perhaps implies more than one phase of construction here. The series of hollows and humps at O-R is more indicative of structures that might be contemporaneous.


7.2.5              The “Village Field” and “Priory Field” or outer bailey (Fig. 20)

This field measures 2 acres and is raised above the level of the surrounding ground by a steep scarp on its west and east sides, with a man-made bank forming part of its southern side and part of the hollow of King Street forming its northern side. There are two portions of its circumference where it does not appear to be defended by any feature. The first is at the west end of the stretch of King Street forming its northern boundary and the second is the location of the butcher’s shop and its immediate environs (i.e. the southeast angle).


Within the “Village Field” the only features apparent are the former boundary at U and a slightly raised area of unknown function at Y.


The “Priory Field” contains the bank – already mentioned – at Z and another small section of bank (AA) at right angles to it running south towards the village (see 7.2.6). Comparison with earlier maps of the site demonstrates that the bank, Z, used to be much longer than it is at present (even from the 1970’s Ordnance Survey map) and ran further towards the present butcher’s shop. The evidence for any continuation of the circuit of defences around the eastern end of the outer bailey was probably lost when the original site of the butcher’s shop was developed prior to 1888.



7.2.6    The fish pond (Fig. 20)

The area known as the fish pond lies between the outer bailey and Prill Lane. Although the majority of the evidence relating to this has been affected by the construction of cottages in the feature, it is still possible to surmise the layout of the features prior to this. Its western extent is defined by a slight dip at the west end of AB, its east end being the bank at AA, which from earlier maps joined the bank Z. Bank AA was probably the pond dam. It also appears that there was a change in the level in the base of the pond at a point where an earlier field boundary was mapped (division between AB and AC). AC is lower than AB resulting in what may have been two ponds with a dam at their east end and fed by the Prill Brook at their west end.


7.3       Geophysical surveys


The results are described below with magnetic methods first, then resistivity and radar.


7.3.1    Magnetic susceptibility (Fig. 21)

In certain cases this method can provide evidence as to areas where human occupation has taken place. In this instance it was hoped that it might assist with the interpretation of occupation areas. Some areas of enhanced susceptibility may be more recent in date.


At field entrances it is common to tip bricks etc - this will result in an enhanced response from the firing of the clay that made the bricks. In the bailey it has been reported that there was a tarmac store, and there is a visible tarmac track. Both these are likely to produce enhanced susceptibility readings. An area of hollow such as the pond may also be filled with material that is different from the natural clay and so this will show up quite clearly too. The following responses are worthy of further note.


On top of the hill to the northwest of the castle on the centre of the ridge there was an area of enhanced susceptibility (MA). This coincided with an area of enhanced resistivity and therefore is probably geological in origin such as an outcrop of harder, rocky material (variations in geology can have a dramatic impact on magnetic susceptibility levels). Another possibility is that it is where cattle have always been fed, and their manure has enhanced the soil’s susceptibility.


On top of the motte there is a slight enhancement in the area where the possible cellars were identified in the earthwork survey (MB), which may be an indication that these are filled with occupation debris.


Despite the presence of tarmac in the bailey there appears to be quite a broad spread of enhanced readings around the foot of the motte (MC). These could relate to structures once located here. There is also a spread of enhanced susceptibility on the lower level within the bailey at MD. This appears to coincide with the known locations of bailey buildings.


In the outer bailey there is an area of enhanced susceptibility (ME) which occurs across a field boundary. In this case it may again relate to the feeding or watering of cattle as there was a trough at this location in the “priory field”. There is also an area of enhanced susceptibility around the butcher’s shop (MF) which probably relates to recent spreads of material.


The fields along the west side of the site demonstrate a suppressed level of susceptibility typical on wet and boggy ground. This is illustrated in Figure 21 by a concentration of blue shading.


7.3.2    Fluxgate gradiometry (Fig. 22)

From the results of the gradiometer survey the reason for some of the areas of enhanced susceptibility become immediately apparent. In the area of the pond and around the butcher’s shop the responses indicate considerable amounts of ferrous (or iron) objects. This implies a spread or dump of rubbish.


In the bailey field the tarmac track is clearly visible, however, there is little indication of the tarmac tip against the base of the motte so the susceptibility responses here may well reflect historic occupation rather than more modern activity.


Anomalies worthy of note are two parallel faint linear features at A. These are likely to be caused by two linear cut features and one possibility is that they are the sill trenches of a timber structure (perhaps one that burnt down). At B in an area where there is the known site of a building there are a number of isolated responses. Where deposits accumulate within the footprint of rooms or buildings they can produce more magnetic responses to their surroundings and certainly the responses here are indicative of that sort of activity. There is also a faint linear magnetic band at C near to the bailey rampart, but interrupted by the stronger responses at D.


Two further responses that may be more recent can be observed at D and E. These show high levels of magnetic activity akin to responses from iron objects or very heavy, uncontrolled burning.


There were some similar responses in the field to the south at F. In particular one small rectangular shaped anomaly here is similar to responses from some types of brick clamp kiln. However, there are no enhanced responses around this area on the susceptibility survey (Fig. 21) to support this interpretation.


Two pipes were located one in the “village field” the other along the line of the continuation of King Street.


Two traverses of gradiometer survey were undertaken across the top of the hill behind the castle but did not reveal anything of significance.


7.3.3 Resistivity survey

The results from this survey have been divided into the various elements of the site where appropriate. Remaining features are discussed using the overall survey plan – Figure 23. In all cases the general data trends are considered first, followed by a more detailed assessment where appropriate.



The keep (Fig. 24)

The resistivity data demonstrates a clear pattern on the top of the motte. This presents itself as a central area containing low readings (blue), surrounded by discrete blocks of much higher readings (red). Remember high readings can be caused by buried masonry or rubble. Some very strong readings near the edge of the top of the motte are most likely to be due to massive buried masonry or rubble. The readings form a clear enough pattern on the south and east sides to allow a reasonable degree of confidence that they represent the internal face of the curtain wall. In those instances where there is an indication of the outside face of the wall then the estimated thickness of the curtain wall would be 3.5m near its base.


With respect to the features identified on the earthwork plot (Fig. 19) the best candidate for a room space (supported by both sets of data) is at F. Here the resistivity readings suggest that there was once a partition wall across the angle of the curtain wall, possibly with an entrance, and also perhaps containing a niche in the curtain wall. The hollow G that runs on from here appears to follow the area of higher resistance and could well represent a robbing trench for the curtain wall. As such this might explain the gap between G and H1, possibly a point where the curtain wall had been completely robbed out; however, there is not sufficient variation in the size of the hollow to support this so another explanation is required.


The other stretches of curtain wall foundation that appear to survive are beneath H1 and H2 and on the motte edge by B and C. In the case of H1 and H2 there is little evidence from the resistivity data that this was anything other than solid curtain wall and the hollows could simply be a remnant from the robbing of the walls. At C and B the position of the wall helps to define the size of structures abutting it, these being c. 4m square. There is some evidence from the resistivity survey for the presence of masonry between the hollows A1 – B. There is also an indication of rubble or a solid base to A1, A2 and B. It appears from the resistivity that there is little surviving evidence for a wall between the courtyard in the centre of the keep and A1/A2. A lack of data between C and B (due to obstructions) obscures any evidence for masonry in the division between these spaces. The hollow at E would appear to relate to the partial robbing out of the curtain wall. The area around D and C has been much more heavily eroded than other points on the circuit possibly implying a greater degree of robbing out here. Perhaps little or no wall footing now survives at this point to hold the edge of the motte in place.


The inner bailey (Fig. 25)

The outlines of modern tracks (1) and historic buildings can be clearly seen in the data.


The most substantial structure within the bailey (2) comprises a building with a rectangular footprint measuring 30m by 10m. It appears to be divided into two parts. The southern half comprising a series of interlinking spaces on the ground floor. There are indications of increased wall thickenings that could be the basses of buttresses or chimney stacks. The resolution of the data is sufficient to identify quite fine partitions (or maybe drains) within the building. The modern track would appear to have been built on top of the line of one of the building’s main walls (this line probably being selected because of the well-drained hard ground it already provided). Given the scale of this structure in relation to others within the bailey it is proposed that this was probably the lord’s main residence at some period during the castle’s occupation. What is also quite striking is the area where no structures are apparent at all around Building 2. The building was situated higher than any others around, its dominance probably being further emphasised by an undeveloped curtilage/garden around it. The main entrance to the building appears to lie on its east side.


The north end of Building 2 has an angled, almost apsidal, wall with what appears to have been a central buttress. On the axis of the buttress and an internal wall line (the latter part of a massive lump of masonry - perhaps a tower base) there appears to be a pillar base (interpreted as such because of its central location. Initial thoughts tend to the chapel of St Nicholas, and although the orientation of the axes of the chapel are N-S and E-W, there is no clear length to the space or truly defined altar end. Also there appears to be an entrance in the north wall. Chapels in castles are always positioned very close to the true E-W alignment expected of such ecclesiastical buildings.


From Building 2 there appears to be a passage way (possibly a private one) leading to the solid-walled (c. 4m thick) structure, measuring about 15m across with multiple sides, and a 3m thick wall leading from it towards the motte. This is situated next to a gap in what now appears to be a defensive bank and was probably part of a massive gatehouse and defence at this location. It is probable that from this position steps accessed the keep on top of the motte.


Within the remainder of the bailey and lying at a lower level the footprints of more than ten structures can be identified. It is likely that those at 4 and 5 (attached to the curtain wall) are the locations of towers. The structure at 6 could be part of an inner bailey gate house, located at the end of the line of a possible road leading across the outer bailey or ”village” field. Other entrance features appear to have been located at 7. The sequence of three structures at 8 indicate that the buildings within the bailey are of more than one phase with what was apparently one building superseding another at this location.


The curtain wall was only apparent on the northeast side of the bailey, the remainder may lie on the other side of the fence in an area that was inaccessible during the survey.


The outer bailey – “village” field (Fig. 27)

Looking at the general trends in the data there are two distinctive bands of high resistance readings. A northwest-southeast band measuring c. 10m in width across the whole field and with a very straight southwest edge, and a roughly east –west band, slightly more irregular in layout, but comprising higher readings and which appears to stop at the aforementioned band. A further area of high resistance occurs in the eastern part of the field. An area of low resistance occupies the central northern part of the field.


In the east-west aligned band of higher resistance, areas of discrete high readings occur in blocks measuring about 15m in length by 6m in width. These apparently have the same orientation.


It is important to consider a geological explanation for the bandings and higher readings at this location before (potentially erroneously) providing an archaeological interpretation of the features. From the geology of the site the south edge of the “outer bailey coincides with a terrace edge (possibly glacial in formation) with the land to the south of this now lying in fluvial catchments and comprising alluvial deposits. The southeast end of the enclosure has been extended into the alluvial plain, thus the need for a constructed bank at this point. From data in the field to the west it is clear that the zone of higher resistance continues, fading out as it approaches the line of the old track and the stream. From this it would appear therefore that across the southern part of the field there is a zone of higher resistance ground. This band may well have been well drained. What the geophysical data may be indicating here is the slight terracing of the hill slope at a point where the ground was well drained for the purpose of constructing free-standing buildings. The areas of slightly higher resistance could be sill walls, but equally could relate to variations in geology. From the results of the survey it is possible to say that occupations of those areas that are higher in resistance may be more preferable to the softer ground in the central northern part of the field.



The outer bailey “priory field” (Fig. 28)

The most striking feature in the data for this field is the 5m wide high resistance response from the large bank that enters the field. From the map evidence it has already been established that this feature was foreshortened within recent times, however, the resistivity provides an accurate indication of its extent as shown on the 1888 First Edition survey (Fig. 12).


Attached to the north edge at the east end of this feature there appears to be evidence for a structure (Fig.28, 1). This measured c. 6m in width and about 25m in length. It appears to have had some internal divisions. North of this there is an area of rectilinear high resistance responses measuring c. 20m by 20m (Fig. 28, 2).


To the south of the bank there is a band of higher resistance separated from it by lower readings. Some indication of rectilinear features here might represent the location of the structure at this location too (Fig. 28, 3).


7.3.4   Radar survey of the motte and bailey (Figs 26 and  29)

Radar survey is responsive to the conductivity of objects and features buried beneath the ground. As such it is likely to produce results on a par with resistivity. It can also provide some vertical section information through the data. With respect to the areas surveyed in this instance it is estimated that the speed of the signal through the ground was about 0.05m per nano-second (this being quite slow). The depth of detection does not therefore seem to have exceeded 0.5m in most places across the site, which is likely to be in part due to the moist nature of the soil and clayey nature of the bed rock. In Figure 26 the impression of structures already identified through the resistivity survey are clearly apparent. It also supports the existence of a blank area between the main residence and other structures on the site (if that indeed is what the structure was). The modern tracks show up clearly above the responses from earlier buildings. The radar data also demonstrated a degree of stratigraphic complexity in the bailey.


On the motte very little in the way of clearly defined internal features showed up. Exceptions to this included the Curtain wall and a possible structure in the north-east angle of the keep. A structure was partially identified here on the resistivity data.


Of the individual profiles surveyed, two demonstrated data worth presenting as well as a further traverse being selected for presentation on the motte top. This latter traverse (Fig. 29, A) shows the position of the well. Also at c. 15m a 2-3m wide structure can be observed in the data. Traverse B on the same figure shows a section through the site of the possible tower on the edge of the bailey the sharp peaks here possibly indicating rubble, whilst C clearly shows the location of a rampart, vertically sliced through by the radar, which would have connected to the base of the south side of the motte with a ditch/hollow either side of it.



8.         Discussion

The study uses archaeological methods - as such it will provide archaeological answers. Further resolution of the history of the site is probably not practical given the nature of the sources researched. It is also quite clear now that the study area required to satisfy fuller interpretation of the evidence relating to the castle is far more extensive than that established at the outset of the project. Any future study would need to consider evidence from the Dulas valley as far as Pontrilas and across to Bilbo and Dore as well as the area of the modern day village. However, the study as undertaken here has provided a very considerable amount of further information about the castle and its surrounding landscape and some relationships between this and the documentary evidence have begun to appear.


Key points are:

-                  The nature of structures associated with the motte and bailey

-                  Past activity relating to the lower bailey

-                  Any chronology relating to these that is discernable

-                  The location of the Priory.


8.1 The early castle


From the documentary account regarding William fitz Osbern’s post-conquest castle there appears little doubt that it reused the foundations of Pentecost’s castle. Physically there does not appear to be any other alternative. The only other potential site on the ridge behind the motte does not show any evidence for a moat (based on the resistivity survey) and would find its vistas partially blocked towards the bailey by the end of the ridge. The motte therefore might be proposed as being the site of the pre-conquest keep. Some debate has been made as to whether this was stone or timber. The traditional view, without worrying about collecting supporting evidence or any other such issues, was that it would be timber.


The topography of the end of the ridge on which the castle had been built probably had a naturally stepped form. So the motte is cut out of the end of the first step. The second, shallower, step delimits the edge of the bailey. It is again likely that this enclosure formed part of Pentecost’s castle. It is not possible to determine how developed the defensive features would have been at that time but it might be expected that some form of ditch surrounded the platform within. The last and lower step is the site formerly associated with the borough. It now has a well developed bank surrounding some of its sides. These natural steps in the hill slope are clearly illustrated in Figure 20.


If we were to accept the association of the borough with the settlement of Mulstoneston it would mean that at the time of the Domesday survey there were 12 smallholdings here. Coplestone Crow (pers comm. ) suggests that Manitone lay on the east bank of the Dulas Brook.


The evidence recovered from the resistivity survey, whilst not entirely conclusive, does indicate the potential sites of what may have been house platforms along an area of better drained ground adjacent to the bank forming the enclosure’s southern boundary. There also appears to be a building against the surviving end of the bank in the field at the rear of the Butcher’s shop (“Priory Field”). The results from other surveys provide little evidence to support occupation here, but archaeological observations in this part of the site (Wills 1983) identified 0.7m of silts overlying medieval occupation deposits in this field so it is likely that responses may be somewhat masked in the lower part of the field at least. There is therefore evidence that this lower enclosure was occupied at one time, and by being attached to the castle it seems likely that this occupation was quite early in origin. If this were the site of Mulstoneston then the arguments against it being long established as an English settlement do not tie in with the reputed defence of the valley in c.915, which implies a considerably more ancient occupation at Ewyas. Where is the evidence for this earlier defence of the Dulas valley? It appears to arise from a postulation on the part of Bannister – so may not have been built there anyway (there is after all only one defended burgh whose location is still not known).


To sum up, at the time of the Conquest, the documentary and survey evidence suggest the existence of a castle, bailey and (defended?) borough.


8.2 The Priory


At some point in the latter part of the 11th century Harold becomes Lord in Ewyas. His significant action was the foundation of the priory. The evidence leans strongly towards the priory being founded at Dulas, and as such most of the descriptions in the cartulary of around this date will also refer to Dulas. There is a strong body of evidence to support the fact that St Michael’s church was located at Dulas, with no church of this name in Ewyas until c. 1195 (when there was a grant relating to its dedication). The “negative” evidence supporting this being that there are no in situ pre 13th century remains at the church. Unfortunately this would also mean that topographic references referring to St Michael’s before this date relate to the topography of Dulas and not Ewyas Harold.


After 1120 the priory was relocated in Ewyas Harold – on the site of Robert’s (and Harold’s before him) barns that were enclosed by a ditch. There was a grant of wood for building houses dated 1130-39 so the priory may have actually moved around then. It has been suggested that Robert relocated the borough to the other side of the Dulas, and that the priory then occupied the area vacated by this. There are two arguments from the documentary evidence against this. Firstly, the original cell, which was small, was relocated to a site formerly occupied by barns that were there in Harold’s time; therefore the ville would have had to have been moved by Harold so he could build the barn in the first place. Secondly, the site (which does not need to be very big at this time) is recorded as containing barns and not a settlement.


So where was the priory? The Cartulary contradicts modern assumptions. These have all been referenced to Walker’s 1976 numbering (in square brackets throughout). They are sorted chronologically and topographically, where possible.


In around 1120 the priory’s nucleus was located in an area where the lord’s barns once stood, enclosed by a ditch and they were given the ditch (perhaps the same one enclosing the barn) and the fishery there [100]. So it was near some ponds, not necessarily ponds that are visible today, and as such must have been near a water course. A later grant seems to support this latter assumption. The priory was given half a burgage (confirmed in 1198-1215 so this must have happened before this date) near the old moat (or ditch) by the water of the mill [112]. There was also an exchange of land for where the “new mill now stands” in c. 1213/14 [106] indicating the new mill had just been built around that time.


Between 1198-1215 Robert de Tregoz confirmed [116] the grant of half the curtilage of Walter of Wellington near the old ditch, to augment their court. This reference implies that the land lies next to where the priory nucleus was (through the use of the term “court”). This was confirmed (1205-15) by Walter (Son of William of Wellington) as being half his curtilage near the barn of the monks, near the old moat, which his father had given to them [113]. In 1206 there is a grant [129] from Walter of Ewyas of 1 acre outside the monks’ court towards Hay this plot being further described as the land near the water. Also in the early 13th century [130] there was a grant from William, son of Walter, of five plots of land (the fifth being confirmation of the former grant:

  1. an acre of arable near the Church of St Michael which extended “as far as the Dulas against the hill”, this latter comment perhaps indicating it followed the lower contour of the hill on this side of the Dulas Brook.
  2. an acre above “Stamhurstam” and below Martin Strigg’s land and the site of the spring called Wndewella.
  3. a plot between land owned by Robert Croc (William Croc’s ancestor?) and Angus Daparel extending as far as his own land.
  4. a perch of land between the Dulas and his furlong near St Michael’s
  5. and an acre of land which his parents had granted to make a yard for their home which lay in his furlong above the Dulas outside the monks’ court.

William also confirmed an acre of land at St Michael’s granted by his brother Roger [131]. Roger confirms an acre next to the acre his father (Walter?) had given to the church. Basilia (Walter’s widow) confirmed his grants and provides useful further topographic descriptions [133]:

1.          “one acre in the furlong outside the monks’ court, lying across her furlong next to the water which falls to the lord’s mill in Ewyas.”

2.          “two acres: one lay in the north part of St Michael’s, with the curtilage which Walter son of Rhydderch (Righered ) held, above the Dulas over against Maes-coed hill (Mescoit ): Walter of Ewyas gave this acre to the monks with her consent to endow the church of St Michael in its dedication (ad dotandam ipsius Sancti Michaelis ecclesiam in dedicacione sua ). The other acre lay below the land which Martin Strigge held above Stamhurstam, which Walter her husband gave to the church of St. James as he lay dying (in extremis suis ), together with the spring called Vndewella.”


Between 1215-44 Robert of Tregoz confirmed the grant of a ditch between the land of Walter of Wellington and the acre called Dulas. It (the ditch?) extended from “the land of Hidiard” to the acre called Dulas near the monk’s barn [117].


There are many descriptive grants from 1265-1300 and they have been divided below into land on either side of the Dulas Brook, again trying to maintain a chronological order.


Descriptions of land on the west (Castle) side of the brook

John de Tregoz confirmed the grant [121] of “an acre of arable land which lies between King Street towards Llangua (Languen) of the left-hand side, and the land which belonged to John Grebin; the larger headland of the acre extends towards Bradelee, that is to the demesne of the lord of the ville, and the smaller extends towards the meadow of Alexander.”



That land on the east side of the brook …

John de Tregoz [119] gave land that once belonged to Basilia, adjacent to those plots held from William and Walter of Wellington, and which extended “9½ perches from the headland to the stream called Dulas, and is 2 perches 2 feet wide”. He also gave a footpath extending from Frog Lane to the cemetery “to augment their court, with the water running there, to be enclosed” [120]. John de Tregoz gave a road [123] which led from his garden (which he held from his uncle) to the road called Frog Street (Vriogistrete) “the upper part lies between the curtilage of the monks, enclosed on both sides, and the lower part lies between the monk’s garden and the curtilage which William Croc once held”. Thomas (son of Henry the Weaver) gave up his claim to a plot of land near the Lord’s garden and land owned by Basilia le Gant [127]. In 1266 Henry (the weaver) was threatened with excommunication if he did not remove his house, which presumably sat in the land proposed for the cemetery of St Michael’s - the bounds of the cemetery being the road to Abbey Dore and the road to Ewyas, the Dulas Brook and the old course of the Dulas Brook. This indicates that the original line of the road to Abbey Dore must run along the side of the cemetery as should another road leading into the village (part of which may still survive running past the post office). There were also 2 acres of arable land formerly owned by Richard of Ewyas towards St Michael’s church and between land owned by himself and his son and extending from land owned by Robert Loue to that of Hugh the miller.


John de Tregoz issued a special licence to Richard le Norman (Prior in 1283) to buy William Croc’s burgage plot that lay between the monk’s curtilage and the road leading to Abbey Dore, and extended from the monk’s garden to the wood called Hay [124]. There is another grant of similar description [125], but as the priory had to buy the first plot and this one was granted it is not unreasonable to assume they may be two different parcels of land. In this case the plot, again formerly owned by William Croc, is bounded by the monks’ curtilage, the road from Ewyas to the lord’s wood called Hay and the monks’ garden. In 1305 there is a quitclaim [126] for a third of the plot of land referred to in no. 125. From this we know that William Croc originally held the land from Richard of Ewyas.


Many of these above descriptions appear to refer to land on the east side of the Dulas. So by the early 13th century, on the other side of the brook from the castle there appears to be St Michael’s Church, the priory court, the Lords Mill and the monks’ barn.


Some thoughts on the site of the priory

The map evidence and local tradition place the priory in the field behind the modern-day butcher’s shop at the foot of the outer enclosure attached to the castle. The resistivity survey has identified odd structures in this location, but only a few, and no clearly interpretable buildings. It is known that the priory was located near a ditch/moat and fishponds when it was originally moved to Ewyas Harold after 1120. Earlier scholars may have naturally assumed that these referred to the ones adjacent to Prill Lane. A field on the other side of the Dulas from the castle was called Piscollony (Sprackling 1988, 23); also on a modern air photo of this field (where the cricket ground now stands) there is a dark area in the grass indicating that a less well draining feature may once have stood here (Plate 1). Could Piscollony derive from piscina colina, fishpond hill? Also, as more land can be identified in the charters on the other side of the Dulas, alongside references to the monks’ court, then an alternative possibility needs to be considered. A key charter in this respect is that of Basilia (Walter’s widow) who confirmed Walter’s grants one of which was [133] “one acre in the furlong outside the monks’ court, lying across her furlong next to the water which falls to the lord’s mill in Ewyas.”


The site of the Lords Mill is well attested (Fig. 9). It lies north of the church and the line of its leat is even now well preserved in hedge boundaries. It is likely that this was the later mill, probably built c. 1200. We know that the gate to the priory was by the castle mill, and the reference for this comes from a charter that post dates 1300. Could the castle mill and the lord’s mill be the same? Whilst it is appreciated that William Croc may have held land in a number of locations around Ewyas Harold, many of the references to grants to the priory seem to refer to an area of land delimited by the Lords Wood, the Dulas, St Michael’s and the road to Abbey Dore, some of which includes his land.


The most difficult part of the charter to fit in relates to the references to Frog Lane. This is commonly believed to have been the road to Dulas based on the fact that the road has no specific name now. What we do know is that Frog Lane is differentiated from King Street and the road to Hay (the lord’s wood). It must also be considered that it might refer to a road that no longer exists. What do we know about Frog Lane? The priory garden may have a frontage on to it. A path leading from the Lords Garden (below the postern gate) divides in two with one branch passing first between two plots owned by the priory and then between the curtilages of the priory (garden) and William Croc before joining Frog “Street” (Lane). The branch not going to Frog Lane, crosses the Dulas to the cemetery. There was also a path from Frog Lane to the cemetery with water running near it (given to the priory to “augment their court”). One possible configuration of these elements and the postulated areas into which the priory expanded is given on figure 31.


None of the above explains why the Ordnance Survey mapped the lower castle field as being the site of the priory. There are two reasons why it may have been mapped here: local tradition or a find spot. Buildings from the mid part of the 14th century onwards could have reused parts of the suppressed priory, so it may be possible for ecclesiastical masonry to be found throughout the castle holdings and thus misdirect antiquarians. The ordnance survey record cards refer to earthworks in the field behind the butchers (15th July 1952 – area seems to be thick with foundations under grass; 30th Aug 1972 – observed a disturbed surface but no building material was apparent). However, the site was marked on the 1st edition OS map of 1888 so these visits were merely to confirm the original identification of the site.


Without further fieldwork it is not possible to confirm or refute the postulated site of the priory court given above. However, it does seem that it expanded across the Dulas, which at that time may have formed three sides of what might have been referred to as a moat on the west side of the brook when it was founded. The description of the extent of the cemetery for the church alludes to an earlier course of the Dulas. There is potential here for the utilisation of natural features for fish ponds. A channel still runs near the postulated line on the tithe map.


8.3 The later castle


It is not possible to say how much, if any, of William fitz Osbern’s castle was made of stone. However, there is clearly a stone keep there now; as well as stone buildings within, and evidence for a stone curtain wall around, the bailey. Whilst we do not know when the early settlement moved from the outer bailey (“village/priory” fields) it is likely that this site would have been needed for housing knights on service, if not during Harold’s lordship, then certainly during Robert’s.


The inner bailey demonstrates a degree of sophistication that might be expected in late 12th -13th century castles. There is a large building surrounded by an area apparently devoid of stone structures (the radar data confirms this). Outside of this blank area is a whole multitude of small structures exhibiting at least three separate phases of construction. It is proposed that the large building at the foot of the motte was the lord’s residence, possibly separated from other buildings in the bailey by an extensive court. The chapel of St Nicholas probably lay at the east end of this building. There also appears to be an exclusive access to the steps to the keep on top of the motte from the rear of this building. It is not possible to determine at present if this latter access was a main through way with the building acting as a lower gate-house or a private access for the lord of the castle. Whatever the case there was clearly a high status building at this location.


There appear to be three significant entrances to the castle. At the north, the gap in the bank containing the road to the modern-day bungalow appears to have a large tower on its west side. This also leads to what was reputedly the Kings Gallop and the area of the park which contained Middle and Lower Castle field and is identified as Park Piece on the 1718 map. There is some suggestion from the internal layout of structures that the entrance on the south-west angle of the bailey, whilst probably not original, had been established during the life of the castle. The possibility of a third entrance (requiring a bridge?) is implied by a line of high resistance readings leading across the “village” field to the south of the bailey, as these may be indicative of the line of a road.


8.4 Mills and fishponds


The only mill site that is clearly known is that of the Lord’s Mill near the northwest angle of St Michael’s Church. It is shown on the 1718 map and its position traceable through the tithe and later maps of the area. The mill leat is a clear landscape feature and is fed from the Dulas opposite Weir Close.


Two early 13th century references to the mill are from 1198-1215 which refers to a plot near the water of the mill by the old moat and also in 1213-14, an exchange of land for where the new mill now stands. In 1307 John le Warr released his tenants from the duty of repairing the stank of the mill which stood near the gate of the priory. One thought was that this referred to the dam for the ponds along Prill Lane which may have driven a mill between the last pond and the Dulas. However, this reference is nearly a century after the new mill was built and therefore more likely refers to the lord’s mill - perhaps the stank was the bank that still survives to the north of the mill site.


With respect to fish ponds there are three possible locations for these. The first is the known site on the north side of Prill Lane where there appear to have been two ponds (although these may have also served to run a mill at an earlier date). The second and third lie in the field that Sprackling states was called Piscollony. A large dark area on an aerial photograph may be one such feature. Also there is reference to an earlier line of the Dulas that bounded part of the cemetery, and references to an “old moat”. These latter references may relate to a meander of the Dulas which once extended to the site of the Lord’s mill and again could have survived to sufficient extent to be used as fish ponds. It may be these latter features that were referred to in the cartulary relating to the siting of the priory at Ewyas Harold.


Any further study lies outside the scope of this report.

9.         Recommendations and potential


As a result of the work undertaken here there are clearly areas of the site that warrant further research and investigation. Future surveys need to target land adjacent to the Dulas Brook near the church to see whether there is any evidence for structures in this area.


Within the surveyed area there is clearly evidence for a good degree of survival of structures. The lower bailey would benefit from some small scale excavation to establish whether there was a road leading to the castle through this area. Also one of the potential sites for a dwelling in this area would warrant investigation to try and establish the length of occupation of the site. Any work undertaken within the area should establish the date of abandonment for the castle.


Perhaps the greatest potential for the site given its historic significance and the level of preservation is its interpretation and presentation to visitors and students. This would, however, require a better means of accessing the motte than at present and the collection of more information about some of the surrounding features to assist in the interpretation of the site.





It is clear from the work undertaken that the site played a very significant role in the development of the County of Herefordshire and the country as a whole. Because it was abandoned before the end of the medieval period there is an opportunity here to establish the status of the early Marcher Lords and a site that reputedly served the function of a “Royal” castle. There is also the possibility of investigating the nature of the livelihoods of Knights in attendance whilst serving their Knights Fees at the castle because of the possibility that they occupied the outer bailey site after the village was moved.


The project has achieved its objectives and identified a range of follow up exercises.



11.       Bibliography


Aitkin, M., 1995, “Ewyas Harold Castle”, Herefordshire Archaeological News 63 , 9 (plan) MS


Bannister. Rev. A.T., 1902, The History of Ewias Harold, its castle, priory and church , Jakeman and Carver, Hereford


Benn, C.A., 1941, “Castles mentioned in the Pipe Rolls under Herefordshire” Transactions of the Woolhope Naturalists Club1 , 132-3


Boucher, A., 2006, Ewyas Harold Castle Project: Archaeological Proposal MS


Bull, H.G., 1869, “Ewyas Harold its name, its castle and its priory”, Transactions of the Woolhope Naturalists Club 1869


Buteux, V., 1996, The historic Marches Towns Survey: Ewyas Harold , Hereford and Worcester Archaeology Service, MS


Clarke, G.T., 1877, “The Castle of Ewyas Harold”, Archaeologia Cambrensis


Copleston-Crow, B., 1986, “The Fief of Alfred of Marlborough in Herefordshire in 1086 and its Descent in the Norman Period” Transactions of the Woolhope Naturalists Club Vol. XLV part 2


Copleston-Crow, B., 1992, “The Castle of Ewyas Harold and its Military Arrangements in the Norman Period”, Herefordshire Archaeological News 57 , 7-11 MS


Copleston-Crow, B., 1993, “Medieval Topography of Ewyas Harold”, Herefordshire Archaeological News 58 , 18-20 MS


Dawkes, G., 1999, Ewyas Harold Castle, Herefordshire. Report on a Watching Brief, Marches Archaeology Internal Report


Duncumb, Rev. J. (n.d.) “Ewias Harold” Extract form hand written MS, Hereford Library


Jackson, J.N., 1954, “The Historical Geography of Herefordshire, from Saxon Times to the Act of Union, 1536” Herefordshire: Its Natural History, Archaeology and History: Woolhope Naturalists Field Club, S.R. Publishers Ltd Gloucester


Marshall, G., 1940, “The Norman occupation of the lands in the Golden Valley, Ewyas and Clifford and their motte and bailey castles”, Trans. Woolhope Nat. field Club Volume for 1936, 1937 and 1938 part III, 141- 158


Morris, J., 1983, Domesday Book: 17- Herefordshire , Phillmore


Page, W. [ed], 1908, The Victoria County History of Herefordshire, Dawson, 237-8


Philips, N., 2002, “Ewyas Harold” from an unpublished study of local castles MS


Robinson, Rev. C.J., 1869, The castles of Herefordshire and their Lords , 58-61


Salter, M, 1989, The Castles of Herefordshire and Worcestershire, 2-9 and 20


RCHM 1, 1931, An Inventory of the Historical Monuments in Herefordshire: Vol. 1 South West , 58 and 62-64


Richardson, R.E., and Sprackling, G., 1987, “A survey of Herefordshire Field-Names” Trans. Woolhope Nat. field Club Volume XLV pt 3, 754-6


Robinson, C.J., 2001, The Mansions and Manors of Herefordshire and their Memories, Logaston (Reprint originally published 1872)


Shoesmith, R., 1996, Castles and Moated Sites of Herefordshire, Logaston, Herefordshire, 104-106


Sprakling, G. [et al ], n.d., Field and Furrow W.E.A.


Sprakling, G., 1988, “Some notes on the topography and medieval layout of the village of

Ewyas Harold” Herefordshire Archaeological News 48, 19-24


Smith, B., 2004, Herefordshire Maps 1577 to 1800 Logaston Press


Stirling-Brown, R., 1989, Herefordshire Castles: A list of classified sites, 6


Strong, G., 1848, The Heraldry of Herefordshire


Walker, D., 1976, A register of the churches of the Monastery of St Peters, Gloucester: “An Ecclesiastical Miscellany” Bristol and Gloucester Archaeological Society, Records Section, Vol.XI


Wills, J., 1981, Archaeology in South Herefordshire District – The Rural West, Hereford and Worcester County Museum, MS


Wills, J., 1983, “Ewyas Harold, Hereford and Worcester” West Midlands Archaeology 26, CBA, 92



12.       Glossary



The king’s tenants in capite who sub let land to others, keeping a portion for themselves in demesne. Some was let to knights in “knight’s fees” for 40 days military service a year; some to free tenants, socmen or yeomen – paying in service or in kind, or in money; some to villains and “borderers”.



These were cottagers holding between 1 and 10 acres. In return for this they would work on the home farm a number of days in the year.



The register book of a monastery. This would keep records of grants/gifts to a priory, or deeds or covenants etc...



Otherwise known as a “hide” was as much land as could be cultivated by one plough and generally to be 120 acres. It should be enough to support one tenement


Clas churches

Early type of collegiate ecclesiastical foundation a bit like the later minsters. They can date from as early as the 6th – 7th century.



Home farm of the lord.


Farmer (Medieval)

A villain who held his land on condition of being bound to provide feorm (food and entertainment) to the non-resident lord when he visited the manor.



Derived from furrow long was traditionally the amount that a team of oxen could plough with out coming to rest. Medieval fields were divided into strips one furlong in length.



Land let to a man in return for 40 days military service.


Manor lands

These comprise the demesne (home farm), lands held by free tenants and land held by villains or borderers.


Perch = 5.5yards



Most important class of unfree tenants (equivalent to Saxon coerds) holding about 30 acres of land. In return for this they would work on the home farm a number of days in the year.



Appendix 1: Sources by reference












OS finance map






Finance book






6” OS map






Castle guide






Notes and drawings on Castles – none for Ewyas Harold






Papists estates survey






Highway alterations – no turnpike – not adopted



Pamph 83



Same as AB61/box12



Pamph 204



“Yesterday in Ewyas Harold”



Pamph 205



WEA local farm study probably Sprackling et al (n.d.)






Court books for Ewyas Harold/Lacy – mainly the latter






Compotus 14th -16th century







Tithe map 4’10” x 6’2”

Award and survey are 1845







Sale of castle meadow






King St and Old Wigga Farms – sales documents – too recent to be of much use






King Street – too recent to be of much use






Building and pasture land






Statutes of Ewyas Harold






John ap Henry bailiff for Ewyas Harold






Title of Lord Abergavenny 1555-1919






Parish council minutes for 1894-1925 (not deemed relevant)






Sketches 1874-7 (nothing of interest)






Right Hon. George Abergavenny – lease dated 29/1/1761






Papist records






Refers to Lord Abergavenny’s title to his heirs.



2820 (1)|



An Exact Map of the Estate adjoining to the Castle yard, lying in ye P(ar)ish of Ewas Herald in the County of Hereford, belonging to ye Right Honoble George lord Bergevenny... Surveyed & drawn July 8th 1718, by Benjamin Fallowes, of Maldon in Essex. 2 ½ chains to an inch (not to scale).



Watkins 322


c. 1900

Photograph of the castle






Oblique views of the castle looking north – shows parch marks of buildings in bailey.





4169-71, 5169-71



1:10000 – not much detail discernable – covers wide area



OS/70398 Exp. 276



Just shows Ewyas Harold common




Exp 34-5, 100



Shows the track across the foot of the motte, and remnants of trees at end of village field as shown on 1920 OS map.






Copy of the schedule. Originally scheduled Nov 1928, amended April 1974. Amendments seem to be mainly to pull monument in line with local government changes in 1974/1998.






Record relating to Ewyas Harold Castle



Source 6026



OS record card – site visit identified uneven ground



Source 11694



OS record card – site visit by A. Rive - 15th July 1952 – area seems to be thick with foundations under grass – referring to priory?



Source 15505



(Dawkes 1999) Watching brief on steps found pottery of 16th /17th century date near top.






Callow Hill – possible civil war earthworks in wood






Blood Field. Civil War Battle Field?



Source 9328

Event 21497



Field walking record identified medieval pottery and animal bone in the area of the priory at Ewyas Harold (outer bailey). SO 3863 2864 – original file not found



Event 21498



WB on water pipe in lower bailey




Sites and monuments records for Ewyas Harold



Pound, S of Church, Ewyas Harold

SO  3878  2870

Post Medieval / Pound


Possible pound, NW of Alma Cottage, Ewyas Harold

SO  3891  2864

Undated / Pound


Benedictine Priory. Ewyas Harold – site marked on ordnance survey maps.

SO  3800  2800

Medieval / Priory
Medieval / Pottery


Churchyard Cross, Ewyas Harold 14th – 15th century

SO  3875  2873

Medieval / Cross-churchyard


Ewyas Harold Castle

SO  3850  2870

Early Medieval / Castle
/ Motte & bailey


Development, Ewyas Harold

SO  3880  2870

medieval / Occupation


Bank, E of Castle, Ewyas H – contained pottery of

SO  3874  2880

Medieval / Defences


Lords Mill

SO  3872  2871

Medieval / MILL


Enclosure, NE of Castle, Ewyas Harold

SO  3800  2800

Undated / Enclosure


Limekiln, Mount Hill, Ewyas Harold

SO  3982  2888

Post Medieval / Kiln-lime


Masonry, in river, Ewyas Harold

SO  3872  2867

Medieval / Stone-carved


Monks graveyard

SO 3880 2879

Medieval / graveyard





BL – British Library

HL – Hereford Library

M – Manuscript

NMR – national Monuments Record, Swindon

PD – Primary document

RO – Hereford Record Office

SD – Secondary document

SMR – Sites and Monuments Record

Var – various formats of information




Appendix 2: Dulas Court capitals (Malcom Thurlby)


The capital with the centrally placed human heads flanked by angle volutes belongs with a type established type in Normandy from the 1030s and continuing into the early 12th century. They appear at Rouen Cathedral – see Maylis Baylé, Les origines et les premiers developpements de la sculpture romane en Normadie: Art de Basse-Normandie , no. 100. Caen (1991), figs 410, 412; see also, examples at Thaon (near Caen) (figs 403 and 406, and at Cerisy-la-Forêt, fig. 534.. Cerisy is probably 1080-1120. For English examples, see that nave arcade at Blyth priory: Peter Coffman and Malcolm Thurlby, ‘Blyth Priory: A Romanesque Church in Nottinghamshire’, Transactions of the Thoroton Society , 105 (2001), 57-72. This is after 1088 but earlier examples can be found in Durham Castle Chapel c. 1072, albeit with chip-carved crosses on the surfaces between the head and the volutes.


To return to the Rouen Cathedral and Thaon examples, the heads and volutes form an upper register of the capital set above a lower register of stylized upright leaves. In the Dulas Court capitals scallops occupy this lower level. I do not know of a parallel for this – at least so far – but it may be significant that in the crypt of Lastingham priory (Yorks.) (1078-85) there is a capital with angle volutes and a row of intersecting arches in the lower half of the capital. Also in the Lastingham crypt there are cushion capitals with volutes, a form that also appears in the nave clerestory at Blyth priory.


Cushion and scalloped capitals – one of the ongoing debates in English Romanesque is whether or not cushions and scallops appear before the Conquest or not. I believe that they do – see:

Malcolm Thurlby, ‘Anglo-Saxon Architecture beyond the Millennium: Its Continuity in Norman Building’, in Nigel Hiscock (ed.), The White Mantle of Churches: Architecture, Liturgy and Art Around the Millennium (Turnhout [Belgium], 2003), 119-137; idem, Romanesque Architecture and Sculpture in Wales (Almeley [Herefs.]: Logaston Press, 2006), 42-45.


It looks as though the scallops have raised shields.


Arch mouldings and jambs: Looks like a soffit in the arch and then a couple of angle fillets? This seems unlikely much before 1100.






The project would like to thank all those individuals who assisted in the collection of data for the project. In particular Margaret and Peter Woodward for their hospitality and assistance (in all kinds of weather), Graham Sprackling and Roger Sterling-Brown for their help and advice, Sue Hubbard for Latin translations, and the stalwart geophysics crew, Roger and Ann Bradley. Our thanks are also extended to the Hereford Record Office, Hereford Reference Library, British Library and Sites and Monuments Record for their help and assistance. The Bartlett-Clark Consultancy for the loan of equipment and assistance with elements of the geophysical survey. The children of Ewyas Harold Primary School and their teachers for assistance with the magnetic susceptibility survey of the bailey. Dr Keith Ray MBE for his support of the project. We would also like to thank the MoD for taking air photos during the project.




Staff of Archaeological Investigations Ltd:


Report:                        Andy Boucher


Illustrations:                Simon Mayes, Sam Porter, Andy Boucher


Editing:                       Niall Oakey





Site reconnaissance:    Simon Mayes


Topographic survey:   Simon Mayes, Dale Rouse


Geophysical survey:   Andy Boucher, Luke Craddock-Bennett, Dale Rouse, Simon Mayes, Catherine Rees


Documentary work:    Andy Boucher, Catherine Rees


Top - Back

Ref: rs_ewh_0007