Medieval Llangwm in Context

By Dai Stephens

Llangwm is a village proud of its history. Long regarded in Pembrokeshire as idiosyncratic and somehow different from its neighbours, it has been the source of many romantic stories, not all of them strictly factual. The natives of Llangwm have told and retold these stories, and lately they have been taken up by more recent residents to promote the village, and, more especially, to obtain funding for much needed renewal projects such as the church. Sometimes, the desire for an interesting story has gone beyond what can be established historically. This essay is an attempt by an amateur historian, to delve into the records and works by professional historians, to discover what is truly documented about Llangwm, and to put the village history in the context of the wider history of Pembrokeshire, of Wales, and of even British and European events.

The Stephens family has a Llangwm association to compete with any of the still existing old Llangwm families, but our history can be traced back only until the mid-eighteenth century (though there was a William Stephens paying a hearth tax in Freystrop in 1670). In historical terms, we, like most others, are recent arrivals, our settlement occurring at a time when many were moving to the Welsh coalfields, perhaps to work as miners, perhaps as mine technicians or managers, perhaps to establish trades associated with mining, perhaps to contribute to the infrastructure on which such communities depended. In the intervening years, the Stephens have intermarried with the Jones, Palmers and, no doubt, other local families, but it is the nature of genealogies that the female side of the ancestry is often lost when the woman adopts the name of her husband. No matter. We all come from somewhere else, originally. Nevertheless, the history of the Llangwm community is of sufficient interest to have become something of an obsession. This is my attempt to bring some evidence and context to Llangwm’s history.

In the beginning…

Although Pembrokeshire has been inhabited from very early times, the Llangwm area carries sparse evidence of human activity. A fairly recent discovery of seventeen hundred and fourteen flint tools, including arrowheads, awls, and a saw in Llangwm Ferry, dates from the early Mesolithic era, perhaps 8000-6500 BC. The people who made these tools were hunter-gatherers, who probably camped briefly in Llangwm before moving on. More recent versions of flint tools dating from 6500 – 4000BC suggest that the site was also visited on later occasions.1 Another collection of Mesolithic flints was unearthed at Nash during the Llangwm Heritage Project’s dig in 2016. Nearby is a Neolithic cromlech, and a hanging stone. But, otherwise, evidence of human activity in and around Llangwm is sadly missing for much of prehistory.

If Arthur, in his pursuit of the Irish Twrch Trwyth, a magic boar and his seven piglets, along the river Cleddau to the first skirmish at Canaston Bridge, passed Llangwm on his way, the Mabinogion gives no hint that he waved in passing, or even noticed our village. And, while at least part of Pembrokeshire was ruled for lengthy periods during the period 400-800 by Irish kings, the extent to which their influence penetrated to the area we call Llangwm is unknown.  Although Danes from Dublin and Waterford entered the Haven, sometimes, like the twenty-three ships of 878 overwintering under way to Devon, sometimes as mercenaries employed by warring Welsh princes, perhaps as raiders, stories of Viking settlement in Llangwm appear to have no evidential basis, and certainly no archaeological finds that support them. Even the existence of nearby Scandinavian place names, such as Freystrop, may indicate later settlements that occurred at the time of Henry I’s plantation of Rhos in the twelfth century with Flemings and English, some of them with Scandinavian ancestry.2 Nor has any Welsh settlement left any record, apart from two Celtic grave slabs from the tenth or eleventh centuries. These suggest that a place of worship of some sort existed on the site of the present church prior to Norman medieval times, which would be consistent with the existence of a llan in the valley that the name Llangwm indicates. But no records exist till Llangwm emerges from its anonymous past in the late thirteenth century.

Medieval Llangwm

It is a common view, based on W. Grenville Thomas’ useful little book Llangwm through the Ages,3 that the first documentary evidence for Llangwm occurs in 1244. This account itself appears to have been based upon a claim in Henry Owen’s Old Pembroke Families. 4 In that year, David de Rupe, lord of “Landegunnie and Maynclochauc”, is said to have made a grant to the abbot and convent of Albadomus [Whitland] of common of pasture over all his land of Pressely for seven years from 1303, and remission of 2s. annual rent minus one penny.5 This text, from the British Museum archives, written in Latin, is commonly cited as the first known written version of the village name. “de Rupe” is the Latinised form of the Norman French “de la Roche”, and thus also apparently provides the first evidence of an association between the de la Roche family and the village. However, there is something curious about the dates: Why would David de la Roche make an agreement with the Abbot in 1244, which only came into force some 60 years later, thus potentially compromising the interests of his descendants? The answer is simple. He didn’t. A search of the British Museum catalogue6 reveals that 1244 refers not to the date of the transaction, but to the paragraph number documenting the entry. However, in an attempt to rationalise the error, Grenville Thomas, following Henry Owen, then cites the same source for a “renewal” of the lease in 1303, ostensibly by another (but presumably actually the same) David de la Roche. Most likely, the first David did not exist.

Interestingly, the British Museum catalogue which documents this transaction provides an index entry for David de Rupe:  “Llandegwning (co. Pemb.), lord of, 1244” similarly referring to the entry number, not the date.    However, there does not seem to be any independent evidence that Landegunnie is actually Llangwm, except that Landegunnie is associated with the de la Roche family, and the de la Roche family is associated with Llangwm.


The Hundred of Rhos was initially settled by the Flemings in the early years of the twelfth century, nearly 200 years before David’s agreement with the Abbot, so why is there no mention of Llangwm until this date?  In this context, it is worth asking what was going on at that time in Pembrokeshire.

Eleventh and Twelfth Centuries

As every schoolchild knows, the Normans invaded England in 1066. At that time, Wales had recently been united under Grufudd ap Llwelyn into a single kingdom (1057 – 1063). The Normans had enough on their plates in subjecting England to their rule, and largely ignored Wales, sealing it off by the establishment of military strongholds along the border (March) under the rule of essentially autonomous barons. Nevertheless, the Normans were recruited by Maredudd ap Owain, King of Deheubarth (of which Pembrokeshire was part), in his campaign against Caradog ap Gruffudd of Morgannwg in 1072. William the Conqueror made an ostensibly purely religious pilgrimage to St Davids in 1081 (though, most likely it was a demonstration to the Welsh of his military power, intended to keep them from interfering in affairs in England). William also used the opportunity to negotiate with Maredudd’s grandson, Rhys ap Tewdyr, now King of Deheubarth, offering support for his rule in Deheubarth, on condition of fealty and payment of tribute.

By 1093 this alliance had broken down, and in that year, Rhys ap Tewdyr was defeated at Brecon, opening the way for a Norman invasion of South Wales. In the same year, Arnulf de Montgomery took Pembroke from the Haven, allowing his father, Roger de Montgomery, to establish a fort, described by the twelfth–thirteenth-century Itinerarium Kambriæ as a “slender fortress built of stakes and turf”. Roger died in 1094 and Gerald de Windsor took over administration of the castle, and was subsequently made steward by the Conqueror’s son, Henry I.  Gerald married Nest, daughter of Rhys ap Tewdyr (and previously lover of Henry prior to his accession to the throne), beginning a tradition of the invaders legitimising their presence by intermarrying with the Welsh aristocracy. In 1100, the Norman, Gilbert de Clare, established a castle in Haverford. It is during this same period that Godebert the Fleming, ancestor of the de la Roche clan, is born (1096), supposedly in Rhos, but more plausibly in Pembroke or even Flanders.

Although the Normans had established strongholds at Pembroke and Haverford, the Welsh had not been sufficiently subjugated, and Henry I boosted his efforts to tame them by bringing in settlers from outside Wales to South Pembrokeshire. In 1108 colonisation was extended across the Haven by the transfer of Flemish settlers into Rhos, possibly under the leadership of William de Brabant.  The first wave of Flemings probably came to Pembrokeshire directly from Flanders, but a subsequent (1155) group came from Northumbria, where they had initially been planted.7 The Flemings, unwanted and troublesome in Northumbria, and disliked throughout England, were attracted by the prospect of land, light rents or services and perhaps a temporary exemption of certain tithes. However, the Flemings were not alone, and Henry I also planted English settlers, mainly from Devon and Somerset who, by the turn of the twelfth century outnumbered both the Flemings and the remaining native Welsh. Indeed, so numerous were the English colonists that by the early fourteenth century, much of south Pembrokeshire was predominantly English, the Welsh, Flemings and Normans having been fully absorbed. The influence of this settlement was vastly greater than that of the Flemings both in language and farming methods. 8

Although the Welsh communities were for the most part evicted and their settlements appropriated, including fertile land belonging to the Bishop of St. Davids in Llanstadwell and St. Ishmael’s, the Bishop’s extensive estates in Burton were unaffected.9 Remember, parts of present day Llangwm, including Guildford and Llangwm Ferry were previously part of Burton, and not incorporated into Llangwm until the 1950s. If parish boundaries were based on pre-existing lordships, parts of present-day Llangwm may well have been within the Bishop’s remit. Indeed, it is remarkable that although Flemish/Norman settlements such as Burton, Llanstadwell and St Ishmael’s are documented, Llangwm, or any precursor name, does not enter the record at this time. Might it be that nothing resembling a village existed where Llangwm now exists, and that the entire area formed part of Burton, or even Rosemarket or Freystrop administrative districts?

How the land was organised amongst the planters is unclear. According to Joyce, the area settled by the Flemings was divided into various lordships, but there are no early records concerning their number and extent.10 Most of Pembrokeshire of the time would have been uncultivated woodland and heath. Joyce suggests that a preliminary scouting of the province was made in order to evaluate its size, topography and resources such as Welsh settlements, arable or pastoral farming, meadow, woodland, waste, streams, lakes, marsh and fisheries, etc. before it was divided into various lordships.  By the late twelfth century, the area under Norman/Flemish control on the northern side of the river was made up of three (or perhaps four) baronies, under the leadership of William de Brabant.

The land belonging to each barony was scattered, possibly to allow each to hold both fertile and less fertile land, to give each one access to the coast, and to allow each barony to take responsibility for defence on the borders with the Welsh areas, with the establishment of forts.  The centres of each barony were decided strategically; Roch lay near the coast and commanded extensive views across the northern district, Haverford controlled the lowest ford of the Western Cleddau river, while Walwyn’s Castle dominated the south western peninsula. By the early twelfth century the three baronies were held by Godebert, Tancred and Walwyn, respectively.

On the basis of documents outlining the military commitments of these lordships to their Pembroke overlord, Joyce suggests a fourth lordship that included Rosemarket, which was commercially important at the time and for many years to come. Twelfth century Rosemarket was held jointly by the barons, presumably as the joint commercial centre of Flemish Rhos. By the thirteenth century the neighbouring parishes of Johnston, Llanstadwell and part of Burton belonged to the baronies of Roch, Haverford and Walwyn’s Castle respectively, but it is possible that they were previously associated with Rosemarket, jointly providing military personnel and access to the Haven and Western Cleddau for the three main lordships.

In keeping with the requirement that each barony should be responsible for holding a position on the borders with the Welsh, the defensive duties of the fourth lordship may have been based on Camrose, where a motte and bailey fort lies between Roch and Haverford. Camrose parish is the largest in the province. Although land locked, it may have gained access to the sea by possessing Lambston and Haroldston West, which were held subsequently in the fourteenth century by Roch and Walwyn’s Castle respectively. Joyce suggests that the baron of Camrose, as well as leader of the whole settlement north of the river, was William de Brabant.

According to this account, then, the northern section was originally split between the lordships of Roch and Camrose, while the eastern part, running along the Western Cleddau to Llangwm, was allocated to Camrose and Haverford. The vulnerable south-western peninsula was protected by Walwyn’s Castle, while responsibility for the coastline from Newgale to St. Ann’s Head and up the river to Burton was defended turn and turnabout by the four lordships.

The dates at which these divisions were established is not clear. Godebert, the baron at Roch, was born about 1096. In 1108, at the time of the first Flemish settlements north of the Haven, he would thus have been only 12 years of age, so we can assume that the baronies were formed later than this date (or that someone else first held the Roch barony). The colonisation was disrupted in 1110 when Prince Owain ap Cadwgan of Powys raided Dyfed, supported by mercenary ships from the Norse towns of Ireland. After Owain’s forces ambushed and killed William de Brabant Henry I removed Cadwgan, Owain’s father and king of Powys, from the lordship of Ceredigion and awarded it to Gilbert fitz Richard de Clare who established forts throughout the region. This reduced the danger to the south and immigration continued, with the lordship of Camrose possibly being granted to William de Brabant’s brother, and Daugleddau newly colonised by Wizo, recently arrived from Flanders.

According to Llangwm Heritage,11 Godebert settled in Llangwm at nearby Great Nash Farm in 1120. I can find no other source for this claim, and it seems questionable since Godebert’s responsibilities lay in Roch, and the area from Haverford down to Llangwm was within the Haverford barony, and thus Tancred’s fiefdom. By 1130, the Camrose barony seems to have been taken into royal administration, with Godebert paying ‘for the land which Lambert Echeners holds’, perhaps the principal estate in Lambston parish. Thus Godebert was extending his influence from Roch towards the south. According to the Llangwm History Society, Godebert died in 1131.

The Welsh fight back

Henry I died in 1135 and the throne went to his nephew, Stephen, not, as promised, to Matilda, Henry’s daughter. The resulting civil war distracted the English crown from affairs in Wales and in January 1136, the Welsh defeated the Normans in the Gower, some 500 men dying on both sides. Although the settlement of the Flemings was supposed to have pacified Pembrokeshire, in October that year Prince Owain Gwynedd and Gruffydd ap Rhys (Tarw) destroyed a 3,000 strong Norman/Flemish army assembled from across south Wales at Crug Mawr near Cardigan. The monk, John of Worcester, continuing the Chronicle undertaken by the better-known monk-historian Florence of Worcester (d.1118), and writing contemporaneously, even suggests that over 10,000 died, partly as the result of a bridge collapsing as the Norman army and its followers fled.12 As a result, Deheubarth (i.e. southwest Wales) was incorporated into Prince Owain’s Gwynedd. During 1137 the now defenceless Rhos was devastated and Norman/Flemish influence never fully recovered in west Wales (though Pembroke and some other castles were not taken).  John of Worcester writes for the year 1137,

“The Welsh, having suffered much in defence of their native land, not only from the powerful Normans, but also from the Flemings, after numbers had fallen on both sides, at length subdued the Flemings, and did not cease to commit devastation on all sides, plundering and burning the vills and the castles, and putting to death all who made any resistance, and the helpless as well as the armed.”

A consequence was the reorganisation of the remains of the settlement. In 1138 the Earldom of Pembroke was established to secure control of Pembroke, and the first Earl, Gilbert de Clare (Strongbow), granted palatinate powers. This was bolstered by a new settlement in 1155 as Henry II banished all Flemings living in England to west Wales. Camrose (and its fees) was divided among the lords of Roch, Walwyn’s Castle and Haverford. The fertile parishes around Rosemarket were shared; Roch received Johnston; Burton (excluding the Bishop’s land) went to Walwyn’s Castle, while Haverford received Llanstadwell. The rest of Rosemarket (village, church, market, etc.) were held jointly, while the remaining lands were divided among the three lordships.

Impact of Events on Llangwm

Whether these events regarding Rosemarket had consequences for Llangwm is not known. According to CADW, Llangwm was a medieval mesne lordship (i.e.the lord had tenants, and was himself a tenant to a superior lord), and was a holding of the de Vales until a Roche kinsman, Gilbert de la Roche, acquired it in the late thirteenth century.13 The de Vales, perhaps better associated with Dale Castle (probably originally Vale Castle), but who also owned other manors, including one Morvil, near Clynderwen, were descended from a knight who had accompanied Robert fitz Martin in his invasion of North Pembrokeshire in 1136 (during which the Norman/Flemish army was heavily defeated near Cardigan ). Robert de Vale was central to the administration of the Earldom of Pembroke under the Earl, William de Valence ( Henry III’s half-brother, by their mother, widow of King John). Robert is mentioned in a letter to the King from the Earl requesting that Robert de Vale be excused attendance at legal proceedings in Shrewsbury as he was needed in West Wales carrying out the king’s service. Note, however, that William de Valence did not arrive in England until 1247. He did not assume the Earldom of Pembroke (through marriage to Joan Marshal, one of the five daughters of William Marshal) until after 1250. De Vale’s main allegiance lay directly with the Pembroke earldom, and it is thus interesting that the relationship of Llangwm to the Lordship of Haverford, within which it lay physically, was always a matter of dispute. (A parallel dispute existed between the de Valence family who claimed the right to refer legal cases within their holdings in Roose (Rhos) to the court at Pembroke, and not that at Haverford). It may be significant that when, in 1295, the de Valence family specified places within the lordship of Haverford over which they claimed jurisdiction, these included land in Llangwm. (It seems likely that although the de Valence family claimed jurisdiction over the Llangwm estate, it would have been leased out to one of his liegemen, perhaps de Vale?)

What then, was the reason that the lordship of Llangwm came into the hands of the de la Roche family? According to Henry Owen’s Old Pembroke Families,

“Roch Castle remained the “caput baroniae”. There is ample evidence that the de la Roches of Langum were a branch distinct from the lords of Roch until they were united by the marriage of David and Johanna in 1315; their residence was either at the Castle House at Langum or at a house where now Great Nash stands. They were buried in the de la Roche Chapel in Langum; the family of Roch Castle were buried at Pill”.

Despite some confusion between Castle House and Great Nash this account indicates that the notion of a link between Godebert and Llangwm is speculative.14  It would then follow that Godebert’s sons Rodebert and Robert who joined the invasion of Ireland by Richard de Clare, Earl of Pembroke, in 1167 were also not from Llangwm. Although it is conceivable that one of the brothers now established himself in Llangwm, while the other ruled at Roch, this leaves the part played by the de Vale family unclear.  Rodebert had three sons, Adam, Henry and David. They adopted the surname de la Roche only after Adam had built a castle on the rocky promontory at Roch in the early thirteenth century, perhaps suggesting that the de la Roche association with Nash developed only after Adam had taken the name.  According to the Llangwm History Society, Godebert’s grandson, David, (b.1160) held Llangwm as his fiefdom. However, in contrast, Grenville Thomas holds that the Llangwm branch of the family is traceable to Robert, (David’s brother?), “probably the son of Richard, Rodbert’s brother, who was also granted land at Talbenny by William Marshall, Earl of Pembroke, who died in 1219”. Further confusion arises from documents indicating that the estate came into the hands of Gilbert de la Roche, presumably the (great?) grandson of Richard, and father of another David, from the previous owners, the de Vales, in the late thirteenth century.15

Sir Adam de Rupe/de la Roche and Llangwm Church

Equally confusingly, as early as 1185, it was a Sir Adam de la Roche, who was responsible for creating an endowment, allowing the building (or possibly re-building) of Llangwm church. Why it was Adam, rather than the postulated Llangwm brother, David, who undertook this work is unclear.  Perhaps it is important that the Normans did not rely solely on military might to subdue the Welsh. They also used religion. The lack of structure of the Welsh church did not fit well with Norman rule, while the aims of the Church of Rome were compatible with the Norman style. The Normans thus undertook reorganisation of the church in Wales, including the establishment of parishes. Might it be that the building of a Norman church in Llangwm and the establishment of a parish associated with the Llangwm lordship was one of the routine responsibilities of the baron at Roch, irrespective of who held the local lordship?

Such an account would fit with Adam’s founding of Pill Priory. The charter establishing the Priory acknowledges that the land granted to the Priory was insufficient to finance the building of the monastery, and the decent maintenance of the monastic community. For that reason, Adam also contributed the income of all the churches for which he was responsible, together with the tithes of his mills. Whether Llangwm was one of those churches is not known. The importance of such grants can be seen from documents of 1291, almost a century later, when Pill’s own temporal possessions (i.e. possessions not directly involved with its spiritual duties) were valued at £24 4s. 10d. while four churches (St Cewydd (Steynton), St Mary (Roch), St David (Little Newcastle) and St Nicholas (New Moat)) contributed £38. If Llangwm contributed at all, it would have been very little, as Llangwm had no ecclesiastical income other than that deriving from the church itself, with no income from lands in its possession.16

de Vales and de la Roches

Robert de Vale married twice, first Avelina of Wideworth, and following her death, Margaret. He left four daughters and his estate was divided into four portions as documented in a subsequent 1303 charter of Geoffrey Hascard, relating to land in Johnston which had been rented to him by David de la Roche. Geoffrey Hascard calls on the heirs of Robert de Vale to recognise this agreement, implying that David de la Roche had been acting on de Vale’s behalf. Importantly, from the point of this essay, among the heirs named were Gilbert de la Roche of Llangum, the husband of one of the daughters of Robert de Vale, and the father of David. Thus, it seems likely that the de la Roche family came into the ownership of Nash, through the de Vale daughter. This account, of course, comes into conflict with the apparent association of David de la Roche with Landegunnie already in 1244, in his agreement with the Whitland abbot, and, since Robert de Vale seems to have died in 1298, the transfer of the Llangwm estate to the de la Roche family cannot have occurred as an inheritance. However, if, as already suggested, the 1244 date is an historical error and the actual date of the agreement with the Abbot of Whitland is 1303, these dates fall nicely into place. David de la Roche acquired Llangwm through his wife’s inheritance, as late as 1298, or, as part of her dowry, a little earlier.

Thirteenth Century

One might imagine that a Flemish lord of the manor at Nash, and the (re?)building of the Llangwm church would be good evidence that the Flemish settlement of Rhos had been accomplished and a period of Norman/Fleming-dominated peace ensued below the Landsker. Not so! The Welsh were continuously fighting back, and in 1189 Rhys ap Gruffydd recovered south Pembrokeshire. Whether, and how, this affected Llangwm and its environs is unknown.

Again, a generation later in 1215, Llywelyn the Great came to Dyfed. After taking Narberth and Wiston castles, he attacked Haverford and burnt the town, (though he failed to take the castle) before spending the next five days ravaging Rhos and Dungleddau. With Haverford only 5 miles away, and the important market of Rosemarket just over the hill, it seems unlikely that a settlement in Llangwm could have been immune from such events. Indeed, according to R.F. Walker, Roose was ravaged down to the northern shores of Milford Haven.17 Were the villagers second and third generation Flemish settlers who feared the Welsh, or Welsh serfs who welcomed a chance of rebellion to recover their own property and lifestyle, or, indeed, surviving Welsh peasantry in constant fear of their Flemish neighbours? Or were they innocent bystanders and victims of the ambitions of lords of either side under whose rule they existed? Was the rout by the Welsh army followed by the pillage that was the usual consequence of such an invasion of the time?  Or a time of liberation? Let us remember, for the ordinary people, in economic terms it probably mattered little who was their lord. For the ordinary people, this was a time of raw existence.

The Normans gain the Ascendancy in South Pembrokeshire

In 1245, following the death of Walter and Anselm Marshal without male heirs, the earldom of Pembroke was partitioned. The barony of Walwyn’s Castle was granted to Richard de Clare, earl of Gloucester; Haverford was divided into three parts, one third to Maud and her husband Roger Mortimer, another to Eva and William de Cantilupe, and the last to Eleanor and Humphrey de Bohun, “homage and service of the barony of Roch” being included in Eleanor’s share.  Sometime between 1252 and 1255 the Cantilupes transferred their portion to the Bohuns who thereafter controlled most of Rhos. The Bohuns held Haverford castle until 1265 when it was given to William de Valence (King Henry III’s half-brother), who also received the Bohun portion of the lordship until 1274 during the minority of the Bohun heir.

This account thus leaves us with several puzzles. Of most concern to historians of Llangwm might be the question, “Why do the de la Roch family turn up in Llangwm so late, and in an area ostensibly within the Haverford lordship?” One potential explanation, that Llangwm’s lord was a de Vale, with primary allegiance to Pembroke castle, has already been mentioned.  Another possible solution is that the Nash estate resulted as a consequence of the death of Walter Marshall, and the inheritance of his Haverford lands by his daughters (and, thus, their husbands).  By 1247 nearby Rosemarket was in the hands of the Knights Hospitaler, while neighbouring Burton (presumably including Llangwm Ferry and Guildford) now belonged to the Stackpole estate, following its purchase by Philip de Stackpole (d. 1257) from Peter de Leia, Bishop of St Davids (1176–1198). Part of Burton, though, must still have belonged to the Bishopric of St Davids, as the building of Benton Castle, in Burton parish, is accredited to one of Peter de Leia’s successors, Bishop Thomas Beck, in about 1293.  In 1307, the associated holding comprised the castle itself, valued at 2s yearly, and a handsome 10 carucates of land (more than a thousand acres), held from Guy de Brian, the baronial lord of Walwyn’s Castle, by homage and knight-service. The knight in question is named as Thomas de Roche, Lord of Llangwm, raising yet another puzzle regarding the succession of the Llangwm estate, as other records indicate David as lord in both 1303 and at his marriage in 1315.

Fourteenth Century

The de la Roche family remained associated with Llangwm over the next few hundred years. The marriage of David to his cousin Johanna (from the senior de la Roche lineage) in 1315 was remembered in Llangwm as part of the Llangwm Heritage celebrations in 2016. Clearly, the de la Roches as a family prospered, but it is unclear what this meant for Llangwm.

In 1324, Aymer de Valence (son of William), (whose property included the lordship of Haverford) died. The resulting Inquisitions Post Mortem,18 drawn up to allow proper allocation of the inheritance, lists the fees payable by the Lordship of Haverford. These amount to nine and one tenth knights fees, based on the manors held through the Haverford lordship. Among them, William de Roche is listed as holding Roch within the Haverford barony. However, it is of considerable interest that Llangwm, although physically located within the Lordship of Haverford, is not among the knights fees listed in the 1324 Inquisition, perhaps indicating that it was separate from the Haverford lordship. Joyce suggests that Llangwm had been instead awarded to the Mortimer family in 1247, and that at the time of the Inquisitions Post Mortem in 1324, Geoffrey de la Roche held two such fees (one of them Llangwm) under the Mortimers. However, by 1376, the Mortimers were in disfavour, their holdings being taken into the hands of the king. Nevertheless, the de la Roches were unaffected and the Llangwm estate was now held by the de la Roche family directly from the king.

The de la Roche family history now becomes confused, but eventually Roch also was taken into the king’s hands following the death of one John de la Roche in 1376. John had likely held Roch during the minority of his niece, Margaret Fleming, who subsequently died, still in minority, in 1382.  John had held extensive estates some of them directly from the Earl of Pembroke and others under various lords. These  included the manor at Dale (consistent with a de Vale inheritance) as well as the manors and advowsons of “Landecombe” under the barony of Roch, and land at Guildford (then part of Burton) under Isabella, widow of Sir John Wogan of Picton, (the first mention of Guildford in the records).

However, with John, the male line of the Roch heirs of the de la Roche family had come to an end, and the Roch holdings were divided among the four sisters of the previous lord, William. One of these was Joan/Johanna who had married David de la Roche of Llangwm, and who nominated her grandson, Thomas, then aged 11, to the inheritance.  Since Thomas was still a minor, he (and his lands) were placed in the custody of Sir Thomas Bermingham in November 1382. Thomas later married Bermingham’s daughter, Elizabeth. However, the Roch inheritance was complex, and was not finally solved until 1392 when Richard II issued a writ partitioning the inheritances of the de la Roches of Roch Castle to the nominees of William’s four sisters. Johanna’s nominee, now Sir Thomas de la Roche, of Llangwm, as representative of the eldest sister, received Roch castle.  However, even this royal decision wasn’t final, and Henry IV, on coming to the throne in 1399, ordered yet another enquiry, which finally settled the matter in favour of Thomas. By 1399, Thomas Roche was listed as a “King’s esquire”, suggesting that he now held Llangwm directly under the King, rather than through the Earl, consistent with Haverford coming under the direct custody of the Crown following the deaths of several Earls and the minorities of their sons, but also consistent with a role of Thomas at Pembroke.


The Glyndŵr Uprising

The hegemony of the Anglo-Normans in Wales was again threatened in 1400-1415. In the summer of 1401, an army of 1,500 raised largely in Pembrokeshire south of the Landsker took on the much smaller 150-strong force of Owain Glyndŵr at Hyddgen, on moorland north of Plynlimon in mid-Wales. The Pembrokeshire force, poorly led and inappropriately equipped, lost over 200 men, giving Glyndŵr his first military success. By 1403, the county was under threat of invasion by Glyndŵr who took Carmarthen, leading an army of 8,000.

On 31st August, Carmarthen was recaptured by Crown forces, and placed under the command of Thomas Roche (of Llangwm), who had been constable of Pembroke Castle since 1399. This success was very short-lived, and Thomas was captured together with his men only a few days later. Given that Thomas Roche was a descendant of Robert de Vale through the marriage of de Vale’s daughter to Gilbert de la Roche, while Owain Glyndŵr was the great-grandson of her sister, Thomas and Owain were cousins, an interesting quirk of Llangwm history. Clearly Thomas was of some importance to the King, since, in October, Thomas’ wife, Elizabeth, was issued with a royal licence to negotiate his ransom, if necessary in return for the release of Welsh prisoners. With south Wales still in rebellion these negotiations had apparently not succeeded by June 1406 when Thomas was allowed to appoint attorneys to oversee his affairs during his imprisonment.

Glyndŵr soon returned westwards and by November 1405 Pembrokeshire succumbed to the Franco-Welsh allies. Against the express instructions of the King, the local lords negotiated a truce with Glyndŵr on payment of £200 in silver, with instructions to every lordship in the county to acknowledge the agreement, and to contribute to the sum. The truce held considerably longer than the six months agreed, allowing some reconstruction of the Pembrokeshire economy to occur without further disruption from the Franco-Welsh allies.

The male de la Roche line ended with the death of Thomas some time before 1413, when his wife is mentioned as a widow. However, given Thomas’ responsibilities first as constable of Pembroke castle, then in the King’s army, then his imprisonment, one may doubt whether he was truly resident in Llangwm, or whether Llangwm was just one of his extensive holdings. One daughter, Elizabeth, married Sir George Longueville, whose descendants retained Roch castle until 1601, while another daughter, Eleanor, married Lord Ferrers of Chartley, thus giving rise some generations later to the Earls of Essex. Perhaps for that reason, the family was associated with the Earl of Essex’s attempted overthrow of Elizabeth I in 1601, and, on its failure, fled to France. Local legend has it that before doing so, they hid many of their valuables in an underground passage said to connect Nash to the church, but no evidence exists for either the treasure or the passage. Grenville Thomas disputes the accuracy of the legend on the basis that the Nash family were already in residence in Nash before this time and that by 1601 the Nashes had given way to the Philipps.

However, it seems possible that Llangwm came into the possession of the de la Roche family through an association with the de Vales of Dale Castle, and it is thus of interest that one of the few records pertaining to the Nash family is an incomplete pedigree discovered in the Dale Castle manuscript.19 Were, perhaps, the Nash family also related to the de Vales, and thus cousins of the de la Roche family?

Fifteenth Century

Following the death of Thomas de la Roche accounts of the Llangwm lordship practically disappear.  From some time before 1582, the Nash family were in residence in Great Nash Sir Richard Nash died at Nash in 1582, though his ancestors came from Jeffreyston, and it is unclear when Great Nash came into possession of the Nash family.  Sir Richard was succeeded by his daughter Janet, who had married Alban Philipps of Picton some time before 1594.  Nevertheless, we are left with a century and a half between Thomas de Roche and Sir Richard Nash in which I can find no account of the manor. Grenville Thomas interprets certain carvings within Llangwm church as representing the coat of arms of the Bowens of Lochmeylir, a family into which an earlier Richard Nash had married. Similar evidence points to an association of the Barri family of Manorbier, but whether either family played  a role in Llangwm’s history is unknown.

Detail from Christopher Saxton’s map of Pembrokeshire, 1578, showing spelling of Llangum.

Some notes on life in the village

In 1376, Llangwm was held by John de Roche, who additionally held extensive lands throughout Pembrokeshire. In Llangwm his holdings consisted of the manor house, a dovecot (worth 4s annually), a garden with “herbage” and fruit (also worth 4s), and the herbage of a wood. Although possession of a wood should have been valuable to the Lord of the Manor in terms of timber Llangwm was unusual in that woods were held in common for use by the estate’s tenants, both free and unfree. Similar common rights extended (but only after haymaking) to the valuable meadow land of about 3 acres, which thus had limited worth to the lord (6s.8d.).  Other agricultural land consisted of 3 carucates (a carucate is defined as the extent that can be cultivated by one plough in one year and a day and is about 120 acres, but whether the Llangwm carucates were that large seems doubtful as they had a worth of 13s. 4d, only twice the value of the 3-acre meadow). Curiously, given the reputation of the Flemings of twelfth century Pembrokeshire for wool production, no mention is made in the records of such activity in the Llangwm estate. In addition to income directly from his land, Sir John received rents from his tenants worth £2 13s. 4d., giving a total income of £4 1s. 4d. Contrast this amount with the income from an equivalent English knight’s annual income of between £20 and £40. Even within Wales, it has been suggested that the fourteenth century, a knight might have an income of £10 to £20, an esquire half that, and a fifteenth century gentleman less.  Clearly, the Lord of Llangwm did not get by on his Llangwm income alone.

However rural and feudal Pembrokeshire had been, by 1500 it was rapidly changing, even on the land.  Enclosed fields were to be found in the Hundred of Roose, around the seats of the gentry and freehold farms. Individual tenants might hold a few small scattered plots, so that in 1623-4 a recommendation was made to exchange and consolidate such plots and allow their tenants to enclose them. Nevertheless, Llangwm was still organised as a manor, paying a knight’s fee as late as the sixteenth and early seventeenth centuries.20

Despite its association with the upwardly mobile de la Roche family, and later with the Nash family, one should not imagine that Llangwm had much importance in its own right. In 1563 the entire population of Llangwm (including Hook, but excluding Guildford) was only 15 households – giving an estimated total population of 60 to 70.  Burton and Rosemarket were perhaps double that size and Freystrop and Johnston a little smaller. Only five of the twenty-eight parishes in the hundred of Roose had fewer people, and Grenville Thomas speculates that Langome was not really a village at all in a modern sense, but a few dispersed clusters of dwellings. If the manor house at Nash possessed a usual number of servants for the time, then one can speculate the number of other villagers was even lower. According to the Black Book of the Household, drawn up by Edward IV in 1471-2, a knight’s household might have 16 servants. An esquire could make do with ten. 21 Thus, Llangwm differed from nearby nuclear villages like Rosemarket which showed the typical Norman structure of a manor house with nearby church and a surrounding settlement. Indeed, the physical separation of the manor house at Nash, and the church on the Green, speaks strongly against Llangwm existing as a nuclear village.

The world was on the brink of change, to which Pembrokeshire, and Llangwm in particular, were not immune. Indeed, Llangwm seems to have been radically affected, for between 1563 and 1670, the number of dwellings increased from 15 to 73, by far the largest percentage increase in Roose (or even Pembrokeshire), where most parishes remained stagnant or increased only modestly in size. In comparison, neighbouring Burton increased from 36 to 65, Rosemarket from 32 to 50, and Freystrop from 12 to 28. The cause was almost certainly the development of coal mining, providing both a need for labour and opportunity to escape from the land. Another factor was the number of Irish refugees pouring into Pembrokeshire. Already in 1528 Sir Rhys ap Grufydd complained that 20,000 Irishmen had landed in Pembrokeshire within the past year. While that may have been an exaggeration, the impact must have been enormous bearing in mind that the total population of Pembrokeshire at the time was less than that number. George Owen remarks that some parishes became entirely Irish, apart from the priest 22. Another wave of Irish came in 1628 and 1629, transported in small boats, one case citing 70 Irishmen being carried in a boat of ten tons. They arrived in an area already hungry from a shortage of corn that was being bought up by merchants and exported. A further wave of Irish refugees came in October 1641.

While south Pembrokeshire and Rhos may have had a strong Flemish influence in the mid thirteenth century, three hundred years later any remaining Flemish influence had been overtaken by English and Irish immigrants, and by influx from neighbouring Welsh areas, partly as the coalfield developed, but also with land ownership. Consistent with such changes, the Hearth Tax returns of 1670, providing the first listing of names of people living in the parish, shows a strong representation of Welsh surnames, while others may be of English origin (though the rector, Henry Purefoy seems to have a French name, consistent with the identity of the church and the aristocracy). The owner of Great Nash at this time was Thomas Corbet, but many of the other Llangwm landowners carried names like Eynon, Jones, Meredith, and Davies. The age of Llangwm as a Flemish village, if it ever truly existed, was long over.


  1. David, et al., Llangwm: a newly identified early Mesolithic site in southwest Wales, in Archaeology in Wales 54, 15 – 24.
  2. Heather James, Pembrokeshire County History, Vol. I, 382.
  3. W. Grenville Thomas, Llangwm through the Ages, ( Haverfordwest, 1991).
  4. Henry Owen, Old Pembroke Families, page 75.
  5. Bob Joyce. “Ancient lordships of Pembrokeshire”, in The Journal of Pembrokeshire Historical Society, 2016.
  6. Llangwm Heritage
  7. John of Worcester page 252
  8. CADW: Hook
  9. Haverford IPM (1324): Calendar of Inquisitions Post Mortem, Edward II, Vol. VI, (Hereford 1910), 336-337.
  10. The Dale Castle manuscript may be a copy of a more general geneology.
  11. Roger. Turvey, in Pembrokeshire County History, Vol. II, 370.
  12. Brian Howells, in Pembrokeshire County History, Vol. III, 67-9CADW: Hook
    1. Haverford IPM (1324): Calendar of Inquisitions Post Mortem, Edward II, Vol. VI, (Hereford 1910), 336-337.
    2. The Dale Castle manuscript may be a copy of a more general geneology.
    3. Roger. Turvey, in Pembrokeshire County History, Vol. II, 370.
    4. Brian Howells, in Pembrokeshire County History, Vol. III, 67-9.
    5. Haverford IPM (1324): Calendar of Inquisitions Post Mortem, Edward II, Vol. VI, (Hereford 1910), 336-337.
    6. The Dale Castle manuscript may be a copy of a more general geneology.
    7. Roger. Turvey, in Pembrokeshire County History, Vol. II, 370.
    8. Brian Howells, in Pembrokeshire County History, Vol. III, 67-9.
    9. Brian Howells, in Pembrokeshire County History, Vol. III, 67-9
    10. 22. George Owen of Henllys, The Description of Pembrokeshire.