Fig. 1. Walls and floors of buildings of the ‘lost village’ exposed in the eroding sand hills below St Ishmael’s Church

Fig. 3. Suggested coastline for the Three Rivers at the end of the Middle Ages. Note how the Gwendraeth then ran south into the sea around Caldecot Point. Compare with Fig. 2.

Fig. 5. Suggested location of principal medieval weirs on the Scars.


Fig. 10. Surviving stakes of the fish trap of the weir know as Y Caj, looking SE with the eastern end of Wall B (arrowed) in the distance.


Fig. 11. The principal weirs on Salmon Scar showing the large natural catchment area (light sand) with ‘leader hedges’ running to the cage. Polyion Ifan (Weir C) is to the right.


Fig. 13. The seaward end of Salmon Scar showing the location of what are assumed to be stake lines for strem nets. The Barrel Post stump and upturned trailer are included to aid orientation on the ground.

Fig. 14. Part of Weir G viewed from Salmon Scar looking S with the peat shelf area known as Y Colbacse beyond.


Fig. 16. Vertical air view of the weirs on the west side of Pastwn Scar.


Fig. 18. Weir L, a spade cut trenching in the peat shelf with packing stones at stakes, looking west.


Fig. 20. Ruts and linear impressions in the peat shelf denoting a ‘transport corridor’ between weirs O and J, looking S.


Reproduced from The Carmarthenshire Antiquary, Vol.xxxix 2003


Fish weirs on the Tâf, Towy and Gwendraeth estuaries, Carmarthenshire.

Heather & Terrence James



Numerous fish weirs have been identified in medieval and early-modern records of the Marcher Lords of Llanstephan and Kidwelly. Both Carmarthen Priory and Whitland Abbey possessed fish weirs in the estuaries of the Tâf, Towy and Gwendraeth (the Three Rivers). Some of these weirs on the Gwendraeth rivers were lost probably in the early seventeenth century as a result of the growth of the Cefn Sidan sand dune barriers and salt marshes and subsequent enclosure by sea walls. The landward edge of stony scars off St Ishmael’s, which have the remains of medieval stone buildings at the edge of high water sealed by sand dunes, appear to have been formed from an eroded promontory. The amount of land lost to the sea cannot have been great. There is little doubt that some of the weirs on the scars today are medieval in origin, and the besanded medieval settlement at St. Ishmael’s (the ‘Lost Village’) could relate to a fishery belonging to Whitland Abbey. The weirs may have fallen out of use because of besandment (and perhaps because of economic change) and this might explain the lack of documentary evidence for their use after the seventeenth century. The loss from local memory of their use may account for weir walls being erroneously identified with the lost village. It is fortunate that we could record the oral testimony of David Phillips the last person to work one of the weirs.



The existence of a 'lost village' in the sand dunes and over the Scars below St. Ishmael's Church is well known in local folklore. These beliefs are partly based on real physical remains of settlement. They consist of stone slabs forming walls and floors exposed on the edge of sand dunes on the seaward side of the main London to West Wales railway line below St. Ishmael's Church. These are often covered or re-exposed by stormy weather and are certainly eroding. The remains are exposed only in section, as a veritable time-slice, being picked at by the ravages of seasonal storms (Fig. 1). They were briefly described with photographs in the Royal Commission's Carmarthenshire Inventory of 1917.1 Also figured but inadequately located is an apparent stone-slabbed well head, exposed between tides. This, it is believed, is positive proof of land loss and a more extensive settlement that had once extended out onto the Salmon Scar.2

We have spent many years since the early 1980s visiting the ever-changing low tide exposures of the Salmon and Pastwn (Pastoun) Scars ('The Scars') and the sand-banks of Cefn Sidan and Carmarthen Bar, firstly in the company of the late W. H. (Bill) Morris, looking for wrecks. At that time we often debated whether part at least of the Scars could be the spread remains of training walls built in the early nineteenth century to improve the navigation of the River Gwendraeth up to Kidwelly Quay. Later we began to identify in the curving lines of stone the remains of stone-walled fish weirs, some with traces of stakes set in them, and especially the very well preserved 'cage' of the most substantial weir. Participation by one of us in a Welsh language TV series, Troed y Fran3about the estuaries of the Tâf, Towy and Gwendraeth involved an interview out on the Salmon Scar with Mr David Phillips, the last man to have worked that weir in 1963. Aerial photography in ideal light conditions and at low water spring tides, provided excellent photographs from the right perspective to enable us to begin to map most of the remains of the fish weirs. Position fixing was made possible by the use of a handheld GPS.4 This article attempts to discover the origins of the fish weirs and their periods of use. 5 We hope that concentration on the present day intertidal archaeology of the Salmon and Pastwn Scars will also serve to throw fresh light on the question of the 'lost village' below St Ishmael's Church—from a seaward rather than a landward perspective.




The quantity, even the range, of fish available today in Carmarthen Bay and its estuaries is a pale reflection of earlier centuries. Statistics compiled by the South Wales Sea Fisheries Committee in 1993 record landings for Milford Haven, the Three Rivers and Llanelli.6 The most commonly landed species were skates and rays, dogfish, cod, hake, shark, plaice, whiting, mackerel, pollack and sole, and also, from deeper waters, the angler or monkfish. Amongst pelagic fish the herring, once so important to Tenby, is no longer present in Carmarthen Bay in great shoals. 7 Its close relatives, the estuarine allis shad and twaite shad, are now rare and protected. Also within the estuaries are flounders, dabs and mullet. Conger eels lurk in pools out on the Bar around wrecks. The regulation of the salmon and sewin fisheries has a long history stretching back to at least the 13th century.8 Today, because of its slow growth and rarity, fishing for bass is also regulated to protect the stock. In addition the estuaries still provide cockles and mussels and in the past crabs and even lobsters. Oysters and prawns were once common off Tenby and in the Bay.

Methods of fishing by rod and line, nets, both drift and fixed, fixed traps and 'stabbing' with spears are all of great antiquity. The celebrated coracle fishery of Carmarthen and to a lesser extent St Clears has been much studied.9 Equally well known are the cockle-women of Llansaint who worked around the Scars and off Ferryside. Salmon and other fish were and are taken by seine netting today out from Ferryside and Llanstephan and by the near-extinct wade-netting from Laugharne in the Tâf. A few stake nets are still set under licence, and nowadays can only be set parallel to the shore unlike the earlier strem nets.10 The evidence from Swansea Bay, where there were numerous stone fish traps in use in the nineteenth century, suggests that stake nets replaced the stone traps. Supposedly stake nets trapped fewer of the young fish useless for eating and essential for future breeding. It is also self-evident that setting a net is less costly than constructing and maintaining an extensive stone weir. Yet there is, to our knowledge, no mention within all the modern 'regulatory' literature dealing with the fish and fisheries in the estuaries of the three rivers of stone fish traps or fish weirs, (goredi in Welsh).11



In recent years there has been a virtual explosion of archaeological data on the coasts and inter-tidal areas of Britain and Ireland. Fish traps and weirs have been recognised, described, surveyed and excavated in many coastal and estuarine locations.12 Timbers from weirs and traps constructed from stake lines can provide radio-carbon date ranges such as those between the 9th and 11th centuries off Sudbrook Point, in the Severn estuary.13 Nearer to Carmarthen Bay, at Whitepool Point, Gower, a 'v' shaped trap, surviving as long lines of stakes, was revealed for a short period. One of the stakes produced a thirteenth century radio-carbon date (710±60BP, CAR-1232).14 Lines of stakes for fish weirs have been noted off Llanelli in the Burry Inlet, the largest concentration being on submerged forest levels at Morfa Bacas. Samples taken from the more complete lines produced radiocarbon dates from the seventeenth and early nineteenth centuries, earlier structures perhaps having been eroded away by the northward migration of the course of the Loughor.15 Documentary sources for medieval fish weirs in the Severn estuary and Bristol Channel are plentiful16 and from the English side of the Channel a Saxon charter of 690 AD records a fish weir off Aust. A charter dated by Professor Wendy Davies to c. 895 AD records goredi at the mouth of the River Troggy at Caldicot, Gwent.17

Many of the examples noted above were constructed solely from timber stakes with wattle fences between, perhaps anchored in lines of stones where these were available. In Cardigan Bay off the shelving pebble beach of Aber-arth are the remains of stone-built fish weirs that enclosed larger areas than the 'v' shaped traps. Weirs at Aber-arth were granted to Strata Florida Abbey by Rhys ap Gruffydd in 1184. They may well predate the 12th century. 'Gored Wyddno', described as being on the beach between the Dyfi and Aberystwyth, is noted in the 9th century Hanes Taliesin. An article in 1924 on the Aber-arth goredi contains valuable oral information on their working by a Miss Davies, the last of the gored keepers. It was hard work rebuilding the walls of a gored after the winter storms and attending the weirs on the tides, often at night, since poachers could take the catch.18 Equally celebrated are the stone-walled goradau of the Menai Straits, particularly the substantial Ynys Gorad Coch set on an island in the middle of the Swellies, the notorious tidal eddies of the Straits.19

Stake and net v-shaped traps and the curved stone goredi work on the same basic principal of trapping fish on the ebbing tide. In river mouths and estuarine channels the point of 'v' shaped traps faced down stream. Fish are guided on the ebb by the stake and wattle walls, or 'leader hedges' (which could be of great length) towards the apex of the 'v' where they were finally trapped in a woven basket or a net. The stone-walled weirs had lines of stakes set along them, supporting nets. The fish would be trapped in the large pool formed by the weir's walls on the falling tide. The ebbing water would guide them towards the mouth of the weir towards a cage (locally caj ) where the fish would be gathered up at low water. Undoubtedly a host of variables gave each and every weir a slightly different structure and catch depending on location. But one point worth emphasising at this juncture is the indiscriminate nature ofthe weirs and traps because they could entrap a great variety of species.



The cartographic record of the area forms an essential component for understanding the parameters of coastal evolution and thus locating fish weirs through time. Such has been the growth of Pembrey/Kidwelly Marsh and Cefn Sidan that it is probable that many early weirs are now buried well inland. Christopher Saxton's topographical surveys of England and Wales carried out between 1573 and 1578 were to more accurate cartographic standards than ever before. They resulted in the publication of an atlas (1579) and a wall map (1583).20 John Speed used Saxton's surveys for his county maps such as that of Carmarthenshire published in 1607. In the area under consideration here Saxton's achievements have been overshadowed by criticism of his inclusion of both 'Llansaint' and 'Hawton' on the map of the St. Ishmael's area as though they were two distinct places. Hawton was traditionally thought to be the name of the lost village, perhaps drowned by storms in the early seventeenth century. However the late Professor Stephens, who lived in Ferryside in his retirement, has shown quite conclusively that 'Hawton' is an English form of 'Llansaint', a fact earlier realised by Mary Curtis. 21 Saxton's 'Calicote Point' (Caldecot—see Fig. 3) is probably a quite accurate representation of the coastline of the sixteenth century, given the limitations of the scale and projections used. To the west, Laugharne Marsh is shown before complete enclosure and the dramatic eastward growth of Ginst Point since the Middle Ages.22 Less attention has been paid to Saxton's depiction of a pronounced pointed peninsula between the Towy and the Gwendraeth labeled 'Sct Ismael point', quite unlike the rounded bluff of the present day coastline (see Figs. 2 and 3). This appears to correspond with the area of Salmon Scar and it supports the assertion that we have suffered land loss there since Saxton's surveys.

Emanuel Bowen's map of South Wales of 1729 is the next map of interest. He gives an indication of areas of 'marsh' by means of horizontal lines. These lowlands seem to have expanded since Saxton's time and were salt marshes inundated only at high tides. In addition to showing the developing 'Kidwelly Marsh' (also known as 'King's Marsh', Fig. 9) a noticeable change from Saxton is Bowen's depiction of the present day rounded bluff of St. Ishmael's. Unfortunately Lewis Morris's individual coastal surveys of Wales of the 1740's did not extend as far south as the Towy estuary. The first detailed chart therefore is that of Murdoch McKenzie in 1770.23 His is the first cartographic evidence for the existence of sand dunes south of St Ishmael's point. Surprisingly, in view of its use as a landmark visible from seaward, he does not mark St Ishmael's church, or the Scars. For the first time we have depicted the channels with low water soundings across Carmarthen Bar. The Gwendraeth Fawr and Fach, conjoined north of 'Pentowen Point' (i.e. Towyn Point, formerly Caldecot Point) and approximately south-west of where the later Bertwn ballast quay was to be built. North west of the confluence of the two Gwendraeths there is an abrupt parting of the channel, with one branch (with soundings) continuing north into the Towy. The other ran south past Towyn Point and parallel to Cefn Sidan to reach the open sea some miles west of Penbre. Although unnamed, this is what was known as the Guy Channel and is shown without soundings.

The next cartographic record is William Morris's most useful 'Carmarthen, Lougharn and Cydwely Bay, Bar and Harbours' of 180024 where changes since 1770 are apparent. The William Morris chart does not record depths of water in his channels, or mark the Scars, but gives more place-names and local detail than McKenzie. The confluence of the two Gwendraeths is shown close to the shore south east of 'Burton'—Y Bertwn—where, by 1800 there was a 'Ballast Quay' marked. Thomas Kymer built this in 1783.25 The Gwendraeth flowed west into the River Towy around the tip of the Salmon Scar (although the Scar is not shown). A blind channel, a stub of the old Guy Channel, is marked as 'Pen Towyn Groats'. South-south-west of Tanylan, and on the edge of the low water channel, there is a 'Barrell Post' presumably to mark the Pastwn Scar. Clearly the Gwendraeth at this date ran hard by the Scars. Two other names are given for the inter-tidal areas. At the confluence of the Rivers Tâf and Towy, below Llanstephan Castle but within the channel, is a spot named as Cerreg Ebware discussed later. In the area of the Salmon and St Ishmael's Scar of today is the legend Pwll yr Ychen ('Ox-pool')—i.e. a place where cattle could drink—which is also discussed below in the section on topography and geology.

Concern about the increasing difficulties of navigation across the Bar which was threatening the export of coal from Kidwelly Quay, led to several schemes being proposed which required fresh surveys. Accompanying the Rennie and Bankes Reports of 1820, are two charts of 1807 and 1814 by John Wedge.26 Most interestingly for our purposes, the 1807 survey, uniquely, marks a 'wear' and a scatter of stone south east of the Pastwn Scar (though no scar is named by Wedge). The 'wear' is separated from the landward scar by a channel. By 1814 the channels had changed and no weir is noted. Two barrel post are shown on the edge of the channel, one of which is marking a scar, probably the Pastwn. Curiously, this chart does not specifically mark the one small-scale piece of improvement work to coastal navigation, which we know to have taken place. This was called the 'canal', a cut across the Salmon Scar to allow a shorter passage for small craft delivering coal from Kidwelly around and up into the River Towy (see Figs. 7 and 11). On April 30th 1813, the Carmarthen Journal reported tenders being invited to restore an ancient channel between the Gwendraeth and the Towy (see Fig. 4) following a meeting of those interested in improving Kidwelly harbour. Work was completed by August 13th of that year. Thisreference might refer to 'the Canal'. The passage is shown with a small mark at its western end on the 1880 Admiralty Chart but its use by colliers from Kidwelly was much reduced by the 1840s when the Gwendraeth had silted up and coals were shipped from Burry Port and Llanelli.27 However small local coal boats continued to make this journey until the 1920s.28 Both charts show the now open 'Guy' Channel, which was buoyed, and passed around Towyn Point across Cefn Sidan sands from which direct access from the sea was gained into the Gwendraeth.

It is only with Lt. H. M. Denham's survey of 1828 and chart of 1830 that we enter the modern era of Admiralty charts with their accurate bathymetry based on many hundreds of soundings. His charts were accompanied by the Sailing Directions for the Bristol Channel published by the Hydrographic Office in 1839 which contain much useful detail.29 By 1830 a low water channel of the Gwendraeth (the 'Kidwelly Guy') ran west and south and did not join the Towy (Fig. 4). Its precursor, the 'Old Guy' channel (perhaps that figured by Wedge) had closed to navigation. Entrance across Carmarthen Bar into the Towy channel to the north was buoyed. The Denham chart shows the Scars as areas of 'large stones' by stippling with a line around the area, distinct from sand or mud banks. The largest of these extended west from Danylan with a detached scar (Pastwn Scar) west of this. The Gwendraeth Fawr still ran hard against the seaward side of these scars. The barrel post on the Pastwn is shown, and also the course of small streams draining seaward across the intertidal scars and mud from below Danylan. It is significant that the Salmon Scar is shown to be detached from the mainland apparently by besandment. The scar appears to be not as exposed as today, which would have conditioned its usefulness for maintaining fish weirs. Denham shows what might be scars west of Kidwelly Quay (bottom r. h. corner, Fig. 4), that could have been a location of earlier weirs. On the Llanstephan side of the Towy there are scars south-east of Wharley point and south of the castle in Scott's Bay, both protected by a white barrel post. These are the prime areas where Llanstephan's main weirs (e.g. Lady Weir, Ebbweir) were formerly sited (Fig. 9). On the Ferryside side there is a black barrel post with scars to the east and south.30 It possible that a weir was located close to the ferry crossing.

The next Admiralty chart based on fresh survey was not published until 1888. It depicts in great detail the by-now tortuous course of the Towy channel across the Bar and the shallow Gwendraeth flowing into it. The ever-advancing lines of sand dunes of Pembrey Burrows are carefully mapped which show a considerable growth of Towyn Point and the salt marshes east of this. For our interest the detail of the Scars is significant showing the Salmon and Pastwn Scars much as they are today. A Barrel Post is shown on the Pastwn with two lesser posts on the Salmon Scar and on the northern end of the 'canal' boat passage. The 1949 Admiralty Chart shows the further dramatic north-westward growth of Towyn Point and the consequent narrowing of the Gwendraeth estuary. The barrel post on the Pastwn and a wreck (the Paul) are shown but not named. The current Admiralty chart, no. 1076 'Linney Head to Oxwich Point', scale 1:75000 still relies on obsolete 1888 and 1941 survey data for the estuary.

A barrel post on the Pastwn Scar is shown on all nineteenth century charts, but only the 1888 chart shows the Salmon Scar post. This tends to further demonstrate that the Gwendraeth navigation was running up to the Pastwn Scar and then at least for some of the time running south into the sea through the Guy channel (i.e. not passing Salmon Scar). It was only when Towyn Point had grown so far to the north-west that the Gwendraeth was forced irrevocably to flow into the Towy. It was then that the Salmon Scar became a danger to navigation. Today the foundations of a substantial Barrel Post are to be seen at the tip of the Salmon Scar, but we have found no physical evidence for the longer-lived Pastwn Barrel Post.31

The fact that the fish weirs have never—apart from a single note by Wedge—been charted is no argument for their non-existence at the dates shown since they were adequately covered as dangers, by the barrel posts. It is however reasonable to suppose that in certain conditions of besandment, (and we have hints of this in Denham's 1830 chart), the fish weirs would become unworkable. Certainly, in our own experience over the past two decades, there have been seasons where parts of the Salmon Scar weirs have been covered in mud and sand. These conditions therefore differ, for example, from those pertaining to the stone fish weirs at Aber-arth in Cardigan Bay or the Menai Straits.

The cartographic evidence and the understanding we gain from their study show how the coastline has changed, and also how the extent of the Scars at the estuary mouth has altered. This will condition our interpretation of the whereabouts of weirs mentioned in early documents discussed next.


Documentary Evidence for Fish weirs in the Tâf, Towy & Gwendraethestuaries

The medieval lords of Wales, both temporal and spiritual, English and Welsh, controlled fishing in both the inland waters of their lands, on the seashore and over the intertidal areas.32 In the Tâf, Towy and Gwendraeth estuaries these were the Lordships of Laugharne, Llanstephan and Kidwelly. Written records of fish weirs for our study area begin in the administrative and financial records of the Crown and of the Marcher Lords. But such rights had been the preserve of the native Welsh princes and, as we have seen, are likely to extend back into the early medieval period. Stone fish weirs were often massive constructions requiring the resources of the major landowners to build, but their upkeep and certainly their operations were leased out to groups or single tenants. We have found no references for weirs in the lordship of Laugharne in the Tâf estuary or indeed where the exact boundary between Laugharne and Llanstephan might lie. But for other landowners we have, around 1170 to 1180, Geoffrey de Marmion, Lord of Llanstephan's gift to the Slebech Commandery of the Knights of St. John of Jerusalem of the church at Llanstephan, a carucate and 50 acres of land, a boat at the Ferry for free transport across the Towy and a fishery in the Tâf. 33

Several weirs are mentioned in the medieval and early modern records of the Lordship of Llanstephan. In 1379, Simon de Burley was Lord of Llanstephan. During Quo Warranto proceedings it was stated that his maritime rights comprised wrecca maris (wreck of the sea), whatever was found floating on the sea shore, all royal fish (sturgeon and porpoise), free fishery in the River Towy with nets and other engines, and weirs on his side (i.e of the River Towy) to [blank in the manuscript] at will.34 In 1388-89 the Lordship was in the King's hands and therefore ministers' accounts have survived in the Public Record Office. They record a payment of fifteen shillings for two weirs 'farmed' (i.e. leased) to 'divers persons, with one weir remaining in the king's hands through want of tenants, though normally yielding two shillings a year'.35 In 1411 the lordship was again in the king's hands and in the ministers' accounts for that year the weirs are named as 'Ladywery', below the Castle, 'La Newere', two weirs called 'Blanchard' and 'Vincent' and 'Edwer'.36 Because of its location away from the estuary, we think that 'Gored Rothe' up river of Llanstephan at the tributary Roath stream in Penrhyn lordship, was of a different type. No income was received from the weirs as they had not been let and were said to be 'destroyed'. In 1413, following receipts from demesne lands around the castle, the weirs Newere, Blanchard and Vyncent, and 'a certain weir called Edwere' are again listed but were still unlet, no doubt because of unsettled conditions in the Glynd uprising.37 Further down the account, following income received from the tenants of 'Morbichurch' (Llanybri), two shillings came from 'a certain weir called Ladwere there'. By the end of the century, an account of 1481-82 records 'of the farm of Vincentiswere and le Roke, late at 3 shillings per annum and the farm of the weir called 'Ladiwere alias Embwer, late at 16d . forsooth . . . nil because they are fallen through want of repairs'. Also 'of the farm of one weir there newly built on the sand on the side of the ferry there' no profits.38

By the early sixteenth century emerging gentry families were leasing much of the lordship lands from the crown. In the Patent Rolls for 1532, amongst lands at Llanstephan leased to Maurice Lloyd and Jankyn Lloyd, including Castle Hill and pasture was 'one weir called Ladywere alias Ebewere' for 12d. per annum and one foundation ( fund' in Latin) of another weir 'newly built by the aforesaid Maurice and Jankyn . . . to be called Neuslode weir'.39 Later in the century, in 1564, in the records of the Court of Augmentations, the Lloyd family continued to lease parts of the Lordship with Thomas 'Luide' taking on, amongst other lands, a 'weir called Arcombewere in tenure of Maurice Lloid, used to be 2 s., now 12d.; another weir called ladywere' also the two weirs which had been leased by Letters Patent to Maurice and Jankyn Lloyd for 12d. per annum.40

It is possible to suggest a general location for these Llanstephan weirs. Firstly their longevity and the notice of repairs (or lack of them) and the phrase 'newly built' all suggest stone fish weirs. It is likely therefore that these were built where there were areas of intertidal 'scars' or strews of stone previously discussed. The present day area of stone off Scott's Bay, due south of the Castle, fits the bill although presently besanded (Fig. 9). It also seems likely that there were groups of weirs such as we find today on the Salmon and Pastwn Scars. Although in 1411 Ladyweir (below the castle) and 'Edwer' are separately listed from then on they appear as 'Ladyweir alias Embwere or Ebweir'. A 'Cerreg Ebware' (Ebbweir Rock) is marked on William Morris's 1800 chart at the then confluence of the low water channels of the Tâf and the Towy. This is the outer extremity of the Scott's Bay scars and probably the same general area as the later Barrel Post. Interestingly, although the name Ebwere drops off the cartographic record, it is still known to seine netsmen. In the late 1980s the late Mr Mostyn Thomas of Ferryside, told us that they occasionally snagged their nets on a 'Carreg Ebwer' but that it never uncovered. The coastal changes here and across at Ferryside and the degree of silting within the estuaries is now so great that the remains of the weirs are probably buried below sand and mud. This does not preclude the possibility of them uncovering at some time in the future.

For Kidwelly, coastal change has been even more extensive (cf. Figs. 2 and 3). Yet surviving out on the Salmon and Pastwn Scars are the remains of several fish weirs, which we shall attempt to identify with named weirs from documentary sources. The medieval records for the Duchy of Lancaster's Lordship of Kidwelly are voluminous and, in the main, unpublished. However some were transcribed from the Public Record Office for the Earls of Cawdor in the nineteenth century, in support of various titles to land and mineral interests. There is no doubt that more could be gained by a complete search through the fourteenth and fifteenth century accounts and court rolls for the Duchy of Lancaster's Lordship of Kidwelly. Such documents became fossilised in form and often content and it is clear that some weirs are being accounted for long after they ceased to exist. Names too may well have changed for a neglected weir when it was rebuilt and relet. There appear to have been two types of weirs—those named over a long span of time andyielding substantial revenues and those worth only 12d or less. Often that rent is paid for the site of a weir which the lessor or his undertenant had then to construct and maintain. The thirteenth century was the highpoint of demesne farming by lords, supported by the labour services of their free and unfree tenants and the latest period perhaps when large capital works like the stone fishweirs might have been constructed. The earliest record we have seen is a copy of Patrick de Chaworth's Inquisition post mortem of 1283, with an English translation, in the Cawdor/Vaughan muniments, that notes weirs (latin gurgites) in the lordship, worth 30 shillings a year.41 These may have included fish traps in rivers.By the later thirteenth and early fourteenth centuries such labour services were being 'commuted' for money payments but the original nature of the service remains fossilised in the ministers' accounts and surveys. So it is of great interest to find in the lists of redditus assisae or rents of assise in the 1488-9 account that amongst other agricultural services now commuted, the tenants of StIshmael's once owed no less than41 'works' (day's labour) for the carriage of osiers to the 'Weir of Towy' and 84 works for the repair of the weir, commuted to payments of 20d. and 7s. respectively. 42 Henry de Lancaster's IPM of 1362 lists 20 shillings revenue from 4 weirs.43

Many weirs are named in the fifteenth, sixteenth and seventeenth century accounts and surveys we have been able to consult. Using relative values, locational indications and place-name evidence, Fig. 5 shows where we think the main weirs were sited, which accords quite well with the surviving physical evidence out on the Salmon and Pastwn scars today.

Broad weir was clearly an important, long-lived and valuable weir. In 1443-4. 'Brodewere' yielded a rent of 20s.44 This was paid together with the rent for Caldecot south of Kidwelly (Fig. 9) where the Butlers of Cwrt, Pembrey, longstanding officers of the Lordship of Kidwelly, rented lands and several other weirs described below.45 In fact, within the scope of this lucrative lease, they had sub-let an area adjacent to Broad weir to one Thomas ap John on a 12 year lease who had made (fact') the weir but still paid a small 12d . rent. In the account for 1488-9, nearly 50 years later, it would seem that considerable coastal changes had taken place.46 Only 6s. 8d. was received for Broad weir. The previous 12d for the weir 'in the water of Towy between Brodewere and the sea' was not received because, along with others, it was 'utterly destroyed and decayed'. A Duchy of Lancaster Survey of 1504 notes that 2 weirs, 'Brodewere' and 'Willy Thomas Skyer' had been abandoned after being engulfed by the sands of the sea.47 But by the time of the 1609 Survey,the sands seem to have shifted once more and Broad weir was again in use.48 It was leased by William Morgan, gent. together with 4 acres of land and marsh in St Ishmael's and a meadow (pratum) called 'Little meade'. 49 This meadow is mentioned in the fifteenth century accounts and it is possible that all these elements of land—meadow, saltmarsh and weir—were contiguous. Our conclusion is that Broad weir was probably sited on the Pastwn Scar. It may have been the weir of Towy to which the labour services of carriage and repair by the unfree tenants of St Ishmael's were owed. The physical remains today(Figs. 16 & 17) accord well with the three weirs near and to seaward of Broadweir referred to in the documents.

The first reference to 'Salmon were' which we feel must be located on the Salmon Scar, is in the 1488-9 account already referred to above. It is grouped with Broad weir, each yielding half a mark totalling 13s. 4 d. Broad Weir had yielded 20s. in 1433-4 but at least they were still functioning when weirs to seaward and landward were in decay or lacked tenants. 'A certain weir between Salmon Wear and the sea' for example did not yield its rent of 12d. 'for want of a tenant'. This central siting suggests to us that Salmon Weir, like Broad Weir, was one of the four major stone weirs of the thirteenth century or earlier. Next in this section of the account, which suggests proximity to the two preceding weirs, and therefore a possible location on Salmon Scar, was a weir called 'Hyeshilly', normally yielding 6s . 8d.—i.e. a reasonable sum, but again in 1488 not let for want of a tenant. Place-name evidence discussed later may further support the suggested location of 'Hyeshilly' on Salmon Scar.

The 1443-4 and 1488-9 accounts note rents for 'one place to make a Wear between Monkswere and the sea'. At this date Monkswere was not part of the Lordship, but by 1609 'Kored yr Abbot' (alias Monk's Weir) evidently was. There are two references to this fish weir in that Survey. Firstly there is a list of freeholders within the 'Fforrenrie de St. Ismaells'.50 This includes 'David ap Jeuan Gwillim' holding a tenement and lands of about 20 acres and 'one ffyshinge weare uppon the River of Towey called Cored yr Abad'. Like the other freeholders his right to title is specified and in this case it was by a lease from John Reede of Carmarthen who had himself leased the weir from William Abbot of Whitland on 16th October 1539.51 Like some others in the list, no rent is specified for this holding. In a subsequent section of the Survey, the jurors answered'The Eighte Artickle' of the Survey relating to the forests, and weirs which were once part of the demesne of the Lordship and what their present extent and degree of 'appropriation' to private interests were. Having detailed encroachment and felling of former forests, the jurors noted that within the foreignry in the parish of St. Ishmael 'there is a certaine tenemente of lands contayninge viii acres or thereabouts called by the name of Sythen Chappell Sainte Leonard with a ffyshinge weare called The Abbotts weare' on the River Towy.52 David ap Jeuan ap Gwillim is the undertenant of John Reade, gent. and details of his lease in 1539 are given—89 years at 10s. a year. The yearly rent of 10s. was now, the jurors reported, paid to the Crown. Finally this section refers to 5 fishing weirs belonging to the Lordship of Kidwelly but actually lists only four.

The Survey names first the weir of Sir John Doune on the Towy, known as Kored Jany, lately held by Richard Cotten esq. and worth 6s. 8d .—but now decayed and no rent received. Second 'Kored yr Abbot' aforesaid 'which lyeth uppon the River of Towye toward the sea, in the tenure of Francis Maunsell esq and worth 5s. p.a. No third weir is given. The fourth weir, called Korred Vache, was also on the Towy, worth 2s. p.a. but now decayed and also in the tenure of Francis Maunsell but the jurors did not know by what right he held the weir. The Fifth weir was 'Kored Shilly', formerly in the tenure of Harry Vaughan, gent, but the jurors did not know the rent and in any case it was 'all decayed'. Can these weirs be identified with the late medieval weirs previously discussed?

Firstly, if we exclude Whitland Abbey's weir, which did not come into the Lordship until the Dissolution of the Monasteries, we have in 1609 four major weirs just as in 1362 on the death of Henry de Lancaster. The missing third weir of the 1609 list must be Broad Weir that William Morgan had leased. Sir John Dwnn was all powerful in Kidwelly in the 1460's and 1470's53 and his weir leased from the Lordship would have been one of the major weirs. His name was still current in the memories of the early seventeenth century jurors even though the lessor was Richard Cotten. In 1609 the weir was also known by the name of its undertenant. A John David Jany is listed as holding a single burgage in Kidwelly in the 1609 Survey and he is likely to be the same man as the John David Jany described as 'yeoman' in 1595.54 Our guess is that this weir was the fifteenth century 'Salmon were'.

Korred yr Abbot alias Monkswere is linked, as we have seen, with a tenement which contained a chapel dedicated to St Leonard, all belonging to Whitland Abbey. How and when Whitland Abbey acquired this property is not clear. The medieval records of the Abbey's lands are sparse but it is surely more than co-incidence that in 1270, Patrick de Chaworth, Lord of Kidwelly, gave Whitland Abbey 19 acres of land within his lordship in return for chantry masses to be said for his soul.55 We suggest two possible locations for the Whitland Abbey holding. There is a field named Bettws and others called Smalls/Swales (? Ishmael's) in the 1841 Tithe Schedule for St Ishmael parish which is on the hill slopes west of Pengay Farm just above the church and not too distant from the 'lost village'.56 Another possibility is a field referred to in a front page advertisement in the Carmarthen Journal of April nineteenth 1844 as 'Parc yr Brodyr (i.e. the Brothers' field) and several houses and gardens . . . constituting the whole of the north side of the village of St. Ishmaels . . . This field is the only eligible site for houses at the Ferryside'. Most of the fields shown in and around Ferryside on the Tithe Map of 1841 (which of course developed at a fast rate following the completion of the railway between Llanelli and Carmarthen in 1852) are un-named. The discovery of this place-name so suggestive of a monastic holding is fortunate particularly since post-1852 residential development spread over former fields and their names were lost. It is worth noting that in the ministers' accounts for 1388-9 in the Lordship of Llanstephan, Whitland Abbey had a half share in the ferry boat57 with the Knights of St John at Slebech. This might suggest that they had a holding close to the landing place on the east side of the River Towy in what later became Ferryside. Chapels either side of ferry crossings or indeed river crossings whether by ford and/or bridge were common in medieval Wales.58 The now lost 'gored' in the list of seine netsmens' names for the pools of the River Towy and estuary might indicate the location of 'Korred yr Abbott'. 59 If so the Whitland Abbey weir was perhaps on St. Ishmael's scar, or further north, but close to the 'lost village'.

The fourth weir is named as 'Korred Vache' also on the Towy and also held by Sir Francis Maunsell, worth 2 shillings a year but now in decay. The jurors did not know by what right Sir Francis held the weir. 'Korred Vache' may well have been much further up the River Towy and not in the foreignry at all. Two late sixteenth century deeds in the Muddlescombe collection refer to a fishing weir called 'Gored Vach' in the Towy, near 'Coed y Gwyddyl' within Idole in the parish of Llandyfaelog.60 The fifth weir was 'Korred Shilly', formerly held by Harry Vaughan, gent. and was also decayed and the rent unknown. This may be the same as 'Hyeshilly' noted above and we suggest was on the Salmon Scar.

A number of other named weirs appear in the fifteenth century Duchy accounts. In 1443-4 20s . was due from 'Shelleswere, Pitteswere and Williamswere' which formed part of the lease of Caldecote, held by the Butlers of Cwrt, Penbre. By 1488-9 they were reported to be utterly waste and decayed. It is likely that these weirs were sited near the tip of the promontory between the two Gwendraeths, to seaward of saltmarshes on the south side of the Gwendraeth Fawr, now buried well inland behind later seawalls (Figs. 2 & 9).61 In 1443-4, within St. Ishmaels, another four weirs are noted: 'Mablyswere', rent 12d ., 'Kyfteswere' valued at only 4d. and 'Herdwykeswere' at 4s. 1d. Another weir called 'Wandreth' was reported to be totally destroyed subm'gunt per cretesia aque maris— flooded by the sea. The first three weirs seem to be named after their operators and their different values must reflect their size. The uncertainties of coastal change no doubt made such ventures less attractive to tenants. Perhaps this explains the 3s. rent received for a weir 'in the water of Towy' leased in this year and the two preceding to Robert ap Gr' Vaghan. No rent was paid until the weir had proved viable over a three year period. Preceding this entry is a rent of 8d . for a place for a weir between Monkswere and the sea. This had formerly been held by the same Robert but was now leased to Meredith ap Gr' Vagan (surely a relative) for 10 years. This, the account notes, had been recorded in the court rolls for the previous year. Without checking the intervening Duchy of Lancaster accounts in the Public Record Office we do not know whether this weir was ever constructed and worked although the willingness to enter on a 10-year lease suggests a potentially lucrative operation. By the end of the century, the 1488-9 account records this and many other weirs as being 'utterly destroyed and decayed'.


Oral Evidence for the use of fish weirs on Salmon and Pastwn Scars

Studies by Geraint Jenkins on many aspects of traditional means of fishing in Wales have shown how important it is to record what practitioners themselves have to say about their craft. 62 We are very fortunate therefore to have been able to have long conversations with Mr David Phillips, the last person to work the Salmon Scar weir known to his family as Y Caj. 63 Mr Phillips worked the weir with his father who died in 1980 aged 70. His father had been brought up by his grandparents Mr and Mrs John Davies at Pant Cottage above Lookout on the coast and they had also worked the weir. Through them his own memories went back well into the nineteenth century. The cottage (SN36680800) has, unfortunately, been demolished. Only the Davies and Phillips families worked the weir and Mr Phillips gives three reasons for this. First the cottage was close enough to the Scars to allow rapid access on the tides and secondly the work was very hard, deterring others in Ferryside and Llansaint who had greater distances to travel. Thirdly and most importantly was the tradition amongst the local fishing families that once a man had staked a claim as it were to a certain area, no-one encroached upon his netting grounds. The family tradition is that a relative returning from Bloemfontein in South Africa saw the possibilities of reworking the old weirs. The large weir, Y Caj, was very successful and productive but only after it had been rebuilt by the returning Mr Billo Davies who, crucially, had the resources to employ masons to build up the stone walls. Mr Phillips told us that there was an attempt to enlarge the weir by the construction of a new wall line (see Fig.10, wall B) but this did not work and the inner curving wall of the weir was then re-used. The oral evidence thus suggests that the weir was not only in a derelict state before the re-use but that any local knowledge of how precisely it had worked had been lost.

Another anecdote, which supports the idea of a technique of fishing long-lost being revived, is a Phillips family tradition of a neighbour seeing a heap of gold sovereigns on the kitchen table at Pant, proceeds from the sale of fish. Thinking to share in this profitable resource, he hired men and transported stone and sawn timber down onto the Scar to build a new weir. It seems to have been a line of stakes set some distance from the shoreline below the Old Battery (see Fig. 11 Weir C). This was known as Polyon Ifan (Ifan's stakes – see Fig. 7). Phillips family tradition holds that it was a complete failure! The only other attempt to rebuild, or build a new cage, we were told, was on the Cwter Draw (Fig. 7) by Mr Phillips' great-grandfather, John Davies but it was not successful either. This last may be our Weirs I/J and K (Fig. 16).

Mr Phillips estimates the heyday of the weir's operations to have been in the 1880's, and the years leading up to the First World War. The availability of rail transport allowed the plentiful catches to be sold far afield and the family dealt with a Mrs Guest of Bristol. Quite possibly it was the provision of the rail link, which we know transformed Ferryside and was vital to the marketing of the cockle fishery, that led the family, aided by the cash influx from the returning Billo Davies, to rebuild the weir. With the arrival of the railway, Ferryside developed into something of a holiday resort providing accommodation for visitors and, equally importantly, a terminus for the ferry to Llanstephan. The Cliff, as H. C. Tierney put it in his local guide of 1900, housed not only 'retired commercial and manufacturing folk . . . but also professional people of good standing, and some members of county families'. Here was a ready market for fresh fish and Carmarthen and Swansea markets were easily available as well. Local sales seem to have replaced the Bristol outlet by the 1920s and 1930s.

Mr Phillips remembers his father finding no less than eleven different types of fish in one bumper catch at the weir. White fish were taken in the winter: herring, whiting and the prized Dover Sole. More humble fish were also taken. Mr Phillips remembers going down to the weir at low tide to find the encircling nets on the side walls as well as that across the caj itself white with innumerable sprats. These were put into a large enamel bucket for transport back to shore since over-handling and the wrong sort of container could turn this fragile catch into a broken-up oily mess. His father made hand nets with which sprats could be scooped up from the remaining pools within the weir.

Records of the South Wales Sea Fisheries Committee for the early twentieth century show John Davies of Pant Cottage to be one of the licensed seine netsmen working mainly from Ferryside.64 In the Towy estuary, to work the seine, one of the netsmen holds one end of net firmly on the edge of the bank of the low water channel of the river. The open boat is then rowed out into the channel and the net paid out over the stern before the boat lands, usually downstream. The net is then slowly hauled in hand over hand. Less common today are the fixed stake nets, which the Ferryside men were also licensed for. These were termed 'kettle nets' in a 1905 application but later records refer only to 'stake nets'. Here lines of wooden stakes were erected across channels and out into the river, which remained in place through the season, and to which the nets were attached. Mr David Phillips uses the term strem for such a net since it was set well out into the channel from the end of the scars. Traces of such stake lines are shown on Fig 13 (E, F, G & H). With the strem only the current held the fish against the net and catches needed to be taken by wading into the water. The agile salmon or sewin could turn and escape but flatfish were firmly caught against the net. The other net used by the Phillips family was the bwr. This was a stake net but with circular ends and could not be worked in a strong current. Mr Phillips remembers his father setting the bwr close inshore across cwter draw (Fig 7) where his great grandfather, John Davies had once attempted to build a new cage. The increasing trend of regulation to conserve fish stocks, but also to reduce the activities of the fishermen, was to prevent nets being set across channels, instead they had to be set parallel to the shore.65

Transport to and from the beach was by pony and cart but bikes were also used; the family had an open boat kept moored down on the mud away from the stony scars. The Phillips family also harvested cockles 'over on the Genst' (Ginst Point at the western mouth of the Tâf), and these might be stored temporarily in the boats. Mr Phillips has stressed the hard and constant work of maintaining the weirs. Stones needed to be piled up onto the walls of the weir and almost every spring the planks which formed the base walls of the weir and the upright stakes to which the nets were attached, had to be reset or replaced (Fig. 6 & 10). Drift wood was scrounged from the beach wherever possible, mainly from Cefn Sidan. Heavy birch planks from the seemingly inexhaustible supply in the hold of the wreck of the Paul across the Gwendraeth, opposite Pastwn Scar, were salvaged and then transported across on the bike. Mr Phillips recalls a shallow patch where the river could be crossed at low tide. Above that however there was a deep pool. This has long since vanished, part of the general silting over the whole estuary. This pool was responsible for a family tragedy when Mr Phillip's father and his two uncles were returning in their boat from Cefn Sidan where they had salvaged a barrel. There was a capsize and the two uncles were drowned.

Mr Phillips has also given us the traditional names of areas of the Scars not recorded in any written or map source known to us and has checked our mapping of these (Fig. 7). They complement the useful list of names preserved orally by the seine netsmen of Ferryside for the river bed of the Towy from Ferryside to the end of the Scars and the confluence of the Rivers Tâf and Towy.66Cwter Glas (Blue Gutter) on the north side of the Salmon Scar was so called because here blue clay was exposed. To the south areas of peat shelf, the so-called 'submerged forest' was exposed and this was called Y Colbacse by older people in Ferryside and Llansaint. The peat itself does not seem to have been dug out as happened in some other areas of peat shelves in Wales.67 Conger eels lurked beneath the peat and above the blue clay and Mr Phillips recalls how they prodded for them with spikes and then grabbed their formidable prey with a large pair of tongs. Another interesting activity down on the Scars for which we have seen no written record was duck shooting. Four circular shelters which Mr Phillips describes as being rather like igloos without a roof were built from the plentiful stone of the scars. Duck have, it seems, only a few flight paths over the Scars and the shooting hides were positioned to intercept these. One hide on Sger Rownd , was close to the shore between Lookout and the present day Holiday Village below The Battery.68 Another, Ty Saethi , was further out, on the Pastwn Scar. There was always a ready market, Mr Phillips recalls, for a brace of duck from the crach at The Cliff. But he also remembers a tragic accident involving the boys from Pen-y-Back Cottage, one of whom accidentally shot the other, the lad dying later of lead poisoning.


Evidence from place-names (Fig. 7)

The Welsh place-name element gored is common in historic forms. 69 We have discussed 'Gored Rothe', 'Cored yr Abad' or 'Korred yr Abbot', 'Kored Jany', 'Kored Shilly' and 'Korred Vache' above in the section on documentary sources. Gored is everywhere translated as 'weir', a term also used extensively in medieval and early modern sources. We have used place-names, in the absence of other evidence, to indicate the presence of lost fish weirs particularly in equating 'Cored yr Abad' or 'Monkswere' with the gored site preserved in the pool names used by Ferryside seine netsmen (see note 59). A less obvious use of 'weir' as a place-name element is in 'Wharley Point', made up of English 'weir'+ley (lçah, 'scrubland' probably here meaning 'scrub wood'), thus the 'weir by the wood' or 'wood by the weir'. It has been assimilated into Welsh as Werle (e.g. Pant y Werle). Its earliest written forms are 'werle' (1481-2) and 'wirle' (1500-01), indicating that assimilation had already taken place by the fifteenth century. Most documented gored and weir names have a personal-named suffix. A few are termed 'new' weirs; 'Broad weir' records the form of the weir. The Llanstephan 'Ebbweir' (variously 'Edwere', Embwer' and William Morris's 'Cerreg Ebwear') records the all-important fact that these weirs were worked on the ebb tides. 'Salmon were' and Salmon Scar alone record the most valuable fish in the weir's catch. To this we may add 'Hyeshilly' and 'Kored Shilly', a weir, which we suggest might also have been located on Salmon Scar. One element possibly derives from sil—'small fry' or sili 'salmon fry' and thus may indicate a specialised weir for catching young salmon, which run down to the open sea in springtime. This element may also be present in the 'Shellwear' within the Gwendraeth recorded in 1488-9. 70

For the remains of the weirs on the Scars today, caj (loaned from English 'cage') is used. Neither gored nor weir has been mentioned orally, although we are confident that some of the gored -named weirs in the Duchy of Lancaster Survey of 1609-13 must have been on the Scars. This apparent dropping of gored could be used as supporting evidence to argue for discontinuity in the use of the weirs themselves sometime after the seventeenth century until their certain reuse in the nineteenth century.

Fig. 8. A tree-stump within the peat shelf near Pastwn Scar.



The place-names used in Fig. 7 were provided by David Phillips and their meanings are for the most part self-explanatory. We are grateful for his suggestions for some of their meanings, polyon refers to the poles or stakes set into the stone walls of the weir to hold nets. Cwter is 'gutter'- the low water streams. The meaning of 'Cwr' is not certain. If it is used in the sense of 'extremity, limit' then, Mr Phillips suggested to us that, 'Cwr Glas' could refer to the 'limit' or 'extremity' of the outcrop of blue clay that underlies the submerged forest layers. This clay was collected by inter alia the local cockle women as a component of pele mân,71 so the locations of the clay would have been important. It may be coincidental that both examples of cwr are also where there are low water ponds. It is interesting to see the use of lloc in Lloc y llong here perhaps meaning a corner or pool of the ship[wreck] rather than an embankment of the ship[wreck], but the latter is of course possible. Pwll yr Ychen, we are told, is a low tide fresh-water spring, at which cattle would water. David Phillips recalls family accounts of fresh water being used from there when sources inland of the Burrows were dry. It was a source of water for those living in Lookout Cottage. YColbacse was used to describe the area of 'submerged forest' deposits and its meaning is unknown, as is Y Cotsier (pronounced 'cotcher'). The sger (scar) place-name element is in origin Viking referring to a half-tide rock or a stack in the inter-tidal zone. There is a 'Sker Rock' off Tenby as well as 'Goskar' on North Beach. It appears to have been adapted locally to refer to the stoney half-tide banks. In the Tâf is 'Black Scar'. On modern OS maps we have 'Pastoun' Scar, but it is probably Welsh pastwn, recalling the club-shaped plan form of the scar. Sger Rownd is another example of the shape of the scar being used as a descriptor in the name. We have no forms earlier than 1891 for 'Salmon Point Scar', but there is a reference to Salmonwere in 1488-9, which with Broadwere was valued at 13s. 4d. Jointly these were the two most valuable weirs and despite the long gap without mention of the 'Salmon' place-name element, we feel this weir must have been on the Salmon Scar and gave it its name. The area known as Y Mine has the remainsof a wartime mine (floating explosive) still to be seen. Of the four shooting hides that were once present, only Ty Saethi ('shooting hide') has a surviving name.



For the sake of clarity we have left the archaeological descriptions of the weirs to an appendix which follows. This inevitably has meant that we have introduced conclusions into the narrative before presenting all of the evidence. However we feel that this was necessary to keep the historical and oral evidence and conclusions free from technical archaeological description which may have confused the reader.

Documentary evidence shows that they were fish weirs established by the thirteenth century. Some are probably much earlier in origin. By the early sixteenth century, when the documentation starts to dry up, it would appear that most weirs were returning no income. Some if not all had become ruinous. We believe that this was a result of a number of factors, principallyit was coastal change that made most untennable; economic factors also played a part. The growth of the dune barriers on the Pendine and Cefn Sidan sides of the three rivers had particularly changed the Gwendraeth side causing besandment of scars near the confluence of the two Gwendraeths. This resulted in the abandonment of a number of weirs.There was also besandment of part of the Salmon and possibly Pastwn scars and of the formerly profitable weirs below Llanstephan castle. Discontinuity between the seventeenth and nineteenth centuries is not certain, but the apparent absence of any documentary references, coupled with evidence for coastal change, seems to allow for this. We have only oral evidence to show that weirs were once again being reused, or new built, in the third quarter of the nineteenth century. We believe that we can identify these nineteenth century works with the most important documented medieval weirs (Salmon weir on Salmon Scar, Broad Weir on Pastwn Scar, and a lost weir, Gored yr Abad, which was probably on a much more extensive St Ishmael's Scar. With regard to the 'lost village' we believed that the quality of quarried stone work of its buildings suggest that what we see must be the work of either a temporal or spiritual Lord. Whitland Abbey would need to have buildings to process its fish from Gored yr Abad, and there is therefore a good possibility that these remains are associated with its sparsely documented fishery on the Towy.




Outline Descriptions: Fish-weirs on the Salmon and Pastwn Scars—topography, geology and archaeology .

To understand the nature and possible dating of these weirs, it is essential to first consider their setting. The weirs discussed in this section are located on and between two extensive natural stony banks known as Salmon and Pastwn Scars that uncover about two hours after high water. These extend from the high water mark to seaward roughly in a south-westerly direction for about 1.2km and from NW-SE about 1.4km. Between the two there is an extensive, lower, area of sand (0.3 km 2) covering submerged forest levels of up to half a metre thick which in turn sit on a blue clay deposit, and area known locally as Y Colbacse. Where this peat shelf has been eroded away, areas of blue clay are exposed or sit directly below sand. In the eastern low water bank of the River Towy, which in 2003 was running fairly close to the Salmon Scar, we observed an unbroken deposit of blue clay, which suggests that the Towy has never run further east than its present course. In much earlier epochs both Gwendraeth rivers ran directly south into the sea but with the north-westward migration of the dune barrier of Cefn Sidan the Gwendraeth has been forced ever westward as the cartographic record described above makes clear (compare Figs. 2, 3 4 and 9). However, the Scars have prevented the Gwendraeth from moving any further north-west, and for at least the last 200 years the river has often run hard against the tip of the Salmon and Pastwn Scars, which was not the case in the Middle Ages (Fig 9). The two Scars are composed of very compact stone, but when the mussel beds upon them are occasionally stripped, the body of the scars can be seen to comprise compact stone in a clay matrix. This clay matrix supports our contention that the Scars are remnants of lateral moraine deposited as the glaciers receded at the end of the last Ice Age. Sea levels rose steadily but the sea level was still low enough some five to six thousand years ago for oakwood forests to develop whose remains can be seen in the peat shelves (Fig. 8).72

It would appear that the remains of the medieval buildings (the 'Lost Village') that are eroding out of the mean HW spring level under the dunes sit on a fairly stone-free subsoil which in turn gives way to boulder clay (Fig. 1). It is important to consider the Scars and their weirs in context with the so-called lost village (Fig. 5). The walls and floors of this settlement are sealed directly by sand dune deposits, and it would appear that a catastrophic storm or storms must have been responsible for the abandonment of the buildings, when they were besanded. We must try to picture the topography that existed before the sand dunes were created. Since the buildings are on just above or on today's MHWS73 line, even with the gradual rise in sea level since the Middle Ages, they must have always been potentially susceptible to flooding. The building remains that we now see appear to be very well constructed, of well-cut quarried sandstone, 74 clearly built to last. We therefore feel the builders would have only placed their settlement at this point if a natural defence or a sea wall protected them. Alternatively the site must have been sufficiently inland of the sea to be safe from inundation. What ever the case we must entertain the view that lost portions of these buildings were sited on land now given up to the sea. How far out this tongue of land extended is of course almost impossible to estimate. We have noted, however, that Saxton's map shows a more pointed 'St Ishmael's Point' than now exists. Supposedly there was a well below The Battery. Tierney, in his local guide of 1900, refers to a green field which remained dry during ordinary tides on the west, that is the seaward side of the railway 'near St Ishmael's, 'within living memory'. A very narrow fragment (a strip, named 'Parc y Burrows' on the tithe map), still exists. But this field is well above sea level and can therefore be dismissed for our purposes. Professor Stephens of Holcwm referred in 1944 to an elderly Llansaint cocklewoman who told him that her forebears milked cows below Lookout. 75 Today the cage of the most substantial weir on Salmon Scar (Weir A, Y Caj) is 650m from MHWS (Fig. 11). It is evident that where our present weirs survive (which we have identified as medieval in origin), their location has always been within the inter-tidal zone, proving that land-loss has not been great in historic times.

The weir structures are infinitely variable. They principally consist of one or more wall type: fairly simple lines of stakes bedded in low stone banks, or substantial stone-embanked walls, and interior training walls leading to cages. The complexity of some of the features, combined with the irregular appearance under shifting sand and mussel beds makes recording, and thus interpretation, difficult. There is never a complete or indeed a clear picture of their plans or extent, let alone the dates or periods from which they were built and worked. In the early years our fieldwork was hampered by an inability to record the position of features accurately. It was only with the use of GPS76 from the early 1990s that we have been able to map features with some accuracy, and it is with hindsight that we realise the importance of recording discrete and disparate areas of stonework even when they appear meaningless. When fragments are strung together over time identification of new weirs or new elements of known weirs has been possible.

Between the Salmon and Pastwn Scars is a natural hollow of about 33 ha. (74.1 acres) into which fish would naturally move as the tide receded to expose the higher scars. It is self evident that the most long-lived and profitable weirs made best use of this natural lie of the intertidal zone with its stony banks. All that was needed was to create a complete barrier and here the builders resorted to the use of timber and wattle walls and later nets set between timber posts or stakes. In no case have we found weirs that use just one form of walling (with the exception of Weir L described below). Indeed a common element is the use of the scars themselves to form one or more sides of the catchment and then resort to stone and timber across the lower ground off the scars. Given the nature of sand movement, their overall extent can only be guessed.

Mr David Phillips, whose information on how his family worked the main weir has been detailed above, sees considerable changes in the Scars since the 1960s. Like may other fishermen and users of the rivers, he believes that the construction of the Llyn Brianne Dam at the headwaters of the River Towy, completed in 1966, and the abstraction of water from the river above Carmarthen, led to reduced flows at the estuary mouth and consequently more silting. The natural basin of the weirs between the Salmon and Pastwn Scars is now besanded and mud-filled. The Scars themselves are flatter too, because the weir walls are no longer built up each spring. When his father operated the weir the top of the scar and the walls began to emerge after only one hour of ebb on the spring tides.


Salmon Scar features (Fig. 11)

Salmon Scar, WeirA 'Y Caj' (cage, SN35870716, Figs. 10A & B). The Davies/Phillips family worked the main weir on Salmon Scar well into the twentieth century. There must have always been a weir in this location because of its natural advantages. This is our preferred location for the historic 'Salmon Weir' of 1488-9. It was sited to make the best use of a great hollow in the scar, enclosing an area 2 hectares (4.9 acres) with an approximate NE-SW length of 270m. Where the natural scar is at its highest, there was no need to construct any walls, but towards the seaward end, approaching the cage, there are low stone banks and timber uprights set in these, with horizontal timbering at ground level to complete the timber part of the barrier. Nets were strung between the uprights (Figs. 6 & 10). Within the catchment area there was a stone-lined channel, which at low tide still contained water, where trapped fish congregated to be driven into the cage (Fig. 10). In addition to this channel there are two other 'secondary' low walls. Both curve in an arc in the opposite direction to the seaward line of the weir itself (Fig. 11). It is difficult to see these operating as a separate, smaller weir, so perhaps they were used to regulate the entrapment and collection of catches at different tidal levels or belonging to a much earlier phase. There is a late addition consisting of two double lines, of sawn 2 x 3 inch stakes, c. 40m long aligned NE-SW, which were inserted as an experiment for oyster or mussel farming in 1991 (Fig 10, arrowed *).

Salmon Scar, Wall B (centred SN35940714). This is a secondary substantial wall on Salmon Scar adjoining on the east side of Weir A extending for about 143 metres. This massive stone bank with coursed stonework in its lower levels, at least 3m wide at its base, running NW-SE. This wall line starts from near the cage mouth of Weir A, running ESE (initially as timber uprights and horizontal planking set in a foundation of stone grounders. After 28m the wall line changes into a massive bank (as already described) which is set on the natural scar. As the scar itself ends 61m further ESE there may have been a change back to timber, but at this junction there is also the remains of a shipwreck (Fig 15).

Salmon Scar, Weirs C/D 'Polyon Ifan' (centred SN36070718). Fig 11 & 12. This is probably one weir sited closer to the land on the eastern flank of the scar. C is a curving low stone line on the scar itself, which then changes to a post line where the scar finishes. In total about 55m combined length has been observed. From the scar (and where the cage is presumed to lie) the weir continues as a good stone bank from SN3612007175 to SN3613407154 and changes alignment, then has been recorded for another 70m to SN3618907154 (Fig. 12). The weir has been identified by David Phillips as Polyon Ifan , which name recalls the owner and the stakes that were set along the course of the wall. A separate weir consists of a low stone spur set on a natural raft of fossilised tree trunks.

Lesser weirs on Salmon Scar (centred SN355069) Figs. 13 & 14. To seaward of the main weirs we have recorded four short lengths of post and stone wall lines (E, F, G, H). They are all located within the same area on the east side of the scar, E-G are 70m apart and H 100m furtherfrom G. All start as low stone banks with upright posts on the scar that change to timber once off the scar. These are interpreted as the remains of strem nets.

The Boat Passage or Canal (Fig. 11).The sides of this straight, narrow channel across the Salmon Scar appear revetted but this may be simply the result of careful packing of stones cleared from the channel in the early nineteenth century (see above). Mr David Phillips recalls his father setting a net across the north-western end of the channel because he thought that fish could escape via the 'Canal' from his weir (Y Caj).

Fig. 15. The wreck of a timber shipwreck at the end of weir B looking N.


Pastwn Scar Features (Fig. 16)

Pastwn Scar, Weir I (centred SN36150682). The most impressive remains on Pastwn Scar consists of about 83m of an E-W stone bank on the scar. The bank becomes more massive as it moves west off the scar running a further 30m. It tails off near where the low water channel (Cwter Draw ) crosses, and although now physical evidence has been recorded, it is assumed that a cage was located hereabouts.

Pastwn Scar Weir J (cage SN3610606844). Weir J is to the west of Weir I. It could be part of Weir I, but has been given a separate letter because of the compelling evidence for separate cages. The evidence is very fragmentary because features are usually obscured by sand. The most interesting feature is a pair of low NE-SW aligned stone banks that form the approach to the cage. Both walls have evidence for upright posts, so these features must have been constructed of horizontal timbers between uprights, forming a most impressive timber cage. The stonework in the 'cage' area sits directly on the peat shelf ( Fig. 17), which, we assume, must prove that the stonework here is wholly artificial and not part of the natural scar. The approach stonework has been recorded running inland for about 53 metres. It should be noted that the two parallel walls leading to the cage align with two banks 200m further to the NE (Weir M). Both I and J are set across the line of Cwter Draw , where David Phillips' great-grandfather apparently constructed a caj. The family tradition is that this was never profitable, and we cannot tell if an earlier weir or weirs were reused, but this is our preferred site for the medieval Broad Weir (Fig. 5).

Pastwn Scar Weir K (SN36180683). The southern E-W wall is shared with I, but Weir K has a perpendicular stone bank running north making it (with Weir I) 'v' shaped. It would have shared much of the same catchment at Weir I and J and may be structurally later. We assume that there was a stake wall line running north as K left the scar, with a cage at this location where the low water channel (Cwter Draw) runs.

Weir L (SN3610306822). This may be a free-standing weir originally unrelated to any others. At a subsequent stage it was linked by a post wall line to Weir J (Fig. 18). Weir L is the least substantial of the weirs, and the only one that does not incorporate part of a scar. It consists today of stumps of stakes set in a narrow curving trench running some 80m cut into the peat shelf subsoil, which has been backfilled and packed with stone. Spade cuts for the trench can still be seen. The trench curves with its apex to seaward, but there is no noticeable cage structure, which suggests that the stonework could be the emplacement for a net of the bwr type. On its western side, the northern continuation appeared to just consist of stakes when recorded in 2001.

Possible Weir M . (Fig. 16) Located west of Pastwn Scar, but probably joined to it. Consists of a roughly NE-SW aligned pair of parallel cambered banks of spread stone .5m high and c. 3m wide with a possible break for the site for a cage at SN3625906971 (Fig. 19) It is possible that these two banks linked with Weir J.

Possible weirs, fragments N and O. These consist of short stretches of stone banking sitting on the peat shelf which suggests they are not natural. N is between 0.5 and 3m wide, about 20m long roughly NS (SN3600506834 to SN3600506820). There is a smaller 5m long spread to the NW at (SN3599606847). Some 30+m to the NE is a near parallel bank of stone 0.4-1.5m wide and 34m long aligned NNW-SSE. Some of the stone appears to sit in a cut into the peat shelf. Both of these banks could form weirs that curve to join Pastwn Scar to the east. There is undoubtedly more hidden evidence for weirs covered by sand at this location.

Linear marks worn into peat shelf (centred SN35990686) Fig. 20 On the NE side of Weir L there is an extensive surviving horizon of peat shelf. Cut, or rather eroded, into this are linear marks running roughly from the land down to the low water Gwendraeth, aligned roughly NE to SW. Modern wheeled vehicles must have made some of these marks, because they appear as parallel lines about the width of a Land Rover. However some of these marks are more irregular, single furrows or are parallels far narrower than a modern vehicle. We suggest that some of these marks could be sledge lines. There is however no record of sledges being used by cockle gatherers or by the Davies/Phillips family working weirs. Mr Phillips suggests that some of these marks could be grooves worn by the keels of the fishing boats brought up between the Scars on the ebb.



We would particularly like to thank Mr and Mrs David Phillips for discussing the oral traditions about the use of the fish weirs on the Scars and for permission to reproduce Fig. 6.



1 Royal Commission on Ancient and Historical Monuments in Wales, Inventory for the County of Carmarthen, 1917, pp. 245-246. The Royal Commission's computerised record number is NPRN 15229. In 1995 an exposed section was cleaned down and recorded by Cambria Archaeology. The 'lost village' is numbered PRN (primary record number) 2113, in the County Sites and Monuments Record (SMR) maintained at Cambria's Llandeilo offices, which can be consulted by appointment, phone 01558 823131. Finds from the site are held by Carmarthenshire County Museum, Abergwili.

2 For the traditional view see H. C. Tierney's Guide to Ferryside 1900, published in Carmarthen by The Welshman.

3 Troed y Fran produced by Rebeca Ltd. 1984, director Phil Edwards, Historical Adviser Terrence James.

4 GPS—global positioning system, originally designed for US military navigation, subsequently developed for civil aviation and marine use, then surveying and vehicle tracking The early maps and charts, as well as recent air photographs, were rectified to modern co-ordinates using Aerial 5 software and these became part of a Geographical Information System; this was used to build up survey data for comparison with earlier coastline information. It is proposed to deposit copies of the data at Cambria Archaeology and the National Monuments Record.

5 Many of the early maps and charts and much of the evidence mentioned in this article are reproduced in Terrence James, 'Where Sea meets Land: The changing Carmarthenshire coastline' in (ed. H. .James) Sir Gâr: Studies in Carmarthenshire History 1991, Carmarthenshire Antiquarian Society Monograph Series Vol. 4. pp. 143-166. Readers are strongly recommended to have this article to hand for cross reference whilst reading this paper.

6 J. Smith and N. Yonow The Coast of Dyfed and South West Glamorgan: An Environmental Appraisal Field Studies Council 1995 pp. 53-7.

7 M. Smylie The Herring Fishers of Wales Welsh Heritage Series no. 6, 1998 Carreg Gwalch, esp. Ch. 1.

8 The 1805 'Act for the Preservation of Salmon and other Fish in the Rivers in the County of Carmarthen' cited a predecessor Act of 1285. The Salmon Fishery Acts of 1861 and 1865 sought to modify existing fish weirs in rivers and forbid new ones. The South Wales Sea Fisheries Committee was founded in 1890 and is based in Swansea; its rôle is the enforcement of National and European fisheries legislation. It works with other bodies to provide cost effective and sustainable management of fisheries and the marine environment. Its well-catalogued archives are held at the West Glamorgan Record Office (D/D SWSF). The current position is set out in the latest (Jan. 2002) Byelaws of the South Wales Sea Fisheries Committee made under Section 5 of the Sea Fisheries Regulation Act of 1966 and the Salmon Act of 1986. Closed seasons, net sizes, their lengths and location, minimum sizes of catch and protected species are all specified.

9 Essential reading on this and many other fisheries are the works of J. Geraint Jenkins: Nets and Coracles 1974, The Coracle 1988 and The Inshore Fishermen of Wales 1991.

10 We take strem to be a Welsh loan word for stream, i.e. tidal or river current; these nets were set on their lines of stakes out into the river channels for maximum effectiveness.

11 We make no distinction between the terms 'fish trap', 'fish weir' or their Welsh equivalent 'gored', which in this article can be accepted to mean much the same thing. See also M. Richards 'Some Fishing Terms in Welsh Place-Names' Folklife 1974 pp 9-19.

12 For a recent overview of the Welsh evidence see R. Turner 'Fishweirs and Fishtraps' in (ed. A. Davidson) The coastal archaeology of Wales CBA Research Report 131 2002, pp. 95-107.

13 S. Godbold & R. C. Turner 'Medieval Fishtraps in the Severn Estuary' Medieval Archaeology XXXVII 1994, pp.19-54

14 Q. Kay & M. Davies 'A medieval fish weir on the beach at Whitepool Point, Gower' Gower 44 1993, pp 6-13.

15 N. Page 'The Llanelli and Loughor Wetlands: an archaeological assessment of the northern shore of the Burry Inlet and the lower reaches of the Loughor Estuary' unpublished report prepared by Cambria Archaeology in 1997 for Cadw-Welsh Historic Monuments, esp. p. 14. This can be consulted in the SMR, for access see n. 1. There is an 1870 plan of Machynis Weir a 'privileged engine' in CRO Trant Unlisted, Carm. Fishery Board.

16 For Cistercian monasteries see David H. Williams Atlas of Cistercian Lands in Wales 1990 and his The Welsh Cistercians vols. I & II 1984.

17 W. Davies The Llandaff Charters 1979, p.123 charter no. 235b; for discussion of other fishing rights see W. Davies An Early Welsh Microcosm 1978 and for the original text of the charter and the early Welsh for the weirs given by King Brochmail ( cumcoretibus suis) see J. G. Evans The Text of the Book of Llan Dâv 1893 p. 236.

18 E. Lewis 'The Goredi near Llanddewi Aberarth' Arch. Camb. 1924 pp. 395-8.For an air photograph, see Davidson, op. cit. note. 12, p. 100.

19 Cecil Jones 'Walls in the sea—the goradau of Menai: some marine antiquities of the Menai Straits' International Journal of Nautical Archaeology and Underwater Exploration 1983 12.1, pp 27-40.

20 D. Huw Owen Mapiau Printiedig Cynnar O Cymru/ Early Printed Maps of Wales 1996 National Library of Wales booklet. S. Tyacke and J. Huddy Christopher Saxton and Tudor map-making 1980 British Library Series No. 2.

21 J. W. W. Stephens 'The Identity of Alkenchurch and Llansaint' TCASFC 29, 1939pp.105-114 citing B. G. Charles 'The Placename Halkenchurch' London Medieval Studies 1938 vol. 1 pt.2; Mary Curtis, The Antiquities of Laugharne, Pendine and their Neighbourhood, 1880, p.256.

22 See T A. James, op. cit. n. 5.

23 Murdoch Mackenzie and his nephew also of the same name conducted numerous coastal surveys of Great Britain for the Admiralty; the Bristol Channel chart was the work of the nephew. Heavy Admiralty demands and lack of resources often meant that the technical innovations pioneered by Mackenzie had to be sacrificed to speed of production.

24 Reproduced in Terrence James, op. cit. n. 5 p. 148.

25 R. Thorne 'Y Bertwn, St Ishmael: an Abandoned Farm' Carms. Antiq. XXXV 1999, pp.30-34; see also Philip Evans' useful unpublished 'Tanylan—the story of a Carmarthenshire Farm' 1997, deposited at Carmarthenshire Record Office and Carmarthen Reference Library.

26 CRO John Francis Collection, Maps and Plans, 41 for 'Kidwelly Bar and Harbour' surveyed by John Wedge in 1807; for a reproduction of Wedge's 1814 chart and the texts of John Rennie and Edward Bankes's 1820 Report on the restoration of Kidwelly harbour see W. H. Morris in Carms.Antiq. XXIV 1988,pp.75-81.

27 There are several references to groundings and shipwrecks in the Index to the Carmarthen Journal in Carmarthen Reference Library, notably from July 22 1814 p. 3(4) 'On Tuesday evening last the sloop Friends Goodwill of Laugharne, Edward Hancock Master, being on passage from Kidwelly to that place, heavily laden with coals, sank off St. Ishmael's Church …'.

28 G. T. Rees Guide to Llansteffan n.d. gives a list & description of the boats. A post survived marking the channel into the early 1990s.

29 These give detailed descriptions of coastal landmarks and directions for crossing Carmarthen Bar. The chart was reissued in 1857 with additions (buoyage and a compass rose) but no changes to the channels or intertidal features.

30 The positions of these two barrel posts are maintained for navigation today by successor port and starboard hand posts.

31 The computer rectification of all the several charts proved conclusively that the barrel post depicted throughout had been on the Pastwn scar.

32 William Rees South Wales and the March 1284-1415. 1924p. 198.

33 William Rees A History of the Order of St.John of Jerusalem in Wales and on the Welsh Border 1947. In 1331 Henry de Gower, Bishop of St David's granted lands, presumably in the area, to the chaplains of St Mary's, Swansea but retained a holding, two weirs and a ferry across the 'Tawe', 'Carmarthenshire'. This is ambiguous and is probably Swansea. Francis Green transcripts of Inquisitions Ad Quod Damnum and Patent Rolls, Green MSS vol XII, Haverfordwest Reference Library. The term 'fishery' almost certainly implies an 'engine' for catching fish, i.e. a fish weir, rather than 'hunting' with nets; see Harold Fox The Evolution of the Fishing Village: Landscape and Society along the South Devon Coast 1086-1550 Leicester Explorations in Local History 2001.

34 In 1906 Sir John Williams, Bart., of Plas Llanstephan, commissioned Dr E A Lewis to compile transcripts of medieval records down to the end of Henry VIII's reign relating to the Castle and Lordship of Llanstephan. These, with additions, were edited and published by Francis Green in vols. XII and XIII of West Wales Historical Records (hence WWHR) 1927 & 1928.

35 WWHR XII p.119.

36 WWHR XIII, pp. 43-54.

37 WWHR XIII, pp. 54-60.

38 WWHR XIII, p. 72.

39 WWHR XIII, p.93

40 E. A. Lewis & J. Conway Davies (eds) Records of the Court of Augmentations relating to Wales and Monmouthshire. Board of Celtic Studies, University of Wales, History and Law Series no. XIII 1954, p. 267.

41 Carmarthenshire Record Office (CRO) Cawdor Vaughan (CV) VI 21/579.

42 CRO CV 21/588

43 CRO CV VI 21/580

44 CRO CVVI 21/585-6

45 F. Jones 'Pembrey Court: An Old Carmarthenshire Manor House' Carms. Antiq. XIX 1983 pp. 17-31.

46 CRO CV VI 21/588

47 PRO DL 43/12/14, quoted by P. Evans with a reproduction of part of the MSS, op. cit. n. 25, p. 51.

48 W. Rees A Survey of the Duchy of Lancaster Lordships in Wales 1609-1613 1953 Board of Celtic Studies, University of Wales, History and Law Series no. XII, Lordship of Kidwelly, pp. 174-303.

49 Idem, p. 217

50 'The foreignry' refers to land outside the 'Englishry' of the lordship and can thus be equated with the 'Welshries' where Welsh customary law and tenures continued after the Norman conquests. The whole of St Ishmaels parish was within the foreignry of the Lordship of Kidwelly.

51 Whitland Abbey put up a spirited resistance to the looming threat of the Dissolution of the Monasteries. Neath, Strata Florida and Whitland escaped suppression until 1539 but only on payment of heavy sums—Ł400 in Whitland's case. Agreeing long leases of their possessions was another attempt to remove their patrimony from the king's grasp. But it failed; the King, as Duke of Lancaster, clearly got hold of Whitland's weir and lands within the Lordship of Kidwelly by 1609.See Williams op. cit.note 16, Vol. I. Section A, 6 'The Suppression'..

52 Syddyn from Tyddyn, originally 'building, house' later 'small farm, tenement' In Thomas Jones' Y Gymraeg yn ei Disgleirdeb Neu helaeth EIR-LYFR / The British language in its Lustre, Or a Copious Dictionary of Welsh and English (facsimile reprint 1977 Black Pig Press, Llanwrda) gives Syddyn as a tenement; Syddyn seems widespread in seventeenth century Carmarthenshire deeds. The difference between the holding of 8 acres and the total of 20 acres held by David ap Ieuan Gwillim may be due to Whitland Abbey owning the tenement with its chapel and remaining acres still in scattered strips in the open fields

53 For a summary of his life and other members of the Dwnn family see R. A. Griffiths The Principality of Wales in the Later Middle Ages: The Structure and Personnel of Government I South Wales 1277-1536. BBCS History & Law series 26, Cardiff 1972, pp. 187-188.

54 NLW Muddlescombe Deeds no. 1255, p.202 in Schedule

55 D. H. Williams The Welsh Cistercians 1984 vol 2, p. 203.

56 'Smalls' is very likely derived from St Ishmael's, in the same way that the rocky islet off the Pembrokeshire coast has got its name, W. A. R. Richardson 'The Smalls, Hats and Barrels: Navigational and Toponymic Hazards' Nomina 17 (1994), pp. 71-97.

57 WWHR XII, p. 124.

58 Some local examples include Llanstephan itself, which the Knights at Slebech held with the ferry boat; the lost chapel of St Hernin at Pontargothi (PRN726), the ysbyty place-name close to Laugharne ferry on the Llanstephan side, and the ysbyty place-name at Yspytti, Llanelli near the Loughor crossing (the Knights at Slebech held Loughor church and a burgage there). There was a chapel on the old bridge at Haverfordwest.

59 Names given with their closest landward locations extending from north of Ferryside down to the estuary in D. Gerald Jones's local guide Introducing Ferryside n.d are: Alma Bank, The Bay, Daboch, Parrog Bank, Barrel Post, Gored, Undercastle, Ebware, Old no. 9, Cwch, Main, Cefn.

60 NLW Muddlescombe Deeds no 1566, p. 132 in Schedule. It is probably the 'Gwar bach y Gored' consisting of poles driven into the ground, joined by woven twigs and facing the tide in the shape of a 'V' noted by Ethel M. Davies The Story of Llandyfaelog Parish Carmarthen 1953. Following the 1861 Salmon Fishery Act (see n. 8) existing weirs needed licences to continue as 'privileged engines'; A certificate and plan of Plas Gwyn Weir on the River Towy of 1870, in the Carmarthen Fisheries Records in the Trant collection (CRO uncat) may be the same weir.

61 See Terrence James, op. cit. n. 5 esp. Fig. 6 p.155.

62 See n. 9.

63 Mr Phillips now lives at Cross Inn, Laugharne.

64 West Glamorgan Record Office DD/SWSF 41/2, Correspondence re Kettle and Stake Nets in the River Towy. H13687 9 Aug 1905 lists 13 fishermen given consents, including Mr John Davies, Pant Cottage, St Ishmael's parish, Ferryside. H14787 lists him in 1915 with another 11 fishermen, mainly from Ferryside, permitted to use stake nets in the Towy.

65 See n. 8.

66 See n. 60.

67 Intertidal peat deposits at Tywyn, Gwynedd were cut for fuel in the post-medieval period, see Fig 4.2 in A. Davidson, op. cit. n. 10. Unfortunately, despite designation of the Tâf, Towy and Gwendraeth estuaries as a Special Area of Conservation, bait digging and the removal of peat for worm farming are accelerating the natural erosion of this priceless paleoenvironmental resource.

68 The Gun Battery, late nineteenth century, is now largely destroyed, only parts surviving in the eroding clay cliff.

69 Melville Richard 'Some Fishing Terms in Welsh Place-Names', Folk Life, 1975, pp. 9-19.

70 For a discussion of sil /sili see Richards, op. cit. note 11.

71 This practice of mixing coal dust with clay to make balls for economic, slow burning fuel on the hearth, was once widespread across south west Wales.

72 See Terrence James op. cit. note 5 p. 144; also 'St. Ishmael's: Submerged Forest' TCASFC 23 1932 p. 56, sample identified as oak by A. Hyde NMW. There are unfortunately no radio-carbon dates for these peat shelves, but the corresponding peat layers on Marros Beach have been dated to around 5,000 b.p., Terrence James op. cit., note 5, p.144.

73 Mean High Water Springs. This is the coastline depicted on nautical charts, and represents the highest point to which spring tides rise (excepting storm surges). This is not the same as the coastline depicted on Ordnance Survey maps which uses an average sea height – Mean Sea Level (MSL) which equates to Ordnance Datum.

74 The red and green marl beds of the Old Red Sandstone cliffs across the river below Llanstephan Castle are the closest source.

75 Cited by Philip Evans from Stephens family papers op. cit.n. 25.

76 For GPS see note 4.



Fig. 2. The present coastline at the estuaries of the Three Rivers. Compare the growth of salt marsh and sand dunes with Fig. 3.

Fig. 4. An extract from Henry Denham’s chart, based a survey data of 1828, showing the Scars as they then appeared with the Gwendraeth not running into the Towy at low water. The dotted line represents the present coastline.


Fig. 6. Y Caj—the last weir worked by the Phillips family in the early 1960s. The wreck of the Paul appears as a dark feature in the distance (left).


Fig. 7. Place-
names from oral evidence in the area of Salmon and Pastwn scars.


Fig. 9. Conjectured coastline in the Middle Ages showing areas of known scars lost to subsequent coastal growth. Some of the weirs of the Duchy of Lancaster were accounted with the Manor of Caldecot, which might mean that there were once weirs located on the Caldecot side of the Gwendraeth.


Fig. 10b. Y Caj viewed looking inland with recent stakes for mussel farming in the distance.


Fig. 12. Remains of the wall of Polyion Ifan running towards Salmon Scar looking NW.


Fig. 17. Parallel lines of ‘leader hedges’ of Weir J. One of the cage’s timber stakes is arrowed. Darker stones can be seen sitting on the peat shelf. Viewed looking N.


Fig. 19. Two spread stoney banks of possible Weir M.