The Glamorgan-Gwent Archaeological Trust Ltd.


Historical Processes, Themes and Background

Historic Administrative Divisions

The historic landscape is situated within the post-medieval County of Glamorgan, which prior to the Norman invasion appears to have been a border region, at various times part of the Kingdom of Dyfed, and at other times held by the early Kingdom of Glywysing, named after an eponymous early King, Glywys; during the 10th century Gower became part of Morgannwg, named after its ruler Morgan (Morcan) Hen (c. 930-74), later Glamorgan (Knight 1995). According to tradition Glywysing or Morgannwg was divided into seven administrative regions or cantrefi, while 12th century sources assert that these were named after the sons of Glywys. Traditionally each cantref was divided into cwmwd, or commotes, each comprising estates or maenorau made up of a number of trefi, or townships.

During the early medieval period Gower formed Welsh medieval Cwmwd of Gwyr, within the Cantref of Eginog; peninsula Gower appears to have formed an area known as Gower iscoed, while the uplands were termed Gower uwchcoed political the upland. The Normans were later to use the latin terms Gower subboscus and Gower supraboscus, for this division, though the term subboscus appears to have become restricted to and synonimous with the area of Llanrhidian Higher, where Welsh cultural identity remained strong (Cooper 1986 and 1998).

The existence of a large early medieval secular estate or 'maenor' based on the lands of the later medieval manor of Landimor, with its church site at Cheriton, has also been conjectured. This large estate or 'maenor', is later visible as several scattered later manors or fees, the much-reduced remnants created following dismemberment under Anglo-Norman control; it is thought that Payn de Turbeville inherited this larger Welsh unit, and that this larger holding was broken up as a result of grants to religious orders, so that by the twelfth century the manor retained only the dispersed sub-manors of Rhossili, Landimor and Llanrhidian (Draisey 2002; Cooper 1998). The surviving area of Landimor itself, and others such as Llanrhidian, may represent the territory formerly associated with one, or more of the former maenor's townships. Moreover, it is possible that the boundaries of the later medieval manor of Llandimor may represent the continuity of a fairly ancient land division; perhaps supported by the fact that the boundaries of the later manor and the parish of Cheriton, make a purposeful extension to include the hillfort of the Bulwarks at the eastern end of Llanmadoc Hill, a contender along with the major hillfort of Cil Ifor, near Llanrhidian, for the origin of the name Landimor. As yet the exact nature of the administrative, ecclesiastical and settlement landscape of the area during the early medieval period is at best speculative and would benefit from further in-depth research and study.

During the medieval period the area of peninsula Gower and the AONB formed part of the medieval lordship of Gower, with included the area currently taken in by the unitary authority of Swansea as well as the parish of Llangiwg and the Rhwngydwyclydach division of the parish of Llansamlet (Morris 2000, 3). A fundamental characteristic of the lordship was its division into the Englishry and Welshry, coming to represent respectively over time the lowland (peninsula Gower or Gower Subboscus), in which Anglo-Norman influence was early and eventually gained predominance, and the upland (Gower Supraboscus) which remained largely Welsh in cultural and, and where Welsh law and custom applied. The latter part of Gower lies exclusively beyond the boundaries of the AONB. The process by which Anglo-Norman control of peninsula Gower was achieved appears to have been far more gradual than has often been previously implied, a fact masked by the political and territorial claims of the de Breos's during the 14th century. There is evidence to suggest that only those knight's fees close to core Norman borough of Swansea, were in fact of early date, or of the 11th century, for example Penrice, Penmaen, or Nicholaston. On the other hand many of the fees to the west and north, such as Landimor, Llangennith and Llanmadoc appear to have remained apart from Anglo-Norman control until various dates between the early and mid-12th century. Once this control had been established, the process of Anglo-Norman domination of both legal, linguistic and cultural spheres occurred, apparently at differing rates, quite often gradual extending in piecemeal fashion to mask native Welsh influences; initially, for example, both Welsh and English Law appear have operated side by side. The divide between Gower Angliciana and Gower Walliciana, between the Englishry and Welshery also appears to have gradually changed overtime, as dominance of English Law and language increased.

It is likely that the given the high quality of agricultural land in south and west Gower, that many of the settlements and administrative centres were established by the early medieval period, if not the late prehistoric period. Indeed it is very likely that the plateau of Gower supported a relatively high population from the prehistoric period onwards: there is a possibility that some of Gower's promontory forts and defended settlements may have continued in occupation from the late prehistoric period into the early medieval period, as at Stembridge, indeed some of these sites may then have been adapted following the Anglo-Norman occupation to serve as manorial centres, given their location amidst the best land. Sites where this may have happened include the defended promontory fort at North Hill Tor (HLCA014), the enclosure at Reynoldston (HLCA037) and Berry (HLCA040), where prehistoric and later remains are found in close proximity. It is also possible that the ringworks at Cil Ifor (HLCA066), Norton Camp, Oxwich (HLCA048), and at Mountybrough, Penrice (HLCA046), also developed from pre-Norman precursors, and were not newly established settlements.

Parish and manorial boundaries use physical features in the landscape, and are likely to be long lasting and conservative characteristics within the landscape (Seyler 1924 and 1925; Morris (ed) 2000). For this reason many of the character area boundaries have adopted former manorial or parish boundaries, where these can be seen to broadly reflect corresponding variations in character.

Back to top of page

Prehistoric Funerary and Ritual Sites and Settlements

Gower contains some of the earliest evidence for human occupation in Wales with finds indicating some level of occupation from the Palaeolithic period (c. 200,000 BC - c. 10,000 BC). Finds from Paviland cave (HLCA029) indicate that people were using the area in the middle of the third millennium BC, and is one of the most important sites for an understanding of the development of the Early Upper Palaeolithic in Britain as a whole. The cave would have been close to the ice-sheet, and some 30km from the sea, and probably used as part of an extended hunting territory. The burial made at Paviland (the 'Red Lady') is of importance on a European scale; this burial suggests that the cave had some sort of particular religious significance. There is also evidence for the utilisation of other caves (Worm's Head, Long Hole) on what is now the south coast of Gower, but was then the edge of a plateau, and at Cat Hole in Llethrid Cwm (HLCA064), and a hand-axe has been found at Rhossili (Lynch et al 2000, 8-9, 11-12, 18-21).

The nature of occupation in the area has always been heavily constrained by its topography and soils, which have dictated a sharp divide between on the one hand the lowlands, which consist of the lower lying agricultural plain of the Gower peninsula and the coastal plain associated with the mouths of the River Loughor, and on the other hand the higher areas of common land, which consist of a series of ridges, and indeed the mountainous lands, which form non-peninsula Gower, to the northeast and beyond the AONB, itself. There is also a climatic distinction between peninsula and non-peninsular Gower, with the coastal belt being one of the more favoured areas of the British Isles, and the uplands lagging several weeks behind in terms of spring re-growth. The uplands, including the upland commons of peninsula Gower, are also now characterised by relatively infertile soils and peat development, but insufficient information is available to be able to determine whether this is the result of soil exhaustion caused by early agricultural activity, compounded with climatic deterioration in the later Bronze Age, as in other areas of Britain. Before the Neolithic period, land use would have been seasonal. Peninsula Gower has produced clear evidence for open settlement of Mesolithic date (c. 10,000 BC - c. 4000 BC), for example at Burry Holms, as well as for the reuse of Cat Hole cave in this period. By contrast the only evidence for a human presence during the same period in the uplands to the north consists of artefact scatters (Lynch et al 2000, 30-1).

The Neolithic period (c. 4000 - c. 2000 BC) is represented principally by chambered tombs, which were communal burial places apparently for whole communities. Neolithic monuments are widespread in the western half of Gower and must represent a significant population, though traces of settlement sites are difficult to find. Although the Neolithic period sees the introduction of agriculture, it is by no means clear whether it was practised to any great extent, and the economy may still have been significantly dependent upon hunting and collection of other foods from the wild (Pollard 1997, 8-9). We have no evidence for contemporary field systems.

Aerial photographic material has revealed a potential henge monument (SAM GM580), a ditched and embanked monument, with the ditch internal to the bank, at Newton (HLCA039). This is considered to be a ritual feature characteristic of the Neolithic period (Aerial photographs: RCAHMW 1964 and 1992).

Two traditions of tomb construction are represented on Gower. Parc le Breos tomb is a fine example of a type which was widespread in southwestern Britain and known as the Cotswold-Severn type from its distribution; this was characterised by a forecourt at one end of the long mound, giving access to a central passage with side chambers. The other type, of which the best-known example is Maen Ceti or Arthur's Stone (SAM GM003) on Cefn Bryn, consisted of a single massive slab or boulder supported on massive upright slabs (orthostats), which formed the sides of the burial chamber or chambers, and originally covered at least in part by a mound. There is another chambered tomb (SAM GM167) on Cefn Bryn, at Nicholaston, another nearby on Penmaen Burrows (SAM GM123), and a pair on the side of Rhossili Down.

The Bronze Age (c. 2000 BC - c. 600 BC) is represented principally by funerary and ritual sites, with a dearth of information on settlement and agricultural use. The cairnfields consisting of groups of small cairns, known from most of the upland ridges and very common on Cefn Bryn, have been interpreted as the result of clearance for agriculture, but they are generally not accompanied by any signs of laid-out fields, and may rather have had a ritual purpose.

The communal tombs of the Neolithic gave way to smaller round barrows and cairns; these were designed for the burial of an individual or small group, and are considered to represent a change in social structure. Ring cairns with an external stone bank and an open interior, were also constructed (Ward 1988); these may possibly have fulfilled similar purposes to stone circles, of which no examples are known in Gower. On the Gower peninsula, as elsewhere, there is an emphasis on siting these monuments on high ground; Cefn Bryn, Hardings Down, Llanmadoc Hill and Rhossili Down, where the main concentrations are, seem to have been treated in the same way as the inland upland (Evans and Lewis 2003), and there is a relatively much lower density of sites in the lower-lying areas. These characteristic concentrations of funerary monuments along Gower upland ridges, such as at Llanmadoc Hill (HLCA012), where14 cairns (originally a group of 19) survive, are considered to be dispersed Bronze Age cemeteries located within a wider Bronze Age landscape; it may be that these upland areas were held as 'sacred mountains'. This type of monument is also characteristic of high ground along the coastal margins, as that at Burry Holms (HLCA006; 00023w).

The construction of ritual monuments increased greatly during the Bronze Age as evidenced by forty or more cairns situated along the ridge of Cefn Bryn (HLCA038). The monuments on Cefn Bryn, characteristic sepulchral structures of the period, belong to a wider Bronze Age ritual landscape, which includes Llanmadoc Hill (HLCA012) and Rhossili Down (HLCA023); the upland areas and ridges appear to have played an important role in ritual and perhaps also defined territories. The results of the excavation at the Great Cairn Ring Cairn (SAM GM196), near Arthur's Stone (HLCA038), now reconstructed, might support the latter hypothesis; the site has been recorded as being a forum for wider ceremonials of symbolic practice (Ward 1981) rather than a repository for the dead. Excavation of a second ring cairn on the common also displayed no evidence for burial (Ward 1982).

Further indications of Bronze Age funerary ritual activity may also be represented by human remains found within caves, such as those excavated at Three Chimneys Cave (SAM GM087), Burry Holms (HLCA006), interpreted as an ossuary. However, the low-lying ground was distinguished by the presence of standing stones, particularly in west Gower, which were probably more widespread than they are now (RCAHMW 1976a, 121-2).

Evidence of Bronze Age settlement is by and large poorly represented, though features such as the midden at Burry Holms (00036w; HLCA 006), and burnt mounds, an enigmatic site type dating to the period, are found on Cefn Bryn (e.g. SAMs GM543; GM436; GM544), and elsewhere, for example near Llangennith (HLCA011) and on Rhossili Down (HLCA023). These sites, comprising mounds of burnt stone, and charcoal usually located in waterlogged areas or close to water sources are considered to be cooking or feasting mounds, but which could equally have had a ritual aspect; other interpretations include sauna sites or sweat houses.

The characteristic settlement type of the Iron Age (c. 600 BC - c. AD 50), as represented in Gower, is the hillfort, with is sub-class, the promontory fort in which a site was defended on two or three sides by natural hill slopes or cliffs and on the other side or sides by man-made defences. These are particularly common on the south Gower coast. Hillforts are now known to have originated in the Late Bronze Age but, in the absence of excavation carried out to the latest standards, this has not as yet been confirmed for Gower. Although some hillforts and promontory forts are heavily defended, in others the surrounding banks and ditches were probably as much boundary markers as fortifications. It is likely that the larger forts, such as Cil Ifor (SAM GM124; HLCA066), some 3ha in area, and the Bulwark (SAM GM061), Llanmadoc Hill (HLCA012) would have provided some form of regional focus, or had an administrative function. Examples of settlement of the period range from the aforementioned large multivallate hillforts to smaller defended enclosures; such as Reynoldston camp within HLCA037 (00161w; 94607; SAM GM195), where occupation appears to have continued into the Roman period, and the three scheduled enclosures (SAM GM060) on Hardings Down (HLCA025), one of which, a small hillfort (00025w; 301323), revealed the foundations of two sub-circular hut platforms and Iron Age pottery. Less well-known sites include two earthwork enclosures recorded near Cilonnen (00235w; 00946w), within HLCA067.

Examples of promontory forts include seven sites, mostly univallate, within HLCA029, including, Lewes Castle fort (SAM GM470; 00140w), Paviland Camp (SAM GM128), Old Castle Fort (SAM GM193; 00139w), Horse Cliff (SAM GM192), and Worm's Head (SAM GM492). Good examples of multivallate types include Thurba Head (SAM GM127) and The Knave (SAM GM128), both within HLCA029. Other coastal promontory sites include that located on Burry Holms (SAM GM088; HLCA006) (HLCA 060, SAM GM132) and others along the northern coastal margins around Tor-gro and North Hill Tor (HLCA014), though the camp at North Hill (SAM GM062) has been argued to be of medieval date (RCAHMW).

Hillforts, and defended enclosures in general, appear to have been an enduring class of monument, as there is some evidence for continued occupation into the following Roman period (c AD 50 - c AD 400), and in at least a few cases probably beyond. The Iron Age promontory fort (SAM GM126) within the Bishopston Valley (HLCA084) excavated by Aubrey Williams in 1939, provided evidence to suggest occupation extending into the Roman period indicated by pottery of 1st and 2nd centuries AD. Elsewhere, for example at Stembridge Camp, continued occupation of sites identified as late prehistoric may even have extended into the early medieval period, or at least indicate re-occupation; recent geophysics and survey carried out by University of Wales, Newport, has confirmed the existence of platforms within the inner part of Stembridge defended enclosure (Kissock 2006, pers comm.).

There is little evidence for any significant degree of Romanisation of peninsula Gower following the Roman invasion, and there was probably little change in the way of life practised by most of the inhabitants of the area. At Scurlage, New Henllys, and Lower Harding's Down, ongoing projects being undertaken by University of Wales, Newport, on Gower, including survey, geophysics and excavation, have revealed rectilinear and circular settlement features strongly reminescent of, for example, the late Iron Age/Romano-British settlement at Whitton (Kissock 2006, pers comm.), settlement type based on the native tradition. The nearest known sites with much in the way of the material trappings of Roman culture were the fort at Loughor and an enigmatic site known only from a mosaic at Oystermouth Church. The fort at Loughor was apparently accompanied by an extra-mural civil settlement, but it is not really possible to determine its nature from the stray finds and observations, which have been made outside the defences (Pearson 2002, 21).

It is likely that the way of life of its inhabitants will have been considerably more Romanised than that in the surrounding countryside. As before, the basis of the economy in both the Iron Age and Roman period will have been agricultural, but no surviving evidence for the contemporary landscape has been identified to date. Pre-medieval field systems of any description are rare in southeast Wales, and work done in the lowland between Port Talbot and the River Wye suggests that they may have been restricted to small fields or paddocks in the immediate area around settlements (infield) and surrounded by extensive areas of undivided land (outfield) (Evans 2001, 34). No comparable work has been done for Gower, or indeed the Swansea unitary authority area.

In circumstances such as these, any new settlements are likely to have been created by enclosing previously unenclosed ground. Many first enclosures have a characteristic curvilinear form, as this is the usually the most economical shape for an enclosure to assume where it is not constrained by other factors, such as pre-existing boundaries or the need to create a fieldscape at the same time (though natural topographical features will of course have to be taken into account).

Back to top of page

Agricultural and Settlement Landscapes

There is little direct evidence for use of the present agricultural landscape in the prehistoric period. Caves in Llethrid Cwm (HLCA064) were utilised in the Mesolithic and Bronze Age, latterly for burial. Also of probable Bronze Age date are the standing stones which cluster in the north and west of the peninsula, such as those located near Old Walls (HLCA020), the remnants of what seem to have originally been a much larger number; RCAHMW (1976a, 121-2) recorded a theory that they might have been associated with early trackways, but were unable to substantiate it. The later prehistoric and Roman periods are represented by numerous small enclosures, but none of these are known to be associated with features in the wider landscape, such as fields.

The fieldscape of the greater part of the Gower peninsula can be traced back to the open field system of the medieval period, (Flatres 1951; Kissock 1986). It has been the consensus of opinion that open fields were first introduced by the Normans (Emery 1971, 155), though Kissock (1991) has argued that they may have a pre-Norman origin. Whilst traditionally held opinion has placed the establishment of the nucleated settlements of Gower with their openfield systems of sharelands firmly within the period of Anglo-Norman control; there are, however, compelling reasons to suggest that these settlement patterns, at least as they appear in north and west peninsular Gower, may have been old established and conceivably have developed from earlier surviving Welsh settlement patterns, models for which are contained in the Welsh Laws. In fact the existence of 'landshares' or rhandiroedd is associated with the trefi (settlements) of free maenol. A type of shareland arrangement which may have some bearing on settlements of the area is nucleal land or tir corddlan, a variant of hereditary land comprising small strips or quillets arranged radially around some kind of nucleus; these were shared gardens, frequently occupied by under tenants. The nucleus could be a churchyard, ecclesiastic land for example being shared in common between brothers. Equally bond tenants would have held equal shares of reckoned land in return for joint obligations and communal renders to the meardref or reeve's vill (Jones 1989).

The settlements of Rhossili (HLCA031; HLCA013), Llangennith (HLCA011), Llanrhidian (HLCA022), Port Eynon (HLCA044), and Llanmadoc (HLCA007) are likely to have initially developed as a nucleated settlements focused on early medieval churches. A similar pattern can be seen at Llanddewi, an Episcopal manor, where there is evidence for a deserted medieval settlement adjacent to the church. Bishopston, a further Episcopal manor, also originated as a nucleated settlement, although this is now masked somewhat by disproportionate 20th century expansion. The settlement of Penrice (Mounty Brough), is focused both on its church, which interestingly may be pre-Norman in origin, and a ringwork, again thought to be Norman, but not confirmed.

It has been considered that the medieval period is the earliest period to which significant areas of the modern landscape can be traced; whilst at present this may be true, future investigation/study may well shed more light on the subject and allow some reconstruction of the early medieval landscapes of Gower. The most important relict landscapes on Gower of the medieval period are undoubtedly the Vile, an area of nationally important still extant strip-fields at Rhossili; the boundaries of the medieval hunting park can still be traced at Parc le Breos, though the area of the park itself was largely divided into fields later on in the Middle Ages, and in the Post-medieval period (Leighton 1999). The Vile is the only substantial relict of a medieval open field, though records only go back to the early 18th century. Although this pattern is generally thought to have been introduced by the Normans to replace the dispersed settlement considered to be the characteristic norm for areas under Welsh domination, it is likely that there was some sort of occupation in the area beforehand, and Kissock (1991, 41-3) has argued that the open field at Rhossili was already a feature of this pre-Norman landscape. Davies (1978, 135; 1979, 97, 124) has also suggested that Rhossili may be the Lann Cingulan of the Book of Llandaff, whose context indicates that it must have been in Gower; however the neighbouring estate of Lann Gemei, whose boundaries are given as 'from the ridge of the Hill, which divides the uncultivated ground and the field, into the sea, and [from the same ridge] right to the spring of the Diwgurach; along [the Diwgurach] downwards as far as the sea' (Evans 1893, 140, 368) fits the topography better.

Originally the Vile and other openfield systems would have been divided into strips separated by turf banks, known as landshares or quillets, (this appears to be a development from the usage of the tithe apportionments, where the word is used to indicate a landholding in a communal field which had still not be fully enclosed from the open field system). The management of openfield systems, such as those of the Vile, and elsewhere (e.g. Llangennith, and Llanrhidian) indicates a characteristic mix of arable and communal grazing, regulated by agreement among the strip-holders; these holdings appear to have been worked communally (Emery 1974, 7-12). Where modern boundaries have been introduced to replace or supplement the banks, these consist of hedges planted on top of the banks, and post and wire fences, with a few stone walls.

Examination of the tithe maps show that, in most parishes, at least some small areas of strip fields survived into the second quarter of the 19th century in a similar form to the Vile (i.e. with no permanent boundaries on the ground between strips, which are shown on the maps as divided by dotted lines); there are also small areas of long narrow fields which are clearly enclosed strips. An example is provided in the undivided strips within enclosed meadows, which appear near Llethryd. Another 'Vile' (known by that name) with vestigial strip fields has been identified, located between Llangennith and Burry Green (Morris 1998, 135), Morris also provides a brief list of other Gower locations where there are indications of former unenclosed strips. Some elements of stripfields survive, for example to the north of Llanmadoc village (SS442936), or in Llanrhidian Higher, between Llanyrnewydd Church and Cefn Bychan (SS855948); there is no apparent difference in this respect between the more highly anglicised south and west Gower and northeast Gower where the place names indicate much more extensive survival of Welsh. A case of particular interest is Bishopston (HLCA086), where the tithe map shows a patchwork of long narrow fields set into blocks running roughly at right angles to each other. Although most of these have solid boundaries, there are some where dotted lines have been used for the boundaries, showing that they were originally strips. Individual fields can still be seen, most clearly east of Longash Farm (SS581887) and south of Murton village (SS588887), but the whole pattern is still clearly apparent and has informed the development of settlement in this part of the parish in the second half of the 20th century, the same is true of 20th century settlement development in Southgate, Pennard. In contrast, the area to the north of the B4436, towards Upper Killay, Wernllath (HLCA080) is more irregular and represents piecemeal enclosure from Fairwood and Clyne Commons (HLCA074 and HLCA079, respectively). Further piecemeal enclosure and clearance of woodland (assart) is characteristic of Llanrhidian Higher (HLCA067); this would appears to largely represent early post-medieval settlement expansion, rather than any notion of the 'traditional Welsh' dispersed holding, as the older agricultural systems to the north west of the area appear to include nucleated settlement and former openfield.

Where an open-field system was in operation, it is likely that it was accompanied by nucleation of settlement, and the nucleated settlement is still the dominant characteristic form in peninsula Gower in apparent contrast to the upland parishes of Llandeilo Talybont and Llangyfelach, beyond the bounds of the AONB, where non-nucleated settlement appears to have been the norm, and farms were created by piecemeal enclosure. Somewhat surprisingly, there is also evidence on the Llanrhidian tithe map for limited areas of landshare, at Wernffrwd, Rallt, Llanmorlais and elsewhere in Llanrhidian Higher (HLCA067). Documentary evidence dating to the medieval period also refers to openfield systems operating in the area (Cooper 1986; 1988).

Non-nucleated dispersed settlements do exist within the area of peninsula Gower (Subboscus), for example in Llanrhidian Higher (HLCA067), where Welsh influence appears to have remained strongest. To say, however, that dispersed settlement equates to a 'Welsh tradition' or influence, or visa versa, is perhaps at best an oversimplification. Dispersed settlement within peninsula Gower appears to be, at least in part, a demographic response to social change and economic prosperity noted during the early post-medieval period, which caused changes in both agriculture and settlement.

The principal unit of land division in the fieldscape as it exists today is the farm; these are largely irregular in outline. The majority of holdings appear to have been enclosed from the medieval/post-medieval open fields, a process, which was more or less complete by the time the tithe maps were drawn up in the 1830s and 1840s. Others appear to have been enclosed from common, or former monastic grange land (see 6.4, below), or other unallocated land, such as the medieval Parc le Breos deer park. The policy of division into fields within individual farms varied. In some farms it is based closely on the individual strips, some of which are recognisable today, and many more were so at the time the tithe maps were drawn up. At others a complete resurvey seems to have taken place, so that the fields within that particular farm are regular. A third option was the creation of small irregular fields; this is most often seen in the hillier parts of the peninsula, particularly the northeast, but map regression makes it clear that the underlying basis was still the open fields. Encroachments on common land are documented from the 16th century, and could range from single-unit fields of perhaps 3 acres, up to blocks of 120 acres. There were significant areas of these in the northern and eastern parts of the Gower peninsula; there appears to have been far less scope for enclosure in west Gower. These encroachments were for a variety of purposes, mostly agricultural (arable as well as pasture and meadow) but could include coalmining or water-mills. (Emery 1971, 155-8; Robinson 1968). Small areas of enclosure can be recognised on the saltmarsh, as for example at Landimore (HLCA015; Locock 1996a, 11). The commonest forms of boundary are hedges of various sorts, particularly hedge banks and hedges with trees. Only limited use is made of stone walls.

Settlement, like the fieldscape, appears to some extent been influenced by Anglo-Norman models, though may be based on older native models, and originally consisted of small villages nucleated on a single centre, together with smaller hamlets and isolated farmsteads. Oxwich village (HLCA048), appears untypical, as a ribbon development (Nuttgens 1979, 7-9), however further analysis of the landscape would suggest that Oxwich Green nearer to the late medieval castle and fortified residence of Oxwich Castle, and with its nucleus of farmsteads set around an open 'green', and a number of outlying 'hamlets' is a better contender for the earlier medieval settlement core. The process of enclosure, and consolidation of holdings, as occurred around Oxwich Green, was frequently accompanied by significant social upheaval and population displacement, resulting in swelling the ranks of the landless poor. The current ribbon development settlement of Oxwich, appears to be the result of encroachment along the edge of the Oxwich Marsh (HLCA054), may in part have evolved as a direct result of enclosure.

Throughout Gower a reduction, or desertion of nucleated settlements, appears to have occurred in favour of settlement within newly consolidated holdings, for example at Kennexstone and Tankeylake (HLCA016). This process may have had its origins in the late medieval period, the result of the Black Death or the Glyndwr rebellion. The relative proportions of this process, which this took place in medieval as against post-medieval times is not clear; by the time the tithe maps were surveyed, very little of the field systems remained to be enclosed. Abandoned settlements are noted throughout north and west Gower in particular, for example at Llanelen (HLCA067), Coety Green (HLCA011), and at Llanddewi (HLCA034), among others. It is likely that other deserted rural settlements of medieval date await discovery, for example recent work by University of Wales, Newport, has lead to the identification of a series of platforms probably associated with a deserted/shrunken medieval village at New Henllys (HLCA033; Kissock 2006, pers comm).

Most of Gower's villages, which survived in to the post-medieval period, were subsequently extended by ribbon development during the 19th and 20th century, and a few have now grown to form substantial settlements. The biggest changes are seen at Pennard, formerly a dispersed village roughly centred on its church (HLCA061), which has later developed a new focus at Southgate (HLCA062), along the road between the golf course and the cliffs. Mapping from the 18th century to the present allows the development of open field systems, such as that the Vile at Rhossili, and indeed other systems, such as those noted at Llanrhidian (HLCA022), Leason (HLCA020), Llangennith (HLCA011), Port Eynon (HLCA044), Horton (HLCA045), and elsewhere. Generally cartographic evidence charts gradual enclosure on the periphery at first, though for many the process of consolidation of holdings is well underway by the mid-19th century, if not the late 18th century. The survival of the Vile system until relatively late, for example, may be in part due to its isolation and conservative attitudes, indeed here enclosure and consolidation was still incomplete in the 1970s; between 1780 and 1845 little change is noted in the occupation of the various holdings shared between farmers from Rhossili and Middleton with holdings ranging between just under an acre to 49 acres (Davies 1956). Gradual enclosure of most strips, coupled with consolidation holdings particularly towards the end of the 19th century and amalgamation into larger fields through removal of divisions, typically during the 20th century appears to have been the norm.

Enclosure of Gower's open-field parishes had not entirely been completed when the tithe maps were drawn up between 1838 and 1848; significant areas of the open field strips remained, either enclosed into long narrow fields or still functioning as landshares. Some of these long narrow fields can still be seen, though removal of boundaries has tended to obscure their origins. Other elements of the medieval landscape can also still be recognised, particularly the medieval hunting park of Parc le Breos and the monastic grange at Walterston Leighton 1999, 76; Toft 1996), but these have been subjected to the same processes as those that have moulded the fieldscape elsewhere in the area. These processes are enclosure and encroachment on the waste, both the inland commons and the cliffs and saltmarshes. The landscape was also diversified by limited areas of woodland, some of which occupy the steep sides of the cwms, which cut the southern side of the peninsula.

In the agricultural areas of the Gower, the 20th century witnessed some removal of hedgerows, modifying the fieldscape, though fields still remain small in the hillier areas of northeast and northwest Gower, where the ground is less suitable for large-scale agricultural machinery; there is also significant survival of small fields in the southeast around Bishopston where the same topographical restraints do not apply.

Back to top of page

Ecclesiastical Landscapes

There is considerable literature linking curvilinear enclosures with early medieval ecclesiastical sites (e.g. Thomas 1972, Brook 1992), and there are a number of churches within Gower AONB where the churchyard could be such an early enclosure. The most convincing examples come from those sites where there is other evidence for an early church, such as at the ancient church of St Teilo at Bishopston, possibly 6th century in date. Llangennith is another early church, and here the churchyard appears as half an oval largely surrounded by roads (another possible indictor of early date). Some of these early churchyards may have lain in the middle of larger curvilinear enclosures, possibly representing the boundary of the associated nawdd or sanctuary; there is a possible site of this type at Pennard (HLCA061). However, such curvilinear patterns are also characteristic of enclosures from common land or other uncultivated areas, which could be made at any time up to and including the post-medieval period; this origin is to be suspected when the enclosure lies on the edge of a common. Some early medieval ecclesiastical sites in Gower, with their associated estates, are documented in the Book of Llandaff, a collection of charters relating to grants of land to the see of Llandaff by the 12th century; these are Bishopston, and, less certainly, Penmaen, Pennard and Rhossili (Davies 1979, 97-8, 124; Evans 1893, 140, 144, 145, 239). As yet, none of the charter boundaries has been confirmed on the ground, but there may be potential for recognising elements here of the existing landscape as going back to the pre-Norman period (Kissock 1991; 2001). It has been argued that early ecclesiastic holdings, such as Lann Cinuur at Bishopston (HLCA086), identified with the episcopal estate held by the see of Llandaff (Davies 1979, 97-8), may survive fossilised in the surviving field systems (Kissock 1991).

Indications of the early medieval ecclesiastic characteristic are common throughout peninsula Gower; St Madoc's Church, Llanmadoc (HLCA007), with its partly curvilinear churchyard for example, though predominantly of twelfth century date, it is regarded as having been founded much earlier in the sixth century by Saint Madoc. Other early medieval Welsh dedications include that to Cadoc at Cheriton (HLCA018), probably the original church associated with the early 'maenor' of Landimor, later between 1135 and 1230, the church at Cheriton was granted along with those at Llanrhidian and Rhossili to the Knights Hospitallers. In relation to the two settlements of Rhossili (HLCA013 and HLCA031), it is considered that these are the two adjoining estates, Lann Cingulan and Lann Gemei, referred to in the Llandaff charters; Davies (1978, 135; 1979, 97, 124) suggests that Lann Cingulan was Rhossili. However, the topography fits best if Lann Cingulan is the Burrows, whose boundaries are given as 'between the two ditches towards the sea and upwards along the two ditches to the mountain along the Cecin, the boundary of Llan Gemei' (Evans 1893, 144, 369), and Lann Gemei is Rhossili village. Documentary evidence suggests that by the early medieval period a monastic cell had been established at Llanrhidian (HLCA022); a reference in the book of Llandaff of c.650 has been interpreted as indicating the existence of a cellula at the site dependant upon the main monastic estate at Rhossili (Davies 1979, 97). Again the documentary evidence is supported by the physical remains at the site a carved stone of 9th-10th century date (also suggestive of an important site of some architectural pretention) and the partly curvilinear shape of its churchyard, as well as its dedication to the native Welsh Saints Illtyd and Rhidian.

A characteristic feature of early medieval Christianity was the offshore hermitage, often established to afford solitary retreat. An example of such a site is to be found in the medieval hermitage/ecclesiastical monastic settlement (SAM GM473) on Burry Holms (HLCA006) known as 'the church of the isle' or 'the hermitage of St. Kenydd-atte-Holme'. It is recorded that Henry earl of Warwick, lord of Swansea and the Gower, granted Burry Holms as well as other properties in northwest Gower to the Abbey of St Taurin in Evreux, Normandy between 1095 and 1115 (Savory 1984). Although the main surviving structures are of twelfth - fourteenth century date, excavations in the 1960s revealed evidence of an old established pre-Norman settlement; the remains at the site included a church, associated dwellings, hall, and possible schoolroom.

Also dedicated to the 6th century St Cennydd is the church at Llangennith (HLCA011); this site retains part of a ninth century cross (RCAHMW 1976c, 46 no 905) and has a partly curvilinear churchyard, indicators of an early foundation. Llangennith is considered to have been the site of a clas church, reputedly destroyed in a Viking raid in 986 AD. By the early 12th century the church, formerly occupied by the hermit, Caradog was granted to the Abbey of St. Taurinus at Evreux in Normandy by the Earl of Warwick. Shortly after a small dependant Benedictine priory cell was established that was separate and distinct from the church (Orrin 1979, 40). St Cennydd, a missionary of St Cattwg, is also credited with founding the church of St Cadoc (Cattwg), Port Eynon (HLCA044), in the 6th or 7th century.

The exact effect on the landscape that the granting of churches and other property to the various monastic orders has yet to be fully examined with regard to Gower. Several of the churches of north and west Gower for example were granted to the Knights Templar or Knights Hospitaller, between the 12th and 14th centuries. The current layout of some settlements, such as Llanmadoc (HLCA007) and Cheriton (HLCA018), may have been influenced by the granting of land to monastic orders, although further detailed study is required to confirm this.

The area of Llanddewi (HLCA034) is significant as a long-standing ecclesiastical centre with early medieval origins through its dedication to St David and curvilinear churchyard enclosure (Evans 2003). The fourteenth century church at Llanddewi is reputed to have been constructed by Henry de Gower, Bishop of St David's (1328-47) in conjunction with a castle or palace, however, it seems that the palace, mentioned in Statutes of St David, was abandoned (Orrin 1979). An earthwork 450m southwest (00160w; 305468; SAM GM334) of the church on a triangular area of common near the centre of the area has been conjectured as relating to the aforementioned palace site, however, others report it as an enclosure of Iron Age date, which is questionable (SAM description).

Other indications of the early medieval ecclesiastic landscape include a Pillar Cross of ninth century date (SAM GM089) found near Stouthall, Reynoldston (HLCA037); this might have marked the boundary of an ecclesiastic estate, possible centred on Llanddewi, or Llangennith. The nearby present church of St George at Reynoldston (HLCA037) is thought to occupy the site of the thirteenth century medieval church founded by Sir Reginald de Breos.

Monastic granges and grangeland appear to have had some effect on the landscape, though detailed analysis of the landscape associated with the granges of Gower remains to be undertaken. These landscapes appear often to have either been the subject to a high level of agricultural improvement, particularly those associated with the Cistercian Order as in HLCA043, Paviland and Monksland, which formed granges of Neath Abbey. Mentioned in the taxation of Pope Nicholas IV in 1291 the monastic grange consisted of one carucate of arable and two mills. The various granges, such as those associated with the Cistercian Abbey of Neath, would have had an important influence on the development of the landscape of the area in terms of agriculture and settlement; stock husbandry (cattle and in particular sheep) was further developed during monastic control of the area (Owen 1989, 213; Cowley 1986; Williams 1990 and 2001). Other monastic lands are to be found at Cillibion (HLCA070) and Walterston (HLCA063), again relating to the Cistercian Abbey of Neath. Elsewhere, large-medium regular enclosures can be seen in association with former monastic grangeland as at Lunnon (HLCA065) and Berry within HLCA040, the latter an ecclesiastic Manor of the Hospitallers. The exact process behind the formation of the landscapes associated with these monastic holdings is unknown, however some may be a result of early post-medieval enclosure/consolidation of the monastic estates as they were redistributed following the Reformation of the 16th century.

There are several chantry chapels in the area including Backingston Chapel in Bishopston (HLCA086) and also reference to the now unknown location of the free chapel of Henllys in Llanddewi (HLCA033 or possibly HLCA034), surveyed in 1545 and 1547 by commissioners appointed by Henry VIII (Evans 2003a).

Historic associations with Non-conformism is a noted characteristic of post-medieval settlement within Gower, particularly in the more industrialised north east; this characteristic is physically represented by chapels, Sunday schools and meeting houses. Non-conformist chapels, such as those within Llanrhidian Higher (HLCA067): Tirzah Baptist Chapel, Llanmorlais, Crwys Independent Chapel, Y Carmel, Penuel Calvinistic Methodist Chapel, Llanmorlais, and the St David's Mission Room, Wernffrwd are generally characteristic of settlements of northeast Gower, and reflect industrial settlement expansion from the latter part of the 19th century. The non-conformist association is also found within south Gower, for example at Ilston (HLCA072) important for early dissent and where in 1649, John Miles established the first Baptist church in Wales, the Baptist congregation using the parish church until the Restoration (Williams 1999, 11-18), and Burry Green (HLCA026) where Bethesda Chapel, opened in 1814, is noted for its association with Lady Diana Barham and the Calvinistic Society and famous Gower preacher, William Griffiths. Oxwich (HLCA048), and Horton (HLCA045) among other areas have strong links to Weslyan Methodism. For example, Oxwich is known to have been visited on numerous occasions by the reknown Methodist preacher John Wesley during the latter half of the 18th century.

Back to top of page

Defensive Landscapes

There are few landscapes within Gower, which can be said to be truly defensive, though several have defensive characteristics or elements. Prehistoric defended enclosures, including promontory forts and hillforts (see 6.2 above), form characteristic elements of coastal and common land areas, dealt with elsewhere in more detail, as their defensive nature is only part of the picture. Character areas where prehistoric hillforts are a characteristic include HLCA012 (the Bulworks, Llanmadoc Hill and HLCA066 (Cil Ifor), of related interest is a possible large enclosure(s) identified on aerial photographs during the current project immediately to southwest of the latter fort.

Promontory forts, which include both coastal and inland sites, appear to characterise not only the prehistoric period, but also the early medieval and possibly the early post-Norman periods, along with simple ringwork defensive enclosures. There may well be some continuity of use/or reuse of defensive sites, specifically enclosures and ringworks, between the early medieval and the post-conquest period, although the lack of recent excavation means the exact chronology and detail is limited, or speculative at best; it should be borne in mind that the ringwork as a site type need not necessarily be exclusively Anglo-Norman in origin. The distribution of ringworks shows a strong correlation to the better land of peninsula Gower's plateau; it is unlikely that this land would not have been exploited during the early medieval period. Given this and the limited evidence of settlement of the early medieval period, beyond ecclesiastic sites and potential continued or re-established occupation of promontory forts, the possibility that early medieval settlement actually occupies the same foci, as those developed under Anglo-Norman control, would not necessarily be unusual, and should not be discounted. Indeed it is often the case that opinion is divided as to whether a particular defended enclosure can be considered to belong to the prehistoric, early medieval, or medieval period, such as the earthwork at Llanddewi (SAM GM334; HLCA034), the promontory site at North Hill Tor (SAM GM062; HLCA014), or the inland promontory site of Pencynes at Stembridge (SAM GM125; HLCA020). The chronology of the site at Berry (HLCA040), for example, where a defended enclosure (SAM GM178) of supposed prehistoric date, exists in close proximity to adjacent 'moated' site, considered as medieval, could be of interest. Character areas where prehistoric or later promontory forts and defended enclosures are known to be located are as follows HLCA014, HLCA020, HLCA025, HLCA029, HLCA034, HLCA040, HLCA042, HLCA048 and HLCA049, HLCA055, HLCA058, HLCA060, HLCA064, HLCA067, and HLCA084.

The chronology of the development of defensive enclosures into ringworks is, as mentioned above, largely unproven, and no recent excavation of a ringwork on Gower has been carried out. The most recent carried out in the 1960s being investigations undertaken on the the ringwork at Penmaen castle (SAM GM129), which has been dated to the late twelfth century; excavation of the ringwork of Mounty Brough (SAM GM053) in 1927 revealed remains of a possible fighting platform. Old Castle, north of Bishopston (SAM GM154; HLCA080), partially excavated in 1899 by WL Morgan (RCAHMW 1991, 81-3 CR1), is in part interesting for its location at the boundary between former common land and the openfield associated with Bishopston (HLCA080); this might indicate a location was chosen particularly to control of the use of the common and the manorial boundary to the north. Similarly the location of the ringwork at Norton (SAM GM157; HLCA048), a partial ringwork with an internal house-platform near Oxwich, at the northern edge of the manor/fee may also reflect territorial protection. Unfortunately little is known of the history this site, which is may have formed part of the de la Mare family's holding, along with Oxwich Castle itself.

Re-use of an earlier defensive site, Cil Ifor hillfort (HLCA066), during the medieval period is indicated by the construction of a ringwork (SAM GM124) at the southeast of the site within the ramparts. It has been suggested that this or the promontory/ringwork (SAM GM062) on North Hill Tor may be the location unlocated castle of the Tuberville family.

Perhaps the most obviously defensive in characteristic of the later stone built castles of Gower, is Penrice Castle (SAM GM047; HLCA047), thought to have a Norman origin, is at least 13th century in date, but almost lacking in datable detail. The site, abandoned in favour of Oxwich Castle in the late medieval period, was slighted by Cromwell in the 17th century, and later, during the 18th century, turned into a picturesque ruin.

Later medieval structures, such as the stone castles of Weobley (SAM GM010; HLCA020), established in the 14th century by the de la Bere family, and Oxwich (SAM GM043; SAM GM472; HLCA048), which may retain fragments of an earlier medieval structure, associated with the de la Mare family are representative of fortified dwellings rather than true castles, though with clear defensive elements (Newman 1995; RCAHMW 2000). The castles at Scurlage (HLCA040), and Llanddewi (HLCA034), of which little now remains, may have comprised smaller fortified residences at best, though current details are scant. Bovehill Castle, Landimor (HLCA019), a construction of the late medieval period with no apparent precursor on the same site, can also be seen in this context.

During the 18th century gun batteries were established on Mumbles Island and Hill (HLCA042), and again in 1860 in response to the threat of French invasion, and the area was refortified during World War II. Twentieth century defensive characteristics also include the Oxwich Point Chain Home Low Radar site (HLCA048) and features established on Fairwood Common, including an aerodrome as part of the World War II defences (HLCA074 and HL075).

Back to top of page

Besanded Landscapes

The natural process of coastal erosion and be-sandment has had an overriding impact on Gower; accumulations of wind blown sand have formed considerable areas of littoral dunes at a number of points along the South Wales coast, with their distribution primarily controlled by coastal topography and by the prevailing westerly winds. They are commonly found in bays; their extent limited by the size of the bay and the relief of the hinterland. Historical evidence indicates that the processes of be-sandment were very active in the later medieval period, during the 13th to 15th centuries, with a deterioration of climate and a noticeable increase in be-sanding as the result of storms, higher rainfall and abnormal tides, all occurring in combination. These conditions led to the erosion of coastal sand dunes and the movement of material inland. Similar processes had a major impact on coastal landscapes elsewhere in Wales, such as at Merthyr Mawr and Kenfig in West Glamorgan, and further a field at Llanddwyn or Newborough Warren on Anglesey, where the lands associated with the 12th/13th century Llys/maerdref settlement of Rhosyr were affected.

The process of besandment can be viewed as the reverse of the enclosure/assart process, as it effectively changed enclosed land to common. This appears to only have occurred in those areas of the coast, which were affected by besandment from the 13th century onwards, and is best documented on the south coast at Penmaen and Pennard, and the west coast near Rhossili but possibly also affecting the northwest at Llangennith and around the fringes of Llanmadoc.

Three notable examples exists on Gower of besandment during this period, these are the medieval settlements and ecclesiastical sites at Penmaen Burrows, Pennard Burrows (HLCA058) and Lower Rhossili (within HLCA013). The historic landscape areas of Penmaen and Pennard burrows also feature the remains of prehistoric monuments, including a Neolithic chambered tomb and possible Bronze Age cairns. This area has been speculated to be the site of two deserted medieval villages, as remains comprise two castles and two churches dating from the twelfth to thirteenth centuries. It is known that by 1316 the burrows at Pennard were already in existence. Encroachment by sand gradually increased leading to final abandonment of the area sometime in the sixteenth century; which was also the case at Penmaen. The burrows were noted as common land and therefore probably used for grazing from the post-medieval period and also used as a rabbit warren. Potential for buried archaeology in the Penmaen and Pennard Burrows is significant and potential for buried landscapes relating to the Neolithic and Bronze Age is high; the extent of medieval remains is however, unknown, though further remains are likely.

References of the 7th to 10th century in the Book of Llandaff refer to two ecclesiastical estates (Llan Cyngualan and Llan Gemei), thought to be within the parish of Rhossili; an, as yet unproven, association has been made between the besanded settlement at Rhossili (HLCA013) and the western most of these, Llan Cyngualan. The exact development of the two settlements at Rhossili, the present village (HLCA031), and the besanded settlement (HLCA013), is complicated by the existence of two churches; St Mary's, in Rhossili Village (HLCA031) and another in the besanded area (HLCA013), excavated in the 1980s. Local tradition states that the Romanesque doorway of the present church (within HLCA031) had been moved from the besanded church with the implication that besandment of the lower village, directly led to the establishment of the present settlement of Rhossili (HLCA031). This scenario is, however, unproven and it is equally possible that the two churches could have existed simultaneously (Evans 1998). Also, it is questionable, whether the besanded settlement with its 12th century church, ever constituted a village; and the possible interpretation of the site (at least in its final stages) as a monastic grange has been muted, (Locock 1996b). It is thought that besandment at Rhossili occurred during the fourteenth century and evidence shows that it was a rapid process in contrast to Penmaen and Pennard.

Other areas on Gower which are now besanded and have unknown archaeological remains, lie mainly along the northwest coast, but are also to be seen on the south Gower coast. The nature or date of besandment in these areas is as yet unstudied. The potential for buried archaeology in these areas is high given suitable conditions for survival and inference from similar areas elsewhere. Such areas identified in the AONB historic landscape characterisation are: Whiteford Burrows (HLCA003), Llangennith, Broughton and Hillend Burrows (HLCA009) and Oxwich and Nicholaston Burrows (HLCA053). At Whiteford Burrows, the presence of medieval remains near or at the surface of the sand indicates that pre-medieval remains may be buried and that the dunes were in existence before this date.

Back to top of page

Common Land

The commons of Cefn Bryn (HLCA038), Ryer's Down (HLCA017), Llanmadoc Hill (HLCA012), Hardings Down (HLCA025), and Rhossili Down (HLCA025), stand out from the lower lying commons, because of their marked upland nature; all are hills or ridges (though Llanmadoc Hill and Cefn Bryn have attached areas of lower lying land), but they are of relatively low elevation, when compared to true uplands. They are essentially areas unsuitable for arable farming, which have been left out of the general enclosure of the surrounding farmland. These commons generally have a high concentration of prehistoric monuments, particularly funerary ones, the only exception being Ryer's Down, where none have been identified (Plunkett Dillon and Latham 1987c, 22). The crests of Llanmadoc Hill and Rhossili Down in particular seem to have been laid out in the Bronze Age as dispersed cemeteries. However, since none of the monuments here have been archaeologically excavated, it is not possible to know how great a period of time was required for these cemeteries to reach their present form. It does not seem unreasonable, however, to postulate that they may have been considered 'sacred mountains'. Ritual use of Rhossili Down had already been established before the Bronze Age, with the construction of two chambered tombs known as Sweyne's Howes on its eastern flank. There are also two Neolithic chambered tombs on Cefn Bryn, but although there are numerous genuine Bronze Age funerary cairns and others whose primary purpose may have been more ritual than funerary (Ward 1987, 1988), the presence of even more undated small cairns makes an appreciation of the form of the Bronze Age landscape uncertain. These are sometimes described as clearance cairns, but there is no evidence for any associated field system, and they may rather have had a ritual function.

Evidence for other prehistoric land use is less widespread. There are undated hut circles and enclosure banks and walls on Rhossili Down, which could be prehistoric, though an early medieval date is also possible (Plunkett Dillon and Latham 1986c, 13-5). The Iron Age is represented by hillforts on Hardings Down and Llanmadoc Hill, although no associated landscape features have yet been identified.

In the medieval and post-medieval periods the commons formed part of an integrated agricultural system in which each village had its surrounding open fields and access to common land for grazing (Emery 1971, 156). Some encroachment is documented from the 16th century onwards, but it was relatively minor compared with encroachments further north, consisting mainly of individual fields enclosed from the waste (Emery 1971, 156-7), though larger areas were occasionally taken, such as Hillend farm which was enclosed from Rhossili Down between 1847 when the tithe map was drawn up and 1885, the date of the 1st ed OS 6" map. There is also some quarrying recorded. There are very few sites on the commons post-dating the prehistoric period, though a series of defensive installations on Rhossili Down date mainly to the Second World War. The area still consists of moorland used for rough grazing.

Of the lower lying commons Welsh Moor and Forest Commons (HLCA069), Pengwern Common (HLCA071), Fairwood Common (HLCA074), Barlands Common (HLCA085), and Clyne Common (HLCA079), comprise a belt of land lying across the southwestern edge of the Coalfield deposits, the largest concentration of lower lying common which was left unenclosed when the field systems of the area were established. There are other more isolated lower lying commons, such as Mynydd-bach-y-Cocs (HLCA068), and other fragments of common, considered too small to be dealt with as separate areas, as is the case with the remnants of common in Pilton Cross and Pilton Green (within HLCA032) and Rallt and Wern Fabian, for example (within HLCA067). Unlike the higher commons, mentioned above, the commons located in the lower lying area are almost continuous, though interrupted in places by intakes, most notably the area around Wernllath (HLCA080) between Fairwood and Clyne Commons, Cilibion Plantation and the area on both sides of the road around Carterford Bridge, all of which were enclosed before Yate's map was drawn up in 1799. Some of these, particularly that at Wernllath, may relate to encroachment which was the subject of complaint(s) issued by the Earl of Worcester in the 1590s (Robinson 1968, 372), 375, 379). Encroachment onto former open common land, along the northern edge of Welsh Moor and Forest Commons (HLCA069), and Pengwern Common (HLCA071), and the western side of Fairwood Common (HLCA074) as evidenced by hedgerow analysis, documentary evidence and farm building episodes, appears to have been a notable characteristic of the 16th and 17th centuries, with large curving intake boundaries relating to farms established in this area, such as Bryncoch, Little Hills, Wimblewood, and Fairwood Corner Farm (HLCA067). These features typical of post-medieval encroachment onto open grassland now define the limits of common land.

They are now used for rough grazing, but there are areas of scrub and trees. Compared with the higher commons, there is little evidence for prehistoric use, with just a few barrows. Today the area is still moorland heath with some scrub and trees, and is still used for rough grazing.

Back to top of page

Deer Parks and Woodland

Two deer parks are known to have existed in medieval Gower: the more well-known being Parc le Breos, which was associated with the Marcher Lordship of Gower, and originated in 1221-32, operating as a deep park until around 1400, though in a reduced state since the demise of de Breos power in 1320 (Leighton 1999); another deer park existed at Clyne wood and is mentioned in de Breos's charter of 1306. The deer park at Parc le Breos (now split between HLCA064 and HLCA065) retains elements of its medieval pale. The other at Clyne is attested to by the place name 'Hen Barc' and documentary evidence alone, is thought to have been located somewhere between HLCA077 and HLCA078, but it is thought unlikely that it was ever physically emparked (Leighton 1997).

Significant areas of ancient woodland are found at Clyne Wood (HLCA078) and within the adjacent area of Hen Barc (HLCA077), within Parc le Breos (HLCA064), and also within Penrice Park and Penrice (HLCA047 and HLCA046), and to an extent within Cillibion (HLCA070). Indicators of formerly more extensive woodland are visible on the lower flanks of Cefn Bryn (within HLCA020, HLCA022 and HLCA028), including remnants of woodland, well-treed hedges, and woodland place name elements. Within Llanrhidian Higher (HLCA067) the highly visible matrix of woodland and enclosure has been shown to be the result of a mixture of assart, i.e. woodland clearance, and enclosure of open ground. This process of farm creation took place largely before c.1300 and after c.1550, with a noticeable upsurge in clearance occurring during the first half of the seventeenth century.

Elsewhere Ancient Woodland survives mainly within the sheltered river valleys, as within HLCA018 Cheriton and Burry Pill, Bishop's Wood (HLCA041), Ilston (HLC072), Ilston Valley (HLCA073), and the Bishopston Valley (HLCA086) and on coastal cliff slopes, for example the wooded cliff top at Oxwich Bay (HLCA049), Nicholaston Woods (HLCA055), and along the northern shore around Tor-gro and North Hill Tor (HLCA014).

Back to top of page

Settlement Pattern

Gower is relatively densely settled by a series of nucleated villages and hamlets, and a network of mainly small farms. The high number of nucleated villages is particularly notable. Many of these have distinctive features such as open space or greens, such as Reynoldston (HLCA037), Llangennith (HLCA011), Llanmadoc (HLCA007), Oxwich Green (HLCA048), and Burry Green (HLCA026); many are grouped around a medieval church. This nucleated pattern is so strong as to suggest that isolated churches may perhaps denote deserted medieval village sites. Church villages are Llanrhidian (HLCA022), Llangennith (HLCA007), Llanmadoc (HLCA007), Rhossili (HLCA031), Llanddewi (HLCA034), Penrice (HLCA046), Port Eynon (HLCA044), Reynoldston, (HLCA037), and Ilston (HLCA072). The presence of a green or open space may have more than one historical origin: some of these settlements also have a medieval church; others, like Burry Green do not, and may perhaps be the result of a later, informal process of settlement formation, as for example common encroachment. There is a suggestion of encroachment in the settlement pattern that occupies the periphery of Reynoldston (HLCA037), Penmaen (HLCA057), and Nicholaston (HLCA056), among others. The positioning of buildings may give some indication of settlement type: in Southgate (HLCA062) and Bishopston (HLCA086) for example are a series of houses set at right-angles to the road; this might indicate late encroachment onto what may have been unconsolidated holdings of a still operating openfield system. This pattern is seen elsewhere on the fringes of besanded landscapes, for example at Kenfig, and might represent post-medieval settlement expansion, or even indicate establishment of new settlement following periods of besandment leading to abandoning of original settlement locations.

Other settlements seem to have been built up as farm hamlets, or from squatter settlements: examples of these are Three Crosses, and West Town, Llangennith (HLCA011), and again, there will be others. The position of settlement is also often distinctive, with farms strung out along the margin of cultivated land and common or open grazing. This is particularly clear for example around Nicholaston (HLCA056) and above Llanrhidian (HLCA022). Some villages also appear to occupy these margins, such as Reynoldston (HLCA037) and Llanmadoc (HLCA007).

Swansea has exerted considerable influence on the shape of settlement at the eastern end of Gower. Here the villages all have strong 19th and 20th century suburban development (Bishopston, Kittle, Southgate in Pennard). These suburbs are characterised by occasional houses in an arts and crafts idiom, inter-war housing, mini-villas and semis and bungalows.

Back to top of page

The Economy of Settlement

The pattern of church villages and farm hamlets is mainly suggestive of traditional agricultural activity sustaining settlement, but there are other settlements where an industrial base is likely, or where for example numbers of chapels give indirect evidence for a late 18th - 19th century influx of population associated with industry, for example Burry Green (HLCA026), Knelston (HLCA035), Llanmorlais and Wernffrwd (both within HLCA067). Informal settlement development, such as at Murton (HLCA086), may also suggest this.

Another influence on settlement came in the later 19th century and was sustained through the 20th century: tourism and the seaside. Perhaps the best example of this is Port Eynon (HLCA044), where the architectural character of several of the house suggests their early use as boarding houses, also evidenced by house names. It may be difficult, however, to distinguish tourism from suburbanisation associated with Swansea: the suburban fringe at Llangennith (HLCA011) for example could possibly be attributable to either.

Back to top of page

The Building Profile

Chronologies of building: in some respects there is a long chronology of building on Gower, with a series of medieval castles and some important late medieval or Tudor houses, but early vernacular buildings are relatively rare. The majority of traditional buildings are 18th - 19th century in date and belong within the Georgian tradition. It is this, which emphatically contributes to the character of settlement in the area. Its characteristic features are compact planning, symmetrical or near-symmetrical facades of 2 or 3 window ranges in 2 storeys. There was a greater variety for later 19th century houses and into the 20th century, presumably under the twin influences of tourism and the growth of Swansea.

Social hierarchies: there is good evidence of high-status houses of medieval and later date, but it is a noticeable characteristic of Gower that there are overall relatively few high-status houses, and therefore quite a restricted range of house types. Thus large gentry houses are relatively unusual, but so too are small cottages. It is the middling small-scale Georgian vernacular, which predominates; particularly good examples of this are Crwys Farmhouse, Llanrhidian Higher, Plenty Farmhouse, Llangennith, and Big House Farm, Llanmadoc.

The principal gentry houses of the area are Penrice, Oxwich, Weobley, Fairy Hill, Stout Hall, and Kilvrough; with the exception of Oxwich and Weobley, these are 18th - 19th century in date. Below these may be ranked a series of villas (Fairwood Lodge, Cwrt Herbert, Glynhir, and Clyne Castle. The latter two were built as the homes of industrialists: whilst the former are also houses with little attachment to land.

There is however very little overt influence of landed estates over the building process, for example, with the pattern of building remaining vernacular even in the immediate vicinity of a gentry seat (Penrice, for example). Visible exceptions to this are the inn and school at Kilvrough, and there are clearly some estate farm buildings (e.g. Penrice Home Farm). Estate influence on architecture is more apparent in the churches, and there are several examples of gentry patronage for 19th century restoration work (Llanrhidian, Rhossili, Llanddewi, Penrice).

There are few examples of smaller vernacular buildings, but early examples include Henbury Cottage, Southgate, and Hareslade, the latter a 17th century single unit house. Others are Margaret's Cottage, The Nook and Briardene Cottage in Oxwich. These are all small storeyed houses; the old cottage in the grounds of Underhill Cottage, Penrice is single storeyed but was apparently estate-built. It is quite possible that others of these small houses were also estate-built or sponsored since it would not be unusual to find estates working as patrons of the vernacular.

Back to top of page

Architectural Development

Early vernacular: a regional Gower house type is identifiable by the 17th century; its distinctive characteristics were its plan and form, which was end-entry, and the presence of bed and stair outshuts. (Internally, the charnel box links this building style to the West Country). Examples of this type are not plentiful but include Tyle House Farm, Llangennith (17th - 18th century), Overton House and Old House, Delvid Farm (perhaps as late as 1750). There is also a series of hall houses, including the 14th century Glebe Farmhouse, Cheriton; Great Pitton Farmhouse, and Old Henllys were storeyed halls of the 16th and 17th century.

Buildings like these lie firmly within a vernacular tradition, and it is interesting to note the introduction of 'polite' architectural influences on local building styles. Pitt farmhouse is a good example illustrating this process as a 17th century centrally planned house. The Georgian style, which dominated building in the area from the mid-18th century through to the mid- 19th century gives strong local character, but is not a truly regional style. Polite architecture is even more obvious in the use of architects for gentry house construction from the mid-late 18th century.

Back to top of page

Building Materials

Stone is the dominant building material, and as in other stone-using areas, the critical points which affect the built character of the area are variations in the sources of stone, and variations in its handling and surface treatment. In Gower limestone predominates, though there appear to be other building stones in some areas (e.g. a sandstone conglomerate(?) in the Llanmadoc area). Significant variation comes mainly through the differential use of limewash and render, and it seems that it was most usual for domestic buildings at least to be limewashed. A distinction was often made between domestic and agricultural buildings in the way that they were finished, and these are subtle variations, which should be noted; distinctions of this sort could be an invaluable ingredient in design guidance for farm building conversions. A recent trend of stripping buildings sounds a jarring note, and may be based on a misunderstanding of traditional building techniques. For roofing, slate predominates, though there are some few examples of thatch surviving. There is also some use of tile and pantile, but this is highly localised, for example in Penrice village (HLCA046).

Back to top of page


There are now few well-preserved farmsteads and farm buildings on Gower, and many recent conversions. Listing has captured at least some of the best of what has survived, (at Crwys Farm, Llanrhidian Higher - a small planned farmstead; Home Farm, Penrice - a planned estate farm; Pitt farm, Penrice, a linear roadside farm). It is clear, though, that there are other farm building groups of good traditional character, which might warrant some form of protection or record (Pitt Sogs Farm, Penrice, The Beeches, Horton, Betlands Farm, Llanddewi, Hills Farm Reynoldston, and others at Middleton, and near Port Eynon).

From surviving evidence it is possible to suggest the strong influence of agricultural improvement during the nineteenth century, with the introduction of well-planned farmyards on mixed farms, evidenced by corn barns and cow-houses in usually small yards. Smaller farms appear to have been planned as linear ranges in which house and farm-buildings were in-line. One unusual feature is a line-up of house and barn, for example at Old Henllys, and Corner House, Rhossili, and an even more unusual example of a barn attached to a church at Llangennith.

A distinctive and relatively rare local agricultural building type is the circular domed pigsty: one example is listed (Pill House, Llanmadoc).

Back to top of page

Post-medieval Estates and Parkland

Gower contains several gentry estates, perhaps the most significant being that of Penrice Park (HLCA047), home of the Mansel Talbot family, the largest and perhaps most influential landowner on Gower during the 18th and 19th centuries. This site is an old established seat of local administrative power, with its medieval fortified precursors located close by. The focus of the park, the listed Penrice Castle Mansion (LB I, 11531), was built in 1773-7 for Thomas Mansel Talbot, to the design of Anthony Keck, architect, of King's Stanley, Gloucestershire. Records refer to Bath stone and stone 'brought from the quarries at Margam'. The masonry is by William Gubbings and the plasterwork by Thomas Keyte. Other characteristics include a late 18th century landscaped park and grounds with enclosed woodland, ornamental fishponds, and walled kitchen gardens, together with an impressive range of well-preserved ancillary estate buildings, such as an orangery, stables and lodges. Included within the area is a large area of associated remodelled farmland, estate buildings and large regular enclosures. Group value with the little altered adjacent village of Penrice (HLCA046) is further enhanced, however, Penrice Park has been surprisingly omitted from the Register of Parks and Gardens.

Other gentry houses of 18th and 19th century date with associated parkland or grounds include Fairyhill (HLCA027), Stouthall (HLCA036), and Kilvrough Manor (HLCA082), Clyne Castle (HLCA078), home to an industrialist, also boasts parkland grounds. All are included in the Register of Parks and Gardens. Fairwood Park, a Regency Villa, now with its golf course (HLCA076) appears to have been built by William Jernegan c.1827 for John Nicholas Lucas of Stouthall.

There are a number of late medieval/early post-medieval manorial estates in Gower, including Oxwich Castle (within HLCA048), Bovehill or Landimor Castle (within HLCA019), and Weobley Castle (within HLCA020); these appear to have retained a strong association with their adjacent settlements, and agricultural landscapes, all reverted to agricultural status by the 18th century, and for this reason have not been dealt with separately. The same is true of lesser manorial foci such as Nicholaston Manor (within HLCA056), and smaller gentry/industrialist houses, such as Cwrt Herbert (HLCA086), and Glynhir (within HLCA067), which effectively failed to develop and remained small by comparison.

Back to top of page

Industrial Development and Transport

Agriculture has always been, and still remains, the principal activity of the area, and related rural crafts and industries form a characteristic of the area's settlements, of which black smiths and sawpits are indicative. Industry is represented largely by coal mining in the northern and eastern section, and quarrying and lime-burning in the south (Toft 1988). The woollen industry was largely a cottage industry, though a mill is documented at Staffel Haegr, Llanrhidian (HLCA022; Cooper 1998, 82-3). Salt was produced at Port Eynon (HLCA044) from the 16th century, and there was an oyster fishery here in the 19th century. Port Eynon seems to have been the only village with a quay; there were numerous small landing places elsewhere at the coast, but at these boats would have beached at high tide and loaded at low tide.

South and west Gower remained largely agricultural, apart from the quarrying and processing of limestone for both local agricultural use and export, while exploitation of the coal deposits to the northeast of Cefn Bryn affected the parish of Llanrhidian (Higher division), as well as the area around Loughor, and the uplands beyond. The coal industry had its roots in the Middle Ages; the charter of 1306 granted the burgesses the right to mine coal for their own use (Williams 1990, 7), whilst excavations at Llanelen (HLCA067) have indicated that industry (charcoal burning, iron production, coal extraction) played an important role in the mixed agricultural economy of the region throughout the medieval period. It is considered that coal extraction during the period was, however, relatively small-scale and appears not to have produced a significant effect on the landscape; commercial mining seems to have begun in a small way at the end of the 16th century (Cooper 1986, 19), but its main development dates to the latter half of the 18th century (Williams 1980, 157). The earliest mines in the area were located in the area between Wernffrwd and Cilonnen and irregular coal working is noted in the area (HLCA067) throughout the early post-medieval period, based on a mixture of shallow primitive crop holes, bell pits and slant mines.

From the 18th century deeper mineshaft workings were sunk around Llanmorlais and Wernffrwd, though problems with flooding proved a major problem. During the early 19th century the coal mines of the area lost out to increased competition from mines closer to Swansea, such as around Kilvey and Loughor, and it was not until the arrival of the railway in the area in the 1860s that coal operations were rejuvenated. The coal industry peaked between 1880 and 1914 (Cooper 1986; Cooper 1998). In all some 16 collieries are known to have operated within area (now designated as the AONB) prior to the closure of the last during the 1950s. Little is now recorded of surviving landscape features relating to the coal industry apart from the early 20th century engine house at Penllwyn Robert; most appear to be generally late in date and do not appear to have been investigated in detail. The associated transport structure of the area, which comprises an 18th century haulage road, and later 19th and 20th century tramroads within the Morlais valley, is again poorly represented (Cooper 1986; Cooper 1998). Other early coal workings have been recorded in the Clyne Valley (Williams 1958, 17-21; Leighton 1997, 135-59), where exploitation of woodland resources, for example charcoal burning also took place. Also of interest is the 19th century chemical works at Lethrid (HLCA070), in close association with a charcoal burning site and surrounding woodland; this was a purely industrial concern, as opposed to agri-industrial.

In the limestone area of south, north and west Gower, lime burning was naturally important: this is reflected in the numbers of limekilns - several of which are listed. The lime burning industry which dates back to the medieval period, appears to have flourished during the post-medieval period, in particular during the first half of the 19th century in line with improvements in agriculture. However from the latter part the 19th century the industry appears to have been in serious decline with increased availability of artificial fertilisers, so that few continued operating beyond the turn of the 19th / 20th century. This characteristic is widespread within rural, though there are unsurprisingly particular concentrations along the coastal outcrops, from where limestone and lime products could be easily exported to other parts of Wales, Devon and beyond (Toft 1988b). Other examples of rural industry include water-powered corn milling, some dating to the medieval period, at Cheriton (HLCA0180, Llanrhidian (HLCA022), and at Park Mill in the Ilston Valley (HLCA073).

The other major component of the area's economy, the shipping industry, was largely concentrated at the mouth of the River Tawe in Swansea itself, and therefore does not register in this study. However there were also smaller shipping facilities at Port Eynon (HLCA044), along the north Gower coast at Llanrhidian (HLCA022), Wernffrwd, Llanmorlais, and elsewhere (HLCA067). The main export from the southern coast was limestone, but this was mainly carried out from open beaches, for example at Pwlldu Bay (Craig 1980, 466, 484; Locock 1996, 16). During the post-medieval period North Gower together with limestone exported coal in particular, from the collieries of Wernffrwd and the Morlais Valley within Llanrhidian Higher (Cooper 1986).

The 20th century saw an overall decline in heavy industry, accelerating towards the end of the century. Mining was particularly hard hit by difficult geological conditions, which led to the closure of pits, and ultimately to the replacement of deep mines by opencast, as being a more cost-effective option (Humphrys 1971). The decline of the coal industry can be traced during the latter half of the 20th century in the map evidence.

Back to top of page

Tourism and Leisure

Two landscape areas in particular are dominated by twentieth century leisure, both golf courses: Langland Bay Golf Course HLCA059, and Fairwood Park and Golf Course HLCA076, have been identified. Whilst some element of impact on settlement, for example at Port Eynon (HLCA044), Horton (HLCA045) and Llangennith (HLCA011) is noted from the late 19th century as being a direct result of the emergence of the coastal leisure 'industry', the major impact of tourism and leisure on the character areas bordering Gower's coast was greatest at the fashionable resorts nearest to Swansea, developed during the mid-late Victorian or Edwardian periods, such as Langland Bay (HLCA081) with its Victorian Villas and later beachfront development and Mumbles with its pier of 1898, which saw its hay day during the early 20th century (HLCA042). More recent leisure developments include caravan parks now features of Llangennith, Broughton and Hillend Burrows (HLCA009), Horton (HLCA045) and Oxwich (HLCA048), in particular.

Back to top of page