The Elan Valley

The Natural Environment

The rocks underlying the Elan valley historic landscape area are mostly composed of Silurian and some Ordovician shales, slates, mudstones and siltstones, together with beds of conglomerate. During the last Ice Age, between about 70,000 and 12,000 years ago, the area became submerged beneath a glacial ice sheet which had a considerable impact upon the present-day topography. Distinctive features of this period of glaciation are the smoothed and flattened upland plateaus, steep-sided, U-shaped glaciated valleys, morainic deposits, and hillside terraces and platforms where glacial meltwater has cut through the layers of stone debris deposited by glacial action. Glacial action disrupted the flow of the river Ystwyth which originally fed a lake in the area of Gors Lwyd, in the watershed between the Ystwyth and Elan, which in turn fed the river Elan, a tributary of the river Wye.

The historic landscape area falls into a number of distinct topographic areas which are closely reflected in the historic landscape character areas defined below. To the east is an undulating lowland area, mostly between 190–250m above sea level, extending from Rhayader to Elan Village and bounded by the mountains to the west and the river Wye to the east.

The lowest dam, at Caban-coch, was built at a natural bottleneck in the Elan valley, between the steep slopes of Craig Gigfran to the north and Craig Cnwch to the south. Formerly, the Elan and Claerwen joined about a kilometre upstream of the dam, the two rivers then diverging and snaking their separate ways into the moorland of Elenydd for a further 20 kilometres in the case of the Elan and 15 kilometres in the case of the Claerwen. The southern part of the moor is drained by tributaries of the Irfon, itself a tributary of the Wye, and the western part of is drained by the Ystwyth to the north-west, the Teifi to the west and the Tywi (Towy) to the south-west. Before the creation of the reservoirs the Elan and Claerwen occupied steep-sided and relatively fairly flat-bottomed valleys, often less than half a kilometre across, the floor of the valley rising from a height of about 250 metres above Ordnance Datum near their confluence to just under 400 metres in their upper reaches.

The moorland of Elenydd, which forms the central portion of the Cambrian Mountains, is dissected into three rough blocks by the river valleys. The southern part of the moor, to the south of the Claerwen valley, is the generally higher and more remote, with extensive upland plateaus at heights of between 400–500 metres and with peaks like as Drum yr Eira, Drygarn Fawr and Pen y Gorllwyn reaching over 600 metres from which both Cardigan Bay and the Brecon Beacons are visible on a clear day. The western part of the moor, between the Elan and Claerwen, again has extensive plateaus with somewhat lower peaks such as Bryn Garw, Trumau and Graig Dyfnant just over 500 metres high. The eastern part of the moor, overlooking the Wye valley, is generally lower though with a few peaks such as Moelfryn and Crugyn Ci of over 500 metres.

Two further distinct topographical areas in the study area are Carn Gafallt, an isolated upland block to the east of Elan Village and Cwm Dulas, a steep sided valley on the southern side of Carn Gafallt drained by a further tributary of the Wye.

The soils in the lowland area of Cwmteuddwr between Rhayader and Elan Village and in Cwm Dulas are generally well drained, fine loamy and silty soils with deeper silty alluvial soils, in places overlying gravelly subsoils, on the flatter land bordering the Elan. Similar soils would formerly have existed in the valley bottom along the lower Elan and Claerwen valleys before the construction of the reservoirs. The small lowland lake of Gwynllyn, to the north-west of Rhayader, has resulted from impeded drainage caused during the last glaciation. The soils on the steep slopes margins of the valleys are generally well drained fine loamy and silty soils, with enclosed fields on some of the flatter areas free of rock outcrops and screes. The Elenydd uplands are principally covered with blanket and basin peat up to 3 metres deep, with areas of loamy upland soils.

The blanket bogs on Elenydd are dominated by sphagnum, cotton grasses and heathers with bog pools and larger clear-water lakes such as Llyn Gynon, Llyn Fyrddon Fawr and Llyn Fyrddon Fach, especially on the northern and western areas of the moor. The lakes, which have formed since the last glaciation, are exceptionally up to 25 hectares in extent and some support populations of brown trout and other fish species. Drier moorland areas on the thin free-draining acid soils beyond the extent of the blanket bogs support a grassy heathland vegetation with bell heather, bilberry and gorse. These upland areas have traditionally provided grazing land of poor to medium quality, exploited during the summer months, and have seen little management other than for the purposes of wildlife conservation in some areas.

The free-draining soils on the more steeply sloping stream and river valleys such as in the lower Elan and Claerwen, in Cwm Dulas and on the eastern slopes of Elenydd overlooking Rhayader, support areas of broadleaved woodland often dominated by sessile oak or by ash in the case of ravines or more rocky slopes, together with heathers and gorse. The less steeply sloping ground in these areas have generally been cleared of natural vegetation and enclosed, perhaps from the medieval period, to create grassland of moderate quality. Some of these areas were planted with conifers, on a smaller scale in the early 19th century but more extensively when the reservoirs were created in the later 19th and earlier 20th centuries.

The well-drained soils on the flatter ground along the unflooded areas of the lower Elan and Claerwen valleys and in the lowland area between Elan Village and Rhayader are almost exclusively maintained as grassland today, though historically they are likely to have been used for the production of cereals and root crops as well as for grazing.

A picture of the vegetational and hydrological change of the Elenydd uplands and the surrounding valleys has been provided by analysis of peat deposits at the head of the Elan valley, at Llyn Gynon (Cardiganshire), in a valley to the north of the Claerwen reservoir at Esgair Nantybeddau, and at upland plateau sites at Pwll-nant-ddu, between Claerwen and Penygarreg and Bryniau Pica to the west of the Claerwen reservoir. These studies have shown that by about 8500 BC, following the end of the last Ice Age, that birch woodland became established over much of the area, with some willow perhaps on damper ground and sedges and grasses forming open heathland perhaps in some areas of high altitude. From about 8200 BC there was a rapid spread of hazel together with the beginnings of mixed broadleaved woodland, including oak and elm, possibly at all but the highest altitudes. Fragments of charcoal appearing in the peat deposits suggest the first indications of human activity in the period about 7500 BC, during the Mesolithic period. From about 6200 BC there was an increase in alder pollen, suggesting the onset of damper conditions and by about 5000 BC there was a transition to more open conditions with a rapid rise in heather, grass and sedge heathland, with continuing evidence of human activity spanning the late Mesolithic and earlier Neolithic periods possibly initially representing the hunting of game animals and later grazing by domesticated flocks. More extensive woodland clearance was taking place during the later Bronze Age, from about 1200 BC and there is some evidence to suggest cultivation at this time and during the subsequent Iron Age. By the later 1st century BC and the 1st century AD the pattern of vegetation had probably come to closely resemble that visible today. There is further evidence for the subsequent intensification of forest clearance affecting particularly alder, oak and birch woodland probably in the lower valleys which it is argued took place during medieval times and accompanied with evidence of pastoral activity representing an intensification of grazing in the area following the establishment of monastic granges of the Cistercian monastery of Strata Florida in the later 12th century.

Administrative Boundaries

During the earlier medieval period the greater part of the study area, to the north of the Claerwen, fell within the commote known as Cwmwd Deuddwr, (contracted to Cwmteuddwr) which derived its name from the two rivers Wye and Elan. The area to the south of the Claerwen fell within the commote of Dinan in the cantref of Buellt, which subsequently formed the hundred of Builth in Breconshire.

The western part of Cwmteuddwr fell within the commotes of Mefenydd and Pennardd which subsequently became the hundreds of Ilar and Penarth in Cardiganshire. From between about the 7th to earlier 10th centuries Cwmteuddwr had formed part of the kingdom of Powys, but in the later 10th and early 11th centuries it formed part of a separate territory between the Wye and Severn, known as Rhwng Gwy a Hafren which along with the kingdom of Brycheiniog showed allegiance to the royal house of Deuheubarth of south-west Wales.

From the end of the 11th century and throughout much of the 12th century ownership of the commote of Cwmteuddwr was to be disputed by the royal house of Deuheubarth and the Anglo-Norman marcher lords. Skirmishes between the Mortimers and Rhys ap Gruffydd (The Lord Rhys) had already occurred by 1176 when Rhys’s two sons-in-law, Morgan ap Meredith and Einion Clyd, Lord of Elfael were ambushed and killed in the woods of ‘Llawr Dderw’ near Rhayader in Cwmteuddwr. Rhys later invaded the territory and established the castle at Rhayader in 1177. In 1184 the greater part of the commote of Cwmteuddwr was amongst the extensive tracts of land that Rhys granted to the newly founded Cistercian monastery at Strata Florida, of which Rhys was the principal benefactor. The grant was made to the abbot of Strata Florida, before Rhys’s army, in the church of St Bridget at Rhayader. Rhys retained a small area of the commote, known as the manor of Cwmteuddwr, in the immediate vicinity of Rhayader, for the defence and provisioning of Rhayader castle. The part of the commote comprising Elenydd and the Elan valley became known as the manor of Grange.

Within a matter of a few years Cwmteuddwr fell into the hands of the Anglo-Norman Mortimer family, following which Cwmteuddwr came to form part of the marcher lordship of the cantref of Maelienydd. On the accession of Edward IV in 1462 it became a crown manor, and remained such until 1825 when it was sold. At the Act of Union of 1536 the former Mortimer commotes of Cwmteuddwr and Gwerthrynion were formed into the hundred of Rhayader in the county of Radnorshire.

Today, the eastern parts of the study area largely falls within the communities of Rhayader and Llanwrthwl, together with parts of the communities of Llanwrtyd Wells, Treflys and Llanafanfawr, in the county of Powys, created in 1974 by the amalgamation of Radnorshire, Breconshire and Montgomeryshire. The western part of the study area falls within the communities of Ysbyty Ystwyth, Ystrad Fflur and Tregaron in the county of Ceredigion.

The Grange of Cwmteuddwr was purchased in 1792 by a Wiltshire gentleman, Mr Thomas Grove, in 1792 who set about developing the estate. After changing hands several times the area of the grange within the watershed of the Elan and Claerwen rivers was acquired by the corporation of Birmingham in 1892 for the construction of the reservoir scheme, upon which work began in 1893. Responsibility for the reservoirs and the Elan Estate passed to Welsh Water in 1974 following the creation of privatized water companies, and remains the largest single area owned by a water company in Britain. In 1989 the Welsh Water Elan Trust was established, with responsibility for protecting the natural environment of the estate and encouraging public access and understanding. The southern part of Elenydd in Breconshire, known as Abergwesyn Common, now belongs to the National Trust. Carn Gafallt, a small separate upland area to the east, also in Breconshire, was purchased by the Royal Society for the Protection of Birds from the Glanusk Estate in 1983.

In ecclesiastical terms most of the study area fell within the medieval parishes of Llansantffraid Cwmteuddwr and Llanwrthwl within the deanery of Buellt, in the diocese of St Davids.

Early Perceptions of the Landscape

Early accounts of journeys by people travelling through the area, early maps and records of place-names provide a record of how the landscape of the Elan valley was described and perceived in the past by those who lived within it or visited it.

The mountainous area surrounding the head of the Elan has been known as Elenydd since early times. The name is probably being derived from the name of the river Elan with the ending -ydd, which implies ‘the area of’ to the name to which it is attached. It is first recorded in Gerald of Wales’s Journey through Wales, composed in 1188s, in which (in translation) he describes ‘the lofty mountains of Moruge, called Elennith in Welsh’. Gerald highlights Elenydd and Eryri (Snowdon) as the two principal mountain ranges in Wales and explains that the English name ‘Moruge’, a form otherwise unrecorded but probably derived from a word similar to the French marais (‘marsh’), relates to the marshlands on its summit. The name therefore has the meaning the ‘Marsh Mountains’, to distinguish it from the English Snowdon, the ‘Snow Mountains’. Gerald was accompanying Archbishop Baldwin’s journey through Wales, urging the faithful to join the Crusades, this particular leg of the journey being between Strata Florida abbey and the church at Llanbadarn Fawr. The Elan valley itself is first recorded as Glan Elan in the late 12th century and subsequently as Dyffryn Elan (Driffyn Elan).

A visit to the area, again in connection with Strata Florida, was made in the 1530s by John Leland, the king’s antiquary, who explored the countryside around the abbey, no doubt with local guides. He visited Llyn Teifi and the other pools on the western side of the moor, noting that they were ‘plentiful of trouttes and elys’. From a point two miles beyond the lakes he surveyed a scene that can have changed little since the 16th century: ‘I standing on Creggenaugllin [Carregyderlwyn?] saw no place within sight, no wood, but al hilly pastures’.

Further across the moor he reached, ‘Llyn y vigin velen’ (Llyn y Figyn), a name which he translates: ‘Y vigin is to say a quaking more. Velin is yellow’. The lake he says is ‘the colour of the mosse and corrupt gresse about it’. From here he went on to ‘Llin creg lloydon’ (Llyn Cerrigllwydion), presumably then taking the route known as the Monks’ Way on towards Rhayader. Just as today, it was the wildness and remoteness of the area which the visitor remarked upon: ‘Al the mountaine ground between Alen [Elan] and Strateflure [Strata Florida/Ystrad Fflur] longeth to Stratefleere, and is almost for wilde pastures and breding ground’.

The first maps of the area began to be published in the 17th century though little detail of the area is generally given. Morden’s map of South Wales published in Camden’s Britannia of 1695 it is simply marked as ‘Gwasted’ (gwastad, ‘waste’). Early 19th-century maps show a number of scarps around the head of the Elan and Claerwen, an area which at this date was often referred to as ‘Cwm Toyddwr Hill’, named after the parish of Cwmteuddwr.

Country ‘seats’ started to be shown on maps published from the early 18th century onwards, Cwm Elan being shown on Bowen’s New and Accurate Map of South Wales first published by subscription in 1729 (Hugh Powell of Cwm Elan being one of the subscribers). Both Cwm Elan and Nantgwyllt houses are shown on Thomas Kitchen’s map of Radnorshire of 1754 and on 1813 edition of Coltman’s Map of South Wales.

The place-names of the Elan Valley area provide a record of an oral tradition that developed over the course of many generations and provide evidence of the ways in which the landscape of the Elan valley was perceived and exploited by those who drew their living from it, complementing evidence from other sources. In the 1860s and 1870s Cwmteuddwr was the only parish in Radnorshire where Welsh was still commonly spoken and by the 1880s it was the only parish where it was actually still spoken. Consequently most early names are in Welsh, though some appear in anglicized forms. The most consistent record of names is that given on the larger-scale Ordnance Survey maps published in the later 19th century but many of the names are likely to have originated much earlier. The names are particularly significant for the mountainous parts of the area, especially since here as elsewhere in Wales the place-names often include early descriptive elements. It was in the valleys and on the valley edges that most people lived, and it is here that many of the names of farms and tenements first appear in written leases, wills and transactions from about the mid 16th century onwards, though some at least were probably much earlier. See the gazetteer for place-names within the study area.

Because of the particular topography of Elenydd, place-name elements distinguishing ridges (esgair, rhestr, trum/drum, cefn, crib) and slopes and ascents (llethr, rhiw, llechwedd, allt and lan) are particularly common, though the common terms fan ‘peak’, bryn ‘hill’, moel/foel, moelfryn ‘bald hill’, and mynydd ‘mountain’ also appear. The terms for hills and ridges appear to be some extent interchangeable, though drum tends to be reserved for the higher hilltops, over 500m, with the other terms generally below 500m, esgair being often used for the ridges between stream valleys and cefn for more extensive flatter upland areas. Terms for slopes likewise appear to be interchangeable, though lethr and rhiw tend to be used for the higher slopes whereas allt and lan tend to be more frequently applied to the lower slopes, below 400m around the margins of the uplands. Terms used to describe discrete eminences include talcen ‘forehead’, pentanau, ‘nobs’, clap ‘lump’, clapiau ‘lumps’s’, cnapiau ‘knobs’, cnwch (cnwc), ‘hillock’, uchelfa ‘high place, copa ‘summit’, crug ‘hillock, heap’, banc ‘bank, hillock’, talar ‘headland’, and more poetically, llofft ‘loft’, castell ‘castle’ and disgwylfa ‘lookout’. The frequent rock outcrops on the sides of steep-sided valleys are called carreg/cerrig ‘stone’, craig and creigiau ‘rock, crag’. Hollows are often indicated by the term pwll.

The broader river valleys are often referred to by the term cwm, several of the narrower stream valleys on the upland edge being distinguished by the terms ceunant or dyfnant ‘ravine’. Some of the shallower upland valleys bear the place-name elements pant ‘valley, dent’ and bwlch ‘pass, gap’, the heads of a number of the upland streams being called blaen ‘point, end, summit, source’.

Major rivers such as the Elan, Claerwen and Ystwyth are distinguished by the term afon, the term aber ‘confluence, mouth’ being restricted to the confluence of the Elan with a number of the major streams feeding it, including the Claerwen, the Afon Gwngu and the Nant Hirin. The majority of streams and brooks are called nant, of which almost 200 are named within the study area, the plural neint ‘streams’ being used for an area with many small streams near the head of the Ystwyth, though several smaller streams have the element ffos or ffrwd. The area includes numerous small waterfalls though none are called pistyll, possibly since they are so common, though the place-name Cwm Pistyll occurs in a single instance, and there is a single occurrence of the terms sgwd ‘flow, fall’. Ffynnon ‘spring, well’ is a surprisingly rare place-name element in the area, and the two named springs Ffynnon Fyw and Fynnon Mary both lie on the southern edge of the Elenydd, in the community of Llanafan Fawr. The small upland lakes and associated topographical features of Elenydd are invariably carry the place-name element llyn, as does the small lowland lake called Gwynllyn to the north-west of Rhayader. The confluence of the Elan and Wye south of Rhayader is known as Llyn Aberdeuddwr.

These topographical features are described by a wide range of descriptive terms including anatomical terms such as safn ‘mouth’, braich ‘arm, spur’, bron/fron ‘breast of hill’, troed ‘foot’ and gwar ‘neck’. Colour attributions are common and include coch/cochion ‘red’, wen/gwyn ‘white’, gwinau ‘auburn, brown’, rhudd ‘red’, llwyd/lwyd/llwydion ‘grey, brown’, melyn ‘yellow’, du/ddu/duon ‘black, gloomy’, and glas/las ‘green, blue’. The colour attributions tend to have reasonably distinct altitudinal ranges, du/ddu/duon tending to be above a height of 500m above sea level, llwyd/lwyd/llwydion and glas/las tending to be above a height of 400m, and coch/cochion, gwen/gwyn, and melyn all tending to be above 300m and below 400m. Relative positions and sized tend to follow the same place-name formulae found elsewhere in Wales, including ucha/uchaf ‘upper’, canol/ganol ‘middle’, isaf ‘lower’, ochr ‘side, limit, border’, perfedd ‘middle’, dan ‘under’, traws ‘across’, pen ‘top’, mawr/fawr ‘big’, bach/fach ‘small’ and bychannau/bychan, ‘little’.

A number of widely-separated place-names are repeated, as in the case of Llethr Melyn which appears twice, Banc Du which appears three times, and Lan Wen which appears four times, though it is perhaps significant that there is little duplication in the naming of streams which tend to provide of a framework for the naming of places throughout much of the area. In other instances closely-grouped names appear in pairs distinguished by colour or hue such as wen ‘white, du ‘black’ as in the case of Afon Claerwen (claer ‘clear, bright’ and wen ‘white’) and Afon Claerddu (ddu ‘black’) or size (bach ‘small’, mawr ‘big’) as in the case of Chawel Bach and Chwarel Mawr. The pairing of hill and stream names is frequent (eg Nant Cormwg and Esgair Cormwg), the sharing of place-names being extended in many instances to other neighbouring topographical to provide a word map of the moorland extending, probably originating as a device to distinguish between discrete unenclosed upland grazing areas up to between 0.5 to 3 kilometres across. The root name is often a stream or river, the means by which the moorland is often approached. The names Afon Gwngu, Abergwngu, Llyn Gwngu, Llethr Gwngu and Blaen Gwngu which define an area about 3 kilometres across near the headwaters of the river Elan. Nant Egnant, Cae Blaenegnant, Bryn Llyn Egnant, and Bryn Caeblaenegnant likewise define an area of upland about 1.5 kilometres across to the west of the Claerwen reservoir. Occasionally, the area names are derived from a particular settlement, as in the case of Treheslog, Creigiau Treheslog and Banc Treheslog, and even upon a colour, as in the case of Creigiau duon, Banc du, Chwarel du and Lan du, which are all to be found within a kilometre of each other.

The condition of the terrain is described by a wide variety of adjectives and nouns which generally emphasise the harshness of the landscape or its degree of exposure to the elements, such as gwyllt ‘wild’, llaith ‘damp’, sych ‘dry’, dyrys ‘difficult’, caled ‘hard’, garw ‘harsh, extreme, coarse’, chwefrin/chwefri, ‘wild’, wynt/gwynt, ‘wind’, eira ‘snow’. The terms melys ‘sweet’ and paradwys ‘paradise’, tawell ‘peaceful’, clyd ‘sheltered’ are less frequent and generally restricted to more sheltered locations, especially on south-facing slopes.

The shapes of the landforms are described by a wide range of terms, including cadenu ‘shaggy’, crychion/crych, ‘wrinkled’, cwta ‘short’, hir ‘long’, pica/bica ‘pointed’, crwn ‘stocky’, lled ‘wide’, cam ‘bent, crooked’.

The names of areas of semi-natural and natural broadleaved woodland together with a number of conifer plantations include the term coed ‘wood’, though this is infrequently used as a place-name element elsewhere, as are llwyn ‘grove, bush’, perth ‘bush, hedge’ and gelli/celli, ‘grove’. The most common tree species name to be mentioned is bedwen/fedwen, (plural bedw) ‘silver birch’ and there are only rare occurrences of onnen ‘ash’, derwen, (plural derw/dderw) ‘oak’, helyg ‘willow’, gelynnen/celyn, ‘holly’, and afallen ‘apple’, the latter predictably only occurring in a lowland context west of Rhayader. Other infrequent references to vegetation include draen ‘thorn’, eithinog ‘gorsey’, cors and mign (in the form figyn/fign) ‘bog’, hesgog/hesg, ‘sedges’, and brwyn ‘rushes’.

Several of animal names occur, particularly in upland areas, including the domesticated animals gaseg/caseg, ‘mare’, march ‘horse’, geifr/geifre, ‘flock of goats’, defaid ‘sheep’, ci ‘dog’, gwartheg ‘cattle’, anner ‘heifer’, ych ‘ox’, moch ‘pig’, hwch ‘sow’. These may provide an indication of the former land use but may alternatively have been used as terms to describe topographical features. Generally, however, the occurrences are so infrequent that few meaningful conclusions can be drawn. Bird names such as aderyn ‘bird’, cywion ‘chickens’, ceiliog ‘cock’, twrci ‘turkey’? (perhaps a corruption of dwrgi ‘otter’), gigfran/cigfran, ‘raven’, gwalch ‘hawk’ likewise only occur infrequently. The names of a number of wild animals also appear. The name of the river Elan, first recorded in the 12th century, is thought to derived from the Welsh elain ‘hind’, to describe the leaping and rushing character of the river before the construction of the reservoir. Other place-names in the area include the elements bwch ‘buck’ and carw ‘deer, stag’.

Historic land use and settlement information is provided by a number of place-names with significantly discrete distributions in a number of instances. Dol ‘meadow’ occurs as an element in a number of place-names on land below 300m in the lower Elan and Claerwen valleys and in the lowlands west of Rhayader, in areas where historically hay meadows are likely to have been created. Upland grazing, by contrast, is indicated by a widespread scatter of place-names with the element waun/gwaun ‘meadow, moor’ which are generally to be found on all but the highest ground, generally within the unenclosed moorland between 400–500m. Rhos ‘moorland’, probably also often indicating upland summer pasture, is represented by a smaller number of names with a similar altitudinal range, but reaching down to between 300–400m in the Elan valley itself, and sometime either within or close to the margins of the enclosed land. The distribution of both waun and rhos names, like that of encroachments are restricted to the more accessible areas of upland grazing on Elenydd. They are rarely found in the remote uplands in the south-eastern part of the area, suggesting that they relate to particular phase of settlement and land use history. The place-name element ffridd ‘mountain pasture, enclosed mountain pasture’ is notably absent from the area, suggesting either that its place is locally taken by another terms or that the particular traditional land use system which the term applies to elsewhere in upland Wales was locally less well developed.

Other place-name elements indicating fields, including cae, maes and erw are very infrequent in the area and appear in both upland and lowland locations. Crops are only rarely mentioned and include haidd ‘barley’ and gwair ‘hay’ both of which are included as elements in place-names around the fringes of the lower-lying ground, at heights of between about 300-400m.

Clearance and enclosure is indicated by several names. Llanerch/llannerch, ‘glade, clearing’ is represented by a number of place-names generally associated with dispersed settlements and generally confined to the now enclosed land in the lower Elan and Claerwen valleys. The terms garth ‘enclosure, garden, hill, ridge’, corlan ‘fold, pen’ fuches ‘herd, fold’, camlas ‘ditch’, clawdd, (plural cloddiau) ‘ditch, barrier’ also occur but are generally too infrequent for any positive conclusions to be drawn, except that as in the case of waun and rhos names, few of these forms are found in the remote uplands in the south-western part of the area. Magwyr (as in the stream name Nant y Fagwyr) may represent ‘wall, enclosure’ or more simply ‘rocky place’.

A number of significant place-names relate to the settlement history of the area which similarly tend to avoid the remote uplands in the south-western part of the area. Tyddyn ‘smallholding’, invariably contracted to ty’n, is often associated with existing dwellings within the enclosed land on the valley edge, generally at a height of between about 200-300m. By contrast, lluest ‘booth’ and less frequently hafod ‘summer house’, caban ‘hut, booth’, by contrast occupy somewhat higher ground above or just above the enclosed land, often at a height of 300-400m. The element hafod occurs much less frequently than in some other areas of Wales, suggesting that locally it is replaced by lluest. Both terms probably indicate seasonally occupied habitations, some of which have remained as permanently occupied farmsteads. The element bod ‘dwelling’, as in Bodtalog, is found in several upland valleys. The less specific place-name element ty ‘house’ occupies a fairly broad altitudinal range extending from valley bottom up onto the higher moors where its occurrence in place-names such as Esgair-y-ty may relate to former seasonal settlements. Higher status settlement place-name elements such as cwrt ‘court’ and neuadd ‘hall’ are infrequent and are predictably confined to the lower lying ground west of Rhayader.

Ownership or an association with individuals is represented by a relatively small number of place-names. Several personal names appears, as for example Dafydd-shon, Iago, Ifan, Ifor, Owen, Madog, Mair and Mary, Siencyn, and Steffan, and in other rare instances denoting associations with people with particular occupations or positions in society appear, such as esgob ‘bishop’, mynach ‘monk’, gweis/gweision ‘servant’, offeiriad ‘priest’, rhingyll ‘bailiff’, though in the case of gwyddel ‘Irishman’ and bleiddiad ‘warrior’ the association is probably legendary. A ridge on the southern part of the moor to the north of Rhos Saith-main is called Rhiw Saeson (‘English ridge’). The elements esgob and mynach are both confined to the lower-lying ground in the lower Claerwen valley and the lowland to the east of Elan Village and are related to the Cistercian grange which existed in this area before the dissolution of the monasteries in the 16th century. Supernatural associations are implied by a mere handful of place-names, with the elements diawl ‘devil’ and cawr ‘giant’. Two terms indicating boundaries, gororion ‘border’ and ffin ‘boundary’ both lie in the western part of the boundary, on the county boundary between Radnorshire and Ceredigion.

A number of antiquities are individually named which include a number prehistoric burial cairns with the place-name element carn (Carn Nant-y-ffald, Carn Wen, Carn Ricet, and Carn Pant Maenllwyd) and several standing stones with the element maen (Maengwyngweddw, Maen Serth, Maen Cam, and Saith-maen) which mostly lie above 400m and probably represent ancient patterns of land use, though in some other instances both of these place-name elements may refer to natural outcrops of rock. The date of most of the names given to antiquities is uncertain, but folklore attached to the cairn known as Carn Gafallt, on the hill to the south-east of Elan Village, is mentioned as one of the Mirabilia Britanniae (‘Marvels of Britain’) appended to the mid 10th-century compilation known as the Historia Brittonum, attributed to the 8th-century Welsh writer Nennius, which provides an early source for the Arthurian legend. In translation, the relevant passage is as follows:

‘There is another wonder in the region called Buelt [Builth]. There is a heap of stones, and one stone laid on the heap having upon it the footmark of a dog. When he hunted the swine Troynt, Cabal, which was a dog of the warrior Arthur, impressed the stone with the print of his foot, and Arthur afterwards collected a heap of stones beneath the stone in which was the print of his dog’s foot, and it is called Carn Cabal [Carn Gafallt]. And people come and take away the stone in their hands for the space of a day and a night, and on the next day it is found on its heap.’

The hunting of the ferocious wild boar Troynt, in this instance known as Twrch Trwyth, also appears in the 11th-century tale of Culhwch and Olwen which appears in the collection of medieval Welsh tales known as the Mabinogion.

Former quarries and mines are indicated in several instances by mwyn ‘ore, mine’, plwm ‘lead’, chwarel ‘quarry’, and gwaith ‘works’, the latter occurring near a known mining site. Other naturally-occurring resources which are possibly indicated by place-names include gro ‘gravel’.

Lines of communication including the roads and tracks and associated structures which cross the mountain are referred to by common terms such as sarn ‘road’, ffordd ‘way’, croes ‘cross, crossroads’, pont/bont ‘bridge’, and llidiart ‘gate’.

Prehistoric and Roman Land Use and Settlement

Early settlement sites have yet to be identified within the historic landscape area, though the environmental evidence and the large numbers prehistoric burial and ritual monuments clearly indicate that the area was already well populated and undergoing clearance during the later Neolithic and early Bronze Age periods, from perhaps at least about 3000 BC. So much so, that it is probable that the patterns of land use which become more familiar at later periods — all-year-round settlements associated with arable and meadows in the lower-lying valleys combined with the exploitation of upland pastures in during the summer months — had its origins during this period.

Characteristic burial and ritual monuments of the prehistoric period are the stone burial cairns, standing stones, stone alignments and stone circles which crown many of the peaks and ridges of the Elenydd plateau, though often absent from some of the less accessible central parts of the moor.

Some of the cairns are quite slight, though others are several metres high and form landmarks visible from afar. Some of the cairns appear to be structurally quite complex, with outer kerbs of upright stone slabs or small internal burial chambers known as cists. In some cases isolated cists are known, as though the burial mound has been largely robbed of stone, and in the case of a monument on Beddaufolau, on the hills above the eastern side of the Garreg-ddu reservoir, there are the remains of a much larger stone chamber, up to about 2.5 metres across, which has the appearance of a Neolithic chambered tomb. Many of the burial mounds remain in good condition, though some have unfortunately been robbed for stone in the past or have been disturbed for the creation of sheep shelters, possible shooting butts, superimposed by modern boundary markers, piled up into modern walkers’ cairns, or disturbed by beacons. It is fortunate that few have suffered the fate of one of the three cairns on Clap yr Arian, carted away for road metalling in 1910, during the course of which part of a stone axe-hammer was discovered.

A number of the cairns take the form of ring-cairns which are known from excavations elsewhere to have been used for both burial and ritual. Prehistoric ceremonial activities are also thought to explain various other enigmatic types of prehistoric stone monument which are known from Elenydd. Alignments of between three and eight upright stones up to about a metre and a half tall are known at Saith-maen, at Rhosygelynnen on the hills above the west side of the Caban-coch reservoir and at Nant y Llyn, on the hills west of Treheslog. One stone circle 25 metres in diameter, composed of 16 stones, is known at Crugian Bach, near Allt Goch, on the hill above the east side of Caban-coch reservoir, and there are the possible remains of a second circle at Bwlch y Ddau Faen. Standing stones have been recorded on various parts of the moor. Some are likely to be later boundary markers, but a number have been found in association with other prehistoric monuments and are therefore likely to be of Bronze Age date. Some of the taller stones now lie where they have fallen, including two large stones, one 3.7 metres tall, near the radio mast on Cefn Llanerchi. A further tall stone at Drum Nant y Gorlan, 2.7 metres tall has also now fallen. The white quartz stone at Pen Maen Wern, 1.5 metres high, is amongst the tallest still standing. The tall and prominently sited standing stone known as Maen Serth, on the hill above the Rhayader-Aberystwyth road across the mountain, is inscribed with a cross, possibly being a prehistoric standing stone ‘Christianized’ between the 7th to 9th centuries. It is traditionally held to mark the spot where Einion Clyd, lord of Elfael, was ambushed and killed at the hands of the Mortimers in 1176, the spot being known locally, according to the Radnorshire historian W. H. Rouse, as ‘The Prince’s Grave’.

Early settlement or other activity is indicated by a scatter of prehistoric artefacts found within the historic landscape area including a number of flint flakes found on the shoreline of the Craig Goch reservoir and several copper and bronze artefacts belonging to various phases of the Bronze Age. Most of the metal artefacts are weapons, and include a dagger or possible halberd found near Glannau Wood, west of the Garreg-ddu reservoir, an early Bronze Age ogival dagger, found in peat digging on Bwlch y Ddau Faen, south of the Claerwen, and a middle Bronze Age rapier found on Drygarn Fawr. Tools from the area include four late Bronze Age socketed bronze axes found in 1895 near the Caban-coch dam during the construction of the Elan valley reservoirs. The axes appear to have been found with part of a stone mould and may represent a bronze smith’s hoard. They were discovered beneath a mass of scree that had fallen from the precipitous valley side, being broken up for road metalling. More prestigious items, perhaps pointing to the presence of an elite within the local Bronze Age population is suggested by a number of middle Bronze Age gold ornaments, found in and on the margins of Cwm Dulas. They consist of a penannular gold ring or earring from Waun Sarn and the hoard of four middle Bronze Age gold torcs, found hidden under a small heap of stones in an area of rough pasture on the edge of the moorland of Carn Gafallt in the 1950s. Objects found associated with prehistoric burial mounds in the area are limited to part of an early Bronze Age battle axe found during the removal of the Clap yr Arian cairn mentioned above, made of dolerite from the Presely area, Pembrokeshire.

The great majority of the known prehistoric burial and ritual sites in the area are on the uplands of Elenydd, though there are suggestions that other similar sites once existed in the lower-lying valleys, having become ploughed down or cleared away in these more heavily cultivated areas. A possible complex of Bronze Age ceremonial sites has been recorded by aerial photography in the Elan valley, to the east of Coed-y-mynach farm, which include two or three ring-ditches, a possible henge monument and a possible pit circle, which it is likely represent Neolithic and early Bronze Age earth and timber equivalents of some of the stone monuments known from the uplands.

Clusters of upland sites are known in various places including those on Carnau Cefn-y-ffordd, Drygarn Fawr, Darren and Bryn. These complexes, together with more isolated burial and ritual monuments within the historic landscape area are likely to have performed a variety of roles within the developing landscape during the fourth to second millennia BC, between about 3500 and 1500 BC. Clusters of monuments may represent ceremonial foci within this landscape, and may indicate the activities of different family or tribal groupings within the area. The distribution of monuments from valley floor to mountain top suggests that a wide range of lowland and upland resources were being exploited by these communities by this time.

These early monuments will have become known and revered within the landscape and in some instances became the subject of folklore which would have helped to fix the place in memory. The 8th-century association of Carn Gafallt with the hunting of the mythical wild boar Troynt has been mentioned above. In the early 16th century John Leland noted that other antiquities on Elenydd were associated with Arthurian legend.

‘The first river that I passed over was Clardue [Claerddu] . . . . hard by were two hillettes, through the wich Clarduy passith, where they fable that a gigant was wont to wasch his hondes, and that Arture killid hym. The dwellers say also that the giant was buried therby, and show the place.’

The sense of place which these mythical and historical associations create is given in characteristic fashion in Ruth Bidgood’s poem ‘Gigant Striding’ in her 1996 collection, The Fluent Moment.

Between two little hills
a gigant striding was wont to wasch his hondes,
till Arthur killed him, for no reason known.Perhaps it was just for his gigantic
striding, that diminished the moor;
his great hands commandeering the stream —

for being huge, anarchic; sharing
ancientness and threat
of the desolate land.

As in later periods, rocks, stones and cairns became a means of defining the territories and resources claimed by neighbouring or rival communities and which through the course of many hundreds of generations would give rise to the pattern of parishes that had emerged by the early medieval period. This, in turn, is reflected in the structure of communities around Elenydd and the Elan valley at the present day, each of which characteristically represents a ‘territory’ extending from valley floor up to moorland plateau. The legacy of these early monuments in providing fixed points from which the landscape might be portioned from an early date is amply illustrated by the fact that the southern boundary of Llanwrthwl community, where it abuts those of Llanwrtyd Wells, Treflys and Llanafanfawr, passes through no less than nine burial cairns, which must have acted as territorial markers until more detailed mapping of the uplands became available in the later 19th century, long after their original purpose had been forgotten.

Burial and ritual customs underwent a dramatic change throughout Britain from about 1500 BC onwards, and until the end of the prehistoric period virtually no burial or ceremonial sites are to be seen within the landscape. The upland pastures and lowland valleys of Elenydd will have continued to be exploited throughout the later prehistoric and Roman periods, though we still await the certain identification of settlement sites of these periods. Early buildings throughout much of this period were probably of timber and thatch, which has left little visible trace at ground level. Early cultivation is generally likely to have focused on the most fertile and hospitable soils in the valleys and on the valley edge and therefore almost certainly obscured or transformed into more recent fields.

Despite the lack of dating evidence, however, it seems likely that settlement and land use during the broad span between the later prehistoric to early medieval periods, is represented by a number of early huts and stone clearance cairns which have been identified on the Elenydd uplands. Two possible settlement sites have been identified by fieldwork to the south of the Claerwen valley, a rounded enclosure 25 metres across on Esgair Gwar-y-cae with a round hut attached to the inside of the bank, and a rounded enclosure about 16 metres across apparently associated with three round huts between 10 and 11 metres in diameter. Clearance cairns have been identified on the hillside south of Cnwch and on the hillslopes to the north-east of Allt Goch. No doubt other sites of these types still await discovery.

The period of the Roman conquest of Wales is dramatically represented by the temporary Roman fort or ‘marching camp’ discovered on Esgair Perfedd as recently as 1966, represented by a low earthwork enclosure of characteristic ‘playing-card’ shape at a height of about 450 metres just to the south of the Rhayader to Aberystwyth turnpike road across the mountain. The camp encloses an area of just over 6 hectares and was built to house a force of about 4,000 men and their supplies in tented accommodation for perhaps only a matter of days. The fort is likely to belong to the period between about AD 74–80, and probably lay on a campaign route into present day Ceredigion or Montgomeryshire.

Little is known of the area during the later Roman and early medieval periods, though it is to be supposed that it much of it was claimed as part of the grazing lands of the emerging communities encircling Elenydd.

The Cistercian Grange

By the late 12th century practically the entire area formed part of the lands granted to the Cistercian abbey of Strata Florida by The Lord Rhys. Most of the historic landscape area falls within the Cwmteuddwr grange, the western part of the area falling within the Cwmystwyth grange to the north-west and the Pennardd grange to the south-east. Although few contemporary records survive there is some evidence which suggests how these monastic granges were managed in the Middle Ages.

As in the case of other similar Cistercian estates, grange centres were established in the most favoured areas for demesne cultivation or grazing by the abbey itself. Two such granges were established within the Cwmteuddwr grange, one in the area known as Llanmadog, now in the vicinity of the Elan Valley Hotel, with a subsidiary centre at Nannerth in a small valley bordering the river Wye just over 5 kilometres to the north. There are no surviving standing buildings associated with either centre, though the site of the grange chapel, Capel Madog (‘Madog’s chapel’), is known, in a field opposite the Elan Valley Hotel. Some remains of the building were still visible in the early 19th century, the architect Stephen Williams being instrumental in causing the track of the Elan Valley Railway to be slightly diverted to avoid it in the later 19th century. Other local place-names which suggest associations with the monastic grange include Nant Madog (‘Madog’s stream’), Llanmadog (Madog’s ‘enclosure’ or ‘church’), and Coed-y-mynach (‘monks’ wood), the name of woodland and a farm about a kilometre to the north. The section of valley in which Elan Village lies, about half a kilometre to the south-west is known as Cwm yr Esgob (‘bishop’s valley’) which also appears significant. Other names in the area which indicate an association with the former grange are Dol-y-mynach (‘monks’ meadow’) and Craig y Mynach (‘monks’ crag’), places which lie close to each other in the Claerwen valley, near Llannerch-y-cawr. Two other names which may possibly have early ecclesiastical associations include Nant Offeiriad ‘priest’s stream’, a tributary of the Nant Cletwr, to the west of Craig Goch, and Nant Rhingyll (‘steward/bailiff’s stream’) to the east of Garreg-ddu reservoir.

The principal resource of these monastic lands was undoubtedly based on the extensive upland pastures upon which cattle and, increasingly, large flocks of sheep were grazed, both by the monastic granges themselves and by those with holdings within its boundaries. As in the time before they were bequeathed to the abbey, the greater part of the income derived is likely to have been derived from customary rents and dues by those with holdings in the area, though because of the nature of the estate the returns from this source were probably relatively low. The greater part of their wealth was probably derived from the wool trade, the monks of Strata Florida being granted a licence by the Crown in 1200 to export their wool free of duty to France and Flanders. The grange centres were ideally placed to manage the flocks of sheep brought down from the eastern side of Elenydd: Llanmadog straddles the principal valley on the eastern side of the uplands, and Nannerth sited on one of the principal stream valleys, the significantly named Nant y Sarn, which gives direct access to the northern areas of the moor from the Wye valley.

It is probable that some crops were grown on cultivated demesne lands at the Llanfadog grange centre and there seems a possibility that the regular field system along the north bank of the river Elan between Coed-y-mynach farm and Noyadd owes its origin either directly or indirectly to the medieval monastic grange. Low ridge and furrow has been noted in the fields opposite the Elan Valley Hotel which may possibly be of medieval date. Other holdings within the grange boundary, especially within the valleys and on the lower-lying ground bordering the Elan and Wye, are likely to have had cultivated fields.

Other resources in the locality available to the grange would have included the oak woodlands which must have once bordered the river valleys, supplying both fuel and building materials, and fish from the rivers and from the upland pools on Elenydd. Both Llyn Teifi and Llyn Fyrddon Fawr supplied eels and trout to the monastery.

Most of the other Cistercian granges in Wales were managed by lay stewards. At the time of the dissolution of Strata Florida abbey in 1539, the Cwmteuddwr grange was unusual in being managed by a monk-bailiff. A further unusual arrangement, evidently current in Cwmteuddwr in the earlier 19th century, which it has been suggested might have had its origin in monastic practice, was the custom of landlords within the parish in effect renting out their flocks on a yearly basis to the tenants living on their land. The rigour with which the monastic lands in Cwmteuddwr were managed at the time of the dissolution of the monasteries is brought into question, however, by John Leland’s observation about Elenydd in the later 1530s that ‘everi man there about puttith in bestes, as many a they wylle, without paying of mony’.

The Break-up of the Monastic Grange

The grange remained in the hands of Strata Florida abbey until its dissolution in 1539 in the reign of Henry VIII, when the crown took possession of all the estates belonging to it. The lands remained in the possession of the crown for a number of years, being leased by various parties until the manor was acquired by Sir James Croft of Croft Castle and Thomas Wigmore of Shobdon, who then proceeded to resell it in lots. Much can be learnt from the extent of the monastic holdings, the tenements already in existence and the nature of their economy from the various deeds and leases relating to the splitting up of the estate in the later 16th century. Some of these documents refer back to tenancy arrangements made by the abbot of Strata Florida in the first decade of the 16th century.

Encroachments on the Elenydd Moorland

Several of the isolated encroachments on the Elenydd moorland were probably also already in existence by the 16th century, a number having possibly originated a matter of centuries earlier. Today, these encroachments are represented by islands of enclosed fields within the moorland between about 3 and 35 hectares in extent. They are confined to the more accessible and agriculturally more favourable parts of the moorland, on the fringes of the moor or within easy access of the Elan and Claerwen valleys, generally on the sheltered south or east-facing slopes of stream valleys. Established in the first instance with or without the authority of manorial courts, these kinds of encroachments on the common were often regularised by the imposition of annual fines which became converted into rents or sometimes into freehold tenure.

Many of these upland settlements possibly originated in the medieval period as a temporary summerhouse (hafod), occupied between about May and August or September, in order to exploit upland pastures at some distance from the home farm. Some were later to become permanently occupied farmsteads, of which a proportion have survived to the present day.

The names of a number of the encroachments on Elenydd include the significant element lluest, which locally appears to take the place of hafod to describe an impermanent dwelling of some kind. Two neighbouring and now abandoned farmsteads on the uplands to the west of Graig Goch reservoir, for example, are called Lluest-aber-caethon and Lluest-Calettwr, named after the streams that they lie next to. Several other upland farms are named more simply after streams or rivers, such as Aberglanhirin, Abergwngu and Claerwen. The place-name element hafod does occur, as in Cwm yr Hafod (‘hafod valley’) and Esgairhafod (‘hafod ridge’), signifying topographical features associated with hafodydd. Other upland habitations are suggested by the place-name element ty (‘house’), as for example in the names Esgair y Ty and Gwar-y-ty.

The named encroachments and upland farms tend to be those which were still inhabited when the first accurate maps of the area were produced in the later 19th century. There is a range of archaeological evidence that there were many more hafodydd and other similar habitations that were abandoned before the first detailed maps of the area were produced. Few if any of these sites and their form and dating is often obscure since in some instances the site of earlier structures have been superimposed by later buildings. The earliest buildings of this period on Elenydd appear to be represented by a number of rectangular platforms, set at right-angles to the slope, where timber buildings were erected. In some instances, as in the case of a platform on Craig y Lluest, traces of drystone walling are visible which seem to represent low sill walls on which timber structures of this kind were built.

A number of hafodydd on Elenydd are first documented in the 16th century, but may have originated much earlier. Typical of the leases of this period was one in 1579 for the tenement and lands of Come Coill [Cwm Coel, now on the western edge of the Garreg-ddu reservoir] which was let for keeping 40 cattle and 100 sheep. The ancient custom of moving to temporary summer homes in the uplands during the summer months to take advantage of the upland pastures is referred to in a lease drawn up in 1585 for ‘a messuage or tenement called Y Brith come [cwm] Ycha, together with one somer house, called Y Clettwr mawr, sometime parcel’, evidently referring to a lower-lying permanent house in the valley of Nant Brithgwm, on the west side of Penygarreg reservoir and a summer house probably to be identified with the ruined upland farm at Lluest-Calettwr, in the hills about 3 kilometres to the north-west, near the head of the Nant Cletwr stream. In the later 1530s John Leland records that near the Claerddu stream, at the head of the Claerwen valley, he saw ‘two veri poore cottagis for somer dayres [summer dairies]’, yet to be identified, which no doubt produced dairy products that were marketed in the surrounding villages.

Late Medieval and Early Post-medieval Farmsteads

Many early timber buildings are likely to have disappeared without trace, though the history of the former farmhouse at Ciloerwynt (Cilewent) may be typical of many of the smaller later medieval smallholdings in the area. The house lay within an encroachment in the Claerwen valley first recorded in a document of 1568 in which the owner is styled ‘yeoman’ farmer. The original house (moved to the Museum of Welsh Life at St Fagans in 1955 and replaced by a new bungalow) perhaps began life as a cruck-built single-bay hall which has been shown by tree-ring dating to have been built in about 1476, though it is uncertain whether the original outer walls were of timber or stone. The later form of the building, with a lintel dated 1734, took the form of a single-storey longhouse with a living-room and fireplace at one end and with accommodation for cattle, calves and horses at the other, perhaps mirroring the form of the 15th-century building.

Stone was probably becoming a more common building material from the later 16th century onwards. Other stone-built upland longhouses appear to have been built during the later 16th to earlier 18th centuries, again with living rooms at one end and with accommodation for animals at the other, a number of which, such as Bryn Melys, Lluest-fach (Llwst-fach), Penglaneinon, continued in habitation until the later 19th and earlier 20th centuries when they were abandoned. Writing in 1880, the Reverend R. W. Banks made the following perceptive comments:

‘the site of the old enclosure, as the names show, was generally selected by the side of one of the brooks, which run from the higher ground, and feed the rivers of the district, with a view to turn out a flock of sheep on the soundest portion of the extensive pasturages which the wastes afford, and at the same time obtain shelter from the steep hill sides. Rude, yet substantial dwellings, constructed of the large schistous flagstones of the district, with a chimney shaft of some pretensions . . . still in many instances remain, and wear the appearance of buildings which may have existed in the sixteenth century.’

Like Ciloerwynt, many of the lowland farms and smallholdings and the fields surrounding them which are shown on maps of the later 19th century were clearly already in existence by the later 15th century and early 16th century. The subsequent 18th- and 19th-century country house at Nantgwyllt in the Claerwen valley, for example, is first recorded as a tenement of ‘Y nauntgwyllt’ in a lease of 1568, its name taken from the adjacent stream. Mention being made of four neighbouring farms in the valley and in a deed of 10 October 1579 by which the then owners of the Grange of Cwmteuddwr, Sir James Croft and Thomas Wigmore esquire, granted to Howell ap John ap Howell, gentleman, for £110 the yearly rent of 6s 8d ‘Aber Nant Guilth’ (Nantgwyllt), ‘Aber Elan’ (Aber Elan), ‘Pen Glan Eignon’ (Penglaneinon) and ‘y Kayhayth’ (Cae-haidd).

The surviving longhouse at Llannerch-y-cawr in the Claerwen valley, now owned by the Elan Valley Trust, may be typical of the more substantial lowland farms of the later 15th and 16th centuries belonging to an emerging late-medieval gentry class. This stone-built longhouse was remodelled in the 17th century and later centuries from a late 15th- to early 16th-century timber cruck-built two-bay hall with a central open hearth and wooden outer walls built on a platform up and down the slope of the hill. One of the most distinctive features of the longhouse was the juxtaposition of living rooms at the upper end and cattle byre or cow-house at the lower end, typical of the later medieval and early post-medieval cattle farms in the area. A smaller, single-bay hall, of similar type is known at Nannerth-ganol in the Wye valley, just outside the historic landscape area, which has a tree-ring date of 1555/56. The farmhouse, again with a byre at one end, is probably more characteristic of the smaller tenanted farms in the area from the later 15th and 16th centuries. Other farmsteads that take this linear form are Nannerth Ffwrdd (just outside the historic landscape area) and at Ty’n y waun, west of Rhayader. The small farm at Cnwch, south of Caban-coch dam, may represent another typical layout, probably of later origin, in which small buildings are grouped informally around a yard.

Growth of Landed Estates in the Early Post-medieval Period

The break-up of the monastic grange in the later 16th century gave rise to a number of the landed estates that were to dominate the history of the area in the 18th and 19th centuries. In 1579/80 the Howells of Nantgwyllt were acquiring land locally including ‘Groy gwonyon’ (Gro Hill lies on the eastern side of the valley), and ‘talke y rose y gelyne’ (Rhosygelynnen lies to the west) probably representing enclosed hill land within a kilometre of Nantgwyllt, on the margins of the moor. In 1581 he was disposing of Aber Elan and Penglaneinon, though in 1585 he was acquiring land further afield including ‘Dolfola’ (Dol-falau, a farm 2.8 kilometres away in the Elan valley, now submerged below the Garreg-ddu reservoir) and ‘Blaenllyngwynllyn’ (a farm near Gwynllyn, about 7 kilometres to the north-east), much of the land here and elsewhere tenanted to other farmers. By the time of his death in 1597 he was a substantial owner of land and mills in Radnorshire and north Breconshire, with a dozen farms in Llanafan Fawr, ten in Llanwrthwl in Breconshire, over twenty properties in Cwmteuddwr. His marwnad (‘elergy’) was sung by the prolific poet and genealogist Lewys Dwnn. His elder son became High Sheriff of Radnorshire. Much of his wealth was derived from his herds of cattle and flocks of sheep, the commercial value of the latter principally based on wool, horses and some grain. His descendant, Howell Powell, was styled ‘Gentleman, of Nantgwyllt’ in the 1670s.

Population Growth During the 16th and 17th Centuries

The local population continued to expand during the 16th century and 17th centuries. New tenements continued to be created from the moorland well into the 17th century, if not later. In 1674, for example, a grant was given for ‘that new cottage in Clarwen [Claerwen], then lately built . . . upon Y Eskyrne y Guion [?Esgair y Guon, not closely located], being a common or waste within the lordship of the Grange, with all inclosures thereto, and with liberty to inclose on the same common, not exceeding sixteen acres’. Despite this, however, the nature of the agricultural economy of the area which had emerged in the later medieval period was probably to remain little changed until the middle of the 18th century. In contrast with the houses of the gentry, however, many households probably survived at little more than subsistence level. Taxation records known as the Lay Subsidy assessment for ‘Comelan’ (Cwm Elan) in 1544 suggests about taxable 21 households, though there were other poorer ones which were not assessed. Twice this number of households are listed in the district of ‘Diffrin Ellin’ (Dyffyn Elan) in Hearth Tax return of 1670, the average number of hearths per household suggesting that it was one of the least prosperous areas of a county which was already poorly off in comparison with its neighbours.

Peat Cutting and Other Common Rights

The 16th- and 17th-century farms and tenements of Elenydd and the Elan valley will have had common rights probably of great antiquity deriving from the Cwmteuddwr grange and the ancient manor of Builth. These will have included grazing rights for sheep, cattle, ponies, and rights for the digging of peat (turbary) and for taking wood for fuel or building repairs (estovers). The date at which the cutting of peat on the moorland commenced is uncertain, but it is likely that it only came into its own as a source of domestic fuel once sources of readily available wood had become exhausted. Evidence for peat cutting is widespread on Elenydd, as for example on Gwar y Ty on the northern side of the moor, on Waun Lydan, to the south of the Claerwen, and between Allt Goch and Y Gamriw, to the east of Caban-coch reservoir, and can often be most clearly identified by aerial reconnaissance. Artificial platforms on which peat may have been dried have been detected by fieldwork in some instances, including ones on Rhos Saeth-maen. It was often the more accessible and consequently often the shallower peat deposits that were exploited in order to lessen the burden of carrying it away from the hill. In some areas it is evident that each farm or a group of neighbouring farms had its own turbary, approached by trackway, which must have been used over the course of many years. Crossing the moor by way of the turnpike from Rhayader to Aberystwyth in about 1820 the scientist Michael Faraday noted the presence of ‘a turfcutter or a peat digger here and there drew the eye for want of a better object’. Jonathan Williams, writing in the decade before noted that the hills in the parish of Cwmteuddwr

‘contain turbaries which supply the neighbourhood with the most excellent peat. This kind of fuel when dried by the joint action of the sun and the wind becomes a black and hard substance, make a cheerful fire, reflects great head, and is little inferior to coal. A peat pit is three feet deep and more, and often contains branches and trunks of trees’.

Peat cutting had all but come to an end during the second half of the 19th century, locally coming into direct competition with the coal from depots at Rhayader following the opening of the Mid-Wales Railway in 1864.

Corn Mills, Fulling Mills and Saw Mills and Corn-drying Kilns

Locally produced grain, wool and timber is likely to have been processed locally up to about the end of the 17th century, all based upon water power. One of the earliest references to the local corn is in a lease of the farm at Ciloerwynt, in the Claerwen valley, made in 1569. According to the custom of the times, the tenant of Ciloerwynt was required to take any corn he had grown to be ground at his landlord’s mill.

The site of this mill is uncertain, but it may have been the same mill in Cwmteuddwr mentioned as belonging to the Howell family of Nantgwyllt in 1597. This may have been on the site of the former corn watermill known at Melin Gwynllyn, now known as Upper Mill, on Nant Gwynllyn stream, to the north-west of Rhayader, which appears to have been in operation from before 1670 up to about 1900, when it consisted of a three-storey timber structure. Melin Gwynllyn operated as a woollen mill, first mentioned in 1710. Another woollen mill is represented by the converted mill building known as Walk Mill, just down stream, about which little has been written. A further woollen mill, a stone-built structure known as Fron Factory, for carding or fulling, was in operation about half a kilometre upstream until about 1840. Ttraces of its leat are still visible.

A second corn watermill known as Gro Mill, belonging to the Nantgwyllt estate, was in operation from at least 1806 until the time it was flooded by the rising waters of the Caban-coch reservoir in 1892. The mill was sited just below the confluence of the Elan and Claerwen rivers, and took its power from a leat drawn from the river Elan to the south of Cwm Elan house. At one period the a corn-drying kiln was attached to the mill, which also operated as a saw mill.

The sawmill complex until recently run by the Elan Valley estate on the western edge of the Caban-coch reservoir, alongside the Nantgwyllt stream and just to the south of the former Nantgwyllt house dates from about the 1920s and includes a complex of timber-framed, brick and corrugated buildings, and includes the remains of a tracked lifting crane, now unused.

Local corn-drying kilns are suggested by place-name evidence were evidently in use by at least the 17th century though no surviving structures have been identified. A late 17th-century deed relating to Rhydoldog, for example, mentions ‘Kaer odin’, a field-name derived from the Welsh cae’r odin (kiln field), which probably refers to a corn-drying kiln.

Earlier Stone Quarries

Small stone quarries are to be seen here and there, particularly around the fringes of the Elenydd uplands as for example near Rhydoldog, on Tremblyd and on the hills above Llannerch-y-cawr which were most probably in use from the later 16th century onwards for the construction of stone buildings and walls. The small quarry at Waun Geufron possibly may have supplied material for the Rhayader to Aberystwyth turnpike road. Other larger quarries at Cigfran and Cnwch, opened for the construction of the reservoirs are mentioned below.

The 18th- and 19th-century history of the Elan valley is dominated by the landed estates and gentry houses that had emerged during the course earlier centuries, focused in particular on the gentry houses that had arisen at Nantgwyllt and Cwm Elan.

The Nantgwyllt estate had passed from the Howells (by now anglicized to Powells) to the Lewises and the Lewis Lloyds and continued to expand, the 18th century witnessing the acquisition of such properties as Cnwch, Cwm Esgob and Perthi Llwydion (Perthyllwydion), and the construction of the small proprietary church serving the estate at Nantgwyllt. A number of the older estates were to be purchased and developed as investments by ‘gentleman farmers’ formerly unconnected with area. The Grange of Cwmteuddwr was purchased by Mr Thomas Grove of Ferne House, Donhead St Andrew, Wiltshire in 1792 from John Jones of Hafod. Grove was described by his contemporary, the Radnorshire historian, the reverend Jonathan Williams, as ‘a Wiltshire gentleman, who purchased 10,000 almost worthless acres, which he is now converting into a paradise’. The estate was no doubt attractive on a number of grounds: Cwm Elan could be made an attractive summer residence, away from the family seat in south-west Wiltshire; its wild scenery of mountains and waterfalls had become fashionable; it offered a challenge in agricultural terms of introducing new farming policies for increased productivity; and it presented opportunities for hunting which would have been attractive to a keen huntsman such as Grove. There was also the prospect of revenue from lead mining which was proving profitable just across the hill in Cwmystwyth.

In the early years of the 19th century Cwm Elan is described by Jonathan Williams as being

‘situated on the left bank of the Elan, in a narrow vale, surrounded by the hills, some of which are inclosed and cultivated, and studded with convenient farm-houses seated at proper distances, whilst others are entirely covered with groves of oak from their summit down to the waters-edge’.

Indeed, so impressive was the scenery that Jonathan Williams himself considered that ‘nor is there another parish in this county or perhaps in the Principality itself, that can exhibit more romantic scenes of Nature than those well wooded, watered, and rocky yet fertile districts’.

The agricultural improvements being introduced by landlords such as Thomas Grove are alluded to in the blank-verse poem of 350 lines entitled ‘Coombe-Ellen’ (Cwm Elan) published in 1798 by William Lisle Bowles, embracing the romantic movement of the late eighteenth and early nineteenth century. In the words of Desmond Hawkins, ‘the poem reaches a climax when the poet ends his contemplation of the awesome solitude and pristine beauty of untouched Nature with the thought that human cultivation can further enhance the scene’, probably to be taken as a reference to the new arable rotation and clover which were being introduced by improving farmers and landowners at this period:

. . . . . . . . . ‘here I bid farewell
To Fancy’s fading pictures, and farewell
The ideal spirit that abides unseen
‘Mid rocks, and woods and solitudes. I hail
Rather the steps of Culture, that ascend
The precipice’s side. She bids the wild
Bloom, and adorns with beauty not its own
The ridged mountain’s tract; she speaks, and lo!
The yellow harvest nods upon the slope;
And through the dark and matted moss upshoots
The bursting clover, smiling in the sun.
These are thy offspring, Culture!’

Thomas Grove of Cwm Elan was to become one of Walter Davies’s local correspondents in his report on Agriculture and the Domestic Economy of South Wales, published in 1815 on behalf of the Board of Agriculture. Indeed, on Davies’s first visit to Radnorshire in 1802, he entered the county via Cwmystwyth and headed straight for Cwm Elan, describing Groves as ‘an improver on the Wiltshire system’, then taking in the estates at Nantgwyllt and Noyadd before proceeding to Rhayader. Notes compiled by Davies make references to some of the improvements that Grove had introduced to sheep farming, including the folding his flock and the crossing of native breeds with the Southdown. He also noted the lead ore discovered and being worked on Grove’s estate. The property was transferred to his son, Thomas Grove junior before about 1809 when he succeeded his father as High Sheriff of Radnorshire. The son, like his father before him, continued to make Cwm Elan his residence for about three months each summer, and maintained the agricultural improvements that his father had set in train. In 1811, Walter Davies noted with some relief that the rumours suggesting that the Cwm Elan estate had deteriorated since the son had taken over were unfounded.

The folding of sheep introduced by Grove upon his estate in the late 18th and early 19th century may have involved the enclosure of selected areas of the moorland and the grazing of crops grown for fodder, perhaps including the ‘bursting clover’ referred to by Bowles, which would have had the advantages of improved nutrition for the livestock, controlling breeding, and improving soil fertility through manuring. These and other improvements had become commonplace towards the end of the 19th century. The Reverend R. W. Banks, writing of Cwmteuddwr in 1880 noted that although the intermixing of herds legally prevailed ‘the practice has been for each tenant to secure a distinct sheep walk and maintain his rights on it by keeping a strong flock, with as little change of sheep as may be’. Writing in about the first decade of the 19th century, Jonathan Williams praised the quality of the animals produced by the local farms:

‘many boast of a produce and stock scarcely to be surpassed for quality and usefulness in any part of the county. The sheep of this parish, which depasture upon the hills the whole of the year, are second to none in the Principality of Wales for symmetry of form, and a sound constitution; and the flavour and delicacy of their flesh are not surpassed by English venison, whilst their wool is admired for its fineness and sought for by the manufacturers of cloth’.

Until about the middle of the 18th century the wool from the district was probably processed locally, but after that date is probably supplied the burgeoning woollen industry in Newtown and Welshpool in the Severn valley in Montgomeryshire.

Most of the cattle in Cwmteuddwr at this period were kept ‘to supply the demands of the dairy’, and significantly a number of farmhouses such as Ciloerwynt were extended in the 18th century to include a separate dairy. Jonathan Williams acknowledged that the ‘small black and brindled cows, the aboriginals of this part of the county . . . have of late years undergone great improvement’ by crossing with Shropshire and Staffordshire breeds, but he was critical of the introduction of beef producing breeds such as the Hereford which (though disputed by other contemporary writers) he considered were ‘ill adapted to contend with the frequent inclemency of the weather, or to thrive with a scanty herbage, or in short to fulfil those purposes to which the climate of this county and the nature of the soil are more peculiarly adapted’.

The period from the 18th to the 20th centuries throughout Wales, however, witnessed an increasing emphasis upon sheep farming and a lessening in importance of lowland cattle farms and upland dairies alike. Longhouses with a byre became redundant, being replaced, if at all, by separate farmhouses and cattle sheds. The more inhospitable hafodydd and lluestau were abandoned, flock management on the Elenydd moorland during this period being marked by the erection of drystone sheepfolds, often with multiple pens, by occasional sheep shelters and by small drystone shepherds huts and shelters, sometimes built out of the ruins of the earlier summer houses. Shearing was carried out at the lowland farms in the spring, various of the larger farms and estate centres such as Nantgwyllt having barns where the wool was stored before being sold to dealers. This is possibly amongst the buildings shown in a drawing showing shearing under way in Eustace Tickell’s book, The Vale of Nantgwilt, published in 1894.

Early Woodland Plantations

Ordnance Survey maps of the later 19th century identify a number of relatively small woodland plantations on the sides of the Claerwen and Elan valleys, that were no doubt amongst the investments being made by landlords and tenants from the late 18th century onwards. Eustace Tickell’s 1894 essay, published just before the flooding of the valleys describes Nantgwillt as being ‘backed by wooded slopes of oak spruce and larch, interspersed with towering groups of Scotch fir’, which are also shown sketches which accompany his essay, some of these plantings evidently having been ornamental.

Hunting

The landscape being developed by some of the larger landowners in the area in the later 18th and earlier 19th century was also taking the huntsman into account. William Lisle Bowles’s poem, referred to above, seems to alludes to Thomas Grove’s love of hunting:

‘All day, along the mountain’s heathy waste,
Booted and strapped, and in rough coat succinct,
His small shrill whistle pendent at his breast,
With dogs and gun, untired the sportsman roams’.

Hunting was a pursuit shared by other gentry of the area. One of the sons of the Lewis family of Nantgwyllt was to be commemorated in the popular early 19th-century hunting song ‘Cwn Squeir Lewis Nantgwillt’ (‘The hounds of Squire Lewis of Nantgwyllt’). Jonathan Williams noted that ‘grouse, both of the black and red species’ were to be found on the moorland, and though little effort appears to have been made in developing a shooting estate in the area at this date a number of former shooting butts are to be seen on the Elenydd moorland which may belong to this period.

Rabbit Farming and Crop Cultivation in the Uplands

Further agricultural innovations that may have been introduced to the area by go-ahead farmers and landowners such as Thomas Grove intent on enhancing their revenue from the land are the groups of artificial rabbit warrens or ‘pillow mounds’ and areas of ridge and furrow to be seen on parts of the moorland of Elenydd.

An impressive group of eight pillow mounds is to be seen on Esgair y Ty, just to the west of Pont ar Elan. A number of the mounds are up to about 40 metres long, 6 metres wide and a metre high, and clearly represent a significant investment by landlord or tenant. Further groups of mounds are to be seen near Glanhirin farm and near Aber Glanhirin farm, respectively about 1.5 and 3 kilometres further to the west. Two further mounds are known on the saddle between Esgair Dderw and Penrhiw-wen, closer to Rhayader. Excavations elsewhere have shown that mounds of this kind were constructed with artificial burrows entering from the sides, which would enable rabbits to be periodically culled with the help of ferrets, to be marketed for both their meat and skins. Artificial rabbit warrens were being built as early as the medieval period, but here it seems more likely that they are post-medieval in date. A concentration of other similar groups of pillow mounds in Radnorshire and in Breconshire has suggested that they may have supplied markets in the expanding industrial towns of south Wales and the Marches. Their siting, near the turnpike road across the moorland, would have helped in getting produce quickly to market.

Former cultivation in the uplands is represented by a number of discrete areas of narrow ridge and furrow some of which has only been identified quite recently as a result of aerial reconnaissance. Some of the ridging is to be found in the same general area as the pillow mounds, as for example on the southern slopes of Esgair y Ty and within the area of the encroachment at Aber Glanhirin. Other areas of ridging are known near the encroachment at Lluest-pen-rhiw and on the slopes of Moelfryn (on the hills above Nannerth), on Cefn Gwair (west of the Craig Goch reservoir), on Cefn Cwm (east of Craig Goch reservoir) and in various places on Carn Gafallt (to the south-east of Elan Village). The ridging is generally fairly low and narrow, being up about 4 metres and as little as 1.5 metres wide and mostly lies between about 300 and 400 metres above sea level, on of beyond the margin of where traces of cultivation might be expected, and mostly lies on the more sheltered, south-facing slopes.

There is no certainty that all the ridging belongs to a single period, and although some might be of medieval or late medieval date, it seems likely that much if not all belongs to the same period of agricultural innovation to which the pillow mounds appear to belong, representing perhaps a relatively short-lived period of cultivation during the late 18th and early 19th centuries. Some of this upland ploughing is enclosed by earthen banks, though in some cases it is uncertain how crops were protected from stock grazing the adjacent moorland. Aber Glanhirin farm significantly still possesses a small stone-built threshing barn of this period in which corn grown on these ridges was threshed.

Fieldscapes

A variety of different field boundary types had emerged within the landscape of the area by the end of the 19th century, some of great antiquity and others of more recent origin.

Most of the lowland area to the west of Rhayader, in Cwm Dulas and in the lower Elan and Claerwen valleys is likely to have been cleared and utilized for pastoral or arable agriculture here from early times and has resulted in a number of different kinds of fieldscape. Most of the area is characterized by irregular, small to medium-sized fields, many of which may have originated from medieval or earlier times, generally with low-cut, multiple-species hedges with scattered more mature trees.

A relatively small number of fields have been amalgamated in recent decades, with former hedge lines represented by low earth banks or intermittent lines of trees or bushes running across a field. In these low-lying areas there are a number of relatively small and discrete areas with more regular field patterns suggesting enclosure of former open pasture in the post-medieval period.

As noted above there seems a possibility that the regular field system along the north bank of the river Elan between Coed-y-mynach farm and Noyadd, some of it associated with ridge and furrow possibly of medieval origin, may have been associated either directly or indirectly with the monastic grange centre at Llanfadog.

The areas towards the wooded valley sides, most notably in Cwm Dulas, around the head of the Nant Madog valley near Galedrhyd, and around the head of the Nant Caethon and Nant Gwynllyn valleys between Rhydoldog and Treheslog, are characterized by smaller irregular fields which have the appearance of assarts, created by felling parts of the adjacent woodland. Similar field patterns clearly once existed in the Claerwen and Elan valleys which are now flooded and likewise probably represent the emergence of freehold farms in the medieval and late medieval periods. Most of these fields remain in use though a number of former hedges have become overgrown and replaced by post and wire fences, and in places field boundaries around the upland margin have effectively been abandoned and are either with overgrown or have intermittent hedges or engulfed in heather or bracken. A number of small areas of former field, such as in the area of The Clyn on the south side of Cwm Dulas, have been overplanted by conifers.

The lowland fields are predominantly used for pasture and fodder crops today, though the presence of field lynchets, particularly on more sloping ground, suggests that a higher proportion of land was cultivated for cereal or other crops in earlier centuries.

The uplands of Elenydd have remained unenclosed apart from the relatively small areas enclosed by encroachments. These generally have a smaller, earlier, core of fields or paddocks around the habitation defined by earth and stone banks, which may once have supported hedges or timber fences, some of which are likely to date from the medieval or early post-medieval periods. The holdings have often been extended by additional fields defined by post and wire fences, a technique dating from the second half of the 19th century to the present day. Sheep could feed all year on the common grazing and the enclosed fields around the upland farmsteads are likely to have been used for temporarily holding cattle overnight, or for calving or milking, or for protecting crops of hay to be fed to cattle or horses over the winter.

A number upland encroachments have polygonal walled enclosures of two or three hectares in extent, as at Lluest-pen-rhiw above Nannerth, Blaen Methan in the Nant Methan valley, on the bank of the Rhiwnant stream south of the Claerwen valley, and at Cerrigcwplau, just below Claerwen dam. The walls, often now dilapidated, are constructed of either large upright slabs (as at Cerrigcwplau) or of drystone walls, or a combination of the two, depending upon the nature of the material available locally. The dating of these walled enclosures has still to be firmly established, but they seem likely to be of later 17th to early 19th-century date.

Probably of similar date are occasional walled boundaries between neighbouring farms or estates, and the wall which isolates the mining operations at the late 18th- and 19th-century Cwm Elan mine, in the Nant Methan valley, west of the Garreg-ddu reservoir, from the neighbouring animal pasture. Walled yards and paddocks can be seen at a number of farms, as at Cnwch.

In contrast with many other upland areas in Wales there appears to have been relatively little late 18th- or early 19th-century enclosure around the margins of the upland area, though late enclosure defined by post and wire fencing are present on Cefn Gwair, west of Craig Goch reservoir and or Rhos y Gelynnen and Gurnos, to the west of Caban-coch reservoir.

Relict field boundaries in the form of banks and ditches can be seen around the shore line of the reservoirs of the Elan valley during periods of low water. Most, if not all of these boundaries are to be seen on Ordnance Survey maps of the late 19th-century, before the construction of the dams. Some of these older boundaries have been replaced by modern post and wire fences which extend down into the waters of the reservoir, to prevent stock from straying in periods of drought.

Development of Farmhouses and Gentry Houses

The association of the Grove family with the Elan valley ended in 1815 when Thomas Grove the younger sold the Cwmteuddwr grange. By the later 18th century the principal landed estates in the area were the Peeles of Cwm Elan, Lewis Lloyds of Nantgwyllt, the Olivers of Rhydoldog, the Evanses of Noyadd, the Prickards of Dderw and the Davises of Gwardolau.

Two of the principal gentry houses of the area, Cwm Elan and Nantgwyllt, were to be demolished to make way for the Elan valley reservoirs. Their appearance and setting, however, is preserved in contemporary photographs and sketches of the 1890s. Cwm Elan, described by Jonathan Williams as having been ‘a neat and elegant mansion’, was a tall mansion of three storeys and three bays, which had been built by Thomas Grove after 1792. Nantgwyllt had been a low stone house with a pedimental gable which had probably been enlarged by Thomas Lloyd about 1770. Despite its picuturesque setting it was considered ‘bitterly cold and damp in winter, for the wooded slope south of the Claerwen shut out all the sunshine and was described by Percy Bysshe Shelley who rented it in 1812 as ‘silent, solitary, old’.

The flooding of the Elan valley removed notable gardens associated with the two gentry houses of Cwm Elan and Nantgwyllt. Jonathan Williams describes the setting of Cwm Elan:

‘The approach to the house is over a handsome stone bridge of one arch leading to a fine verdant lawn, which forms a curve with the course of the river and unites a ‘singular combination’, as a certain elegant author describes the situation, ‘of natural and artificial beauties, of wild scenery, and elegant ornament, of a foaming river, and rugged rocks, perpendicular precipices, and lofty mountains, contrasted with rich meadows and neat enclosures, leaving apparently nothing deficient to complete this singular and picturesque scene.’

Nantgwyllt was accompanied by ornamental tree plantings and a lawn stretching down to the banks of the river Claerwen. Unlike the house, the walls of the large, polygonal walled garden behind the house and the adjacent 18th-century road bridge across the Nant Gwyllt were left intact when the Caban-coch reservoir was created and are still to be clearly seen when the water level of the reservoir drops.

Rhydoldog is one of the few surviving 18th-century gentry houses in the area, ‘built somewhat in the cottage style’ in about the middle of the 18th century but subsequently enlarged, replacing a 17th-century house. The present Dderw house was built in about 1870, replacing an earlier brick-built house of 1799, which in turn replaced an earlier house of perhaps the 16th century.

The Dderw house lies with a small park in a striking valley-bottom setting adjacent to the deep gully of the Nant Gwynllyn. The park appears to date from about 1800, being enclosed on the west by belts of ornamental woodland, with water features, late 19th-century kitchen garden and a formal Arts and Crafts style terraced grass garden and orchard added in the 1920s.

A number of the both upland and lowland farmhouses within the area were being rebuilt or enlarged during the course of the 18th and 19th centuries. The now-ruinous farmhouse at Lluest-aber-caethon, probably built on the site of an earlier house, has a stone in the chimney stack inscribed ‘D. E. 1814 CLETTWR’, perhaps signifying that it was built by the owner of the adjacent Lluest Clettwr. New farmhouses were built at Cerrigcwplau, Hirnant and Rhiwnant in the later 19th-century, sometimes of local stone with brick window and door openings. Several farmhouses were evidently replaced or superseded during the 20th-century, the longhouse at Llannerch-y-cawr being superseded by a new bungalow, the Victorian farmhouse at Cerrigcwplau demolished and replaced shortly after the construction of the Claerwen dam, and the longhouse at Ciloerwynt which was demolished and replaced by a modern, single-storey farmhouse.

The historic landscape includes a number of distinct mining landscapes of the late 18th and early 19th century which, though peripheral to and on a smaller scale than the more extensive workings in the upper Ystwyth valley in north Ceredigion, just to the west, are important in terms of landscape history of the historic landscape area. Most of the mines within the Elan valley are fairly well hidden from view, though occasionally the heaps of mineral waste and leats for harnessing water power for processing ores had a somewhat broader impact upon the landscape.

Despite a short-lived attempt to smelt the ores in Aberystwyth in the late 1780s and early 1790s, most of the processed ores from these small Radnorshire mines, like those from Cwmystwyth, were initially transported across the mountains to the coast, to be taken by sea to Deeside or to Neath or Swansea for smelting, though from 1864 onwards rail transport for ores was available from Rhayader.

There is no evidence of prehistoric or medieval workings at any of the mines in the Elan Valley historic landscape area though it is likely that early mining in Cwmystwyth will have had a visible and environmental impact upon the western side of Elenydd by the later 16th century when John Leland records that mining ‘hath destroid the Woddes that sometime grew plentifull thereabout’.

The mine known as the Cwm Elan mine produced lead and zinc ores, appears to belong to a single phase of operation, and provides what is perhaps the best example of late 19th-century mining technology and planning in Powys, with the structures still remarkably well preserved. It lies on the western slopes of the Nant Methan stream in an upland valley on the edge of the moorland area to the west of the Garreg-ddu reservoir and originated from the discovery of lead ore during the digging of a drainage ditch in 1796, no doubt as part of the agricultural improvements being introduced on the estate of Thomas Grove. The early workings, which may have included a series of shallow open-cuts along the banks a stream flowing into the Nant Methan, were initially worked Grove who subsequently leased out the operations. The main phase of working, responsible for the majority of surviving structures, began in 1871 with the formation of the Cwm Elan Mining Company and by the following year included shallow and deep adits and shafts. A processing mill began operations in 1873 with equipment supplied by William Thomas of Llanidloes Foundry, powered by three waterwheels, the largest of which was thirty-six feet in diameter. The waterwheels were powered by a 16-kilometre leat running on a carefully surveyed course across Elenydd moorland from Llyn Cerrigllwydion Isaf, 170 metres higher up and just 7 kilometres away as the crow flies, which took three months to complete. Drought and lack of funds forced the company into liquidation in 1874. Visible surviving remains include partially collapsed shafts, ore-bins for the storage of ore, platforms for stone houses and jiggers and a buddle for processing ore and settling pits. Surviving ruined stone buildings survive which were associated with the mine, including an explosives’ magazine, a smithy and a mine manager’s house and office, probably all built from stone obtained from the adjacent small stone quarry. There is also the remains of a red brick house built in the 1890s by the Birmingham Corporation Waterworks after they had acquired the Elan Estate to construct the existing reservoirs.

A second mining complex which produced copper and lead ores is to be seen in the valley of the Rhiwnant and Nant y Carw, to the south of the Claerwen valley, comprising a number of different small mines. Dalrhiw lay on gently sloping ground on the south side of the Rhiwnant. The earliest workings are marked by of an adit driven south from the banks of the stream in 1850, but a greater impact upon the landscape resulted from the subsequent sinking of a shaft and the development of on-site processing driven by waterwheels drawing power by water from leats taken off the Rhiwnant. Visible remains of the mining operations include a shaft from which ore was raised by means of a horse whim, ore bins, and a small crusher house powered by waterwheel, a wheel pit to house a fifty-two foot wheel for pumping, and a another wheel pit which probably powered jiggers. As at Cwm Elan there are the ruins of a number of stone buildings including the mine office or manager’s house, a possible smithy, and a small walled enclosure, possibly a garden. A low earthwork enclosure just south of the mine buildings may have been used as a pound for horses employed at the whim or for the transporting the processed ore away from the mine. The workings here continued until 1881.

The third mine sett known as Nant y Car South, on the north bank of the Rhiwnant stream opposite Dalrhiw mine, produced copper, lead and zinc ores. As at Dalrhiw, the earliest workings appear to have been a series of adits driven into the hillside close to the stream. The main development of the mine took place during the 1860s and 1870s and involved the sinking of a shaft and the construction of an impressive crusher and wheelpit, along with other related structures which form the main features of the site. The spoil tips of development and processing waste create a distinctive landscape feature with the characteristic ‘fingers’ of spoil radiating below the shaft and ore-bins. The Nant y Car mine originally operated at a site on the south side of the Claerwen valley, to the north, where trials had been worked before the middle of the 19th century. From 1844 working developed on a much larger scale and by the 1850s a rich vein of copper ore was being exploited, although by 1854 this had proved disappointing. Prospecting elsewhere in the sett in 1855 revealed a promising lead lode in an adit on the north side of the Rhiwnant stream, at the mine to become known as Nant y Car South. The following year additional machinery was brought in to deepen the workings, although the lode ultimately proved disappointing and the company was wound-up in 1859. Visible remains at Nant y Car South consist of a series of adits driven in from the banks of the stream, an engine shaft with the adjacent foundations of a winding house, a tramway and ore-bins, waste tips, and platforms for a crusher and jiggers, and two circular buddles for ore processing, powered by a waterwheel.

A new and richer load was discovered in about 1883 further up the valley at the Nantygarw mine, leading to the abandonment of Nant y Car South. The mine occupies the only available ground on a natural terrace at the mouth of a hanging valley high above the Rhiwnant. Below the site the main valley sides are precipitous and rocky, while above more gently sloping ground rises to the moorland plateau to the north. The view eastwards from the site is spectacular, looking down the Rhiwnant Valley to Dalrhiw, Nant y Car South and beyond. The surviving remains which have transformed this remote upland location are the result of more than one period of tenure, although they are dominated by the latest phase of activity which saw the installation of a processing mill for which the site is perhaps best known. Other substantial structural remains survive, including the shaft, wheelpit, and smithy. Extensive spoil tips of processing waste spread outwards from the processing mill, while at the bottom end of the site spoil tumbles down the steep valley sides to the Rhiwnant below. Visible structures include a substantial crushing mill, still standing to a height of over 4 metres, powered by a leat constructed in 1893 carrying water fron Llyn Carw, about 100 metres higher up and 2 kilometres away to the west, jigging platforms, and a small circular buddle, reservoir and settling pits. Building remains include possible barracks for housing workers during the week, a smithy, a mine office and possible manager’s house and explosives’ magazine which survives intact but without a roof, to a height of over three metres.

The Nantygarw mine, the latest and most remote of the Elan valley mines was wound up in 1893, and although little is known of the workings, 50 men were employed, suggesting a reasonably sizable enterprise. Despite large returns the company went into liquidation in 1897 and although some activity continued until 1899, all further mining was discontinued due to the risk of contaminating the Elan valley water supplies.

Early Tracks and Drove-ways

Various tracks and paths became established across the historic landscape area from early times. Early routes no doubt included ones which gave access to early settlements the Elan and Claerwen valleys from the Wye valley, as well as the mountain route from the Wye valley to Cwmystwyth via the upper Elan valley, a line of communication which the Roman marching camp on Esgair Perfedd suggests has been in use since at least Roman times. This route is shown on a number of late 17th and early 18th century maps, including Morden’s map of South Wales published in Camden’s Britannia of 1695. Other early routes across Elenydd were one linking Claerwen valley with the upper reaches of the Teifi valley and Tregaron, and another which strikes off across the mountain from the direction of the upper Teifi valley towards Aberglanhirin in the upper Elan valley and then directly down to the Wye valley near Llangurig.

Many of these early routes were only suitable for those travelling on foot or on horseback, though a horse-drawn wheeled sledge known as a ‘wheel-car’ (sled olwynion) is known to have been used in the area before the advent of mechanised transport which was ideal for carrying loads of peat, hay, heather or even fern across the wet moorland and steep valley sides. A sledge of this kind, a typical form of agricultural transport on early farms in Montgomeryshire and Radnorshire, was photographed in the Elan valley by Iorwerth Peate in the early 20th century.

A number of the routes across the mountains formed part of recognised drovers’ ways across Radnorshire linking Carmarthenshire and Ceredigion farms with the market towns in either the English Midlands, a trade in its heyday in the 18th and early 19th centuries. Four routes have been suggested across Elenydd from the account books of 19th-century drovers, early maps, earthwork and place-name evidence. The most northerly drove-way across Elenydd, leading to Shrewsbury and the north Midlands, went from Cwmystwyth (Ceredigion), along the upper Ystwyth valley and across Yr Allt to Llangurig (Montgomeryshire). A branch from this route probably also struck off along the turnpike road along the upper Ystwyth and Elan valleys to Rhayader. A further drove took cattle either via Llangurig or Rhayader, on a route from Ffair Rhos (Ceredigion) via the upper Claerddu and Claerwen valleys and the ‘Ancient Road’ to join the Cwmystyth-Rhayader road near Aber Glanhirin or to continue northwards into the Wye valley by way of Cefn Bach. The further drove-way from Pontrhydfendigaid (Ceredigion), appears to have gone via Strata Florida, the upper Irfon valley, across Drygarn Fawr, and along the Rhiwnant valley to the lower Claerwen valley and thence to Rhayader via the Elan valley. This appears to have shared part of the same course as a drove from Tregaron (Ceredigion) which crossed the southern tip of Elenydd before taking a course along the Irfon valley to Abergwesyn (Breconshire).

A number of the routes across Elenydd are popularly held to have monastic associations, and although this historical association is uncertain a number of the routes are undoubtedly of considerable antiquity. The ‘Ancient Road’ mentioned above is shown on Ordnance Survey maps of the 1890s and is known colloquially as the ‘Monks’ Way’ or ‘Monks’ Trod’, supposedly linking the Cistercian abbey at Strata Florida with its daughter house at Cwmhir. It was popularised in John Williams’s History of Radnorshire, compiled in the early 19th century, which envisaged the monks of the monastic grange at Nant Madoc, near Elan Village, visiting their mother house at Strata Florida,

‘either for the purpose of their mutual peace and edification, or for consulting together on their temporary interests; and it is recorded that the inhabitants of this religious establishment were accustomed on certain periodical seasons, to visit their brethren in the abbey of Strata Florida . . . marching over the hills in procession, and making the rocks re-echo their loud and chaunted hymns. Their road over the mountains may at this date be traced.’

Toll-roads

A few of these early tracks and routes across the mountains were to be improved during the course of the 18th and early 19th century to enable them to carry wheeled transport, in response to the increasing demands of trade, tourists and mail. The most prominent of these was the most direct route between Aberystwyth to Rhayader road across Elenydd, via Devil’s Bridge and Cwmystyth, which was made into a toll road in about 1790. Surviving from this period are two typical early 19th-century round-headed roadside milestones, one to the east of Pont ar Elan and the other to the south of Dderw, inscribed with distances from Aberystwyth and Rhayader.

For many English travellers the sudden transition from lowland to upland scenery as they progressed on their coach journeys across Elenydd west of Rhayader often came as a shock. To quote Richard Moore-Colyer ‘Henry Skrine wrote of his state of ‘perpetual alarm’ as he ascended the rocky road out of the town and subsequently lamented the dreary, treeless expanse of open hill which he traversed ‘in mournful silence’. A more measured tone is registered in the diary of the scientist Michael Faraday in 1819: ‘After a while we got among more mountains and nothing but large concave forms met the eye for a long time. Lively little catle with myriads of sheep now and then diversified the general monotony’. By 1829 this mountain route had been replaced as the main road between Aberystwyth and Rhayader by the more lowland road along the Rheidol and Wye valleys via Ponterwyd and Llangurig (A44/A470) which is now main route.

For a time, between the late 18th and the coming of the railways in the later 19th century there were occasional conflicts between the turnpike roads and the drovers and other travellers. Where possible, drovers would endeavour to keep to the unpaved mountain routes to avoid the tolls. The charges levied on other road users became very unpopular and gave rise to the ‘Rebecca Riots’ in parts of Wales. Trouble arose locally around Rhayader in 1842. In early September two gates where the Aberystwyth road branched off beyond the bridge over the Wye were destroyed and in the following month one of the most violent incidents of the riots took place halfway across the mountains at the Bodtalog tollgate, near Abergwngu Hill, when the lone woman who kept the gate was nearly blinded a powder-loaded gun. The riots resulted in a change in the law which established the County Road Boards which took over the responsibilities of the turnpike trusts from 1845.

Railways

The drovers’ roads across Elenydd were to come to an end with the coming of the railways. This was marked locally by the Mid-Wales Railway from Llanidloes to Builth Road which opened in 1864 and which runs across the historic landscape area to the west of Rhayader. The railway closed in 1962, just under a century later, due to competition with road transport, the disused line of the railway still being marked by a cutting and by the short, 271-yard tunnel about half a kilometre south of Rhayader. The junction with the former Elan Valley Railway, just to the south of the tunnel, is described below.

The history of the Elan valley today is dominated by two gentry mansions of Nantgwyllt and Cwm Elan, both now submerged below the reservoir, the families associated with them, and the inspiration they gave to several English poets of the romantic movement of the late 18th and early 19th century. The literary associations, though short-lived, have a profound and enduring effect upon our appreciation of the drowned landscapes of the Elan valley at the present day.

As noted in the section on landed estates, the first of the poets to commemorate the Elan valley in verse was William Lisle Bowles, friend of Thomas Grove senior of Cwm Elan house. An edition of his extended blank-verse poem entitled ‘Coombe-Ellen’, published in 1801 is accompanied by an engraving ‘from a drawing by Mrs. Grove’. In the words of Desmond Hawkins, the poem begins with an ‘invocation to the spirit of wild untamed Nature’:

‘Call the strange spirit that abides unseen
In wilds, and wastes, and shaggy solitudes,
And bid this dim hand lead thee through these scenes
That burst immense around! By mountains, glens,
And solitary cataracts that dash
Through dark ravines.’

The Groves subsequently became related by marriage with the Bysshe Shelley family, one of whom in 1784 was High Sheriff of Radnorshire. The poet, Percy Bysshe Shelley came to Elan valley and stayed a month as the guest of his cousin, Thomas Grove, at the ‘neat and elegant mansion’ of Cwm-Elan in 1811. This was several months after the young poet had been sent down from Oxford following his co-authorship of the pamphlet called Necessity of Atheism and consequent estrangement from his family. In view of his reputation for eccentricity is it no surprise that Shelley was still remembered locally in the late 1870s.

His thoughts on the Elan valley, mostly contained in letters he wrote whilst staying there, were to become known from two biographies published before the end of the 19th century, after his death in 1822 at the age of 30. In a letter to a friend, Elizabeth Hitchener, he wrote from Cwm Elan in July 1811,

‘This county of Wales is excessively grand; rocks piled on each other to tremendous heights, rivers formed into cataracts by their projections an valleys clothed in woods, present an appearance of enchantment but why do they enchant, why is it more affecting than a plain, it cannot be innate, is it acquired?’

In a note to Thomas Jefferson Hogg later in the month he wrote of ‘waterfalls midst the umbrage of a thousand shadowy trees—form the principal feature of the scenery. I am not wholly uninfluenced by its magic in my lonely walks.’

Shelley was back in the valley with his first wife Harriet and sister-in-law in April the following year, and was this time staying at Nantgwyllt, a mile away from Cwm Elan, which he hoped to acquire. He wrote to William Godwin in April 1812 (a matter of only two years before his notorious elopement with Mary Wollstonecraft),

‘We are not yet completely certain of being able to obtain the house where we now are. It has a farm of two hundred acres, and the rent is but ninety-eight pounds per annum. The cheapness, beauty, and retirement make this place in every point of view desirable. Nor can I view this scenery — mountains and rocks seeming to form a barrier round this quiet valley, which the tumult of the world may never overleap.’

A further letter describes the farm at that time consisting of about 200 acres was about 130 acres arable and the rest wood and mountain. On 1 May he wrote, ‘Give me Nantgwillt, fix me in this spot, so retired, so lovely, so fit for the seclusion of those who think and feel. Fate, I ask no more!’

The only mention of Cwm Elan in verse is in a poem of 1812 called ‘The Retrospect, Cwm Elan, 1812’, which though largely concerned with other matters, includes several descriptive passages:

‘Ye jagged peaks that frown sublime,
Mocking the blunted scythe of Time,
Whence I would watch its lustre pale
Steal from the moon o’er yonder vale:
Though rock, whose bosom black and vast,
Bared to the stream’s unceasing flow,
Ever its giant shade doth cast
On the tumultuous surge below’

Shelley tried to secure the lease of the estate for ‘a little colony of enlightened souls’ but protracted negotiations came to nothing. Despite his appreciation of the secluded and picturesque nature of the Elan valley he was not oblivious to social deprivations of the area. On his first visit to Cwm Elan in 1811 he was troubled mentally by an encounter with a beggar who claimed to have suffered at the hands of the better off. Shelley also spoke somewhat disparagingly about the local community: ‘I have been to church to-day: they preach partly in Welsh, which sounds most singularly. A christening was performed out of an old broken slop-basin’, and elsewhere ‘I am all solitude, as I cannot call the society here an alternative to it’, speaking elsewhere of missing letters and ‘the pillage of the Rhayader mail’. Later, in 1812, after leaving the Elan valley he was to write of Wales, though not perhaps exclusively of Radnorshire, in the following terms:

‘It is the last stronghold of the most vulgar and commonplace prejudices of aristocracy. Lawyers of unexampled villany rule and grind the poor, whilst they cheat the rich. The peasants are more serfs, and are fed and lodged worse than pigs. The gentry have all the ferocity and despotism of ancient barons, without their dignity and chivalric distain of shame and danger.’

The cultural significance of Shelley’s appreciation of the secluded and picturesque nature of the Elan valley was recognised before the end of the 19th century, exemplified by the inclusion of an essay entitled ‘Shelley at Cwm Elan and Nantgwilt’ by William Rossetti, a member of the Pre-Raphaelite Brotherhood, in Eustace Tickell’s The Vale of Nantgwilt. A Submerged Valley, published in 1894, as the Elan valley reservoir scheme was being constructed.

Attention was being drawn to the antiquities of Elenydd and the surrounding area early in the 19th century, in a fashion which combines the picturesque and romantic with a stab at a more rational and historical interpretations. In the second volume of his History of the County of Brecknock, published in 1809, Theophilus Jones speculates on the purpose of a number of antiquities on the southern portion of Elenydd in the parish of Llanwrthwl.

‘A great part of this parish consists of lofty hills, bogs and commons; among the first is the Drygarn or Derwydd garn, (Mount Druid or Druid’s rock) . . . . On the top of this are many Carnau or Carneddau, large heaps of stones, as there were also upon a less elevated eminence not far from hence, called Gemrhiw [Gamriw].On the road from Llandovery and Llangamarch to Rhayader, are seen stones placed irregularly in the ground, which have given a common, partly in this parish and partly in Llanafan, the name of Rhôs saith maen, or Seven-stone common; whether they are sepulchral, military or druidical remains, is not known, but from the name of Rhos y beddau, the common of the graves, not far from hence, nearer to the river Wye, it should seem that they commemorate a battle, most likely that of Llechryd and the slaughter in the flight of Riryd and Cadwgan.’

Jonathan Williams in his General History of the County of Radnor, published posthumously in 1859, likewise tried to associate what is in reality most probably a prehistoric burial monument in the neighbouring parish of Cwmteuddwr with historical events with which he was more familiar:

‘Near to Gwaith-y-mwynau there is a considerable tumulus, or barrow [probably the monument now known as Clap yr Arian, above the head of the Nantgwynllyn valley] . . . . from thence may distinctly be seen the Castle of Rhayader, to which fortress therefore, it must have served as an outpost to give intelligence to the garrison of the approach of an enemy.’

Jonathan Williams also, as we have seen above, drew attention to the remains of the monastic grange in Cwmteuddwr, associated with the Cistercian abbey at Strata Florida.

In the 1840s, Lady Charlotte Guest, translator of the Mabinogion, became excited by the realisation that the cairn known as Carn Gafallt was mentioned in the 10th-century manuscript known as Nennius’s Historia Brittonum recording early folklore associated with the Arthurian legend that she ‘prevailed upon a gentleman to undertake a pilgrimage . . . to the summit of Cefn Carn Cavall. Her correspondent, who remains anonymous, prepared the following account of his expedition which combines romanticism with a rational interpretation of the association with one of Arthur’s hounds.

‘Carn Cavall, or, as it is generally pronounced, Corn Cavall, is a lofty and rugged mountain, in the upper part of the district anciently called Buellt, now written Builth, in Breconshire. Scattered over this mountain are several cairns of various dimensions, some of which are of very considerable magnitude, being at least a hundred and fifty feet in circumference. On one of these carns may still be seen a stone, so nearly corresponding with the description in Nennius, as to furnish strong presumption that it is the identical object referred to. It is near two feet in length, and not quite a foot wide, and such as a man might, without any great exertion, carry away in his hands. On the one side is an oval indentation, rounded at the bottom, nearly four inches long by three wide, about two inches deep, and altogether presenting such an appearance as might, without any great strain of imagination, be thought to resemble the print of a dog’s foot . . . As the stone is a species of conglomerate, it is possible that some unimaginative geologist may persist in maintaining that this footprint is nothing more than the cavity, left by the removal of a rounded pebble, which was once embedded in the stone.’

The monastic associations of Elenydd and the Elan valley were being elaborated upon in the 1880s by the researches of Stephen Williams, the Radnorshire architect, who contributed a chapter entitled ‘The Grange of Cwmteuddwr’ to Eustace Tickell’s book, The Vale of Nantgwilt, published in 1894. Both Williams and Tickell were to be engaged upon the Elan valley reservoir scheme, Tickell being the civil engineer responsible for the construction of the Penygarreg dam. He was also a competent writer and artist, the numerous sketches in his book of the Elan valley before it was flooded by the reservoir, including views of Capel Nantgwyllt, and the houses at Nantgwyllt and Cwm Elan, amply illustrating its picturesque qualities. The object of Tickell’s book was to

‘commemorate scenes in one of the most charming valleys in Great Britain. Scenes which are soon to be lost for ever, submerged beneath the waters of a series of lakes, which, by a colossal engineering undertaking, are about to be constructed for the purpose of supplying water to the city of Birmingham . . . Beautiful lakes they will doubtless be, winding up the valleys with sinuous margins, wooded promontories such as are seen on Derwentwater, frowning crags and screes which will remind one of Wastwater. But their construction dooms many a picturesque and interesting spot to destruction, and it would be indeed a pity if they should be allowed to pass away without some record, however inadequate.’

There were few visitors to the valley, since as Tickell observes ‘It lies hidden away amongst the mountains and leads to nowhere. . . . The valley is visited by few, there being no inn for the tourist to put up at’. Tickell was conscious of the inevitable march of progress, however, and in counting the losses, observed that ‘it must be remembered that, sad as it is, it would be difficult to find in this island a place where more than 70 square miles of land could be taken for a public purpose without dispossessing very many more people, destroying many more homes.’

In the same book, William Rossetti, perhaps with a vision of Millais’ iconic Ophelia before him, was to introduce a more melodramatic air, prefiguring a theme which was to recur in a number of romantic novels of the early 20th century:

‘Harriet Shelley died by her own deed in the Serpentine in 1816, Shelley in the Mediterranean waves in 1822; and now a watery doom effaces the scenes of their short-lived love, Nantgwilt and Cwm Elan. A world of waters, a world of death.’

These literary associations were well known when work began on Birminghams’s reservoir scheme and are likely to have had a subtle influence on various picturesque aspects of its design.

The landscape character of the Elan valley underwent a dramatic change with the construction of the reservoir scheme built between 1893 and 1906 for the corporation of Birmingham, but officially opened amidst much pomp by King Edward VII in 1904. The stunning new landscape that was created during this period, for which this area of western Radnorshire is justly famous, is thus mainly a Victorian to Edwardian in origin, created as a civic project by a distant city, though the valley’s literary and historical associations continue to play an influential role in how the Elan valley is perceived. The qualities that contributed to the allure of the Elan valley’s picturesque landscape — its the gushing streams, steep-sided valleys, and remoteness — were the very same as those which were to make it an ideal site for the Birmingham’s reservoir scheme.

Birmingham, in common with other industrial cities in Britain, came under increasing pressure in the later 19th century to provide adequate water supplies for its rapidly expanding workforce, the population of the city at the time numbering about 650,000. Local water supplies were proving inadequate in terms of volume and had often become polluted, resulting in typhoid and cholera epidemics. As in the case of Liverpool Corporation’s scheme a few years earlier at Lake Vyrnwy, Montgomeryshire, both the quantity and purity of the water were important factors, and Birmingham, like Liverpool, was to seek a solution to these problems in the Welsh hills, some 75 miles from the city.

The Elan valley had been first being examined as a suitable source for Birmingham’s water supply in 1870 by Sir Robert Rawlinson. Nothing further was done for twenty years until 1890 when James Mansergh, the most distinguished water engineer of the day, was appointed as consultant engineer for the project. Mansergh had become familiar with the Elan and Claerwen valleys when working on the Mid-Wales railway in the 1860s. Birmingham Corporation had committed itself to the substantial investment that would be needed to ensure its water supplies to meet both present and future needs and accepted their consultant’s recommendation to acquire the Elan valley watershed which a member of the corporation’s Water Committee described as ‘treasures of untold value’.

The Elan valley scheme, like Liverpool’s early scheme at Lake Vyrnwy, was to be free of the political controversy that was surround the construction of reservoirs to supply English cities in Wales in the second half of the 20th century, as in the case of the Tryweryn reservoir near Bala, built for Liverpool in the early 1960s. Such opposition as there was, was largely concerned with the question of financial compensation and the potential disruption to existing water supplies. Some opposition was to be anticipated from the lord of the manor, Robert Lewis Lloyd of Nantgwilt though the committee in charge of the scheme ‘very prudently therefore came to terms with that gentleman early in the proceedings’. There was no serious opposition to the construction of the 73-mile aqueduct to the Frankley reservoir, near Birmingham, Mansergh noting somewhat sardonically in 1894 that ‘landowners being well enough aware nowadays that they have little chance to stop a great and useful scheme of this character, and that their prudent policy is to acquiesce, with the chance of bleeding the promoters heavily for interfering with their property. Experience showing that in this process they are perhaps more than fairly proficient’. An extensive inquiry was also made into the rights and privileges of the local commoners.

The parliamentary bill granting the necessary legal powers to Birmingham Corporation to undertake the scheme became law in 1892. The determination of the area of the watershed to be acquired was in Mansergh’s own words ‘a comparatively easy problem (considered from a water engineer’s point of view), because the contraction of the valley at Caban-coch, and the opening out above of the wide expanse of flat land, fixed at once the position of the dam of the lowest reservoir’. The extent of the watershed was mapped out by accurate survey onto Ordnance Survey plans and a three-dimensional scale model of the area was created, the watershed also being marked on the ground by stone pillars set at intervals. Birmingham was to purchase to entire watershed, an area of about 71 square miles, which Mansergh considered by the standards of 1894 to be ‘an abnormally extensive tract of country to be secured in a mountain district’, representing a large portion of the lands that had been granted by Rhys ap Gruffydd to the newly founded Cistercian monastery at Strata Florida in 1184.

Records of rainfall had been kept by the Lloyds of Nantgwilt since 1870, which indicated an average annual rainfall in the area of the watershed of about 70 inches, which over the area of the watershed would amount to an average of about 100 million gallons a day. Of this figure, the parliamentary bill provided for an average of 27 million gallons ‘compensation water’ to be passed into the Elan every day, although the act allowed for a proportion of this to be reserved to create periodic spates to help the fish to run up the river.

The scheme as originally envisaged, allowing for future expansion, comprised six reservoirs, three on the Elan — Caban-coch (partly on the Claerwen), Penygarreg and Craig Goch, and three on the Claerwen — Dol-y-mynach, Ciloerwynt and Pant-y-beddau. It had always been the intention that the work would be ‘carried out by instalments to meet the growing requirements of the districts to be supplied’. The three Elan reservoirs were sufficient to meet the city’s needs in the first decade of the 20th century and consequently the projected Ciloerwynt and Pant-y-beddau dams were never built. The foundations of Dol-y-mynach dam and the Dolymynach tunnel from the Claerwen to the Elan were built at an early stage, however, since the site of the dam would be flooded once the Caban-coch reservoir had filled up, but the dam was never completed as originally envisaged and only a small reservoir was created. The original proposals in the Claerwen valley were to be superseded by the construction of a single much larger dam, built in the late 1940s and early 1950s, described below. A symmetrical pair of hydroelectric power houses below the Caban-coch dam retaining their original turbines and generators which formed an integral part of the Elan valley reservoir scheme.

Contemporary technology favoured the construction of relatively high, stone-built dams across steep-sided valleys. A feature of the scheme considered by Mansergh to be ‘novel and unique’ was that in order to maintain a sufficient fall in the aqueduct between Elan and Birmingham it was necessary for the water to be drawn off above the lowest dam at Caban-coch. Water is therefore taken off at the Foel tower, at a point just above a weir below the Garreg-ddu viaduct, normally submerged 40 feet below the surface of the reservoir, the viaduct being necessary to carry the road further up the Claerwen valley, replacing the earlier road in the valley bottom. The water is carried from Foel tower in a tunnel excavated below Foel hill and running to the filter beds opposite Elan Village and from thence in the direction of Rhayader. The average height of the three complete dams, Caban-coch, Penygarreg and Craig Goch and was about 120 feet, the thickness of their bases being designed to be just about equal to their height.

The basic construction of the dams was of large irregular blocks of rubble embedded in concrete with a concrete lining six feet thick, faced upstream and downstream with facings of shaped stones arranged in snecked courses. Stone for the core of the dams at Caban-coch, Penygarreg and Craig Goch was obtained from the two quarries specially opened on the same outcrop of conglomerate on opposite sides of the river Elan near Caban-coch — Cigfran quarry to the north and Craig Cnwch to the south, and the Aberdeuddwr quarry on Cerrig Gwynion, just to the south of Rhayader. Facing and other dressed stone was obtained from the quarries at Llanelwedd near Builth Wells and at Pontypridd. Cement came from works on the river Medway in south-east England, being delivered by sea to Aberystwyth and then by rail to the Elan valley. The cost of the Elan valley scheme ran slightly over budget due to the failure to find suitable building stone near any of the dams apart from Caban-coch, and the necessity of bringing additional stone in from further afield.

Building materials were delivered to the Elan valley by means of the Elan Valley Railway, specially constructed by Birmingham Corporation for the waterworks, the first section of the railway to be built, being the three-mile length from Rhayader Junction on the mid Wales section of the Mid-Wales Railway to the depot below the site of the Caban-coch dam, on the north bank of the Elan, which included a cement cooling shed, general stores, coal depot, workshops for carpenters, smiths, fitters and waggon builders, and sawmills. The various railway tracks leading to each of the dams had a total length of 33 miles and was worked by six locomotives capable of transporting 1,000 tons of materials a day.

In addition to the transport of materials, the railway was also used for transporting workmen from the temporary settlement of wooden huts constructed on the south bank of the river, on the site now occupied by Elan Village, occupied by skilled and unskilled workmen and clerical workers engaged on the scheme. Careful enquiry was made of other large contemporary civil engineering schemes to establish best practice in maintaining discipline and lawfulness, avoid drunkenness and illness in a large, temporary, and predominantly male workforce. Problems of nature had arisen just a few years before when a large influx of workers had arrived in a similarly remote rural setting to build the Liverpool Corporation’s reservoir scheme at Lake Vyrnwy, Montgomeryshire. The village was provided with hospitals, schoolroom, public hall, fire brigade depot and canteen. The canteen was unique for its time in being a municipal public house, all the profits of which were devoted to the social welfare of the community, covering the costs of the mission room, recreation room, gymnasium, free library, recreation grounds and bathhouse. Access to the village was controlled by means of a guarded suspension bridge across the river Elan.

Like other large engineering works, the workforce was drawn from all over Britain. The resident engineer throughout the works was George Yourdi, ‘an expert in cement work’ of Greek and Irish parentage, who ‘has tramped miles up and down the valley, inspecting, directing, controlling, everywhere, on the coldest of frosty winter days or in the most torrid summer heat’. Other engineers engaged upon the work included Eustace Tickell who supervised the construction of the Penygarreg dam, and who also, as mentioned in an earlier section, produced the book entitled The Vale of Nantgwilt: A Submerged Valley, published in London in 1894. Yourdi occupied the house at Nantgwyllt throughout much of the works, the principal offices occupying a site below the Garreg-ddu weir, near the confluence of the Elan and Claerwen.

Survey work and question of valuation and compensation within the Elan valley undertaken on behalf of Birmingham Corporation by the local surveyor and architect Stephen W. Williams, appointed James Mansergh, who had previously worked with him on the railway schemes in mid Wales. Williams was also made responsible for building the village to house the workmen working. Williams had long been engaged upon his researches on the Cistercian houses of Wales and being familiar with local antiquities was instrumental in effecting a realignment of the Elan Valley Railway to avoid the site of the grange chapel near the Elan Valley Hotel. Williams also designed what became the Birmingham Corporation Water Board’s offices in South Street, Rhayader, built originally for his own use in 1893, and was the architect of Nantgwyllt church near the southern end of the Garreg-ddu viaduct, which replaced one drowned by the rising waters of the reservoir, under construction in 1898 and opened in 1903. Sculptured corbels inside the church are thought to include representations of Stephen Williams and possibly of James Mansergh.

Demolition and flooding of existing houses and buildings within the Elan valley was inevitable. As noted by Mansergh in 1894

‘In the execution of these works (in addition to Nantgwilt already mentioned), there will be submerged the residence of Cwm Elan, the little church at Nantgwilt, the school, the Baptist chapel, and twenty farm and other buildings. With these, practically the whole of the valley lands now worked for agricultural purposes will be covered, leaving the area from which the water will be collected a vast tract of nearly uninhabited moorland, used only as sheepwalk.All the manorial and other rights have been acquired so as to stop mining or quarrying of any description, and ample powers are possessed in the Act to prevent any chance of pollution, and ensure the collection of the water in its pristine purity’.

Careful attention was evidently given in the design and planning of the works to cause as little blemish as possible to the landscape of the Elan valley. Where possible, work yards, processing plants, railway cuttings and other essential works of this kind were sited in places where they would eventually be flooded by the rising waters of the reservoirs. Some elements of the works remain visible, however, including parts of the course of the railway tracks to Penygarreg and Craig Goch dams, the brick railway bridge across the Nant Hesgog stream as well as the course of the link to the former Mid-Wales Railway at Rhayader.

A foundations of number of the former houses are exposed during periods of low water, including those of Cwm Elan and Nantgwyllt, the walled garden and road bridge at Nantgwyllt, and the former farmhouses at Ty Nant below Penygarreg reservoir and Dol-faenog below Garreg-ddu reservoir. Various elements of the construction works are likewise exposed during periods of drought, notably the mason’s yard to the south-west of Caban-coch dam, and the stone and timber foundations of a workmen’s hut just to the west of the Craig Goch reservoir, similar to the complete hut which survives near the Elan Valley Visitor Centre.

The engineering-architecture of the project lends the area a strong unifying quality, as it is clear that the whole landscape of the reservoirs was carefully designed, not only in the interests of utility, but also consciously as spectacle. The lakeside roads and plantings, the carefully contrived viewpoints, especially that below the Penygarreg dam, and picturesque effects such as the siting and juxtaposition of Nantgwyllt church, Garreg-ddu viaduct and the Foel tower, a visual focal point of the entire scheme, suggest a strong governing aesthetic. The Garreg-ddu viaduct, a is one of the important visual element of the scheme and although well above the submerged weir, in Haslam’s words, ‘gives the illusion on of crossing a shallow lake’. Mansergh’s intentions are quite clear, stating that ‘when more than full water will overflow from all the reservoirs in picturesque cascades down the faces of the dams’, Caban-coch dam being calculated in times of maximum flood to form ‘probably the finest waterfall in this country’.

New roads were built to replace the roads and lanes linking the surviving farms around the Elan and Claerwen valleys, but they are more than this. The railed lakeside carriage rides are part of the original conception, giving access to a landscape designed to be seen, and which like Lake Vyrnwy before it would attract numerous visitors, perhaps especially from Birmingham. Each of the dams was to carry a plaque proudly quantifying its dimensions and the volume of water that it impounded. The Elan Valley Hotel, on the road towards Rhayader, appears to be broadly contemporary, its large function-roomed wing suggesting an early role in the promoting tourism in the area.

In many ways, this was a totally designed landscape: the dams and valve towers are pre-eminent in it, but the engineered roads with their railings, bridges and retaining walls are all part of a scheme whose reach can be mapped in the distribution of even minor architectural features such as culverts and parapets. Not least of these is the series of retaining walls which line the roads: these are variously rock-faced stone, random rubble, and orthostatic blocks infilled with drystone walling, the latter probably derived from a local vernacular tradition of field walling. It was a landscape designed to blend in with the remnant of what had gone before, however. The richer valley-bottom fields and farms, roads and bridges were to disappear below the water. What survived was the remnant broadleaved woodland on the steep, uncultivated valley sides and the more marginal fields and smallholdings on the fringes of the moorland.

As in the case of a number of other large late 19th-century and early 20th-century reservoir schemes, Birmingham Corporation’s Elan valley scheme was accompanied by large-scale afforestation, as an alternative to animal grazing and as a means of managing the purity of water runoff, particularly in the immediate vicinity of the reservoirs. William Linnard has stated that ‘over 1,000 acres had been planted before 1918, mainly with Scots pine and European larch, but also with Japanese larch, Douglas fir, Sitka spruce and Corsican pine. One plantation of European larch, with Scots pine for shelter at the higher elevations, planted in 1904–05, was awarded the Silver Medal at the Royal Agricultural Society’s show in 1919’. The plantings were carefully blended in with the existing ancient and replanted broadleaved woodland which formerly existed around the valley sides, of which remnants survive around the fringes of the Caban-coch, Garreg-ddu and Penygarreg reservoirs. Most of the catchment area for the Elan reservoirs was to remain unplanted, however, experience elsewhere having shown that the loss of water by transpiration resulting from afforestation, unlike much of the neighbouring hill land in Ceredigion and north Breconshire which is now cloaked in forestry planted in the early 20th century.

This engineering-architecture has a strong stylistic signature and constructional vocabulary. It is a kind of civic baroque, and is very theatrical, which has become popularly styled ‘Birmingham Baroque’. It relies on exaggeration: the use of over-scaled detailing on the stonework of the dams, for example, is used symbolically to suggest strength; the materials were presumably largely local, and it seems likely that the designers of the scheme were seeking a style that would somehow be apt for the rugged terrain; its context therefore is at least partly in the long-fascination with the picturesque in English architecture and landscape architecture.

Elan Village, much of which is dated 1909, was built to house maintenance workers, replacing the earlier timber settlement built to house the workforce engaged on the construction of the reservoir scheme. The village represents a perfect arts and crafts ensemble including houses, estate office, and other ancillary structures. It again uses local stone and possibly slate, but employs styles which were common currency in the arts and crafts movement, and whilst seen as a vernacular revival had little to do with specific regional traditions. The village is best seen as another manifestation of the picturesque, very carefully composed and integrating open spaces and planting as part of its overall design. Its gentler architectural language is perfectly adapted to the domesticity of its purpose, contrasting with the robust muscularity of the architecture of the reservoirs themselves, again fitting into notions of fitness for purpose which would have been common currency at the time.

The works in the Elan valley themselves were carried out by direct administration. The aqueduct, by contrast, was carried out largely by contractors and is generally shows less concern at blending in with its surroundings. The 73 miles of aqueduct linking the Elan valley with the service reservoir at Frankley, about 7 miles from the centre of Birmingham, at 600 feet above sea level, a drop of about 170 feet. For about half its length the water flows at atmospheric pressure in a brick-lined conduit laid in a cut-and-cover trench, but where the aqueduct is below the hydraulic gradient, where it crosses valleys for example, it flows in cast iron pipes under pressure and included, when first constructed, about 13 miles of tunnel. The initial scheme used two 42-inch diameter pipes, to which two more 60-inch diameter pipes were added in 1919 and 1961 which increased the capacity of the aqueducts to 75 million gallons a day. The distinctive and somewhat obtrusive brick-built ‘Washout Chamber’ to the north-east of Coed-y-mynach farm is characteristic of the structures that mark the line of the aqueduct that reaches across Radnorshire and the Midland to the Frankley reservoir, skirting Rhayader, Nantmel, Bleddfa and Cleobury Mortimer.

‘In every childhood there are, I suppose, certain features in the physical environment which exercise a preponderating effect on the imagination. Such, for me, without doubt, was the building of the Elan Valley reservoirs which impounded the wild waters of the Rhayader Massif in Radnorshire, diverting them from their natural outlet, which was by way of the Wye and the Bristol Channel to the Atlantic, into the sewers of a city which lay on the eastern side of the central watershed, and discharging them finally, by way of the Trent, into the North Sea.’

The House Under the Water was one of a series of Midlands novels by Young set ‘like beads along the string’ that was the Elan valley scheme’s pipeline and in this instance set in the border country popularized by late 19th and early twentieth century authors as A. E. Housman and Mary Webb. Its precursor, Undergrowth, written in collaboration with his brother and published in 1911 when he was in his twenties, also describes the building of a dam in a Welsh valley. His novel, The Black Diamond, takes place along the track of Birmingham’s Elan valley aqueduct. The House Under the Water which charts the romantic entanglements of Griffith Tregaron and his family from their removal from Worcestershire to the ancestral seat at Nant Escob, set in a deep Welsh valley, and the eventual abandonment of the house and valley as it was slowly drowned by the flooding of the river Garon. One of the underlying themes of the novel, by a novelist widely admired during his lifetime but considered by some to be over-sentimental, was the contrast between the harshness of the Welsh valleys and mountains within easy reach of the more hospitable farming landscape of the Midlands. The nature of the place had irrevocably changed, but Phillipa, the heroine of the novel, no doubt expressed a popular opinion about the Elan valley, that the spirit of the place ‘resurgent, inviolable, had perfected out of man’s disfigurement, a new loveliness surpassing any that conscious man could achieve’:

‘a new earth, if not a new heaven. For the earth that she knew and loved had passed away and the waters lay everywhere . . . in two shining lakes whose clear surface, swept by the draughts curling through the valley, danced with crystalline wavelets, which lapped their shores in an innocent gaiety, or, when flaws of wind passed, spread mirrors of indigo in whose depths the reflected mountains appeared to dream, as though lost in the contemplation of their own still beauty’.

The severe drought of 1937 provided a warning that the city of Birmingham would need to increase capacity, and although plans for the new reservoir, requiring a further parliamentary bill, were at an advanced stage in 1939 construction work was delayed due to the onset of the Second World War. The dams and reservoirs in the Elan valley were seen as obvious targets for sabotage or bombing during the Second World War since this would threaten Birmingham’s principal water supply and were guarded by employees and units of the Home Guard throughout the war, visible remnants of this period being two hexagonal red brick pillbox gun emplacements in Coed y Foel at the junction between the Caban-coch and Garreg-ddu reservoirs. A small, 35-foot high masonry dam which had been built across the Nant-y-gro stream, upstream of the Caban-coch dam, at an early stage in the late 19th century to supply the workmen’s village was to play a notable role in May and July 1942, being used during the early and highly secretive preparations for Barnes Wallis’s ‘Dambusters’ raids on the dams of the Ruhr valley in 1943. The bombed Nant-y-gro dam survives in much the same condition in which it was left during the war.

Work on extending the capacity of the Elan valley reservoirs was taken up again after the end of the war. The advances in engineering and mechanisation that has taken place during the course of the earlier 20th century permitted the construction of a broader and taller dam than had originally been envisaged, about 1.5 kilometres downstream of the dam proposed for the Pant-y-beddau reservoir. The reservoir, started in 1946 and officially opened by Queen Elizabeth in 1952 as one of her first official engagements, almost doubled the supply of water to Birmingham from the Elan valley.

About 470 men worked on the construction of the dam, 56 metres high and 355 metres in breadth, who unlike the workforce employed on the earlier reservoirs were all housed in the local community and transported to the site by road. The workforce included about a hundred Italian stonemasons due to a shortage of skilled workmen because of the repairs due to war damage being undertaken in a number of British cities at the time, including the restoration work being undertaken on London’s Houses of Parliament. Building materials were mostly transported by road from the railway depot at Rhayader.

The reservoir was to be the largest in the Elan valley complex. It is built of concrete but at considerable extra cost in terms of both materials and labour the dam was faced in rock-faced stonework from South Wales and Derbyshire and incorporated other design features in order to harmonize with the aesthetic standards of earlier dams in the Elan valley. Water is released from the reservoir by either of two 1.2 metre diameter pipes, each side of the dam base, which discharge into the river Claerwen.

Further proposals to extend the capacity of the Elan valley reservoirs were under active consideration in the early 1970s, which focused upon a new ‘High Dam’ which would in effect replace the Craig Goch dam. The scheme failed to go ahead, though it is evident from a report presented in 1973 that this new engineering works would have broken with tradition:

‘at the time of their construction these structures were technological achievements of the highest order. A new and higher dam at Craig Goch would, in our view, best be designed in a style appropriate to its size and the times in which we live.’

In addition to the original turbine houses below the Caban-coch dam, hydroelectric turbines were added to the Claerwen, Craig Goch, Penygarreg dams and the Foel tower in 1998. The generating sites have been concealed as far as possible and are connected by underground cables.

The Elan Estate has been managed to protect the quality and quantity of the water supplies since 1892 and partly as a consequence its moorlands, bogs, woodlands, rivers and reservoirs are of national importance for plants and wildlife. Today, the estate is divided into 43 holdings covering some 17,402 hectares, five managed directly by the estate and the remainder are tenants of the Elan Valley Trust. The moorland within the watershed of the reservoirs is predominantly used for sheep farming, together with a limited number of cattle and a handful semi-wild Welsh mountain ponies.

Abse, J., 2000. Letters from Wales, Bridgend.

Anon, 1813. Cambrian Traveller’s Guide, London.

Baker, A.R.H & Butler, R.H. (eds), 1973. Studies of Field Systems in the British Isles, Cambridge.

B[anks], R.W., 1880. ‘Grange of Cwmtoyddur’, Archaeologia Cambrensis 2, 4th ser., 30-50.

Banks, R.W., 1885. ‘On a bronze dagger found at Bwlch y ddey faen, Breconshire’, Archaeologia Cambrensis 2 , 156.

Barclay, T., 1898. The Future Water Supply of Birmingham.

Baughan, P.E., 1991. A Regional History of the Railways of Great Britain, Volume 11, North and Mid Wales, Nairn.

Bidgood, R., 1986. ‘Early families of Rhydoldog, Llansanffraid Cwmteuddwr, Part 1’, Transactions of the Radnorshire Society 56, 35-51.

Bidgood, R., 1987. ‘Early families of Rhydoldog, Llansanffraid Cwmteuddwr, Part 2’, Transactions of the Radnorshire Society 57, 34-48.

Bidgood, R., 1992. Selected Poems, Bridgend.

Bidgood, R., 1995. ‘Nantgwyllt, Part 1’, Transactions of the Radnorshire Society 65, 33-46.

Bidgood, R., 1996a. ‘Nantgwyllt, Part 2’, Transactions of the Radnorshire Society 66, 16-30.

Bidgood, R., 1996b. The Fluent Moment, Brigend.

Bidgood, R., 2003. Abergwesyn: Bwlch y Ddau Faen, Places. y Man a’r Lle 6, Gwasg Gregynog.

Bowen, E.G., 1950. ‘The monastic economy of the Cistercians of Strata Florida’, Ceredigion 1.

Bowen, E.G., 1965. Wales: A Physical, Historical and Regional Geography, London, 2nd edn.

Bromwich, R., & Evans, D. Simon, 1992. Culhuwch and Olwen, Cardiff.

Browne, D. & Hughes, S., 2003. The Archaeology of the Welsh Uplands, RCAHMW.

Buckley, S.L. & Waler, M.J.C., 2001. ‘The Flandrian vegetation history of upland Mid-Wales — Bryniau Pica’, in M.J.C. Walker & D. McCarroll (eds), The Quaternary of West Wales: Field Guide, Quaternary Research Association, 93-102.

Bull, D.E., 2003. Aberglanhirin, Elan Valley: Tir Gofal Historic Environment report – HE2, CPAT Report 602.

Burgess, C. B., 1962 , ‘The Bronze Age in Radnorshire: a re-appraisal’, Transactions of the Radnorshire Society 32, 7-24.

Burnham, H. , 1995. A Guide to Ancient and Historic Wales: Clwyd and Powys.

Cadw 1999. Register of Landscapes, Parks and Gardens of Special Historic Interest in Wales. Part 1: Parks & Gardens, Powys.

Cadw 2001. Register of Landscapes of Special Historic Interest in Wales, CCW/Cadw/ICOMOS UK.

Cadw & CCW 2003. Guide to Good Practice on using the Register of Landscapes of Historic Interest in Wales in the Planning and Development Process, Cadw Welsh Historic Monuments and Countryside Council for Wales, July 2003.

Cambrian Archaeological Projects 2001. Uplands Survey Initiative – Landscape No: 55. The Elan Valley, Cambrian Archaeological Project Report No. 161.

Carter, D., 1840? Rhaiadr Gwy: a poem illustrated by notes historical and topographical, antiquarian and archaeological, Llanidloes.

Caseldine, A., 1990. Environmental Archaeology in Wales, St David’s University College, Lampeter.

CCW 2001. The LANDMAP Information System/System Wybodaeth LANDMAP, Countryside Council for Wales, 2001, version 1.

Clark, J., 1794. A General View of the Agriculture of the County of Radnor, London.

Colyer, R.J., 1966. The Welsh Cattle Drovers, Cardiff.

Colyer, R.J., 1977. ‘Welsh cattle drovers in the nineteenth century’, National Library of Wales Journal 17, 392-409.

Coppack, G., 2000. The White Monks. The Cistercians in Britain 1128–1540, Stroud.

Countryside Commission 1988. The Common Lands of Wales [Map and Gazetteer].

Cragg, R., (ed.) 1997. Civil Engineering Heritage: Wales and West Central England, London.

Davies, D. R., 1948. ‘Old Music Manuscript’, Transactions of the Radnorshire Society 18, 69.

Davies, E., 1980. ‘Hafod, hafoty, and lluest: their distribution, features and purpose’, Ceredigion 9.1, 1-41.

Davies, E., 1984-85. ‘Hafod and lluest’, Folk Life 23, 76-96.

Davies, J. L. & Webster, P., 1977. Archaeological Survey, Site of Enlarged Craig Goch Reservoir.

Davies, Walter, 1815. A General View of the Agriculture and Domestic Economy of South Wales, London, 2 vols.

Dowden, E., 1886. The Life of Percy Bysshe Shelley, 2 vols.

Drake, A.J., 2000. Cambrian Way, Tewkesbury.

Faraday, M.A., 1996. ‘The Radnorshire Lay Subsidy of 1543-1545’, Transactions of the Radnorshire Society 66, 82-129.

Faraday, M.A., 1989. ‘The Radnorshire Hearth Tax Return of 1670. Part One’, Transactions of the Radnorshire Society 59, 29-58.

Fenn, R.W.D. & Sinclair, J.B., 1989. ‘ “Our ubiquitous friend”. S.W. Williams, F.S.A., F.R.I.B.A., F.S.I., 1837-1899’, Transactions of the Radnorshire Society 59, 116-33.

Fenn, R.W.D. & Sinclair, J.B., 2003. ‘Stephen W. Williams of Penralley, 1837–1899’, Archaeologia Cambrensis 149 (2000), 89–121.

Fenn, R.W.D., Fenn, A.C.W. & Sinclair, J.B., 2002. ‘Radnorshire: some polite perceptions of its landscape’, Transactions of the Radnorshire Society 72, 84-120.

Flemming-Williams, L. & Myhill, C., 2003. ‘The Monks Trod: a path of time and place’, Rural Wales, Summer 2003, 12–17.

George, T.N., 1970. British Regional Geology: South Wales, HMSO, 3rd edn.

Godwin, F. & Toulson, S., 1977. The Drovers’ Roads of Wales, London.

Gregory, D., 1994. Radnorshire: A Historical Guide.

Grimes, W. F. , 1937. ‘Recent finds of prehistoric implements’, Bulletin of the Board of Celtic Studies 8, 96.

Grimes, W. F. , 1951. Prehistory of Wales.

Guest, C., 1849. The Mabinogion from the Llyfr Coch o Hergest and other ancient Wesh Manuscripts, with an English Translation and Notes, 3 vols, London.

Hall, G.W., 1993. Metal Mines of Southern Wales, Herefordshire.

Hankinson, R., 2003. Elan Valley (NE) Upland Survey, Powys: Field Survey, CPAT Report 516.

Haslam, R., 1979. The Buildings of Wales: Powys, Cardiff.

Hawkins, D., 1985. ‘The Groves of Cwm Elan’, Transactions of the Radnorshire Society 60, 45-9.

Hogg, A.H.A. & Jones, G.D.B., 1979. ‘The Roman marching camp at Esgairperfedd (Radnor)’, Bulletin of the Board of Celtic Studies 22, 274-6.

Hopkinson, C. & Speight, M., 2002. The Mortimers, Lords of the March, Logaston.

Hoskins, W.G. & Stamp, L. Dudley, 1963. The Common Lands of England and Wales, London.

Houlder, C., 1974. Archaeological Guide to Wales.

Howse, W., 1952. ‘The Turnpike System’, Radnorshire Society Transactions 22, 12-21, 64-5. Howse, W.H., 1949. History of Radnorshire, London.

Jarrett, M.G., 1969 (ed.). The Roman Frontier in Wales, Cardiff.

Jenkins, J.G., 1962. Agricultural Transport in Wales, National Museum of Wales, Cardiff.

Jenkins, J.G.., 1969. The Welsh Woollen Industry, National Museum of Wales, Cardiff.

Jerman, H. N. , 1936 , ‘The Bronze Age in Radnorshire. Part I’, Transactions of the Radnorshire Society 6 , 33-48.

Jones, F.L. (ed.) 1964. The Letters of P.B. Shelley, Oxford University Press.

Jones, I.E., 1977. ‘The mapping of Radnorshire before the Ordnance Survey’, Transactions of the Radnorshire Society, 47, 13-24.

Jones, N.W., Walters, M. & Frost, P., forthcoming. Mountains and Orefields: Metal Mining Landscapes of Mid and North-East Wales, Council for British Archaeology Research Report.

Jones, S.R. & Smith, J.T., 1963. ‘The houses of Breconshire. Part I. The Builth District’, Brycheiniog 9, 1-77.

Jones, T., 1909-30. History of the County of Brecknock, 4 vols in 2 vols, Glanusk edn.

Jones, T. Thornley, 1951 , ‘The Giant’s Grave’, Transactions of the Radnorshire Society 21, 50 .

Judge, C.W., 1997. The Elan Valley Railway: The Railway of the Birmingham Corporation Waterworks, Usk.

Kidner, R.W., 2003. The Mid-Wales Railway, Usk.

Lapworth, H., 1899-1900. ‘The construction of the Elan Aqueduct: Rhayader to Dolau’, Min. Proc. Instn Civ. Engrs, 140, 235-48.

Lewis, M. Gwyneth, 1977. The Printed Maps of Radnorshire 1578-1900, National Library of Wales.

Lewis, S., 1833. A Topographical Dictionary of Wales, London.

Lewis Lloyd, L.B., 1950. ‘Squire Lewis’s hounds’, Transactions of the Radnorshire Society 20, 65-6.

Lewis, W.J., 1967. Lead Mining in Wales, University of Wales Press, Cardiff.

Linnard, W., 2000. Welsh Woods and Forests, Llandysul.

Lloyd, J.E. & Jenkins, R.T., 1959. he Dictionary of Welsh Biography Down to 1940, The Honourable Society of the Cymmrodorion.

Lloyd, T., 1989. The Lost Houses of Wales, 2nd edn, London.

MAFF 1985. ‘Agricultural Land Classification of England and Wales’, Map and Technical Report No. 11, HMSO, London.

Mansergh, J., 1894. ‘The Birmingham Water Scheme’, in R.E.Tickell, The Vale of Nantgwilt: A Submerged Valley, London, 11-17.

Mansergh, E.L. & Mansergh, W.L., 1911-12. ‘The works for the supply of water to the City of Birmingham from mid-Wales’, Min. Proc. Instn Civ. Engrs 190, 23-25.

Maxfield, C.E., 2003. Penglaneinon, Elan Valley, Tir Gofal Historic Environment Report – HE2, Clwyd-Powys Archaeological Trust Report 602.

Miles, D. H. and Haddon-Reece, D., 1996. ‘Welsh Dendrochonology Project’, Vernacular Architecture 27, 106-13.

Miles, D. H. & Worthington, M. J., 1999 , ‘List 103: Welsh Dendrochronology Project – phase three’, Vernacular Architecture 30.

Milton, D., 1999. The Elan Valley Way, Oldbury.

Moore, P.D., 1994. ‘The history of vegetation in Cardiganshire’, in J.L. Davies & D.P. Kirby (eds), Cardiganshire County History. Volume 1. From the Earliest Times to the Coming of the Normans, Cardiff.

Moore, P.D. & Chater, E.H., 1969a. ‘Studies in the vegetation history of mid Wales. I. The post-glacial period in Cardiganshire’, New Phytologist 68, 183-96.

Moore, P.D. & Chater, E.H., 1969b. ‘The changing vegetation of west-central Wales in the light of human history’, Journal of Ecology 57, 361-79.

Moore-Colyer, R., 2001. Roads & Trackways of Wales, Ashbourne.

Moore-Colyer, R., 2002. Welsh Cattle Drovers, Ashbourne.

Morgan, R., 1998. A Sudy of Radnorshire Place-Names, Llanrwst.

Morgan, R. & Powell, R.F.P., 1999. A Study of Breconshire Place-Names, Llanrwst.

Morton, R., 1997. The Building of the Elan Valley Dams.

Nash-Williams, V.E. , 1950. The Early Christian Monuments of Wales.

Newman, C. W., 1960. ‘Shelley and Nantgwyllt: three new letters’, Transactions of the Radnorshire Society 30, 17-30.

Nicholson, G., 1813. The Cambrian Traveller’s Guide in Every Direction.

O’Dwyer, S., 1936. The Roman Roads of Wales. IV. Cardiganshire and Radnorshire, Newtown.

Oliver, R. C. B., 1983. ‘The diary of Captain Frederick Jones. Part I, 1789 to 1799’, Transactions of the Radnorshire Society 53, 28-56.

Owen, D. E., 1911. ‘On a stone axe hammer from Clap-yr-arian, Rhayader ‘, Archaeologia Cambrensis 11, 153.

Owen, D. H., 1989. ‘The Middle Ages’, in D. H. Owen (ed.), Settlement and Society in Wales, Cardiff.

Owens, M., 2000 , ‘Llanerch-y-Cawr — An old Welsh longhouse restored by the Welsh Water Elan Trust’, Rural Wales, 13.

Parker, W. K., 1977. ‘The visits of Walter Davies to Radnorshire, 1802-11’, Transactions of the Radnorshire Society 47, 52-61.

Peate, Iorwerth C., 1944. The Welsh House, 2nd edn, Liverpool.

Peate, Iorwerth C., 1972. Tradition and Folk Life: A Welsh View, London.

Pennant, T., 1784. A Tour in Wales, London.

Pierce, T. Jones, 1950. ‘Strata Florida Abbey’, Ceredigion 1.

Platt, C., 1969. The Monastic Grange in Medieval England, London.

Price, J. E., 1936. ‘Schedule of Radnorshire Place names. Parish of Llansantffraed Cwmdauddwr’, Transactions of the Radnorshire Society 6 , 66-78.

Prince, H., 1989. ‘The changing rural landscape’ 1750–1850’, in G.E. Mingay (ed.), The Agrarian History of England and Wales Vol. 6, 1750–1850, Cambridge University Press, 7–83.

Pugh, J. A., 1936. ‘Notes on the parish of Cwmteuddwr’, Transactions of the Radnorshire Society 6 , 61-5.

RCAHM, 1913. Inventory of the Ancient Monuments in Wales and Monmouth III – County of Radnor.

RCAHMW 1992. An Inventory of the Ancient Monuments in Glamorgan, Volume III: Medieval Secular Monuments, Part II: Non-defensive, Royal Commission on the Ancient and Historical Monuments of Wales.

RCAHMW 1997. An Inventory of the Ancient Monuments in Brecknock (Brycheiniog); The Prehistoric and Roman Monuments, Part 1, Later Prehistoric Monuments and Unenclosed Settlements to 1000 A.D., Royal Commission on the Ancient and Historical Monuments of Wales.

Rees, S. E., 1997. ‘The historical and cultural importance of ponds and small lakes in Wales, UK’, Aquatic Conservation 7 (1997), 133–9.

Rees, W., 1924. South Wales and the March, 1284-1415, Oxford.

Rees, W., 1968. An Historical Atlas of Wales, 3rd edn, London.

Richards, M., Welsh Administrative and Territorial Units, Cardiff.

Ridyard, G.W., 1994a. ‘Water power in Radnorshire’, Transactions of the Radnorshire Society 9, 36-48.

Ridyard, G.W., 1994b. ‘Supplementary notes on watermills of Radnorshire – part II’, Melin 10, 2-39.

Ridyard, G.W., 1997. ‘Supplementary notes on watermills of Radnorshire – part IV’, Melin 13, 55-87.

Ridyard, G.W., 1998. ‘Supplementary notes on the watermills of Radnorshire – part V’, Melin 14, 15-50.

Rossetti, W.M., 1894. ‘Shelley at Cwm Elan and Nantgwilt’, in R.E. Tickell, The Vale of Nantgwilt: A Submerged Valley, London, 17-35.

Rowlands, D.W.L., 2003. ‘David Jones’s reminiscences on old Llanwddyn and the construction of Lake Vyrnwy’, Montgomeryshire Collections91, 109-36.

Rudeforth, C.C., Hartnup, R., Lea, J. W., Thompson, T. R. E., & Wright, P.S., 1984. Soils and their Uses in Wales, Bulletin No. 11, Soil Survey of England and Wales, Harpenden.

Savory, H.N., 1958. ‘The Late Bronze Age in Wales. Some new discoveries and intepretations’, Archaeologia Cambrensis 107, 3-63.

Savory, H. N., 1980. Guide Catalogue of the Bronze Age Collections.

Sheail, J., 1971. Rabbits and their History, Newton Abbot.

Silvester, R.J., 1991. ‘Tomen Llansantffraid: a motte near Rhaeadr, Powys’, Medieval Archaeology 35, 109-14.

Silvester, R.J., 1994a. Elan Valley Uplands Field Survey 1993/94, Clwyd-Powys Archaeological Trust Report 97.

Silvester, R. J. , 1994b. Radnorshire Historic Settlements , Clwyd-Powys Archaeological Trust Report 92 .

Silvester, R.J., 1995. ‘Pillow mounds at Y Foel, Llanllugan’, Montgomeryshire Collections 83 (1995), 75–90.

Silvester, R.J., 1997. Deserted medieval and later rural settlements in Radnorshire, Clwyd-Powys Archaeological Trust Report 227.

Silvester, R.J., 2000. ‘Deserted rural settlements of medieval and later date in north-eastern and central Wales’, in Atkinson, J.A., Banks, I. & MacGregor, G., Townships to Farmsteads: Rural Settlement Studies in Scotland, England and Wales, British Archaeological Reports, British Series 293.

Smith, L.T., (ed.), 1907-10. The Itinerary of John Leland the Antiquary, in or about the years 1535-45, 5 vols, London.

Smith, P , 1963 , ‘The longhouse and the laithe house: a study of the house-and-byre homestead in Wales and the West Riding’, in L. Alcock & I. Ll. Foster, Culture and Environment, 415-37.

Smith, P., 1988. Houses of the Welsh Countryside, 2nd edn.

Smith, R.T. & Taylor, J.A., 1969. ‘The post-glacial development of vegetation and soils in north Cardiganshire’, Transactions of the Institute of British Geography 48, 75-95.

Soil Survey 1983. 1:250,000 Soil Map of England and Wales, Sheet 2, Wales, Soil Survey of England and Wales.

Stephens, M., 1986. The Oxford Companion to the Literature of Wales, Oxford University Press.

Suggett, R., 1996. ‘The chronology of late-medieval timber houses in Wales’, Vernacular Architecture 27, 28-35.

Suggett, R., forthcoming. Houses and History in the Welsh Marches: the Medieval Houses of Radnorshire and their Transformations, RCAHMW.

Taylor, C., 1975. Fields in the English Landscape, London.

Thirlwell, A., 2003. William and Lucy: the other Rossettis, Yale.

Thomas, S.P., 1984. ‘Twelve miles a day: some thoughts on the drovers’, Transactions of the Radnorshire Society 54, 58-66.

Thorpe, L., 1978. Gerald of Wales: The Journey through Wales and The Description of Wales, Harmondsworth.

Tickell, R.E., 1894. The Vale of Nantgwilt: A Submerged Valley, London.

Timberlake, S., 2002. ‘Medieval lead-smelting boles near Penguelan, Cwmystwyth, Ceredigion’, Archaeology in Wales 42, 45-60.

Timberlake, S., 2003. Excavations on Copa Hill, Cwmystwyth (1986-1999): An Early Bronze Age Copper Mine within the uplands of Central Wales, British Archaeological Reports 348.

Tomos, D., 1980. Michael Faraday in Wales, London.

Wade-Evans, A.W., 1938. Nennius’s History of the Britons.

Walters, M., 1993. Powys Metal Mines Survey, Clwyd-Powys Archaeological Trust Report 89.

Ward, C., 2002. Cotters and Squatters: Housing’s Hidden History, Nottingham.

Welsh Water (undated). Metal Mines of the Elan Estate, Elan.

Wiliam, E., 1986. The Historical Farm Buildings of Wales.

Wiliam, E., 1992. Welsh Long-houses. Four Centuries of Farming at Cilewent, National Museum of Wales, Cardiff.

Williams, D., 1954. The Rebecca Riots, University of Wale Press, Cardiff.

Williams, D. H., 1990. Atlas of Cistercian Lands in Wales, University of Wale Press, Cardiff.

Williams, D. H. , 1999 , ‘Stephen William Williams: The Archaeologist and Soldier: The Man and his Family’, Transactions of the Radnorshire Society 69, 116-42.

Williams, D. H., 2001. The Welsh Cistercians, Leominster.

Williams, G., 1991. ‘Romantic and Realist: Theophilus Evans and Theophilus Jones’, Archaeologia Cambrensis 149, 17-27.

Williams, J., 1905. A General History of the County of Radnor, Brecon.

Williams, S.W., 1889. The Cistercian Abbey of Strata Florida, London.

Williams, S.W., 1894. ‘The Grange of Cwmteuddwr’, in R.E. Tickell, The Vale of Nantgwilt: A Submerged Valley, London, 36-40.

Wiltshire, P.E.J. & Moore, P.D., 1983. ‘Palaeovegetation and palaeohydrology in upland Britain’ in K.J. Gregory, Background to Palaeohydrology, [John Wiley & Sons], 433-51.

Winchester, A.J.L., 2000. The Harvest of the Hills, Edinburgh University Press.

Young, Francis Brett, 1932. The House Under the Water, Combridge, Birmingham.

British Dams Sociey: information about British dams.

Elan Valley Trust: natural history, wildlife and conservation.

Gathering the Jewels: archive photographs, drawings and text (search for Elan Valley)

Powys Digital History Project: archive photographs, drawings and text (follow Six Powys Communities Online, Rhayader and the Elan Valley).

671 Squadron: The “Dambusters”: about wartime activities in the Elan valley.

The Lost Club Journal: article by J. Howard about Francis Brett Young (follow Lost Club Authors, Francis Brett Young).