WELL RESEARCHED: Holy Wells and Other Holy Places



Holy Wells and Other Holy Places

by Jeremy Harte


To what extent can place-names be used to trace a pre-Christian sacred landscape? And what differences, if any, can be identified in the ‘holiness’ of different types of well-name? The author examines the use of ‘halig’ in place-names, and shows how it cannot, contrary to popular belief, indicate elements of a sacred pagan topography. Firmly separating ‘halig’ from other, demonstrably pagan, place-name elements, the author identifies a chronology of holy well-name development and sets it in the context of a specifically Christian veneration which developed naturally to form part of the cult of saints in the later Middle Ages.


Editors’ note

We regret that the current limitations on the capabilities of fonts in html coding mean that we are unable to show the macrons over vowels in certain place-name elements used by Jeremy Harte in the article which follows.   Please note that in all cases where the following words are used in the text, they should be read as follows:-


word cited in text correct form


Editors’ note


The significance of ‘holy’

Pagan or Christian sanctity?

A chronology of well-names

Cultural development and well-names

Final thoughts



A thousand years and more have passed since Anglo-Saxon surveyors, working their way through the parish of Halse in Somerset, passed by ‘an ash tree which the ignorant call holy’: fraxinum quem imperiti sacrum vocant (Grundy 1935; Turner 1953, p.118). Ignorant they may have been, but they probably had a clearer understanding of the local landscape than many of their descendants. Ash trees are not uncommon in the vale of Taunton Dene. So what made this one special?

Actually, the tree was known to the locals simply as �an halgan �sc. Turning a holy tree into a tree ‘which the ignorant call holy’ is a significant gloss, apparently added by a sniffy scribe from Winchester who was given the job of translating the boundary marks, which derive from an original of 854, into Latin. The tree lay in a rather remote part of the abbey’s domains, (OSGR ST 127 277 on the modern map), so the anonymous monk was not disallowing the possibility of its holiness from personal experience. He must have had reasons for believing that trees, generally speaking, could not be holy and that anyone who thought they were was a credulous rustic.

Other people thought differently. There was another holy tree in Somerset north of the Mendips, at Hallatrow, Helgetrev 1086. In the East Riding there is Hallytreeholme, Halitreholme c1180, ‘island with a holy tree’. While the ash may have been sacred in Somerset, there is also Holyoaks in the Rutland parish of Stoke Dry, Haliach 1086 (Ekwall 1960, pp.212,246-7).

In fact holiness was quite common in the Anglo-Saxon landscape. Besides the trees, there were holy places of one sort or another at High Halstow and Lower Halstow in Kent, Hastoe Farm in Hertfordshire and Austy Wood (Halwestowe) in Warwickshire (Reaney 1961, p.126). Forms with stow are normal, but there are two with stoc, Halstock (Halganstoke 998) in Dorset (Mills 1986, p.81) and Halstock Farm (Halgestok 1240) on the edge of Dartmoor (Gover et al. 1931-2, p.203).

Still, the most common element to be compounded with halig commemorates neither a place nor a tree, but a spring of living waters: wella, or often wylle in the West Saxon forms, welle in the West Midland ones. In the North Country the equivalent Old Norse kelda is often used (Smith 1956, 2.3, 2.250). Significant settlement names of this kind occur in Huntingdonshire, Kent, Lincolnshire and Northumberland (Gelling 1984, p.31). Half of the halig compounds in Ekwall refer to wells or springs, and the proportion would probably be higher if a systematic survey of field and other minor names could be added to the reckoning. Out of the nineteen forms with halig found by the English Place-Name Survey for Devon, fifteen relate to wells (Gover et al. 1931-2, pp.67, 105, 109, 122, 141, 203, 245, 247, 250, 268, 309, 313, 323, 352, 367, 478, 483, 537).

Holy wells have been a minor theme in landscape history ever since there was such a thing. To say that a spring was a holy well was to classify it, without much hesitation, as a focus of popular beliefs – superstitions which, under a thin Christian veneer, preserved the immemorial traditions of the pagan past. And there are still people who write like this.

But of late the idea of the holy well has become a little more complicated. The tightly knotted conceptual bundle which held together healing wells, wells featured in popular customs, haunted wells, wells named after saints and so on has come undone. If we no longer believe in ‘pagan survivals’ then we have lost the unifying concept which made sense of almost all traditions about springs and fountains by regarding them as aspects of a single, archaic cult of wells.

The significance of ‘holy’

In this deconstruction of ‘the holy well’, one strand of tradition is left behind: the existence in large numbers of sites called halig wella and the like. If there is no longer a self-evident, timeless model for what a holy well is, then how are we to understand the motives which made people speak of a spring as the Holy Well? In this context the Winchester scribe’s casual remark about the ash tree of Taunton Dene takes on a new significance. It is unusual for a mediaeval writer to reflect on the interpretation of halig as an element in toponymy.

He wasn’t doing it as an academic exercise, of course. Neither were the draftsmen of thirteenth-century diocesan instructions when they insisted that ‘not stones nor woods, trees or wells… should be worshipped as if they were holy’. This is part of a longstanding tradition of pastoral control over laymen’s claims to have discovered holiness in the landscape. In the 990s, probably the same time that the Winchester scribe made his point, Aelfric was ticking off his audience for offering gifts to ‘some earthfast stone or tree or well-spring’. This three-fold denunciation – of wells, trees and stones – was already routine when Anglo-Saxon preachers derived it from Continental sources (Rattue 1995, pp.78, 87).

Christian clergy were definitely condemning the idea – even the name – of the holy tree and the holy well. It suggests that halig might bear some kind of non-Christian, pre-Christian sense. The search for underlying meanings of this kind has become almost a scholarly convention, despite the reluctant acceptance in English Place-Name Elements that ‘it is generally not possible to find any heathen allusion in the use of this word’ (Smith 1956, 1.225). The Rutland place-name Holyoaks has certainly been interpreted in this way: recent discussion would have it suggest ‘an enclave where pagan practices flourished’ (Cox 1994, pp.298-9). But this is more than the evidence will bear.

There were wells in Anglo-Saxon England which fell on the wrong side of the boundary between religion and superstition, but they were not called holy. At Elwell in Dorset and (confusingly) Holywell in Lincolnshire, wella is compounded with h�l, ‘omen’; and Rumwell in Somerset and Runwell in Essex, it appears with run, ‘secret’ (Smith 1956, 2.250),. These names suggest practices which it would be very difficult to bring within a Christian interpretation, and it is quite clear from the laws prohibiting frehtwellas, ‘springs used for divination’, that these places were linked with paganism in the official mind. Nonetheless, they flourished, examples being known from Oxfordshire, Shropshire, Nottinghamshire and the West Riding (Hough 1996).

But ‘paganism’ is not a neutral term in the primary sources. It is meant to condemn the practices against which the author writes his sermon or treatise. The very persistence of this literature of denunciation – and it stretches from the fifth century to the twelfth, and beyond – suggests that its authors were marginal to the society which they had so obviously failed to influence. As Valerie Flint has suggested, these purist preachers often took their stand because they disapproved of the work of their contemporaries in dedicating and consecrating landscape features. One priest’s pagan abomination might very well have been his colleague’s popular shrine (Flint 1991, pp.204-216). If we are to pin the label of paganism onto Holyoaks in Rutland, what are we to make of Cressage in Shropshire, ‘Christ’s oak’? There is a Christow in Devon, which bears the same relationship to the Halstow forms (Cameron 1996, p.124).

An oak tree, or a well, has a direct physical presence. You can walk around it, listen to it, respond to its atmosphere. But a stow is something rather more abstract – a place where people assembled, in the earlier records, and in the later ones simply a place. Saints own their stows, and have given their names to them at Bridestow, Marystowe, Petrockstowe and other places in Devon and Cornwall, but the emphasis seems to be on their ownership, rather than their presence (Smith 1956, 2.158-161). Christow, then, is not ‘the place where Christ can be found’, but ‘the estate property offered up to Christ’. The same goes for stoc. Visitors to Halstock may be disappointed if they expect a revelation from the sacred landscape. Its name seems to be linked to the cult of St Juthware there, afterwards reinforced by the possession of the property by Sherborne Abbey. It is true that the place of St Juthware’s cult, with its holy well and tree, may have inspired a tingle of religious awe. But is that what mediaeval holiness was really about? Another local saint has his stoc at nearby Stockwood, once Stokes sancti Edwoldi, and there is Kewstoke on the Somerset coast. But then a little to the east we find North and South Stoke, so called because they belonged to Bath Priory, and in Dorset there is Stoke Abbott, the property of the Abbot of Sherborne (Smith 1956, 2.153-5). The abbot may have been a good man but name-givers were interested in him as a landlord, not as a sacred presence. The holy names which govern stoc and stow – and, by imputation, halig itself – are words used to talk about ownership, not about feelings.

We accept these things, perhaps with some regret, but we would still like to feel that there was something special about �an halgan �sc at Halse, or about the numberless holy wells – something evocative, numinous, which made people call them holy. This may be a distinctly modern outlook. Certainly we find that later in the Middle Ages, people could be very matter-of-fact about holiness. There is a Holy Brook at Reading – le Granators Broke als le Hallowed Broke 1552; it seems to have given a name to le Haliwatereslane 1301 (Gelling 1973, 1.11, p.172). It took its name, not from any veneration of the water, but because it flowed past Reading Abbey: in fact it looks suspiciously as if the water did its holy service operating the abbey mills. An excellent purpose, to be sure, but not the sort of thing to get them all worked up at the Sacred Land Project.

Holiness in place-names is not necessarily something to do with experience: it may be more to do with ownership. After all, ownership is what most place-name records are about. Holy Island succeeds Lindisfarne, not as a statement about the haunting quality of the place, but to indicate that it belongs to a monastery and no longer to the Lindisfaran. Holywick in Buckinghamshire belonged to Medmenham Abbey (Smith 1956, 1.225). A similar interpretation of Halliford in Middlesex, halgan forde 962, and Hallington in Northumberland, Halidene 1247 (Ekwall 1960, p.212) would see them as the ford and valley which led to, or belonged to, a religious body. Other landmarks include the halig stan of Halsdon House at Dolton in Devon (Gover 1931-2, p.367) and Holystone in Northumberland, Halistan 1242 (Ekwall 1960, p.247). It is tempting, but probably misleading, to regard these as further enclaves of pagan activity, furtively assembling at megalithic sites. In fact the Northumberland Holystone gets its name from the monastery on the site.

Halmpstone farm in Devon lies near the parish boundary of Bishops Tawton. This village is said to have preceded Crediton as the diocesan see for Devon, which would account for the name, Halgmerston 1285 (Gover 1931-2, p.352). The ‘holy boundary stone’ is holy, not because the stone itself is to be revered, but because it marks the land of the holy man. A general respect for boundaries would explain the name of the Holybourne in Hampshire, Haliburne 1086 (Coates 1989, p.94), which demarcates two manors.


Pagan or Christian sanctity?

The description of natural features as ‘holy’ does not necessarily mean that they received veneration from Christians, let alone from pagans. But there is a group of place-names which can be described, without ambiguity, as pagan: those which relate to the worship of the old Germanic gods. The corpus of these place-names has been subject to revision, but there are still enough of them to offer a picture of pre-Christian practice (Gelling 1978, pp.158-161). They are quite different from names of the Holyoaks type. The names of the discredited gods – Woden, Thunor, Tiw and Frig – typically compound with words for open spaces, feld and leah; sometimes with beorh and hlaew, which might mean either ‘hill’ or ‘barrow’. The principal word for temple, hearg, belongs exclusively with dun, ‘hill’. Weoh or wig, which seems to mean a temple of smaller and more local status, follows the same pattern, although there is a single instance of it combined with wella, at Wyville, Uuiuuella 1106-23, in the Lincolnshire parish of Saltby (Ekwall 1960, p.541). Apart from that, the pagan elements are not used to qualify stones, trees or wells, while halig is never found with a word for open spaces and only once with one for a hill (Hillborough, halig beorh, in Kent – Smith 1956, 1.225). Clearly there are two separate religious traditions at work here.

This is unusual, since all of these words have a common origin in pre-Christian Germanic worship. Wig is a substantive formed from weoh, ‘sacred’, which has a semantic field close to that of halig. The early missionaries in England chose halig to translate Latin sanctus, and rejected weoh as pagan: but Wulfilas, faced with the same linguistic or doctrinal problem among the Goths, chose their equivalent word weihs and firmly rejected hailags (Green 1998, p.360). At first sight the words might appear interchangeable. But there is more to it than that.

The adjective heilag is found in every Germanic language, and takes its meaning from a complex of ideas represented by two forms of heil, a noun meaning ‘good fortune’ and an adjective meaning ‘healthy, whole’. Other associated senses include ‘omen’, ‘blessing’, and in a Christian sense ‘salvation’. The adjective heil, ‘fortunate’, can yield a substantive meaning ‘talisman’ – just as in English we can speak of ‘the Luck of Edenhall’. But this is not just the kind of luck which we hope for as we toss a coin. Heil was the noble power, handed down from the gods, which carried kings to victory in battle. Through this divine quality, priests could see into the future and let the land grow rich. Healers, too, could be strengthened to banish the demons of disease. And the tools of their trades – swords, ritual staffs, amulets and so on – they were heilag too (Green 1998, pp.16-20).

Places in the landscape, including water features, could be heilag. This is certainly the case for the Norse form heilagr. The ninth-century poem speaks of ‘the holy places of the powers’ (Davidson 1964, p.76). When the hero Helgi was born, ‘the holy streams were flowing from the hills of heaven’ (Helgakvi�a 2). Place-names of this sort were given in formal ceremonies: the hill called Helgafell in Iceland was consecrated to the cult of Thor by the settler Thorolf Moster-beard (Hreinsson 1997, 5.134).

There must have been similar processes at work in Britain, although the only evidence we have for halig in a pre-Christian context refers, not to a place, but a time. Bede obligingly tells us that September was Halegmonath in the old calendar, and that ‘Halegmonath means ‘month of sacred rites’’ (De Temporum Ratione 330, in Bede 1999, p.53). Presumably this was something on the lines of a harvest festival.

But the holy trees, stones and wells of Anglo-Saxon topography are not likely to perpetuate a pre-Christian geography. One good reason for doubting this is their distribution: there is no concentration of sites in the old pagan districts, the ones to which the names with hearg and wig are confined. Quite the reverse. We have already met with �an halgan �sc in the fertile, late-settled farmlands of Taunton Dene.

Can we trust the apparent evidence for geographical distribution? There is little doubt that the concentration of pagan place-names in eastern England reflects real geographical facts – even though there is no simple, one-to-one relationship between the original worship of the gods and the survival of references to them in place-names. But the evidence for an uneven distribution of halig is much less certain, because this element is commonest in field names and the names of small settlements: the earliest volumes of the English Place-Name Survey, those which cover the heartland of Old English paganism, didn’t include names of this kind.

All that can be said is that in Dorset, Somerset and Devon, where names indicating pagan worship are almost unknown, we find numerous place-names of the Holywell type. Taking into account the dates at which the western shires of Wessex were annexed into an English-speaking kingdom, this suggests that halig was being used to form local names after the eighth century, in many cases after the ninth. And this is to a large extent confirmed by the chronology of wella as a place-name

A chronology of well-names

Wella is one of the words which are not employed in the earliest phase of name-formation in English (Cox 1976). When it does appear in early instances, it is usually a simplex name, as in Wells, Wielea 766. Despite the importance of this site, the religious associations which went back to (or were grafted onto) a Roman origin, and the later tradition which would make it St Andrew’s Well, there was no question of designating it as halig. This point is made by James Rattue, who follows it by at least four similar instances – Willesden in Middlesex, Wilton in Somerset, Welton in the East Riding and Well in the North Riding. In each case wella forms either a simplex name or the first element of a name: and it refers to a spring which, in later tradition at least, we find as a holy well (Rattue 1995, pp.58, 64, 84; cf. Scherr 1986, p.82).

So the first class of place-names for holy wells consists of names in wella with no overt reference to the sanctity of the waters. The second class, the halig wella names, is later than this. Where kelda rather than wella is the second element, these names clearly postdate Scandinavian settlement. But even in the untroubled South-west, there are signs of a relative chronology of name formation. Fontmell Magna in Dorset, Funtemel 877, is a surviving Celtic name from funton, mailo-, ‘spring by the bare hill’. The survival of the name suggests that Fontmell was an estate passed over from the British to the West Saxon Church in 700 or thereabouts (Barker 1984). The name would refer to Springhead, source of the Fontmell Brook and clearly a significant central place, even though there are no records of it being venerated. Instead we meet with a halgan welle in 932, on the parish boundary (Mills 1989, p.110), evidently a later and marginal formation.

The third class of place-names, those in which holy wells are named after saints, originates at the same time as the halig wella type. Rattue has located a possible Eadburgeswelle of 795, a Cynburge wella of 926 and a Ceollan wylle of 952 (Rattue 1995, p.63). Formally, there is nothing to distinguish these sites from the very large number of wells that are named after landowners. It is only with the benefit of local records that we can identify Edburga, Cyniburg and Ceolla as saints, their ownership of the wells being spiritual rather than territorial. It is possible, given the imperfect survival of traditions about local Saxon saints, that many other apparently tenurial well names could actually be reflecting lost local cults – the sort that, further west, were kept alive by the flickering fires of Cornish separatism. What is certain is that, at the time of Domesday, place-names of the halig wella type were much more common than any which can be identified as preserving the names of saints.

By about 1300, the use of ‘Saint’ as a prefix makes it obvious that a holy man or woman is being referred to. Plemstall in Cheshire, Pleymundestowe 1291, is named after an Archbishop of Canterbury who retired here to a hermit’s life and died in 914. The name of Seint Pleymondes Well, 1302, shows as much (Cameron 1996, p.125). By the turn of the thirteenth and fourteenth centuries, ‘Saint’ had become a required title in formations of all kinds, and place-names of the Kirkbride type gave way to forms like that of St Bees.

Until recently, with the publication of studies such as The Living Stream, there was very little emphasis on how well cults had changed over time. Perhaps because of this, the literature on holy wells – extensive though it is – does not contain many early forms for their names. There are few pre-Reformation records for place-names of the St X’s Well type: indeed, there are comparatively few records for mediaeval holy wells at all.

But it seems likely that this, the fourth class of place-names for holy wells, became current after the late thirteenth century. The saints who most often give their names to wells are not obscure local figures, but the sacred protectors found throughout Christendom – James, John, Catherine and Peter, with the Virgin Mary heading the list. The dedication of wells to St Anne reflected an interest in the Holy land from 1150 onwards (Morris 1989, p.90) although the first recorded instance of St Anne’s Well as a place-name seems to be at Brislington, outside Bristol, in the 16th century (Quinn 1999, pp.147-9). Physical evidence for veneration, in the form of well chapels and so on, also dates from this period (Hutton 1991, p.167).

This new, more personal interest in patron saints led to onomastic speculation, as secular well names were rewritten to put them under the protection of a sacred figure. Babwell in the Somerset parish of Cucklington, derived from the OE personal name Babba, had become associated with St Barbara by the fifteenth century (Scherr 1986, p.85). Chad’s well, a pre-Reformation dedication at Chadshunt in Warwickshire, owes its sanctity to a reinterpretation of the name of the village, Caedelesfuntan 949 (Rattue 1995, p.86).

In fact the national popularity of St Chad’s Wells probably owes much to mangling of OE ceald ‘cold’. In the same way one suspects that many examples of St Helen’s Well derive from a reworking of h�l ‘omen’. But there are insufficient data to determine this point: not because her cult was obscure – there are 233 sites dedicated to St Helen in the British Isles, according to one survey (Jones 1986, p.64) – but because so little work has been done to locate its manifestation in time, either through documents or archaeology. Of course, there is little doubt that many cults recorded in the nineteenth century are likely to be much older. But inferences about pre-Reformation cults are not the same as actual mediaeval evidence, and we need a national survey of the latter.

In the absence of this, it is not at all certain when place-names of the second class ceased to be given to wells. Holywell Wood in Monmouth (Halliwell 1343) must postdate the arrival of the English language with an English castle garrison in 1071 (Charles 1938, p.262). In the thirteenth century, folk-etymology could create new names of the halig wella type, through the alteration of initial elements such as hol, as easily as it could discover doubtful saints. But there was a growing prejudice in favour of saints. The well at Stainland, south of Halifax, must have been known by a halig wella form up until 1279-1324, when it was the home of Henry de sacro fonte: after that it appears as St Helen’s Well (Haigh 1986). In the North Riding, at Middleham, a saint owes her very existence to the name of a well. This spring in this village appears to have been halig kelda, but by the late thirteenth century, when kelda had ceased to be a vocabulary word, the name was reinterpreted as that of a St Alkelda. The parish church is dedicated to her along with the well (Whelan 1986, p.4).

Halig wella forms survived where they had become transferred from the original spring to a settlement – as in the most famous case, Holywell in Flintshire. After it was granted to an abbey at Chester in 1093, this became the English name for the settlement which was Treffynnon in Welsh. The holy well itself was Ffynnon Gwenfrewi to the locals (Charles 1938, p.224). Clearly, in the eleventh century at least, English conventions of place-name formation discouraged direct reference to a saint, using halig instead. The contrast with the native language is very obvious, for neither Welsh nor the other Celtic languages have any form of well-name analogous to our halig wella; and there are no names of the fourth class, with a word for ‘saint’, either. ‘It should be noted that wells bearing English names usually retain the title ‘saint’, e.g. St Mary’s Well, but those bearing Welsh names rarely do so, e.g. it is Ffynnon Seiriol but never Ffynnon Sant Seiriol’ (Jones 1954, p.6).

There is no onomastic history to the Welsh well-names; they are all of one type. But the names of English holy wells do form a chronological series, even if exact dates are hard to come by. This suggests that they can act as clues to the cultural evolution of the holy well, and of other holy places.


Cultural development and well-names

Halig names typically refer to obscure places. The ‘ignorant’ inhabitants of Taunton Dean had chosen the most out-of-the way corner in which to venerate �an halgan �sc. Central places, under the watchful eye of major clerics and landowners, were less likely to develop irregular devotions of this kind. Several of the halig wella forms refer to uncultivated areas or upland farms: this may explain the frequency of names in Devon, where – due to the dispersed pattern of settlement – more early mediaeval farm names survive than elsewhere. And this sense of the marginal continues in more recent sources. Holy Well, at Ditcheat in Somerset, is typically obscure: ‘a small patch of water under a hedge, on the right side of the road leading from Evercreech Junction to Castle Cary, and about half-a-mile distant from the station’ (Horne 1923, p.28).

Impressions can be misleading, especially when several of the early halig wella forms are derived from boundary marks in charters. This may make the class in general seem more peripheral than it really is. Nevertheless, regional comparisons can be made for the topography of well-names from the second and fourth class. The evidence from Dorset (Harte 1985; Rattue 1992) show that the county has place-names of both kinds, some being halig wella forms, while others are of the St X’s Well type. Keeping to those place-names recorded before 1800, there are eight called holy wells, and six dedicated to saints.

It is true that an exclusive citation of early forms will underestimate the evidence for mediaeval traditions. There are saints’ wells which only appear in recent records, such as Lady Well at Hermitage and St Andrew’s Well at Bridport, but which nevertheless appear to go back to the Middle Ages: others, however, may come from the imagination of antiquarian vicars. Similarly there are Holywell place-names recorded after 1800 – at Abbotsbury, East Stoke, Fifehead Neville and Trent – which seem unlikely to be romantic coinages, although they may well have come into being through the corruption of an element other than halig. But where there is so much uncertainty in the evidence, it seems best to take 1800 as a cut-off point for both classes of place-name.

The two sites recorded in charters as halig wella – Portesham and Fontmell Magna – are naturally enough at a distance from the main settlement, but most of the other forms also appear as the names of small farms; in the later references, the wells seem to be in outlying fields. This is true of Child Okeford, Gillingham, Hazelbury Bryan, and Radipole. Two of the sites, however, are found in towns – Shaftesbury and Sherborne. There is also a Haliwell at Wilksworth Farm near Wimborne, recorded in 1389 (James Rattue pers. comm.): this would seem to be another small farm site.

On the other hand, all but one of the saints’ wells are centrally located. Alhallon well 1545 is just off the main streets of Wareham. St Andrew’s Well in Lyme Regis, which appears in fifteenth-century documents, lay within the town. In Beaminster, St Mary’s Well, on record since 1692, lies in the next street to the church.

Two saint’s wells, those at Cerne Abbas and Corfe Castle, are associated with incidents in the life of saints. At Corfe the well first appears in the Early South-English Legendary (�17 lines 119-121) in a retelling of the events surrounding the death of St Edward, King and Martyr;

‘�e put �are he was feorst i-founde: a welle �are gan springue,
Fair and cler, �at i-last: and is ofte gret botningue,
�at men cleopieth  to �is day: ‘seint Edwardes welle”.

This text of the Legendary was compiled in the 1280s, apparently at Gloucester. This would make it one of the earliest references to a well incorporating ‘saint’ in its name.

At Cerne Abbas a mediaeval well chapel dedicated to St Augustine survived long enough to be recorded by the antiquary Thomas Gerard. It lay in the grounds of the Abbey, just north of the village. William of Malmesbury’s twelfth-century life of the saint introduces a miracle at a well, but again this is a standard story from saints’ lives, rather than being inspired by a pre-existing spring. St Augustine’s Well was identified, book in hand, after the local tradition had been created – perhaps as late as the fifteenth century, which from a few surviving details of the stonework would seem to be the date of the chapel (Castleden 1996, pp.87, 112).

All of these wells are associated with churches or chapels: it is because of this that they are central to their settlements. The odd one out is St Whyte’s or Wite’s Well – there is no early form for the name, but in 1545 it was being pointed out on the slopes of a hill, over a mile south of Whitchurch Canonicorum, where the saint used to retreat for prayer and contemplation. This well is unusual also in being dedicated to a local Anglo-Saxon saint, rather than one of the major late mediaeval intercessors, and seems to be a survival from an earlier state of things.

The tendency to choose a central location – and the wells of Wareham, Lyme, Beaminster and so on appear to have been chosen, rather than spontaneously emerging through some popular process – suggests that these wells are part of the late mediaeval passion for expressing Christianity through minor monuments. The fourteenth and fifteenth centuries were a golden age for market and wayside crosses, bridge and hilltop chapels, hermitages and so on. All of these traditions had existed earlier, but the great majority of sites date from after 1300. It is tempting to see the saints’ wells as part of this pattern.

On the other hand, a case has been made by James Rattue for dating many wells in this class to a much earlier date – they could be from before 900 or thereabouts, in the formative years of the Anglo-Saxon church (Rattue 1995, pp.53-62). Although actual documentation is lacking, there is a close correlation between cults at wells and early church foundations. In fact all the ten sites which he identifies as Dorset minsters have a later tradition of wells. Admittedly the designation of any given site as a minster is a matter of debate among historians, as it was no doubt amongst clergy at the time. The two senses of mynster, ‘monastery’ and ‘central church with daughter foundations’, do not always match up: besides, there is no reason to suppose that the whole landscape was neatly divided into separate, non-competing minster parochiae.

Still, the fact remains that there were a great many central places in the early English church, distinguished by monastic houses, territories with daughter churches, boundaries co-extensive with those of villae regales and so on. And it is at these settlements that we find the saints’ wells. Among the Dorset examples, for instance, Wareham has a major early church, Whitchurch Canonicorum an early cult site and an extended parish, and Beaminster another extended parish with a mynster place-name.

The correlation of wells with early church centres is certainly real, but it does not prove that the wells are early too. Most of the minster sites of the eighth and ninth centuries were well chosen geographically and administratively. For years afterwards, they continued to form a focus for the people of the surrounding countryside: in most cases, they still do. Wareham, Lyme and Beaminster are towns, while Cerne, Corfe and Whitchurch, if not exactly urban, nevertheless occupy a geographical niche for which we seem to have no word – that of a large village providing services to a hinterland of smaller settlements. We are more likely to find holy wells in these central places, just as we find pubs or shops, because people are around to make use of them.

What use did people expect to make of these sites? Other places of veneration which grew in popularity after 1300 – gild chapels and so on – depended on a mutual bond between saint and patron, on a prominent role in the daily business of society, and on a constant traffic of vows and offerings. Saints’ wells were seen in the same way: people intended to make gifts at these shrines and so it made sense to have them by the roadside near inhabited places. The locations selected for typical modern wishing wells, as put up by Rotarians, Lions or other charitable groups, have a lot to tell us here. Access is important, and so is security. Not everyone respects charity now, and presumably there were people prepared to risk offending saints then. So the sacristan of the church or chapel would have to be on hand to clear away the money at regular intervals.


Final thoughts

This is all a far cry from the lonely springs known as halig in the earlier period – say between the eighth and thirteenth centuries. No chapels marked them: as far as we can tell, they received no processions, no offerings. Apart from being venerated and being wet, they have little in common with monuments of the other class and it may be misleading to group them together in the uniform concept of ‘holy well’. Of course it is our detachment from the usual sources of water in the ground that encourages us to find a special mystique in all these sites – even the Rotarian wishing well. Mediaeval observers might have taken a quite different view. A thousand years on, it is not the Anglo-Saxon peasantry who are ignorant: it’s us.

name element.


Barker, Katherine (1984). Institution and landscape in early medieval Wessex: Aldhelm of Malmesbury, Sherborne and Selwoodshire. Proceedings of the Dorset Natural History & Archaeological Society, 106, pp.33-42.

Bede (1999). The reckoning of time. Translated by Faith Wallis. Liverpool: Liverpool University Press.

Cameron, Kenneth (1996). English place-names. London: Batsford.

Castleden, Rodney (1996). The Cerne Giant. Wincanton: Dorset Publishing Company.

Charles, B.G. (1938). Non-Celtic place-names in Wales. London: University College.

Coates, Richard (1989). The place-names of Hampshire. London: Batsford.

Cox, Barrie (1976). The place-names of the earliest English records. Journal of the English Place-Name Society, 8, pp.12-66.

Cox, Barrie (1994). The place-names of Rutland. Nottingham: English Place-Name Society.

Davidson, Hilda Ellis (1964). Gods and myths of northern Europe. Harmondsworth: Penguin.

Ekwall, Eilert (1960). The concise Oxford dictionary of English place-names. Oxford: Clarendon.

Flint, Valerie (1991). The rise of magic in early medieval Europe. Oxford: Clarendon.

Gelling, Margaret (1973). The place-names of Berkshire. Nottingham: English Place-Name Society.

Gelling, Margaret (1978). Signposts to the past: place-names and the history of England. London: Dent.

Gelling, Margaret (1984). Place-names in the landscape. London: Dent.

Gover, J.E.B., Mawer, A. and Stenton, F.M. (1931-2). The place-names of Devon. Nottingham: English Place-Name Society.

Green, D.H. (1998). Language and history in the early Germanic world. Cambridge: Cambridge University Press.

Grundy, G.B. (1935). The Saxon charters and field names of Somerset. Taunton: Somerset Natural History & Archaeological Society.

Haigh, Mike (1986). St Helen’s Well, Stainland. Source (First Series), 4, p.9.

Harte, Jeremy (1985). Dorset holy wells. Source (First Series), 1, pp.3-8.

Horne, Ethelbert (1923). Somerset holy wells. London: Somerset Folk Press.

Hough, Carole (1996). The place-name Fritwell. Journal of the English Place-Name Society, 29, pp.65-69.

Hreinsson, Vi�ar (1997). The complete sagas of the Icelanders. Reykjavik: Leifur Eiriksson.

Hutton, Ronald (1991). The pagan religions of the ancient British Isles: their nature and legacy. Oxford: Blackwell.

Jones, Francis (1954). The holy wells of Wales. Cardiff: University of Wales Press.

Jones, Graham (1986). Holy wells and the cult of St Helen. Landscape History, 8, pp.54-75.

Mills, A.D. (1986). Dorset place-names: their origins and meanings. Wimborne: Roy Gasson.

Mills, A.D. (1989). Place-names of Dorset III. Nottingham: English Place-Name Society.

Morris, Richard. (1989). Churches in the landscape. London: Dent.

Quinn, Phil (1999). Holy wells of Bath and Bristol region. Woonton Almeley, Herefordshire: Logaston.

Rattue, James (1992). An inventory of ancient, holy and healing wells of Dorset. Proceedings of the Dorset Natural History & Archaeological Society, 113, pp.265-268.

Rattue, James (1995). The living stream: holy wells in historical context. Woodbridge: Boydell.

Reaney, P.H. (1961). The origin of English place-names. London: Routledge & Kegan Paul.

Scherr, Jennifer (1986). Names of springs and wells in Somerset. Nomina, 10, pp.79-91

Smith, Albert (1956). English place-name elements. Nottingham: English Place-Name Society.

Turner, A.G.L. (1953). Some Old English passages relating to the episcopal manor of Taunton. Proceedings of the Somerset Natural History and Archaeological Society, 98, pp.118-126.

Whelan, Edna (1986). Holy wells in Yorkshire (part two). Source (First Series), 4, pp.3-7.

Page designed by Rich Pederick (© Living Spring Journal, MM)
Written & maintained by Katy Jordan & Rich Pederick
Created  May 1, MM

%d bloggers like this: