The application of archæological theory to the study of "Celtic" water cults, with particular reference to holy wells studies.

A talk given to the Perth Source Moot, April 1999

by Feòrag NicBhrìde

I can only find a couple of Holy Wells in Perthshire so, rather than making this talk rather too short, I thought I'd look at the way holy wells and the so-called "Celtic" water cult have been studied throughout history with particular reference to works on Scottish sites. In particular, I want to look at the way that folkloric practices associated with holy wells are assumed to be survivals of a prehistoric (usually defined as 'Celtic') water cult.

I would like to start by stating that the study of the relationship between 'Celtic' water cults and holy wells appears not to have progressed since the mid-19th century. This extract from an article published in 1974 sums up the generally held belief:

"Streams, rivers, fountains, springs and wells have been accounted sacred from earliest times. Each possessed its nymph or deity who exacted tribute, sometimes even sacrifice. It seems now to be generally accepted that well-worship in Britain originated before the Christian era; that the Christian missionaries found it in vogue on their arrival, tolerated it at first and utilised it afterwards for their own ends. Consequently, with a veneer of Christianity, and the substitution of a saint's name, water worship has held its own to our day." (Bonser 1974, 29).

This assumption permeates nearly all modern writing on holy wells and seems to be accepted without question--something of which I admit guilt myself in my own study of the wells and springs of Leeds--and attempts at an alternative explanation are rare. Congès notes the reliance on folklore in the compilation of a corpus on Gallo-Roman water cults and warns against the citation of modern practices in explaining the past. I assumed you probably didn't want to hear my attempts to speak French at this point!

Alternative interpretations can be found in closely related studies. Although not involving a holy well site, archæologists Knüsel and Carr (1995) offer a scientific explanation for the deposition of crania in the Thames and its tributaries. The debate over the function of the Bronze Age Wilsford Shaft is also illuminating. Both of these will be discussed later. The definition of 'Celtic' has been debated for some time and I have absolutely no intention of discussing that tonight!

The tradition of description

One of the major problems of holy well studies is an enduring tradition of cataloguing and description with little or no attempt at explanation. When an explanation is offered, it is frequently influenced by the religious beliefs of the author, rather than from an archæological basis. Hence, Roman Catholic authors seem to uncritically accept the tales of the saint associated with the site. This can be particularly seen in the two guidebooks sold at St. Winefride's Well in Wales (Charles-Edwards n.d.; David 1971). David suggests the site had been used as a baptistry in the 7th century, (David 1971, n.p.) Charles-Edwards (n.d., 2) makes no mention of anything prior to the twelfth century. Both authors then concern themselves with the history of pilgrimage at the site and in description of the architecture.

One of the earliest surviving descriptions of a holy well, Matthew Mackaile's (1664) account of the Balm Well at Liberton in Edinburgh begins with a simple description of the site and its history. After criticising 'that execrable Regicide and Usurper, Oliver Cromwell, with his rebellious and sacrilegious complices" (Mackaile 1664, 118), for defacing the well, Mackaile discusses the origin of the spring. He considers a number of possibilities derived from Aristotle, Ecclesiastes and contemporary scientific research. In the end, Mackaile decides his Biblically derived hypothesis is clearly the right one as "it be according to the express Word of God" (Mackaile 1664, 126). After this, he asks and answers number of questions concerning the chemical composition of the water.

In his address to the Scottish Ecclesiological Society in 1925, William Burnett felt confident enough to assert that any pre-Christian origin for these wells was irrelevant and that their creation anew in Christian guise (born-again wells?) was all that existed now. "There are no Holy Wells outside the Christian era", he proclaimed, before explaining that beforehand, the wells had been occupied by evil spirits and people were scared of them and that they had only acquired healing characteristics because of the appropriate saint!

Neo-pagan and earth mysteries-oriented authors similarly offer explanations based on their beliefs. Whelan and Taylor (1989, 3-7) claimed, in "Yorkshire Holy Wells and Sacred Springs" that "the sanctity of water has been acknowledged since remotest times" and "water taken straight from the earth has a mysterious quality… This is the quality of sanctity. Whether or not we associate these sites with specific ancient deities (or their saintly equivalents) does, perhaps, for us today, make no great difference--they are places where the Spirit of water is experienced, through which a contact can be made between ourselves and the formative forces of the cosmos". Although Whelan and Taylor emphasise that their belief is a contemporary one, John Michell, in The Earth Spirit, its ways, shrines and mysteries., is happy to extrapolate backwards in time:

"Where fresh water runs there runs spirit, and this is particularly so wherever water springs up from below the earth, for it comes from the realm of the earth goddess and bears her gifts. Properly every spring has its season of efficacy when its virtues are most generously displayed. In times before doctors, psychiatrists, marriage guidance officials, newspaper horoscopes, drugs and artificial fertilisers, all their functions were exercised by the spirits of the local springs, who required no payment but respect and attention" (Michell 1975, quoted in Whelan and Taylor 1989, 8).

The bulk of reports of holy wells in the antiquarian journals of the nineteenth century can be divided into two types: catalogues of all the known sites in a geographical area (see, for example, Walker 1883) and descriptions of single sites. In the latter category falls an early example of rescue work. Prior to moving St. Margaret's Well from below a railway depot to Holyrood Park, a thorough survey of the buildings was made and published (Laing 1858). According to a plaque attached to a stone marking the original site of this well, the site was also excavated in the early 1970s prior to the building of Meadowbank stadium. Unfortunately, this excavation appears to be unpublished. Most publications of single sites from this period (the late nineteenth century) seem to take the form of a thorough survey of the remains, such as Irvine 1898, and some sites are even excavated (Collier 1906) It appears as if the sites were surveyed simply because they were there!

There have been several attempts to catalogue wells in Scotland over time. The best-known example of such a work for England is Hope's "Legendary Lore of the Holy Wells of England", published in 1893, in which, county-by-county, he listed the wells he knew about and described them and the folklore associated with them. What is less known is that this book was compiled from a series of articles he wrote in the Antiquary, and that these articles covered Scotland as well.

A few years earlier in 1883, J. Russel Walker's "Holy Wells in Scotland" appeared in the Proceedings of the Society of Antiquaries of Scotland. He starts in now-familiar style: "The virtues of water seem to have been recognised by every tribe and people..." before describing a hotchpotch of lore from various times and places and re-telling the obligatory Columba and the druids story as his explanation of how old pagan wells became holy. After much discussion of the sites' Christian aspects, he describes and illustrates a number of wells he has visited and gives a list of wells by saint's name and other attributes, with snippets of lore.

Another interpretation found in the nineteenth century involved colourful descriptions of purported human sacrifice, based uncritically on the Roman writings of Caesar and Strabo. Cuming Walters tells us:

"Wherever the Celtic element is, there will be found the superstition concerning water leading to strange rites, and sometimes to appalling sacrifices.

"The Franks offered human sacrifices to those rivers they were about to cross, and in Wales horses were annually sacrificed at St. George's Well, near Abergelan." (Walters 1991,7)

This view of bloodthirsty savages still persists in the popular mind, and is frequently invoked by certain Christian fundamentalists when denigrating neo-pagan beliefs.

An example of the persistence of this view can be seen in the interpretation of a panel on the Gundestup Bowl which depicts a person being dipped into a similar bowl. The bowl is widely believed to be 'Celtic', and the scenes to depict 'Celtic' religion. A discussion of evidence to the contrary is not appropriate here, and it suffices that the writers attribute the bowl to the Celts. Piggott refers to a Roman account of a sacrifice to drowning and proposes that the bowl shows " a representation of this very rite" (Piggott 1968, 84). In a rare moment of scepticism, Anne Ross points out that, assuming the bowl is Celtic, the scene is just as likely to depict a cauldron of rebirth as seen in the Irish and Welsh myths (Ross 1968, 279).

As an aside, Anne Ross is one of the few researchers in the field who has used any kind of theoretical background. Her view is a diffusionist one --the Celtic water cult, and hence holy wells, originate in Hallstatt and La Tène Europe (Ross 1968, 257). Her persistent reference to "Belgic Britons" (Ross 1968; Ross 1974, 53) suggests she regards the cult as something brought over by an incoming group.

The archæological record itself does not support the idea that the ancient British practised human sacrifice. Despite the widespread presence of human remains in wells and shafts which can be dated to the Iron Age and Romano-British periods, almost all of them have been put there after the death of the individual concerned. The most convincing interpretation of them is that it is a special form of burial.

Most recent work remains very much in the tradition of cataloguing sites (see, for example, NicBhrìde 1994, Shepherd 1994, Whelan and Taylor 1989)--the only obvious difference between this work and the notices of the Victorian antiquarians is that the middle-class men have been replaced by middle-class women! There is a large body of data, yet rarely is any use made of it. Ross (1974, 58) sees a need for further work: "The survival of custom and belief about wells into later tradition, and the fact that they form a visible focus for ritual reflects their religious importance in pagan times. This topic requires an independent study..." This statement is interesting. The assumption that folkloric veneration of holy wells is a pre-Christian survival is there.

Given the predominance of research using an earth mysteries framework, a contextual approach is very appealing--the relevance of water is not as obvious as it may initially appear, given the distribution of holy well sites. As Hartnett pointed out in 1947: "Certainly, the existence in Pagan Ireland of water worship and its survival in an emasculated form in the cult of the holy well requires some such explanation in view of the abnormally wet climate obtaining here" (Hartnett 1947, 5). There are many questions that could be considered by trying to get inside the heads of Iron Age people: "Why is water important in a wet climate?" is a single example.

There are very few examples where the existence of a holy well has been directly related to Iron Age or Romano-British remains. An interesting observation is made by Childe--the (Neolithic) cairns on Rousay are all near springs, including a site called "St. Mary's Well" (Childe 1942, 141). He notes that Crawford has observed a similar phenomenon for the Cotswold tombs. This has interesting implications for Anne Ross's view. Can this phenomenon be observed anywhere else? Is there some symbolism relating water and death that has been missed?

There are a few examples where questions have been asked of the data and explanation sought. The deposits in Coventina's Well were considered from a number of perspectives. Originally, it was thought the deposits were a result of a long, continuous process of deposition of offerings (Allason-Jones and McKay 1985, 9), with a final deposition as the result of Theodesian edicts, the material being put into the well to hide it from Christian iconoclasts. Others suggested the Christians did the damage themselves (ibid.)

One important question is whether watery places really were considered holy in pre-Christian Britain and Ireland. Well-known is the debate between two of the authors of the report on the Bronze Age Wilsford Shaft (Ashbee, Bell and Proudfoot 1989, 128-138). Using environmental evidence, the functionalist Bell concludes that the shaft is a water supply. Ashbee, "has developed his ideas into a social archæological narrative of British prehistory" and "is particularly concerned therefore with disentangling the non-material component of prehistoric monuments" (Ashbee, Bell and Proudfoot 1989, 128). He prefers a ritual interpretation.

Finally, Knüsel and Carr's (1995) challenge to the suggestion that human cranial remains in the Thames and its tributaries represent deliberate ritual deposition as part of an Iron Age water cult indicates the value of a scientific approach. The original article by Bradley and Gordon associated the crania with late prehistoric metalwork, but acknowledged problems with the association caused by both crania and metalwork being recovered by dredging or as opportunistic finds (Knüsel and Carr 1995, 163). The original report also used spatial distribution and craniometric analyses as supporting evidence (ibid.). Nine of the crania were submitted for accelerator dating and were found to range in date from the Neolithic through to the Anglo-Saxon period, casting doubt that the rest of them could be from a single period of time. Knüsel and Carr studied the distribution of human bones found in the Thames to see if there was another explanation for the apparent association and found that different parts of the skeleton are affected differently by moving water. Complete crania move fast in a current and can be deposited a long way from the original point of deposition (Knüsel and Carr 1995, 154-5). Jawbones do not move so far, and are often lost due to rolling. A lack of signs of gnawing or knife marks suggests the crania were part of a complete head when they entered the water (ibid.). Modern data for suicides in the Thames was used to suggest that the crania could be the result of a normal rate of drowning, the distribution being the result of natural processes. A similar study reported in British Archaeology (CBA 1994, n.p.) concluded that apparent Bronze Age ritual deposits in wet places could be explained by accidental loss.


So, to summarise, the study of prehistoric water worship, and in particular the question of the origins of holy wells, has not been touched upon by archæological theory. There is a tradition of cataloguing sites that began in the nineteenth century and continues to this day in which the origins of holy wells in prehistoric water worship has been accepted uncritically. There are important questions, which need to be asked with respect to these sites. Functional and scientific approaches can also cast light on whether or not these watery sites really were sacred in prehistory.


ALLASON-JONES, L. AND McKAY, B. (1985) Coventina's Well: a shirine on Hadrian's Wall. Hexham, Trustees of the Clayton Collection, Chesters Museum.

ASHBEE, P., BELL, M. AND PROUDFOOT, E. (1989) Wilsford Shaft: excavations 1960-62. London, English Heritage.

BONSER, K. J. (1974) "Spas, wells and springs of Leeds". Publications of the Thoresby Society: Miscellany vol. 16 part 1, pp 29-50.

BURNETT, W. (1925) "Holy Wells in Scotland". Transactions of the Scottish Ecclesiological Society, vol. viii, part 1. pp 57-69.

CHARLES-EDWARDS, T. (n.d.) Saint Winefride and Her Well. The Historical Background. Holywell, W. Williams and Son.

CHILDE, V. G. (1942) "The chambered cairns of Rousay". Antiquaries Journal, 22 pp 139-42.

COLLIER, 'MRS' (1906) "St. Clether: his chapel and holy wells". Journal of the British Archaeological Association, new series vol. xii, pp 83-90.

CONGÈS, A. R. (1994) "Culte de l'eau et dieux guérisseurs en Gaule romaine". Journal of Roman Archaeology, 7, pp 397-407 (book review).

DAVID, C. (1971) St Winefride's Well: a history and guide. Naas (Co. Kildare), Leinster Leader.

HARTNETT, P. J. (1947) "Holy wells of East Muskerry, Co, Cork". Journal of the Cork Historical and Archaeological Society, 52, pp 5-17.

HOPE, R. C. (1893) The Legendary Lore of the Holy Wells of England. London, Elliot Stock.

IRVINE, G. G. (1898) "The church and well of St. Doulough, County Dublin". Journal of the British Archaeological Association, new series vol. iv, pp 329-338.

KNÜSEL, C. J. AND CARR, G. C. (1995) "On the significance of the crania from the River Thames and its tributaries". Antiquity 69, pp 162-9.

LAING, D. (1883) "Statement relative to the removal of St. Margaret's Well". Proceedings of the Society of Antiquaries of Scotland, new series, vol. iii, pp 365-367.

MACKAILE, M (1664) "The Oyly-Well: or A Topographico-Spagyrical description of the Oyly-Well, at St. Catharines-chappel in the Paroch of Libberton" in Mackaile, M. Moffet-Well. Edinburgh, Robert Brown, pp 117-138

MICHELL, J. (1975) The Earth Spirit, its ways, shrines and mysteries. London, Thames and Hudson.

NICBHRÌDE, F. (1994) The wells and springs of Leeds - a survey. Preston, Pagan Prattle Publications.

PIGGOTT, S. (1968) The Druids. London, Thames and Hudson.

ROSS, A. (1968) "Shafts, pits, wells - sanctuaries of the Belgic Britons?" in Studies in ancient Europe (eds J. M. Coles and D. D. A. Simpson), pp 255-285, Leicester.

ROSS, A. (1974) Pagan Celtic Britain. London, Sphere Books.

SHEPHERD, V. (1994) Historic wells in and around Bradford. Loughborough, Heart of Albion Press

WALKER, J. R. (1883) "Holy Wells in Scotland". Proceedings of the Society of Antiquaries of Scotland, new series, vol. v.

WALTERS, C. (1991) Holy Wells. Loughborough, Heart of Albion Press (reprint of a chapter from Walters' 1898 book The Church Treasury)

WHELAN, E. AND TAYLOR, I. (1989) Yorkshire Holy Wells and Sacred Springs. York, Northern Lights.

<-- Wells index.