The discovery of the Wilsford Shaft, a 30m (100ft) deep pit, underneath what was thought to be a pond barrow (Wilsford 33a), during excavations in 1960-62 led to disagreement between the excavators, Paul Ashbee and Edwina Proudfoot, on the purpose of the shaft: was it simply a well to access water or used as a place of ritual deposition, perhaps a portal to the underworld? 
Lake Down Barrow Cemetery, Wilsford cum Lake, Wiltshire
The discovery of this deep shaft beneath a pond barrow raised the question of whether all pond barrows have shafts beneath them. Yet magnetometer and resistivity surveys over a sample of three pond barrows at the nearby Lake Down barrow cemetery near Wilsford in Wiltshire in 1996 failed to reveal any indications of further shafts. 
The purpose of the enigmatic pond barrow is still unclear. Pond barrows differ considerably from the broader category of round barrows, although some do contain burials, albeit only in small quantities; a sepulchral role is not thought to have been their primary function. Pond barrows usually occur in association with other Bronze Age monuments, usually within the context of funerary or ceremonial monuments of barrow cemeteries, typically located in prominent locations on rolling downlands, rarely on hilltops. However their positioning within the barrow cemetery, whilst not totally random, does not appear to be influenced by any strict governance. Consequently, as their purpose is little understood, the pond barrow’s association with barrow cemeteries has led to a ritual interpretation.
Less than 100 pond barrows have been identified in Britain and these are generally confined to south and east England, with concentrations in Dorset and Wiltshire, although a few outliers are known further north and east. There are three main concentrations around Stonehenge, Avebury and Winterbourne Abbas, although other examples are known in Berkshire and as far afield as Norfolk and recently (2010) a possible pond barrow has been identified amongst a rare group of earthworks consisting of seven monuments near Llanfyrnach, Pembrokeshire.
If the purpose of the pond barrow remains unclear then the function of the Wilsford Shaft is equally puzzling; clearly “ritual shafts” need to be differentiated from “wells”. By definition a “well” is a deep hole or shaft sunk into the earth to obtain water.
Whereas a “ritual shaft” appears to have been purposefully dug expressly for the ceremonial deposition of objects apparently for reasons other than storage or disposal. Deposits in shafts have been found to include metalwork, pottery, ash and charred wood, human and animal remains. Yet, ritual deposits tend to consist of only a few species of animal and sometimes only a few parts of the animal, but these sacrificial offerings bear no evidence of being butchered for human consumption.
In England and Wales the skulls of cattle and skulls or bodies of horses and dogs are often found underneath Iron Age houses or deposited in nearby pits and in the ditches or ramparts of enclosures. Sheep are rare finds in this context.  A ritual shaft may be seen as a portal to the underworld where votive offerings where deposited to placate the gods.
But what happens when a well receives ritual deposits such as in the widespread custom of dropping votive deposits, such as human skulls or coins, into the abyss – should the well then become classified as a ritual shaft?
Special difficulties arise in the case of shafts and pits that were undoubtedly used as wells or cisterns, that also contain deposits, perhaps from just one single occasion, suggestive of a deliberate act of ritual deposition. Furthermore, there would appear to have been different deposits for significant occurrences in the water supply and just to complicate matters further different deposits are found in different geographical regions. Human skulls are common in wells and possibly testament to a Celtic Cult of the Head closely associated with sacred waters and wells of wisdom. The Celts were well known for making votive offerings into lakes and pools, wells and springs. They placed similar objects into shafts and pits. Rarely are skulls found in holy wells, more usually they occur in wells that have previously been used to supply water for domestic or industrial purposes. This ritual deposition is emphasised at Caerwent where a cistern was given two skulls and some pottery. An adult human skeleton was laid over it as if to seal it before the wall of a house was built above it. However, well deposits are typically interpreted as accidental loss or refuse disposal.
In London there is evidence of the practice of ritual deposition when a human skull was found in the well in Queen Street. The skull appears to have been deposited in the first century after the well had silted up and became unusable. In Cannon Street, also in London, a box was found in a well containing a human cranium, animal bones, including those of two dogs, and pottery dated to the first century. Another skull was found in a third century Roman well at a villa in Northwood, Herts. This appears to have been a complete decapitated head, sandwiched between layers of building debris, suggesting the head was deposited at the time of demolition.
The bones of certain animals are more common than human skulls as deposits in wells. Five dog skulls were found in a well at Caerwent and numerous dogs were found in a deep well associated with a first century shrine at Muntham Court, Sussex. A complete Samian bowl and the remains of sixteen dogs were placed in a second century well at Staines near London. Similar finds have been discovered in Avon, Warwickshire, Wiltshire, Gwent and Bedfordshire; the practice was evidently widespread in southern Britain throughout the Roman period. Pairs of dogs feature in many well closing ceremonies, yet archaeologists invariably interpret this as the fatal ending of a dog fight; an interpretation that just does not stand up to scrutiny. The repetition of canine deposition is so common that it indicates a traditional custom in the sequence of well deposits. In the Romano-Celtic period the remains of dogs seem to be found frequently in association with closure of wells. The regular occurrence of these deposits after the well ceased to be used and covered by demolition material strongly suggests a rite of termination. 
Bronze dog figurine, a probable votive offering from the ‘Romano-Celtic’ temple of Nodens at Lydney Park, Gloucestershire.
Evidently, there appears to be some significance of dogs in the rite of termination. The dog has been man’s closest companion for millennia, the first domesticated animal, yet providing a very specific role in spiritual guardianship throughout the mythologies of the world.  It seems very probable that dogs were linked with some chthonic or underworld ritual. Dogs seem to have fulfilled a special role in Celtic mythology; for example, in the First Branch of the Mabinogi it is Pwyll’s hunting dogs, in pursuit of a stag, that lead him to an Otherworld meeting with Arawn, king of Annwfn, indicating a curious connection between dogs and travel to the realm of the dead.
At the temple to Nodens at Lydney on the Severn, a funnel shaped structure was placed above a pit in the central part of the temple. The shaft was found to contain twenty one coins and a bronze representation of a dog. Nodens was a god of healing and the association of dogs with hunter gods with healing attributes is well attested throughout the Romano-Celtic world. A small bronze dog was also found at Coventina’s Well at Carrawbrough on Hadrian’s Wall, amongst 14,000 coins, pins, bronze heads, a bronze horse, brooches, glass, pottery, shrine bells and a human skull. 
A panel on the Gundestrop cauldron depicts a god, usually identified as Teutates immersing a man head-first into a cauldron with a dog underneath. According to Lucan in the Pharsalia, Teutates, the Gaulish god of the tribe, was propitiated by drowning victims in a cauldron, or vat of water. The iconography is usually interpreted as a ritual act and may be related to the tale of the Second Branch of the Mabinogi in which we see the Cauldron of Rebirth in which “a man who is killed today and thrown in the cauldron, arises the next day he will be as good as he was at his best, except he will not be able to talk”. The dog presumably appears on the panel in his role as guide for the spirits of the dead warriors.
|Teutates panel of Gundestrop cauldron.
The Neolithic- Bronze age complex at Flag Fen, near Peterborough, was a site that was a major focus for funerary rituals for many centuries, from around 1,200 to 200 BC, with most deposition made in the middle of this period. Amongst the objects recovered from the site are three hundred pieces of very fine metalwork of weapons or ornaments, most broken or bent before being thrown into the marsh. Significantly all the deposited material was placed to one side of an impressive alignment of several thousand wooden posts driven in to the marsh between an island and the dry land. The skeletons of two dogs were found buried around the base of the timbers. The posts clearly act as a divisional marker and the dogs seem to have been ritually killed and deposited to serve as guardians of the liminal boundary. 
Another Bronze Age site at Caldicot in Gwent provides evidence for a dog buried in a manner which strongly suggests a role as ritual guardian. Large numbers of animal skulls have been found in wells across the country. Often interpreted as refuse, but the large numbers of specific animals supports a structured deposition of a particular rite rather than a haphazard rubbish tip. Skulls of thirteen oxen, three sheep, as well as two dogs, and a complete skeleton of a Great Northern Diver were found amongst the fourth century fill of a well at Portchester Castle. At another Roman military site, the auxiliary fort at Vindolanda, just south of Hadrian’s Wall, the skulls of ten oxen were found in the late fourth century fill of the well. If the Roman military had an obsession for putting multiple quantities of ox skulls in wells, then in Essex it was horse skulls. Five horse skulls were deposited in a well in Wickford in the fourth century, signifying the commencement of a period of neglect, although the well was subsequently reused. At Chelmsford, a dry shaft that was re-cut in the second century, was found to contain five horse skulls. Termination indicating the well had failed or political change resulting in the desertion of a site.
The situation is further complicated as not all deposits in wells were associated with termination and many deposits are recognised as rituals of commencement. Offerings connected with the rite of commencement, such as complete pots, are usually found at the very bottom, or half way down the shaft. However, as we have seen above, rites of termination may start at the bottom of the shaft and continue in the fill in structured layers.
Clearly it is difficult but important to distinguish between shafts that seemingly serve no practical purpose and appear to have been used from inception for a succession of ritual deposits, and those that were dug and used purely as wells, possibly over a long period of time, and underwent ritual deposition as a rite of termination or commencement, as we have seen above. The offerings may have been intended for the same underworld gods but the purpose of the ritual may have been quite different. 
By the mid-1980s, twenty one “ritual shafts” had been identified. About half of these are concentrated in Kent and Surrey, the rest spread across England from as fas as Exeter and Warwick. Many of these were given the classification simply because they failed to possess any other obvious purpose. Notably, the shafts of south-east England contained more dog and bird bones, whereas those from other regions contained more horse remains, otherwise they held much the same material: pottery, ash, ox bones and human remains. 
We find a classic ritual shaft within the Romano-Celtic temple at Jordan Hill on the South Dorset Downs. The earliest feature on the site was a dry shaft containing a repetitive succession of deposits. At the bottom of this shaft was a rough cist formed of two stones, containing two Roman pots, a short sword, a spearhead, a knife, two pieces of iron and a steelyard. Above this was a layer of stone tiles on which was a layer of ashes and charcoal. Above this was a double layer of tiles, arranged in pairs, with a skeleton of a bird and a Roman coin between each. Another layer of ashes, thereafter in alternating succession a layer of tiles containing bird skeletons and coins and then ashes again. This repetition continued for sixteen layers between the top and bottom of the well but interrupted in the middle by another cist, like that at the bottom of the shaft containing a sword, a spearhead and pots. The birds represented were raven, crow, buzzard and starling, ceremoniously deposited in a succession of rites as the shaft was filled. 
Jordan Hill Romano-Celtic Temple
There are many other examples of dry shafts containing ritual deposits. At Keston, Kent, a sixteen foot shaft was found to contain the articulated skeleton of a dog at the bottom. To the side of the shaft were a pile of animal bones that seem to have been distributed for the next layer of three skeletons of horses, articulated and complete, carefully arranged in a triangle, the second with his head to the tail of the first, the third head to head with that of the first. There were a further seven animal deposits, making ten in total. Pottery evidence indicates the deposits were taking place in the first and second centuries. Nearby, a similar shaft was found to contain the cremated remains of two small dogs in a pile at the bottom, covered by sherds of red pottery. Both shafts were in close proximity to the mausoleum of a Roman villa. It seems likely that other shafts may remain undiscovered in this area. The deposition of dogs in the first layer would again appear to be significant as noted above.
The concept of the deep shaft as a means of communicating with the gods of the underworld is widespread and ancient, it was known in the Mediterranean world to the Greeks and Romans. The device has certainly been used since the Bronze Age in Britain. Whereas there may be some doubt as to the purpose of some of these shafts discussed above, there can be no doubting the function of the shaft at Swanwick in Hampshire.
The evidence for Wilsford Shaft may appear ambiguous, but a similar shaft discovered in clay-digging near the river Hamble at Swanwick near Fareham in Hampshire, was clearly used for ritual purposes.
The Swanwick shaft is a circular shaft, 24ft deep, 14ft diameter at the funnel mouth but at a depth of 9ft becoming cylindrical and about 7ft across. At the top of this cylindrical part of the shaft a compact horizontal band of charcoal (oak, hazel and elder) ran across the filling. Above the band of charcoal were 20 clay loom-weights, typical of Deveral-Rimbury cultures of Bronze Age settlements in Susses and Wessex , of between 1,200 – 1,000 BC. Fragments of a saddle quern were found scattered through the fill. Below this, again the shaft narrowed slightly to a diameter of 4ft and contained a central wooden post 5ft high and 8in diameter, packed with estuarine clay.
A brown deposit on the walls of the lower part of the shaft and the post proved on analysis to have a composition that included phosphorous and nitrogen compounds, such as could most likely derive from dried blood or flesh, indicating the post was used for sacrifice. Swanwick can safely be considered a ritual shaft and clearly quite different from a well containing votive offerings.
A similar discovery was made at Holzhausen in Bavaria. Investigations in Germany of the quadrilateral earthwork enclosures known as the Viereckschanzen type of late La Tene period within the last century BC, concentrated in the Upper Danube and northwards toward the Main. These seems to be ritual sanctuary enclosures with smaller square ditches around graves or Romano-Celtic temples.
The enclosure at Holzhausen had three distinct phases with a ritual pit or shaft as a feature of both the first and last. The shaft of Phase III lay in the northern corner of the earthwork and was cylindrical, 20ft deep and 6ft diameter, with a slightly funnel shaped mouth. In the centre of the lower part of its filling a wooden post 4in diameter and 6ft 6in long had been carefully set up and packed into position. Around the upper part were lumps of brown clay found to contain high concentrations of albuminous and nitrogen compounds, and traces of animal fats, which have been interpreted as the result of the decomposition of blood, flesh and viscera.
The similarities in size, proportion and details between Swanwick and Holzhausen is quite spectacular. Other Late Bronze Age ritual enclosures similar to Swanwick and Holzhausen with shafts containing posts at the base have been found in Southern Germany and one in Vledder in the Netherlands. 
Above we have seen examples to illustrate a complex picture of what may be defined as a true “ritual shaft“. There are dry shafts that seem to have been used purely for ritual purposes and wells that have received ritual offerings at certain periods in their use. Yet the waters become quite muddied when the function of the shaft may have changed during its serviceable life. At Wilsford the deposit has been interpreted as an accidental accumulation. However, we should not dismiss out of hand the possibility that the shaft was originally dug to provide access to the underworld realm of the gods, possibly in the Neolithic period, and abandoned as such when flooded and then later re-cut and then used as a well during the Bronze Age. 
Prehistoric shafts terminating in wells have been found in England, their intended original purpose unclear. One such shaft found in Dunstable, Bedfordshire, was found packed full with coins, pottery, Roman tiles, sandstone slabs, charred wood, animal and human bones, to a depth of 116ft. A human skull was lodged in a niche in the shaft when it was first dug. Another 37ft deep shaft at Biddenham contained part of an altar slab, a mutilated statue, fragments from about fifty urns, bones of horse, hogs, ox, fox and dog, and a human skeleton. In Warwickshire a 4ft square shaft (?) was discovered in 1689. At a depth of 20ft a large square stone, with a hole in the middle, was found supporting several grey-ware urns. Twelve were removed whole while another twelve had been broken apparently by the fall of a stone from above. The shaft was sounded to a depth of 40ft but the bottom was not reached. 
At Ashill in Norfolk, a 40ft deep shaft constructed with wooden square sides contained pottery, bones including those of a toad and a wooden bucket amongst the fill of the upper 19ft. Below this level the nature of the fill changed to layers of perfect urns embedded in hazel leaves and hazel nuts. A layer of flints covered the bottom of the shaft. The hazel tree’s association with sacred wells in Celtic literature requires no further explanation here.
At Fir Tree Field, Down Farm, Dorset another prehistoric shaft, complete with weathering cone and containing artefacts, was found but unlike the man-made Wilsford shaft this was a natural solution hollow utilised for ritual purposes. A pond barrow had been constructed close to the shaft at Fir Tree Field, perhaps making some reference to it and certainly suggesting a link. The close proximity of the Fir Tree Field shaft to the Dorset Cursus and numerous sites of Bronze Age activity suggest this natural shaft was a significant feature in the prehistoric landscape. As with the Wilsford shaft, two alternative proposals have been made for the function of the shaft. Firstly, owing to the presence of human and animal bone and other artefacts such as a shale ring, amber beads, and bone pins, in the fill, that it was primarily ritual in nature. Secondly, the environmental evidence suggests that it was used as a well. Perhaps both explanations are correct?
The environmental evidence from the fill of the Wilsford shaft indicates it was surrounded by grassland, indicating that the shaft, perhaps originally dug in the Neolithic period, was in use around 1,450 BC as a very deep well used by the local community for obtaining water for livestock and other purposes. The finds including grass cropped by sheep, sheep dung and two aborted lambs , together with the absence of typical deposit material, when compared to Iron Age and Roman period shafts in England and mainland Europe that could be construed as offerings, tends to support this interpretation. As we have seen above sheep rarely feature in ritual deposition
The rare find of prehistoric rope, and a wooden container dated to 3,650-3,100 BC from the bottom of the Wilsford shaft, suggest this was an ancient well, that may have been first cut in the Middle Neolithic period and was merely re-cut, to a wider diameter, in the 2nd millennium BC. Alternatively, the wooden bucket may already have been ancient when deposited at a later date, perhaps the Early Bronze Age. Yet significantly amongst the other finds from the bottom of the shaft were a flint scraper, a selection of flint flakes and blades, amber beads, a notched bone and ring-headed bone pins; refuse or typical of deposits found at consecrated Neolithic sites such as monument ditches and ritual pits and at Fir Tree Field?
After all, is a shaft no more than a deep, narrow pit?
1. Paul Ashbee, M.G. Bell, E.V. Proudfoot, Wilsford Shaft: Excavations, 1960-62, English Heritage Archaeological Report, 1989.
2. Paul Linford, Report on geophysical survey of three pond barrows, March 1996.Lake Down, Near Wilsford, Wiltshire.
3. Ronald Hutton, The Pagan Religions of the Ancient British Isles, Blackwell, 1993, p.192.
4. Ralph Merrifield, The Archaeology of Ritual and Magic, BCA, 1987, pp.46-47.
5. Black Dogs, Guardians of the Corpse Ways, Bob Trubshaw, Mercian Mysteries. No.20, 1994.
6. Anne Ross, Pagan Celtic Britain, Chicago Academy, 1967.
7. Hutton, op.cit. p.186.
8. Merrifield, op.cit.
9. Hutton, op.cit.p.192.
10. Merrifield, op.cit.
11. Stuart Piggott, The Antiquaries Journal, September 1963 43 : pp 286-287.
12. Ashbee & Proudfoot, op.cit.
13. Ross, op.cit. pp.53-54
* * *