Academic Papers

These Papers or essays may be corrected or updated following ongoing research. They are intended as a means to maintain a flow of information in advance of the longer term publication of Archaeological Reports for the Littledean Site. Source Papers through links


Occasional Paper 1.

Occasional Paper 2. The Incidence of Oyster Shell in the Springhead Pool

Occasional Paper 3.

Occasional Paper 4.

Occasional Paper 5. An Introduction to the miniature rock art of Littledean

Occasional Paper 6. Interim report of finds from water associated contexts Littledean Water Shrine



Occasional Papers No.6

(c) 31st May 2019 D.M.Macer-Wright

NB. 21st November – Elements of this report have moved forward as a result of discovering the Water Channel in July 2019.

A remarkable number of significant finds have now accumulated from this site from secure period contexts which indicate the spring head area became a focus for both ritual and iron metal working activity. This report catalogues these finds and concludes that there is now sufficient evidence to probably reach a reasoned conclusion that this site has seen the deposition of man made artefacts since the Mesolithic period. The Neolithic evidence is particularly important while an Iron Age metal working period embraced significant ritual activity. Lying between the ritualistic functions of the Neolithic/Bronze Age and the establishment of a Romano British water shrine, the Iron Age metal working period has important sealed artefacts which, given a growing consensus of understanding concerning Celtic practices, argue convincingly for associated ritual activity. A possible parallel conclusion is that the late Neolithic/early Bronze Age function of the site, likely based around seasonal forecasting for agricultural purposes, whilst essentially routine, may have embraced daily or seasonal ritual associated with recording the solar and lunar movements. If these connections can be made, one can conclude that Littledean provides a compelling case for a ritual continuum since the beginnings of agricultural activity until some point in the centuries following withdrawal of Roman administration.


Six distinct small pools, interlinked through a small water course, have been identified within the water shrine focal area, which also include an outlying one identified within the system of the Iron Age boundary bank 12 metres to the west. For ease of description I include all of these under the terminology ‘Pool’. They are all now dry, excepting Pool 1 (P1) which, apart from hot/dry summers, retains a small volume of water for much of the year. Although this term suggests a collective unity of function, this is not the case and the functions of the pools vary in time and space. Although a collective cohesion does appear to exist, establishing whether the purpose of the pools was geared towards either manufacturing or ritual at any one time, is not apparent in the evidence. The boundaries between the concepts of profane and sacred, in as much as we understand those concepts today, appear highly blurred.

It is clear from the contexts, layers and artefacts that the natural spring fed water features were manipulated from some point in the Neolithic period until final abandonment in the later Roman period, probably when Christian influences brought an end to pagan water worship.

The collection of finds are remarkable and excepting one unique find, a possible figurine, can be identified by their typology as falling into discrete groups of artefacts. All these artefacts, excepting the unique group from Furnace 6 (Fe6) have one thing in common which is that they have been ‘placed’ in a watery context. Given the small size of the pools, under 3 metres in diameter, it is clear none of the artefacts have been accidentally lost. I have not included flint artefacts, which are sufficiently numerous and demand their own report. Likewise the iron artefacts. The three reports should therefore be read collectively along with the report for the pebble lithics. However the flint material is particularly significant and is covered in the appropriate contexts within the DISCUSSION.

The archetypal artefacts are stone heads/facial images of which there are 18 including an outlier from the Dean Hall water garden. The exotic artefacts are polished axes and associated possible ceremonial objects. Of these there are 15. The more mundane and easily overlooked artefacts are quern stones, both grinding platforms (saddle querns) and hand grinders of which combined there are >50. A small group of 5 artefacts with similar morphology are identified as loom weights. Finally this report covers one possibly significant item which may be a ‘mother’ figurine.

This report records and describes 143 prehistoric artefacts which, excepting 6, all came from water contexts. Removing the 54 pebbles, which as a collective group are assigned their own report, along with the flint lithics and metal work. Thus 83 artefacts have been recovered from the pools and water feed, all of which make up the focal area of the Littledean Water Shrine. No other prehistoric artefacts have been found anywhere else from the excavated areas.

The total area excavated since 1984 is circa 1145 sq metres. The 65 artefacts, excepting the polished pendant axe, two heads and two possible heads, came from a total area of circa 53 sq metres of water features and their immediate environs. Expressed as a percentage, all the prehistoric artefacts so far recovered have come from 4.6% of the total excavated area. Given the proviso that prehistoric layers yet to be excavated over the remaining 95.4% area, where potential features have been identified, may yield further artefacts, these figures are strikingly loaded. It is fundamentally clear that deposition of artefacts in the water contexts of the Littledean Water Shrine was a matter of choice and highly suggestive of custom.


Two heads and two face masks came from the spring head pool (P1) and a possible head fragment from the stone and gravel packed surface surrounding the west end of the pool . One head from pool 4a (P4a) and 5 face masks from pool 5 (P5).

The face masks from P5 are small votives measuring between three and five centimetres. The smallest, on a round disc of yellow/green sandstone has a distinct face with lentoid eyes, whilst others all on old red sandstone have facial features defined by ‘pin hole’ eyes. Several of these masks could be seen as anthropomorphic with bird like features. The largest however is a blue slate polished disk with an incised celtic face, with lentoid eyes, wedge nose and double lines for the mouth. As a group they are varied and two pieces with incised lines are presumably fragments of larger masks. From the same pool a broken neolithic polished axe of possible Langstone origin was embedded in the bank slope. A neolithic flint auger/scraper was recovered from the eastern bank top. Whereas the flint may be residual to the IA surface the broken axe would seem to be a deliberate re deposited item. A ‘plate’ of gypsum was also found associated with P5 which has faint outlines, possibly of a face and another rounded piece of gypsum or kaolin material also has faint outlines. Given the soft easily washed away texture it is doubtful that enough survives to reach a certain identification or conclusion, whether these two items have natural or contrived facial features..

The head from P4a was recovered from the same context as the ‘figurine’ (see below) and three other small pieces of sandstone, all of which are notable for rounded and smoothed surfaces. The head was gently washed of loose dirt and sediment in pool P1 at which point no facial details were seen. Following subsequent re assessment and study of this group it became clear that all the five pieces were encrusted with a fine matrix of sand deriving from old red sandstone erosion. It seems clear therefore that pool 4a which lies ‘in the line’ of the supplementary feed had significant flows of water capable of carrying fine sand in suspension, which was deposited on the pool’s bed and banks. The head from P4a is 7 centimetres x 3 centimetres with damage to the left side of the face, but leaving the eye recognisable. The right eye is intact and the nose substantially so. The mouth is crudely scored and may have received abrasive damage. Traces of red ochre are present below the nose point and in the eye crevices, with traces of colour elsewhere suggesting the whole frontal area was coloured with red ochre. The rear is unworked. Remarkably there is evidence for a flattened base and the piece, when placed on a flat surface, stands vertically and is stable. The eyes and the general visage is very similar to the head from P1.

The head from P1 is extraordinary. It’s discovery was equally extraordinary. For almost 35 years it has lain encased in clay in one of the clay sample bags at Manchester University, recorded as “clay samples from pool section (4)”. These samples are in a varied state of preservation, some having crumbled and broken down and others intact. Four samples in the bag in question were essentially intact but with some breakdown of the uncut faces of clay ‘blocks’. As Fig 1. illustrates, the clay encasing the head when prised away started to reveal facial features. After diluting the clay under a slow running tap the whole head was revealed and was striking for the powerful red colouration across parts of the face.

FIG. 1 The process of discovering and revealing an unique red ochre stained head in a clay sample from the sacred pool of the Littledean Water Shrine metal working phase Iron Age

A. stone head encased in clay
B. stone head partly revealed
C. stone head revealed with ochre stain

The head is worked into a yellow/green local sandstone. It has lentoid eyes, a pronounced long wedge shaped nose, small mouth with a pronounced lower lip and a small rounded chin. Across the lips , the cheeks and bridge of the nose and also on the forehead, but excepting the eyes, is a strong colouring of red ochre. The head is 15cm high and 7cm broad. The clay context from which it comes is unequivocal and probably belongs to point X marked on the late Professor Jones’ sectional drawing, where he identifies the central pool. He equated this pool with the stone overflow drain under the caveat that the pool belonged to an earlier Romano Celtic phase.

The possible head fragment from the outlying stone and gravel surface of P1 appears to be the nose from a substantive head, significantly larger than the head just described. It also has traces of red ochre in the indented line below the nose and in a groove, which is probably damage to the right hand side of the face. This leads me to suspect that although there appear to be traces of ochre, there are also traces of clay. The context from which the head was found had no clay at all. This suggests therefore that this broken piece may have been redeposited and packed into the outer surface of the pool bank, having previously been in the clay pool deposits. Accordingly any clay adhering to the fragment had either washed away or more likely was leached away from the many actions taking place in the soil of the outer surface, to leave just small traces in the crevices and more persistent traces of ochre elsewhere. It is therefore the ochre that leads one to conclude this is a fragment of a stone head, directly comparable with the other ochre stained heads.

A third ochre stained head came from the possible neolithic stone hole SH3 immediately in association with a small slag piece distinctly shaped like an head or skull. The slag shape is unique in the collective experience of slag waste from the Littledean site as a whole. It would seem that it was deposited as a reference to its coincidental shape to keep the manufactured head ‘company’. The third ochre head is similar in overall shape and style to the others except for the eyes which are round and sunken with pin holes for the pupils.

A fourth ochre head is totally different. It is inscribed and etched onto a piece of ochre stone and was found in the sandy silts of P1 in association with two further red ochre stones, both with evidence of rubbing to obtain ochre powder. This head is remarkable for its detail. There are traces of pecking to represent an hair line and a moustache on the upper lip, while both sides of the mouth have long flowing moustache features. The left side is square edged with a distinct and pronounced ear feature, whilst the right side is lost.

The two O.R.S face masks from P1 are particularly interesting as they appear to have been re worked on earlier stone artefacts. One of these stone masks can be identified as both a hand quern grinder/crushing stone and a ritual object. It is a slightly oval disc in shape 11cm x 11cm x 3cm thick. Both flat sides show signs of rubbing wear, with one side having a distinct polish from fine rubbing wear. The 3cm edge is finely pecked all around possibly from an impact crushing action. Two circular ‘pin hole’ eyes are cut into the side with the polished wear and two out curving scored lines above to the rim, possibly representing horns. This mask was sealed in the clay deposit of P1 amongst the central deposits.

The other mask is a grainy old red sandstone ovoid disc circa 12cm in diameter. It is a unique shape to the site. Although the evidence for use as a quern rubber stone has been camouflaged by the facial features, it looks reasonably certain that it was so used.. It can best be described as a discus shaped disc having a double bevelled rim. All surfaces, except a discrete area more smoothed than pecked, show small pecking marks, suggesting it was used for both crushing and grinding. Incised in the slightly flatter side are two close set ‘pin hole’ eyes 1cm in diameter and 2cm apart, below two incised lines 5cm long and 1.5cm apart forming a symbolic mouth. Above the eyes and centrally placed is a small ‘pinhole’ with two smaller ones above left and two above right, with associated incisions possibly representing horns. However as Bryn Walters observes if this is reversed the mouth lines may represent a headband and the central small hole a mouth. On the alternate side there is also an inscribed face with two close set ‘pin hole’ eyes 2cm apart and immediately below a slit mouth >2cm wide incised by joining small pin holes. Above the eyes are two vertical lines also incised by connecting ‘pin holes’. This mask was redeposited in the top fill of the spring head pool P1. These two masks are notable for their seemingly crude and minimal design, neither having any representation of a nose and one wonders if they belong to a period anterior to the Iron Age.

Further to the above, two esoteric heads/faces may be seen in two pebbles. The first is a flint pebble circa 6cm high x 4.5cm broad x 2.75cm thick with natural ‘dimples’ which can be seen to represent eyes, nose and mouth. This flint pebble came from P1. The second is 3.25cm high x 2cm broad x 1.5cm thick with two small natural dimples and smaller central vertically linear mark. The back of the pebble is flat. The image and dome of the ‘head’ looks remarkably like an owl looking left.


FIG. 2A Moulded clay head base of post hole PH10

At the time of writing this report a significant post hole, PH10, the last post hole surrounding P1, has yielded the decayed image of a head at its base Fig 2. The sub rectangular post hole, with vertical chocking stone in its north face and stone/clay walls, had a soft stone carved outline at its base. Light trowelling and brushing caused minimal damage. However following detailed drawing, an attempt to lift the ‘head’ failed and it disintegrated into pieces. It had been placed in a slight depression central to the base of the post hole. The base of the hole is small flat stones set in clay. The soil fill of the post hole was entirely ORS red/brown soil with small stone pieces. The head image was comprised of a matrix of clay/sand/stone which had attached to the base thin pieces of sandstone. This is the only example of a shaped moulded matrix from the site and it seems reasonable to consider it was a head. Flotation and analysis by Robin Holley produced remarkable results. The matrix was clay, sand, small stone and oyster shell fragments and dust. The shell would seem to have undoubtedly acted as a setting agent through its calcium content. Sealed within the homogenous matrix were 6 salmonid fish bones, possibly salmon, a small polished sandstone ‘fish head’ with inscribed circular eye and a small sherd of arretine pottery (yet to be confirmed) probably in the range 30 BC to 40 AD.

This find is extraordinary as the post hole appears to belong to the portal of an open air proto water shrine,being the first phase of water shrine development on the site and dated within a range either side of the Roman Conquest.


A small perforated polished axe (see SHOP page for illustration) was found in post pit 1 (PP1), which pit was backfilled with larger and smaller pieces of old red sandstone and covered by IA metal working, layer 5a. The axe was positioned head down in the natural red clay, probably pushed in until the tail of the axe was just covered and buried under a largish piece of old red sandstone with significant grinding wear on the face lying above the axe. It has a hole drilled near the tail end by which to suspend it. The blade is perfect and remarkably sharp. The stone may be Cornish greenstone with striations of yellow colour. There is a small area of corroded damage to the stone near the drilled hole. It is highly polished.

fig. 2 polished stones deposit at base of stone hole 1st showing
fig. 2 polished neolithic stones base of stone hole 1

On 24th September 2018 I made a remarkable discovery of polished stone artefacts at the base of the stone/post pit (SH1) which I had excavated in July 2018 and identified as a possible placement hole for a standing stone. Excavation revealed four layers of structured deposits, which were exclusively IA quern stone material and possible loom weights. The excavation was ceased at the point where the natural clay was too hard to trowel further. By mid September rain had followed the extremely hot summer and the base of the pit had become soft enough to trowel. Having placed a ranging pole in the Stone Hole to take an alignment on the equinox I hit a ‘ringing’ stone and later that morning established that there were coloured/polished stones below the clay at the base of the pit. Following a phone call to Bryn Walters for advice I photographed and recorded the in situ stones and covered them back over until Bryn could establish a strategy to take forward the lifting of the stones. On 23rd October Bryn Walters, Luigi Thompson, Robin Holley and Anthony Beeson came to the site and witnessed the excavation and lifting of these stones.

In order of removal the 1st was a polished blue circular domed ceremonial mace, 2nd a large yellow/white quartz pebble, 3rd old red sandstone with significant quern wear, 4th a polished blue stone hand axe, 5th a polished brown stone axe, 6th an axe blank/ hand roller quern, 7th an axe blank/hand roller quern. These artefacts had clearly been laid into the bottom of the Stone Hole as a ritual deposit. However Robin Holley considered it unlikely to be a deposit contemporary with the positioning of a neolithic standing stone and advised excavating below the deposit of stones. This proved to be correct and below the stones were the pad stones for a standing stone. Another polished split blue stone, probably a quern stone grinder was recovered from the side of the Stone Hole and also a broken polished stone grinder.

Two further exotic stone finds came from Pool 4a (P4a). The first a polished ‘hammer’ stone with significant percussion chips to one end and the second a broken stone circular mace head. This artefact is less polished and may be granite.

Robin Holley considered it likely there could be more stones and we decided to protect SH1 pending further investigation in 2019. On 13th May Robin targetted further discrete excavation to establish the the original construction evidence for the stone hole and the construction of the contemporary pool. This proved to be highly successful, rewarding and vitally important work. It also produced two further polished axe blanks and a further stone mace shaped from a block of Old Red sandstone. This extraordinary artefact has a fish image pecked on one side with sufficient evidence to conclude it represents a salmon.


There were three ORS bun/disc shaped hand quern rubber/grinder stones redeposited in the pre nymphaeum levelling of Pool 1 (P1) layer 4a and one from layer 5b, the first of 5 structured layers in the fill of the Stone Hole (SH1). The two others have already been described under HEADS. All fall within the range 10-12 cm in diameter and 4-5cm thick except one disc shaped example 3cm thick with rubbing wear on both sides. The domed top of the querns and rims of all, show pecking marks possibly from percussion/crushing of iron ore. Many smaller hand quern rubber/grinders were found in the three layers of ‘structured deposit’ excavated from SH1 of various shapes and wear patterns, from curved or rounded to flat surfaces. Notably an ORS hand quern rubber/grinder stone was amongst and in the same context of the polished stone hoard layer 5g. This example is diagnostically different in shape and size to the bun/disc shaped examples. Also there were two other flattish pieces of ORS quern material in layer 5g. All these types, of hand quern rubber/grinder stones, were only found in or immediately associated with SH1 excepting two larger curved hand quern rubbing stones from the fill of Pool 2b (P2b). 8 random sized hand quern rubber/grinder stones were recovered from SH1, all except one in a soft grainy sandstone are ORS.

Broken flat quern pieces were deposited throughout the four structured layers of SH1. Broken saddle querns were used, with larger random small boulder shaped stones, to fill Pool 2a (P2a) as a single deposit. P2a was notable for its broken stone fill of which circa 50% was of stone pieces with rubbing/grinding wear and numerous pieces with distinctive dishing features consistent with saddle querns.. The broken pieces from SH1 were generally lacking in diagnostic shape excepting 5 pieces with distinct curvature throughout the grain strata of the ORS. The remainder were flat pieces heavily smoothed on one side only. Two pieces may be from broken saddle querns. A number of these pieces have iron staining and/or traces of iron/slag residue. Conversely the broken querns from P2a are more convincingly saddle quern, much larger pieces and with greater evidence of curvature and none of the pieces have any evidence of iron staining. One saddle quern, although showing a break across its width, is otherwise virtually intact and potentially important as an unused/unground-in quern, following manufacture. 71 pieces of quern material were recorded from SH1 excluding small pieces unaccounted for. 52 pieces of Old Red Sandstones from P2a include 22 pieces of distinctive quern material of which 8 are distinctive broken saddle querns. No obvious hand quern rubbing/grinding stones came from P2a.

The best preserved and most complete saddle quern, with extensive wear and dishing on both sides, indicating it had been long used, was recovered from the clay deposits of P1. A further extraordinary quern has been identified which may have measured 90cm x 60cm and is broken across its middle. This example has a deeply dished grinding surface and may be identified as lying between the typology of the ‘trough quern’ and the ‘saddle quern’.

Two broken quern stones of particular significance are possible proto rotary querns. One broken piece from the stone/gravel surface surrounding P1 has wear and distinctive evidence for a ‘hopper’ hole. The other from P2a is important as it has a break across the potential ‘hopper’ hole and a grooved grinding surface which has been partly worn down. There is also evidence for another hole, if not two, which may have taken a handle to turn it around a central pivot into the base quern. To match these examples is a possible proto rotary quern base, probably significantly reduced in size, which has a central placement hole for the vertical rod to sit in, around which turned the rotating top segment. All three pieces are somewhat removed from the recognised and accepted rotary querns of the Celtic record. The base piece was sealed in the clay in the central cross section of the pool clay sediments, amongst other larger pieces of stone Fig 3.

Of particular significance are several large reused stones/boulders laid in the foundation of the cult statue podium in the apse which have reasonably certain wear from quern use. These include one very large conglomerate boulder which has a flat surface where the quartz stones have been worn flat. Of further significance is another huge conglomerate stone built in to the foundation for the apse. This stone bears the tell tale grooves created by sharpening polished stone axes and is a remarkable testament to the Neoloithic origins of the site. Further stones bear distinctive marks which may also be of the Neolithic age. One further large sandstone has features which may possibly be axe casting moulds.


Two mace heads have already been mentioned. The one from the SH1 hoard and the other broken half from P4a.

A large ORS mace broken across the shaft hole is of primitive form. The half is a wedge shaped point 12cm long from edge of the percussion point to the break x 11cm broad x 6cm wide x 7-8cm deep. It was recovered from deep within the clay deposits of P1. The shaft hole may have been drilled from both sides given its irregularity. The proof of drilling is evident from the circular ribs in the shaft hole section. If the other half reciprocated the shape it would have been very heavy and several kilos in weight.

A probable mace head was also recovered from the bank top in the stone/gravel surface at the west end of P1, which was also broken at the shaft hole. Although the shaft hole is not so distinct as the previous example enough of a smooth core survives to suggest it was drilled right through. The half is 9-10cm long from the break to the percussion point x 6cm broad and 7cm deep. The shape is similarly wedge shaped but deeper than broad thus reversing the form of the previous described example. This example is of particular interest as a natural water pitted sandstone has been utilised to make what must have been a rather unusual and decorative artefact. There is a similarity for a decorative finish, as seen in the remarkable example of the intricately carved white flint mace head from Maesmor, North Wales. Very near to this find spot a fragment of the same stone was found. The two pieces do not join and if they do belong to each other, there must be other pieces, suggesting that it may have been deposited in a broken state and then became broken further from treading.


These are a small group of ORS stones with similar diagnostic features of single opposing worn grooves. They are peculiar however as some have graffiti incisions, some do not . Thus it is the grooves alone which the wear suggests were created for hanging from threads of some sort. The most likely explanation would seem therefore that they were loom weights.

Three of these objects came from SH1, one in each layer 5b, 5c and 5d. The first from layer 5b has worn surfaces on both sides and odd incisions and measures 12cm x 8cm x 2cm and 1.5 cm wide grooves half way along the long axis and is shaped a little like an axe. The second from layer 5c is 13.5cm x 7cm x 2cm with 3cm wide grooves. One side is relatively flat and worn and the other side is coarse and not flat. This piece has a distinctive humanoid shape with torso, deep and widely grooved waist, with the groove running around the flat side, so that both side grooves are joined and wide hips, from which the lower half runs down to a point. A deep incised groove runs vertically from the tip of the point to below the grooved area, ending in a ‘spear like’ head. The reverse side also has an incision, less pronounced ending in a pecked hole, which may be damage and not contemporary. The likeness to a figurine is striking. The third piece from the lowest level is the most like a weight. It measures 10cm x 7cm x 3cm. The waisted area is circa 3cms below the top which was probably square before a piece broke away. The bottom is slightly rounded. The grooves are circa 2cm wide and pronounced, whilst in contradistinction to the other two pieces from SH1 the rims/sides are worn and rounded.. Both faces have wear and random linear incisions. The piece has traces of iron staining. Whereas these three pieces are essentially flat and two dimensional the two pieces of ORS from the water feed to P4a are altogether chunkier and random in shape.

The first of the pieces from the Water Feed WFP4a is circa 10cm x 5cm x 5cm with a well worn waist 2cm from one end circa 4cm and 3cm with a third groove 2cm. This piece is distinguished by an incised six pointed cross on the top end and evidence of further incised lines and wear along the sides. It has clearly been damaged and was possibly a four sided broadly rectangular piece of stone in its original form. The second is very random in shape and heavily incised over several surfaces, indicating its current shape is not due to damage. 12 cm long with three sides 5cm, 6cm and 7cm in width indicate this piece differs significantly from the others. The waist features are 2cm below the top and 3cm wide. One face has two incised lines crossing it, with evidence of further lines, all angled rather than at right angles to the long axis. The other face has a graffiti stylised image of possible eyes nose and mouth, intersecting at right angles and overlying another similar image of eyes and nose. Two faces have smoothed areas from wear.

A sixth ‘loom weight’ is described below as a figurine. It will be seen the diagnostic features are very similar to the object from SH1 layer 5c. It is dealt with separately as a ‘figurine’ due to its small size. However see DISCUSSION for the author’s conclusion on this object.


This ORS object is 7.5cm x 4cm x 1.5-2.5cm. The waist is 2.5cm from the top and circa 1.5cm broad. The bottom half is more or less slightly curved and then leading to a point. The back is flat and the front has a prominent chested torso with traces of breasts and cleavage. The right hand hip, frontal view is pronounced, the left is not, due possibly to wear.

An incised groove runs from the point vertically to a small hole below the waist. A similar and opposing incision is found on the rear with clear evidence of a smooth abrasion The top of the stone is flat and possibly a piece has broken away, leaving a slight trace of a smoothed area. Most striking is this object’s small size and water rounded erosion. It came from the east bank of the pool 4a, near the bank top and with several other water worn old red sandstone pebbles


Flint and quartz pebbles were recovered from the following pools. P1 (25), P2 (7), P2a (8), P2b (5), P4 (2) and P5 (3). None were found in P3. The greatest concentration therefore was from P1, with a significant cluster covering the north bank and slope of the pool. These included both flint and quartz pebbles. The quartz pebbles from P2 and P2a were generally white and fresh looking. Two pieces from P2b were cut quartz rather than pebbles. The few from P5 were all yellow with iron staining. Apart from one yellow stained pebble 6cm x 6cm from Pool 2b, one 7cm x 6cm from pool 2 and one 7.5cm x 7cm from Pool 2b which although slightly tinted yellow, is a very clean unabraided white quartz pebble which has a flat, possibly ground back, most pebbles, both flint and quartz, are in the 3 – 5cm range. It is notable that by far the majority are from Pool 1 and also that the pebbles from Pool 1 are more consistent in size. It is also notable that in both Pool2 and 2b all the pebbles were found in the depressions of each pool which probably signify spring source points. Of a total of 54 pebbles only 4 came from a dry context, and those few were from the east side of the pool.


Five fossils were recovered from the clay sediment and redeposited layers of P1. These comprise two pieces of stigmaria fossil, tree roots from the Lepidodendron tree of the Carboniferous period, and a third piece and two ‘devil’s toenails’ Gryphaea.


The first evidence for structures associated with water at Littledean is a possible henge type monument encircling the Late Neolithic/Early Bronze Age pool P1. It is at this point in time that the contemporary polished artefacts were first deposited. However the Mesolithic flint material makes it clear that flints (microliths) were being deposited in the wet area of the spring head, which was the fore runner of P1. The presence of Neolithic and Bronze Age flint scrapers in the contexts of P1 demonstrate the practice of depositing flints in water had a long ancestry.

The sandstone fragments of saddle querns may represent a long tradition of use from Neolithic times right through to the Iron Age. The near complete example from the clay deposits of Pool 1 is an Iron Age deposit, however an expert opinion based on its wear pattern is required. No examples of Romano British rotary querns have been found from the site and yet the two fragments of proto rotary quern, if indeed this is what they are, suggest the quern was constantly used for grinding grains by the local community well into the Iron Age.

The incidence of sandstone hand grinder/rubber querns and flat fragments of sandstone quern found in the structured deposits of Stone Hole 1 and elsewhere is identified as being consistent with metal working use, e.g., sand grinding for clay paste. A number of these stones have iron staining deposits, which are not evident on the saddle quern material. However analysis is required to establish whether this is natural from lying in wet iron rich conditions or from contact with iron ore/metal working activity.

The archaeological contexts demonstrate that any henge type monument, as indicated by the evidence of stone holes, was removed during the Iron Age and probably, given the excellent slag pieces in the fill of SH3, after metal working had commenced. The structured deposits of SH1 demonstrate that the polished artefacts were re deposited following removal of the standing stone and four layers of sandstone quern artefacts were deposited in soil fill probably over a protracted period of time. SH1 was finally capped with path material to the proto water shrine, thus demonstrating a continuum since the removal of the stone. SH3 is a particularly important stone hole as the deposits include 3 pieces of significant slag, including the slag ‘skull’, which lay immediately adjacent, in the soil fill, to the third of the three ochre stained Celtic heads found from the water shrine.

The red ochre coloured heads are of the first importance. The association of one of these with the slag “skull” demonstrates that the deposition of these heads was occurring during the metal working phase. The heads are fundamentally ritual objects and provide remarkable evidence that metal working and ritual did go side by side. This confirms Dr Anne Ross’ views that smithing was a sacred activity likely under the protection of the smith god Goibhnu. For period dating they are vital. The head from SH3 is clearly broadly contemporary with the head from the 1985 clay samples of the pool 1 section. They typologically belonging to the Celtic Iron Age.

The artefacts from the supplementary feed are particularly enigmatic. However the sandstone pieces from P4a identified as possible loom weights are directly comparable with the pieces in the structured deposits of SH1 and therefore, if not redeposited, are probably Iron Age. The broken mace and polished ‘hammer’ stone from P4a are more problematic as they would seem more likely to be Neolithic.

A 2 – 4cm layer of black compacted silt at the bottom of P1, sealed under the clay and sand sediments which filled the pool, is identified as all that was left of the Late Neolithic/ Bronze Age silts following dredging in the Iron Age when a functional change took place for metal working. This is established by the fact that the flints were only found in association with pool bank deposits and Romano British levelling over the pool in advance of building the Nymphaeum. No flints were found in the clay deposits. The clay and fine sand deposit, built up during the metal working period, remains the most difficult causative effect to explain. If it occurred as a result of inflowing water high in suspended clay, then pools 4 and 4a and the feed should also be filled with it. They are not however. This leaves the only rational explanation that the clay deposit in the springhead pool was the result of the metal working processes themselves (see Occasional Paper 2). Given the group of water worn sandstones in pool 4a, which include the possible loom weights, possible figurine and the third ochre head, it would seem that during the Iron Age a significant flow of water passed through the pool at certain times of the year. A factor is the third ochre staine head from pool 4a, which demonstrates convincingly that all the artefacts were either deposited, or re deposited, in the Iron Age at a point in time when there was a shift to metal working.

The coincidence of artefacts and water at the Littledean site is compelling evidence for deliberate deposition. The recurrence of specific type groups of artefacts appears to be non coincidental and deliberate. Can these deposits be regarded as a choice driven exercise and the result of ritual purpose? To reach a positive answer it must be seen that the relationship between water and artefact is securely interpreted and not skewed by a failure to recognise group related deposits of similar type elsewhere on the Littledean Water Shrine site. This point is satisfied by the fact that roughly as little as 0.6% of the site has yielded the groups of artefacts described in this report. It currently seems very unlikely that, excepting other possible IA metal working areas, there are any other significant deposit areas or features. A series of possible pits to the south side of the water shrine, which may be ‘stone holes’, were identified in 1986 and are clearly seen on a SkyScan. These may have similar structured deposits, in which case it would be reasonable to conclude that the practice of filling pits with quern material was not specific to water contexts. Although such a case would not remove the concept of ritual motivation it would break any thesis at Littledean that water was the motivating force for deposition.

Whether or not the hoard of polished stone artefacts in the Stone Hole (SH1), were deposited as a ritual associated with the placement of a post or stone in the Late Neolithic or early Bronze Age periods, is impossible to say although it seems likely. This deposit has no immediate parallel in time or space. The unique nature of the hoard may be that it was ritually motivated by the positioning of a vital standing stone over the actual spring head water source, rather than a stone raised in the more studied arrangement relating to henges. The polished stone hoard at Littledean, as pointed out by Robin Holley, was probably a secondary deposit placed after a standing stone was removed, following which there were four further layers of structured quern deposits of probable Iron Age date which eventually filled SH1. Ultimately the filled in stone hole was covered by a pathway which terminated at the entrance to the first phase of the late Iron Age water shrine. This pathway also had deposited quern material in its surface. The recovery of two further polished stone artefacts from the west ‘wall’ of the earth/stone revetment for the standing stone may be confirmation that the original deposit had been around the standing stone, rather than under it. Robin Holley’s point was endorsed when the layer below the hoard revealed the in situ flat stone pad, upon which the standing stone would have stood. The relevant point is probably that both the standing stone and the artefacts were placed in a wet deposit, which today is still either saturated or ‘boggy’. It would seem likely therefore that the polished artefacts were come upon accidentally when circumstances demanded the removal of the standing stone, at which point in time the artefacts were laid at the bottom of the empty hole as a clear expression of ritual and thus started a period of further ritual deposits ending with the first structural phase of the water shrine. This activity is defined by the fact that the polished stone ‘hoard’ was placed in a position from which it was not recoverable, as distinct from a ‘treasure’ hoard hidden for future recovery.

The Mesolithic and Neolithic flints recovered from the pool and the north bank are particularly significant. These are not the product waste from knapping, as was first thought likely in 1984. If the early flint material was knapping waste there would be hundreds if not thousands of pieces rather than a total of approximately 150 of all periods. This point is endorsed by the fact that casual flint finds from the site in general, although relatively evenly spread are numerically low. It seems clear that during the Mesolithic and certainly during the Neolithic periods, flints were occasionally deposited in the wet areas around the spring water source. Following the creation of an artificial pool in the Late Neolithic or Early Bronze Age flint scrapers in particular were deposited in the pool. which in the Bronze Age included more valued items such as polished and finely finished round scrapers.

Whereas the layers of clay sediment in pool 1 may be explained, the thin layer of black silt whilst consistent with a build up of decayed organic material within pools whose only source is springs, is insufficient to account for a timespan of two millenia. Even in an hilltop location silt deposition is a rapid process turning springs and pools into boggy wetland, as any one familiar with such places will know. During the 1000 – 1500 years after the formation of the pool, until the onset of the Iron Age, a very significant depth of silt would have built up. On the north side of the pool the sloping pool bank was found to be exaggerated by a low mound of dark earth adjacent to the pool edge. This was topped by a layer of iron rich soil coloured orange with a mix of red clay. Roughly 75% of the multi period flint material including flint pebbles came from this bank of dark earth below the orange soil. The remainder of the flint material was re deposited over the pool centre with all the other types of artefact at some time before the Romano British redevelopment of the nymphaeum took place. Significantly no flints were recorded from the layers of clay silt build up.

It is vital therefore to comprehend that the build up of clay within the pool is a secondary feature and represents a new phase in the pool’s significance. The water rounded sandstone pieces from pool 4a provide solid evidence which attest significant periods of high flow. Where water flow slows as a result of impounding siltation will quickly take place. The thin black silt layer at the base of the pool combined with the lack of flints found in the later clay deposits argues convincingly that the pool was ‘dredged’ in advance of the metal working phase and the silt deposited on the north bank. The recorde covering of iron rich orange soil and clay over this bank deposit was most likely burnt clay or collapsed furnace lining, from the metal working phase. Unfortunately in 1985 no samples were taken for analysis.

The non occurrence of flints in the clay deposits may be profound while the group of sandstone finds from within the clay deposits may reveal evidence of a symbolic shift. The proto rotary quern, if that is a correct identification, is probably diagnostically Iron Age, the hand quern rubber/grinder is matched by the others from IA contexts and the broken sandstone mace has no direct parallel with the neolithic examples and its shape suggests it was an Iron Age weapon. The broken saddle quern is a well worn, double sided quern which could date any time from Neolithic to Iron Age. These artefacts were undoubtedly recovered during excavation because they could immediately be recognised as significant finds of some sort. Other stones, several of which can be recognised in the 1985 photographs can, given the recent recognition of quern stones, also be considered probable artefacts. (In 1985 not one of the stones from pool 2b was recognised as a quern, nor the rubbing wear recognised. They were all considered random sandstone fill.) The question, why did people throw into pool 1 large pieces of random stone with no symbolic value, has an obvious answer. They did not. The artefacts which were recovered demonstrate such things were deposited there for a serious purpose and, given the failure in 1985 to recognise in pool 2b the quern deposit, it is now reasonably certain that numerous other quern artefacts were not recognised. That these escaped the archaeological net is demonstrated by the photographic record.

When the supplementary feed was created is not clear. The polished stone hand hammer in pool 4a is probably neolithic, whereas although the broken mace may also be Neolithic it could equally be a broken Iron Age deposit. The water worn sandstone pieces also from P4a, if loom weights, are comparable to the similar objects from the structured deposits of SH1 and therefore probably Iron Age in date. At this point of the excavations it seems reasonable to assume the supplementary feed was made to supplement the spring head which was perhaps not flowing so strongly. However when this occurred seems currently impossible to grasp.

The overall evidence is so far compelling however and should allow for a few conclusions to be reached regarding life around the springhead at Littledean. During the hunter/gatherer Mesolithic period flints were deposited in the watery context of the springhead. Four find spots from outside the immediate confines of the springhead were clusters of 3 to 5 microliths, as distinct from the single flint find spots. These four find spots are unique at Littledean and highly significant as they may mark the spots where Mesolithic arrows were lost, having missed their prey. One can imagine the scrub covered environment within which the hunters concealed themselves downwind of animals coming to the spring to drink. Are the flints discarded in the spring offerings from a successful hunt? The nearest source for restocking flint weaponry was the southern chalklands or through trading with other groups. As a hunter one could replenish the arrow shafts with ease but not the microliths which gave them potency. One can further imagine those microliths carried in a bag to replenish lost arrows. Today any archer will know how staggeringly easy it is to lose an arrow, even in short grass! The microliths were vital to successful survival and one must assume the vital currency of life. They were certainly not objects to casually discard. Is the depositing of microliths in the spring head the genesis of water worship at Littledean? It is an attractive idea with an extraordinary resonance for modern day humans.

The continuation of this practice for depositing flints in the water is also seen through the Neolithic phases with the deposit of flint scrapers and flint pebbles. Notably finer artefacts are missing. There are no arrowheads for example or finely knapped flint tools. One could even suspect that the deposits became less fashionable, until a sea change took place on or after the cusp of the agricultural revolution.

The creation at Littledean of a man made pool associated with monoliths or posts, marks the turning point from Neolithic proto farming to a settled community based lifestyle, which is driven by the seasonal parameters of day length/light and heat. These ideas may have arrived suddenly with experienced farmers seeking new land or more slowly, evolving through local experimentation, possibly influenced by trading groups. Which ever process one accepts towards our modern interpretations of early agriculture one fact stands out. Success in growing food, whether grains or meat, depended on the ‘growing season’. Anthropologists are aware that mankind in Northern Europe became aware of this fact by some point in time following the end of the last Ice Age.

The relationship between rainfall/water and yield is too profound to rationalise quickly. The movement of the sun and the seasonal pendulum of increasing and decreasing day length can be observed on a daily basis from the Littledean Water shrine, as the sun moves between the midwinter and midsummer standstill points, tracking back and forth along the hilltop of the eastern Cotswold sky line. The western horizon provides the same daily sunset observations. The evidence is mounting at Littledean for deliberate alignments in the laying out of the pool and associated man made features, both in terms of solar and lunar alignments. It will be these alignments that will ultimately confirm a mind set which deliberated to produce a water related agricultural calender. The synthesis of season and water to achieve a spiritual oneness in daily life, which recognised the equation of light and dark, hot and cold, wet and dry may have been the most likely function of the late Neolithic/Early Bronze Age spring head pool. However one attempts to define the purpose of this pool it cannot be denied that the spiritual element involved a huge leap of faith. The inherent spirituality no doubt manifested as magic and superstition. Is it safe to conclude that this site at littledean was used to forecast when to sow? Quite likely it is, because one can throw away all modern time keeping mechanisms and successfully start again from the Littledean site. Everything ultimately depends on water and prehistoric man was no doubt better prepared for this fact of life than we are today. Maybe water was taken from the pool at certain times of the year to nurture crops and animals. Today much of modern society would struggle with this sacred concept. There is however no reason to doubt the minds and skills of a contemporary society which fully grasped the concept of sanctity in water. Although the rituals they pursued are lost to us, one thing is evident from the worn cobbled slope facing the pool’s outlet, people entered the pool. Whether this was to drink, bathe or collect water, this evidence provides a tangible link with our prehistoric predecessors of 4500 years ago.

The plan and appearance of the archaeological features of the original pool suggest it had two outfalls. However vertical photographs taken on 12th April 2019 suggest a different interpretation warning of drawing the wrong conclusions from visual impressions at ground level. On first inspection the late Neolithic/Early Bronze Age pool, when observed from the east end, does appear to have had two outfalls, largely on the basis that both are significantly aligned, the south west towards the major lunar standstill and the north west towards the summer solstice. The vertical photo Pic2 gives a different and significant change of perspective. It indicates that the south west outfall is in the centre of the west bank and the pool surround is more rectangular than circular. This would be consistent with the north west outflow being a secondary feature following the blocking up of the SW outfall, before the onset of metal working in the Iron Age. The metal working layer covered the backfill of the SW outfall. It is clearly very significant therefore that no clay and sand deposits are evident in the SW outfall.

The perspective obtained from the vertical photography allows a more comprehensive analysis for the evolution of the pool P1 and allows a more consistent interpretation. The combined evidence suggests the late Neolithic/Early Bronze Age pool had an outfall in the centre of its south west facing bank. An area of worn sandstone ‘cobble’ surface to the sloping bed of this south west facing bank is strictly contemporary with this outfall and the original pool. The wear to this cobble indicates the pool was accessed by either humans or animals or both. At some point in the Iron Age the outfall was blocked up and became covered by a layer including small slag pieces from metal working as well as structural elements of an associated wooden metal working structure. A new or modified outfall, most likely new, was created to discharge the water to the watercourse on the north west side. Whether this was strictly contemporary with a new supplementary water feed into the pool at the north east corner is uncertain.

The earliest quern type artefactual material found at Littledean is fully consistent with a neolithic/early bronze age agrarian society. The collecting of this material in broken form and deposition, particularly in pool 2b, albeit in a late IA/RB context, demonstrates the probability of deposited caches from the Bronze Age. Whether these had been deposited in the pre IA pool is impossible to say, although the survival of a particularly fine example, recovered from the IA clay silts, may be an token of the practice.

Before leaving the somewhat esoteric pre Iron Age history we should discuss the evidence, if any, for Bronze Age settlement in the Littledean district. Since Lidar, Historic England have been able to authoritatively confirm that Welshbbury IA hillfort a mile to the north of the Littledean Water Shrine was first established in the Bronze Age in association with field systems. Professor St Joseph the eminent geologist of Cambridge University obtained seminal aerial photographs of the Littledean region in the 1960s. These were provided for me in the 1970s and clearly show an extensive field system across Dean Hill (Littledean Water Shrine) and running under the Forest of Dean to the south towards Blakeney. The photographic aerial survey coincided with extensive forestry fellings which revealed the Bronze Age field banks lining up precisely with the Dean Hill system. Given the Forest boundaries to the south of Littledean are pre Domesday Survey, these fields are clearly ancient.

The Littledean landscape evidence is now becoming clearer. There was probably an extensive Bronze Age ‘estate’ covering the eastern old red sandstone ridge at Littledean, where the Littledean Water shrine is located, with a ‘fortified’ settlement site at Welshbury. Whether or not the ringwork known as Littledean Camp (“Roman Camp”) which is the site of the Norman “Old Castle of Dene”, situated roughly half way between the Littledean Water Shrine and Welshbury hillfort is part of this Bronze Age landscape remains uncertain. Although securely dated to the Norman period by the 1950s excavation and supported by both field names and description in the 17th century estate records, opinions have always been divided as to this earthwork’s original construction. It would be helpful, if not pivotal, to see Lidar for Dean Hill, which might resolve once and for all just how significant this hill is in the evolution of modern man.

A putative Bronze Age origin is supported by finds of axe heads and other Bronze Age material found in the vicinity of the Camp by the late Professor Major John Penberthy of Dean Hall prior to 1938. Formerly housed in a collection of the late Dr. Hubbard at Hill House, Pleasant Stile, these artefacts have unfortunately long been dispersed and disposed of. Other significant Bronze Age finds have been made in recent years in the Littledean region by metal detectors. There are no known Bronze Age axe heads from the Littledean Water Shrine site. Family anecdotes suggest however that both Captain Corbet Singleton of Dean Hall and his predecessor Penberthy used early metal detectors from the 1930s.

We leave the Neolithic/Bronze Age secure in the knowledge that the the spring head pool P1 sits in a Bronze Age landscape. Both Neolithic and Bronze Age flints deposited in the pool suggest an empathy inherent since Mesolithic times. The hoard of ‘polished stones’ is our first real evidence of ritual and the evidence from quern deposits, albeit re deposited at a later period, evidence of accumulation of a significant artefact type, for which there is a large corpus of academic study world wide indicating ritual hoarding. The evidence for both solar and lunar alignments requires extensive study and testing. Early signs indicate this may prove to be a rich field of enlightenment and central to studies of early agriculture. I believe we can enter the Iron Age at Littledean secure in the probability that ritual had become an established element attached to the function of the spring head pool and its immediate environs.

The typology of this class of artefact is determined by shape and wear, both separate parameters. Querns are typically recognised as either more primitive Saddle Querns of the Neolithic/Bronze Ages or Rotary Querns stemming from the later Bronze/Early Iron Age. Generally they are associated with grinding grain or other seeds and nuts. The querns from the Littledean site include grinders and grinding surfaces identified for uses associated with metal working and ore grinding.

I must make the caveat that there is little comparative information for ore grinding/crushing stones/querns. The Forest of Dean iron ore field lies half a mile to the west, across the valley of Dean and around the 800 ft contour of Littledean Hill where many surface scowles are visible from prehistoric mining. The ore present in the limestone measures is called ‘brush ore’ which is very high grade and existed anciently in almost pure form which crushed easily for direct injection into the furnace. Querns are recognised by iron staining and ridges adhering to hand quern grinders and quern ‘plate’ surfaces and extrapolated accordingly. Saddle quern pieces are identified by wear and

Fn1. The possibility that the clay deposits were the result of artificial activity, for example case hardening of steel blades during the metal working period, whilst an attractive solution, has no parallel in archaeology. Peter Crew at Plas Tan y Bwlch, who visited Littledean in the 1990s, the leading authority on early iron smithing technology, has the opinion that it is unlikely the clay would build up from this process. Given the evidence establishes beyond reasonable doubt that a secondary earth cut water feed was installed to the pool, siltation from that source is most likely and indeed inevitable. The perennial flow of water from other spring sources would be clear/translucent. However every time it rained run off would cause erosion in the artificial earth cut channel. This is the district norm arising from the geology and local soils, whereby water courses run red during rainfall.