Rediscovered Stonehenge Aberdeen Monolith Link

Abstract

In 2024 the central Altar Stone of Stonehenge (Stone 80) was shown to be Old Red Sandstone from the Orcadian Basin of north-east Scotland, carried at least 750 km to Salisbury Plain. That result has revived an old intuition: that Stonehenge’s great recumbent slab and the recumbent stone circles (RSCs) of Aberdeenshire belong to a single, far-flung megalithic conversation. This paper tests that intuition against the most detailed record we have of an RSC under excavation — V. Gordon Childe’s final report on Old Keig and the engineering study H. E. Kilbride-Jones published alongside it, both in the Proceedings of the Society of Antiquaries of Scotland for 1934. The reports document deliberate quarrying and shaping of the monoliths, a standardised, asymmetric profile with a pointed base, and a lever-and-fulcrum method of erection that Kilbride-Jones explicitly carried south to the Stonehenge trilithons. Set beside the modern provenance work and the reading of the Altar Stone as a deliberately laid recumbent — an interpretation that goes back to Gordon Freeman’s alignment observations and that I developed in 2015 — these papers supply the concrete archaeological substance for comparison. I argue that the parallels are real and worth taking seriously, but that they are most safely read as evidence of a shared insular repertoire of practice and symbolism rather than of direct cultural transmission in either direction. I also show where the 1934 chronology, on which Kilbride-Jones built an explicit sequence, has since been overturned.

Introduction

On 14 August 2024, Clarke and colleagues reported in Nature that the Altar Stone at the heart of Stonehenge — Stone 80, a six-tonne slab of Old Red Sandstone — is not Welsh, as had been assumed for a century, but came from the Orcadian Basin of north-east Scotland, at least 750 km away and most plausibly moved by sea (Clarke et al. 2024). The finding did two things at once. It severed the Altar Stone from the Welsh bluestones with which it had long been bracketed, and it placed its geological birthplace in the one corner of Britain that possesses a stone-circle tradition built around a great horizontal stone: the recumbent stone circles of Aberdeenshire and its neighbours.

That coincidence is the subject of this paper. It is not a new thought that Stonehenge’s recumbent Altar Stone and the Scottish recumbents might be related; what is new is that we now have a geological reason to ask the question seriously. I have my own stake in it. Building on Gordon Freeman’s observation that the Altar Stone’s long axis is skewed to a secondary solstitial line — from midwinter sunrise to midsummer sunset (Freeman 2009; 2012) — I argued in 2015 that the stone was never a fallen upright but was laid flat deliberately, as one element of a designed and skewed central setting (the “Twisted Trilithon”; Daw 2015; 2025). Parker Pearson and colleagues (2024) have since reached the same conclusion about deliberate placement and gone further, proposing that the Altar Stone’s position — recumbent, in the south-western arc of the monument, at the foot of the two tallest uprights of the Great Trilithon — echoes the plan of a recumbent stone circle.

Are the echoes based in the actual monoliths? How a recumbent stone circle was actually built: how its great stones were quarried, shaped, raised and bedded. For that there is no better source than the excavation of Old Keig, the circle with the largest recumbent of them all, dug by Childe in 1932 and 1933. His final report appeared in the Proceedings of the Society of Antiquaries of Scotland for 1934 and — uniquely among excavation reports of its kind — it was accompanied in the same volume by a dedicated engineering study by his site assistant, H. E. Kilbride-Jones, who had watched the Old Keig monoliths come out of the ground and who explicitly carried his conclusions south to the trilithons of Stonehenge (Childe 1934; Kilbride-Jones 1934). These two papers, ninety years old and written long before anyone could have guessed where the Altar Stone came from, turn out to be the natural starting point for the comparison.

I will argue that the parallels are real and worth taking seriously, but that the honest reading of them is more cautious than the headlines. The most defensible conclusion is that Stonehenge and the Aberdeenshire circles drew on a shared insular repertoire — of deliberately shaped monoliths, of a clever method of erection, and of the symbolism of a great recumbent stone — rather than that one region taught the other. I will also show where the 1934 chronology, on which Kilbride-Jones built an explicit sequence from Stonehenge to Scotland, has since been turned on its head.

The recumbent stone circles of north-east Scotland

Recumbent stone circles are a regional speciality. Something over seventy survive in Aberdeenshire and the adjoining counties, with none of the type found elsewhere in Britain (Burl 2000; Hill 2021). The defining feature is a single massive stone laid on its side — the recumbent — set between the two tallest uprights of the ring (the “flankers”) in the south to south-west arc, with the remaining stones graded down in height towards the opposite side. Most enclose a low ring cairn, and excavation repeatedly finds cremated bone, broken pottery and white quartz within them. The recumbent and its flankers frame a section of the southern sky, and since the work of Burl and others the circles have been read as devices for watching the moon — the recumbent acting as a level “sill” over which the low southern moon could be seen to roll at its standstills (Burl 2000). Richard Bradley’s excavations at Tomnaverie and elsewhere have nuanced this picture: he found that the circles were often raised around pre-existing ring cairns, and questioned whether any precise lunar sightline was ever intended, with a general orientation towards the southern sky and midwinter sunset doing much of the work (Bradley 2005). The lunar association remains the dominant interpretation, but it is now held more cautiously than it once was.

Crucially for what follows, these are monuments of the late Neolithic and the very start of the Bronze Age. Modern dating — anchored by Bradley’s fieldwork and by the broader chronology of the associated ring cairns and the earlier Clava cairns — places their construction around 2500 BC, in the last few centuries of the third millennium (Bradley 2005; Hill 2021). That is a point to hold on to, because it is not what Childe and Kilbride-Jones believed in 1934.

Old Keig itself is the giant of the class. Its recumbent is the heaviest known, on the order of fifty tonnes — a single block whose transport and placement would have been a feat in its own right (Burl 2000; Hill 2021). It is precisely because the circle is so large, and was so carefully recorded, that Childe’s report repays close reading.

Old Keig as Childe found it

Childe stripped the south-eastern half of the circle to bare subsoil in 1933, having cut his first trench across it in 1932 (Childe 1933; 1934). He recovered the bases of the surviving uprights, traced the bank, and — most importantly — established the relationship between the stones, a “compaction layer” of trampled subsoil, and a central cairn.

The recumbent itself sits almost dead level, its underside “remarkably smooth,” about 1.1–1.2 ft above his datum. Strikingly, over most of its length it rests on neither bedrock nor undisturbed soil: east of his baseline the only solid support for the heavy eastern end is a single block of rock wedged tightly between the stone’s base and a ridge of bedrock (Childe 1934, 376–7). The great stone was, in effect, balanced and packed into place rather than founded on rock.

The flankers and the one surviving circle-pillar tell the same story of careful seating. The east flanker (Pe) stands 9½ ft high; about a foot above its lowest point it is 4 ft 6 in wide, and below that it tapers to a roughly triangular point lying about a foot back from the edge nearest the recumbent. Its base sits in a shallow groove quarried into the bedrock — only some 4 to 5 inches deep and about 3½ ft long, with only the inner edge clearly cut, because the bedrock shelved away on the outer side. The outer edge of the stone was then held by two wedge-stones driven in beneath it, and the inner side packed with large stones bedded on the compaction layer (Childe 1934, 377–8). The pillar P3, the only stone of the great circle still standing apart from the flankers, was set in a socket just nine inches deep and held by wedges, two of them placed obliquely so as to act as “skid-stones,” guiding the base down into the hole as the stone was raised (Childe 1934, 379). Childe noticed that all the pillars he examined were “more or less pointed at the base,” and referred his readers to Kilbride-Jones’s paper for the explanation (Childe 1934, 385).

What the circle enclosed was a burial. Towards the centre Childe found the disturbed remains of a ring cairn — a rough kerb of slabs about 15 ft in radius, set on the compaction layer — and beneath it a patch of subsoil baked brick-red by intense heat, cut by a small east–west trench that, in his words, “looks suspiciously like a grave.” Cremated human bone was spread through this area, together with fragments of several large coarse urns, which Childe suspected were containers for offerings rather than for the ashes themselves (Childe 1934, 386–8). The relationship of the cairn to the compaction layer let him reconstruct a sequence: the uprights were set up first, the trampling that produced the compaction layer happening during their erection, and only later was the cairn heaped over the burial — “just as a Pharaoh would build his pyramid before his death,” as he put it (Childe 1934, 388).

Then, the dating — and here the report shows its age. The only chronological evidence was the pottery: a coarse, flat-rimmed ware which Childe matched, on the one hand, to the fabric then assigned to “Iron Age A” in England and, on the other, to material associated with Late Bronze Age objects at Covesea, with a lignite armlet pointing the same way (Childe 1934, 389–91). He therefore placed Old Keig in a “Late Bronze Age” horizon. We now know this is far too late: flat-rimmed wares of this general kind were long-lived and undiagnostic, and the recumbent circles as a class belong to the late third millennium, some fifteen centuries or more earlier than Childe supposed (Bradley 2005). The excavation is excellent; the calendar attached to it is obsolete. That distinction matters for every comparison with Stonehenge that follows.

Illustrations from Kilbride-Jones

Deliberately shaped stones: Kilbride-Jones on the monoliths

Kilbride-Jones’s contribution begins with an observation that still carries weight: the stones at Old Keig were not rough boulders but were split from the living rock and shaped to a deliberate, repeated form. He measured three of them in detail (Table 1). Each is thicker towards the base and along one face; each has a pointed base, with the point set not in the middle but below one edge, so that the profile resembles a right-angled triangle; and in each the apex is carried vertically above the point of the base. The straight, thicker side and the pointed base together throw the centre of gravity towards one side and keep it low (Kilbride-Jones 1934, 83–8).

The finest of the three, the prostrate Pm, clinched the argument. Its four sides are comparatively straight, it is “a very regular trapezium,” and — most tellingly — its base bears careful tooling that Kilbride-Jones took as positive proof that the shape of the base was intentional, “the result of a preconceived idea and plan” (Kilbride-Jones 1934, 85). He thought the tooling had been done with a metal tool, probably bronze, and drew an explicit comparison with the dressing of the Stonehenge sarsens, which were worked with sarsen mauls (Kilbride-Jones 1934, 83, 85). The roughest stone, Px, reused as the footing of a field-dyke, shows the same design crudely executed; the form, he argued, was a standard, more or less well achieved according to the care taken.

Table 1. Principal dimensions of the three Old Keig monoliths recorded in 1934 (after Childe 1934 and Kilbride-Jones 1934). Figures are as given in the reports; the two authors’ measurements for Pe differ slightly.

Stone	Length / height	Max. breadth	Notes
Pe (east flanker; in situ)	9½ ft (Childe); 9 ft base–apex and 5 ft 11 in above turf (K-J)	4 ft 9 in (K-J)	Triangular pointed base; mass above the base; held by two outer wedges and inner packing
Pm (prostrate pillar)	7 ft 7 in (K-J); 7½ ft (Childe)	3 ft 8½ in (K-J)	“Finest example”; a regular trapezium; tooled, pointed base
Px (prostrate; reused in a dyke)	8 ft 6 in	3 ft 7 in	“Rudest” of the three; the same form, roughly executed
P3 (circle-pillar; in situ)	9½ ft	4½ ft at turf	Socket only 9 in deep; held by wedges and oblique skid-stones

 

A method of erection

Why shape a stone this way? Kilbride-Jones’s answer is the heart of his paper, and it is genuinely clever. He proposed that the shaped monoliths were not tipped upright from flat on the ground, nor levered up from outside or inside the ring, but were laid on their shorter, curved side; a triangular socket was cut to take the pointed base; and the stone was then rolled upright, pivoting on the point of its base as a fulcrum, with the push applied near the apex, as nearly as possible at right angles to the line from fulcrum to top (Kilbride-Jones 1934, 88–9).

The mechanics were worked out for him by an engineer, A. Regnauld, and printed as an appendix. For an idealised stone of even thickness weighing five tons, the initial push needed is about 1.575 tons; as the stone rises this force falls away rapidly, reaching zero once the stone has turned through 48°, at which point the centre of gravity passes over the fulcrum and the stone balances. For the remaining 42° to vertical the problem is reversed — the stone now tends to over-topple, and all that is required is a restraining rope around the apex (Kilbride-Jones 1934, 89, 94–6). Kilbride-Jones contrasted this with the conventional assumption of raising a symmetrical stone from flat on the ground, which on Regnauld’s figures would demand a push of about 2¼ tons sustained through almost the whole 90° arc. The shaped stone, raised his way, needs nearly a ton less force and roughly half the total work (Kilbride-Jones 1934, 89–90). The asymmetry is not decoration; it is mechanical advantage, designed in.

It is worth being clear about the status of these numbers. They are an idealised statics calculation for a five-ton stone, not a measurement of the Old Keig monoliths, which are larger and irregular; and they assume the push can be applied at the top in a particular direction — an assumption Regnauld himself flagged as questionable, since the relevant point can be well above head height (Kilbride-Jones 1934, 95–6). The value of the argument is qualitative, and it remains sound: a pointed, asymmetric, base-heavy stone is markedly easier to stand up than a symmetrical one, and the people who shaped these stones evidently knew it.

Childe’s report supplies the independent, practical check. His contractor, J. C. Milne of Bents, “experienced in handling large stones,” described essentially the same operation from the mason’s side: the stones brought up on log rollers with wooden levers, a socket dug, the base raised by leverage with packing stones slipped underneath after each small lift until it could slide into place, the oblique packers acting as skid-stones to guide the base down (Childe 1934, 384–5). The skid-stones Childe actually found in the sockets of Pe, P3 and the lesser uprights are physical confirmation that the bases were guided into prepared holes as the stones came up, exactly as a controlled, levered erection requires.

Kilbride-Jones takes the theory to Stonehenge

Kilbride-Jones did not stop at Old Keig. The second half of his paper applies the same analysis to Stonehenge, and it is here that the 1934 volume becomes directly relevant to the Altar Stone debate. He argued that the great sarsens were raised in the same way as the Scottish flankers, not tipped from horizontal as William Gowland and others had supposed. Stone 56, the tallest upright, has a pointed base resting against a wedge “in the same manner as the east flanking Stone Pe at Old Keig,” and Kilbride-Jones read Gowland’s puzzling cut-away in the chalk as the mark of a fulcrum that had slipped during raising (Gowland 1902, via Kilbride-Jones 1934, 90–2). He pointed to Lt.-Col. Hawley’s discovery of post-holes set against the broad faces of Stones 1, 7 and 30 as the footings of temporary timber buttresses — lines of saplings that steadied each stone as it was hauled past the vertical and adjusted sideways to receive its lintel, then cut off at ground level, which (he suggested) explained the decayed wood Hawley found in them. He even noted that several Stonehenge uprights (Stones 29 and 30, 6 and 7) taper to a point on both faces, a refinement that avoids the base striking the edge of the hole as the stone is brought up (Hawley 1921, via Kilbride-Jones 1934, 92–3).

Whatever one makes of the details, two things are notable. First, an experienced field observer in 1934, comparing the two monuments stone for stone, was struck by how much the engineering had in common. Second — and this is usually forgotten — Kilbride-Jones read the relationship in the opposite direction to the modern story. He believed Stonehenge “antedates the stone circles of the north-east of Scotland by several centuries,” and that the standardised, pointed-base form was an idea that appeared first, experimentally, at Stonehenge and only later became perfected and routine in Scotland (Kilbride-Jones 1934, 93–4). On his chronology, if anyone influenced anyone, Wessex influenced the north — the reverse of the implication many have since drawn from the Altar Stone’s Scottish origin.

That chronology cannot now stand. Stonehenge’s sarsen settings are dated to around 2500 BC — the sarsen circle to 2580–2475 cal BC (Parker Pearson et al. 2024) — which is essentially the same horizon as the floruit of the recumbent circles. Neither tradition is clearly the parent of the other. Kilbride-Jones’s “Stonehenge first” sequence dissolves; but so does any simple “Scotland first” replacement. On present dating the two were broadly contemporary, and the shared engineering is most economically explained as a common solution to a common problem, reached within a connected island.

The recumbent and the Altar Stone

This brings us back to the Altar Stone. It is a slab of Old Red Sandstone 4.9 m long, 1 m wide and 0.5 m thick — about six tonnes — lying flat in the south-western part of the centre of Stonehenge, immediately in front of the Great Trilithon and now pinned beneath that trilithon’s fallen upright (Stone 55) and lintel (Stone 156) (Atkinson 1956; Clarke et al. 2024; Parker Pearson et al. 2024).

Whether it was ever anything but recumbent has been debated for a century. Atkinson thought the bevelled dressing of one end showed it had once stood upright and later fallen. The contrary case rests on alignment. Aubrey Burl, in a careful discussion, calculated that the odds against the stone coming to lie on the solstitial alignment by a chance fall are about one in 165, and concluded that it was placed flat on purpose (Burl 2006). My own contribution (Daw 2015; 2025), building on Gordon Freeman’s work, was to notice that the Altar Stone is not merely lying along an alignment but shares a specific skew — about 80° from the main midsummer-sunrise axis — with the tallest trilithon (Stones 55 and 56) and the bluestones beside it, the whole group encoding a secondary solstitial axis from midwinter sunrise to midsummer sunset. A stone that participates in that designed geometry was not dropped there by accident. Parker Pearson and colleagues (2024) have independently come to the same view: the Altar Stone was set deliberately as a recumbent, and its solstice-aligned position in front of the Great Trilithon is the work of design, not collapse.

Once the deliberate placement is granted, the resemblance to the Scottish circles is hard to ignore, and Parker Pearson’s team make it explicit: a great stone, laid flat, in the south-western arc of the monument, at the foot of its two tallest uprights, “recalls the plans of recumbent stone circles of north-east Scotland” (Parker Pearson et al. 2024). The recumbent and its flankers at, say, Old Keig, and the Altar Stone beneath the towering Great Trilithon, occupy structurally similar places: the focal, southern, low-stone-between-tall-stones position around which the whole monument is organised.

It would be easy to overstate this, and rigour demands that the differences be set out as plainly as the likeness. A recumbent stone circle is a ring of graded uprights with two tall flankers immediately bracketing the recumbent; Stonehenge has no such graded ring, and no flankers in the Aberdeenshire sense — the Great Trilithon is a lintelled pair, an architectural form entirely absent from the Scottish tradition. The alignments differ in kind: the recumbents frame the low southern moon (or, on the cautious reading, the midwinter sun) across a level sill, whereas the Altar Stone lies along the solar solstitial axis of a monument famous for it. And the engineering similarities that so struck Kilbride-Jones — pointed bases, levered erection, timber buttresses — are, on his own argument, generic insular practice, the kind of thing competent stone-raisers anywhere in Britain might arrive at independently. The Altar Stone parallel is genuine and suggestive; it is not, by itself, proof that the builders of Stonehenge had a recumbent stone circle in mind.

What the provenance does and does not tell us

The geology is now the firmest part of the argument, and also the most often over-read. Clarke and colleagues (2024) fingerprinted the Altar Stone from the ages and chemistry of its detrital zircon, apatite and rutile grains, and matched it to the Old Red Sandstone of the Orcadian Basin — a match they tested against, and which excluded, the rest of Britain and Ireland. The implied journey is at least 750 km, and because hauling a six-tonne block that far overland across the grain of the country is so forbidding, transport by sea is the likeliest option (Clarke et al. 2024). A follow-up study, applying the same methods to Orcadian rocks and to the monoliths of the Stenness and Brodgar circles, ruled out Mainland Orkney specifically, pointing instead to the Scottish mainland (Bevins et al. 2024).

The Orcadian Basin is large. As Parker Pearson and colleagues (2024) note, it runs from Orkney and Shetland down to Inverness and eastward along the coast to Banff, Turriff and Rhynie. The last two names matter here: Turriff and Rhynie lie in the heart of recumbent-stone-circle country. So the provenance is at least consistent with a source in or near the region of the recumbent circles — but it does no more than that. It identifies a geological formation, not a quarry, still less a particular monument or a particular community of builders. The precise source remains unestablished, and the authors of the provenance study were careful to say so. To move from “the rock is from the Orcadian Basin” to “the Altar Stone was taken from an Aberdeenshire stone circle” is to add a great deal that the science does not supply.

What the provenance does establish is connection at a distance, and that is not trivial. A six-tonne stone did not travel several hundred miles by accident. Whether it came as raw rock, as a stone already standing (or lying) in a Scottish monument and later dismantled, or as a gift sealing some alliance — all of which have been suggested (Parker Pearson et al. 2024) — it testifies to sustained, organised, long-range contact between northern Scotland and Salisbury Plain in the centuries around 2500 BC. That contact is corroborated from a quite different direction by the close similarities between the timber house floors at Durrington Walls, beside Stonehenge, and those of Late Neolithic Orkney (Parker Pearson et al. 2024). The Altar Stone is one thread in a wider weave of northern connection.

Discussion: a shared repertoire, not a simple borrowing

Set side by side, the 1934 reports and the 2024 science point in the same broad direction without quite meeting. From Old Keig we learn that the builders of a recumbent circle deliberately shaped their stones, understood how to stand a heavy monolith up with the least effort, bedded and packed their stones with skid-stones and timber stays, and organised the whole monument around a recumbent stone and a cremation burial. From Stonehenge we learn that its central recumbent slab was placed deliberately, in a structurally comparable position, and that the rock for it came from the very region where recumbent circles are found. Kilbride-Jones, looking only at the stones, already thought the two traditions shared a method; Parker Pearson’s team, looking at position and provenance, think they shared an idea.

The cautious synthesis is that both are glimpses of a connected Neolithic Britain with a common stock of practices and symbols — the worked monolith, the economical erection, the meaningful recumbent stone — drawn upon differently in different places. That reading honours the genuine parallels without inventing a migration of architects from Aberdeenshire to Wessex (or, in Kilbride-Jones’s version, the reverse). It also fits the present chronology, in which the sarsen Stonehenge and the recumbent circles are broadly contemporary rather than parent and child.

Ninety years ago, two men crouched over the upturned base of a stone at Old Keig and saw in it the same problem the builders of Stonehenge had solved. They could not have known that the central stone of the monument they were comparing it to had itself come from their own corner of Scotland. That is the kind of coincidence that ought to make us look again — carefully — at what they wrote.

References

Atkinson, R. J. C. 1956. Stonehenge. London: Hamish Hamilton.

Bevins, R. E., Pearce, N. J. G., Hillier, S., Pirrie, D., Ixer, R. A., Andò, S., Barbarano, M., Power, M. & Turner, P. 2024. ‘Was the Stonehenge Altar Stone from Orkney? Investigating the mineralogy and geochemistry of Orcadian Old Red Sandstones and Neolithic circle monuments.’ Journal of Archaeological Science: Reports 58: 104738.

Bradley, R. 2005. The Moon and the Bonfire: An Investigation of Three Stone Circles in North-East Scotland. Edinburgh: Society of Antiquaries of Scotland.

Burl, A. 2000. The Stone Circles of Britain, Ireland and Brittany. New Haven & London: Yale University Press.

Burl, A. 2006. A Brief History of Stonehenge. London: Robinson.

Childe, V. G. 1933. ‘Trial Excavations at the Old Keig Stone Circle, Aberdeenshire.’ Proceedings of the Society of Antiquaries of Scotland 67: 37–53.

Childe, V. G. 1934. ‘Final Report on the Excavation of the Stone Circle at Old Keig, Aberdeenshire.’ Proceedings of the Society of Antiquaries of Scotland 68: 372–393.

Clarke, A. J. I., Kirkland, C. L., Bevins, R. E., Pearce, N. J. G., Glorie, S. & Ixer, R. A. 2024. ‘A Scottish provenance for the Altar Stone of Stonehenge.’ Nature 632 (8025): 570–575. doi:10.1038/s41586-024-07652-1.

Daw, T. 2015. ‘The “Twisted Trilithon”: a secondary solstitial axis at Stonehenge.’ Wiltshire Archaeological and Natural History Magazine 108. 15-24 

Daw, T. 2025. ‘The Geometry of the Horseshoe Stones at Stonehenge: Design, Alignment, and Cultural Significance.’ sarsen.org. doi:10.13140/RG.2.2.23505.65127.

Gowland, W. 1902. ‘Recent Excavations at Stonehenge.’ Archaeologia 58: 37–118.

Hawley, W. 1921. ‘Stonehenge: Interim Report on the Exploration.’ The Antiquaries Journal 1: 19–41.

Hill, J. 2021. The Recumbent Stone Circles of Aberdeenshire: Archaeology, Design, Astronomy and Methods. Newcastle upon Tyne: Cambridge Scholars Publishing.

Kilbride-Jones, H. E. 1934. ‘Stone Circles: A New Theory of the Erection of the Monoliths.’ Proceedings of the Society of Antiquaries of Scotland 68: 81–96.

Parker Pearson, M., Bevins, R., Bradley, R., Ixer, R., Pearce, N. & Richards, C. 2024. ‘Stonehenge and its Altar Stone: the significance of distant stone sources.’ Archaeology International 27 (1): 113–137. doi:10.14324/ai.27.1.13.