The Use and Reuse of Stone Circles

by Richard Bradley

  Four hundred pieces of worked quartz came from the area surrounding the stone circle. The raw material was generally of higher quality than the finds from its interior and included small flakes which probably result from artefact production; others could have been utilised. In contrast to the area within the ring cairn, there was less evidence of deliberately broken slabs or boulders of quartz and quartzite, and pink quartz was not distributed far beyond the limits of the monument.

  The character of the principal monument

  Richard Bradley

  This section draws on the results of excavation and field survey in north-east Scotland to shed more light on the nature of this remarkable monument. Greater detail will be provided in Chapter 9 which discusses the amount of variation between the stone circles excavated in the Howe of Cromar (Tomnaverie, Hillhead and Waulkmill), and the Blue Cairn which was surveyed by Adam Welfare.

  Although the site had been disturbed in modern times and also in the first millennium AD, there seems little doubt that it was originally a recumbent stone circle similar to those investigated in recent years at Tomnaverie and Cothiemuir Wood (Bradley 2005, chapters 2 and 3). It must have been one of the largest structures of its kind with an overall diameter of 26 m, and seems to have been the second highest of those whose exact positions are known. It shares the same components as other monuments – a ring cairn which was filled with rubble during a secondary phase; an outer stone circle; an alignment towards the southwest; an apparent association with major mountains in the region; and evidence that the central area was reused. Several local monuments illustrate the special importance of quartz. The new project suggests that the stone setting at Hillhead was later than the construction of the ring cairn: an additional characteristic shared with other sites.

  Figure 2.28. Beaker sherds and a barbed and tanged arrowhead found together in a forestry trench 16 m south of the monument (Jane Summers).

  Some features are more unusual. One is the location of Hillhead. It contrasts with that of most comparable sites because it is on such high ground. This is very different from the positions of both its neighbours which were associated with the basin to the west. At the same time, neither of those monuments was in an area with many surface finds; the same is true of Cothiemuir Wood (Bradley 2005, chapter 4). Hillhead, on the other hand, is surrounded by finds of worked quartz which outnumber those found during fieldwalking on more sheltered land in the Howe of Cromar. It seems likely that the deposition of worked quartz on the hillside began after the ring cairn had been built. The quartz artefacts cannot be dated exactly but they were not associated with the Late Bronze Age roundhouse excavated in 2014.

  The form of the Hillhead monument is most unusual. The original enclosure was defined by a rubble wall approximately 2.4 m wide and at least 45 cm high. It may have extended around the entire perimeter, with an entrance to the southwest. It was only later that the width of that boundary was increased along part of the circuit and an outer kerb was added. The final ring cairn was massive. The point can no longer be established with any confidence, but the fact that the existence of this site was overlooked when so many other monuments were recorded suggests that the stone circle associated with it may have been inconspicuous. In that case there would have been a striking contrast between the large scale on which the cairn was constructed, and the less prominent role played by a ring of monoliths. It is obvious that the cairn increased in height towards the northeast. If this was a recumbent stone circle, the monoliths would have been tallest at the southwest.

  Recumbent stone circles were sometimes built on carefully levelled platforms, but Hillhead was located on a slope and this effect was emphasised by the structure of the ring cairn which was built on a much larger scale on its uphill side. The outcome was a secluded space that appeared to be slightly hollowed. This effect may have been enhanced by the stepped interior of the enclosure towards the north and northeast. A similar feature was observed at Hatton of Ardoyne in the nineteenth century (Coles 1901, 241–46), but this may be the only example suggested by a modern excavation. On the other hand, it could be one way of interpreting Childe’s plan (1934, fig. 9) of Old Keig which depicts a line of upright slabs running parallel to the outer kerb of a ring cairn but underneath the centre of a bank of rubble.

  Other structural features have counterparts elsewhere in north-east Scotland. The enclosure wall is particularly distinctive. It lacked the usual kerbs of upright stones but was not unlike a series of monuments in Buchan. Hillhead would have been approximately 20 m in diameter, which is considerably larger than any of those structures. A more appropriate comparison may be with the stone circle at North Strone. The two are strikingly similar to one another in size and ground plan, but in this case there is no evidence of a wall (Welfare 2011, 88). Another comparison might be with the bank of large boulders interpreted as the inner kerb of a ring cairn at Cothiemuir Wood (Bradley 2005, 57). This could have been a free-standing element, and again it might have been supplemented by a deposit of rubble and an outer kerb during a later phase.

  That is a tentative suggestion, but it is clear that such a sequence did occur at Hillhead. In fact the distinctive character of its outer kerb is closely matched by its counterpart at Cothiemuir Wood which was held in place between two deposits of large boulders. In their final forms the plans of these monuments were similar to one another. Hillhead was 26 m in diameter and its court measured 15 m across. At Cothiemuir Wood the equivalent measurements were 16 m and 10 m respectively. The ratio of the court diameter to that of the ring cairn was 1:1.6 at Cothiemuir and 1:1.7 at Hillhead.

  There is also a contrast with the well-preserved structure at Tomnaverie which will be considered in more detail in Chapter 9. Its outer kerb was built by exactly the same method, but, unlike that monument, Hillhead did not show any sign of patterned stonework on the surface of the cairn. There was no indication of radial divisions, nor was there anything to suggest the choice of raw materials according to their colours. On the other hand, the putative recumbent and flankers at Hillhead conform to a more general pattern in north-east Scotland where the large horizontal stone has a different colour and texture from the pillars to either side (Welfare 2011, 117–18).

  At Hillhead the features taken to be sockets for the flankers cut through the material of the ring cairn, suggesting that the stone circle was erected afterwards. That was also the case at Tomnaverie and Cothiemuir Wood, but here the monoliths were bedded in an external platform abutting the outer kerb. There was little sign of a similar feature at Hillhead.

  A last anomaly also recalls early observations of these monuments. This is the deposit of broken quartz spread across the interior of the ring cairn. It was the latest component of the monument and has no equivalent in the more extensive excavation at Tomnaverie. Something similar was recognised in the recumbent stone circle at Strichen (Phillips et al. 2006), and it also happened at Croftmoraig where the results of a new excavation are considered in Chapter 4. That is just one of the elements discussed in more detail in the second part of this book.

  CHAPTER THREE

  Excavations at Waulkmill, Tarland, Aberdeenshire: a Neolithic pit, Roman Iron Age burials and an earlier prehistoric stone circle

  Richard Bradley, Amanda Clarke and Fraser Hunter

  Introduction

  The site at Waulkmill was selected for excavation in 2012 because the stone circle recorded there in the nineteenth century was probably the neighbour of Tomnaverie, which is only 2 km away. The aim of the work was to compare these two monuments. At the time it took place nothing was known about an example at Hillhead just 4 km to the northeast; in the event it was investigated one year later. That project has already been considered, and the relationship between all three sites is discussed as a separate study in Part 2 of this monograph.

  The exact position of the Waulkmill circle had been lost and it was necessary to establish it by excavation. The site was located on the low ground of the Howe of Cromar on t
he edge of the modern village of Tarland. It was intervisible with Tomnaverie and Hillhead (Fig. 3.1). Little was known about the monument or its surroundings, but about 700 m to its west a settlement consisting of three enclosures or roundhouses was identified. Approximately 350 m to its east a number of Roman Iron Age artefacts were discovered in the nineteenth century (Coles 1905, 214–17; Ogston 1931, figs 62, 63). They were associated with at least one grave, but they were too far away from the stone circle for any connection to be inferred.

  Figure 3.1. The position of the monument at Waulkmill (arrowed) seen from the Tomnaverie recumbent stone circle (Aaron Watson).

  Figure 3.2. The relocated monolith at Waulkmill, with the position of the stone circle under excavation in the background (Aaron Watson).

  Figure 3.3. The site of the stone circle under excavation, seen from the north. The picture emphasizes its location on top of a shallow ridge (Aaron Watson).

  The Waulkmill stone circle was demolished by the farmer early in the nineteenth century, but one of the uprights was allowed to remain in place (Fig. 3.2). The monument was located where the ground fell away at the western end of an inconspicuous glacial ridge (Fig. 3.3). Until recently the lower ground was boggy and poorly drained, meaning that this feature would have appeared more prominent than it is today. The eastern end of that ridge is cut by a stream, and in between it and the circle there was a quarry. It was here that the Roman Iron Age artefacts were discovered in 1898 and shortly after. Early Ordnance Survey maps supply important information, for over time that quarry expanded towards the west. When the artefacts were found, the part closest to the stream was being worked. This provides an accurate estimate of the original findspot (Fig. 3.4).

  Figure 3.4. The location of Waulkmill stone circle in relation to the local topography, the position of a former sandpit where burials were found in the nineteenth century, and the settlement site at Melgum (contours at 50 m intervals).

  Nineteenth century discoveries and observations

  The stone circle

  Little is known about the Waulkmill stone circle. When Fred Coles (1905, 214) was writing his account of the site in 1905 he was told that it had been destroyed ‘about seventy years ago’; ‘ten or eleven’ monoliths were removed, leaving the one that survives today. More recently it was moved to the edge of the field. No artefacts have been found there, but, in 2000, scatters of worked flint and quartz were recorded from the hillside north of the site (Bradley 2005, 87–92). They were separated from the monument by an area of low-lying ground.

  Roman Iron Age artefacts from the quarry

  More is known about the finds from the quarry, but the earliest records raise the possibility that two separate groups of artefacts have been conflated in subsequent accounts.

  The first discovery was in 1898 and was documented seven years later in Fred Coles’s account of the Waulkmill stone circle (1905, 214–17). He drew on a contemporary newspaper account and his own examination of the finds.

  According to the Aberdeenshire Evening Express for August 6th 1898, a gamekeeper came upon

  four undressed stones, each measuring 2 feet broad and 18 inches high, placed one above the other, and at a depth of 4 feet from the surface … On making further search he discovered eight stone and glass buttons … lying at regular intervals as if they had fallen from the garment worn by the occupant of the tomb. A silver buckle … and two or three pieces of steel or wrought-iron, apparently the remains of some weapon, as well as a number of small bones, were also found. The bones, when lifted, crumbled into dust.

  The landowner told Coles that ‘at a later date, some time in 1899, a number of silver articles were found in the same sand-pit … I fear all trace of them is lost’.

  The artefacts discovered in 1898 were a penannular brooch, a dagger and 13 gaming counters (Fig. 3.5). There was also ‘a piece of almost colourless glass ... cylindrical in form’. In a recent paper Hall and Forsyth (2011) date the grave to the second or third century AD, and Hall considers them in greater detail in a later section of this chapter. Since these artefacts were gaming pieces rather than buttons, the observation that they were ‘lying at regular intervals as if they had fallen from the garment’ needs to be reconsidered. Perhaps they were laid out on a board.

  Figure 3.5. The miniature cauldron, penannular brooch, iron dagger and gaming pieces found in the sandpit at the end of the nineteenth century (NMS).

  Sixteen years later some of the subsequent discoveries were traced to a private collection and published by J. Graham Callander (1915). In his words they included

  a small cup of cast bronze, the crown of a human molar tooth found in the cup, a disc of translucent blue glass broken in two, and twelve pebbles of brown, grey, and whitish quartzite. The owner said that the glass disc was all that remained of a number of similar objects and other relics found in a sand-pit near Tarland, which were once in his possession but which he had been compelled to give up to the Laird; that he had kept the glass disc, it being broken, and that the cup, tooth and pebbles had been found afterwards in the same sand-pit and he had secured them, their discovery being kept secret.

  (Callander 1915, 204; our emphasis)

  It is possible that these artefacts came from more than one deposit. Those found in 1898 were in a cist containing bones. Since it was bounded by massive stones it would have been easy to remove its contents. The 1899 collection included a tooth, the miniature cup and a second set of gaming counters made of different materials from the other group. They could have accompanied a second burial. Lastly, there is the reference to ‘silver articles’ being found in 1899. It cannot refer to any of the artefacts catalogued by Callander, raising the possibility that a third deposit was discovered in the quarry and subsequently lost. It may not have been associated with a burial and could date from a later period (as discussed further below).

  The 2012 excavation

  The aim of the 2012 excavation was to relocate the stone circle and, if possible, to establish its plan and chronology. That would allow it to be compared with the well-preserved monument at Tomnaverie.

  Geophysical survey performs badly on the glacial geology of north-east Scotland, and at Waulkmill the likely position of the stone circle was estimated by considering the local microtopography in relation to the siting of similar monuments in the area. Although the structure was levelled as an obstacle to the plough, the surviving monolith was recorded on the first edition Ordnance Survey map; it is also apparent on a photograph taken in 1915, but has been moved since then (Fig. 3.6). Some of the other stones were probably built into a field wall, as it includes some exceptionally large fragments just south of the supposed location of the monument.

  The character of the excavation was guided by three considerations. The first was that the stone circles of north-east Scotland occur in a limited range of sizes and usually include about a dozen monoliths. In most cases they command an uninterrupted view towards the south or southwest (Welfare 2011). A second feature is the presence of a ring cairn inside the circle of monoliths. Experience at Tomnaverie and other sites showed the kinds of subsoil feature that would result from the removal of a stone setting. Work at Broomend of Crichie (Bradley 2011, 20) and the Hill of Tuach (this volume, Chapter 5) also documented the ways in which monoliths were broken up during the nineteenth century.

  At Waulkmill the modern ploughsoil was removed mechanically over an area measuring 25 × 15 m, revealing the glacial till which was a mixture of coarse gravel, clay, boulders and patches of sand (Fig. 3.2). The ploughsoil at the end of the glacial ridge contained an unusually high density of rubble, perhaps from a levelled cairn. Beyond conforming to the estimated position of the monument, it no longer retained any structure.

  The excavation identified subsoil features dating from three separate phases (Fig. 3.7). In chronological order they were:

  • The base of an isolated pit associated with Neolithic pottery;

  • Traces of the inner and outer kerbs of a demo
lished ring cairn, the sockets of up to four standing stones, and, possibly, a large pit in the centre of the monument. These features should date from the Bronze Age.

  • Two graves associated with Roman Iron Age artefacts, and two deposits of cremated bone with radiocarbon dates in the same period.

  Figure 3.6. An earlier twentieth century postcard of Tarland showing (arrowed) the surviving monolith at Waulkmill in its original position on top of a low ridge (Photographer unknown).

  Figure 3.7. Plan of the excavated features at Waulkmill.

  The Neolithic pit

  The earliest feature was a steep-sided pit which had been truncated by the plough and survived to a depth of only 20 cm. It had a single layer of filling with a broken nodule of waterworn quartz, 10 cm in maximum dimension, in its centre. The feature contained sherds of Carinated Bowl. No other artefacts were found there. A few shallow postholes were identified nearby. One contained a flint flake, but they remain undated and do not define any recognisable structure.

  In view of the fieldwalking project undertaken in 2000, it is worth drawing attention to the absence of worked flint or quartz in the pit itself. It is clear that this method may not always locate the sites of this phase. That is particularly true in Cromar where two long cairns are recorded but no surface finds can be assigned with any confidence to the same period. The same disparity has been noted on other Early and Middle Neolithic sites in Scotland (Sheridan and Sharples 1992, 7–8).