Perhaps the most striking parallel is with a monument at Monquhitter which was levelled in 1894. Joseph Anderson (1902, 676) who visited the site at the time it was destroyed refers to ‘a ring of stones, some of which were nearly three feet high, enclosing a circular area of about twenty-eight yards in diameter’. It is not clear whether he was referring to a kerb, a circle of monoliths, or to both. At the centre of the cairn was ‘an approximately circular enclosure defined by rudely-built walling’. He observed that there was no passage linking the centre of the monument to the perimeter. For that reason the ‘chamber’ has been identified as the court of a ring cairn. A pit containing Beaker pottery was probably contemporary with this structure.
There were two small cists in the body of the cairn. One contained cremated bone, whilst its neighbour included a remarkable collection of artefacts. A flint knife and a flint axe were presumably associated with the first use of the monument, but they were found together with a jet ring, two decorated glass counters, a bracelet, a glass bead and a sardonyx intaglio of a dancing satyr. With these objects was a series of 44 pebbles which had obviously been brought to the site, as well as six pieces of rock crystal and one of amber. Like those from Waulkmill, the pebbles were smooth and may have been collected from a beach. They occurred in a variety of materials and colours. The deposit is dated to the second or third century AD. This group is discussed by Mark Hall on p. 51. Robert Stevenson (1967) suggested that both cists were associated with the original monument and that one of them had been reopened to receive these artefacts. That would be consistent with the small proportions of these structures. Similar features are associated with Early Bronze Age cemeteries in the northeast of Scotland.
Only 6 km from Monquhitter was another deposit which included a distinctive Roman artefact. This was a glass flagon discovered about 1857 in the building of a railway line at Turriff. According to the earliest account, it was ‘in a sandy hillock ... There were no bones nor ashes nor the remains of any urn, but with the bottle was found a large number of dark-brownish, wine-coloured glass beads from half an inch to four-fifths of an inch in diameter’ (Dunbar 1930, 147–8). They have since been lost. In this case there was no direct link with a prehistoric monument, although the ‘hillock’ could have been an artificial mound. This deposit was probably in a burial, and Anne Robertson (1970) dates it to the second or early third century AD. Dominic Ingemark (2014) suggests that the vessel was between about AD 65 and 125.
Another early source refers to a large cairn at Boghead, near Kintore, 33 km from Waulkmill. It was removed in the late eighteenth century, but it is known to have contained a stone coffin. Again the monument was associated with artefacts of more than one date, for there were parts of an Early Bronze Age jet necklace, which still survive, as well as a Romano-British glass bracelet (a variant of type 3A) inside a ceramic vessel containing human hair (Wilson 1849, 11–12; Kilbride-Jones 1938, 382, fig. 5.2). Presumably this was a case in which a Roman deposit, dating to the first or second century AD, was inserted into an Early Bronze Age cist. In other cases the available evidence is even less detailed, but it includes two finds from Aberdeenshire: a glass bead perhaps from a stone circle at New Deer, and a necklace of glass beads apparently from an Early Bronze Age cist at Dukeston (Guido 1978, 193; Ralston and Inglis 1984, 44).
There is a further site where burials dating from this period may have been associated with an earlier monument. In contrast to the finds from Monquhitter, Turriff and Kintore, they contained bronze artefacts made in a local style. Hillockhead is about 8 km northeast of Waulkmill. Here two late Iron Age terrets were found in a short cist associated with a substantial cairn (Simpson 1943, 76–77). The other finds included bones (presumably human) and an urn. The metalwork dates between the third and seventh century AD on current estimates (Hunter 2010), but the reference to a short cist is puzzling as this kind of structure is usually (although not always) a feature of the Early Bronze Age; the same may apply to the ‘urn’. For that reason Simpson suggested that an earlier burial may have been reopened.
All these sites were destroyed a long time ago, and in no case were they recorded in much detail. Before the new excavation at Waulkmill there would have been little reason to bring them together. Even now the connections between the primary and secondary uses of these places are not equally strong. They carry most conviction at Monquhitter and the Hill of Boghead and are least satisfactory at Turriff. On the other hand, the original structures included cairns that may have been built in the Early Bronze Age, and the monument at Monquhitter could have been similar to that at Waulkmill.
The finds from these places are of comparable dates. None is assigned to the late first millennium BC, but the deposit from Hillockhead dates to the third–seventh century AD. The reuse of the cairn at Hill of Boghead is dated to the first–second century AD and the finds from Monquhitter and Turriff to the second–third centuries. The diagnostic artefacts from Waulkmill are probably of third century date.
The local contexts of the finds from northeast Scotland are distinctive in a number of ways. Those around Cromar are located in inland areas well beyond the distribution of Roman military sites (Jones 2011), but that does not apply to the other finds. The dating of these burials is not precise and it is impossible to relate them to the presence of Roman soldiers in this part of Scotland. It may be more significant that these deposits were found between 150 and 200 km beyond the northern frontier established by the Antonine Wall in the mid-second century AD. They share their distinctive distribution with a local style of bronze metalwork dating from the first and second centuries AD. Massive arm rings are a special feature of this region and belong to a tradition of metalworking which is found in the area north of the frontier (Hunter 1997; 2014). Both types have been discovered in Cromar. Terrets, on the other hand, are markedly more widespread.
Many of the artefacts were made of recycled Roman metal. They were ‘an entirely local invention’ (Hunter 2007a, 291). Nor are they found in the same hoards as Roman items. Imported objects were collected as a source of raw material and were melted down to create artefacts of a different type, but even when they were discarded they were kept apart from objects associated with the world beyond the frontier. Roman artefacts are almost absent from the hoards in this part of Scotland, but they do occur in graves where they can be associated with glass beads (Hunter 1997; 2001). They were also relatively common on settlement sites, especially on higher-status ones (Hunter 2013b, 23–24).
There is more information at a local scale. Little is known about the settlements inhabited during this period, but a number of sites in Cromar had a distinctive character. The only detailed surveys were produced in 1911, but they were not published until later (Ogston 1931). One group of unusually large structures was at Melgum Lodge, 700 m from Waulkmill, but only one survives today (Welfare 2011, 527–28) (Figs 8.6–8.7). Sites of the same kind are better preserved at Old and New Kinord, which are between 5 and 6 km from the Waulkmill stone circle (Ogston 1931, chapters 2 and 3).
At Kinord the surface evidence shows that some of the domestic buildings were associated with souterrains. This observation is important as the houses themselves remain undated. Souterrains, on the other hand, had a finite history in the east of Scotland, and particularly in Perthshire, Fife and Angus where they developed towards the end of the first millennium BC and went out of use by the early third century AD (Harding 2004, 196–99). Those in Aberdeenshire are less well-known, but an important site is at Castle Newe, 12 km northwest of Waulkmill (RCAHMS 2007, 88–92). Here the entrance to one of these structures was covered by a deposit containing a Roman coin dating from the first century AD.
Like the Waulkmill stone circle, these settlements referred to the past. One feature of the souterrains had escaped notice until Richard Hingley published his study of the Scottish Iron Age (1992, 29). It was comparatively common for the walls of souterrains to incorporate pieces of carved stone dating from a much older period.
These fragments featured cup marks or cups and rings of Late Neolithic or Early Bronze Age origin and were sometimes placed in prominent positions. This practice is also known in Aberdeenshire but has received little attention. Two examples are recorded in surviving structures not far from Waulkmill: at Culsh 3 km to its east (Jervise 1865), and at Clova 7 km to its north where it happened in two separate souterrains belonging to the same group (RCAHMS 2007, 90–91). At Culsh a cup-marked stone was located close to the beginning of the passage, while at Clova cup marks and cups and rings were identified on the underside of two of the roof slabs. Inside a second example on the same site there were two cup-marked stones, one of them not far from the entrance. In each case pieces of ancient rock art were incorporated in a new context. This could have happened during the same period as the reuse of the Waulkmill stone circle as a cemetery, and both involved an explicit reference to the past.
Hall and Forsyth describe the grave goods found at Waulkmill in the nineteenth century as a ‘clear mixture of indigenous and Roman in a high status non-Roman burial context’ (2011, 1331). The same would apply to the graves excavated in 2012 and perhaps to the contents of other burials considered in this paper. It is obvious that local people were in contact with communities much further to the south. Perhaps they were the elite who occupied the monumental roundhouses of Cromar, but this idea has yet to be tested. They acquired imported goods and adopted unfamiliar practices. The three gaming sets from Waulkmill illustrate the point.
At the same time, they kept their distance from the Roman world. They refashioned Roman metalwork to make objects in an extravagant but essentially local style. At Waulkmill they buried their dead in a cemetery whose contents and organisation referred not to the Classical world, but to a form of ancient architecture that was particularly well-represented in the immediate area. By these means people were able to capitalise on their links with the Roman Empire, but emphasised their independence from it.
Figure 8.6. Outline plans of the settlements at Old and New Kinord in relation to the surviving structure at Melgum. Information from Ogston 1931 and Welfare 2011.
The reuse of older monuments in the Pictish period
Richard Bradley
Two or three of the sites considered in this chapter saw another phase of activity in the first millennium AD (Fig. 8.8). One was the large Early Bronze Age ring cairn at Balnuaran of Clava. Excavation in 1994 located a small setting of slabs in between its outer kerb and the ring of monoliths that enclosed it. This feature covered a deposit of cremated human bone. Associated charcoal has a date of AD 560–980 at two sigma and AD 610–770 at one sigma (Bradley 2000, 56–58). The other example is the henge monument at Broomend of Crichie which has a Class I symbol stone in its centre. It is known that the monolith was moved there from a nearby site and was used to mark the position of the Chalcolithic shaft grave excavated in 1855 (Dalrymple 1884). Its original position is reasonably well-documented, allowing the late Ian Shepherd to identify the crop mark of a square barrow at the original findspot. Radiocarbon dates from Hillhead suggest that the monument was disturbed between AD 690 and 967, but the nature of this activity is uncertain.
Figure 8.7. The sockets of the Waulkmill stone circle are indicated by the yellow buckets. The features to the left are Graves 1 and 2. The site of the Melgum settlement is in the wood on the near horizon (Aaron Watson).
Figure 8.8. Distribution of prehistoric monoliths and other structures apparently reused in the Pictish period. Information from Clarke 2007.
Figure 8.9. The Pictish symbol stone from Dingwall showing later motifs apparently superimposed on a distribution of cup marks. Information from Mack 2007.
Two of these examples shed some light on Pictish mortuary practices, but it is important to recognise that they may well be exceptional. Although a few symbol stones did mark the position of barrows or flat graves, others could have been associated with boundaries. The majority have probably been moved from their original locations, making these extraordinary sculptures particularly difficult to understand. It follows that any account of their relationship to older monuments must be approached with caution.
A useful starting point is an important article by Jim Inglis which focused on the small number of decorated stones that are likely to remain in their original positions. There are comparatively few, but Inglis (1987) found that 50% of them were located within 300 m of a prehistoric ‘ritual’ monument. He compared this estimate with a random sample of locations within the same area. Here the equivalent figure was only 2%. The contrast is remarkable.
There are other ways of investigating the relationship between prehistoric monuments and the earliest symbol stones – those assigned to Class 1 and dated to the sixth and seventh centuries AD. One is to consider individual sites. There are some cases in which these designs were created at monuments which can still be seen today. The most convincing example is at Nether Corskie (Welfare 2011, 16). Another is the Brandsbutt Stone which may have formed part of the ring of monoliths whose plan has been recovered by excavation (Shepherd 1984). Alternatively, the decorated stone could have been located nearby. Another well-documented case is at Edderton where a symbol stone is about 150 m southwest of the Ardavannie stone circle which originally included ten monoliths and was associated with an Early Bronze Age cist (information from the Highland Council Historic Environment Record). In other cases the identification of this relationship depends on documentary sources which may be unreliable. That is why Mack (2007) lists six examples where this could have happened, while Clarke (2007) counts only three, although he suggests that the decorated stone at Broomend of Crichie might have formed part of the avenue linking a recumbent stone circle, the henge and a group of Beaker cist burials.
Another method is to look for reused standing stones. In this case analysis of the Pictish carvings follows two procedures. The first is to consider the significance of irregular or unusually tall uprights. Unlike most of those assigned to Class 1, their surface had not been dressed. Again a good example is the monolith at Edderton, but it is unusual because it was so near to an older monument. The argument is necessarily subjective, but it is more persuasive where the decorated surface features an array of cup marks as well as the characteristic repertoire of Pictish art. That method is especially important where the two groups of symbols occupy different parts of the decorated surface, or where the cup marks are more eroded than the remaining images (Clarke 2007). In other cases a design which clearly dates from the first millennium AD overlaps a distribution of cup marks but takes no account of their distribution. Good examples are recorded from Wester Balblair and Dingwall (Fig. 8.9). There are 45 examples in Clarke’s list. Both Clarke (2007) and Mack (2007) make the interesting observation that stones which seem to have been significant in the prehistoric period are most often found at sites where the Pictish designs themselves were created in more than one phase.
Some of those distinctive images are also represented on metalwork (Henderson and Henderson 2004, chapter 4), and it may be no coincidence that two of the most important hoards of decorated artefacts come from Bronze Age monuments. One is the collection from Gaulcross, where the original structure seems to have been a circle of six monoliths (Stevenson 1966). More silver objects have been found there recently, and excavation suggests that the monument was also reused during the Middle Bronze Age and the Early Iron Age (Goldberg et al. 2015). Another collection of Pictish metalwork comes from Norrie’s Law (Graham-Campbell 1992). This structure has been described as a ring cairn, but in its final form it was probably a large round barrow surrounded by a ditch (RCAHMS 1933, 185–86). During its prehistoric phase it had been associated with a series of cists, cremations and an urned burial. It may be no accident that hoards which have so much in common should come from similar locations. It is conceivable that the ‘silver articles’ found at Waulkmill in the nineteenth century were of similar age. Whether or not that was the case it seems likely that prehistoric monuments were delibe
rately selected for reuse.
How should these observations be interpreted? David Clarke provides the most persuasive explanation. The Pictish designs attributed to Class 1 are usually described as ‘symbol stones’. As he says, it would be more helpful to think of them as ‘stones with symbols’. In that sense they recall a much more ancient tradition:
The use of undressed boulders seems clearly intended to evoke associations with the landscape’s most enduring symbol of past belief systems, the standing stones or the individual components of a stone circle. The point is that we are here seeing more the use of the imagery of the past than the re-use of the actual monuments.
(Clarke 2007, 33)
The designs attributed to Class 1 do not contain any Christian images. They are very different from those of Classes 2 and 3, which include carvings of the cross and scenes from the Bible. Unlike the reused fragments that have been considered so far, these stones had been carefully shaped, and in some cases the traditional designs are found on one face and those connected with the new religion on the reverse. The treatment of the stone makes it harder to identify earlier cup marks, but, even where they should be visible, they are quite unusual.
Clarke interprets the reuse of prehistoric monoliths as a reaction to the adoption of Christianity in northern and eastern Scotland and a restatement of traditional beliefs at a time when they were being questioned. It was ‘an explicit renunciation of the purpose of Christian missionary activities’ (Clarke 2007, 35). That is why older monuments acquired a new importance in the regions where the new religion was accepted only slowly.
The Use and Reuse of Stone Circles Page 22
