At the level of the stone setting it was possible to identify the discontinuous stain left by the sides of a timber coffin. They seem to have buckled as the filling of the grave settled against them, but enough remained to suggest that it was originally 2.5 m long and about 60 cm wide. Part of the base was still preserved. It was in very poor condition, but samples of the wood were taken where it survived intact. They have been examined by Steven Allen who reports that they come from a single radially faced board, most probably of oak. Human bones were not represented, but Dr Aleks Pluskowski (pers. comm.) was able to identify three heavily mineralised teeth of bos close to the western end of the grave (Fig. 3.19). They may represent all that remains of a skull, traces of which could be identified as a patch of discoloured soil. Beside them were two bronze rings.
Figure 3.13. Plan and elevation of Grave 1.
Figure 3.14. Vertical view of Grave 1 after excavation (Aaron Watson).
This feature seems to have been filled in two stages. In the first, soil accumulated around the coffin. The upper part of the pit may have stayed open long enough for the faces of the cutting to be affected by erosion. The secondary filling consisted of loose sediments which had probably weathered from the sides. They included a distinct lens of dark humic material, suggesting that these deposits had time to stabilise before more material accumulated.
Figure 3.15. Outline plan of Grave 2.
Discussion
The second grave was obviously much larger than its neighbour, and there were other contrasts between them. Each contained a setting of boulders towards its base, but in one case this took the form of a drystone wall built against the edge of the pit to create a kind of cist – there was no sign of any wood. In the second case the stones were laid around a timber coffin on the bottom of the grave.
The first grave was refilled rapidly, but not before one wall of the burial chamber had given way. After that, it was completely covered. Its neighbour had a different history. Again the burial was enclosed, this time in a coffin surrounded by a setting of boulders. Its wooden sides could be traced through the lowest filling, but there was no trace of the lid, although it might have been level with the surface of the stone setting. The upper part of the pit seems to have been left open for some time, perhaps to allow the coffin, or even the body, to be viewed. Its sides were exposed to natural weathering and it filled by natural silting.
The positions of both graves appeared to acknowledge the presence of the stone circle. They were just outside its perimeter – the distances were 1.25 m and 3 m respectively – and to some extent their alignments echoed that of its outer kerb, although the relationship was only approximate. Still more important, large flakes which may well have been removed from the stones had been incorporated in the fillings of both these features. That happened as the first grave was closed, and here the main concentration was at a higher level than the cist. In Grave 2, however, they were associated with the setting of boulders around the coffin. Here the flakes were in the same deposit as pieces of quartz – a distinctive material that was associated with the earlier ring cairn and may have been taken from that structure.
Figure 3.16. Detailed plan of the lower level of Grave 2.
Cremation burials
Two deposits containing human remains were discovered in the excavation of the stone circle and were initially thought to date from a period of reuse during the Late Bronze Age, when many monuments in the region received deposits of burnt bone. Nine metres outside the stone circle at Waulkmill was a shallow pit, 40 cm in diameter and only 8 cm deep. It was capped by small pieces of stone. Others lined the sides and bottom of this feature, which contained a cremation burial. What was unexpected was its age. A sample from this deposit returned a date in the Roman Iron Age.
The second deposit raised more problems. The existing feature in the centre of the ring cairn was recut by a steep-sided pit, 60 cm in diameter and 55 cm deep. It had been filled by a deposit of soil and charcoal which included a small amount of cremated bone. When it was refilled, its position was marked by a circular setting of rounded cobbles. They were between 7 cm and 12 cm in maximum dimensions. Many of them were quartzite. They had been placed at the exact centre of the older monument. Again the sample of burnt bone produced a radiocarbon date in the Roman Iron Age.
Figure 3.17. Grave 2 showing the setting of boulders towards the left flanking the remains of the wooden coffin (Aaron Watson).
Figure 3.18. The first indication of a coffin stain within Grave 2 (Aaron Watson).
Figure 3.19. Two spiral rings and the stain left by the animal skull in Grave 2 (Aaron Watson).
RADIOCARBON DATES
Samples from both the cremation burials were dated, but the bos teeth were too heavily mineralised to be studied by the same method.
An isolated pit outside the stone circle (Context 1031):
SUERC-44435 Cremated bone. 1802±29 BP/AD 130–323 at 95.4% probability
The recut pit in the centre of the stone circle (Context 1057):
SUERC-44434 Cremated bone. 1673±29 BP/AD 259–428 at 95.4% probability
Cremated bone
Fiona Shapland
Cremated human bone was recovered from two contexts. Context 1031 was a shallow pit which contained a Roman Iron Age cremation burial. Context 1037 was a shallow feature capped by a setting of pebbles dug into the filling of an earlier pit in the centre of the ring cairn.
Methodology
The methodology of analysis used was the same as stated in Chapter 2 (p. 22).
CREMATED BONE FROM CONTEXT 1031
The total weight of bone fragments recovered from this burial was 2063.8 g, which is consistent with the cremation of a single adult (McKinley 1993, 285) but would indicate unusually good post-cremation recovery and virtually complete burial of the remains. No clearly non-human bone fragments were identified, although the majority of fragments were too small for definite identification.
BONE COLOUR AND CONDITION: THE CREMATION PROCESS
The cremated human bone from this burial was consistently white in colour, with very little variation between elements. Most recovered fragments measured less than 10 × 10 mm, although a significant proportion of fragments measured over 20 × 10 mm. Fracture lines were curved and irregular, with considerable warping. This indicates that the body was fleshed at the time of cremation, and that a temperature of over 900°C was maintained (Walker et al. 2008, 133).
IDENTIFIED FRAGMENTS
The majority (over 80%) of bone fragments from this burial were unidentifiable, but 391 fragments were identified as belonging to an area of skeleton or, in a small proportion of cases, a specific bone (Table 3.1). These identified fragments provide a minimum number of individuals of one, as there were no repeated elements identified, although given the large weight of bone recovered the presence of a second individual is a possibility.
EVIDENCE FOR AGE-AT-DEATH AND SEX
Estimation of age-at-death was impeded by the fragmented and incomplete nature of the skeleton. No complete teeth were recovered, the state of the pubic symphysis and auricular surface of the pelvis were not recordable, and sternal rib ends were not observable. Where observable, cranial sutures were partially fused. All observable long bone epiphyses were completely fused, as was a single upper lumbar vertebral ring. In addition, one upper third molar root was observed to be complete. All these factors indicate that this individual was an adult over 19 years of age. One site of degenerative joint disease (see below) may suggest that this is an older adult (over 30 years).
Sex assessment was particularly problematic due to the poor survival of the pelvis, from which no pelvic criteria were recordable. Two cranial features, the orbital margin and occipital process, were observable, and both of these were possibly female. This does not provide strong enough evidence to assign a sex to this individual.
PATHOLOGY
Only one pathological feature was identified on this skeleton: slight marginal osteophyte developme
nt on the articular facet of the tenth or eleventh left rib. This is evidence of joint deterioration which would presumably have affected the corresponding vertebral facet as well. Degenerative joint disease of the costovertebral joints is common in individuals over the age of 40 years, and is usually associated with other degenerative changes in the spine (Ortner 2003, 549, 555).
Cremated bone from Context 1037
This deposit is also dated to the Roman Iron Age. The surviving cremated bone from this deposit consisted of nine small unidentifiable fragments, only two of which were over 5 × 5 mm in size. It is impossible to be certain that these fragments represent human bone, although given the context of discovery this seems most likely. These fragments are predominantly white in colour indicating a high cremation temperature. No information on sex or age-at-death could be ascertained.
Roman Iron Age artefacts
Catalogue of the finds from the 2012 excavation
Fraser Hunter and Penelope Walton Rogers, with surface X-ray fluorescence analysis by Maureen Young
GRAVE 1: A PENANNULAR BROOCH
Fraser Hunter
Description: small penannular brooch of Fowler (1960) type Aa, with a circular-sectioned hoop, slightly oval in plan, expanding gradually to squared-off terminals (Figs 3.20–3.21). Humped pin, circular in section, the tip flattened into a narrow leaf shape (its extremity lost). Filemarks from finishing on the upper surface; very slight circumferential facets on the hoop from hammering into shape. Hoop: external diam. 21.8 × 24.7 mm; rod: diam. 2.7 mm (terminals: diam. 3.8 mm). Pin: length 38.4 mm, diam. 2.2 mm. Alloy: both hoop and pin are leaded bronze with trace levels (probably <1%) of zinc, arsenic and antimony.
Table 3.1. Identified human bone fragments from Context 1031.
Area of skeleton No. fragments Specific bones identified
Cranium 79 Frontal, parietal, occipital, temporal, zygomatic, maxilla, mandible, 3 tooth roots
Vertebrae 27 2 cervical (including axis), 4 thoracic and 3 lumbar vertebrae
Ribs 49 Left and right ribs, upper, middle and lower
Upper limb 85 Humerus, radius, ulna, clavicle, scapula
Hands 38 Scaphoid, lunate, pisiform, hamate, capitate, 5 metacarpals, 2 proximal phalanges, 5 middle phalanges, 6 distal phalanges
Pelvis 9 Ossa coxa, sacrum
Lower limb 89 Femur, tibia, patella
Feet 15 Calcaneus, talus, 2 metatarsals, 2 distal phalanges
The brooch lay towards the north-east end of the grave. No body traces survived, but the location is consistent with a position around the neck or upper chest. There were no preserved textile traces to confirm whether it had been attached to clothing; its small size would be suitable only for fine textile. Type Aa is widespread in Scotland, with examples both in iron and copper alloy (iron: Laws of Monifieth, Angus; Craigie, Angus; Dunbar, East Lothian; copper alloy: Dun Fheurain, Argyll; Culbin, Moray; Camelon, Falkirk; West Mains of Ethie, Angus; Luffness, East Lothian) (Fowler 1960, 171–72; Wilson 1980, fig. 3, no. 2; Longworth 1966, 180, fig 4.2; NMS X.FA 93). Fowler argued that this was the earliest form of penannular brooch; there are securely dated Iron Age examples (for instance from Wetwang Slack in Yorkshire, probably of third century BC date; Dent 1982), but its currency clearly continued into the Roman period as the finds from Camelon and Craigie demonstrate: the former from a Roman fort, the latter with a burial radiocarbon-dated to the Roman Iron Age (Sheridan 2004b, 176). Thus the Waulkmill brooch cannot be closely dated typologically. The alloy type used is also not chronologically characteristic. However, there is a remarkably similar penannular brooch from earlier finds at the site (Fig. 3.22). Size and pin form are very close, but it differs in the use of silver and the decorated (ribbed) hoop, both of which are unusual features for penannular brooches.
The use of silver puts the 1898 find into the Roman Iron Age, as there is no evidence of pre-Roman silver use in Scotland. Direct parallels are elusive, however. The closest comes from a late Roman hoard of jewellery and spoons from Great Horwood, Buckinghamshire, buried in a small silver beaker (Waugh 1966). The brooch is a similar size and has a ribbed hoop, but the terminals and pin are different. Accompanying spoons suggest a late fourth or fifth century date for deposition of the hoard. This is not impossible for the Waulkmill find, as penannulars were a long-lived form which persisted throughout the Roman period, but it is methodologically suspect to apply the dating from one near-unique find to another, and we must turn to a closer study of the associated finds. As argued above (p. 30), the earlier finds from Waulkmill should be seen as separate burial groups. The brooch was found with the remains of a fine dagger (catalogued below) and gaming pieces, some of Roman origin. The parallel for the brooch is consistent with the terminus post quem provided by the stem of a glass cup which was probably reused as gaming piece. Dominic Ingemark (2014, 52–54, 252) dates the manufacture of the original vessel to between the mid-second and early fourth century AD; he argues that the glass colour suggests the later end of this range, though Hilary Cool (p. 47) prefers late second/early third century. The glass counters are harder to date, as the translucent blue ones are atypical and the polychrome millefiori ones are very rare. She notes, however, that the colours may stem from reusing millefiori glass of late second/early third century date, and there are similar counters in a gaming set from a rich female burial at Bonn, Josefstraße of later third century date (Haberey 1961, 328–30). In sum, the polychrome gaming counters are likely to be late second– third century, the reused vessel no earlier than this date, while the only parallel for the silver penannular brooch is fourth or fifth century AD. On balance, a third century date is most likely, taking the glass as the most securely dated material.
Figure 3.20. The mica schist counter and brooch from Grave 1, and the rings from Grave 2.
Figure 3.21. The penannular brooch from Grave 1 (NMS).
Figure 3.22. The silver penannular brooch found in 1898
Penannular brooches are Iron Age in origin, but their popularity ran through the Roman occupation and they flourished again in the early medieval period, especially in Scotland and Ireland. This makes study of them rather complex. It is tempting to view them as essentially indigenous, but this is hard to argue for the Roman Iron Age in Aberdeenshire where examples are few. Table 3.2 reviews the occurrence of penannular brooches in pre-Roman and Roman Iron Age Scotland. Securely pre-Roman examples are rare, and all bar one are of the very plain Aa type which continued into the Roman Iron Age, as noted above (Fowler 1960; the exception is a B2 from Moredun). Early examples are plain and predominantly iron, rather different from the Romano-British forms. Certain types (including all versions of type A) were commonplace on Romano-British sites (Fowler 1960; Snape 1993, 9, 28–29). The consistency of these brooches across Britain and their differences from earlier penannular brooches in decoration and predominant material make it unlikely that they represent Scottish products. Rather, they are best seen as a Romano-British development stemming from an Iron Age tradition. As such, in a Scottish context they are best classed as imported material from Roman Britain.
Roman brooches from north of the frontier are dominated by types with stylistic links to Iron Age traditions, fitting local tastes in style or form, such as penannulars, trumpet and headstud brooches (Hunter 2013a, 275, fig. 7). Penannulars make up 14% of the brooch assemblage from Roman forts in Scotland and 13% of the Tyne-Solway assemblage (Snape 1993, table 5), but they account for 36% of those from the major hillfort of Traprain Law and 19% of those from other Iron Age contexts. The relatively plain A3 brooches were most common on Iron Age sites, whereas on Roman sites the A2 styles (with milled terminals) are as common as A3. Traprain Law differs markedly in detail from other Iron Age sites, with A2 making up two-thirds of the assemblage. Previous work has noted that Traprain’s brooch assemblage diverges from typical Iron Age sites (Hunter 1996, 122–23), although the reasons are unclear.
In sum, the new Waulkmill find has striking parallels in size and form to the earli
er brooch from the site. Both are best seen as Romano-British rather than local products although, as discussed below, it is also possible that the copper alloy one emulated the silver one.
GRAVE 2: TWO ENIGMATIC RINGS
Description: Spiral ring (sf.22), once complete but now in 13 fragments. Cannot be refitted neatly owing to corrosion of fracture surfaces and loss of surface, but it appears to be essentially identical to sf.23. Associated with it are five fragments of mineralised organics (see Walton-Rogers, below). The ring is: 7.0 mm (hoop); 2.8 mm (terminal). Alloy: leaded bronze (trace zinc, arsenic, antimony).
Table 3.2. Pre-Roman and Roman Iron Age penannular brooches from Scotland divided by different site types, from Fowler (1960) with additions (Fraser Hunter).
Description: Spiral ring with a tall D-section (sf.23). The ends taper neatly in thickness to rounded points and overlap over c. 15% of the circumference. The contact surfaces are very flat, and the join would have been flush. Cast and bent into shape. External diam. 34.0 × 34.8 mm; internal diam. 28.8 mm; 7.0 mm (hoop); 2.8 mm (terminal). Alloy: leaded bronze (trace zinc, arsenic, antimony).
The Use and Reuse of Stone Circles Page 8