2008:AD1 - Knock Dhu, Ballyhackett, Antrim

A programme of excavation was undertaken between September and November 2008 at the promontory fort of Knock Dhu, Ballyhackett, Co. Antrim. The site was the subject of a series of small-scale excavations that, in their initial stages, formed the subject of a Time Team television programme. The excavations were intended to initiate a long-term research project investigating the archaeological remains of the surrounding landscape.
Knock Dhu is a triple-rampart promontory fort located on the east-facing scarp of the Antrim Plateau. The fort is situated on a protruding ridge at the northern end of the Sallagh Braes. This promontory forms a distinct feature along the east Antrim coast and commands an impressive range of views both southwards towards the Mourne Mountains and eastwards across the North Channel towards the Mull of Kintyre, Ailsa Craig and the southern uplands of Galloway and Ayrshire. The fort is bounded by precipitous cliffs to the north and south and a steep slope to the east, whilst the site’s western side gently extends out on to the Antrim Plateau. It is in this western direction that the site is bounded by its defining triple-rampart earthwork. The enclosed part of the promontory consists of an area of c. 8ha; however, all of the possible hut circles and house platforms that have been identified to date within the interior of the fort are concentrated in the area closest to the ramparts.
Aligned approximately north–south and turning through a slight angle midway along their length, the site’s defining earthworks consist of three separate banks that extend across the promontory of Knock Dhu for c. 300m. At the northern end of the site these banks are set relatively close together, but, towards the southern half of their length, the line of the outer bank runs progressively further out towards the west to create a relatively wide open space within the earthworks. Although visually impressive, all three banks are relatively slight features that have been carefully positioned to exploit the promontory’s natural topography and thereby maximise the impact of their appearance. Throughout its length the inner bank is augmented with a narrow external ditch, whilst, along the centre of the earthworks’ length, the partially silted remains of an external ditch associated with the outer bank is also visible. The 2008 excavations and a ground-penetrating radar survey conducted in July 2008 suggest that this outer ditch may have extended further along the length of the outer bank than it now appears to do, although it still remains unknown whether the outer ditch extended along the entire length of the earthwork defences. No topographic expression of an external ditch associated with the rampart’s middle bank is visible, nor did the ground-penetrating radar survey or the excavation of the single cutting across the earthworks (Trench 1) reveal any evidence for an external ditch associated with the middle bank. Consequently, it is considered unlikely that any part of the middle bank was augmented with an external ditch. The site appears to have had only one original entrance, which is located towards the southern end of the earthworks. The entrance consists of a break through all three lines of banks and the inner ditch close to the point where the distance between the outer and middle bank reaches its maximum width. The results of the excavation of a small trench across the entrance (Trench 4), coupled with the change in character of the middle bank to the south of the entrance, suggest that the feature is original.
The crest of the northern part of the inner bank has been used to define the townland boundary between Ballyhackett and Linford. This boundary has been formalised by a stone wall of uncertain date. A number of apparently subrectangular features have either been cut into, or been built against, the inside of the inner bank. It is assumed that these features are of medieval or later date and represent the remains of booley houses, or other temporary structures.
In total, seven relatively small and narrow trenches were excavated during the course of the 2008 season. A staggered cutting consisting of two adjacent trenches was made across the fort’s earthwork defences in the northern part of the site (Trenches 1 and 7), whilst the character of the original entrance in the southern part of the site was partially investigated in Trench 4. Two trenches were opened up across the interior of two apparently isolated hut circles located in the northern and central parts of the site (Trenches 3 and 6 respectively), whilst a complex of several adjacent hut circles and platforms towards the southern end of the promontory was also partially excavated (Trenches 2 and 5). In addition, part of a circular house-like structure was overlain by the secondary extension of the inner bank (Trench 7).
The earthworks
The earthwork defences were excavated in two near-adjacent trenches. The outer part of the inner bank, the inner ditch, the middle bank, the outer bank and the outer ditch were all investigated in Trench 1 (length 20.8m, width 2m), whilst the remainder of the inner bank was excavated in Trench 7 (length 5.2m, width 2m). Originally, it was intended to excavate a single continuous trench across the fort’s defining earthworks (Trench 1). To minimise the length of Trench 1, it was located it in the northern half of the defences where the three lines of banks and their associated ditches are set closest together. In this part of the site, the townland boundary wall is located upon the top of the fort’s inner bank. As a potentially historic structural element within the scheduled monument it was important that the integrity of the wall was not compromised. Consequently, the eastern end of Trench 1 was located at a mid-point on the western edge of the inner bank. As this meant that a full cutting across the inner bank could not be excavated in Trench 1, another cutting (Trench 7), across the inner bank at a point 3.6m to the south of Trench 1 where a modern gate had been inserted into the townland boundary wall, was excavated. This arrangement meant that a single, albeit staggered, section through the entirety of the site’s defences could be constructed. At this point along their length, excavation demonstrated that the defences had a total width of 23.8m, the inner and outer ditches being 1.1m and 1m deep, and 4.2m and 3.1m wide respectively. The banks were similarly slight structures: the inner bank was 5.2m wide and 0.7m high, the middle bank was 3.7m wide and 0.7m high and the outer bank was 4.1m wide and 1m high.
It was not possible to meaningfully relate the stratigraphy of the three banks and two ditches into a single sequence. However, excavation indicated that there were two main phases of construction across the earthworks. Initially, the inner and middle banks were constructed from material that was, at least in part, quarried during the creation of the adjacent inner ditch, suggesting that these three elements of the site’s defences were contemporary. Similarly the outer bank was apparently built with deposits derived from cutting the adjacent outer ditch. The relatively flat area between the middle and outer banks was not augmented by a ditch. Buried soil horizons were preserved under all three banks; however, an extensive programme of AMS radiocarbon dating of charred twigs and other plant macrofossils demonstrated that significant amounts of already ancient organic material had been incorporated into, and preserved within, these buried soils. Consequently, the dates quoted below only provide crude termini post quem for the construction of the banks. As noted above, the inner and middle bank were both apparently contemporary, therefore the latest date derived from the buried soil beneath the inner bank (UB No. 10920, 3053629 bp, calibrated at two sigma to 1409–1221 bc) is superseded by the latest date derived from the buried soil beneath the middle bank (UB No. 10948, 2792622 bp, calibrated at two sigma to 1008–858 bc). The latest date derived from the soil horizon preserved beneath the outer bank had a calibrated range of 1426–1293 bc (UB No. 10337, 3086627 bp). It is not certain whether the outer bank and ditch were contemporary with the inner bank, inner ditch and middle bank. There was no stratigraphic evidence to indicate that all five elements of the site’s defining earthworks (that is the inner, middle and outer bank, and the inner and outer ditch) were not built in an initial, single episode of construction, although the deviation in the line of the outer bank suggests that the outer bank and ditch could have been the product of a separate episode of construction. That in the southern half of their length the inner and middle banks run along the crest of a natural ridge above the line of the outer bank suggests that, if they are not all contemporary features, the outer bank and ditch are likely to be the later elements of the site’s defences. That this would not be the order suggested by an initial reading of the available AMS radiocarbon dates is best considered a reflection of just how potentially misleading such crude termini post quem can be.
The second definite phase of activity of the earthworks is represented by an extension of the inner bank, which was achieved by using material derived from digging out a substantial amount of silt that had accumulated in the inner ditch. This recutting of the inner ditch presumably took place after a prolonged period of silting. The only meaningful dating evidence recovered from the ditch sequences was the AMS radiocarbon dating of plant macrofossils taken from pollen monoliths and bulk samples, although these demonstrated that the silting sequences of the ditches were compromised by the residual deposition of organic material. The latest date from towards the base of the recut inner ditch sequence provided a crude terminus post quem with a calibrated range of 740–389 bc (UB No. 10286, 2386638 bp) for an early stage in the process of silting up of the recut ditch. There was some stratigraphic evidence to suggest that the outer ditch had also been recut, although the displaced silts had not been used to modify the outer bank. The latest date (UB No. 10335, 2209628 bp) from towards the base of the probably recut outer ditch had a calibrated range at two sigma of 373–200 bc. That the recutting of the inner and outer ditches was carried out at the same time is an unproven, albeit reasonable, assumption.
The entrance
The fort’s entrance was investigated in a small cutting (Trench 4: 5.5m by 2m) located in the base of the break within the inner bank. The entrance was essentially a hollow way whose base and sides had eroded through use. Consequently, slump deposits of bank material and erosional surfaces formed significant elements of the trench’s stratigraphy and there were few contexts from which meaningful samples for radiocarbon dating could be retrieved. Evidence to indicate that the entrance was significantly modified at least once in antiquity was, however, recovered during the course of excavation.
Removal of the modern topsoil revealed a roughly cobbled surface surviving at the base of the entrance. The cobbles were set within a bed of thin silty loam that physically overlay the polished surface of the bedrock. This bed of cobbling was formed by cutting through the natural clay subsoil and its width (1.9m) probably represents the ancient width of the modified entrance passage. On the northern side of the entrance a retaining curb of large stones was laid directly upon the surface of the cobbles. No evidence for a comparable curb on the southern side of the entrance survived. The presence of a polished bedrock surface beneath the in situ cobbles suggests that the cobbles, and by extension the northern curb and the bank deposits it retained, were all secondary modifications to an original entrance which had been worn down to the surface of the bedrock. This interpretation was supported by the recognition of the truncated base of a large post-hole sealed beneath the lowest of the inner bank deposits at a distance of 0.55m to the north of the curb. This post-hole was only partially exposed within Trench 4, but excavation demonstrated that it was cut through the natural subsoil to the surface of the underlying bedrock and had an estimated diameter of 0.4m. Presumably, this post-hole accommodated a structural timber associated with the original entrance. A radiocarbon date derived from charred twigs recovered from the post-hole’s fill provides a crude terminus post quem, calibrated at two sigma, of 1118–930 bc (UB No. 10960, 2855627 bp) for the dismantling of this structural timber.
On the northern side of the entrance the bank, retained by the curb and sealing the post-hole, consisted of a series of dumps of silty clay and clay deposits. These in situ dumped deposits were distinguishable from the overlying deposits of slumped bank material because they had been subject to an episode of burning that had been intense enough to create a surface of heat-modified clay and charcoal. It is uncertain at what date this conflagration took place and the surface that it preserved could represent that created after a considerable period of erosion, rather than the form of the entrance during the site’s occupation. Although no evidence for a buried soil horizon survived on the northern side of the entrance, a buried soil horizon was preserved underneath the in situ bank material on the entrance’s southern side. On the southern side of the entrance a comparable sequence of dumped silty clay deposits was used to construct the inner bank, although some evidence for the use of turves within the bank’s construction was also uncovered.
The houses
Trench 3 (dimensions 4m by 1.4m) was opened up across part of an isolated circular structure located on an apparently artificial rock-cut terrace in the northern part of the promontory. Excavation demonstrated that the terrace had been used three times, each phase being separated by a horizontal discontinuity. The earliest phase of occupation was apparently represented by a truncated hearth and a slightly curved, narrow slot-trench, both of which were cut into the natural loamy soil. A significant quantity of burnt wattle was recovered from the slot-trench which presumably accommodated a wattle wall. Two charred hazelnut-shell fragments from the hearth’s fills produced, when calibrated to two sigma, date ranges of 2454–2201 bc and 2459–2206 bc (UB No. 10914, 3834622 bp, and UB No. 10915, 3851632 bp respectively), whilst a charred hazelnut shell from the wall slot returned a date of 2459–2206 bc (UB No. 10954; 3850627 bp). Although not meaningfully stratigraphically related to each other, the consistency of these dates suggests that the hearth and the slot-trench are elements of a single phase of Late Neolithic or Early Bronze Age occupation. Not enough of the slot-trench was exposed to accurately estimate the shape or size of this structure, however. That the centre of the truncated hearth was located only 1.2m from the slot-trench suggests that the hearth was unlikely to have been located in the centre of a round house. Following an extended period of abandonment, the terrace was reused twice. In both of these episodes of reuse, earlier deposits were heavily truncated and a new earth floor was deposited. Erosion has removed most of the features associated with these two later phases, but superimposed hearths associated with each phase and a stake-hole stratigraphically associated with the earlier of these two hearths did survive. Charred twigs recovered from the stake-hole associated with earliest of these two hearths produced a calibrated date range of 972–825 bc (UB No. 10957, 2746626 bp), whilst charred twigs from the basal fill of the later hearth produced a calibrated date of 836–784 bc (UB No. 10955, 2635627 bp). The possible base of a heavily denuded earthen wall directly overlying the natural subsoil was located c. 3.6m to the north-east of the centre of the hearths. The circular, topographic expression of part of this wall could be traced beyond the edges of the excavation trench. The Late Bronze Age or Early Iron Age hearths were located within the centre of the circular structure defined by this wall, suggesting that it formed part of a round house some 7.2m in diameter, which was associated with one, if not both, of these two hearths.
Trench 6 (dimensions 5.5m by 2m) was excavated across part of a hut circle (diameter 6.5m) located towards the centre of the promontory. The excavated hut circle was apparently associated with a smaller circular structure (diameter 4.5m) located about 4m to the south-west. Excavation demonstrated that the hut circle was built upon a terrace and was defined by an earthen wall built upon a stone footing. Only a short section of this much-denuded wall was exposed. However, the trench was positioned across a south-eastern entrance that had apparently been augmented with cobbles to prevent erosion at this point of frequent egress. Two phases of occupation, represented by truncated hearths stratigraphically separated by a thin silt apparently deposited during a period of abandonment, were identified. A post-hole and a stake-hole associated with the earlier phase of occupation were also excavated. Radiocarbon dating of charred twigs associated with the post-hole provides a date range, calibrated at two sigma, of 969–814 bc (UB No. 10965, 2735630 bp) for the earlier phase, whilst charred twigs recovered from the basal fill of the later hearth provide a date range of 822–768 bc (UB No. 10964; 2609628 bp) for the second phase of occupation.
Two adjacent hut circles, which formed part of a complex of five circular structures located towards the southern end of the promontory, were investigated in Trenches 2 and 5. The first house was excavated in two quadrants (Trench 2: 5.5m by 5m and 4.5m by 2–4.5m). It was built on a platform that had been created by cutting a terrace into the natural slope and using the displaced spoil to extend the terrace by constructing a platform retained by a now denuded stone revetment. This platform sealed a buried soil horizon from which two samples of charred twigs were dated to provide terminus post quem of 1626–1516 bc (UB No. 10952, 3297621 bp) and 3692–3528 bc (UB No. 10953, 4830628 bp). Part of the house’s earthen wall (original diameter 7m, width 0.8m, surviving height 0.25m) and an external drainage gully (width 0.4m, depth 0.09m) survived. Slight traces of an internal gully were also observed. Erosion had removed all occupational surfaces, and by extension any potential evidence for multiple phases of occupation, associated with this house, however, the truncated remains of a near central hearth and a series of stake-holes did survive. Charred twigs recovered from the fill of the hearth returned a date range of 802–555 bc when calibrated at two sigma (UB No. 10950, 2557628 bp). The second house was partially excavated in one of the Trench 2 quadrants and partially investigated at one end of Trench 5 (dimensions 8m by 2m). It too was built upon a platform similar to that of the first house, which sealed a buried soil horizon from which two samples of charred twigs provided terminus post quem of 1494–1318 bc (UB No. 10961, 3136630 bp) and 1737–1535 bc (UB No. 10962, 3357627 bp). The topographic expression of a circular wall (diameter 6.5m) survived at the back of the platform, but this had been truncated by erosion in the excavated part of the site. No other features associated with the second house survived. Although no stratigraphic relationship could be established between the two houses, that the back of the terrace cut to accommodate the first house coincided directly with the stone revetment built to retain the platform upon which the second house was set, and that the external drainage gully associated with the second house fed directly into the drainage gully associated with the first house, suggests that they were probably contemporary.
The final structure identified as a possible house was partially exposed in Trench 7. Removal of the deposits associated with the extension of the inner bank (see above) revealed part of a curving wall of large stones, which presumably formed part of a circular house. Two courses of the wall survived intact (to a height of 0.25m) and, although it was not fully exposed, the undressed stones appeared to be arranged to form two crude faces that retained an infill of smaller stones that were apparently bonded with grey clay. Only a 1.2m length of the wall was excavated; it emerged from the corner of the trench and ended in an apparently deliberate terminal that may have formed the north-eastern side of a north-west-facing entrance. No topographic expression of the structure survived on the ground surface beyond the edges of the trench, and not enough of the wall was exposed to accurately estimate the structure’s size. The wall was built directly upon a levelling deposit of silty clay that overlay the truncated surface of a buried soil. No internal layers or features associated with the occupation of the structure survived; however, a series of possible midden deposits had accumulated against the wall’s external face, either during the structure’s occupation or following its abandonment. That the structure was overlain by deposits associated with the extension of the inner bank indicates that it is ancient, however, whether the structure predated the original construction of the inner bank remains uncertain.
Discussion
Despite the small size of the trenches and the evaluative character of the investigations, the 2008 excavations at Knock Dhu have produced impressive results. Provisional analysis of the results is hampered by problems of dating. Given the paucity of diagnostic finds recovered during the course of the excavation, assessment of the site’s chronology is dependent upon the radiocarbon dating of short-lived organic material such as charred twigs and hazelnut shells. Evidence for the truncated remains of a Late Neolithic or Early Bronze Age house in Trench 3 suggests that the promontory was periodically exploited for occupation prior to the construction of the fort’s defining earthwork defences at some point during, or soon after, the Late Bronze Age. That said, the bulk of the hut circles and house platforms excavated within the interior of the fort also appear to date to the Late Bronze Age and are presumably broadly contemporary with the construction and maintenance of the fort’s defences. A variety of house types were recognised despite the small sizes of the trenches. Some of the structures were built on terraces, whilst others were constructed upon artificial platforms. Various types of wall were also identified, including earthen walls, earthen walls built on to stone footings and clay-bonded stone walls. These variations provide an insight into the diversity of apparently contemporary architectural practices at Knock Dhu.