Figure 12.01: Building 1. Features associated with a possible timber barrack phase, with plan of later stone barrack shown in outline at 1:200
The Daniels excavations uncovered only limited evidence for earlier phases of timber partitions in the praetentura barracks, of the kind which clearly emerged in the retentura barracks during both the 1978–9 and 1997–8 excavations and which may be taken to indicate the existence of an earlier, timber barrack phase. The relevant features uncovered in each of the barracks are described below and their significance assessed in a concluding discussion.
Several features which could represent the remains of a possible earlier timber barrack were recognised in the eastern half of the Phase 1 stone barrack. These remains took the form of three clay-filled slots which probably held timber sleeper beams and did not fall comfortably into the successive series of slots associated with the two sub-phases of the Phase 1 stone barrack. The three slots were located within Contubernia 1 and 2 (P05:34, N05:49) and inside the western room of the officer’s quarters (P05:48). Only their northern terminals were exposed, evident as streaks of yellow clay. Slots P05:34 and N05:49 were spaced 3.10m apart, a feasible contubernium spacing, whilst P05:34 was 6.80m from slot P05:48, further east, which could conceivably represent the width of two contubernia assuming there was an intermediate slot which was not located by the excavators in 1975. All three terminated between 0.20m and 0.40m to the south of the stone barrack’s north wall, with no trace of an earlier north wall – whether of timber or stone – directly linking-up these partitions. In this they resembled the partitions identified in the Phase 1 levels of Building 3 (M07:20, N07:19, 22, 23), which all similarly terminated between 0.4m and 0.5m short of the south wall (see Fig. 13.10). Moreover, the three anomalous slots in Building 1 were completely at variance with the positioning of the other slots defining the contubernium layout of the Phase 1 stone barrack (see Fig 12.01). The slots ranged in width from 0.20m in the case of P05:34 to up to 0.30m in the case of N05:49. None was observed to extend the south of the secondary north wall of Building 1 and hence the recorded surviving lengths of these slots from north to south nowhere exceeded 0.80m. N05:49, the longest of the slots, as revealed, also featured a dogleg at its northern end giving it an inverted L-shape.
In addition to the three north-south aligned examples described above, a very short length of east-west aligned slot (Q05:57), of uncertain phasing, could be relevant in this context. Found in the west room of the officer’s quarters, this was not assigned to any particular phase by the excavators and could conceivably be associated with a timber barrack block on the site of Building 1, although, equally, it could also belong to the Sub Phase 1 of the stone barrack.
One problem with identifying these slots as part of an earlier timber barrack is that they apparently cut into clay floor levels (e.g. N05:47, P05:33, 45) which, in the course of post-excavation analysis, were firmly assigned to Sub Phase 1 of the Phase 1 stone barrack by Daniels and Moffat, as set out below. Thus, as might be expected, the building’s stratigraphic phasing cannot now be disentangled in a fully satisfactory manner, which would enable features such as the partition slots described above to be confidently reinterpreted in the light of more recent discoveries elsewhere in the fort.
If the slots do represent an earlier timber barrack it is not clear how much weight to put on the apparent lack of a north wall associated with them. If it is not simply a factor of differential survival, it could signify that the barrack frontage featured a different form of construction, perhaps employing a sill-beam resting on the ground surface rather than being set in a narrow trench like the side walls of the contubernia. There could conceivably have been a functional reason for using a different type of construction here, if, for example, the wall panels along the barrack frontage took the form of shutters which could be removed during the heat of summer.
Figure 12.01: Building 1. Features associated with a possible timber barrack phase, with plan of later stone barrack shown in outline at 1:200
Only one partition (M05:41) was identified in Building 2 which may have belonged to a timber phase preceding the stone-walled barrack (see Fig 13.09). Not only did it appear to be out of step with the other timber contubernia partitions associated with the stone building (too close to L05:55 to the west and too distant from M05:42 to the east – 3.10m and 4.50m respectively), but also site plan P216 suggests that this partition continued northwards under the line of the north wall of the building, being visible at the base of robber trench M05:31, and cannot therefore have formed part of the stone barrack (see below Chapter 13, Building 2, Contubernium 6). Although no evidence was recovered for timber outer walls which might have been associated with partition M05:41, the analogy with the Period 1 barracks revealed in the retentura in 1997–8, suggests that Building 2 may similarly have been constructed entirely of timber in its primary phase.
Evidence for two phases of timber partition walling can also be identified in Building 3, although once again the earlier phase, which might be associated with a timber barrack, is represented by only by a single slot, N07:23 (see Fig 13.10). This was 0.15m wide and filled with stone chippings. It was traced over a distance of 1.50m, being cut by the rubble foundations of the south wall of the stone barrack (N07:24) at one end and overlain to the north by a 1.70m long stone alignment (N07:18), which ran parallel to the south wall and probably also represented a feature of the early stone barrack. The slot was located 1.70m from the west wall of the centurion’s quarters (P07:23) and 1.90m from the next partition (N07:22). It clearly did not fall into the same pattern as the three other partition slots (M07:20, N07:19, 22), which were fairly evenly spaced, 3.40–3.60m apart, and convincingly define the four easternmost contubernia of a subsequent stone barrack block.
As a consequence of its anomalous position N07:23 was interpreted by the excavators as a possible cordrig cultivation furrow, pre-dating the construction of the fort and similar to those recognised elsewhere on the fort site. However it appears to have been an isolated slot rather than one of a whole series of such features, running parallel with one another, as was the case with the other cultivation furrows, and was also perhaps rather too narrow, though some irregularity in its width was noted. It might therefore be interpreted as a primary partition instead, associated with an initial timber barrack like those clearly identified in the retentura. No construction slots which might represent the exterior walls of a timber barrack were uncovered beneath the stone foundations of Building 3 in the short lengths where this was checked, but an earlier timber barrack might not have occupied precisely the same footprint as the stone barrack block and hence may conceivably not have been detectable by that means.
There was clear evidence for the existence of two phases or sub-phases of partition slot in Building 4. As was the case in Building 1, both sets of partitions were attributed to the stone barrack in Daniels’ initial post-excavation analysis. The two successive phases are clearest towards the west, with the Phase 1 and Phase 2 slots lying alongside one another, separating Contubernia 1–4, and in this part of the block it is certainly the case that both phases of partition can be fitted comfortably into the plan of the stone barrack. Further east, however, only a single set of partitions formed a coherent pattern within the stone block. Four additional partitions (J04:29, 31, 32, K04:25) at this end of the building were more difficult to fit into the standard layout of a north-facing barrack contubernia, with the two easternmost (J04:32 and K04:25) being spaced one contubernium width apart, along the central north-south axes of Contubernia 8 and 9 respectively (see Fig 13.11).
The most westerly of these slots (J04:31) was uncovered in Contubernium 6, some c. 1.10m from west wall and 2.10m from the east wall of the contubernium, just north of the stone barrack’s south wall, and was rather unusual in form. It consisted of two adjacent lines of yellow clay and small stones clearly visible in the clay make-up (J04:25). On the site plan (P62a), one smear of clay, c. 0.50m north of the south wall, is shown branching off eastward for at least 0.20m, and this interpreted on the inked all period plans as an arm of the slot although it is not mentioned in the relevant context entry in the site notebook.
Another slot (J04:29), some 0.30m wide, was uncovered only 0.65m from the west partition slot of Contubernium 8 (J04:28), marked by a line of grit with very occasional flecks of yellow clay. A further possible slot (J04:32) was traced in the same contubernium, 1.90m from the latter’s west wall and c. 1.25m from its east wall (J04:26). This slot was only identified after excavation, on the basis of details recorded on site plan P62b, and was depicted on the inked plan showing all the features associated with Building 4 (reproduced here as Fig 13.11), but never given a numbered entry in the site notebook. The site plan shows a thin, 1.25m long, yellow streak, presumably meant to represent clay, associated with small patches of charcoal and pink clay in the southern part of the contubernium.
The easternmost of the four anomalous partitions was located in Contubernium 9 and took the form of a north-south slot (K04:25), 0.10m wide, situated 2.00m from the east wall of the building and c. 1.70m from the slight remains of the timber partition slot (J04:26) forming the west wall of the contubernium. Intermittent traces of the southern 2.50m of this were found.
None of the four slots described above corresponded to a recognisable component of a typical barrack contubernium plan and it is unclear what function they could have performed within the stone barrack block. These anomalous slots might therefore have been associated with a timber barrack block, with contubernia occupying different footprints from their counterparts in the stone barrack. However, but it is not easy to integrate these into a coherent pattern with the Phase 1 slots in the western part of the building, so there may have been a timber barrack and two phases of contubernium partitions relating to the subsequent stone block.
Unlike the barrack blocks previously considered here, there were no features in Building 5 which could not be quite plausibly attributed to the stone barrack phases. The description of those phases set out in Chapter 13 follows the outline originally arrived at by the Daniels post-excavation team in the structural sequence in the context database and outline plans which accommodated all the known features in a two-phase scheme for the stone barrack. However it is noticeable that an east-west oriented partition (E04:27, F04:39, F05:59), extending the length of the officer’s quarters, lined up perfectly with a slot in the adjoining Contubernium 1 (G05:29), on the other side of the stone wall separating the two parts of the barrack, both slots lying some 2.00–2.20m from the south wall of the building (see Fig 13.21). This raises the possibility that, rather than representing two separate partitions, as set out in Chapter 13, they might have formed a continuous wall slot associated with a timber barrack preceding the stone block. Such a wall slot might have been an internal partition but it is also possible that it represented an exterior wall – the north wall being most likely – for a timber block which occupied a more southerly footprint, only partially underlying its stone successor. This could also potentially affect the interpretation of other partition slots in the southern part of the officer’s house (e.g. E04:25, F05:58). A further slot (E05:16), interpreted as a post trench, could represent part of the south wall of such a putative timber block. It was located in the area of Alley 4, between Buildings 5 and 6, just south of unused wall foundations E05:04/F05:12. The slot was traced for a length of 1.80m and was 0.25m wide, 0.05m deep, with a bowl-shaped bottom, and was filled with yellow and pink clay. If E05:16 was the southern counterpart of wall slot E04:27/F04:39/F05:59, it would give the timber barrack an internal width of 6.30–6.40m, which compares very closely to the equivalent dimensions of the timber cavalry barracks in the retentura excavated in 1997–8 (c. 6.40–6.60m).
According to this reconstruction of these fragmentary remains, the initial timber version of Building 5 would have lain around 6.30m south of Building 4, presumably separated from it by a street (assuming the timber south wall of 4 lay somewhere within the footprint of the corresponding later stone wall). If it is assumed that another, as yet undiscovered, timber barrack occupied the site of Building 6, that barrack would in turn have lain only c. 2.70–3.00m from the Period 1 timber Building 5. Thus it is possible that the disposition of the Hadrianic barracks in the praetentura – or the western half of it at least – was somewhat different from the later arrangements there. Specifically, the two southerly barracks – equivalent to Buildings 5 and 6 – would initially have been paired more closely together and separated only by a narrow alley, whilst the most northerly barrack – equivalent to 4 – stood in a more isolated position on the other side of a street from 5, an arrangement which was clearly reversed when the barracks were rebuilt in stone in Period 2.
Daniels and his team were themselves conscious of the possibility that the barracks might have been built entirely in timber initially. However they were unable to find evidence for timber outer walls, except in the case of Building 12, and this caused them to discard the idea. The fragmentary traces described above do not conclusively prove that the barracks in the praetentura were timber-built, though they are, cumulatively, very suggestive. The evidence is essentially based on the anomalous positioning of certain features, rather than a series of clearly recorded stratigraphic relationships. In the case of the slots identified in and around Buildings 5 and 6, in particular, the interpretation outlined above remains only a tentative hypothesis and it is still very uncertain whether they should indeed be attributed to a Period 1 timber barrack or to a subsequent stone barrack. Ultimately, therefore, it is recent discovery of more extensively preserved timber barracks in the retentura which underpins the overall interpretation. Otherwise the anomalous slots would could simply have been dismissed as construction features or cultivation furrows.
The sparsity of features in the eastern half of the praetentura can be explained by the better preservation of the stone barracks there, which limited the scope of the excavators to investigate the underlying levels. This is less true of the western praetentura, where excavation was taken right down to the natural subsoil in substantial areas of Sites 2 and 4, revealing parallel rows of pre-Roman cultivation furrows on a slightly different alignment from that of the fort buildings. Hence it is perhaps somewhat surprising that the layout of the primary timber barrack phase did not emerge more clearly here.