Preparations for the construction of the secondary fort included backfilling the open ditch of the primary fort with compact, light brown clay (078, 248; Fig.3, Sections 2 and 3). The machine section demonstrated that there was little, if any, natural sedimentation present in the ditch beneath this fill. There may also have been some levelling of the site, accounting for the possible absence of the front of the primary rampart. Similarly, the absence of mixed rampart material from the ditch fill may indicate that it was backfilled before the rampart was levelled.
The main structure of this phase was a barrack block with a stone officer's quarters and timber accommodation for the men. To its north was part of the via sagularis and, to the east, the via vicinaria.
This structure comprised clay-bonded dressed stone walls with external dimensions of 10.90m (north-south) by 10.08m (east-west), the eastern wall more-or-less corresponding with the line of the ditch of the primary defences. The siting of the building evidently necessitated some unusual preliminaries to the standard construction process, for the eastern foundations were strengthened. First, a rectangular pit (sloping upwards from east to west) was dug into the backfilled ditch and neighbouring ground surface. The foundations of the building were then built up to ground level against the edge of this pit, and the interior backfilled with mixed material.
The eastern wall had three levels of faced foundation below a levelling course approximating to the Roman ground surface and up to two courses survived above this (Fig.5). Both the north and south walls has foundations which sloped upwards from their respective eastern corners, but both of these used undressed sub-rounded boulders in their lowest courses below dressed masonry.
The main fabric of the walls was the customary wedge-shaped dressed facing stones with a clay and rubble core and clay interstices. The eastern wall (0.52m) was narrower than its northern (0.60m) or southern (0.56-0.60m) neighbours. No sign of locating holes or slots for a timber framework were noted, nor were any doorways identified. In the case of the eastern wall, this was sufficiently well-preserved for any doorways present to have been evident. In the north wall, damage by modern drains towards the centre, and later levelling across the western half (where only the foundation course was present), meant any entrances in this part would not have survived. Traces of burning at one point on the north wall were dated by archaeomagnetism to A.D.230-70 (see Noel below). The burning of the wall fabric was of such a character as to indicate that it had not been standing to its full height at the time of the fire.
Two main internal clay-bonded stone partitions were located within the building. An east-west wall (140), 0.50m broad, divided the building in two, although much of the western part was removed by the later levelling. In the central area, the lowest course of stones were only evident as a series of impressions in the pinkish-brown clay (180) used in the shallow foundation trench (177) for the partition. A further partition (145) - 0.50m broad, 2.80m long, and terminating 1.20m short of the inner face of the north wall - ran north from this, effectively dividing the building into one large room to the south and two smaller ones to the north. Above the fill of the east-west partition wall construction trench, but apparently contemporary with the north-south partition wall, was a hard, dark clay deposit (155) containing much of a BB1 bowl (see Dore below, No.25), sealed by the subsequent floor makeup.
Within the north-eastern room thus created, there was a slab-lined and covered drain against the inside face of the eastern wall. This flowed southwards, beneath partition 140, and then turned south-westwards, diagonally across the southern room. It had been removed by later disturbance in the western half of the building. The drain had been cut (109) into the makeup material (111), lining slabs placed vertically within it (the bottom was not lined) and cover slabs (106) placed over the top. However, the slabs then stood proud of the surface of the room, so a floor of yellowish clay (107) was laid at least to the same level as the tops of these slabs. There was a slight enlargement of this drain at its northern terminal, but no sign of any associated structures. Within the north-eastern room, a series of extremely thin, yellowy sandy clay layers with decayed sandstone chips were found, interleaved with darker sandy layers.
The surfaces within the north-western room had largely been removed by later disturbance, but an east-west section across the room revealed a dark, purplish layer with extensive traces of burning (159), beneath a hard-packed, yellowy clay with sand and charcoal inclusions (165). The latter closely resembled the floor surfaces in the north-eastern room.
Excavation of the lowest levels of the southern room, which lay below the depth threatened by the foundations of the proposed building, was not attempted, although the course of the drain was identified in O'Brien's trial trench and this was re-examined during the 1990-91 excavations, identifying a series of slab-paved floors (174-6) contemporary with, and immediately above, the drain cover-slabs.
At a number of places around the inner face of the main external wall, and also around the partition walls, fairly small flat slabs (usually less than 200mm square) were noted resting vertically against the wall face, embedded within flooring material.
The contubernium area of the first barracks was not fully excavated for similar reasons, but it was clearly of posthole construction and separated from the officer's quarters by a narrow alleyway floored with a single line of stone slabs (259). Reflecting the original Roman ground surface, these slabs sloped upwards from east to west (a slope of around 9%) and were removed by subsequent levelling in the western half. At least four postholes (264, 268-70) that belonged to the north wall of the contubernium area were identified.
Two hearths of this phase were also identified. Hearth 225 was associated with a spread of black ashy loam (226) and a fragmentary floor of slabs (202) which covered an area of 4.80 by 1.20m.
A partition between the first and second contubernia may have been indicated by a marked linearity of rubble running east-west (203, 205), including a pivot stone (see Coulston below) within it, and with a parallel band of yellow clay (206).
Immediately to the north of the officer's quarters, and at the extreme northern end of the site, a small portion of the earliest surface of the via sagularis was discovered and found to slope upwards from east to west, with an incline of around 9%. It too had been truncated by later levelling. The surface (151), composed of small cobbles, butted directly against the north face of the north wall of the barrack.
The via vicinaria was more adequately explored. The primary surface (151, 267, 337) had a marked camber and appeared to rest directly on the underlying material (partly rampart, partly natural) with no foundation. It was apparently bordered on its western side by a kerb of rounded river boulders sunk into the ground.
The fill of Rainbird's Trench 5 produced two stone gutter blocks (see Coulston below) which presumably came from the eavesdrip of the officer's quarters.
Before construction of the secondary fort, careful attention was paid to backfilling the ditch of its predessor, packing it with clay. Moreover, the foundations of the officer's quarters show that the possibility of subsidence had been anticipated and the minimal amount of slumping visible in the walls even now attests the success of this operation.
If correct, the archaeomagnetic date for the burning of part of the north wall of the officer's quarters might be indicative of a bonfire lit during the construction process, since it could not have burned both sides of the wall (and its core) after it was built, and would be much too early for any later activity on the wall once demolished. Lack of wear on any of the burned stones, together with the extent of the burning, makes it unlikely that this was a fire lit in a doorway.
The combination of a stone officer's quarters and timber contubernia is unusual, but not without parallel. Likewise, other examples of barracks with separate officer's quarters are known (Davison 1989, 82 notes examples from Strageath, Maryport, Watercrook, Caernarvon, and Hod Hill). In the case of Chester-le-Street, the alleyway may have afforded access to an entrance in the south wall.
It has become common to speak of clay-bonded stone walls as footings for timber-framed superstructures, but there are a number of reasons why this may not have been the case for the Church Chare building. Since the east wall appears to survive for at least three courses above Roman ground level and no evidence of locating holes for a timber framework could be found within its fabric, the means of attaching the superstructure is lacking. The whole question of clay-bonded Roman military architecture is poorly served in the literature and its capabilities generally under-estimated. Parallels in more recent vernacular architecture suggest that full-height, load-bearing walls would present no insurmountable structural problems (cf. Fenton and Walker 1981, 57) and that, providing the integrity of the fabric was maintained by effective rendering, a serviceable and durable building could be produced. Indeed, a collapsed, 14-course-high, clay-bonded stone barrack wall, recently found at South Shields (and of similar width to the Church Chare officer's quarters), would seem to support this interpretation (P.T. Bidwell, pers. comm.). The widespread use of clay-bonded walling in military buildings in Britain indicates that the Romans seem to have thought it as valid a constructional technique as 'conventional' mortared walls.
The small, flat, upended slabs noted against the walls may have been connected with the required rendering (perhaps as a guide to the preferred thickness?). Similar uprights have been noted in recent excavations in the north-west corner tower at Birdoswald (T. Wilmott, pers. comm.).
With regard to the men's quarters, it is possible that the positioning of the hearths may offer some indication of the size of the contubernia, although a suggested width of 4.60m (15.5Rft) seems on the large side by comparison with other barracks (cf Niederbieber at 4.60 x 5.40m and 4.60 x 6.25m; Heidenheim at 4.50 x 4.50m; Aalen at 4.50 x 4.38 - ibid. Table XIX - the first a numerus fort, the other two for alae).
The unusual width of the via vicinaria can either be interpreted as being due to the absence of a twin building for the excavated barrack block, or simply the fact that the barracks within the secondary fort were widely separated.