The identification of the western rampart and one ditch of the primary fort serve to modify existing notions about the positioning of this base. The excavations in Middle Chare (Evans et al 1991, 16, Fig.5) located timber buildings beneath those of the stone fort and sought to explain the stratigraphical sequence of the rampart as if its earliest phase belonged to a primary, timber fort. However, it now seems clear that the area examined in 1978 lay outside the defences of the primary fort, in which case the buildings located may have belonged to a vicus complex or annexe.
Both the coin and pottery evidence point to the primary fort at Chester-le-Street belonging to the second half of the 2nd century A.D. (King, using the samian, asserts after A.D.175, in Evans et al 1991, 45), presumably post-dating the abandonment of the Antonine Wall, perhaps related to the construction of the fort at Newcastle, since it would help secure the lines of communication to the latter. It is, however, curious that the mid-Antonine forts at Newcastle and South Shields should be built in stone, but Chester-le-Street, their contemporary, in turf and timber.
The excavation at Church Chare has served to demonstrate that the retentural barracks in the north-west corner of the secondary fort were laid out per strigas (cf Davison 1989, Fig.8,B, C, and H).
The date of the construction of the secondary fort is not easily determined. The inscription referring to construction work in A.D.216 (RIB 1049) need not even have belonged to the secondary fort, as there are Antonine parallels in Britain for forts with turf-and-timber defences and stone internal buildings (cf Johnson 1983, 274-9), so it could have been associated with additions to the primary fort. However, the archaeomagnetic date for the burning on the north wall of the officer's quarters, which has here been presumed to be related to the construction of this phase, falls within a range of A.D.230-70. Pottery from the ditch fill (see Dore above, Nos. 52-6) indicates a construction date in the later part of this range. Likewise, the presence of a little-worn antoninianus of Victorinus/Tetricus on the floor surface of the officer's quarters would seem to accord with occupation during the second half of the 3rd century A.D.
The fact of the construction of a new fort probably betokens the arrival of a new garrison and it may even be indicative of a new type of unit occupying the site, since the old fort was demolished (apparently immediately before the secondary one was begun): garrison changes in the secondary fort at Corbridge were apparently marked by the replacement of old buildings but the retention of the former defences (Bishop and Dore 1989, 88), the implication being that the available area within the fort was thought adequate. The per scamnum alignment of the praetentural stone granary in the secondary fort is unusual, however, and it is conceivable that it may have been retained from the primary fort (I am grateful to Mr P.T. Bidwell for this suggestion).
If the barrack was allowed to decay at the end of Phase 2, this may have been because the whole fort was abandoned, although this seems unlikely, given the Roman practice of systematic demolition of abandoned military establishments. It would be more likely, as elsewhere (cf. Boon 1972, 56), that the centuria or turma that occupied it were removed, although the main body of the unit remained, leaving the building vacant. Its return, or perhaps a change of garrison, might have been the occasion for the rebuild that characterises Phase 3. The fact that the timber contubernia were also rebuilt may be a hint that it was not a change of garrison, if constructional styles are in any way indicative of individual unit preferences.
Termini post quem for the beginning of this phase are provided by mid- 3rd century pottery from beneath the flagged floor of the officer's quarters and by the coin from the Phase 2 floor surface. There is an interesting implication here: taking the archaeomagnetic date from Phase 2 together with the pottery from Phase 3, it might be argued that there could have been a fairly short initial occupation of the barrack before its being abandoned and allowed to decay. Moreover, Phase 3 need not necessarily correspond with changes elsewhere within the fort.
The function of the Church Chare barrack seems to have changed: with the contubernia defunct and only the officer's quarters retained, apparently unpartitioned, the building now served a different purpose. If Phases 2 and 3 belonged to occupation by the same unit, then it might be speculated that Phase 4 witnessed the arrival of the Notitia's numerus vigilum. Hodgson (1991, 90) argues that they could have been late arrivals and may not have been the first numerus at Chester-le-Street, but this interpretation is dependent upon a number of assumptions, not least that the N CON on tile stamps at Binchester (Richmond and Crawford 1949, 29) really does represent an otherwise-unattested n(umerus) Con(cangiensium). It may even be that the numerus vigilum and numerus Concangiensium were one and the same unit, perhaps split between two bases.
The stamped tile, with its possible reference to that numerus vigilum, was associated with late 3rd century pottery. Late 3rd century material also came from at least one demolition pit and was associated with tumble from the walls of the officer's quarters. The coins, on the other hand, seem to suggest continued occupation to the middle of the 4th century, a conclusion supported by appropriate pottery from a few (unfortunately not very diagnostic) contexts.
The complete absence of evidence relating to the early ecclesiastical occupation of Chester-le-Street was disappointing, but not surprising, given the amount of disturbance to the site in subsequent periods. A solitary medieval coin and a few sherds of pottery are the only evidence of activity between the Roman period and the 16th century.
The archive of the 1990-91 Church Chare excavations, which has been deposited with the Bowes Museum, includes the following specialist reports:
Thanks are due to the diggers, supervisory staff (Graham Young, Colin Lofthouse, Mick McElvaney, Alan Teasdale), and specialists who have contributed to this report. Alan Rushworth provided some useful discussion on the numerus vigilum. Special mention must be made of Stewart MacPherson who helped with the preparation of archive plans, and Graeme Stobbs, who produced the plans and small find drawings for this report. The project was overseen by Colm O'Brien of the Archaeological Practice, David Stocker of English Heritage, and Mary Lakin and John Pickin of The Bowes Museum. Practical (and enthusiastic) assistance for the project was provided by Mrs Doreen Yates and the ladies of the Parish Centre, Jim Crow, John Dawson, Durham County Fire Brigade, and the Rev. G. Walker and the Parochial Church Council. Paul Bidwell and Colm O'Brien very kindly read and commented upon draft versions of this report, although any errors or omissions necessarily remain the responsibility of the main author.