Excavation

Date 31 January 1993 - 28 February 1993

Event ID 633039

Category Recording

Type Excavation

Permalink http://canmore.org.uk/event/633039

Two seasons of rescue excavation by GUARD have now taken place at St Michael's Churchyard, Inveresk, on this Roman auxiliary fort site. Work preceded the extension of the cemetery by East Lothian District Council, into the field immediately outside the cemetery wall at its western extremity. Richmond calculated that the western defences of the fort should run obliquely through this field, aligned roughly NE to SW. Both trenches, the first opened in April and May 1991, the second in March 1993, measured c25m by 20m and were located contiguously just within the putative line of the fort defences.

The state of preservation encountered in both seasons complemented and confirmed observations made by Hanson (Hanson 1984) following exploratory trenches opened in this field in 1981. In general, structural remains associated with the occupation of the fort were found to survive in a reasonable state of repair in the northern half of the 1991 trench; the southern half of that trench by contrast contained only scant remains, and these in a poor state of repair; the 1993 trench was found to be largely devoid of remains, save for a series of pits and several patches of rough cobbling plus one roughly flagged area (see further below). This situation may be related to the differences in topsoil depth recorded across the two trenches, being greatest to the N and E where survival is best, and slightest to W and S where survival is poorest. Consequently one might anticipate reasonably good survival in the northern half of the field, which remains unexplored.

The 1991 excavation uncovered a stretch of some 20m of the via praetoria, as it approached the W gate of the fort, together with its junction with the via sagularis and a distance of some 20m of that road. The well-laid cobbled sub-surface of the via praetoria ran across the full width of the trench, bordered in places by dressed kerb stones. It was c7.50m wide and when sectioned was found to survive only to the depth of one or two layers of cobbles towards the E, and two or three layers to the W. The via sagularis by contrast survived poorly, having been completely robbed out or badly disturbed at several points along the exposed length. Consequently its width could only be estimated, at c5m or perhaps a little more, and its depth, where surviving was no more than that of the visible stones. Both it and the via praetoria were disturbed by a large pit which seems likely to have been cut in antiquity (see below) at the point where the two roads meet. Stratigraphic evidence encountered on the via praetoria suggests at least two phases of road building activity, one layer of road cobbling having been found to run under the in situ kerb stones on both sides of the road.

Evidence for two phases of occupation was also provided by the stratigraphic sequence recovered at the W gate. Although only partially revealed during the excavation, the remainder lying outwith the excavation area, traces of mortared stone construction were uncovered on the S side of the road, presumably representing the foundation levels of a gate structure. Also found were two square-cut pits, one adjacent to and S of the road, the other centrally placed within it. Each was filled with carefully-laid alternate layers of clay and cobbles, and though neither pit lay fully within the excavation trench, their dimensions are likely to have been c1.25m by 1.25m. These presumably represent the foundations of a gate superstructure, providing firm bases in the loose sand subsoil to take the weight of large uprights; a third example presumably lay on the N side of the road, again outside the excavated area. The position of the central example, sealed beneath the undisturbed uppermost surviving layer of road cobbling, adds weight to the contention of more than one phase of occupation.

The fort interior, as exposed in both trenches, proved to be largely empty. The adverse effects of past cultivation were evident through the lines of plough marks visible on the cleaned surface and the scarring detectable on numerous pieces of dislodged building masonry, as well as the kerb stones of the via praetoria. No trace of the presence of negative features indicative of the foundations for wooden barrack blocks or other rectangular-plan buildings was evident on the cleaned surfaces and even deep, cross-site sections cut well into the natural subsoil in both seasons failed to reveal any indication of post holes or post-trenches. It is possible that these may have existed as highly-truncated features, which were missed and removed during cleaning (see also Hanson 1984), though the level at which cleaning took place was commensurate with that of the surviving road. It is perhaps more likely that no wooden structures existed in the areas opened. Certainly, the considerable quantities of masonry debris found across the site suggest that building in stone may have been more prevalent here than has previously been supposed.

What little structural evidence there was within the fort interior is all of debatable provenance. One area exposed in the 1991 trench, close to the S edge, and measuring approximately 3.50m E to W by 2.50m N to S, may represent an occupation surface. Here an irregular patch of stone flagging incorporating two quernstones, and possibly associated with a line of three dressed masonry blocks, was found in a heavily disturbed state. Considerable quantities of Roman material were recovered from this area, but mixed in with modern finds and topsoil; no finds came from securely stratified contexts. It is possible that this relates to a post-Roman site occupation. A similar patch of stone flagging was uncovered at the southern end of the 1993 trench; of similar extent, this too had suffered considerable distubance and insecurely stratified finds ranged in date from Roman to modern.

Evidence for the possibly deliberate destruction/demolition of the fort was presented by a large pit which had been cut through both roads (noted above) and which measured up to 3.60m in diameter, and a maximum of 1.70m deep. The upper half of the fill was packed with Roman masonry, and a few sherds of Roman pottery were recovered from different levels throughout the feature, including one from near the bottom. This, coupled with the lack of any modern material from the pit fill, suggests, though does not prove, that it was dug in antiquity. Several smaller pits, on average between 1.50m and 2m in diameter by on average 1m deep were encountered scattered across the area opened in 1993. Some of these also contained quantities of animal bone. All were discrete features cut into the natural sand and lacked any telling stratigraphic relationships.

Small find recovery was high, as would be expected on such a site, though at least half of the material came from the surface or the topsoil. Preliminary examination of the pottery suggests nothing inconsistent with an Antonine occupation date.

Sponsors: East Lothian District Council, Historic Scotland

A Leslie 1993.