In 2007, NAA was commissioned to carry out a programme of archaeological investigation and recording during construction of a new golf course at Hurworth-on-Tees near Darlington, County Durham. The site consisted of six fields extending across 113ha on a floodplain surrounded to the west, south and east by a loop of the River Tees. Most of the area in the southern part of the site was covered in deposits of alluvium resulting from repeated flooding by the river and was, unsurprisingly, devoid of evidence for past settlement, although some agricultural features were recorded.

There was already some evidence for Roman activity in this part of the Tees Valley, including a settlement (probably a villa) near Dalton-on-Tees on the opposite bank of the river, a possible fort recorded from aerial photographs at Hilltop Farm to the north of the site, and a Roman coin found at Hurworth Grange. A Roman sarcophagus in the cellar of Rockliffe Hall, immediately to the north of the site, may have been brought here from elsewhere. Despite this evidence, the site revealed at Rockliffe Park was completely unexpected. Archaeological features were mainly restricted to one field towards the northern edge of the site on slightly higher ground nearest Hurworth and adjacent to Rockliffe Farm. This area contained significant remains of a Romano-British rural settlement dating from the 3rd and 4th centuries AD.

In the area containing the Roman settlement, partial excavation was undertaken in three areas where soil-stripping had exposed archaeological remains and their preservation would be compromised by the development plans. The remaining areas were protected with geotextile and retained undisturbed within the new golf course. One consequence of this methodology was that there was no physical, stratigraphic link between the three investigated areas, which compromised interpretation of the results. To the south, Area A was 60m by 50m, the central Area B was 70m by 15m, and Area C at the northern edge of the site was 45m by 40m.

All three areas contained linear and curving ditches defining parts of a series of enclosures. Extensive recutting and intercutting showed that these enclosures had been repeatedly remodelled. Dateable pottery found discarded in the ditches, and a small number of associated pits, showed that the enclosures were mostly of 3rd to 4th century date.

In Area A, we also found three crop-drying kilns, two heavily truncated although the third better preserved example had a classic Roman T-shaped form. A circular stokehole filled with charcoal led to clay- and stone-lined flues, with a stone wall at the cross of the T which would have supported a drying platform. Carbonised grain recovered from the flue shows that the kiln had been used to dry crops of barley, emmer or spelt wheat, naked wheat and oats. Two elongated pits found in Area B were filled with burnt clay and heat-affected stones. These were tentatively interpreted as having been quenching troughs associated with metalworking, the burnt deposits representing in-filling with waste.

Area C in the northern (and highest) part of the site area was the only area where convincing structural remains were identified. The heavily truncated remains of another crop-drying kiln were sealed by a floor surface constructed of large limestone slabs. Several postholes around its perimeter presumably represented the walls of a timber superstructure; a posthole and beam-slot cut into the surface indicated the position of interior partitions. The surface was covered in a layer containing a large quantity of later 4th-century pottery and a broken millstone. To the south of the structure, passing from east to west, was a cobbled pathway that had possibly been flanked by a wooden fence represented by several postholes.

The excavations produced a large assemblage of Roman pottery, all dating from the mid-3rd to 4th century. Other finds were also mostly utilitarian. A ceramic spindle-whorl would have been used in processing wool or other fibres, while a broken saddle-quern had been reused as part of a floor surface. The broken millstone found in the Area C building was of a size and type probably used in an animal-powered mill and might suggest a function for the structure.

Against this apparent background of agricultural processing within the excavated area were several more intriguing finds. Two fragments from an iron stylus indicate literacy (and perhaps a high social status) among the inhabitants. Seven fragments of Roman tile, mostly roofing materials but including pieces of box-flue tile, hint at a high-status building somewhere in the vicinity. The most spectacular find was a large copper-alloy dish that had been formed by spinning on a lathe and then tin-plated on the inside. This piece would have been tableware in a high-status or wealthy household. These vessels are very rare finds, the only complete bowl of a similar design is in Germany, although fragments of similar pieces have been found in Britain at Fishbourne and Colchester. Another curving piece of copper-alloy found during the excavation at Rockliffe Park may be a fragment of a second vessel.

The Roman archaeological remains identified during construction of the new golf course were, in general, reminiscent of farmstead sites of this period found throughout the Roman province. However, there were strong hints at Rockliffe Park that the complex of enclosures may have lain at the periphery of a much higher-status site as indicated by the ceramic building materials and copper-alloy dish which would have been an expensive and uncommon item in rural Roman Britain. It seems quite likely that there is a villa waiting to be discovered somewhere in the vicinity.

A report on the full results of the project can be downloaded for free from the Archaeology Data Service website:

NAA (2009) Rockliffe Park, Hurworth-on-Tees: Post-Excavation Report; Northern Archaeological Associates Report 09/03. https://doi.org/10.5284/1030293