Excavations revealed evidence of the fort's defenses and also showed that it was considerably larger than the Cornish forts of Nanstallon and Restormel.

A view from the outer rampart across the two v-shaped ditches and inner rampart to the interior of the fort

Calstock Roman fort

Dr Chris Smart, Professor Stephen Rippon and Dr Peter Claughton

Geophysical survey carried out by Chris Smart and Dr Peter Claughton, Research Fellows in the Department of Archaeology at the University of Exeter as part of the Bere Ferrers Project (directed by Professor Stephen Rippon) has led to the discovery and excavation of a previously unknown Roman fort. The site, situated on a spur above the River Tamar on the border between Devon and Cornwall, is adjacent to St. Andrew’s church in the parish of Calstock.

The Bere Ferrers Project has, over the past two years, been investigating the development of the royal silver mines which were worked on the Devon bank of the Tamar between 1292 and the mid 16th century. While the mining of ore was restricted to the localised deposits in Bere Ferrers parish, the processing of that ore was carried out away from the mines close to woodland which supplied fuel for smelting and refining. Documentary evidence suggested that smelting was carried out near to the church of St. Andrew in the parish of Calstock on the Cornish side of the Tamar when woods there were allocated to the mines as there are references to smelting and refining at a curia ‘next to the church’ and vetus castrum de Calistok (‘at the old castle/fort at Calstock’). It is known that this curia was an earthen-walled and gated administrative centre containing the King’s hall and the refinery, although previously the vetus castrum de Calistok has been interpreted as the Iron Age hillfort in the north of the parish.

In the summer of 2006 a high-resolution resistivity survey was conducted on Church Hill to the south of the new burial ground in search of the medieval activity. Faint traces of the banks and ditches of an enclosure were identified and were interpreted as either the site of the curia, or an earlier prehistoric hillslope enclosure.

In the autumn of 2007, a magnetometer survey was conducted of a larger area with the specific goal of identifying the locations of furnaces associated with the medieval industry. This survey gave greater clarity compared to the resistivity results, and revealed the distinctive outline of a Roman fort enclosed by two ramparts and two ditches,. Significantly, a number of anomalies also hinted at the potential sites of smelting furnaces known as ‘boles’. Our sponsors, The Leverhulme Trust, kindly agreed to fund an archaeological evaluation of the site, led by Chris Smart, over two weeks in January 2008. This evaluation had two key aims: first to date the construction of the fort and second to examine the possible medieval smelting bole.

A single evaluation trench measuring 2m by 45m was positioned to expose a section across the fort’s defences as well as the potential smelting bole. As the field in which the excavation took place was under permanent pasture and had not been ploughed in the recent past, it was expected that the preservation of in situ archaeology would be good. Overall, the fort measures c.170m by 160m, with substantial defences leaving an internal area of c.140m by 130m (1.82 hectares). Comparison with the internal areas of the two other Cornish forts, Nanstallon (0.72 hectares) and Restormel (0.42 hectares), demonstrates that the fort at Calstock was constructed on a much larger scale. The trench was cut a short distance into the interior of the fort, revealing a possible oven and stoking chamber. Its position, against the inner rampart, would have minimised the risk to the timber buildings, and is a typical feature of Roman forts. Pottery of various types was recovered from the oven and overlying deposits, and may give an indication of the length of use of the site.

Excavation revealed an inner rampart approximately 5m wide, constructed of clay and shillet cast from the digging of the first ditch, and revetted with timbers on both faces. The outermost revetment was constructed of worked timbers 30cm square, suggesting that they may have formed the basis for a timber super-structure or palisade. A band of blue-grey clay has been interpreted as the degraded remnants of a stack of turves which had been piled upon the natural geology abutting the timber revetment. The two ditches which sat between the inner and outer rampart both had a characteristic v-shaped profile with a square-cut base, typical of Roman military sites. Each measured 2.8m deep from the modern land surface and was approximately 3.5m wide. The outer rampart was also approximately 5m wide and appears to have been constructed of the same intermixed deposit of clay and shillet. Geophysical survey and excavation shows that the outer rampart on the western and southern sides of the fort was capped with large angular and sub-angular sandstone rubble. The date and purpose of this stone capping has yet to be determined. Acid soil conditions will have limited the survival of animal bone but sherds of pottery, including Samian, were recovered. Significantly, what have been tentatively identified as pieces of mineralised vein material were recovered from one of the lower fills of the outer ditch, indicating that mined material had been deposited there. This material, which has yet to be examined in detail, is perhaps the first certain evidence of a Roman military interest in Cornwall’s rich mineral resources.

Investigation of the geophysical anomaly to the south of the fort defences may corroborate this hypothesis. Prior to excavation it was conceived that this may be the location of a medieval smelting bole referred to in fourteenth century accounts. The feature presented itself as a square or rectangular (it was not fully revealed) stone-lined furnace. No slag was recovered from the feature but its fill has been sampled and is awaiting further analysis. The furnace sat within a deposit of intermixed heat-affected natural, charcoal and furnace-lining that extended for approximately 10m south of the outer rampart. Samian pottery was also found in this deposit. Furnace-lining and unabraded Roman pottery, including a sherd of stamped Samian, was also found in a shallow pit at the southern end of this ‘working area’. Radiocarbon dating will hopefully confirm whether the furnace is Roman or medieval, although, at this stage it appears that it was a focus of Roman industrial activity.

Post-excavation analysis and preparation of the written report is now underway. It is anticipated that full details of the excavation will be published in the national journal of Roman archaeology Britannia, as well as the county journal Cornish Archaeology.

Back to Research projects