Frocester, probably the 'ceaster by the Frome' (Smith 1964), Gloucester, Cirencester and Woodchester are the only Gloucestershire place-names to incorporate the Old English element 'ceaster'. The use of this term, applied by the Anglo-Saxons to prominent remains of Romano-British settlement, implies that those who named Frocester, its Stanborough 'stone fortified enclosure' or 'mound' (of villa ruins) Field, and Borghullsweye, the 'way to the fort hill' (possibly the high point on which stood a Romano-British complex (SITE 5), the remains of which underlay the old parish church of St Peter's) must have seen evidence of substantial Roman buildings. No tradition of them survived, and Frocester did not rate an entry in Witts Archaeological Handbook of Gloucestershire (1883) despite the recognition, in the late eighteenth century (Lysons 1797) and in Kelly's Directory of 1856, of Frocester as the site of a Roman settlement. Both the 1856 reference and a later note by St Clair Baddeley (1931) were probably based on finds made during the mid 19th century rebuilding of the parish church. When this was demolished in 1952, it became clear that it had been built in part on the footings of a major Roman building (Gracie 1963).
In 1959, Eddie Price noted an extensive scatter of Romano-British occupation-debris in Big Stanborough, a field on his farm. He brought it to the attention of Capt H S Gracie, who in 1961 followed up work at St Peter's (SITE 5) with what was intended to be a small excavation to confirm the existence of this second, previously unknown Frocester site (SITE 1). The nature of the discoveries and the evidence of attrition by the modern plough encouraged him to continue the excavation. This he did for eighteen successive seasons, until his death in February 1979. From then to the present day the research has been directed by Eddie, who had worked alongside Capt Gracie from the beginning.
As with all sites found in Frocester,SITE 1 does not exhibit crop marks and is invisible from the air, probably because of the nature of the soil, but it is now known to cover approximately 7 acres (2.8ha). Half of it has been excavated, and additional exploratory trenches have helped to determine its full extent. The published interim reports (Gracie 1970, Gracie & Price 1979, and Price 1983) have been revised and brought up to date for incorporation in this report.
Small scale investigations on other sites, mostly on the farm, together with a programme of systematic fieldwalking over the rest of the parish, have resulted in imformation to supplement that from SITE 1.