4. Sites

4.1 Dorchester

Type 18 vessels have been recovered from a number of sites within and around the Roman town of Durnovaria (Dorchester). At Greyhound Yard two Type 18 vessels came from late Roman shafts (Seager Smith and Davies 1993, figs 149.309 and 151.350) (Figs 2 and 3) and also from debris layers overlying late Roman structures at Colliton Park (Aitken and Aitken 1982, 121) (Figure 4) but associated coins were lacking. More usefully, Type 18 vessels were identified in the latest assemblages of pottery from the still unpublished Dorchester bath-house excavation in association with coins of AD 388-402 (Andrews forthcoming).

Figure 2: Type 18 vessel from Greyhound Yard, Dorchester (after Seager Smith and Davies 1993, fig. 123)

Figure 3: Type 18 vessel from Greyhound Yard, Dorchester (after Seager Smith and Davies 1993, fig. 123)

Figure 4: Type 18 vessel from Colliton Park, Dorchester (after Aitken and Aitken 1982, fig. 17.6)

The most important locations within the walls of Durnovaria for understanding the chronology of Type 18 vessels are probably the Dorchester Hospital sites. The first excavations were carried out in 1969 and reported on by Patrick-Greene (1993). One of the Type 18 vessels mentioned in that report (Patrick-Greene 1993, fig. 19.32) was found in a late 4th-century demolition deposit (Gerrard 2004, table 8.1). Of greater significance is a vessel from this site that was not discussed in my earlier paper (Gerrard 2004) (Figure 5). A Type 18 bowl was found in Pit 6, which clearly cut an internal wall in Building 1 (Patrick-Greene 1993, 87). This pit also contained a white-painted OXRS bowl of Young's (1977) Type C51.1 (AD 350-400+) and 'several late 4th century coins' (Patrick-Greene 1993, 87). The published table of coins (Patrick Greene 1993, 93) does not list any from Pit 6 but does list late 4th-century coins from Pit 5 (Patrick-Greene 1993, 87). Examination of the site notebook (Richard Breward pers. comm.) demonstrates that the published table is in error and it ought to read 'Pit 6' instead of 'Pit 5'. Pit 6 thus contained the Type 18 bowl along with coins of the House of Valentinian (AD 364-378) and Theodosius (AD 388-402) and post-dates the abandonment of the late Roman Building 1.

Figure 5: Type 18 vessel from Dorchester Hospital, Dorchester (after Patrick-Greene 1993, fig. 19.32)

The 1969 excavations also recovered a dispersed hoard of Theodosian coins (AD 388-402) and some late 4th- or early 5th-century belt fittings (Patrick-Greene 1993). Given this evidence it is unsurprising that the redevelopment of the site in 2000-2001 also uncovered significant traces of late Roman activity (Trevarthen 2008), including a large aisled structure interpreted as a barn (Building 12). A feature internal to this structure apparently contained a coin of AD 388-402, suggesting that it was being modified at the very end of the 4th century or into the 5th century (Cooke 2007, 65). An oven within the building contained fragments of a Type 18 bowl (Trevarthen 2008, 40 and fig. 76) and a dark soil layer sealed the demolition or collapse of the building. This dark layer contained 997 fresh sherds of pottery and 409 copper-alloy coins, 223 of which were dated AD 388-402. The pottery assemblage had typical late Roman BB1 vessel forms (Types 2, 3, 20 and 25) as well as Type 18 vessels and Type 12 jars in South East Dorset Orange Wiped Ware (SEDOWW) (Trevarthen 2008, 40; Gerrard 2010). A 5th-century date for the use or deposition of these vessels seems certain.

Figure 6: Type 18 vessel from Poundbury, Dorchester (after Patrick-Greene 1993, fig. 19.32)

Poundbury on the outskirts of Dorchester is famous as the site of a large late Roman cemetery and a univallate Iron Age hillfort (Sparey-Green 1987). It is also the location of an important 5th- and 6th-century settlement and some Roman-period settlement features too. At Site G the skeleton in Grave 1411 contained a coin of AD 321 in its mouth and a coin of AD 364-378 in the gravefill. Nearby was a stakehole that was also found to contain a coin of AD 364-378 (Sparey-Green 1987, 63). This stakehole was overlain by a cobbled surface that was in turn cut by a ditch that also cut Grave 1411. The cobbles were also apparently cut by a pit (G1104) containing 47 sherds (1.433kg) of BB1. The large sherd size is indicative of the fresh and unabraded nature of the pottery in this feature, which included typical late Roman BB1 bowl Types 20 and 25 alongside large fragments from two fresh Type 18 vessels (Figure 6) (Sparey-Green 1987, fig. 88.41). A terminus post quem of AD 364-378 seems appropriate for this pit and its contents.