Verulamium is thought to have lain within a postulated sub-Roman enclave centred on London and is of particular interest in being referred to as preserving elements of organised civic life as late as 429, when St Germanus visited Britain. Evidence for this sub-Roman continuation of town life has been encountered in several excavations and in particular that of Frere on Bluehouse Hill during the 1950s in advance of road construction (Frere 1983). A building at the south-east corner of Insula XXVII had been burnt down during a widespread town fire in 155-160: the building plot was thought by Frere to have remained abandoned for two centuries before construction of a substantial house with mosaic pavements during the late 4th century. This house was enlarged and altered before going out of use c. 430-440. A large stone barn with wall buttresses, measuring 43.40 by 17.00m was then constructed over its levelled remains and was thought by the excavator to have stood until c. 450-70. After this barn was demolished, a wooden pipeline of Roman type with iron collars was laid across its site, running south from the direction of the theatre at Verulamium: this pipe was thought to be no earlier than the late 5th century.
Frere's late dating of the house has recently been questioned because of the Antonine style of its mosaics and the conclusion drawn that it was constructed soon after the fire (Faulkner and Neal 2009). I re-examined Frere's pottery from the various phases of occupation some years ago and came to a similar conclusion.
The coarse pottery from the make-up for Building XXVII, 2, lacks anything that need be later than the third quarter of the 2nd-century. Frere cites the presence of fragments from two (BB1) beaded-and-flanged dishes in the chalk floor below Room 3 and dates them to c. 350-400 on the strength of Gillam's dates for his similar form 315 (Frere 1983, 212-14). This latter form (Gillam 1970) is, however, undecorated whereas Frere's examples are decorated with acute latticing. My work on the (BB1) pottery from the late Roman kilns at Bestwall Quarry, Wareham, and elsewhere (Lyne 1994 and 2012a) shows that the beaded-and-flanged dish in that fabric has a very long history, appearing c. 150, long before the appearance of the developed beaded-and-flanged bowl. Acute latticing on (BB1) open forms, including the beaded-and-flanged dish, disappears c.200 and is replaced by burnished arcading. The form itself changes little, apart from its flange becoming stubbier, until its disappearance during the early 5th century. These fragments from two dishes were, therefore, probably deposited in the floor during the period c. 150-200.
The only pottery assemblages of any consequence associated with occupation within the house come from a corn-dryer inserted during its third phase, dated by Frere to c. 400-435 (Frere 1983, 223-24). The substantial pottery assemblages from the fills of the corn-dryer include much fresh material and are of 3rd-century date. The back-filling of the corn-dryer seems to have taken place over a period of time, in that there is little or no (BB1) present in the pottery assemblages from the various tips until we reach the uppermost fill (Context X XXIII, 3). This yielded fresh fragments from a beaded-and-flanged bowl of Bestwall form 6/4 and three straight-sided dishes of forms 8/3 and 8/5 (Lyne 2012a, c. 240-290/300, 200-270 and 220-300 respectively). This evidence all points to the house being built during the late 2nd century and its three recognised construction phases being over by c. 300. The coin evidence does, however, indicate refloorings and other modifications taking place throughout the 4th and into the 5th century. What this means is that the later barn structure could easily be early 5th rather than mid- to late 5th century in date.
The theatre was excavated by Kenyon in 1933-34 (1935) and was thought to have been abandoned towards the end of the 4th century before being used as a rubbish tip. This author decided to restrict re-examination of the pottery to that from the 1933 season of excavation as the precise subdivision of the theatre fills into layers was abandoned during the following season (Lyne 1994, 206-9). Mixed 4th-century coinage, up to and including Theodosian issues, occurred in large quantities throughout the deposition of the layers of rubbish.
There are three copper-alloy nummi post-dating 402-3 from Verulamium (Abdy and Williams 2006). They are a Victoria Augg issue of Honorius (421-3) and two Victoria Augg issues of Valentinian III (425-435), suggesting that a full monetary economy was still functioning there well into the second quarter of the 5th century and perhaps later. We may, therefore, have to re-date the use of the theatre as a rubbish tip to the mid-late 5th century.
What does the pottery in the tips tell us? It can be seen that there was a steady increase in the significance of open forms in relation to cooking-pots throughout the successive phases of dumping. This could indicate increasing difficulty in getting new pots, in that cooking-pots would become fouled and unusable owing to food residues absorbed into their fabrics, whereas bowls and dishes used for eating and often with area-burnished interiors would last longer.
Most of the pottery in the theatre dumps comes from the Much Hadham kilns in east Hertfordshire and indicates that they were still producing pottery well into the 5th century. Local sandy black wares are also present and include a distinctive straight-sided bowl type with its top simply bent over to form a weakly flanged rim. An example of this distinctive type is also present in the pottery assemblage from the early 5th century squatting in the Billingsgate bath-house at London. The uppermost dumping in the Verulamium theatre has much higher percentages of Oxfordshire (OXRS) and other finewares than the lower ones and suggests total pottery loss, with fine pottery lasting longer than kitchen wares, as was the case at Richborough and elsewhere.
The evidence for 5th-century occupation within the walls of Verulamium has been reviewed by Niblett et al. 2006. They see the 5th-century settlement as comprising simple wooden houses concentrated in the core of the Roman town and interspersed with more substantial patched-up 4th-century stone buildings. The town does not appear to have had a violent end and there were still people living in the ruins as late as the 12th century.