It appears that only York, Caerwent, Carmarthen, Aldborough, and Wroxeter have produced no definite pre-Roman evidence as yet. Aldborough, Caerwent, and Wroxeter have produced Iron Age coins, but these may have been post-conquest arrivals. Some of the sites lay outside the area of habitual coin use in the late pre-Roman Iron Age (Haselgrove 1987), so it may be that the locations were visited, used, or important in some way in prehistory, but this is more difficult to prove with certainty; excavations at these sites have tended to not go below the first - and second-century layers.
All of the towns except Caerwent were located next to rivers, flood plains or marshland areas, crossing points, and sometimes islands, indicating that water, movement, communication, and experience were important factors in the significance of the places. Springs were also features at a number of the sites. At Caerwent, the small Nedern Brook flowed around 250-300 m to the south of the Roman town (Brewer 1993: 56). The various watery contexts will have had different practical values: marshland for food and material resources and grazing, springs for sources of clean water, and rivers for resources and transport (cf. Rippon 2006). These uses will have remained important throughout the Roman period, demonstrating an aspect of the continued functioning of the sites. However, they will also have been associated with other meanings, adding considerable significance to the experience of these landscapes. A number of sites consisted of more than one focus
Of activity with different areas of apparent importance, including earthworks, enclosures, and industrial activity. The ways in which people moved in and negotiated these places constituted a major part of the meaning with which they were imbued. The topographies in which the towns and public buildings were constructed were already highly ritualised. Towns appropriated these landscapes and were, in some respect, shaped by them.
There is evidence of metalworking at several of the sites including moulds that were probably used in coin production, whereas in the case of Cirencester there is evidence of coin production relatively nearby at Bagendon in a marshy context.2 In some cases there is less evidence of late Iron Age material than in others. J. D. Hill (2007: 30) has highlighted the massive increase in material culture in the late Iron Age, especially at oppida, compared with earlier settlements but this need not necessarily mean that sites that lack similar quantities of material culture were regarded as inferior; instead this lack of surviving material may have related to cultural preference and values.
4.3 Dating
Dating is important when we consider these sites. Several scholars have emphasised that most of the oppida were late constructions - the late first century b. c. and continuing into the early first century a. d. - and are on sites that do not exhibit much evidence of earlier occupation (e. g., Creighton 2006; Moore 2006). This need not necessarily mean, however, that the sites were not being used prior to the construction of earthworks. At some oppida in northern France, for instance, it has been suggested that the earthworks represent the monumentalisation of places with earlier significance where activities such as meetings and ceremonies took place (Metzler, Meniel, and Gaeng 2006). This would be difficult to identify with certainty in the archaeological record, but features such as rivers and marshland that attracted religious veneration indicate that at least some of these sites were meaningful before the construction of earthworks.
The passage by Strabo (IV.5.2), written at the end of the first century b. c., and describing pre-Roman settlement in Britain as spaces in forests made for huts and animals but not 'with the purpose of staying a long time’, might be useful here (though he was probably drawing on what Caesar recorded). It is likely that there will have been some seasonal movements with cattle and sheep, according to the agricultural year, which were important activities in the landscape, and people will have met other communities for feasts and ceremonies at settlements or 'empty areas’. Moore (2006), for instance, has demonstrated that Bagendon was later in date than the surrounding settlement system. It was not part of earlier settlement but it may already have been important as a luminal watery place used for meetings, exchange, and religious activity. Over time, some places of these periodic meetings, markets, and rituals perhaps became more permanent and monumentalised. Meanings attached to these places will have gradually accumulated and continued into the Roman period.
The Bagendon earthworks near Cirencester were not established until the early first century a. d. (Trow 1990: iii). At Verlamion (Verulamium), construction of the dykes did not begin until about the mid-first century b. c., with the earliest possibly at Wheathampstead,
2 There has been debate (as there has with the role of the coinage itself) about whether these moulds or trays represent coin production or simply the working of precious metals (e. g., Haselgrove 1987: 28-9; Niblett 2001: 42-3; Tournaire et al. 1982: 429-32). Either would indicate activity of considerable importance.
And the building of these earthworks continued into the early first century a. d., together with the enclosures at St. Michael’s and Gorhambury (Haselgrove and Millett 1997: 284; Neal, Wardle, and Hunn 1990; I. Thompson 2005: 27-32). The majority of the dykes at pre-Roman Camulodunum (Camulodunon) were of the first century b. c. or immediately pre-conquest, although there do appear to have been two that were constructed postconquest (Hawkes and Hull 1947:45).3 At Silchester, the dates of the two main earthworks remain problematic but the 'Inner Earthwork’ is thought possibly to be of late-first-century b. c. or very early-first-century a. d. date and the 'Outer Earthwork’ is thought to have come later (Fulford and Timby 2000: 545). The 'Entrenchments’ at Chichester were also of the first century b. c. or first century a. d., although there is some evidence of the infilling and redigging of ditches after the Roman conquest (Bradley 1971; Magilton 2003:156-9).
Oram’s Arbour may have been the earliest of the oppida on or near the sites of Roman towns. Its construction date is still uncertain, but Middle Iron Age saucepan pottery4 from the primary fills of the earthworks could indicate a date between the end of the fourth and mid-first centuries b. c. (Qualmann etal. 2004:4,90). The pottery may fall nearer to the end of this period or relate to disturbed earlier activity on the site (early Iron Age pottery may also indicate preceding activity here). A later construction date would fit with the evidence from other earthwork sites and currently a late-second - or first-century-B. c. date seems most likely. Traces of an earlier earthwork constructed at Canterbury around 300 b. c. (P. Blockley 1987) have been found, but its extent, and the way in which it related to enclosures of the first century b. c. and early first century a. d. here (as at the Marlowe car park site; K. Blockley et al. 1995: 27-51), are unclear.5 The dykes and enclosures at Canterbury do, however, indicate that this area remained a focus of activity over a long period in prehistory.
At many sites where there is pre-Roman activity, but apparently no oppidum, the excavated evidence is equally late, as at Leicester, Exeter, Lincoln, and Gloucester, where it is of the first century b. c. and early first century a. d., but the landscapes do indicate longer histories. Within the River Witham at Lincoln, and further downstream at sites such as Fiskerton, there have been a number of late Bronze Age metal finds, including swords, axes, and spearheads (Jones and Stocker 2003:24-6). These indicate that this was a place of religious veneration of greater antiquity. At London, too, it can be inferred from the finds from the Thames, such as human skulls of Bronze Age date (Bradley and Gordon 1988), that there was religious activity over a long period of time. The significance of these places pre-dates the construction of earthworks and other activities. The absence of an obviously continuous archaeological sequence need not indicate discontinuous importance during this time.
World History
