- Celts and Romans -
This would help to explain the apparent primate importance of the civitas centres
during the later first and second centuries Gones 199 Ia: 59-60), and the relatively slow
development at all but a few 'small towns' (Burnham and Wacher 1990: 10-14).
In the countryside, the villa has traditionally been seen as a symbol of change,
usually involving a more capitalistic approach to agriculture and a close economic
symbiosis with the cities (Applebaum 1972). For reasons already outlined, it now
seems less easy to link villas directly with innovation in agriculture and as such they
are best seen not so much as evidence of successful and profitable exploitation of
large estates, but rather as symbols of surplus wealth and conspicuous display. More
important, perhaps, the increased emphasis on continuity in the landscape and
among rural settlements spanning the later Iron-Age to Romano-British periods has
influenced modern opinion on early villa development, with a growing acceptance
that many villas must have emerged out of, rather than being imposed upon, the
existing native infrastructure. This has long been known or suspected for sites like
Park Street and Lockleys, but has recently been further reinforced by the evidence
from Rivenhall (Rodwell and Rodwell 1986) and Gorhambury, where the villa
was clearly constructed on a site previously occupied by a native settlement in close
proximity to the dykes associated with the oppidum of Verulamium (Neal et al.
1990); a similar link between an existing centre of power and an early villa has
also been noted at The Ditches, part of the oppidum complex at Bagendon (Trow
1988). Although the question of ownership is hard to establish with any certainty,
such examples serve to reinforce the notion that most belonged to natives, thereby
helping to explain the rash of villas which emerged in the vicinity of such towns
as Verulamium at a time when the city itself was being progressively elaborated
with public buildings and amenities (Neal 1978). This would be the simplest inter-
pretation as well of the early and large villas which developed along the south coast
at sites like Fishbourne, Angmering and Southwick, in what had probably been the
client kingdom of Cogidubnus, though the direct association of Fishbourne with
Cogidubnus now seems far less secure than was once thought (Cunliffe 1973).
Similar concentrations are known in Kent (Detsicas 1983) and Essex (Rodwell 1978),
likewise in areas with a well-developed native infrastructure on the eve of the
conquest. This would reinforce the argument that the elite adopted the villa as a
convenient vehicle for expressing their attachment to the new romanized order
(much as they were doing in the towns), while still retaining their power within the
existing socio-economic and agricultural landscape (Blagg 1991).
Moreover Smith's (1978, 1984) work on villas has consistently identified distinctive
types of villa which do not conform easily with the traditional classical norm of
individual ownership. The most important from the point of view of early villa
development and the emergence of a Romano-Celtic society, is his recognition of so-
called 'unit system villas' involving two houses set at an angle, two or more houses
focused around a courtyard with no obvious, axially placed entrance, and elongated
villas incorporating groups of rooms seemingly forming separate units. Smith has
argued that such villas reflect some form of multiple land ownership enshrined within
the plan, even though it can sometimes be skilfully hidden beneath the classical
veneer. Although there are problems with this interpretation, there seems to be little
doubt that the explanation lies firmly in the native social order and infrastructure,
135