present, providing the opportunity to investigate the relationship of
settlement hierarchy to ceramic distribution patterning. Moreover, this
material has been subjected to little synthesised study by comparison with,
say, central southern England (e.g. Fulford 1975a; Hodder 1974a; Lyne and
Jefferies 1979) and the south Midlands (Young 1977a) although site reports
on pottery are abundant. Many of the exotic fine wares of the region have
been studied in depth (e.g. Arthur 1978; Fulford 1977a; Greene 1979a; Rigby
1973; Young 1977a) as have certain of the exotic coarse wares (e.g. Hartley
1963; Lyne and Jefferies 1979; Williams 1977; Young 1977a). These studies
have tended to be focused upon the fortunes of individual industries with
little attempt being made to relate single industries to the broad spectrum
of ceramics of the appropriate period (Greene 1979a provides a notable
exception). All of these factors make the study of the pottery of Roman Kent
one that is full of untapped potential.
II. THE SELECTION OF SITES AND ASSEMBLAGES
FOR ANALYSIS
The number of sites within the study region for which published information
exists is considerable, but the quality and extent of reports are inevitably
variable. The study region contains two walled towns — Canterbury (Durovernum
Cantiacorum) and Rochester (Durobrivae) — four major military bases —
Regulbium, Rutupiae, Dubris and Lemanis — and three other named sites
along Watling Street — Noviomagus, Vagniacae, and Durolevum. The status of
these latter sites is not certain, but they may all be thought of as ‘small
towns’ — the name of the first may imply a trading function, while
Vagniacae (Springhead) incorporated a temple complex and a number of other
buildings (Penn 1965; Harker 1980) and the large cemetery at Ospringe
(Whiting et al. 1931.) may well be related to the
|
|
Durolevum community (cf. Rivet 1970). The Ordnance Survey (1978) lists 21 sites of ‘villa’
status revealed by excavations, 32 ‘other substantial buildings’ and
four rural temples and shrines. In addition a large number of cemeteries and
individual burials have been recorded and a smaller number of occupation
sites and deposits devoid of structures of ‘villa’ or other buildings
status (cf. Jessup and Taylor 1932; Pollard 1977; Sheldon and Schaaf 1978,
for surveys of portions of the evidence). The iron-working sites of the
Weald must also be taken into account in any survey of the study region, for
the road system and the presence of stamped tiles of the Classis Britannica
alone are sufficient evidence to imply strong links between these sites and
Roman Kent (e.g. see Cleere 1974, 1977; Peacock 1977b). The classification
of sites is discussed below.
The pottery reports that accompany published accounts of
excavations are as varied in quality as the excavations themselves. The
selection of sites for examination at first hand was in the first instance
dictated by the existence of a published pottery report with details of
stratigraphic relationships. These sites give an uneven spread over space,
time and functional ranges, and some were rejected to minimise duplication;
for example, of the two villas fully published in the middle Darent valley
— Lullingstone (Meates 1979) and Farningham Manor House (Meates 1973) —
the former alone was examined, although the material from both was
accessible. Sites with published, but unstratified pottery provided some
infilling of the framework established with the stratified sites. In many
cases the amount of pottery recorded suggested that further investigations
of the assemblages would not be worthwhile. Consultation with museum
authorities and excavators revealed a considerable body of unpublished
material, some of it destined for future reports. It was not always possible
to ascertain the quantity of material
|