|
assemblages studied, it is reasonably clear that all were complete by the
second half of the fourth century. Certain elements of continuity with the
late third century can be detected, in particular the intensive utilisation
of reduced sandy wheel-thrown plain necked jars, and the predominance of
Oxfordshire wares amongst the mortaria, achieved with the wane of ‘East
Kent’ oxidised wares in the late third (or possibly early fourth) century.
However, it seems possible that consumption of both mortaria and sandy
wheel-thrown plain wares declined in the late fourth century.
The study of west Kent pottery in this period is inhibited by
the dearth of fourth-century assemblages uncontaminated by residual
material. Springhead and Rochester have only provided deposits of a broad
third- to early fifth-century date (S. Harker, pers. comm.; Flight and
Harrison 1978), the latter covered by medieval layers. The build-up of
material within the ‘cellar’ at Chalk over the late third/early fourth
century collapse (Johnston 1972, layers 5 and 6) includes some samian
(unpublished). Lullingstone villa material is mostly unstratified, but two
pits (Meates et a!. 1950, 18—19, Group III — Room 8 has been
renumbered 15 (Meates 1979); Meates et al. 1952, nos. 82—90,
from Room 14, renumbered 11 (Meates 1979)) include quantities of mid
fourth-century coins (Meates 1979, 56—7, 87—8) associated with small
pottery assemblages (Pollard 1987, Groups XLVI and XLV), and the pottery
from destruction and post-destruction levels in particular of the ‘Deep
Room’ and bathing establishment would appear to be free of contamination
from material at least of residual third- to early fourth-century date (ibid.,
Groups LI, LII, XXXVII). Further evidence of the range of pottery in use
in the fourth century is provided by a small assemblage from Bexley (Tester
1963) associated with coins from A.D. 222—235 to 337—341; the pottery
does not include BB2 or third century ‘north-west Kent’ fine sandy
burnished ware, suggesting that it is dateable to within the final
century
|
|
of the Roman occupation of Britain. Varying quantities of material of late
third- to fourth-century type have been recovered from numerous other sites
in unstratified contexts (cf. Figs. 7 and 8). No fourth-century kilns are
known in west Kent.
The quantified pottery sequences from Chalk, Springhead and
Rochester allow very broad trends in the consumption and discard of wares in
the third and fourth centuries to be detected (Appendix 5). On this basis a
decline in BB2 can be postulated; the ‘fourth century’ accumulation in
the Chalk ‘cellar’ incorporated some 30 per cent BB2 (vessel rim
equivalence) which comprises 31 per cent of the coarse wares alone, a lower
proportion than was present in the late third-century layers, wherein BB2
comprised over 44 per cent of the coarse wares. Comparisons of the figures
from late second to mid-third and third- to fourth-century assemblages from
Rochester, and mid-second to mid-third and mid-third to fourth-century
assemblages from Springhead, also show an overall fall in the amounts of
BB2. There is no evidence for any additions to the formal range of this ware
in west Kent later than the mid-third century adoption of the
bead-and-flange dish. It is quite conceivable that BB2 continued to be
produced into the first half of the fourth century; indeed, there is no
evidence to conflict with this view. However, the absence of the ware from
the Bexley group, and apparently from both Lullingstone pits (this pottery
could not all be traced by the present author — Pollard 1987; there is no
obvious BB2 in the original publication, except perhaps Meates et al. 1950,
no. 14), suggests that it was passing out of usage in the second quarter of
the century or thereabouts. The pottery attributable to the destruction and
post-destruction groups of late fourth to early fifth century from
Lullingstone, and to an occupation level close to the Temple-Mausoleum on
its south side (Pollard 1987, Group XL, associated with three coins of
Magnentius and two copies
|