Celts and Romans -Figure 8.4 Native resistance or romanization as a two-way process? (Cartoon courtesy of
Albert Rusling.)
paralleled by works on broader continental issues like Brandt and Slofstra (1983),
Barrett et al. (1989) and Blagg and Millett (1991). Various factors lie behind this shift
in emphasis.
The first has been the significant increase in the number of new sites being
discovered in the countryside, with considerable implications for the size of the
provincial population. One of the most influential surveys was that undertaken in
Northamptonshire in the 1970S (Royal Commission on Historical Monuments
1980), which recorded some 900 sites or find spots, including just 59 villas and
over 600 rural sites. To put this into perspective, Taylor (1975) quantified the rate
of discovery in just one area, the Nene valley south-west of Water Newton;
here there were 36 sites on the 19JI Ordnance Survey map of Roman Britain, 130
by the 1956 edition, and 434 by 1972 - a twelvefold increase in discoveries over 40
years (Figure 8.5). The density of sites across the county as a whole was one per
3.6 square km; in the Nene valley this rises to one per 1.5 square km. Comparable
densities have been noted elsewhere (Millett 1990, table 8.3). The implications for
the population are enormous, with past estimates progressively moving upwards
from Collingwood's conservative estimate of half a million (1929) to Smith's inflated
figure of 5 to 6 million (1977). The modern consensus favours a figure of 3 to 4
million, with Millett (1990: 185) suggesting 3.7 million, certainly more than
Domesday levels, but not as high as estimates for the fourteenth century prior to the
Black Death.
A second factor has been the realization that something like 90 per cent of
the population occupied the diverse range of non-villa settlements now known in
the countryside, outside the romanized towns and villas. This point has been empha-
sized by Hingley (1989: 4) who calculated from the ratio of villa to non-villa
settlements in three counties that, even in the south and east, villas may have formed
no more than C.I 5 per cent of the total of rural sites. This increasing emphasis
(^127)