116 THE ROMAN EMPIRE
presence of a settled garrison frequently stimulated the growth of an ‘army’
of local producers and suppliers. This was not always a spontaneous
development. The village structure of fi rst- and second- century Dobrogea in
the province of Lower Moesia was the product of a series of enforced
colonizations of conquered tribesmen, which had the specifi c object of
supporting the lower Danubian frontier with supplies and manpower.
Moreover, it is signifi cant that elsewhere in Lower Moesia the Roman
occupation had very little effect on local settlement patterns. Similarly,
British archaeologists have noticed that the proximity of the army did not
stimulate agriculture in the highland zone or in Wales; and that for that
matter it did not provoke the appearance of a new, indigenous northern
pottery industry on a scale to compete with or replace imported pottery. The
situation in the south- west of England was different: here it seems that
the locality, within a radius of about 30 miles, played an important part
in the supply of the garrison at Gloucester. But to return to the North, where
the bulk of the army was stationed: it was the South that shouldered the
main burden of supplying the army of Britain. To put it in another way,
the category of middle- distance supply was important in Britain.^12
The British army also received goods from Gaul or further afi eld. In the
period from the invasion to the end of the fi rst century, a very signifi cant
level of imports was sustained in artefacts. Not only in artefacts: pottery and
other small manufactures typically travelled pick- a-back, in the gaps left by
a primary cargo, and can therefore be taken as proxy for the bulk movement
of raw materials and perishables. The latter items are invisible, and their
identity and relative importance can only be guessed at: in the British case
perhaps iron, cloth, hides and some foodstuffs. Should we include cereals in
the list of long- distance imports? The discovery by archaeologists of grain
pests such as Sitophilus granarius on British sites, unattested in earlier
periods and therefore foreign, can tell us nothing whatever about the
regularity of grain imports or the quantities involved. It is still therefore in
principle open to advocates of British self- suffi ciency to argue that the
province paid its own way in respect of the most important staple of all. The
case is stronger for the second century, if the fall- off in the import of (visible)
manufactures and (invisible) primary products can be correlated with
increased agricultural production in Britain. A partial correlation can be
admitted. Another factor, perhaps the primary factor, in the decline of
imports was the reduction in the size of the military establishment in the
province in the second and third centuries.^13
The British case illustrates the diffi culty of gauging the relative importance
of neighbourhood, regional and long- distance supply in any particular
instance, especially as the balance between the various categories was likely
to alter over time. But the supposition still stands that an army by preference
supported itself from the locality and region where it was stationed. If
necessary, goods were ordered in bulk from long- distance suppliers, and the
order might well have been substantial in the early period of military