SUPPLYING THE ROMAN EMPIRE 117occupation. But the balance tilted away from long- distance imports as
suppliers closer to hand grew in number and capacity, and as the state
authorities grew more interested in and more expert at exploiting them. The
signifi cance of this last factor, not touched upon thus far, will emerge when
we come to discuss the means by which supplies were extracted from the
civilian population.
The status of the goods supplied to the army and the methods by which
they were introduced are a second index of the degree of state involvement
in army supply.
The army to some extent supplied itself.^14 This was a matter of practical
necessity, in as much as the environment was unfriendly, the resources of the
civilian population insuffi cient, and imports inadequate. Thus, for example,
in the Rhineland, it was standard for military personnel to make cooking
pots, mortaria and other pottery, iron articles and implements, leather goods
and certain weapons, and for that matter, to graze animals on land attached
to the legion. We are a long way from the more organized state production
system of Diocletian, but self- supply was not a negligible factor in earlier
times, and would have reduced the army’s dependence on and interaction
with the local economy.
The goods it could not produce, or not in suffi cient quantity – grain,
fodder, meat, a wide variety of processed foods (beverages, milk products,
salt and so on), clothing, armour and weaponry – had to be acquired in
other ways. Almost from the fi rst in the history of Rome’s relationship with
a frontier area, food and equipment came in as tax, tribute or contributions
under some other name from defeated enemies and other peoples who
acknowledged Roman supremacy.^15 Collection and transport might be
supervised by soldiers or civilian offi cials. In more settled times and
environments, these exactions characteristically took the form of obligations
imposed on the civilian population through the agency of city offi cials.
To tax should be added requisitions. Where requisitioned goods were
paid for rather than merely seized, as in times and places where good
relations with civilians were accorded some value, the price was presumably
fi xed by the buyer and therefore usually below the market rate. Compulsory
purchase may be supposed to have been a fundamental source of supplies
everywhere, as it can be shown to have been in Egypt. This is not surprising,
for the following reason. The return of taxes in kind, if levelled as a
percentage of the harvest, was unpredictable, while goods travelling from a
distance might be held up or lost. While in the matter of vital staples, in
particular grain, the authorities stockpiled supplies (British military granaries
were built to hold one to two years’ supply of grain), the precise needs of a
garrison in other items might have been underestimated. Topping up must
have been a common necessity; in which case, it was better diplomacy to
buy than to impose a supplementary tax. This is what the detachments of
soldiers from Stobi on the lower Danube and from Dura-Europus on the
northern Euphrates were doing when they went in search of grain for men