Modern maps of North Africa offer a clear representation of the course of the Roman frontier: a sinewy defensive line stretches across it in three broken sections, across southern Numidia, with a fourth clinging to the eastern foothills of the Hodna Mountains in eastern Mauretania. Known to modern archaeologists as the fossatum Africae, this wall-and-ditch system was first established in the earlysecond century AD, and it seems to have remained in use until at least the early fifth century (when the emperor Honorius issued laws regarding its upkeep). In some stretches, especially around Maserfelta and Gemellae, the wall was studded with forts and watchtowers, and the complex as a whole was relatively close to the main base of the Legio III Augusta at Lambaesis – the principal military presence in the region. On one level, this was clearly ‘the frontier’. But this still leaves lots of questions unanswered. How do we account for the great gaps in this defensive system – like that close to Zarai, for example? How did the border of Roman authority work in the large parts of North Africa where we cannot detect these grand defensive lines? And how do we explain that peculiar loop around the Hodna Mountains, which seems to bear little relation to the wider defensive geography?
One important clue to this mystery can be found in a carved inscription dating to AD 202, which was found close to the garrison outpost at Zarai. This lists the duties to be paid on various (the pungent sauce made of fermented fish). This inscription is a reminder that the frontier system was not concerned with keeping barbarians out (or provincials in), but with directing the movement between different regions, and extracting taxation where possible.
