more than one distinct system of fortification, and where we can trace an early fort, defended by earthen ramparts, passing through a gradual process of evolution in its defences into something larger and stronger. The Saalburg, for instance, has been more which probably carefully investigated than any other Roman work in Germany, presents evidence of no less than six periods of occupation.[1] Similarly, at Kapersburg[2] we have three successive forts occupying very nearly the same ground: the earliest a simple earth fort with a single ditch ; the second slightly larger but otherwise much the same, except that the rampart, instead of being simply of earth, had consisted of two parallel dry-stone walls with earth packed in between; the third, again somewhat larger, defended by a ditch and a solid stone wall five feet four inches in thickness. So, too, at Zugmantel[3] we can follow three stages in the existence of the fort. In the first it was small in size, and had a ditch and an earthen rampart. In the second stage it was considerably enlarged, while round it there ran a stone wall about four feet thick. Finally, it was still further enlarged, and the surrounding wall was made no less than six feet six inches thick. Illustrations of this evolutionary process can be cited from our literary sources. Arrian, in his Periplus, written in the time of Hadrian, describes the fort on the Phasis on the south coast of the Black Sea, a little post garrisoned by some four hundred men and surrounded by a rampart and two broad ditches. 'Formerly,' he writes, 'the rampart was of earth, and the towers planted upon it were of wood. Now both ramparts and towers are made of brick.'[4] Closely analogous is the well known inscription from Bumbesti in Roumania, commemorating the replacement in stone of the turf walls of a dilapidated fort. The alteration took place in the year A.D. 201.[5]
It is this very process of evolution that we have to follow at Newstead—a simple earthen fort, early in date and irregular in its ground-plan, passing into a later stone-walled fort of what was no doubt a standard type. But although the earthen fort here preceded the stone-walled fort, and is usually found to do so elsewhere, an earthen fort is not necessarily an early fort, nor a
1 H. Jacobi, Führer durch das Römerkastell Saalburg, p. 16.
2 Der Obergermanisch-Raetische Limes, Lief. 17, Kastell Kapersburg.
3 Ibid. Lief. 32, Kastell Zugmantel.
4 Macdonald and Park, Bar Hill, p.31, where the passage is quoted.
5 IMP. CAES. L. SEPTIMIUS SEVERUS . . . ET IMP. CAES. M. AVR. ANTONINUS . . . MVROS CESP[IT(ICIOS)] CASTRORUM COH. I. AVRELIAE BRITTONUM (MILLARIAE) ANTONINIANA VETVST(ATE) DILA[PSOS] LAPIDE EOS RESTITUERUNT PER OCTAVIUM IULIANUM LEG IPSO[RUM] PR. PR. C.I.L. vol. iii. 14485a.
28