Eastbourne beach, there was no prey for the freebooter such as lay open farther west on the fertile coast between Brighton and Chichester. But as a post for ships to watch the Channel, Pevensey had its plain advantages. It would seem that, if the classis Britannica ceased to exist as a whole in the fourth century, ships must have been attached to each of the forts in some manner which has not been recorded.
   It is not easy on historical grounds to fix the precise date when this system of defence was organized. Its general character, as sketched in the ‘Notitia,’ shows that it is not earlier than the reforms of Diocletian. That is, it belongs either to the very end of the third century or to some part of the fourth century. The one reference to it which occurs in the ancient historians helps us a little further. Ammian states that in A.D. 368 Nectaridus, the ‘comes maritimi tractus,’ and Fullofaudes, the ‘dux,’ were killed or captured by barbarians in Britain.¹⁹  The former is clearly the Count of the Saxon Shore, and the latter the Dux Britanniarum who commanded the rest of the British troops. This is proof that the Saxon Shore had been organized before 368. Unhappily we have little further evidence. No safe inference can be drawn from the general course of contemporary events. It might be urged that Frank and Saxon pirates were plaguing all western Europe by A.D. 300, and Britain can hardly have escaped. But ancient writers mention Gaul as the chief victim; no word occurs of attacks on Britain till about A.D. 350, and the general prosperity of the island in the Constantinian period proves that, if attacked, it did not suffer much. This leaves us at liberty to infer either that it was not attacked or that it was already well protected, and in the end tells us little.
   Archaeology is more helpful. Burgh Castle, Bradwell, and Pevensey seem not to have been inhabited till the Saxon Shore was established, and their coin-lists are fairly full. The earliest issues discovered here date from about A.D. 260 or 280, and coins of all parts of the fourth century abound. But the decisive evidence comes from Richborough, where the excavations carried out since 1922 by the Society of Antiquaries have gradually narrowed down the probable date of the building of the Saxon-Shore fortress to a period of twenty years on either side of A.D. 280. We may reasonably conclude that the Shore was organized in the latter part of the third century, but whether by Constantius Chlorus after the fall of Allectus in A.D. 297 or by one of his Gallic predecessors still remains to be determined (see below, p. 41).
   Two further questions arise. One is whether the Saxon Shore, as then created, included pre-existing forts. The other is whether it was afterwards enlarged by additional forts. These questions are not easy to answer. The fact that the ‘Notitia’ names nine forts while the remains of eleven or twelve have been probably detected, may suggest that either Felixstowe or Porchester (to say nothing of Carisbrooke) was not yet in existence or had passed out of use when the ‘ Notitia’ was compiled. But the date of this list is uncertain, and it does not therefore help us to any conclusions. Again, the absence of uniformity both in the outlines and in some other details of the forts may suggest a difference of period. But this want of uniformity, as has been said, meets us elsewhere among the remains of the fourth century and should not perhaps be stressed. It may be that the complete surrender to the contour-principle
   ¹⁹ Ammianus Marcellinus, xxvii, 8, 1.