It seems probable that Maximus retained in Gaul a considerable part of the Britannic army; and if so, it was doubtless one of the cares of Stilicho to restore to Britain these contingents. We may assume that from A.D. 390 two legions (IInd and VIth), if not the XXth also, were in Britain.
(3) At the end of 401 Alaric entered Italy (for the chronology see Appendix 17 to Bury’s edition of Gibbon, vol. iii.), and a legion was summoned for the defence of Italy. Claudian is again our source, De Bello Gothico, 416-8:—
Venit et extremis legio praetenta Britannis
Quae Scotto dat frena truci ferroque notatas
Perlegit exanimes Picto moriente figuras.
This description of a legion which might be called upon to act against both Irish and Pict certainly suggests the legion which was stationed in the northern military district, of which the headquarters was York.[366] Now in the Notitia Dignitatum, which represents the state of the civil and military service in the Empire in the first years of the fifth century, we find two legions in Britain, the VIth in the northern districts under the dux Britanniarum, the IInd in the south-eastern district under the comes litoris Saxonici per Britannias. No legion appears in the western district under the comes Britanniae. It has therefore been supposed that it was the XXth legion, stationed in the west, which was summoned to Italy; and the fact that this legion does not appear in the Notitia at all has suggested that this document was drawn up after it had left Britain, but before it had been permanently assigned to any other station. Claudian’s words need not be fatal to this theory, for the legion whose headquarters was at Chester would have to defend Britain against the Scots, and might have to defend it against Picts if they broke through the wall; and in any case the words of the poet in such a matter could not be precisely pressed. But the argument cannot be regarded as conclusive. It is perfectly possible that the XXth legion had been broken up before this time, that there were only two legions in Britain, and that it was the VIth which went to Italy. Its departure, it should be remembered, did not leave north Britain defenceless; there were large forces, cohorts, and alae, distinct from the legion. It is, moreover, possible that the British section of the Not. is of much older date.
(4) The elevation of the tyrant Constantine and his crossing into Gaul would seem to have happened in the first half of A.D. 407. This seems to follow from the account in Olympiodorus (fr. 12, ed. Müller). The second half of A.D. 406 is filled by the episodes of Constantine’s predecessors, Marcus and Gratian. Zosimus (who, in regard to the dating, misunderstood Olympiodorus, as was pointed out by Freeman[367]) says that the invasion of Gaul by the Vandals and their fellows was one of the causes of the rebellion in Britain (vi. 3, 1). This cannot be true, if the text is sound in the entry of Prosper Tiro, which states that the Vandals and Alans crossed the Rhine at the end of December A.D. 406.[368] On the other hand, two considerations might tempt us to suspect this notice, namely, the unlikelihood of a migration in the middle of winter, and the existence of two edicts for a levée of provincials contra hostiles impetus dated April 18 and 20 at Ravenna (cp. Hodgkin, i. 739). But neither of these considerations is weighty enough to justify us in assuming without further evidence anything inconsistent with the testimony of Prosper’s MSS. In any case, the crossing of Constantine into Gaul will fall in the year A.D. 407.[369] We can have no doubt that most of the soldiers in Britain accompanied Constantine when he departed to secure Gaul. But he must have left garrisons; it was important for him to retain his hold on the island if, as is probable, his corn supply depended on it. Mr. Freeman has summed up the event thus in its consequences for Britain: “Britain might be saved by a campaign in Gaul. But if this was the motive, the thought of saving Britain must soon have passed away from the minds of Constantine and his soldiers. Whether they cared for such an object or not, the course of things on the mainland soon made it hopeless for them to think of keeping up any relations with the great island. The crossing of Constantine into Gaul thus became the end of the Roman power in Britain” (p. 56).
It is possible that Constantine, before he departed, may have set up a colleague, named Carausius, to safeguard his interests in Britain. This at least is the inference drawn by Mr. A. J. Evans from a coin found at Richborough (see Numismatic Chronicle, third series, vii. 191 sqq. 1887, and App. 19 to Bury’s edition of Gibbon, vol. iii.).
(5) We may now briefly consider the account of this last period of Roman domination which is given in the Epistle of Gildas, De excidio et conquestu Britanniae. Gildas, a native of west Britain, wrote in the first half of the sixth century, and his description in cc. 13-20 represents the confused memories of that troubled time, surviving in his own day. As Mommsen[370] says: “haec ut in universum rerum statum in Britannia qui tum fuit recte repraesentare videntur, sic ut narrantur magis famam incertam reddunt quam sinceram rerum gestarum narrationem.”
In this account three great devastations are distinguished, and each is attributed to the withdrawal of the Roman garrison. The first was caused by the revolt of Maximus (cc. 13, 14); the island was devastated by Scots and Picts;[371] and it was not until the islanders sent an embassy with letters to Rome that “a legion,” praeteriti mali (the disloyalty of Britain) immemor, was sent to Britain, defeated the enemy, and built a wall across the island. The details are of course inaccurate, but the general fact that the rebellion of Maximus brought invasion upon Britain, and after his fall measures were taken (by Stilicho) for fortifying it, is correct.