Early in the 4th century it was necessary to establish a special coast defence, reaching from the Wash to Spithead, against Saxon pirates: there were forts at Brancaster, Borough Castle (near Yarmouth), Bradwell (at the mouth of the Colne and Blackwater), Reculver, Richborough, Dover and Lymme (all in Kent), Pevensey in Sussex, Porchester near Portsmouth, and perhaps also at Felixstowe in Suffolk. After about 350, barbarian assaults, not only of Saxons but also of Irish (Scoti) and Picts, became commoner and more terrible. At the end of the century Magnus Maximus, claiming to be emperor, withdrew many troops from Britain and a later pretender did the same. Early in the 5th century the Teutonic conquest of Gaul cut the island off from Rome. This does not mean that there was any great "departure of Romans." The central government simply ceased to send the usual governors and high officers. The Romano-British were left to themselves. Their position was weak. Their fortresses lay in the north and west, while the Saxons attacked the east and south. Their trained troops, and even their own numbers, must have been few.
It is intelligible that they followed a precedent set by Rome in that age, and hired Saxons to repel Saxons. But they could not command the fidelity of their mercenaries, and the Saxon peril only grew greater. It would seem as if the Romano-Britons were speedily driven from the east of the island. Even Wroxeter on the Welsh border may have been finally destroyed before the end of the 5th century. It seems that the Saxons though apparently unable to maintain their hold so far to the west, were able to prevent the natives from recovering the lowlands. Thus driven from the centres of Romanized life, from the region of walled cities and civilized houses, into the hills of Wales and the north-west, the provincials underwent an intelligible change. The Celtic element, never quite extinct in those hills and, like most forms of barbarism, reasserting itself in this wild age - not without reinforcement from Ireland - challenged the remnants of Roman civilization and in the end absorbed them. The Celtic language reappeared; the Celtic art emerged from its shelters in the west to develop in new and medieval fashions.
The principal references to early Britain in classical writers occur in Strabo, Diodorus, Julius Caesar, the elder Pliny, Tacitus, Ptolemy and Cassius Dio, and in the lists of the Antonine Itinerary (probably about A.D. 210-230; ed. Parthey, 1848), the Notitia Dignitatum (about A.D. 400; ed. Seeck, 1876), and the Ravennas (7th-century rechauffé; ed. Parthey 1860). The chief passages are collected in Petrie's Monumenta Hist. Britann. (1848), and (alphabetically) in Holder's Altkeltische Sprachschatz (1896-1908). The Roman inscriptions have been collected by Hübner, Corpus Inscriptionum Latin. vii. (1873), and in supplements by Hübner and Haverfield in the periodical Ephemeris epigraphica; see also Hübner, Inscript. Britann. Christianae (1876, now out of date), and J. Rhys on Pictish, etc., inscriptions, Proceedings Soc. Antiq. Scotland, xxvi., xxxii.
Of modern works the best summary for Roman Britain and for Caesar's invasions is T.R. Holmes, Ancient Britain (1907), who cites numerous authorities. See also Sir John Evans, Stone Implements, Bronze Implements, and Ancient British Coins (with suppl.); Boyd Dawkins, Early Man in Britain (1880); J. Rhys, Celtic Britain (3rd ed., 1904). For late Celtic art see J.M. Kemble and A.W. Franks' Horae Ferales (1863), and Arthur J. Evans in Archaeologia, vols. lii.-lv. Celtic ethnology and philology (see Celt) are still in the "age of discussion." For ancient earthworks see A. Hadrian Allcroft, Earthwork of England (1909).
For Roman Britain see, in general, Prof. F. Haverfield, The Romanization of Roman Britain (Oxford, 1906), and his articles in the Victoria County History; also the chapter in Mommsen's Roman Provinces; and an article in the Edinburgh Review, 1899. For the wall of Hadrian see John Hodgson, History of Northumberland (1840); J.C. Bruce, Roman Wall (3rd ed., 1867); reports of excavations by Haverfield in the Cumberland Archaeological Society Transactions (1894-1904); and R.C. Bosanquet, Roman Camp at Housesteads (Newcastle, 1904). For the Scottish Excavations see Proceedings of the Society of Antiquaries of Scotland, xx.-xl., and especially J. Macdonald, Bar Hill (reprint, Glasgow, 1906). For other forts see R.S. Ferguson, Cumberland Arch. Soc. Trans. xii., on Hardknott; and J. Ward, Roman Fort of Gellygaer (London, 1903). For the Roman occupation of Scotland see Haverfield in Antonine Wall Report (1899); J. Macdonald, Roman Stones in Hunterian Mus. (1897); and, though an older work, Stuart's Caledonia Romana (1852). For Silchester, Archaeologia (1890-1908); for Caerwent (ib. 1901-1908); for London, Charles Roach Smith, Roman London (1859); for Christianity in Roman Britain, Engl. Hist. Rev. (1896); for the villages, Gen. Pitt-Rivers' Excavations in Cranborne Chase, etc. (4 vols., 1887-1908), and Proc. Soc. of Ant. xviii.
For the end of Roman Britain see Engl. Hist. Rev. (1904); Prof. Bury's Life of St Patrick (1905); Haverfield's Romanization (cited above); and P. Vinogradoff, Growth of the Manor (1905), bk. i.
(F. J. H.)