CHAPTER XIII

ROMAN BRITAIN

 

THE character and history of Roman Britain, as of many other Roman provinces, were predominantly determined by the facts of its geography. To that cause, or set of causes, more than to any other, we must attribute alike the Roman desire to conquer the province and the actual stages of the conquest, the distribution of the troops employed as permanent garrison, the quality and extent of the Romanized civilization, and, lastly, a great part of the long series of incidents by which the island was lost to Rome and Roman culture.

Geologically, Britain forms the north-west side of a huge valley which had its south-east side in northern and central France. Down the centre of this valley ran two rivers, the one flowing south-west along a bed now covered by the English Channel, the other flowing north-east through a region now beneath the German Ocean. From these rivers, the land sloped upwards, south-east to Vosges, Alps, and Cevennes, north-west to Cornwall, Wales and northern Britain. The two rivers have long vanished. But the configuration of their valleys has lasted. Though unquiet seas now divide England from north-western Europe, the two areas, that were once the two sides of the valleys, still look to each other. Their lowlands lie opposite; their main rivers flow out into the intervening sea; their easiest entrances face; each area lies open by nature to the trade or the brute force of the other; each has its most fertile, most habitable, and least defensible districts next to those of the other.

Hence comes the peculiar configuration of our island. In south-east Britain there is little continuous hill-country that rises above the 600 foot contour line. Instead, wide undulating lowlands, marked by no striking physical feature and containing little to arrest or even divert the march of ancient armies or of traders, stretch over all the south and east and midlands. For hills, we must go north of Trent and Humber or west of Severn and Exe. There we shall find almost the converse of the south-east. Throughout a large, scattered region, extending from Cornwall to the Highlands, the land lies mostly above, and much of it high above, the 600 foot line; its soil and climate are ill-suited to agriculture; its deep valleys and gorges and wild moors and high peaks oppose alike the soldier and the citizen. Behind this upland lies the Atlantic, and an Atlantic which meant of old the reverse of what it does today. To the ancients, this hill-country was the end of the world; for us—since Columbus—it is the beginning

Roman Britain: Conquest [BC 55-A.D. 140

These physical features are reproduced plainly in the early history of Britain. It was natural that about BC 50-A.D. 50 southern Britain should be occupied by Celtic tribes and even families which had close kindred in Gaul, and that a lively intercourse should exist between the two. It was no less natural that, even before Rome had fully conquered Gaul, Caesar’s troops should be seen in Kent and Middlesex (BC 55-54) and Roman suzerainty extended over these regions; and when the annexation of Gaul was finally complete, that of Britain seemed the obvious sequel. The sequel was, indeed, delayed awhile by political causes. Augustus (BC 43-AD 14) had too much else to do: Tiberius (14-37) saw no need for it, just as he saw no need for any wars of conquest. But after 37 it became urgent. Changes in southern Britain had favoured an anti-Roman reaction there and had even perhaps produced disquiet in northern Gaul; Caligula (37-41) had made some fiasco in connection with it; when Claudius succeeded, there was need of vigorous action and, as it chanced, the leading statesmen of the moment favored a forward policy in many lands. The result was a well-planned and deservedly successful invasion (AD 43).

The details of the ensuing war of conquest do not here concern us. It is enough to say that the lowlands offered little resistance. In one part of them, near the south-east coast, Roman ways had become familiar since Caesar's raids. In another part—the midlands—the population was then, as now, thin. Nowhere (despite the theories of Guest and Green) were there physical obstacles likely to delay the Roman arms. By 47 the invaders had subdued almost all the lowlands, as far west as Exeter and Shrewsbury and as far north as the Humber. Then came a pause. The difficulties of the hill-country, the bravery of the hill-tribes, political circumstances at Rome, combined not indeed to arrest but seriously to impede advance. But the decade 70-80 saw the final conquest of Wales and the first subjugation of northern England, and in the years 80-84 Agricola was able to cross the Tyne and the Cheviots and gradually advance into Perthshire. Much of the land which he overran was but imperfectly subdued and the northern part of it—everything, probably, north of the Tweed—was abandoned when he was recalled (85). Thirty years later (115-120) an in­surrection shook the whole Roman power in northern Britain, and when Hadrian had restored order, he established the frontier along a line from Tyne to Solway, which he fortified by forts and a continuous wall (about 122-124). Fifteen or twenty years later, about AD 140, his successor Pius, for reasons not properly recorded, made a fresh advance; he annexed Scotland up to the narrow isthmus between Forth and Clyde and fortified that with a continuous wall, a series of forts along it variously estimated at 12 or (more probably) at 18 or 20, and some outposts along the natural route through the Gap of Stirling to the north-east. This wall was not meant as a substitute for Hadrian's Wall, but as a defence to the country north of it.

Rome had now reached her furthest permanent north. But the advance was not long accepted quietly by the natives. Twenty years after Pius had built his wall, a storm broke loose through all northern Britain from Derbyshire to Cheviot or beyond (about 158-160). A second storm followed 20 years later (about 183); the Wall of Pius was then or soon after definitely lost, and disorder apparently continued till the Emperor Septimius Severus came out in person (208-211) and rebuilt the Wall of Hadrian to form, with a few outlying forts, the Roman frontier. With this step ends the series of alternating organization and revolt which make up the external history of the earlier Roman Britain. Henceforward the Wall was the boundary until the coming of the barbarians who ended Roman rule in the island.

AD 85-211] Garrisons

The force which garrisoned this fluctuating frontier and kept the province quiet consisted of three (till AD 85, of four) legions and an uncertain number of troops of the second grade, the so-called auxilia, in all perhaps some 35-40,000 men, mostly heavy infantry. The three legions were disposed in three fortresses, Isca Silurum (Caerleon on Usk, legio II Augusta), Deva (Chester, legio XX Valeria Victrix) and Eburacum (York, legio VI Victrix): from these centers detachments (vexillationes) were sent out to form expeditionary forces, to construct fortifications and other military works, and generally to meet important but occasional needs. Outside these three main fortresses, the province was kept quiet and safe by a network of small forts (castella), varying in size from two or three to six or seven acres and garrisoned by auxiliary cohortes (infantry) or alae (cavalry), some 500 and some 1000 strong. These forts were planted along important roads and at strategic points, 10 or 15 or 20 miles apart. Their distribution is noteworthy. In the lowlands there were none. During the early years of the conquest we can, indeed, trace garrisons at one or two places, such as Cirencester. But, as the conquest advanced, it was seen that the lowlands needed no force to ensure their peace, and the troops were pushed on into the hills, beyond Severn and Trent. Eighteen or twenty forts were dotted about Wales, though many of these seem to have been abandoned in the course of the second century, as having become superfluous through the growing pacification of the land. A much larger number can be detected in Derbyshire, Lancashire, the hill-country of Yorkshire, and northwards as far as Cheviot: Hadrian's Wall, in particular, was principally defended by a series of such forts. We cannot, however, give precise statistics of these forts until explora­tion has advanced further : it is doubtful not only how far the known examples provide us with a fairly full list of them, but, still more, to what extent all the forts were in occupation at the same time and to what extent one succeeded another.

The troops which garrisoned these military posts were Roman, in the sense that they not only obeyed the Roman Emperor but were in theory and to a great extent in practice, even in the later days of Roman Britain, recruited within the Empire. The legionaries came from Romanized districts in the Western Empire; the auxiliaries, naturally less civilized to begin with but drilled into Roman ways and speech, were largely drawn from the Rhine and its neighborhood: some probably were Celts, like the native Britons, others (as their names on tombstones and altars prove) were Teutonic in race. To what extent Britons were enrolled to garrison Britain, is not very clear; certainly, the statement that British recruits were always sent to the Continent (chiefly to Germany), by way of precaution, seems on our present evidence to be less sweepingly true than was formerly supposed.

From the standpoints alike of the ancient Roman statesman and of the modern Roman historian the military posts and their garrisons formed the dominant element in Britain. But they have left little permanent mark on the civilization and character of the island. The ruins of their forts and fortresses are on our hill-sides. But, Roman as they were, their garrisons did little to spread Roman culture here. Outside their walls, each of them had a small or large settlement of womenfolk, traders, perhaps also of time-expired soldiers wishful to end their days where they had served. But hardly any of these settlements grew up into towns. York may form an exception: it is a pure coincidence, due to causes far more recent than the Roman age, that Newcastle, Manchester, and Cardiff stand on sites once occupied by Roman ‘auxiliary’ forts. Nor do the garrisons appear to have greatly affected the racial character of the Romano-British population. Even in times of peace, the average annual discharge of time-expired men, with land-grants or bounties, cannot have greatly exceeded 1000, and, as we have seen, times of peace were rare in Britain. Of these discharged soldiers by no means all settled in Britain, and some of them may have been of Celtic or even of British birth. Whatever German or other foreign elements passed into the population through the army, cannot have been greater than that population could easily and naturally absorb without being seriously affected by them. The true contribution which the army made to Romano-British civilization was that its upland forts and fortresses formed a sheltering wall round the peaceful interior regions.

AD. 43-96] History of Romanization

Behind these formidable garrisons, kept safe from barbarian inroads and in easy contact with the Roman Empire by short sea passages from Rutupiae (Richborough, near Sandwich in Kent) to Boulogne or from Colchester to the Rhine, stretched the lowlands of southern, midland, and eastern Britain. Here Roman culture spread and something approximating to real Romanization took place. The process began probably before the Claudian invasion of 43. The native British coinage of the south-eastern tribes and other indications suggest that, in the 100 years between Julius Caesar and Claudius, Roman ways and perhaps even Roman speech had found admission to the shores of Britain, and this infiltration (as I have said) may have made easier the ultimate conquest. After the conquest, the process continued in two ways. In part it was definitely aided by the government which established here, as in other provinces, municipalities peopled by Roman citizens, for the most part discharged legionaries, and known as coloniae: these, however, were comparatively few in Britain. Far greater was the automatic movement. Italians flocked to the newly opened regions—traders, as it seems, rather than the laborers who form the emigrants from Italy today: how numerous they were, we can hardly tell, but such commercial emigrations are always more important commercially than for their mere numbers. Certainly a far more notable movement was the automatic acceptance of Roman civilization by the British natives.

We can to some extent trace this movement. Quite early in the period AD 43-80, the British town Verulamium, just outside St Albans in Hertfordshire, was judged to have become sufficiently Romanized to merit the municipal status and title of municipium (practically equivalent to that of the colonia manned by veteran soldiers). The great revolt of Boudicca (less correctly called Boadicea) in AD 60 was directed not only against the supremacy of Rome but also against the spread of Roman civilization, and one incident in it was the massacre of many thousands of "loyal" natives along with actual Romans. Romanization, it is plain, had been spreading apace. Nor did this massacre check it for long. The Flavian period (AD 70-96) saw in Britain, as indeed in other provinces, a serious development of Roman culture and in particular of Roman town life, the peculiar gift of Rome to her western provinces. In the decade AD 70-80, the Britons began, as Tacitus tells us, to speak Latin and to use Latin dress and the material fabric of Latin civilized life. Now towns sprang up, such as Silchester (Calleva Atrebatum) and Caerwent (Venta Silurum), laid out on the model approved by Roman town-planners, furnished with public buildings (forum, basilica, etc.) of Roman style, and filled with houses which were Roman in their internal fittings (baths, hypocausts, wall-paintings) if not in ground-plan. Now the baths of Bath (Aquae Sulis) were equipped with civilized buildings suited to their new visitors: the earliest datable monument there belongs to about 77. Two coloniae also were planted. Hitherto there had only been one, established by Claudius at Colchester (Camulodunum): now one was added at Lincoln (Lindum) and in 96 a third at Gloucester (Glevum). A new Civil Judge (legatus iuridicus) begins to make his appearance beside the regular legatus Augusti pro praetore who was at once commander of the troops and judge of the chief court and governor of the province, and the appointment is doubtless due to increasing civil business in the law courts. When Tacitus praises Agricola because he encouraged the provincials to adopt Roman culture, he praises him for following the tendency of his age, not for striking out any novel line of his own. It is probable that by the end of the first century, Roman civilization was laying firm hold on all the British lowlands.

Subsequent progress was slower, or at least less showy. Little advance was made beyond the lowlands. Towns and ‘villas’ were rare west of the Severn, and save in the vale of York they were equally rare north of the Trent. The uplands remained comparatively unaffected. Their population, as recent excavations in Cumberland and in Anglesey have shown, used Roman objects and came to some extent within range of Roman culture. But it seems impossible to speak of them as fully civilized, even if, in the later years of the Roman occupation, they did not remain wholly barbarian. In the lowlands we may ascribe to the second and third centuries the development of the rural system and the building of farmhouses and country residences constructed in Roman fashion. It is very difficult to date these houses. But the evidence of coins seems to show that the end of the third and the first half of the fourth century were the periods when they were most numerous and most fully occupied, and when, as we may fairly argue, the countryside of Roman Britain was most fully permeated with Roman culture. For such a conclusion we shall have the support of a neighboring parallel in Gaul.

The administration of the civilized part of Britain, while of course subject to the governor of the whole province, was in effect entrusted to the local authorities. Each Roman municipium and colonia ruled itself, including a territory which might be as long and broad as a small English county. Some districts probably belonged to the Imperial Domains and were ruled by local agents of the Emperor; such, probably, were the lead-mining districts, as on Mendip or in Derbyshire or Flintshire. The remainder of the country, by far its largest part, was divided up, as before the Roman conquest, among the native cantons or tribes, now organized in more or less Roman fashion: each tribe had its council (ordo) and tribal magistrates and its capital where the tribal council met. Thus, the tribe or canton of the Silures, the civitas Silurum, as it learnt to call itself, had its capital at Venta Silurum, Caerwent (between Chepstow and Newport); there its council met and decreto ordinis, by decree of the council, measures were taken for the government of the tribal area which probably covered much of Monmouthshire and some of Glamorgan. This, we know by epigraphic evidence, occurred at Caerwent and we shall not be rash in assuming, on slighter evidence, that the same system obtained in other tribal areas in Britain. It is just the system which Rome applied also to the local government of Gaul north of the Cevennes: it illustrates well the Roman method of entrusting local government to a restricted form of Home Rule.

Towns

In the social fabric of Romano-British life, the two chief elements were the town and the country house or ‘villa’. Both are mainly Roman importations. The Celts do not appear to have reached any definite urban life, either in Gaul or in Britain, before the coming of the Romans, though they no doubt had, even in Britain, agglomerations of houses which came near to being towns. But with the Roman conquest a real town life arose. In part, this was directly created by the government under the Roman forms of municipium and colonia, noticed above. Colchester (Camulodunum), Lincoln (Lindum), Gloucester (Glevum), York (Eburdcum), were coloniae; the first three were founded in the first century by drafts of time-expired soldiers and the fourth, York, probably grew out of the ‘civil settlement’ on the west bank of the Ouse which confronted the legionary fortress under the present Cathedral and its precincts. One town Verulamium (St Albans) was a municipium, ranking with the four coloniae in privilege and standing but different (as explained above) in origin. All these five towns attained considerable prosperity, and in particular Camulodunum, Eburacum, and Verulamium, but none can vie with the more splendid municipalities of other provinces.

Besides them, Roman Britain could show a larger number—some ten or fifteen, according to the standard adopted—of country-towns which varied much in size but possessed in their own way the essential features of urban life. The chief of these seem to be the following: (1) Isurium Brigantum, capital or chef-lieu of the Brigantes, now Aldborough, some twelve miles N.W. of York and the most northerly Romano-British town properly so called, (2) Ratae, capital of the Coritani, now Leicester, (3) Viroconium—so best spelt, not Uriconium—capital of the Cornovii, now Wroxeter, on the Severn, five miles below Shrewsbury, (4) Corinium, capital of the Dobuni, now Cirencester, (5) Venta Silurum, already mentioned, (6) Isca Dumnoniorum, capital of the Dumnonii, now Exeter, (7) Durnovaria, capital of the Durotriges, now Dorchester in Dorsetshire, (8) Venta Belgarum, capital of the Belgae, now Winchester, (9) Calleva Atrebatum, capital of the Atrebates, close to Silchester, (10) Durovernum Cantiacorum, capital of the Cantii, now Canterbury, (11) Venta Icenorum, capital of the Iceni, now Caister by Norwich, and perhaps—for the limits of the list are not easily drawn with rigidity—Chesterford (Roman name unknown) in Essex, Kenchester (Magna) in Herefordshire, Chesterton (Durobrivae?) on the Nen, Rochester (also Durobrivae) in Kent, and even one or two which have perhaps less right to inclusion. Many of these town are indicated by the Ravenna Geographer as holding some special rank and nearly all are declared by their remains to be the sites of really Romanized town-life. What exactly their status or government was, has yet to be defined. But it is fairly probable—especially from the Caerwent monument erected by the ordo civitatis Silurum—that the authorities of town and tribe were one.

The general fashion of these towns has been revealed to us by excavations at Silchester and Caerwent. At Silchester, the whole 100 acres within the walls have been systematically uncovered during the last twenty years and the buildings studied with especial care. At Caerwent, a smaller area (39 acres) has been excavated so far as the buildings of the present village permit. Both show much the same features, with certain differences in detail which are both natural and instructive: (I) Both have been planned according to the Roman method, which obtained in many parts of the Empire: that is, the streets run at right angles, so as to form a chessboard pattern with square plots for the houses. At Silchester, where space was obviously abundant, the sanctity of the street frontages seems to have been in general observed: at Caerwent, which is of smaller size and more thickly crowded with buildings, the street plan has suffered some encroachments, but not so much as to obliterate its character. (II) Both towns had near their centre the Town Buildings known as Forum and Basilica. At Silchester the Forum was a rectangular plot of two acres, with streets running along all its four sides. It contained a central open court, nearly 140 feet square, surrounded on three sides by corridors or cloisters with rooms—presumably shops and lounges—opening into them; on the fourth side was a pillared hall, 270 by 58 feet in floor space, decorated with Corinthian columns, marble lined walls, statues, and the like, and behind this hall a row of rooms which probably served as offices for the town authorities and the like. The Caerwent Municipal Buildings were very similar: so (as far as we can tell from imperfect finds) were those at Cirencester and Wroxeter. They are indeed examples of a type which was represented in most large towns of the western Empire and in Italy itself. (III) Both towns had in addition small temples in different quarters within the walls and at Silchester a small building close to the Forum is so similar in every detail to the early Christian church of the western basilican type, that we can hardly hesitate to call it a church. (IV) Both towns, again, seem to have had Public Baths: those at Silchester covered an area of 80 by 160 feet in their earliest form and in later times were much extended. Both again had more direct provision for amusements. At Silchester an earthen amphitheatre stood outside the walls: at Caerwent there are traces of the stone walls of one inside the ramparts. (V) Of dwelling-houses and shops and the like both towns had naturally no lack. The private houses are built like most of the private houses in the Celtic part of the Empire, in fashions very dissimilar from anything at Pompeii or Rome, but are fitted in Roman style with mosaics, hypocausts, painted wall-plaster, and the like. They are especially noteworthy as being properly ‘country houses’, brought together to form a town perforce, and not ‘town houses’ such as could be used to compose regular rows or terraces or streets. Even the architecture thus declares that the town life of these cantonal chef-lieux, though real, was incomplete.

The civilization of the towns appears to have been of the Roman type. Not only do the buildings declare this: inscriptions, and, in particular, casual scratchings on tiles or pots which can often be assigned to the lower classes, prove that Latin was both read and written and spoken easily in Silchester and Caerwent. Whether Celtic was also known, is uncertain: here evidence is totally lacking. But it may be observed that if Celtic was understood, one would expect to meet it, quite as much as Latin, on casual sgraffiti, while the total disappearance of a native tongue can be paralleled from southern Gaul and southern Spain and is not incredible in towns. Nor do the smaller objects found at Silchester and Caerwent show much survival of the Late Celtic art which prevailed in Britain in the pre-Roman age and which certainly survived here and there in the island. But while Romanized, these towns are not large or rich. It has been calculated that Silchester did not contain more than eighty houses of decent size, and the industries traceable there—in particular, some dyers’ furnaces—do not indicate wealth or capital. The Romano-British towns, it seems, were assimilated to Rome. But they were not powerful enough to carry their Roman culture through a barbarian conquest or impose it on their conquerors.

Country Houses, Villages

From the town we pass to the country. This seems to have been divided up among estates commonly (though perhaps unscientifically) styled ‘villas’. Of the residences, etc. which formed the buildings of these estates many examples survive. Some are as large and luxurious as any Gaulish nobleman's residence on the other side of the Channel. Others are small houses or even mere farms or cottages. It is difficult, on our present evidence, to deduce from these houses the agrarian system to which they belonged, save that it was plainly no mere slave system. But it is clear from the character of the residences and the remains in them that they represent the same Romanized civilization as the towns, while a few chance sgraffiti suggest that Latin was used in some, at least, of them. A priori, it is not improbable that, while the towns were Romanized, the countryside remained to some extent Celtic or bilingual. But all that is certain as yet is that scanty evidence proves some knowledge of Latin. These country houses were very irregularly distributed over the island. In some districts they abounded and included splendid mansions: such districts are north Kent, west Sussex, parts of Hants, of Somerset, of Gloucestershire, of Lincolnshire. Other districts, notably the midlands of Warwickshire or Buckinghamshire, contained very few ‘villas’ and indeed, as it seems, very few inhabitants at all. The Romans probably found these latter districts thinly peopled and they left them in the same condition.

Besides country houses and farms, the countryside also contained occasional villages or hamlets inhabited solely by peasants; such have been excavated in Dorsetshire by the late General Pitt-Rivers. These villages testify, in their degree, to the spread of Roman material civilization. However little their inhabitants understood of the higher aspects of Roman culture, the objects found in them—pottery, brooches, etc.—are much the same as those of the Romanized towns and "villas" and are widely different from those of the Celtic villages, such as those lately excavated near Glastonbury, which belong to the latest pre-Roman age.

The province was, on the whole, well provided with roads, some of them constructed for military purposes, some obviously connected with the various towns: whether any of them follow lines laid out by the Britons before AD 43 is more than doubtful. In describing them, we must put aside all notion of the famous ‘Four Great Roads’ of Saxon times. That category of four roads was a medieval invention, probably dating from the eleventh or twelfth century antiquaries, and the names of the roads composing it are Anglo-Saxon names, some of which the inventors of the ‘Four Road’ plainly did not understand. If we examine the Roman roads actually known to us, we discern in the English lowlands four main groups of roads radiating from the natural geographical centre, London, and a fifth group crossing England from north-east to south-west. The first ran from the Kentish ports and Canterbury through the populous north Kent to London. The second took the traveler west by Staines (Pontes) to Silchester and thence by various branching roads to Winchester, Dorchester, Exeter, to Bath, to Gloucester and south Wales. A third, known to the English as Watling street, crossed the Midlands by Verulam to Wall near Lichfield (Letocetum), Wroxeter, Chester (Deva) and mid and north Wales: it also, by a branch from High Cross (Venonae) gave access to Leicester and Lincoln. A fourth, running north-east from London, led to Colchester and Caister by Norwich and (as it seems) by a branch through Cambridge to Lincoln. The fifth group, unconnected with London, compromises two roads of importance. One, named ‘Fosse’ by the English, ran from Lincoln and Leicester by High Cross to Cirencester, Bath and Exeter. Another, probably called Ryknield street by the English, ran from the north through Sheffield and Derby and Birmingham (of which Derby alone is a Roman site) to Cirencester and in a fashion duplicated the Fosse. There were also other roads—such as Akeman street, which crossed the southern Midlands from near St Albans by way of Alchester (near Bicester) to Cirencester and Bath — which must be considered as independent of the main scheme. But, judged by the places they served and by the posts along them, the five groups above indicated seem the really important roads of southern or non-military Roman Britain.

AD 286-296] Roads, Sea Communications

The road systems of Wales and of the north were military and can best be understood from a map. In Wales, roads ran along the south and north coasts to Carmarthen and Carnarvon, while a road (Sarn Helen) along the west coast connected the two, and interior roads—especially one up the Severn from Wroxeter and one down the Usk—connected the forts which guarded the valleys: these roads, however, need further exploration before they can be fully set out. In the north, three main routes are visible. One, starting from the legionary fortress at York, ran north, with various branches, to places on the lower Tyne, Corbridge, Newcastle (Pons Aelius), Shields. Another, diverging at Catterick Bridge from the first, ran over Stainmoor to the Eden valley and the Roman Wall near Carlisle. A third, starting from the legionary fortress at Chester (Deva) passed north to the Lake country and by various ramifications served all that is now Cumberland, Westmorland and west Northumberland. Several of these roads appear, as it were, in duplicate leading from the same general starting-point to the same general destination, and no doubt, if we knew enough, we should find that one of the two routes in question belonged to an older or a later age than the other.

Communications with the Continent seem to have been conducted chiefly between the Kentish ports and those of the opposite Gaulish littoral, and in particular between Rutupiae (Richborough, just north of Sandwich) and Gessoriacum, otherwise called Bononia, now Boulogne. There was also not infrequent intercourse between Colchester and the Rhine estuary, to which we may ascribe various German products found in Roman Colchester, though not elsewhere in Roman Britain. On occasion men also reached or left the island by long sea passages. Troops, it appears, were sometimes shipped direct from Fectio (Vechten, near Utrecht), the port of the Rhine, to the mouth of the Tyne in Northumberland, while traders now and then sailed direct from Gaul to Ireland and to British ports on the Irish Channel. The police of the seas was entrusted to a classis Britannica, which intermittent references in our authorities show to have existed from the middle of the first century (that is from the original conquest or soon after) till at least the end of the third century. Despite its title, the principal station of this fleet was not in Britain but at Boulogne, and its work was the preservation of order on either coast of the Straits of Dover. This fleet appears to have been a police flotilla rather than a naval force, but for once it emerged into the political importance which fleets often assume. About 286 a Menapian (i.e. probably, Belgian) by name Carausius became commandant, possibly with extended powers to cope with the increasing piracy; he set himself up as colleague to the two reigning emperors, Maximian and Diocletian, enlarged his fleet, allied himself with the sea-robbers, and in 289 actually extorted some kind of recognition at Rome. But in 293 he was murdered and his successor Allectus was crushed by the Emperor Constantius Chlorus in 296. Carausius was apparently an able man. But in his aims he differed little from many other pretenders to the throne whom the later third century produced: his object was not an independent Britain but a share in the government of the Empire. His special significance is that he showed, for the first time in history, how a fleet might detach Britain from its geographical connection with the north-western Continent. Twelve centuries passed before this possibility was again realized.

AD 300-380

The preceding paragraphs have described the main features of Roman Britain, civil and military, during the main part of its existence. In the fourth century, change was plainly imminent. Barbarian sailors, Saxons and others, began, as we have seen, rather earlier than 300 to issue from the other shores of the German Ocean and to vex the coasts of Gaul and probably also those of Britain. Carausius in 286 or 287 was sent to repress them. After his and his successor's deaths, some change, the nature of which is not yet quite clear, was made in the classis Britannica, and we now hear hardly anything more of it. A system of coast defence was established from the Wash to the Isle of Wight. It consisted of some nine forts, each planted on a harbor and garrisoned by a regiment of horse or foot. The ‘British Fleet’, so far as Britain was concerned, may have been divided up amongst these forts or may have been entirely suspended. But it is difficult to make out (owing to the general obscurity) whether the change was made in the interests of coast defence or as a preventive against another Carausius. The new system was known — from the name of the chief assailant — as the Saxon Shore (Litus Saxonicum).

Whatever the step and whatever the motive, Britain appears for a while to have escaped the Saxon pillages. During the first years of the fourth century, it enjoyed indeed considerable prosperity. But no Golden Age lasts long. Before 350, probably in 343, the Emperor Constans had to cross the Channel and drive out the raiders—not Saxons only, but Picts from the north and Scots (Irish) from the north-west. This event opens the first act in the Fall of Roman Britain (343-383). In 360 further interference was needed and Lupicinus, magister armorum, was sent over from Gaul. Probably he effected little: certainly we read that in 368 all Britain was in evil plight and Theodosius (father of Theodosius I), Rome's best general at that time, was dispatched with large forces. He won a complete success. In 368 he cleared the invading bands out of the south: in 369 he moved north, restoring towns and forts and limites, including presumably Hadrian’s Wall. So decisive was his victory that one district—now unfortunately unidentifiable—which he rescued from the barbarians, was named Valentia in honor of the then Emperor of the West, Valentinian I. For some years after this Britain disappears from recorded history, and may be thought to have enjoyed comparative peace.

Such is the account given us by ancient writers of the period circa 343-383. It sounds as though things were already “about as bad as they could be”. But a similar tale is told of many other provinces, and yet the Empire survived. When Ausonius wrote his Mosella in 371, he described the Moselle valley as a rich and fertile and happy countryside. Britain had no Ausonius. But she can adduce archaeological evidence, which is often more valuable than literature. The coins which have been found in Romano-British ‘villas’, ill-recorded as they too often are, give us a clue. They suggest that some country houses and farms were destroyed or abandoned as early as 350 or 360, but that more of them remained occupied till about 385 or even later. It is not surprising to read in Ammianus that about 360 Britain was able to export corn regularly to northern Germany and Gaul. The first act in the Fall of Roman Britain contained trouble and disturbance, no doubt, but few disasters.

AD 380-410

The second act (383 to about 410) brought greater evils and of a new kind. In 383 an officer of the British army, by birth a Spaniard, by name Magnus Maximus, proclaimed himself Emperor, crossed with many troops to Gaul and conquered western Europe: in 387 he seized Italy: in 388 he was overthrown by the legitimate Emperors. Later British tradition of the sixth century asserted that his British troops never returned home and that the island was thus left defenseless. We cannot verify this tradition. But we have proof, both that Britain was sore pressed and that the central government tried to help it. Claudian alludes to measures taken by Stilicho, prime minister to the then Emperor Honorius, about 395-8. Archaeological evidence shows that the coast-fort of Pevensey (Anderida) was repaired under Honorius, and that a fort was built high on the summit of Peak, overhanging the Yorkshire coast half­way between Whitby and Scarborough, by an officer of the same period who is known to have been in Britain a little after 400. These efforts were in vain. Troops—not necessarily legionaries though Claudian calls them legio—had to be withdrawn for the defense of Italy in 402. Finally, the Great Raid of barbarians who crossed the Rhine on the winter's night which divided 406 from 407 and the subsequent barbarian attack on Rome itself cut Britain off from the Mediterranean. The so-called ‘departure of the Romans’ speedily followed. This departure did not mean any great departure of persons, Roman or other, from the island. It meant that the central government in Italy now ceased to send out the usual governors and other high officials and to organize the supply of troops. No one went: some persons failed to come.

How far the British themselves were responsible for, or even agreeable to, this sundering of an ancient tie is, even after the latest inquiries, not very certain. The old idea that Britons and Romans were still two distinct and hostile racial elements has, of course, been long abandoned by all competent inquirers—for reasons which the preceding pages will have made evident. But we have the names of three usurpers who tried to seize the imperial crown in Britain (406-11), Marcus, Gratian, and Constantine, and it seems that, as Constantine went off to seek a throne on the Continent, the Britons left to themselves set up a local autonomy for self-protection. Unfortunately, our ancient authorities are less clear than could be wished, especially on the chronology of these events. One thing which seems certain is that Britain did not conceive herself as breaking loose from the Empire and that in the years to come the Britons considered themselves ‘Romans’. If we may believe Gildas, they even appealed for help to Aetius, the Roman minister, in 446.

Saxon Conquest [AD 300-446

The attacks of the ‘Saxons’ had begun before 300 and though at first their brunt fell more heavily on the Gaulish than on the British coasts, they were felt seriously in Britain from about 350 onwards. At first, they were the attacks of mere pillagers: later, like the later attacks of the barbarians elsewhere, they became invasions of settlers. When exactly the change took place, is unknown, nor is it clear what incident gave the stimulus. It seems probable, however, that the Britons of the early fourth century, harassed by attacks of all kinds, adopted the common device—even more familiar in that age than in any other—and set a thief to catch a thief. The man who set is named in the legends Vortigern of Kent; the thieves who were set, are called Hengest and Horsa. We need not attach much weight to these names, nor can we hope to fix a precise date. But the incident is sufficiently well attested and sufficiently probable to find acceptance, and it obviously occurred early in the fifth century. It had the natural result. The English, called in to protect, remained to rule: they formed settlements on the east coast and began the English invasion. But they began it under conditions altogether different from those which attended the barbarian conquests on the Continent. The English were more savage and hostile to civilization than most of the continental invaders; on the other hand, they were far less overwhelmingly numerous. The Romano-British culture was less strong and coherent than the civilization of Roman Gaul, but the Britons themselves—at least those in the hills—were no less ready to fight than the bravest of the continental provincials. The sequel was naturally different in the two regions.

The course of the invasion is a matter for English historians. But part of it depends on Romano-British archaeology. This seems to contradict violently the chronology which the Anglo-Saxon Chronicle sets out in suspiciously precise detail. We know that Wroxeter was burnt and we have evidence that the burning occurred soon after (if indeed it was not before) AD 400. We must treat this evidence cautiously, since not a fiftieth part of the site has yet been explored. But at Silchester, which has been all uncovered, the spade has told us that the town was abandoned (not burnt), and as a limit for the date, we find no coins which need be later than about AD 420. The same absence of fifth century coins may be noted on other sites which have been sufficiently explored to yield trustworthy testimony. It would seem as if the invaders, entering Britain on its eastern and least defensible side, were able, like the Romans four centuries earlier, rapidly to sweep over the lowlands, but were not able to maintain their hold. Thus for several generations this region became a debatable land, where neither Romano-British city life could safely endure nor the English take firm hold and settle. In the long confusion, the Romano-British civilization of the lowlands perished. The towns, burnt or abandoned, lay waste and empty. Even Durovernum (Canterbury), presumably the capital of Vortigern, whom the legend mates with a Saxon wife, ceased to exist, and at the healing springs of Aquae Sulis (Bath) the wild birds built their nests in the marsh which hid the ruins. The country houses and farms perished even more easily: not one is known in which we can trace English inhabitants succeeding to British. The old native tribal areas and the Roman administrative boundaries were alike lost: today we have no certain knowledge of any of them. The Roman speech vanished; the Romano-British material civilization, and the house-plans and house-furniture, hypocausts and mosaics, even the fashions of brooches and pottery, vanished with it. Only the solid aggeres of the roads remained still in use, and in these, too, there were gaps and intervals. All else was but the scattered débris of a ruined world.

Meanwhile the Romanized Britons, in losing the lowlands, lost their towns and all the apparatus of town life. They retired into the hills, to Wales and to the north—the later Strathclyde—and there, in a region where Roman civilization had never established itself in its higher forms, they underwent an intelligible change. The Celtic element, never quite extinct in those hills and reinforced perhaps by immigrations from Ireland, reasserted itself afresh. Gradually, the remnants of Roman civilization were worn down: the Celtic speech reappeared and, as sequel, the Late Celtic art was strong enough to pass on an artistic legacy to the Middle Ages.