English Heritage


Year published


Supporting documents




Comment on this article
Double click anywhere on the text to add an annotation in-line

Citation Show another format:

'Introduction', An Inventory of the Historical Monuments in City of York, Volume 1: Eburacum, Roman York (1962), pp. XXIX-XLI. URL: Date accessed: 19 April 2014. Add to my bookshelf


(Min 3 characters)



The Roman Legionary Fortress and City at York

The first known appearance of Roman troops in permanent quarters at York, under the Emperor Vespasian, goes with the extension of the province of Britain brought about in A.D. 71–4 by the suppression and conquest of the rebellious client-kingdom of the Brigantes. (fn. 1) The thrust was made by the governor Petilius Cerialis and his army; and the contemporary occupation detected at Brough and Malton suggests (fn. 2) that the main line of advance lay through the East Riding, then the territory of the Parisi, along the western slope of the Yorkshire Wolds, which lay high and dry above the lower basin of the Ouse (Fig. 1). In that basin the position occupied by York is of the greatest strategic importance. For there the river Ouse cuts through a glacial moraine (fn. 3) which forms a low but substantial natural causeway across the wide and marshy valley. The crossing not only affords easy access to the West Riding of Yorkshire and its dales, but furnishes the necessary south-westward connection with the Welsh March and the Roman legionary garrisons (fn. 4) stationed along it. York is thus the key-position first for the control of the Brigantes, secondly for surveillance of the hinterland of East Yorkshire and Lincolnshire and thirdly for connection with the western military area. It was picked with that unerring eye (fn. 5) for the pre-eminent strategic position which places Roman reconnaissance throughout the Empire in a class of its own. In this case, as in others, it is, however, very likely that the choice was no snap decision but that the advantages of the position were known beforehand. The client-kingdom had been penetrated at least three times (fn. 6) by expeditionary forces, quite apart from visits made to the Brigantes by Roman military officers and their escorts on this or that peaceful mission. But whatever the preliminaries, admiration for the choice remains unstinted and is further enhanced by the reflection that the Roman initiative once and for all determined the position of the Northern Command in England through every century between then and now.

The name of the Roman site was Eboracum, so spelt (fn. 7) by the geographer Ptolemy and in three inscriptions giving the second syllable; the form Eburacum, found in the Antonine Itinerary and the Ravenna Cosmography, being once attested epigraphically. The canting emblem of the town was the boar which appears on the Bordeaux altar, (fn. 8) and evidently this was the popular derivation, as in the Saxon form Eoforwic. But the form Eburacum, accepted by philologists, (fn. 9) is cognate with Irish ibhar, iubhar, 'yew'—and this, combined with -acum, may mean 'the place of yews' or 'the field of Eburos', a personal name derived therefrom.

From the first the fortress was of normal size for a legion, covering 50 acres and designed as a rectangle with rounded angles, 1370 ft. by 1590 ft. over the rampart (Fig. 3). There is good evidence that the rampart of the earlier fortress, which has been identified on the south-west front and at the east angle, was rapidly constructed with materials immediately available. Green boughs were lopped, trimmed and laid down as a 'corduroy' foundation for a rampart of clay or sand with turf front (p. 20), all rapidly constructed by the legionaries in the fashion so vividly portrayed in the reliefs of Trajan's Column. (fn. 10) That the occupying force was the Ninth Legion, whose title Hispana derived from distinguished earlier service in Spain, (fn. 11) is not to be doubted. This regiment had been commanded by Cerialis a decade earlier, during the black days of the Boudiccan revolt; and he was now choosing it, as governor and commander-in-chief of the province, to be the spearhead of his attack. Support was given by the Twentieth, (fn. 12) whose commanding officer was Julius Agricola, destined himself to be governor of the province in due course. Lincoln, the old fortress of the Ninth, was now allotted (fn. 13) to the Second Adiutrix, a newly-levied legion brought to Britain by Cerialis from its first bout of service in the Rhineland revolt: and it has been pointed out that the earliest coinage (fn. 14) from Roman York, distinctively rich in Claudian small change, may represent the old savings-deposits of the Ninth, brought to York on removal. Outside and beyond the fortress there was much to do. New forts must have been planted in the dales. The courses for a new road-system penetrating the territory of the Brigantes must have been chosen, surveyed and cleared. Such roads cannot yet have been the metalled arterial routes, up to 30 ft. wide, which later fashioned the garrisons of northern Britain into a web from which there was no escape. But the first preparations, in the form of limites or cleared tracks through forest and undergrowth, with embankments or 'corduroys' in wet spots, must have seemed sufficiently raw and frightening, representing hard and bitter work for the native corvées that laboured under relentless Roman supervision. Tacitus, who no doubt had heard of such operations at first hand from his father-in-law Agricola, gave them due prominence in his imaginative picture (fn. 15) of the burdens of the vanquished. For it was in fact Agricola himself who between A.D. 79 and 85 completed the main lines of the system and initiated much more to north of it, carrying Roman permanent occupation to the Highland gates. Agricola also gave more permanent character to the defences of the York fortress, levelling the rampart of Cerialis and bedding upon its spread remains a new foundation of squared oak strapping carrying a solid rampart of stiff clay with turf-work front (p. 20), the whole at least 16 ft. wide and comparable with the 18 ft. rampart at Caerleon. (fn. 16) The work was fitted at intervals with timber towers some 11 ft. wide—if this was not, as at Inchtuthil, (fn. 17) half the width only—and about 23 ft. from back to front (p. 14). Internal buildings were also of timber constructed not with boarded walls, but with timber-framed wattle-and-daub walls and shingle roofs, insulated by the daub walling so as to be cool in summer and warm in winter. Great cooking-ovens at the back of the rampart (p. 43b, Figs. 9, 31) kept all but the smallest hearths or braziers away from the buildings and thus reduced the danger of fire. Nothing is known of gates at this period, but they must have occupied the positions taken in due course by those built in stone. If the Greek inscription (INSCRIPTIONS etc., No. 142) mentioning the governor's residence belongs to this time, as is likely, its findspot indicates a situation for this Residency on the right bank, across the Ouse from the legionary fortress. It would represent Agricola's advanced headquarters from which his campaigns were organised.

The rebuilding of the defences and internal buildings in stone was part of a general consolidation applied to the three legionary fortresses of Britain in the first half of Trajan's reign. The operation had begun at Caerleon, (fn. 18) the fortress of the Second Legion Augusta, in A.D. 99–100; it was in progress at Chester (fn. 19) after A.D. 102, while at York it is recorded by an inscription of A.D. 107–8 (INSCRIPTIONS etc., No. 1). As regards the defences at least the chronological order indicated by the inscriptions matches that of their original construction, and it is likely that all the timber structures were by now in need of renewal. The consolidation in stone, however, did not mark merely a routine repair: it signalised the final selection of the three fortresses in Britain as the permanent headquarters of their respective legions. The advantages of York as a centre of land-communications have already been described. But it must be emphasised that the Ouse (fn. 20) was long navigable up to this point for sea-going ships of the day, so that bulk supplies of heavy goods could reach and leave the fortress by water. Good though Roman roads were, the vehicles which used them were springless (fn. 21) and were hauled by means of harness (fn. 22) engaging the neck instead of the chest, with consequent loss of efficiency. Water transport was thus the more valuable, and in its development internal waterways had an important part to play. By means of the Trent, Fossdyke, Witham and Car Dyke the Ouse was linked (fn. 23) with the rich reclaimed Fenland which grew so much of the corn needed by the army for its daily bread. These and other streams could bring the local raw materials, such as iron from Lincolnshire, (fn. 24) lead from Derbyshire, (fn. 25) building-stone (fn. 26) and roofing slates (fn. 27) from the West Riding and timber from the Vale of York itself. Into the Humber came from overseas the heavy cargoes of wine and oil; from its mouth troops and supplies coming down river from York could leave for the ports and store-bases of the East Coast, to succour or to revictual the inland garrisons. It is in this context that the soldier-pilot of the Sixth Legion is to be understood in a somewhat later age (see INSCRIPTIONS etc., No. 36). The development of these facilities must have sealed the choice of position.

The inscription (ibid., No. 1) which records the Trajanic building operations came from the porta principalis sinistra, or south-east gate of the fortress. It is a work of art in its own right (Pl. 41), of especial interest because the stone-cutter has copied with fidelity and skill the letter-forms of an elegant draught written with pen or brush and reflecting the superb standards of the legionary clerks, draughtsmen or signwriters when their work attained monumental scale. The altar dedicated to Silvanus by a cornicularius, or staff-clerk, of the Ninth reflects comparable standards (fn. 28) applied with individuality to a less conspicuous stone (INSCRIPTIONS etc., No. 32). The gateway connected with the Trajanic inscription lies buried below the modern houses and street surfaces of King's Square, and only at the south-west gate, on the river front, have remains identifiable with the Trajanic gateway been discovered (Fig. 6). These reveal one half of a twin portal and one of the two guard-chamber basements, and the whole plan suggests a massive work with upper storey carried right across the gate and not necessarily carried higher by towers. Interval towers and an angle-tower of stone are also known, while the curtain-wall, 5 ft. thick on a foundation of grouted cobbling, is built of magnesian limestone, no doubt from the Tadcaster (fn. 29) area. Nothing certain is known of internal buildings, excepting stores-buildings situated on the intervallum, or open space just inside the rampart, and identified on both the south-west and north-west sides of the fortress as of Trajanic date. The barrack-buildings found in Davygate appear to be later, and it may be recalled that at Chester (fn. 30) the Trajanic barracks were still built with timber frames, though the foundation-sills for these were set high and dry upon stone sill-walls.

The Trajanic building-inscription is the latest dated record of the Ninth Legion. Under Hadrian it was replaced by the Sixth Legion, (fn. 31) brought across from Germany by Platorius Nepos, whose governorship (fn. 32) began in A.D. 122. No ancient source describes how the Ninth Legion came to disappear from the Roman army list, and the matter has given rise to varied conjecture. The circumstantial evidence, such as it is, may be summarised as follows. There was successful Roman campaigning in Britain in A.D. 119, important enough to be commemorated on Hadrian's coinage (fn. 33) with the types of Iuppiter Victor and Roma Victrix, and Britannia. This is all consequent upon a state of affairs at his accession (fn. 34) in which 'the Britons could not be held under Roman rule'. The conditions are adequate to cover a disgraceful defeat of the Ninth, meriting at the hands of disciplinarian (fn. 35) Hadrian the cashiering of the unit. The later the event is placed the more difficult it is to account for, and no strain is involved in associating the disbanding of the unit, whose staff officers (fn. 36) at least survived, with the visit (fn. 37) of Hadrian to Britain in A.D. 121–2. To assume, as has been done, that the legion was annihilated or that it continued to exist, toiling in unrecorded obscurity upon the works of Hadrian's Wall, is an unnecessary postulate. Finally, it is certain that the Sixth Legion was henceforward stationed at York, until at least the 4th century. (fn. 38)

The next historical event reflected in the buildings of the fortress is the reconstruction of its defences and internal buildings in 197 under Severus, following the devastation of the North early in that year by the tribes of central and north Scotland. This disaster had been brought about by the withdrawal of the garrison of northern Britain to fight in Gaul for Albinus, (fn. 39) governor of Britain and rival of Severus for the principate. When Albinus was defeated, and before any troops could return, the tribes took their opportunity. As a result of their destruction the Trajanic wall seems to have been removed to its foundations almost everywhere, except at one known point where its lower courses were useful for levelling up. A new and more imposing wall was built in its stead (Fig. 4). This wall, nearly 6 ft. thick, and buried below the north-east defences of the mediaeval city, survives almost to full height at the east angle, where it is seen to emerge from below them. It consists of a single course of footing-flags, upon which stands a front footing-course and chamfered plinth, both of massive masonry. These carry thirty or more courses of small ashlar, capped by a chamfered string-mould and a single course at parapet-walk level, 15 ft. high. The rear face of the wall rises in small ashlar courses, surviving to a height of 12½ ft. at most, with offsets at about 4 ft., 4½ ft. and 7 ft. Solid grouting occupies the space between the faces, and the wall was reinforced at the back against battering-rams by a massive clay bank, which carried a cobbled rampartwalk, some 5 ft. lower than the parapet-walk, from which the latter could be serviced. This was a very necessary provision if active defence was required. The towers also were rebuilt and equipped for artillery defence, though this, as at High Rochester (fn. 40) beyond Hadrian's Wall, may not have come until a little later in the century. The principal rebuild is to be dated as soon in A.D. 197 as it could take place. In the spring of that year Virius Lupus recovered (fn. 41) Britain for Severus. The first essential step was the reconstruction of the legionary fortress, and since forts (fn. 42) were being rebuilt in 197 at Ilkley, Bowes, Binchester, and Brough-under-Stainmore, it is reasonable to date the York reconstruction to this year also. Inside the fortress the stone barracks identified in Davygate belong to this period. Outside it, traces of massive walling detected at three widely separated points on the north-west side of the fortress have been interpreted as the boundary of an annexe (Monument (13), p. 45). Since archaeological examination has been possible at one point only, the evidence is slighter than could be wished; in particular, it is not known whether the walling ran north-east of the main road through Bootham. But the suggestion remains the more interesting and perhaps the more inherently likely in view of the comparable reserved space on the corresponding side of the fortress at Caerleon. (fn. 43) The principal developments at York during the 3rd century belong, however, to the civilian sphere and are described below.

The 3rd century closed with a disaster and restoration closely comparable with that of 197, although on this occasion the army of Britain was not withdrawn overseas but gathered in the south (fn. 44) of the island under the usurper Allectus in order to prevent its recovery for the central government by Constantius Caesar. The defeat of Allectus and the impossibility of immediately reorganising and recovering the whole province brought upon the North a second devastating invasion. Inside the fortress the subsequent reorganisation seems to have involved the rebuilding of barracks, but positive evidence of such rebuilding is confined to the four known barrack-blocks in Davygate (Fig. 29) and what happened to other barracks in the fortress is uncertain. It should, however, be emphasised that, whatever the effect of the permission to marry granted to legionaries in the 3rd century or of the reduction in size of legions in the 4th century, such legionary barracks as are known to be occupied continue in much the same form as before, both at Caerleon (fn. 45) and at York. The 4th-century bath-house in St. Sampson's Square (p. 42), which is sometimes taken to indicate the erection of new buildings where barracks once stood, does not necessarily occupy a barrack-block position. Concerning the fortress defences much more is known. They were drastically remodelled on the north-west side of the fortress and on the south-west front, where they now assumed superlatively imposing form. Six majestic interval towers and two vast angle towers, all with polygonal fronts, here dominated the river (Fig. 3., Pls. 2–9): and it must be borne in mind that so stately a reconstruction implies a new porta praetoria of the same scale, despite that fact that no intelligible evidence now survives. No fortress front in the Empire was more splendid. The scale far outweighs that of the contemporary new legionary fortresses, such as Kaiseraugst. (fn. 46) The massive emphasis calls for a special explanation. It is not to be explained by defensive precautions comparable with the bastioned forts of the Saxon Shore, for it far surpasses them in size and strength. It is certainly true that across the river the fortress confronted the town which was the capital of Britannia Inferior, and that a major architectural expression of military dominance was demanded. But what dominance was thus expressed? If it were the headquarters of the Dux Britanniarum, (fn. 47) the new commander of all the land-forces of Britain created in the 4th century, this would be an answer worthy of the staging which the immense fortifications provide. The internal arrangements of this great place-forte are obscure. The bath-house in St. Sampson's Square has already been mentioned. It is of large size (fn. 48) as such buildings go, and may well represent the principal bath-house of the fortress. There is in fact nothing abnormal in the presence of such a building within a legionary fortress: Vindonissa and Lambaesis provide good examples (fn. 49) of the practice, though it is not universal. But if all too little survives of major internal buildings, the splendid head of Constantine I (INSCRIPTIONS etc., No. 8, Pl. 42), carved in local stone and found in Stonegate, gives a hint of their official complement of Imperial statues. (fn. 50) Constantine, whose father Constantius I died in York in July, 306, after a great punitive campaign in the far North, (fn. 51) was proclaimed in York itself. It is recorded that one of his principal supporters was an Alemannic King, (fn. 52) Crocus, whose presence was probably due to levies made during the Alemannic war of Constantius. But the fact that a barbarian chief could wield such influence anticipates the Gothic tenure of the high military commands a century later, and emphasises where power was ultimately to lie. In 367–9 the Dux Britanniarum, who commanded the inland defences of the four provinces into which Diocletian had divided Britain, has himself a name with a distinctively Frankish suffix, (fn. 53) whatever its precise form. This was a time of social revolution, and one of its most marked effects is seen in the burials of the period. The remarkable feature is the use of stone sarcophagi rifled from their original tomb-chambers, where they had once been disposed in such a manner that their decoration or their inscriptions could be seen. These were now buried below ground as ordinary stone coffins, the original skeleton having been ejected to make way for a new one. This treatment was accorded to at least three sarcophagi connected with the centurionate, two from the south cemetery of the canabae and one from a position north-west of The Mount (INSCRIPTIONS Nos. 104, 107–8). The indication that these tombs were thus rifled and destroyed strongly suggests a complete rupture in tradition such as might well have come with the reorganisation of the legions under Diocletian and the break caused by the destruction of A.D. 296.

The Civil Settlement (Canabae) Outside The Legionary Fortress

The extramural settlement, or canabae (fn. 54) (meaning apparently 'the booths'), which housed the motley crowd of tradesmen and purveyors wont to gather about any large military force, lay chiefly outside the north-west and south-east sides of the fortress, including quays along the Foss, and spread in a less concentrated fashion to the north-east (see Map at end of book). Such folk were settled in the territorium legionis, (fn. 55) that is, on land legally assigned to the legion in garrison, and their holdings were on short leases, confirmed or terminated every lustrum or five-year period under supervision of the primus pilus or senior centurion of the legion. (fn. 56) The territorium covered many square miles (fn. 57) and included pasture and woodland as well as arable. The lands and farms further afield within it were held under similar conditions. As the settlement grew and prosperity increased, its buildings took on the aspect of a town, as had the extramural settlement at Vetera (fn. 58) in A.D. 69 when it was demolished in expectation of a siege. Few structural remains of the canabae at York have as yet been identified, but outside the south angle of the fortress (Fig. 37) buildings of all kinds were densely packed and here inscriptions speak of temples to Hercules and also to the Emperors' Divinities coupled with a Celtic goddess whose name is incomplete (INSCRIPTIONS etc., Nos. 52–3). Dedications to another god, otherwise unknown, named Arciaco and to a Genius Loci, together with a votive ring of silver inscribed to the Gallic god Sucellus, widen the variety (ibid., Nos. 40, 51, 140). The dedication to Hercules further mentions two of the dedicators who were not only Roman citizens but were officially connected with Eburacum, probably holding administrative offices in the vicus or civil settlement itself. (fn. 59) A collegium, probably a burial-club, is mentioned in a graffito scored on a tile (INSCRIPTIONS etc., No. 24). But the growth of the crowded modern city over exactly these areas has precluded any systematic archaeological examination of the canabae and only the distribution of relics found by chance enables an estimate of their extent to be made. Major buildings, such as the ludus or amphitheatre, (fn. 60) remain as yet undetected.

The tomb-monuments of the fortress and canabae on the left bank denote a prosperous community. Most notable is the 3rd-century head of a lady's portrait-statue from Fishergate and the monumental corner-piece in bust form from Castlegate (INSCRIPTIONS etc., Nos. 113, 130). But the architectural fragments from Bootham Bar and its vicinity, decorated with the funeral symbolism (fn. 61) of sea-monsters, a sphinx, a Dioscurus and perhaps a hunting scene, denote important monumental tombs, as does the tablet of Flavinus from Clifton (ibid., Nos. 126, 127, 122, 128, 78). The two reused sarcophagi from Castle Yard (ibid., Nos. 104, 107), one with most elegant amorino supporters, are important indications of the wealth and even the taste of the centurionate, and their ultimate fate has already been discussed. But the record of soldiers' tombstones is incomplete without certain items from the right bank; namely, the tombstones of the standard-bearer Rufinus of the Ninth Legion, of Baebius Crescens of the Sixth, and of the centurion now lacking an inscription (ibid., Nos. 75, 72, 95); also the tablet from the monument of the soldier of the Ninth from Novaria, the sarcophagus of a centurion's daughter, and a possible fragment of a cavalry tombstone (ibid., Nos. 91, 108, 99).

The Chartered Town (Colonia)

Across the river Ouse from the legionary fortress there grew up another settlement, which seems to have begun in the form of ribbon development along the south-westward road leading to Tadcaster. Two early roadside monuments associated with it are the dedication to Silvanus by Lucius Celerinius Vitalis, a clerk (cornicularius) of the Ninth Legion, and the tombstone of the standard-bearer Lucius Duccius Rufinus from the same regiment (INSCRIPTIONS etc., Nos. 32, 75), and the Sixth Legion is also represented by the tombstone of Baebius Crescens and the altar of the pilot Minucius Mudenus (ibid., Nos. 72, 36). By the middle of the 2nd century, however, the settlement was evidently growing and timber-framed buildings had spread some distance north-west of the main road and south-west of the river. Whether this settlement had lain originally in the territorium legionis or in the cantonal lands of the Brigantes, there came a moment when, having grown to maturity independently of the canabae, it received a charter making it independent. (fn. 62) A dedication by a York merchant in Bordeaux, (fn. 63) dated to A.D. 237, mentions the existence of a colonia, that is, the highest rank of chartered town, and analogy (fn. 64) would suggest that it had been upgraded from a municipium, though no epigraphic evidence has yet attested this stage in development. That the colonia (Fig. 37) lay on the south-west river-bank, however, is certainly to be deduced from the site of its cemetery (Fig. 62), which was there situated and has yielded three sarcophagi, of a decurion, of a sevir Augustalis and of his wife (INSCRIPTIONS etc., Nos. 105, 110, 106). Little is known of its buildings (Fig. 38), but they included a temple of Serapis and large Baths (Monuments 32, 34). Its status as a colonia is no doubt connected with the fact that it was the capital (fn. 65) of Britannia Inferior, the second province into which Britain was divided (fn. 66) not later than A.D. 213, which embraced the cantons of the Brigantes, the Parisi and the Coritani. It should be understood as an honorary status of rank, sharply separated in conception from the earlier military coloniae at Colchester, Lincoln and Gloucester, which, though towns of importance, were primarily land-settlements for time-expired legionaries.

The place must also have derived prestige from the choice of York as the seat of the Imperial Court of Severus and Caracalla (fn. 67) during the three years spent in mounting and conducting punitive campaigns against the Caledonii and Maeatae. A rescript (fn. 68) of 5th May, A.D. 210, came from Eboracum. Here also lived the Empress Julia Domna (fn. 69) and the Caesar Geta, who became the third joint-Emperor in A.D. 209 and administered (fn. 70) Imperial affairs with a Council while the senior Emperors were on campaign. The whole state of affairs closely resembles that under Marcus Aurelius on the Danube, when Carnuntum, (fn. 71) containing both a legionary fortress and a municipium, was for three years (172–4) the seat of the Imperial Court during the Marcomannic wars. Then, if not before, Carnuntum received an Imperial palace, (fn. 72) erected in the area between fortress and town, which was maintained thereafter and served more than a century later for an Imperial conference. At York nothing is known archaeologically of a Severan palace, but it is mentioned (fn. 73) in the Life of Severus, the more convincingly because the reference to it is incidental, and is styled palatium and domus Palatina. Thus even if its situation is unknown, its existence is sure. It is likely that it was used again by Constantius I, who died in York; (fn. 74) and perhaps even later by Constans, if his winter campaign of 343 took him so far North in person. (fn. 75) Carnuntum thus affords a valuable parallel for conditions at Eburacum and helps to make their various terms intelligible.

As the capital of Britannia Inferior Eburacum was also the seat of an Imperial governor, with his praetorium, (fn. 76) or official residence. But this legate was also legionary legate and it may well be that the legate's house in the fortress continued to serve for the purpose. Under the later Empire, however, when civil and military commands were divorced, there must have existed a civil governor's residence. For this a riverside site, like that of the praetorium of the legate of Lower Germany at Cologne, (fn. 77) would afford the most attractive amenity. The Cologne praetorium was not divorced from the colonia, but lay within its walled circuit. This arrangement might be expected at York, where there is reasonably strong evidence for fortification of the colonia, though it relates to the north-west side of the town only, where the Roman work underlies the early mediaeval bank and Plantagenet City Wall, and perhaps extended slightly beyond them at the W. corner (Monument 16). This was quite certainly the boundary of the colonia in the 4th century, when burials in quantity are known to have crowded close up to these limits (see BURIALS: IV Region, Area (a), p. 80). Within the town there is good evidence for stone buildings of this period, with good mosaic pavements and heated rooms, spreading widely on both sides of the Roman main street (Fig. 38). But the southward extent of the town remains undetermined, and excavation is much to be desired. Monumental buildings are in evidence, but knowledge of their plans is fragmentary and their date within the Roman period remains unsure. Certain it is that the large Baths (Monument 34) in the north-west area of the town were rebuilt, but it can hardly be determined whether the reconstructed establishment was as spacious as its forerunner. The date of the great colonnaded building (Monument 30) in Railway Street is unknown, though it presumably belongs to the colonia and is manifestly part of a great public building of majestic proportions, such as a basilica: it contains, however, if the meagre record is correct, an architectural solecism in the form of a central column in the short side where convention would require an intercolumniation. Other substantial buildings were found under the Railway Offices (Monument 31), but the recorder of detail was always lacking. The elaborate architectural fragments (INSCRIPTIONS etc., Nos. 13, 14) from the foundations of the mediaeval Ouse Bridge show that in the Middle Ages important fragments were still available to be robbed for their stone, though no mediaeval chronicler mentions them.

Religious dedications from the area of the colonia are prolific and it is likely that the majority belong to the 3rd century. An exception is the Greek dedication linked with the grammar-school master Demetrius of Tarsus, in Britain shortly before A.D. 84 (ibid., No. 142). It will be noted also that major official dedications connected with the public worship of the town are absent, the only approach towards this sphere of religious life being an altar which couples the Emperor's divinity with the Genius of Eboracum (ibid., No. 35). This is not because such things were lacking: seviri Augustales, for example, are well attested (ibid., No. 110). But time and man have reduced the principal buildings to the point of extinction. The surviving altars are all private dedications. The Matres are honoured four times, Mars twice, the German god Huitris and a Genius loci once each; special mention should be made of the altars to Jupiter, with the deities of hospitality and home, by a commander of an auxiliary cohort and to Fortune by the wife of a legionary legate, this a 2nd-century stone reused (see Index, under ALTARS passim). Statues of Fortune and of Britannia, the latter dedicated by an Imperial freedman, are known from inscriptions (INSCRIPTIONS etc., Nos. 56–7). The very fine life-size Mars in stone is almost complete (ibid., No. 59). The worship of Mithras has evoked a notable bull-killing tableau, with scenes alluding to grades of initiation, and a remarkable figure of Arimanius, (fn. 78) whose lion-headed figure is not unfamiliar in sculpture but to whom named dedications are rare (ibid., Nos. 67, 58). The findspots of a relief of Mercury and of a remarkably powerful Celtic head (fn. 79) are unknown, but unlikely to lie outside York (ibid., Nos. 68, 70).

The cemeteries spread widely. While the official burials already mentioned probably belong to the 3rd century, 4th-century interments extend all over the Railway Station area, and for a great distance along the south-westward road, where the burial-places of poorer folk are represented by the Trentholme Drive cemetery, those of wealthier citizens by sarcophagi found closer to the town (Figs. 62, 70). Fortune has indeed been lavish in discoveries of funerary stones associated with the colonia or its area. Funeral-banquet scenes are represented by the stones of Aeliana, Velva, Mantia and her relatives, and by a stone without a text (INSCRIPTIONS etc., Nos. 71, 82, 84, 98). Standing figures are the Augustini family, Julia Brica, Corellia Fortis, the man and boy, the togate family, the smith at work, the man with writing-tablets, and a male figure, all but the first three lacking texts (ibid., Nos. 77, 80, 73, 85, 97, 96, 100–1). Busts are known of Decimina and Julia Secunda, and of an anonymous male, also two fragments of male heads and of an imago clipeata (ibid., Nos. 74, 81, 117, 116, 118, 125). Upright tombstones with conventional ornament only are those of Hyllus the foster-child, Julius and Felix, ...]abinia, Manlius Crescens, Minna, and the child of Vitellia Procula (ibid., Nos. 79, 83, 86, 88–90). Fragments exist of a large altar-tomb (ibid., No. 131). Tablets originally walled into monumental structures are those of Eglecta, with its amorino supporters, the legionary of the Ninth from Novaria, and that with legal directions (ibid., Nos. 76, 91, 94). The figure of a mourning Atys (ibid., No. 124) probably flanked a door of a monument, as at Chester, (fn. 80) and it must be recalled that many of the pieces which now have the look of freestanding upright grave-stones were in fact walled centrally into the façades of small monumental tombs of round or rectangular plan, as discoveries at Carnuntum (fn. 81) show. Then come the sarcophagi, themselves often intended for a tomb-chamber: of Aelia Severa (reused), of Flavius Bellator the decurion (possibly reused), of Julia Fortunata, the wife of the sevir Diogenes (whose own sarcophagus is recorded but now lost), of Simplicia Florentina (reused), and of Theodorianus (ibid., Nos. 103, 105–6, 110, 108–9). The fragment of a sarcophagus with amorino supporters should be noted (ibid., No. 112); also, among the tomb-furnishings or decorations, the large sphinx, and the three fragments of a great funerary candelabrum, of a table-leg, of a capital with retiarius, of lions, and of a marine monster, not to mention numerous pine-cone finials, the symbols of life without end (ibid., Nos. 120, 137–8, 134, 123, 121, 133). The whole gives a picture of a well-to-do community, not sufficiently wealthy or tasteful to attract first-class monumental masons, yet insistent upon standards considerably above those of Chester or Caerleon.

The heterogeneity of this bustling and confident world is well attested by the sevir Diogenes, who himself came from Aquitaine, while his wife was a native of Sardinia, or by Theodorianus, whose home had been the Italian country-town of Nomentum. These were civilians, probably traders, who fared far and wide over the Empire. But Britons also could do this. Marcus Aurelius Lunaris, whose third name is British, (fn. 82) was a sevir Augustalis not only of York but of Lincoln, and his altar (fn. 83) at Bordeaux describes him as 'having been ship-borne from York (ab Eboraci avectus)'. His British nationality further emerges in his dedication to Tutela, the patron-goddess of Bordeaux, for he gave her the title Boudiga, in the British language equivalent to Victoria. (fn. 84) Lunaris must be one of the earliest attested exponents of British commercial enterprise. What he imported from Bordeaux can hardly be in doubt: olive oil or wine were its staple exports, and in a country so rich in animal fats as Britain the profit to be made upon imported wine was much more tempting than that upon oil. It may well have been this trade which attracted Diogenes to York from Aquitaine. But what went in exchange? Among raw products in bulk, wool or hides are a possibility, anticipating monastic trade a thousand years later. But the governor Claudius Paulinus was choosing as a handsome present (fn. 85) a British tossia, a word of uncertain meaning (fn. 86) judged to be some kind of mantle, as if manufactured clothing was of exportable quality; and there were other rarities like furs or sealskin, (fn. 87) and jewellery of jet or objects of walrus ivory. The jet (see pp. 141–4) was a special Yorkshire product, highly prized for its supposed magical properties as well as for its rarity, tractability and intrinsic beauty, while the walrus and the seals were to be found only in British waters. Livestock were also exported: hunting dogs, (fn. 88) for example, sold for a high price, and the official connection of Lunaris with Lincoln makes horses (fn. 89) another possibility. Whatever his cargoes Lunaris was a wealthy man, for only a man of substance, whether freedman or full citizen—a sevir Augustalis (fn. 90) might be either—could afford to take up this religious service, with concomitant public obligations, in two cities. His achievement was typical of the age and may serve as an epitome of the enterprise encouraged by the Roman peace.



In preparing the Inventory use has been made of the work of many scholars, living and dead. The following short biographical notes will indicate the status and the approach of the principal scholars of the past to Roman York. The concluding paragraphs give the briefest bibliography of some important published sources and direction to collections of York finds.

Martin Lister, M.D., F.R.S. (1638?–1712), the eminent zoologist, practised medicine from 1670 until 1683 in York with considerable repute. He devoted his leisure to the study of natural history and local antiquities, contributing papers on Roman York to the Philosophical Transactions of the Royal Society. He was the first to recognise that the Multangular Tower was Roman. Before receiving his M.D. at Oxford in 1684, he presented some Roman remains from York to that University.

Ralph Thoresby, F.R.S. (1658–1725), wool merchant of Leeds, antiquary and collector, paid several visits to York at the end of the 17th century when the Roman cemetery between Bootham and the river was being excavated for brick-earth and gathered a notable collection of antiquities. He published accounts of several of the discoveries in the Philosophical Transactions of the Royal Society and in Musaeum Thoresbyanum, the catalogue of his museum. After his son's death his collections were sold in London in 1764.

John Horsley, M.A., F.R.S. (1685–1732), the Northumberland antiquary, wrote a history of Roman Britain, Britannia Romana, published posthumously in 1732, which included an account of Roman York. This served as a basis for Francis Drake's account of Roman York in Eboracum.

Francis Drake, F.R.S., F.S.A. (1696–1771), a York surgeon, published his Eboracum in 1736. In this he included all the material published by Lister, Thoresby and Horsley and much new matter resulting from his own observations. He describes in detail the collection of Roman antiquities from York in the possession of Dr. Langwith, Beckwith and other York antiquaries. After 1736 Drake added notes of new discoveries to his own interleaved copy of Eboracum, now in the York Public Library, and published accounts of some in the Philosophical Transactions and local newspapers. He corresponded with Stukeley, and the correspondence was published in the Surtees Society publication LXXX (1885).

John Burton, M.D., F.S.A. (1710–1771), gynaecologist and author of Monasticon Eboracense, Drake's friend and fellow antiquary, also collected Roman antiquities. His collection is long since dispersed and lost. The only memorial of his interest in Roman York are two tracts in Archaeologia. He was the model for Dr. Slop in Tristram Shandy.

The Rev. Charles Wellbeloved (1769–1858), Unitarian divine and archaeologist, was one of the founders of the Yorkshire Philosophical Society, and until his death in 1858 honorary curator of antiquities in the Yorkshire Museum. In 1842 he published his Eburacum or York under the Romans, the first systematic account of Roman York; this includes first hand and accurate records of the fortress defences breached in 1835 during the construction of St. Leonard's Place and of the fragmentary remains of bath buildings found on the site of the old Railway Station. He was fortunate in his illustrator, John Browne (1793–1877), the historian of York Minster, who provided measured drawings of many of the principal discoveries. As well as writing several papers, Wellbeloved in 1852 produced the first edition of the book, A Descriptive Account of the Antiquities in the grounds and in the Museum of the Yorkshire Philosophical Society, that in subsequent editions became the Yorkshire Museum Handbook.

James Cook (1785–1872), in business as a brazier and tinsmith with his father, Robert Cook, until 1838 when he retired, amassed a large collection of Roman antiquities found at York. On his death the collection was bought and presented to the Yorkshire Museum. James Cook made a catalogue comprising annotated drawings of his collection, which is now in the Yorkshire Museum. His brother, Robert Cook, also collected but on behalf of Thomas Bateman, the Derbyshire antiquary, whose museum eventually passed into the Sheffield Public Museum. The objects are described in the catalogues of the Bateman Collection, with further details available in the Bateman correspondence preserved in the Sheffield Museum.

William Hargrove (1788–1862), the York newspaper proprietor and journalist, was a topographer and collector. In 1818 he published his History of York, in which he included the substance of Drake's account of Roman York with much new material. Full notes on many discoveries were published in his newspapers the Herald and Courant; these are useful in spite of their often highly coloured language. He also published a series of guides to York that contain references to casual finds. A collection of his miscellaneous MS. notes was presented to the Yorkshire Museum by his son.

The Rev. John Kenrick, M.A., F.S.A. (1788–1877), was tutor in classics and history under his father-in-law, C. Wellbeloved, at Manchester College, York (now Manchester College, Oxford). He succeeded Wellbeloved as honorary curator at the Yorkshire Museum, was responsible for revisions of the Museum Handbook and bought and presented James Cook's collection to the Museum. He published many papers on local finds but preferred to relate them to a wider background rather than record the details of their discovery.

The Rev. James Raine, D.C.L. (1830–1896), chancellor of York Minster, archaeologist and historian, was curator of antiquities in the Yorkshire Museum from 1873 until his death. He was responsible for observing excavation work done during the building of the present Railway Station and its ancillary buildings, for collecting the finds, and for their arrangement in the Museum. In addition to a general history of York, he published numerous papers relative to this subject, and left notes in diary form recording his observations of the Railway excavations. These notes, gathered together and copied in type by his son, the Rev. Angelo Raine, are now in the York Public Library. James Raine was responsible for the final edition of the Yorkshire Museum Handbook, 1891.

George Benson, A.R.I.B.A. (1856–1935), architect, was a prolific writer who published numerous papers on York antiquities. He was an early and constant observer on any building or drainage site. His approach was that of the geographer as well as the historian, an attitude reflected in the title of one of his papers The forming of the Landscape of York, being an introduction to the study of the archaeology of the district. His principal work was a history of York in three volumes; the first volume contains an account of Roman York, excellent for its date, which, though superseded by the results of archaeological activity over half a century, is still useful. Additional items appear in the later volumes.

Steuart Napier Miller, M.A. (1880–1952), lecturer in Roman History and Archaeology at Glasgow University, in 1925–8 directed the first large-scale archaeological excavations in the city for the York Excavation Committee. Though the excavations were limited to the problem of the fortress defences, they were extensive and their publication in the Journal of Roman Studies provided the fundamental basis for all subsequent work on Roman York. The accumulated archive, which includes unpublished detail, is preserved in the Yorkshire Museum.

Since the early 18th century the York newspapers, the Chronicle, the Courant, the Herald, the Yorkshireman, the Yorkshire Gazette and the Yorkshire Evening Press, are the main sources of information for casual finds. The complete runs are preserved in the York Public Library, where they are being indexed for material relating to the history of the city. The Gentleman's Magazine and the various Guides to York also contain useful facts, though very often at second hand from the newspapers.

The Yorkshire Philosophical Society, founded in 1823, in its Annual Reports, published from that date and continuing, and in its Communications or Proceedings, published from time to time separately from the Reports, includes notices of large and small finds. During the last decade the Yorkshire Archaeological Journal has contained several reports on York excavations, and 'Roman Yorkshire' published annually in the Journal from 1929 to 1943 included notes on York finds. The Yorkshire Architectural and York Archaeological Society published Proceedings spasmodically between 1933 and 1954, which also dealt with Roman York, and in 1956 published a Short Guide to Roman York. Each year since 1921 the Journal of Roman Studies has included short interim reports of Roman excavations and discoveries in Britain.

Apart from the Roman remains in situ, the basis for any study of Roman York must be the mass of York material housed in the Yorkshire Museum. These collections are of great variety and extent. A comprehensive catalogue of them has not been published since 1891 and the archives relating to them are far from complete; a modern publication is an early need. The Yorkshire Philosophical Society collections were moved to the Museum in 1829. In 1961 York City Council took over responsibility for the Museum. Other substantial collections of finds from Roman York are in the Bateman Collection at the Sheffield Public Museum, and in the History Department at St. John's College, York. The British Museum and the Ashmolean Museum at Oxford also contain finds from Roman York.


1 Tacitus, Historiae, III, cap. 45, regnum Venutio, bellum nobis relictum; Agricola, cap. 17, terrorem statim intulit Petilius Cerialis, Brigantum civitatem ... adgressus.
2 Brough (E. Riding): P. Corder and I. A. Richmond, BAAJ, ser. 3, VII (1942), 1–30. Malton (N. Riding): P. Corder, The Defences of the Roman fort at Malton (1930), 64.
3 For the moraine and its significance, G. Benson, York I, 1–2, figs. 7–8.
4 At Gloucester Legio XX (CIL, VII, 1339) probably succeeded by Legio II Augusta; and at Wroxeter Legio XIV Gemina, which left for the Rhineland and Mainz in A.D. 70 (Tacitus, Historiae, IV, cap. 68) and was probably succeeded by Legio XX.
5 Tacitus, Agricola, cap. 22, praising Agricola's eye for the opportunitates locorum, no doubt develops a common form of praise.
6 c. A.D. 48, Tacitus, Annales, XII, cap. 32; c. A.D. 57, op. cit., XII, cap 40; A.D. 69, Tacitus, Historiae, III, cap. 45.
7 Archaeologia, XCIII (1949), 33; CIL, VII, 236.
8 JRS, XI (1921), 107.
9 Sir Ifor Williams, Archaeologia, XCIII (1949), 33; E. Ekwall, The Oxford Dictionary of English place-names, s.v. York; M. Förster, Der Flussname Themse und seine Sippe (1941), 249.
10 Brit. School Rome Papers, XIII (1935), 18–22.
11 E. Ritterling, Pauly-Wissowa, Realencyclopädie, XII, 1664.
12 Tacitus, Agricola, caps. 7–8.
13 Arch. Journ., CIII (1946), 28–9.
14 C. H. V. Sutherland, 'Romano-British imitations of the Bronze coins of Claudius I', Numismatic Notes and Monographs, no. 65, 23.
15 Tacitus, Agricola, cap. 31, corpora ipsa ac manus silvis ac paludibus emuniendis inter verbera et contumelia conteruntur.
16 V. E. Nash-Williams, Excavations in the Prysg field, 1927–9 (National Museum of Wales), 8–10.
17 JRS, XLIV (1954), 84, fig. 8.
18 JRS, XVIII (1928), 211.
19 JRS, XXXIX (1949), 114.
20 As in Alcuin's day, when it was emporium terrae commune marisque.
21 For illustrations of carriages and waggons, see M. Rostovtzeff, Social and economic history of the Roman Empire (edn. 2) I, pl. lxxiv.
22 Lefebvre des Noëttes, L'Attelage, le cheval de selle à travers les âges (1931), 83–8: for clear examples of Gallic harness, see W. von Massow, Die Grabmäler von Neumagen, 141, Abb. 91, and H. Dragendorff and E. Krüger, Das Grabmal von Igel, 79, Abb. 49 and taf. 12.
23 J. G. D. Clark, Ant. Journ., XXIX (1949), 159–62.
24 For ironstone mining and smelting in N.W. Lincolnshire, see H. Dudley, Early Days in N.W. Lincolnshire (1949,) 142–3, 194–6.
25 For Derbyshire lead concentrated at Petuaria, see BAAJ, ser. 3, VII (1942), 28: eight inscribed pigs have so far been found.
26 Building-stone, M. Kitson-Clark, Gazetteer of Roman East Yorkshire, 25.
27 Roofing-slates, ibid.
28 For another inscription reflecting the legion's official script, see JRS XXXV (1945), 20, fig. 4.
29 The name Calcaria, applied to Tadcaster (Itinerarium Antonini, 468, 5), is probably connected with limestone quarries, as is the Gallic Calcaria in Narbonensis: but the word can also mean 'lime-kilns'.
30 Chester Arch. Soc. Journ., XXXVIII (1951), 3.
31 For the presence of Legio VI under Platorius Nepos soon after A.D. 122, see JRS, XXVII (1937), 247: the altars, EE, III, 99, and IX, 1162, explained as dedicated on the arrival of the legion in the Tyne, are equally explicable as foundation-dedications of the bridge.
32 CIL, XVI, 69 of 17 July, 122.
33 B.M. Catalogue, III (1936), p. clxiii.
34 Scriptores Historiae Augustae, Vit. Hadr., 5, 2, Britanni teneri sub Romana ditione non poterant.
35 For Hadrian as a disciplinarian, see the coin-type of Disciplina, B.M. Catalogue, III (1936), 466; also Scriptores Historiae Augustae, Vit. Hadr., 10, 2–8, 22, 1: and ILS, 2487, 9133–5.
36 E. Birley, Roman Britain and the Roman Army (1953), 26–7.
37 On the date of the visit see B.M. Catalogue, III (1936), pp. cx–cxii: it is not fixed by the diploma cited in n. 4 above.
38 Notitia Dignitatum, Occidentis, xl, 18, praefectus legionis sextae: but Sextae is also used as a place-name, ibid. 2–4.
39 Herodian, Hist. III, 7: for the forces involved, Cassius Dio, LXXV, 21, 6.
40 CIL, VII, 1044–5 (A.D. 219–20), 1046 (A.D. 225–35).
41 Cassius Dio, LXXV, 5, 4, for conditions on arrival; for the date, see next note.
42 Ilkley: CIL, VII, 210 (A.D. 197, Caracalla Imp. destinatus). Bowes: op. cit., 273 (legatus Augusti) with the commander of the ala at Binchester (EE, VII, 979, 980) in charge. Brough-under-Stainmore: JRS, XIV (1924), 248 (consuls of A.D. 197).
43 JRS, XLV (1955), 122, fig. 6.
44 Panegyrici Veteres, III (1828), v, 16, nec omnes copias quas trahebat instruxerit, describing the final resistance of Allectus.
45 Arch. Camb., XCV (1940), 121–2.
46 F. Stähelin, Die Schweiz in römischer Zeit, 278, Abb. 59.
47 Notitia Dignitatum, Occidentis, xl, 17: Ammianus Marcellinus, xxvii, cap. 8, 1.
48 This is implied by the size of its caldarium, at least 32 ft. wide (see Fig. 32).
49 Vindonissa: R. Laur-Belart, Römisch-Germanische Forschungen, X, Vindonissa, 46–56, Taf. 17. Lambaesis: R. Cagnat, L'armée romaine de l'Afrique, II, 457, 514–6. Also, in a secondary stage, Novaesium.
50 Ant. Journ., XXIV (1944), 1–5.
51 Panegyrici Veteres, III (1828), vii, 7, fruiturus exinde luce perpetua iam videret illic diem paene continuum; this implies the far north. Eutropius, Breviarium, X, 1, obiit in Britannia Eboraci.
52 Aurelius Victor, Epitome de Caesaribus, 41, 3: for the Alemannic wars of Constantius in 298–305, see Panegyrici Veteres, V, 2.
53 Fullofaudes is the form generally accepted, but the manuscripts also give Bulchobaudes, Ammianus Marcellinus, xxvii, cap. 8, 1.
54 See Thesaurus Linguae Latinae. s.v.
55 A. Mócsy, 'Das territorium legionis und die canabae in Pannonien', Acad. Scient. Hungaricae, Acta Archaeologica, III (1953), 179–99; also H. von Petrikovits, Das römische Rheinland, 63–7.
56 ILS, 9103, for a soldier leaseholder, and the primipilate lustrum: and for the latter, op. cit., 4222, 9104; also L. Barkóczi, Brigetio, no. 48 Dissertationes Pannonicae, Ser. II, No. 22 (1951)), and J. Szilágyi, Acad. Scient. Hungaricae, Acta Arch. I (1951), 196.
57 For boundary-stones 60 km. distant from the fortress, see ILS, 2454, 2455, cf. Petrikovits, op. cit.
58 Tacitus, Historiae, IV, cap. 22, longae pacis opera, haud procul castris in modum municipii extructa.
59 Cf. ILS, 2473, 2474.
60 For the significance of the ludus, see Brit. School Rome Papers, XIII (1935), 31.
61 For the funerary symbolism, see F. Cumont, Lux Perpetua, 290–2.
62 For the complete physical separation, topographical as well as juristic, see Mócsy, op. cit., 197, Abb. 1, 2, illustrating Aquincum and Brigetio, and E. Swoboda, Carnuntum (edn. 3, 1958), 189, Abb. 18, for Carnuntum.
63 JRS, XI (1921), 102: the text reads Deae Tutel(a)e Boudig(ae) M(arcus) Aur(elius) Lunaris IIIIIIvir Aug(ustalis) col(oniarum) Ebor(aci) et Lind(i) prov(inciae) Brit(anniae) Inf(erioris) aram quam vover(at) ab Eboraci avect(us) v(otum) s(olvit) l(ibens) m(erito), Perpetuo et Corne(liano consulibus).
64 Cf. Carnuntum and Aquincum, Cambridge Anc. Hist., XII, 19. If Aurelius Victor, Epitome de Caesaribus, cap. 20, 27, in Britanniae municipio, cui Eburaci nomen, ... exstinctus est were to be taken strictly the point would be proved.
65 CIL, XIII, 3162, leg(ato) Aug(usti) p(ro) p(raetore) in Brit(annia) ad legionem sext[am] refers to the post of Paulinus as governor: cf. CIL, VII, 1044–5 (219–20). The capital goes with the combination of governor and legionary legate.
66 Archaeologia Aeliana, ser. 4, XI (1934), 136–7.
67 The choice of York follows from the palace (see note 12) and the death of Severus there; Eutropius, Breviarum, VIII, 19, decessit Eboraci.
68 Cod. lust., iii, 32, 1, dated III Non. Mai. Eboraci, Fausto et Rufino Coss.
69 Her presence is attested by Cassius Dio, LXXVI, 16, 5, with the wife of the Caledonian chieftain Argentocoxus.
70 Herodian, Hist., III, 14, 9, referring to the whole Empire not to a province.
71 Eutropius, Breviarum, VIII, 13, 1, apud Carnuntum iugi triennio perseverasset, bellum Marcomannicum confecit.
72 E. Swoboda, op. cit., 131 ff., Abb. 5.
73 Scriptores Historiae Augustae, Severus, 22, 7.
74 Eutropius, op. cit., X, 1.
75 The connection of Constans with the North follows from the crossreference to the frontier-scouts in Ammianus Marcellinus, xx, cap. 1, 1: for the campaign, Firmicus, de error, profan. relig., 28, 6.
76 For praetorium in this sense, ILS, 2298.
77 O. Doppelfeld, 'Römische Grossbauten unter dem Kölner Rathaus,' Germania, XXXIV (1956), 83 ff.
78 For Arimanius, see F. Cumont, Les religions orientales dans le paganisme romain, 141–8.
79 Cf. the stone from Machen; JRS, XXXVI (1946), 133.
80 R. P. Wright, Catalogue of the Roman inscribed and sculptured stones in the Grosvenor Museum, Chester (1955), 53, nos. 154–5, pl. xxxviii, also fig 1b.
81 Der Römische Limes in Osterreich, XVIII (1937), 25–31.
82 For Lunaris as a British name, see EE, III, 87, and Cumberland and Westmorland Ant. and Arch. Soc. Trans., N. S., XXXII (1932), 133.
83 See note 2, p. xxxvi, above.
84 It is the same name as Boudicca: for the derivation see A. Holder, Alt-celtischer Sprachschatz, s.v.
85 The presents (CIL, XIII, 3162) comprise a Canusian cloak, a Laodicean dalmatic, a gold jewelled brooch, two coats, a British cloak (tossia) and a sealskin.
86 O. Hirschfeld (CIL, XIII, 3162) equates the word with the mediaeval toxa meaning 'cloak'.
87 CIL, XIII, 3162, pellem vituli marini: for the equation of this with the seal (phoca), see J. Cotte, Poissons et animaux aquatiques au temps de Pline, 32–6.
88 For a wolf-hound, see R. E. M. and T. V. Wheeler, 'Report on the excavation of the Prehistoric Roman and Post-Roman site in Lydney Park, Gloucestershire', 88–9, Soc. Ant., Research Report 9; cf. Symmachus, Epistolae, II, 77, on seven Irish wolf-hounds in iron cages; pointers, Oppian, Cynegetica, I, 470–80; hounds, Strabo, Geographica, IV, 5, 2; greyhounds, Nemesianus, Cynegetica, 124–5; bull-dogs, Claudian, Stilicho, III, 301.
89 Arch. Journ., CIII (1946), 68, for horses at Horkstow and Lincoln.
90 G. Chilver, Cisalpine Gaul, 204–7: they might be incolae as well as liberti, von Premerstein, in Ruggiero, Dizionario epigrafico, s.v. Augustales, 831.