Our knowledge of ancient artillery derives primarily from several ancient technical texts. The most important among them are the works of Philon, Heron and Vitruvius. Taken in combination, these texts are sufficiently detailed to allow reconstruction of functioning artillerypieces, and the first reconstructions of catapults were made as early as the middle of the nineteenth century, long before any actual find was known. The second source for our knowledge of ancient artillery consists of representations of catapults found on Greek and Roman reliefs. In spite of their small number these reliefs are a significant complement to the written texts, not only because they show technical detail which might otherwise be controversial, but also because most of them can be dated within narrow bounds and in this way give some hints at the direction of development of ancient artillery. But the chances of finding new texts or new reliefs are very small. So actual finds of ancient catapults in archaeological excavations constitute a third source, which will gain increasing importance, because there are fair chances of finding remains of catapults at a number of sites.

The project was a rescue excavation before housing development, mainly between Hadrian's Wall and the Vallum, close to the fort of Petriana at Stanwix. Traces were found of a pre Wall field-system. A minor Roman road was found running parallel to and 80 m south of the Vallum. A substantial Roman boundary of postsettings for posts over 2 m high was found running between, and at right angles to, the Wall and Vallum. In one area south of the Vallum a series of ditches of the second century A.D. suggested meadowland associated with a vicus of the nearby fort. To assist in its conservation the fragmentary remains of the Wall were traced and their position recorded. No trace was found of a Military Way; it may have been obliterated by an historically-recorded medieval road found parallel to and immediately south of the Wall. The position of Milecastle 65 was located by geophysical survey and found by test excavation to survive at foundation-level.

The fort of Hayton, south-east of Pocklington, North Humberside (Yorkshire), was first located from the air by Professor J. K. St. Joseph in 1974. His photographs showed a small fort surrounded by double ditches: the crop-mark of these was distinct on the SE, NE and NW sides (PL. IX) and at the rounded northern and eastern angles. Further reconnaissance in 1975 produced more photographs which introduced the possibility that the fort also had an annexe to its north-east. A single ditch appears to continue the line of the inner fort ditch on the south-eastern side and form an enclosure with rounded angles to the north-east of the fort. Examination of Ordnance Survey photographs taken in 1968 showed that the platform of the fort had been just visible then. In view of the sudden clarity of the fort cropmarks, it was decided to undertake a small-scale excavation to discover whether the remains of the fort were being eroded by the constant ploughing to which thearea is subject. The first campaign, in February 1975) took place in the field south of the minor road which runs through the fort site (FIG. I, Areas A–F) and located the eastern angle, parts of the angle-tower and the south-east gateway. A second season, in October and November of the same year, examined the state of preservation of the remains in the northern field after removal of the crop (FIG. I, Areas G–S) locating the line of the ditches on the three remaining sides as well as the north-east gateway, the north angle-tower and remains of two of the interior buildings.

Excavations in 1972–75 on behalf of the Department of the Environment revealed an extensive Iron Age settlement and traces of widespread Roman agricultural and industrial activity at Wakerley, Northamptonshire (FIG. 2). The settlement was situated in Wakerley parish, immediately to the south of the road running between the villages of Wakerley and Harringworth and nine miles north-northeast of Corby (FIG. 3). It was sited on sloping ground, overlooking the valley of the River Welland, and some ½ mile from the river itself. From the site there are extensive views of the river valley to the north and west and of the hills and dales of Rutland that lie beyond. A deep natural gully occurs in the hillside, just to the west of the settlement and, as a result, the site is in an open position and fully exposed to the westerly winds. The settlement was located between the 250 and 300 ft. contours on a wide expanse of Lower Lincolnshire Limestone. Clays of the LowerEstuarine Series and outcrops of Northampton Sand and Ironstone occur on the lower slopes of the valley below the site and in the adjacent gully to the west. It is likely that a convenient supply of water would have been available in this gully in earlier periods, but this has been piped away in modern times.

Vexillations from all three British legions – II Augusta, IX Hispana and XX Valeria Victrix – took part on the Vitellian side in the second battle of Bedriacum in A.D. 69 (Tacitus, Histories 3, 22), but their main bodies had remained behind in the province under the cautious governorship of Vettius Bolanus (Histories 2, 97). After the capture of Fabius Valens, however, when the cause of Vitellius was plainly becoming hopeless, the army of Britain declared for Vespasian (Histories 3, 44): et Britanniam inditus erga Vespasianum favor, quod illic secundae legioni a Claudio praepositus et hello clarus egerat, non sine motu adiunxit ceterarum, in quibus plerique centuriones ac milites a Vitellio provecti expertum iam principem anxii mutabant. The meaning of this passage is surely clear enough: II Augusta, which Vespasian had commanded with distinction during the Claudian invasion and the following years, took the lead in swinging the whole army of Britain on to the Flavian side, despite the resistance of the other legions. On the attitude of XIV Gemina there is room for uncertainty: it had been sent back to Britain by Vitellius (Histories 2, 66), but it can hardly have been very enthusiastic for him, nor is it easy to suppose that many of its centurions and men can have owed their advancement to him. However that may be, II Augusta was manifestly the prime mover in securing the adherence of the army of Britain to Vespasian.

The Roman auxiliary fort at Doncaster, the Danum of the Antonine Itinery, lies roughly half way between Lincoln and York (FIG. I). Over the past ten years, the centre of the modern town has been largely redeveloped and the Doncaster Museum, in conjunction with the Department of the Environment, has carried out extensive rescue excavations. The rectangular fort, which could be denned from modern property boundaries, was defended by a stone wall, 2 m thick, which enclosed an area of c. 237 ha (5.85 acres); it was found to be of Antonine and later date. Beneath this and extending further to south and east, was a Flavian to Hadrianic fort, covering an area in excess of 2-6 ha (642 acres). Two phases were recognised in the earlier fort, separated by widespread evidence of burning. The charcoal and burnt daub of the fires, however, were concentrated into clearly defined bonfires, implying an orderly demolition. The constructiontrenches of the buildings of the second phase were dug through the remains of the bonfires in such a way as to show that reoccupation was immediate. From these construction-trenches three asses of Vespasian were recovered and an as of Domitian, minted in 86-87, was found in an associated clay floor; asses of 86-87 are the commonest first- or second-century coins from the site. The available evidence would seem to suggest that the fort was remodelled after a period of neglect perhaps while the main body of troops was fighting in Scotland under Agricola and the fort itself may have had a caretaker garrison. Historical and archaeological evidence combine to suggest a partial withdrawal from Scotland in or shortly after 87 (Hartley, 1972) and it is at this time that some lowland forts would again be required to house full garrisons. The fort at Doncaster, in common with others on the edge of the Highland Zone, was abandoned under Hadrian; the smaller stone-walled fort,situated in the north-west corner of its predecessor, is one of a Pennine and circum-Pennine group initially constructed in the late 150s or early 160s.

The discovery of a large Roman camp at Durno, six miles NW of Inverurie, in Aberdeenshire, was made on 26 July 1975, in the course of a reconnaissance flight from Scone airfield to the Moray Firth. Only about half the perimeter, including two gates (each with a traverse) was visible at the time of discovery, but this was enough to show that the area exceeded 50 acres. Further observation from the air in 1976-77, and field-work, including the digging of a number of ditch-sections, have established the line of the ditch round the greater part of the perimeter, proving the area to be about 140 acres (57 ha.).

Professor C. Northcote Parkinson has demonstrated that government officials increase by a process of quasi-organic growth, unrelated to the burden of work for which they are responsible. The phenomenon is as apparent in the Roman Empire as in modern Britain; but modern British scholarship working on the evidence from the Roman Empire has carried the process even further by producing an official who never existed - though in fairness it must be admitted that no new post was created. It would be interesting and instructive to follow the process by which this apocryphal administrator became accepted into the canon of governors of Roman Britain, but problems of space mean that we can do no more than show why he must be declared redundant.

The Roman army has long been regarded as the epitome of organised self-sufficiency in both ancient and more recent literature. That the army acted as engineers as well as soldiers is so amply attested that it needs no comment here. The manufacture of tiles by the military is well known and the quarrying of stone both by the soldiers themselves and under military supervision is attested; but there is little evidence for the supply of building timber to the army. Since throughout the first and for a large part of the second century the majority of forts in Britain were built wholly or partly of timber, this represents a considerable gap in our knowledge of military building. This article is an attempt to consider the problems of timber-supplies for military purposes in Britain during that period.

The written sources for Roman Britain, especially after the first century A.D., are so scant that our information must come very largely from archaeological discoveries. While these have thrown light on many aspects of Britain during the Roman period, we are still very much in the dark about one important feature of Romano-British life, the intellectual and cultural activities of the well-educated Briton. It is in the fourth century that the problem becomes particularly acute; Britain escaped much of the unrest that plagued the continent and thus enjoyed an era of peace and prosperity, a prosperity reflected in the stimulus given to construction of villas on a grand scale and often with fairly lavish interior decoration. It is a period when we should naturally expect a parallel flowering in the level of cultural life. The sources, however, are virtually silent on this topic and we must base our knowledge on the deductions that can be made from the evidence of physical prosperity; a useful criterion of cultural and intellectual activity is interest in the literary classics, but while it may be reasonable to assume that in his schooling the Romanized Briton would receive some grounding in the major works of Latin literature, archaeology has provided relatively little evidence that this was in fact the case. The known examples have been so frequently cited and illustrated that they have produced a general impression that they are copious. In fact, there is prima facie evidence of only one writer, the poet Vergil; the allusions are limited to one of his works, the Aeneid, and are far from numerous. My present purpose is to provide a catalogue of the allusions in Roman Britain to Vergil's poem, to assess the evidence that they provide for a general knowledge of the Aeneid, and to show what I consider to be the special significance of one example, the mosaic from Lullingstone Villa in Kent.

The date of 441/2 for the decisive point in the series of events which constituted the adventus Saxonum has frequently been quoted as deriving from a contemporary source. As a matter of fact, however, it is a ghostdate, born in 1892. It originated in a marginale by Mommsen to a chronicle-fragment which is one of the continuations used to extend the Carolingian edition of Jerome. This fragment was edited and annotated by Mommsen before the modern editions of Jerome were available. The information required to correct Mommsen has, however, been readily accessible since 1923, and the manuscript basis of his text has been open to consultation in London for much longer. It is not my purpose here to wage any wars, even upon a long neglect of the primary source, but merely to set out some considerations of Form-kritik which are the historian's equivalent to the archaeologist's typological criteria.

The enamelled skillet-handle illustrated (FIG. I) was found in 1972 at Brough-on-Fosse, Nottinghamshire, by Mr John Hart. The handle adds a further example to the rare group of vessels with cham plevé enamelling which include the Rudge cup and Amiens skillet. The Brough skillet-handle measures 3 in. (7·8 cm) in length and is enamelled on the upper side; the reverse is plain. The enamelled decoration consists of a panel of vine-leaf ornament repeated five times. The enamel has flaked, but it appears that red was used in the centre of the vine leaf surrounded by a deep blue and then a green. These vessels are conveniently referred to as being enamelled, but recent analyses have shown that, in common with late Iron Age enamelwork, these are not true enamels but only coloured glass, which has not fused with the bronzework.

The Roman relief sculpture at Chedworth Manor, Gloucestershire (PL. XX A), discussed in this paper, was first brought to my attention by Dr Martin Henig of the Institute of Archaeology, Oxford, and it is published here for the first time by kind permission of its owner, Mr J. D. F. Green. Traditionally brought to the Manor by the Knights of St John during their occupation of it in the Middle Ages, the relief was in position on an outer wall of the house some fifty years ago. But after World War II Mr Green moved it inside the house, where it now adorns one wall of the principal reception-room. At Mr Green's invitation I saw and studied it there.

The fragment of skillet-handle illustrated (FIG. I) was found during the 1973 excavations in Goss Street, and is the first example of a Roman bronze vessel with a manufacturer's stamp to be found in Chester. Unfortunately the handle occurred in a medieval robber-trench, but is likely to be derived from Roman military levels. Bronze skillets would appear to have been used by legionaries and the civilian population for mulling wine, for cooking and as capacity-measures. It has been suggested by Wright that a group of legionaries would each have carried one of slightly different size which belonged to a graduated set, and often there are internal markings within the bowl to measure a known amount. Skillets of this type were brought into Britain during the first and second centuries. Most of the earlier examples were probably made at the Campanian centres of bronze-working in Southern Italy, at Capua and Nola, but after about 60 it would appear that Gaulish manufacturers progressively captured the market. Campanian metal-workers produced a wide range of vessels, but it is mainly skillets that are stamped with the maker's name. To date, forty different vessels are known with stamps from the British Isles (see Appendix).

Recent excavations in Romano-British towns have produced evidence that property-boundaries, once fixed, often persisted over long periods. St. Albans, Cirencester, Leicester, Winchester, and Gloucester have furnished examples, while the results of earlier excavations at Caerwent, Wroxeter and Caistor-by-Norwich can probably be interpreted along similar lines. European parallels may be cited from Augst, Kempten, Xanten and, possibly, Cologne. On those sites where modern excavation has taken place the property-boundaries appear to have been fixed at an early date in the history of their respective towns. It would be of great interest to establish on what basis the original distribution of building-plots was made, especially those sited on street-frontages close to town-centres, where building-land was at a premium and where a systematic allocation of property might be expected.

Professor Branigan, in his recent article ‘Villa Settlement in the West Country’, interprets the plans of six British villas as a series of rooms ranged around a small courtyard which he calls an intra-mural yard. They will be discussed later; for the moment it is enough to point out that no evidence is adduced in support of the interpretation. Branigan, however, not content to leave the matter there, goes on to say that the intra-mural yard is ‘a design which is common among the smaller villas in northern Gaul’ and points to six plans in support of this notion. Again, no evidence is offered. Now what is so astonishing is that three of the villas cited as evidence for this bland statement were interpreted as hall-type villas in 1921 by Franz Oelmann in the classic article which established the existence of the type, and have been so regarded ever since. The reversion to the orthodoxy of nearly sixty years ago is not defended, Oelmann's arguments are not refuted, nor does his article even appear in the bibliography; we are expected to take Branigan's word that it is so. For all that there are no arguments to attack, the point at issue is of such fundamental importance for villa studies in all the northern provinces that I propose to pass in review the essential facts on which judgement has to be based.

Since the publication of Professor Grimes's account of the excavations at Holt (Grimes, 1930), in which he includes a general survey of tile-production and a list of tile and pottery kilns, no detailed research on the tile industry has been carried out in Britain apart from work on patterned flue-tiles (e.g. Lowther, 1948). In Italy much more has been learnt about the organisation of the tile and brick industry from an examination of the many stamps from Rome, Ostia and elsewhere (Bloch, 1967; Helen, 1975; and various volumes of CIL) and further research is being undertaken by Dr Steinby of the Finnish Institute in Rome. In this country, however, no overall survey of the production and distribution of tiles has been undertaken and one is not aware that such research is currently in progress. For this reason it has been necessary to examine the evidence from the whole of Britain when trying to understand problems concerned with the tile industry in the Cirencester region.

This analysis of the stamps of Legio IX (or VIIII) Hispana found at Lincoln, Templeborough, Old Winteringham (Lines), Malton, Aldborough and mainly York, or at Carlisle and its legionary tilery at Scalesceugh forms a sequel to the present writer's article in Britannia vii (1976), 224-35 on ‘Tile-Stamps of the Sixth Legion found in Britain'. It has been based on rubbings and squeezes made from the originals. Much of the earlier bibliography has been omitted because stamps when published merely in printed capitals cannot be identified with specific dies. The die used for the one example known from Lincoln (Lindum) seems to have been taken north for use among others at one of the tileries, on sites not yet located, which supplied York (Eboracum). The rare examples from Aldborough (Isurium Brigantum), the capital of the canton fifteen miles north-west of York, match two of the dies from York. But in contrast one example, Type I, assigned with probability to Aldborough, has no parallel. The outlying post at Malton (Derventio) has produced two dies unmatched elsewhere, even at York.