Dr. J. C. Mann has given us (Britannia ii (1971), 218–24) a rich listing of variants attested in the Latin inscriptions of Britain. His list is very complete and based on up-to-date sources, keyed where appropriate to RIB. As such it is a very valuable collection.

The variants are classed by Latin letters which are found in the spellings in relation to what is expected in classical Latin. In the commentary which is offered below variation is classed not so much by the letter(s) involved as by the mechanism or phenomenon which is seen as accounting for the observed variation. Such a classification, if correct and fully explicated, may hope to lead ultimately not only to a recognition of the truly popular and spoken characteristics of the Latin of Britain but also to a grasp of the basis of this differentiation and of the social forces at work.

Brockworth is situated in the Severn Valley between Gloucester and the Cotswold escarpment. The parish is divided by the line of Ermin Street, with the church and Court lying to the north. The Horsebere Brook winds its way past the site from the slopes of the Cotswolds to reach the Severn about 8 km away near Gloucester. Three other westward-flowing streams cross the low-lying undulating country between Churchdown Hill, 4-5 km due east of Gloucester, and Robinswood Hill to the south, both outliers of the Jurassic escarpment. Over the Lower Lias clay are spreads of Superficial Deposit, here composed of gravel and sand intermingled with a yellow clay.

As a schoolboy in the early 1960s the writer discovered a series of Iron Age and Roman salt-production sites along the shoreline of Chichester and Portsmouth Harbours (FIG. 1). These areas were visited intermittently over the next ten years and many of the observations made at that time were published in 1975. That study described the topographical setting of these sites and the character and chronology of the surface finds. It also included some discussion of the economic context of salt-making and its place in the settlement pattern.

The report is focused on excavations in a suite of four rooms in the villa dating to the late third or early fourth century, associated with a plunge-pool excavated in the nineteenth century; the excavations were undertaken in advance of consolidation work. The rooms overlay a second- century bath-house also known from earlier work. The most important find was a hoard of fourth-century counterfeit coins which had been deposited in one of the rooms after it had gone out of use.

The mosaics of fourth-century Romano-British villas constitute a rich but relatively neglected resource. In this paper, I propose to demonstrate how figured pavements can provide indications of the function of the rooms they decorated.

Hitherto, an approach to the identification of different types of room using mosaics as a prime indicator is one that has not been fully utilised. Major studies of the Roman villa are valuable in their own right but barely touch upon room function. Some important and still-cited articles have concentrated on form rather than function. Other studies have focused upon social aspects of the villa. In recent years comments have been made about how decoration can be used to differentiate space but with little reference to mosaics. As against those studies which hint at the use of mosaics but do not make much of them, Thébert has produced a seminal work addressing function, architecture and decoration, especially mosaics, showing that many illuminating deductions can be made.

The theology of the Imperial Numen continues to stir vigorous debate. In an article recently published in this journal J.C. Mann has proposed a radically different interpretation of the concept as seen through Celtic eyes in the Roman provinces of the North-West. He lays down in the first place that it is epigraphically impossible for numinibus Aug. to refer to multiple living emperors, who would rather have been indicated by the reduplicated ‘g’. Nor can the reference be to the living emperor in combination with the deceased: deified dead emperors would certainly have had the divus title and could not in any case be credited with numen since they possessed no ‘divine power’. As a result, the formula n[u]minibus Aug. on an altar from Benwell (RIB 1330), for example, must be taken to refer to the numina of the reigning emperor, who in popular thought was considered to have multiple numina, powers ‘in every human field’ (p. 177). Long before its publication, an oral version of this thesis had convinced E. Birley, who on the strength of it completed both RIB 707 and 1330 in such a way as to refer likewise to the numina of the living emperor. Since this expansion affects the interpretation of epigraphical texts that are of interest in ways other than just recording the cult of the numen, the whole question merits re-examination.

A half-life-sized copper-alloy head of a bearded man was published in the Portable Antiquities Scheme's report of finds from Roman Britain in 2009.1 The head was purchased by the Ashmolean Museum in 2011. In this paper evidence for the identification of the subject as a portrait of the emperor Marcus Aurelius is reviewed by comparison with metropolitan and other certainly identified heads of deities and portraits of the emperor. The technique and likely function of the head are compared with those of similarly worked Roman copper-alloy heads of emperors and deities found in South-East Britain. Finally, a brief account is given of geophysical survey and trial excavation conducted in 2012–13 in the field where the head was found. This offers a unique opportunity to explore the head's archaeological context.

The Romano-British theatre at Gosbecks is situated in the parish of Stanway, three miles west of the colonia at Colchester on a small promontory between the Roman river and a lesser tributary. It lies on the eastern edge of a thirty-acre area which was covered in Roman times by numerous buildings including a possible bath-building, the large Gosbecks temple standing within its own temenos, and a massive walled ‘fairground’. The theatre lies 125 yards due south of the temple and may well have stood within the ‘fairground’. The entire area (FIG. I) falls into a class of rural site widespread in Roman Gaul where large temples are associated with theatre, mansio and bath-buildings, but where private houses and town walls are absent.

This paper assesses the alternative models currently proposed for the original design of the Antonine Wall. It argues that there is sufficient evidence of a pattern in the structural relationships between the Wall and its various garrison posts to confirm that there was a major change of plan during the construction process. This resulted in the addition of a series of secondary forts, several of which can now be shown to have replaced primary fortlets, and a concomitant delay in the construction programme. An explanation is offered for this rapid change of design, which resulted in one of the most intensively garrisoned linear frontiers in the Roman Empire.

This paper assesses the effects which the building of Hadrian's Wall had on the patterns of supply and communication from the continent. Existing systems were strengthened rather than altered, and Hadrian's reign saw the full development of ports and military installations on the North Sea and Channel coasts. Navigation to Britain and sailing conditions on various routes are discussed, comparing their importance in the transport of wine, oil, exotic plants and samian ware and the movement of military personnel. Use of the Rhône–Rhine axis is emphasised for the movement of goods from Central Gaul and the Mediterranean, but other rivers in western and north-western Gaul were of some importance, as the details of samian distribution demonstrate. Finally, non-state organisation of the acquisition and distribution of commodities supplied to the army on Hadrian's Wall is strongly favoured.

In his discussion of the nomenclature of auxiliary regiments, G. L. Cheesman stated his belief that the adjectival form of a provincial name such as Cyrenaica, Moesica, or Syriaca implied not original recruitment in those provinces, but recent, perhaps lengthy, service in them. The existence of such units as cohortes II Hispanorum scutata Cyrenaica, I Lusitanorum Cyrenaica, and I–IV Thracum Syriaca lend solid support to his argument. It would have been entirely reasonable for a unit serving on an expedition to be known by the province with which it was usually associated, e.g. ‘the Syrian regiment of Thracians’. This would be especially true where there were a number of other regiments with the same ethnic title and otherwise indistinguishable, e.g. the numerous regiments of Thraci and Hispani. Here, the sense is rather, ‘the Syrian regiment of Thracians’. Subsequently, Cheesman qualified his generalisation when he observed that there must have been occasions when a form such as Cyrenaica would equal Cyrenaeorum? His modification was prompted by difficulties associated with regiments lacking an ethnic title and bearing only an adjectival form of a provincial name. He cites as examples the units simply designated as I–III Augusta Cyrenaica. It is to be assumed that he saw the same argument as applicable when he appended the ala I and the cohors I Britannica to the list of units whose origin he gives as Britain.

The site, which was in medieval times the precinct of St Oswald's Priory, is close to a now-vanished channel of the River Severn, on which, a little to the south, was the Roman quay (FIG. I). The site is outside the NW corner of the Roman colonia of Glevum. It was excavated in 1975-6 in advance of re-development, with the aid of a grant from the Department of the Environment. Building plans were cancelled in 1976 and the site ceased to qualify for ‘rescue’ grants. Only limited archaeological work was therefore done on the Roman levels, which are still available for future research.

Tacitus's account of the Claudian invasion of Britain and the campaigns that followed it down to the year 47 is unfortunately lost, and historians have to be content with Dio, who gives the only consecutive account, supplemented by scattered references in other sources. This short note is an attempt to extract a few more fragments of information from these literary sources and in particular it concerns the part played by Rome's Batavian auxiliaries.

Roman Scotland provides a unique opportunity to examine archaeological sites which represent uncomplicated, often single, phases of activity. Rather than the usual palimpsests seen in the sites in the rest of the island, resulting from centuries of occupation, Scotland has been the archaeological equivalent of a highwater mark, only briefly inundated by the Spring tides of the Roman Empire during the Flavian, Antonine, and Severan periods.