Several publications give the impression that Roman coarse pottery in the north is closely datable. But Hadrian's Wall and northern forts and the Antonine Wall are dated by historical records, inscriptions and structural periods, and not by any potsherds found in their strata: it is the periods which give their context to the potsherds. The history of this old error is reviewed in two parts: (i) the misuse, since 1930, of Gerald Simpson's work at Haltwhistle Burn fortlet; (ii) the misuse, since 1934, of the evidence for dating the great Corbridge Storehouse and the consequent misdating of Antonine pottery at Corstopitum and Mumrills.

A unique linchpin depicting a human thumb, discovered in North Lincolnshire in the 1980s, is here reconsidered and particular emphasis is placed on its potential as an apotropaic symbol. This function is explored in terms of the wider material and literary basis for magical gestures in Roman Britain and the Roman world. The relationship between this image and the space in which it was used is also explored.

In addressing a CBA conference in 1965 Mr. C. E. Stevens performed his now famous ‘strip-tease act’, whereby he arrived at the conclusion that, save in a few exceptional areas, south-eastern England from Hengistbury Head to the Wash rarely produces Romano-British villas with patterned pavements. He related this phenomenon to the system of land-tenure which he believes he can recognize operating in south-eastern England during the Roman period—tenure in tir gwelyawg. Hence, the occupiers of the villas did not lay mosaic pavements because they ‘do not seem to have had the power of capital accumulation sufficient to place an order for them’. South-eastern England thus becomes a zone of ‘poor men's villas’ (to use Mr. Stevens' term), and sitting in the very centre of this zone are the villas of the Chilterns. Mr. Stevens generously recognized that High Wycombe was an exception to his rule, and rather more grudgingly admitted that Totternhoe might be another, but he was confident that ‘critics might find one or two more, but not many’. I have previously challenged, in passing, this harsh assessment of the Chiltern villas, but in the last two years additional information has come to light and the importance of well-dated mosaic pavements from villas has been emphasized by Dr. D. J. Smith's brilliant chapter on villa-mosaics in The Roman Villa in Britain.

Cunobelinus was the most significant figure in Britain during the decades leading up to the Roman invasion, though his reign has received relatively little attention. Cunobelinus' coinage is of great importance to understanding the socio-political structure of South-East Britain prior to the Roman invasion and whilst studies of his gold and silver have been published in previous editions of Britannia (Allen 1975; de Jersey 2001), his bronzes have been subject to surprisingly little work, particularly considering that they are by far the most common struck bronze issues known from Iron Age Britain, with a total of 2,608 examples currently recorded in the Celtic Coin Index and on the PAS database combined. This study proposes a broad typological scheme with which Cunobelinus' bronzes can be ordered and demonstrates that, like Cunobelinus' silver, but unlike his gold, they can be divided into three regional groupings, which it can be argued correspond to three different political sub-groupings under Cunobelinus' control. In addition, the bronze's metallurgy and metrology and the mints at which they were struck are investigated. This article examines the contribution of coinage to understanding Cunobelinus' political history, and how he used imagery to reinforce and legitimate his power in the different regions under his control at different times during his reign. The types of sites at which Cunobelinus' bronzes have been found are also outlined and the likely function of the coins discussed.

The large hillfort at Hembury, near Honiton (Devon) is one of the most impressive late prehistoric sites in South-West England. Occupied in the Neolithic and Iron Age, it was taken over by a Roman force about or shortly before A.D. 50. Substantial timber buildings were constructed, including a probable fabrica, in which iron from the adjacent Blackdown hills was worked. The Roman site was abandoned by the early Flavian period and not reoccupied. Though not evidently a conventional fort, Hembury joins a list of hillforts in South-West England which were used by the Roman army in the early decades of conquest. These include Hod Hill and possibly Maiden Castle (Dorset), Ham Hill and South Cadbury (Somerset).

Modern Usk is a small market town situated within the county of Gwent on the east bank of the River Usk at its confluence with the tributary River Olway, and on the flood plain of both rivers. It lies at a height of c. 17 m O D, on a major bridging-point of the River Usk, and commands two natural routeways. The first, formed by the valley of the Usk itself, runs from the central highlands of southern Wales around Brecon, south-eastwards to Abergavenny and Usk, and from there southwards to reach the sea at Newport. The second, minor, route follows the valley of the Olway north-east to Raglan.

Fragments of two robust wool textiles with an unusual knotted blue pile were recovered from a Period I (late Flavian) fort ditch at Vindolanda. Their knotted structure — unknown hitherto in the western Roman provinces and only partially paralleled in the eastern — is discussed, together with questions about their possible production centre and actual function. The Supplementary Material available online (https://doi.org/10.1017/S0068113X18000259) contains technical details of the textiles, an investigation of the raw materials and a comparison of the wools used.