Britannia 5

NX 903884. Complete excavation in 1971 confirmed the Roman date and provided a complete plan of the fortlet within its single ditch and turf rampart. It had a six-timber gate with ?tower, two timber buildings, a latrine and two ?hearths. The single structural phase is dated Antonine I, c 142-58. On abandonment the buildings were probably burnt, the rampart slighted, and pottery dumped in the ditches. Au(abr)

[NT 1977]. The fort is of milliary cohort size at 4.8 acres (1.92ha) within the ramparts. It originated c 140 and had a second Antonine occupation followed by extensive Severan reconstruction. Limited exploration revealed a vicus, and a fourth, post-Severan occupation of the fort may be civilian. Unusual features included a very narrow praetentura. Appended reports treat coins, samian and other pottery, small finds, and (summary) bone report.

SS 958699. A small re-excavation of part of this villa has removed some apparent discrepancies in the evidence for the structural development of the buildings, and taken in conjunction with the published material has made a new and more detailed interpretation possible. There are also reports on a Fortuna statuette and the mosaic. Au(adp)

Brings together new information on pre-Flavian kilns, whose products rarely travelled over 25 miles and therefore show considerable variation, as do the kilns themselves (eight types figured and described). Kiln furniture is seldom integral to the structure; fabrics are almost invariably oxidised until well into the Neronian period.

Questions the refusal of the appellation villa to a town house with associated farm buildings and land. In reply, Rivet maintains the need to restrict the term to a country estate. See also abstract 75/3051.

Recent attempts to redate the Carpow fragmentary inscription are rejected on both archaeological and epigraphic grounds; the inscription itself and the coarse pottery would allow occupation to continue until 215 or even 216.

TQ 161052. A cremation at the Marquis of Granby inn was accompanied by several vessels including plain samian and a glass flask. It had probably been in a nailed wooden container. Light emission spectroscopy of the samian agreed with macroscopic attributions to Lezoux and Rheinzabern potteries. The glass, analysed by X-ray fluorescence, was of Group II (antimony- rich).

Using collections of distinctive pottery from S England, two possible marketing models were set up. In Model I, a kiln is sited near a town and its products lie within that town's sphere of influence but may also travel along main roads to more distant locations (e.g. Savernake ware, Rowlands Castle ware); this model also applies to other goods such as roller-stamped flue-tiles. In Model II, large-scale industries producing coarse pottery show distributions largely independent of main towns and roads. Two further models, one for small-scale rural production (e.g. Congresbury kilns) and one biased by army requirements (e.g. Severn Valley wares) are also described.

[NY 7766]. Interim report on the tablets excavated in 1973 from levels dated c 90-100. The thin leaves bear ink writing and were probably originally fastened together concertina-fashion (intermediate form between roll and codex?). The writing concerns deposits of money and supplies of food including meat and wine apparently disbursed to military personnel; many points of interest arise.

New additions to the 1961 catalogue are mapped. Over 40 Type I buckles from various kinds of site began manufacture in Britain c 370 and some may have been made as late as early 5th century. The mainly southerly distribution has a small, isolated northern group. Their frequent association with female burial calls for further investigation, particularly in well-dated contexts. Some 30 Type II buckles, also British-made, are more diffusely distributed than Type I; some could have been worn by civilian officials but many have clear military associations starting under ?Constans (342/3).