During the course of the excavation it became apparent that the furrows of the mediaeval field system were particularly shallow in this north-eastern part of the field. The resulting slight buildup on the ridges meant that almost all traces of earlier occupation, except of features dug into the subsoil, had been destroyed by the action of the plough. The disappointingly barren, slightly stony, disturbed deposits which survived on the ridges were cleared, but most furrows were not excavated (Fig. 1a). However, all was not lost! At the SE edge of the excavation the earliest feature, Ditch 64, continued NE, but unexpectedly turned N before reaching the boundary of the site (Fig. 1b). A junction with either Ditch 49/62 or 38/39, a major E-W ditch known from the geophysical survey to lie beyond this year's trench, is almost certainly in the strip of woodland alongside the field. Two paving stones, set against and carefully laid level with the top of Zz30 F3, formed an extension to the SE. The first, of sandstone, measured 1.1 m x 0.78m, was carefully bedded on a skim of lime mortar, and may have been part ofthe structure. The other, a less regular 1.02m x 0.47m slab of limestone, 60mm thick, had been sunk into the natural gravel. It may originally have served as an anvil on which to crush clay roof tile for making the opus signinum used in the replacement kiln Zz29 F4-175. A large piece of tile, stamped with ARVERI (Fig. 2), the name of its maker, was found underneath the stone, and many tiny fragments in the infill around its edges. More came from an irregular hollow dug into the fill of Ditch 64, as did another ARVERI stamp and a nearly complete, but broken TPLF tegula. Part of a second example of the latter was found in the stony backfill of hollow XxYy 28 F1,

The rubble seen in the 1993 exploratory trench proved to be mostly demolition infill. This included fragments of thin flat sandstone slabs, which presumably once overlay the main flues of an underlying double T-shaped malting or drying kiln Zz29 F4-F5. The surviving lower courses of its walls consisted of up to two courses of limestone blocks, laid on and in part consolidated with opus signinum mortar; their upper surfaces were scored in places by the plough. The much-damaged cross-flues of both survived at the bottom of a mediaeval furrow. That of F5 was cut through by a mid 19th century field drain. Ditch 65, dug at the rear of the kilns to prevent them being flooded, ran NW towards Ditch 49/62. Both F4 and F5 were fired from the south-west from a deeply eroded, interconnected double stoke-hole. These heavily burnt hollows extended into both main flues, where they were infilled with ashy rubble. This had been levelled and covered with two C 0.7m square sandstone slabs, one set between thejaws of each furnace. Such large nearly complete paying stones are a noteworthy exception to the scattered fingments recorded elsewhere on the site. The close proximity and construction of Zz29 F4-F5 suggests that it was built as a single unit. This would probably have included an overlying rectangular working floor covering c.5.2m x 4m. The paving stones laid to the SE of Zz30 F3 may have been its point of access. It can take up to three weeks to process grain into malt, but no post-hole evidence of a roofed building covering the kilns was found. One can only postulate on the existence of sill beams set at or above present ground level. The purpose of the two isolated post-holes found close to Xx30 F6 and Yy30 F5 is also not known.

Five unstratified 4th century coins came from the disturbed layers over this part of the site but with the exception of some pottery neither the truncated spread of dark occupation soil between Xx/Yy28 F I and Yy30 F5 nor the kilns produced any firm dating evidence. Despite this, it is worth noting that in the villa, Building A, an opus signinum mortar was more commonly used for building purposes in the latter part of the century. In Room 5 the conversion of the T-shaped kiln A2 Fl, into a trench furnace (Vol. 1, 167), may also have been related to the construction of kiln Zz29 F4-F5.

The short length of ditch which ran diagonally across the N part of the excavation was an extension of a major landscape feature incorporating both late IA & RB alignments and labelled Ditch 48/49. This was last sectioned about 50m further west, and it is not known how far the earlier phases of it continue beyond that point. Its junction with Ditch 62, just beyond this year's trench, has yet to be examined so as to establish the relationship between the two. The present length is now tentatively listed as Ditch 49/62. It was sectioned twice and near the bottom produced clear evidence of at least two distinct, but undated alignments. The dating evidence provided by the twenty-two coins from its overall 0.5m thick abandonment upper backfill of black soil and domestic rubbish, suggests that it was probably deposited between about AD 350-375. It produced exceptionally large quantities of pottery, notably Oxford colour­ coated bowls, Dorset black-bumished vessels, and local micaceous greywares, but only three sherds of late 4th century Midland shelly ware. The forty artefacts from it included a black glass gaming counter with one red and three white spots, a square piece of bone inlay, several bronze bangles, a small graduated balance arm, and a pair of tweezers. The ironwork was mostly fragmentary off-cuts, the best piece probably the shank and greater part of the blade of a pruning hook. Evidence of metalworking consisted of a scatter of corroded wire and strip off-cuts, a fragment of a slag encrusted crucible, and a few small basal slag masses.

Structure 27, part of which included a stony rubble surface uncovered in an extension at the east corner of the excavation trench, lay almost directly under the modern ploughsoil. It was associated with a shallow post hole, BB30 Fl, and bounded on its N side by what are interpreted as two U-section sill beam slots, BB30 F2-F3 (Fig. 3). it post-dated the mid 19th century field drain which ran W across the site, but was almost certainly contemporary with Structure 28. This consisted of an irregularly spread alignment of shallow stone holes which emerged from the NE baulk and continued N for about 13 m. These are interpreted as the slight remains of a shallow dry stone wall footing. The soil associated with both structures produced late 19th-early 20th century china and coarse glazed earthenware, clay pipe fragments and thin brown bottle glass. Neither were completely uncovered so their full extent is not known, but their proximity close to the field's boundary combined with my own farming experience suggests that both were once part of some semi-permanent livestock housing, probably intended to shelter ewes and lambs in the early spring.

My thanks to those who took home the large bagfuls of pottery and duly returned them with the contents carefully washed and dried. 1 am particularly grateful in this respect to Peggy Fowler and Angela Newcombe, who as in previous years, shouldered the task of dealing with most of this material.

Next year we propose to examine the area immediately to the NW of this year's trench. It should cover both the E & W extensions of Ditch 49/62, with the chance to check the junctions with Ditches 62 and 65. The geophysics suggest that the E part of the prehistoric trackway lies close alongside, and the furthest extent of the excavations may just reach Ditch 38/39.

If all goes well it is hoped to reopen excavation on Saturday 22nd April 2006 and continue every weekend until the autumn.