The 400plus pages of the 2001-2002 Research Reports were completed by the end of March (available only on CD-ROM - with nearly 300 illustrations), and excavation resumed early April. As before work took place in two areas.

Justin Blake's team continues to investigate the far western boundaries of the site, ultimately hoping to locate the western defences of the large pre-Hadrianic forts. The enlargement of the 2002 trench has revealed more of the multiple stone buildings, ovens, roadways and drains of what appears to have been orderly second century occupation, with later stone and timber structures. But the buildings do not appear to be normal civilian housing, and the presence of much brick and tile suggests that there may have been yet another baths building nearby. Finds have been fewer than usual (so far), but they have included a fine (and complete) enamelled dragonesque brooch, a large millstone, several gemstones, much pottery and more coins than usual.

Volunteers working on the remains of second century buildings near the western limits of the Vindolanda site

Andrew Birley's team have been working immediately to the south of the Severan Commander's residence, dealing with the plough-damaged remains of third century civilian buildings, and the underlying spread of rampart clay from the Severan defences, before dropping into the anaerobic remains of the superimposed timber structures associated with the early timber forts. The major timber structure may prove to have been the fort's hospital, although it will be next year before we are certain of the identification. So far the team have located a major north to south roadway, with substantial side drain, and on the west side, disappearing under the trench edge, what may prove to be a workshop. The potential hospital lies on the eastern side. Running below the remains, right across the site, is a large wooden water pipe, consisting of 1.66metre lengths of alder wood, with a 5cms bore through the centre, and jointed together by oak cross-pieces. Water still runs in the pipes. The underlying timber buildings have yet to be examined.

There have been few writing tablets so far - and none from the potential hospital - but there has been some fine leather, including a lady's stamped shoe, a few probable surgical instruments, and much high quality decorated Samian ware from South Gaul.

Excavation will continue until the end of August 2003, and will resume in April 2004, in the same areas.

A familiar sight at Vindolanda - lines of wooden posts, marked with white tags, and plenty of water, some two metres below the walls of Severan Commander's residence

The timber water pipe (below the ranging pole) heading towards the eastern side of the possible hospital

We still haven't found those western defences of the early wooden forts, and that must mean that they lie beyond the Trust's land, to the west, making the forts over 7 acres in size. It has been a frustrating experience for all concerned, but it has had its compensations - and it has thrown up a major puzzle. At least two periods of early wooden military buildings are certain in this area, to be succeeded by two successive stone structures during the second century, and then both a timber and a stone period in the third century (definitely civilian). The earliest of the second century stone periods may well have been Hadrianic, judging by the pottery and the coins, and the huge commercial ovens associated with them may have prepared the food for the legionary building gangs
on the Wall.

But underneath all these structures we found four parallel fort ditches, complete with the remnants of defensive stakes and pits - but running west to east, and thus not associated with any previously known Vindolanda forts! These ditches can only have been the southern defences of an early fort which stretched northwards, across the line of the presumably later Stanegate road. So far we have little positive dating evidence, and we cannot say that it must have been an earlier fort than our existing 'period I', which dates to the late AD 80s. As is customary, the excavations of 2003 in this area have solved a few problems and created more.

One of six major ovens associated with the early stone buildings

Excavators stand in the four newly discovered fort ditches - the deepest is on the left (north)

The second excavation area was devoted to the examination of the remains to the south of the Severan commander's residence, roughly half way between the other excavation site and the later stone fort. Post-Roman ploughing had all but obliterated the third century civilian buildings, which had probably been two domestic houses, and they had been constructed over the spread remains of the massive southern clay rampart of the Severan establishment. The creation of that rampart had, in turn, damaged the underlying mid-second century remains, which appear to have been connected with workshops. Below that, some two metres from the present ground surface, there were considerable remains of the periods II/III and IV pre-Hadrianic fort buildings, although Roman demolition had removed all door thresholds. The finest of the surviving timber was the wooden water-main, which ran across the site, being associated with the earliest occupation there. Large alder trunks, with bark remaining, had been bored through with an augur, to create a 5cms pipe for the water, and the individual lengths were connected with rectangular oak slabs - without any use of iron or lead fittings. The source of the water was probably the major spring at the western edge of the Vindolanda site.

A solid 3.20m wide roadway ran north to south through the site, and on either side we were able to examine something like a third of two significant buildings. That on the western side may prove to be a military workshop, judging by the quantities of iron slag in one part, and by the piles of leather off-cuts and old boots and shoes in another. On the eastern side lay the remains of buildings much more difficult to interpret. Their laminated or plank floors were unusually clean, and the plan so far recovered looks more like a military hospital block than anything else. But as usual, it will take another year of excavation before we have sufficient of the plans to be more positive about identifications.

Lengths of the wooden water-pipe, awaiting transport to the Museum laboratory

Surviving planked flooring, on wooden joists. Elsewhere, Roman demolition teams had done a thorough job

Overall, finds were not as voluminous as usual, although the boot and shoe count exceeded 80, but writing tablets were very few. One small uninscribed altar, a few bronze medical instruments, an enamelled seal-box (with contents), some remarkable Republican denarii, and an inscribed silver goblet were the outstanding pieces - but the Museum laboratory has plenty of work ahead during the winter months.

Excavation should resume in adjacent areas to those of 2003 in early April 2004.

We are most grateful to the efforts of a fine bunch of volunteers, drawn from several nations, who worked exceptionally hard in most unusual Vindolanda weather, where the normal dangers of constant water and mud were often replaced by sunburn - and clay baked like brick.