Week 12 (30 August- 5 September)

Update: A decision whether to run a dig in 2005 will be made in January and will be posted on this website. Also, check back here shortly for the final updates to the 2004 web diary.

The final public week of the project! Over 1,000 people visited the site on the last day, helping to boost the weekly total to almost 2,500. The season total was just under 19,500, almost the same as last year despite being the generally poor weather and being closed for the better part of a week. Many former trainees and placements returned this week, hoping to witness the major discoveries that always turn up at the end of a dig. The weather was good again, allowing us to make us much progress as possible.

Finds work under the public gaze	Roman glass gaming counter (diameter 13mm)

In Trench 3, it now seems that a low bank had been laid down at the beginning of the rampart construction to form the rear of the rampart base. This may have had the dual role of marking the limit of the rampart for the builders, and retaining the body of the rampart. The concave surfaces of the rampart deposits being excavated are partly due to their lying against this bank, although it still seems they are also subsiding into a cut beneath the rampart.

The concave rampart deposits in Trench 3. The water-filled hole is the base of the medieval column foundation pit

In Trench 4 the mortar-filled robber trench was emptied, revealing a cobble foundation. This foundation resembles the cobble foundation found in Trench 1 in 2001, and apparently formed the south-east end of the medieval hospital infirmary from about 1100 until the building was extended to the south-east in around 1250. As this foundation was at the maximum safe depth limit, it was decided to concentrate our efforts on investigating the layers lying between this structure and the stone-lined drain to the south-east, which presumably pre-date the infirmary.

Excavating possible pre-infirmary deposits, cut by the stone-lined drain (to rear of trench)

In Trench 5 the bottom of the wall plaster layers 'inside' the Multangular Tower was reached. Beneath were various thin layers including gravel and clay, indicating these were floor deposits within the tower. Hopefully, study of the finds from these layers will tell us when these floors were laid down, and what kind of activities took place in this building.

Excavating floor layers inside the Multangular Tower. The concrete foundations forming the south-east corner of the tower are clearly visible.

The mortarium found in Trench 5 last week, after cleaning (about 200mm across)

A narrow cut was found, evidently the north-west side of the 'L-shaped' feature found in Trench 3. This is thought to be the base of a flight of steps that allowed the defenders of the fortress to reach the battlements quickly. However the corner of this feature overlay the south-east corner of the Multangular Tower foundation, which means the rear of the tower must have been dismantled before the timber staircase was constructed. The dating of this staircase feature is now crucial; if it is Roman in date, it means the rear of the Multangular Tower was demolished during the Roman period, while the front of the tower remained and is still standing!

The trainee on the left is standing in the 'L-shaped' feature, and the trainee on the right is stood in the medieval column foundation cut

At the west end of Trench 6 our attention turned to the deposits through which Miller had dug his trench to expose the timber piles. These deposits, well over 1m deep, are clearly not natural; indeed they are dumps of Roman waste material. It seems the original Roman ground surface, which consisted of sand over clay, sloped down steadily westwards to the River Ouse from this point. It therefore proved necessary to form an artificial terrace at this point before the Multangular Tower could be constructed. Hopefully the finds recovered from these deposits, including leather and wood, will help to date the construction of the Tower more closely - almost compensating for the failure to obtain tree-ring dates from the piles!

Roman dump deposits, cut through by the foundations of the Multangular Tower (left), visible in the side of Miller's trench. The scale rests on probable natural sand

At the east end of the trench, a large mass of mortared limestone rubble was found. Much of this had apparently been removed during Miller's excavations, but it seems likely that this material originally filled the entire compartment. The best explanation we can offer at present is that the Roman deposits into which the foundations of the Multangular Tower had been dug into at this point proved unstable, and had to be replaced with something more solid. The construction and maintenance of this tower is proving to be a much more complex operation than we had previously thought.

The mortared rubble layer is exposed