We worked hard in the lime tree avenue of Forty Hall today tying up loose ends and backfilling all but one trench, on what turned out to be the last day of excavation for this year's summer dig on the site of Elsyng Palace.

Although we still haven't found evidence for the elusive inner gatehouse, we have instead built a picture of what may prove to be quite a substantial moat separating the inner and outer courts of the palace.

We pressed on with excavation in trench 4, which we opened a few days ago on the alignment of the western end of the range of buildings which define the northern side of the palace's outer courtyard, in order to find out how far west the range extends, and whether or not the new 'moat' feature we identified in trenches 1-3 extends out this way to meet it.

By the end of the day yesterday this trench had revealed an alignment of rubble which closely matched where we would have expected the building range to cross the trench, and early this morning we began excavating this rubble deposit to find out whether there was evidence of any structure or not.

By mid-afternoon however, the rubble had been lifted to reveal what at first appeared to be a sloping brickearth surface - precisely like the sloping brickearth cuts in our other trenches which define the edge of the 'moat' - but further work in the afternoon showed that the slope was illusory, and the brickearth contained a substantial amount of brick and roof tile rubble, meaning that it was itself the fill of a larger cut.

We pressed on excavating this context throughout the mid to late afternoon and eventually came to the conclusion that it probably is the fill of the 'moat' feature, which does then indeed seem to be extending out towards the western end of the outer courtyard's north range.

Unfortunately, having been forced off site early in the week due to the severe weather has left us short of precious time that could have been spent fully excavating this trench and exploring the theory further, but it does give us a good starting point for next year's work which will probably include an effort to locate the actual end of the north range and to examine how it meets this year's 'moat'.

Despite the absence of palace structures in trench 4 we did have the consolation of plenty of small finds, including a large sherd of another Bartmann jug, (see day 6) this time bearing a smaller armorial fragment of what may have been the arms of Amsterdam.

Trench 4 also produced a complete clay tobacco pipe bowl of a form we've seen before, and is closely dateable to circa 1640-60, meaning it could well have been deposited by workmen during the final demolition of the palace.

All that remains to do on site tomorrow is backfilling trench 4 and clearing the site, which we expect will only take at most half the day.