Day 10 of the dig saw us soldier on in the extensions to Trench 2. The circular brick void at the south end of the trench is continuing to cause much head-scratching, not least when it became apparent that there is in fact no properly constructed lid to it - it is covered only by an extremely haphazard collection of mortar and brick and tile rubble.

How on Earth this has survived intact without collapsing for the better part of four centuries is anybody's guess, and how it got there as it is now, is also more than a little puzzling.

Our only plausible interpretation at the moment is that the feature maybe once had a wooden lid which was buried in rubble, and has since decayed away.

We spent much of the morning and afternoon carefully clearing away rubble from the top of the feature, hoping to find a properly constructed roof, but once holes began to open up in the rubble we stopped and recorded the surface of the mortar rich deposit, about the thickness of one brick, which defines the top of the void. (Click here for an annotated version of the above image).

We then dug down on the outside of the feature, on its east side, hoping to be able to expose its external face but the rubble here was so loose that another gap opened up, and so we again paused work to reconsider our strategy.

It's pretty clear now that the covering of the void is not an in-situ structural feature, and so tomorrow we will begin to carefully remove it and finally get a good look at the circle of brickwork inside, and hopefully get a better idea of what it is - it could be a well, or a garderobe chute, a cistern or something else again.

Its function is likely closely related to the adjacent brick 'floor' surface - which today we dug an ever-growing sondage down one face.

To our surprise it has now revealed at least seven courses of brickwork, which means that this is something more than just a floor. Together with the appearance of the well-like void we're now suspecting that the brick surface is perhaps a culverted drain - maybe taking water from the south end of the trench and channeling it north, maybe into a moat.

If this is the case, we can expect the brickwork down the side of the 'floor' to continue down to a considerable depth.

It is also possible, if this is a culvert, that the nearby circular 'well' feature is connected to it, though there's no evidence for that yet.

Meanwhile in Trench 4, the wall that we've been working hard to reveal all week finally appeared. This trench was placed on the location of a wall junction we only saw part of late on in last year's dig, and was intended to confirm whether the wall was a corner or a junction.

We can now confirm it is a junction, and a spur of the wall continues south-west, which means our gatehouse is larger than we previously thought. It is a very substantial wall, as you might expect for the foundations of a four-storey building.

We also continued to excavate Trench 4, in which we have abandoned hope of finding signs of the 1960s trenches, but in recompense it is producing one or two nice finds including another large chunk of the brilliant blue Westerwald jug we found part of on day 8.