Perhaps fifty years after the last use of the Causewayed enclosures, human settlement resumed on the hill, at about 3,550BC. This time there is no doubt that it was about defence. The bank surrounding the site followed the line of the inner causewayed ditches on the East side of the site, but extended about 40 metres further out towards the hill point to take in the flat area there, where the Sacred Circle would later be constructed. Outside the bank, which was about a metre high and carefully constructed, was a ditch, this time continuous except for three entrances. These were carefully defended with gates, one of which is shown, right.
There was a small settlement of houses on the top of the hill again, and once again a large scatter of finds. In cutting O1 there was a large flint working site, yielding thousands of flakes. Unfortunately the shallow depth of soil right on the top of the hill makes certainty about exact house locations difficult.
One difference to the earlier Neolithic period was that this time a track extended into the small valley at the far side of the hill where the Long Mound would eventually be constructed. This appears to be the first time this part of the hill was used, and it apparently contained a shrine with square building and cobbled area, surrounded by fences (see photo below). This part of the hill is well hidden from elsewhere, which may have been why it was selected.
Whatever was going on during this period at the hill, it seems pretty certain that it ended with a battle. Careful measuring in of the many hundreds of flint arrowheads found on the site appear to show evidence of an attack on the site (see distribution diagram, right). This appears to correspond to the end of the settlement in the period, presumably rather violently, at about 3,450BC.
