lay below the level of the road constructed for the second occupation, where it passed between the later ditches on the south. We cannot, of course, rely altogether upon such a find as evidence of the prolongation of the first occupation. All that we can say with certainty is that the latest coins found in the ditch of the early fort date from A.D. 86, and that it is well nigh certain that a garrison was still there at that date. But indications already alluded to suggest that the earlier of our two main periods did not end with the evacuation of the Agricolan earthwork, but included also the earliest occupation of the enlarged fort.

Here again the strongest of these indications is obtained from the pottery, particularly that taken from the overlapping ditch in front of the West Gate, a ditch which, as we have seen, must have been filled up at a comparatively early period. It is a mere handful of fragments—some twelve pieces of Terra Sigillata. Four belong to decorated bowls. One is the rim of Dragendorff, Type 29. Two show the well-known arrow points. A fourth has a border of long, pointed leaves, resembling those illustrated on page 211, Fig. 4. The undecorated vessels comprise pieces of Dragendorff, Types 18, 27 and 35, as well as of the bowl-like vessel with flat rim, ornamented with lotus buds, figured in Plate XXXIX., Type 11. Of the coarse dishes there is a fragment of a dark-grey colour, exhibiting the technique of 'rustic ware,' and also a portion of the neck of a large urn-like vessel of a close-textured reddish-brown ware, such as was noted in Pit LIV. There is not among these pieces a single one which is characteristic of the late period. Unfortunately the corresponding ditches in front of the North and South Gates yielded little, though 'rustic ware' came from the former. It must further be noted that most of the types of Terra Sigillata mentioned above were present in that portion of the inner ditch of the enlarged fort cleared out on the west front. As a rule, the evidence of ditches is less to be relied upon than that of pits. The latter do not seem to have been open for any length of time. After they had once been filled up with rubbish, they were probably closed for sanitary reasons. The clay stopping near the' surface was no infrequent feature. Ditches, on the other hand, must have lain open for a long period. Yet in this case the pottery of the ditches entirely supports the theory put forward. The ditch of the enlarged fort, open no doubt for a comparatively long time, produced not only the early fragments, but also many of later date, including one or more which seem to belong to the export from Rheinzabern, while the branch of the same ditch in front of the gate, filled up early, showed only early pottery.