ON EXCAVATIONS ON ROMAN SITES IN BRITAIN. 205 



again very near to the present surface; they were for the most part 

 covered by a foot of humus only. 



The main rampart at the S.W. side of the hill-fort was then investi- 

 gated, together with some 55 square yards of the interior area adjacent 

 — the whole of the latter with the riddle. Here, what had come to be 

 known as our ' relic bed ' was again encountered, between 1 and 1^ feet 

 below the present surface. The bed extended from the level interior 

 area almost up to the crest of the rampart, although objects upon the 

 flat were far more abundant than upon the slope of the rampart. The 

 relics found were practically a repetition of those already recorded 

 from a similar level elsewhere in the hill-fort, with certain additions. 

 These additions included part of a bronze penannular brooch, an 

 interesting bronze gilt ornament (possibly a plate brooch), several 

 earrings and other articles of coiled bronze wire, a bone toggle, a very 

 small iron sickle, part of a horseshoe, a good flint scraper, and, among 

 pottery, fragments of a red incense bowl with fringed rim, fragments 

 of painted white mortaria with hammer-head rim, fragments and a 

 rim of ' calcite ' or ' vesicular ' ware and several portions of black bowls 

 mended with iron rivets ; the latter would seem to show that even the 

 common pottery used here was imported from a distance, or it would 

 hardly have been worth the trouble of repair by riveting. 



The rampart here was then cut through from back to front; there 

 was no facing- wall, only a few stones in line which may have been the 

 remnant of one. Some 30 feet behind these stones a low wall appa- 

 rently marked the back of the rampart. The core was again found to 

 be of mixed material, mostly dry rubble-stone with thick layers of clay 

 in places ; its construction 'seemed to show a later and a larger rampart 

 covering an earlier and a smaller one. At a point 5 feet in front of the 

 present crest and 4 feet below it, the top of a parallel dry stone wall 

 was encountered ; this was excavated and found to stand 5 feet high ; it 

 was probably the back wall of the earlier rampart. Further, at 11 feet 

 behind this wall and 10 feet below the present crest, the top course of 

 yet another parallel wall was discovered on the closing day of our work ; 

 this would seem to belong to a still earlier structure. We regretted 

 the necessity of postponing its exploration, which will require the 

 removal of a great weight of material in order to reach it. 



Many fresh cuttings were also made at various points in front of the 

 rampart, in further investigation of visible ditches and ramparts and in 

 search for other ditches suspected but hidden from view by dibris. Two 

 long cuttings approximately W.S.W. (on which side of the hill-fort 

 it will be remembered that the main rampart is entirely thrown down) 

 revealed three parallel ditches cut in the rock ; two of these had been 

 hidden from sight by fallen debris. These three ditches, like others 

 previously excavated, were filled, the first completely and the second 

 and third less so, according to the slope of the hill-side, with clean 

 dry rubble mixed with some larger wall-facing stones. These stones lay 

 upon the rocky bottoms of the ditches with hardly any silting below 

 them, again showing that the main rampart must have been thrown 

 down soon after these ditches were cut. 



Part of the rocky berm excavated between the site of the main 



