More information : [Area: SU 6385 6230] Belgic Defences of Calleva?
Air photographs taken in 1949 revealed a previously unsuspected and wholly buried defensive system. The line is known precisely for about two-thirds of its circuit. The area enclosed must be approximately 90 acres. The Roman Wall [SU 66 SW 46] in its northwest sector shows a pronounced subsidence caused by the passage of the ditch below. From here it runs to an almost right-angled corner just inside the North Gate, and thence straight to the southeast corner of Insula 27. In Insula 22 (1899) it was found, but described as a large rubbish pit of uncertain limits, and it also passed under a building of Insula 27 (1901) with buttressed walls and a relieving arch. The outer edge of the ditch at the extreme corner of Insula 27, where a change of direction takes place, is marked by a buttressed boundary-wall. On the south side of the modern track, the ditch was found by electrical earth-resistivity survey in 1955, but its course thence towards the Wall is obscure. It must have influenced the siting and alignment of the Public Baths [4.9] and it is suggested that the defence passed hard by their eastern side and (following a very indeterminate crop-mark on some aerial photographs) then crossed the Wall near its south-south-east angle. In the meadow immediately south of this point a strong cropmark appears on the photographs, and slight traces of the ditch and bank are to be seen near the field-gate into the lane. Faint indications of the course take us as far as the north edge of Rampier Copse, where another strong cropmark, substantiated by slight surface relief, carries the line round again to the subsidence in the wall, by way of another angle just to the south of the modern track. There are four entrances at the north, south, east and west which appear on the aerial photographs as slight inturnings from the line of the barrier.
A trench across the ditch on the west of the Wall showed it to be 45ft. wide and 11ft. deep, cut with sloping sides and a 12ft. flat bottom at the very base of the gravel. In 1955 the Silchester Excavation Committee opened a section within the walled area. Although the bank had been almost entirely obliterated by the early 2nd century, enough remained to show that it had been constructed on a base of about 45ft. There were also traces of a counterscarp bank, so that in all the defence was over 100ft. across. Two dozen sherds found in and beneath the bank showed no Roman features. In the first layer of infilling of the ditch pre-Roman pottery outnumbered Roman by more than two to one. There was no reason to suppose that the filling of the ditch contained material later than 50A.D. if so, a pre-Roman date for the defence is assured. However, the number of entrances and the almost exact correspondence with the main east-west and north-south lines of the Roman street plan is a serious obstacle to the pre-Roman dating.
Coins of Epaticcus and Cunobeline have been found at Silchester and the Belgic earthwork may have been built by the Atrebates or the Catuvellauni. Evidence of pre-Roman occupation is slight and consists of about 30 coins (some of which may have circulated in the early Roman period); sherds of 40 vessels of Arretine ware, 17 of which are stamped, some by potters known to have been active circa 11BC - AD 9; terra nigra and rubra occur as sherds with pre-Conquest potters' stamps; butt-beakers of the earliest Augustan pattern can be included. Fragments of distinctive flan moulds together with the abbreviation for Calleva on pre-Roman coins indicate a mint here, although no such coins have been found at Silchester. (1-3)
Eight fragments of flan moulds were found at Silchester during the Society of Ants. excavations. They are of a type identified at Camulodunum as moulds for casting flans for the coinage of Cunobelinus. The Silchester specimens contained traces of gold, silver, copper and lead. Only two have an exact provenance - one from a pit in Insula II and the other from a pit outside the West Gate. Both pits contained sherds of the earliest type but also contained Claudian to 2nd century pottery and even some 4th century sherds. The only certain coinage attributable to the mint - of Eppillus - is too small to have been made in these moulds. (4)
The line of the Inner Earthwork in the southeast sector is based on very little evidence. The line must be substantially correct but at its east end, near the supposed entrance, may not be a megger sounding made by Case in 1955 suggests that the ditch may be slightly to the west of the line shown on the 1956 plan [A.O. 25"] and, in so doing, would make the supposed entrance rather less of an inturn. (5)
'The O.S. line of 1956, marked on the plan over which we collaborated, derives from a trace on A.P No.276, but I am by no means certain that the eye of faith has not been deceived over this view. My idea is that there may be a re-entrant caused by the adverse ground hereabouts...There may not be an entrance on the east'. (6)
'Much more definite evidence is required before we can accept as Belgic an enclosure with four vertically opposed entrances and with cardo and decumanus at right angles to one another. The entrances themselves are not, as drawn, "inturned" in the meaningful Iron Age sense of the term, and since the ditch was filled from the rampart, which in turn contained whatever was in the gravel from which it was constructed, the arguments based on the proportion of Iron Age and Roman sherds recovered from the ditch filling is unconvincing'. (7)
Earthwork clearly visible as crop and soil marks. (8-13)
As a result of combined fieldwork with Mr.G.C.Boon, the circuit of the Belgic earthwork has been established. The primary authority for this circuit is the air photographic evidence with certain confirmatory evidence [see Authty.1]. Little physical evidence survives, other than an extremely spread, unsurveyable bank on either side of the south entrance and south of the west entrance. The plan of the earthwork is polygonal and it was formed by a single bank and ditch. It occupies the high ground at the end of a spur at a height of about.300 ft.OD. Although no direct contradiction of a Belgic date for this earthwork is possible, on the evidence available, there are certain features which are not easily explained:
i The relationship of the later Roman street plan to the entrances.
ii The alignments of the Roman roads to the entrances.
Photographic reduction of MS Section of 1955 Excavation. (Supplied by Mr. Boon). (14)
The Inner Earthwork cannot be dated precisely, but a date of circa AD25- AD43-44 seems likely. Boon briefly discusses the alignment of the roads and street plan (15-17)
Additional bibliography. (18)
The Inner earthwork probably dates to the mid first century AD and encloses an area of 32 hectares. It is situated within the circuit of the Outer Earthwork and partly underlies the later town. Boon argues construction took place in the later 40s AD, but it is also feasible to have been pre-Conquest. Excavations have found that the Roman Road to Winchester partly overlies the ditch and that the main infilling of the ditch had taken place by 60 AD and was completed by the mid-late second century AD. It has been also argued by Boon that the earthwork was constructed by Cogidubnus after the establishment of a client state shortly after 43 AD. However, evidence from recent excavations has not clarified to which side of AD43 the earthwork was constructed. It is possible it was constructed pre-conquest in a response to either the threat of a Roman invasion or to the turmoil following the death of Cunobelin circa 40 AD. Excavations have also shown the Inner Earthwork marked the extent of the oppidum of Calleva prior to the Roman invasion. (19-20)
Sections of the Inner Earthwork of Calleva were recorded during the Silchester Iron Age Environs Project from aerial photographs. A broad bank with an external ditch were visible as cropmarks within the town walls extending from Insula 27 to Insula 24. Another section was identifed on the western side of the town outside the walls, also as a bank with an external ditch. It was previously recorded as part of the Aggregate Landscape of Hampshire NMP project. (21-23) |