Calleva Atrebatum
The cantonal capital of the civitas Atrebatum 13 km SW of Reading (Berkshire) where the collection of antiquities is housed in the borough museum, loan of the Duke of Wellington; there is a small museum near the site.
Calleva (town in the wood) was an oppidum of the Belgic Atrebates, probably as early as Commius (ca. 50 B.C.). Under the king Eppillus (ca. A.D. 5-10) it had a mint, and a high standard of living among the richer inhabitants is denoted by imported pottery. It was overwhelmed by the Catuvellauni under Epaticcus ca. A.D. 25, but after the Roman invasion of 43 became for a few years a bastion of the client kingdom of Cogidubnus (cf. Tac. Agric. 14). A polygonal earthwork of ca. 34 ha is attributable to this phase, while other earthworks S of the town relate to earlier periods. The town developed apace from the Flavian period on, largely because it was an important nodal point of the road system. The chief mark of transition to cantonal-capital status is the central forum-basilica of principia type (82.5 x 94.5 m). The basilica had a single row of Corinthian columns ca. 8 m high, and its apsidal aedes and other rooms were lined with imported and local marble.
Also of the transitional period is the street grid, laid out within a polygonal outer earthwork which may in part be of Cogidubnian origin; the grid (ca. 660 x 720 m) was aligned N-S, producing originally 41 insulae of different but related sizes. Of other important public buildings, the baths (Reihentyp, 63 x 28.5 m) were a creation of the regal period, lying at an angle to the street grid and adapted to it, like many other early buildings of lesser consequence. The earthen amphitheater on the NE (arena ca. 48 x 39 m) has not been excavated. A large official residency (praetorium?) lies on the S side of the town. Sacred buildings include three square and one polygonal Romano-Celtic temples, two in a walled temenos under the present churchyard, and the polygonal temple in another, S of the forum; simple cellas, including one in a walled enclosure, were probably also temples. There is yet another, or more probably a guild schola, on the main street; inscriptions naming a collegiuin peregrinorum Callevae consistentium were found in one of the Romano-Celtic fana.
Silchester has also yielded remains of a small church, ca. 12.6 m long over-all, with narthex, aisled nave, transept, and apse to the W containing a mosaic panel; there is no formal proof of its nature, though a 4th c. date is likely, and the implication of the plan is inescapable. Unlike other early churches excavated N of the Alps, this had no cemetery context, and its position next to the forum suggests that the cult had official support.
There were about 180 buildings assignable as houses or shops, the latter mainly of the strip type aligned along the main E-W street N of the forum. Some of the smallest insulae are not fully built up, and the outlying ones are largely vacant. It has been questioned whether the plan accurately represents the density of building, since the excavators undoubtedly missed some structures and could identify those of timber only with difficulty; excavation elsewhere (e.g. at Verulamium), however, suggests that the picture presented is essentially true when the results of recent aerial photography are added. It is interesting to compare Silchester with Caerwent (Venta Silurum, q.v.), where the insulae of a much smaller town are more densely built up. The houses range from cottages of two or three rooms to large establishments as much as 39 m square, often lying well back from the streets and in some cases evidently of earlier origin. About 16 larger houses were probably the residences of curiales: they have hypocausts, mosaics, and painted walls, although these are also often found in smaller dwellings. A difference between Silchester and Caerwent can be seen in the extreme rarity of the courtyard house, i.e. ranges of rooms completely enclosing a courtyard or garden. There is only one such example, and that is an organic development from a simpler building.
In general, the foundations of houses and other buildings were of flint masonry, and in many cases the superstructure was of half-timbered work filled in with clay; windows were often glazed, and the roofs were tiled (in the late period stone slabs were used in place of tiles). It is not likely that many houses had an upper story, and on the basis of the plan it seems unlikely that the townsfolk could have numbered more than ca. 2000 souls. Since Silchester occupies a somewhat elevated position at the end of a gravel spur, piped water was not so much in evidence as at Caerwent, and was probably arranged only in specific cases such as the baths. Water came from wells, and the rubbish buried in pits all over the town must sometimes have led to contamination of the supply.
At least some of the houses were farms, which accounts for the unusually spacious plan. While agriculture and allied crafts were undoubtedly the main occupation of the townsfolk, retail trade and small manufacturing are also in evidence, and general trade along the roads which converged on the town probably accounts for the presence of the collegium peregrinorum. Pottery was made; an imperial tilery existed ca. 3 km S during the reign of Nero; metal-working included the refining of silver on a small scale (from argentiferous copper, probably for jewelry), pewter manufacture, and blacksmiths' work (notable finds of iron tools). Other industries included glass making (from scrap), tanning, and bone-working.
The inner defenses consisted primarily of a bank and ditch drawn polygonally around the built-up area ca. 196-197 (?) during the Albinus crisis. The bank was faced with a substantial wall in the 3d c., 2.7 m thick at the base and ca. 7.5 m high—the circuit of the wall is all that can be seen at Silchester today (and the site of the amphitheater)—and a new ditch was dug farther out to accompany it. There were double gateways on E and W, single on N and S, posterns SE, SW, and NE, the latter leading to the amphitheater. In terms of standard wagonloads of 1200 librae, it can be calculated that the wall required over 126,000 loads of flint (from several km away) and 35,000 loads of öölitic bondingstone (which had to be brought at least 56 km).Unlike the walls of Caerwent and some other Romano-British towns, those of Silchester do not appear to have acquired external bastions in the 4th c., though various gateways came to be blocked with rubble and architectural fragments at a late date.
Occupation flourished throughout the 4th c. and well into the 5th; the town at this period seems to have had a garrison of Germanic mercenaries, to judge from items of equipment found. Later still, the lands of the last Callevans were demarcated from those of Saxon settlers on the NW by an earthwork ca. 3.2 km from the town. The latest relic is a sandstone pillar, reused as a funerary monument, bearing an inscription in Irish Ogham characters, found farther E than any other (ca. 600?).
BIBLIOGRAPHYExcavation reports: Archaeologia 40, 46, 50, 52-62, 92, 102 (1866-1969) MPI; JRS 52 (1962) I; O. Cuntz, ed., Itineraria Romana (1929-) 73-74; G. C. Boon, A New Guide to the Roman Town Calleva Atrebatum (1972) MI; id., Silchester: the Roman Town of Calleva (1974) MPI
G. C. BOON