Great Witcombe

GREAT WITCOMBE

(10 miles N.W. of Cirencester)

(1) Roman Villa (SO 899144), discovered in 1818 near Cooper's Hill Farm, was in the succeeding years largely uncovered by Samuel Lysons and Sir William Hicks. Modern excavations for the Inspectorate of Ancient Monuments were conducted by Mrs. E. M. Clifford in 1938–9 and, since 1961, by Mr. E. Greenfield. The situation is a moderate slope facing S.S.E., on Lias Clay, near the head of a broad re-entrant valley, with sand, then Oolite rising steeply above it. Springs emerging above the villa make the site naturally wet and unstable. There is a stream 100 yds. to S.E.; beside it was a small Roman building (2). Map, p. 41.

Slight indications of nearby Iron Age occupation include a ditch under the N.E. corner of the villa. Early Roman pottery and other finds point to occupation in the 1st and 2nd centuries. Foundation trenches show that a walled structure earlier than the villa lay immediately to the S., its N.W. angle under room 6. The villa is terraced. A ramp, formerly with steps, now an incline of 13°, joins rooms 1 and 34a. The main structure dates from c. 250 and occupation probably continued into the 5th century. The dominant feature is a room (15) at first rectangular (12 ft. by 21 ft.) with its N.W. corners rounded internally, later converted to an octagon of 21 ft. with an apsidal projection from the N.W. side (Plate 24). The room was fronted by a tessellated gallery, 12 ft. wide, with a 'portico' and buttresses on the downhill or courtyard side. The wing E. of the gallery contained a kitchen with an oven (34) and a latrine (35) adjacent. No mosaics have been reported in the E. wing; the S.E. rooms lay, partly at least, under post-Roman ridge-and-furrow (C.U.A.P., OAP ATZ 79). The W. wing was devoted largely to bath arrangements, with projecting plunge baths, hot and cold. There were seven tessellated floors and an elaborate threshold between rooms 6 and 8. The surviving mosaics are fragmentary; representations of fish and geometric patterns recorded by Lysons are illustrated on Plates 10–12. Materials included cut samian sherds. There are no recognisable heated 'living' rooms. Room 3 was a latrine.

The thick walls and buttresses (the latter otherwise found in this area only at Woodchester) to some extent reflect the natural difficulties of the site, as do the numerous drains, but the substantial nature of parts of the structure might also suggest that there was an upper storey over certain rooms. A possible reconstruction of the villa, made by Mr. D. S. Neal of the Department of the Environment and kindly communicated in advance of publication, shows the gallery as a dominant feature connecting with upper floors in the N. part of the W. wing and extending over most of the E. wing. Room 1 in the W. wing has interior projections integral with the N. wall; these are generally thought to have formed niches, though a buttressing function is also likely, the floor of the room on the N.W. being 5 ft. higher. A small cistern in the middle of room 1, projecting 6 in. above the Old Red Sandstone floor, was fed with water which drained into the system serving the adjacent latrine (3). Ritual usages were deduced from the niches, the water, and from a pottery model of a fir-cone found in the room.

Extensive additions and rebuilding included the conversion of room 5 from apodyterium to caldarium and the re-use of parts of Oolite columns in the structure of the hypocaust. A corner passage or 'slype' joined rooms 5 and 8. The door between rooms 5 and 6 was blocked; its doorstep, apparently reused, is grooved. Fragments of Oolite columns loose in room 6 display a variety of simple mouldings, including the common double torus base. Two sizes are represented. The smaller ones (Plate 29) have shafts about 6 in. across and are unweathered; the larger, with shafts over 9 in. across, are markedly weathered (Plate 28).

Materials include Oolite and tufa, the latter mostly from the W. wing. Most roof tiles found are earthenware, but some are of Old Red Sandstone. Imported white marble was used in some moulded cornices (a fragment is preserved in the Gloucester City Museum) and painted pieces of fine sandstone were also found. Window glass, both green and colourless, appeared in quantity in the rooms of the S.W. bath block. Painted plaster has been reported from the W. wing only, where it was extensive and notable; polychrome designs were reported up to 6 ft. high in room 1. In the plunge bath E. of room 6, opus signinum occurred as a cove between the floor and the sides. Around room 34a, inserted ovens and a spread of rubbish belong to the final phases. Fragments of fretted balustrade, 2⅓ ft. high, with Smotifs, have decoration on one face only, with keying on the reverse (Plate 28).

Early Roman finds include a coin of Domitian, much first-century Glevum ware, a Hod Hill brooch and 2nd-century samian ware. Coins are predominantly late Roman; of 26 found, about half are later than 367; one belongs to the House of Theodosius. There is much 4th-century pottery. A penannular bronze brooch (Fowler, type F) may attest activity in the 5th century. Other finds include an earthenware fir-cone from room 1, a bronze 'dog' box-handle, a bronze steelyard, a key and pins. An iron knife-coulter, now in the B.M. and once thought to provide secure evidence of a heavy plough, could have belonged to a coultered ard, though doubt must remain because of the set of the blade. There are a number of querns, one made of puddingstone. Glass includes millefiori, snakethread, beaker fragments and an intaglio. Bones include ox, sheep, pig, hare, domestic fowl and other birds. Much wood includes cherry and other species new in the Roman period. A piece of coal was found. Paint pigments were noted and analysed in two pots; an oyster shell had served as a palette. Graffiti include unintelligible lettering on a tile and the scoring of a 'lyre' on two stones.

Arch, XIX (1821), 178–83. JBAA, I (1846), 56 (intaglio). Witts (1883), 66–7, No. 22. Ant J, XIX (1939), 194 (fir-cone). Antiquity, IX (1935), 339–41; cf. JRS, LIV (1964), 54–65 (coulter). TBGAS, XXX (1907), 246; 57 (1935), 275; 73 (1954), 5–69; 74 (1955), 171–2 (graffito). JRS, LI (1961), 186; LIII (1963), 41; LVI (1966), 212; LVII (1967), 194 (graffito). Toynbee (1964), 92, 271, 334. RVB, 143, 146–7 (plaster). Lysons's plan is in the library of the Society of Antiquaries (Red Portfolio).

(2) Roman Villa (SO 90021417), on Lias Clay, lies on a gentle slope below and 100 yds. S.E. of (1). A stream flows N.E. about 10 yds. to the east. Excavations by Sir William Hicks uncovered the major part of a building with at least eight rooms. Floors of 'red plaster' were recorded in four rooms and 'some brown tiles for roofing' were found, together with coloured wall-plaster, white tesserae and further tiles. It seems that the entrance to room 1 was marked by a grooved doorstep. Although no scale accompanies the original plan, on the ground a slightly sunken platform approximately 100 ft. by 50 ft. suggests the maximum possible size of the building uncovered.

TBGAS, 73 (1954), 13–15.

(3) Romano-British Settlement (so 90121372), is indicated by sherds and roofing tiles found over approximately ½ acre on a natural terrace in the Oolite, some 100 yds. above the spring-line marking the junction with Lias Clays.

TBGAS, 81 (1962), 214.

(4) Romano-British Settlement (so 90751370), is suggested by sherds and tiles found in black soil, apparently slipped from terraces which may have been used as building platforms. Springs marking the junction between the Oolite and Lias are adjacent.

TBGAS, 81 (1962), 214–15.