Pages 24-29

Ancient and Historical Monuments in the County of Gloucester Iron Age and Romano-British Monuments in the Gloucestershire Cotswolds. Originally published by Her Majesty's Stationery Office, London, 1976.

This free content was digitised by double rekeying and sponsored by English Heritage. All rights reserved.



(6 miles N.E. of Cirencester)

Roman finds from points imprecisely located in the area of the village include late Roman coins, bronze dividers, a plumb-bob (said to have been found at Manor Farm) and, allegedly, a stone bas-relief of Aesculapius. Tesserae have been reported in the gardens of Church Row Cottages, about 50 yds. S.E. of the church. (fn. 1) From the crest of a hill N.W. of the village come an inscribed potsherd, a 4th-century coin and samian and mortaria sherds. (fn. 2) An extension of the 'White Way' probably ran N.E. across the area now occupied by the airfield and Chedworth Woods (SP 0413), but no trace of it has been found.

Several Roman sites occur in the vicinity of the villa (1), some of them in adjacent parishes. About ½ mile N., at two points 300 yds. apart in Withington (SP 05321432, 05021429), pottery has been found by Mr. A. N. Irvine. Romano-British potsherds and an iron gouge, probably Roman, were noted during excavation on and near a round barrow ¾ mile S.W. of the villa, at SP 04081285. (fn. 3) A terraced track in Streetfold (Yanworth (2)) aligns on the villa, and another track extends in its general direction S.W. from the settlement at Yanworth (1).

A description (fn. 4) of eight alleged Roman 'cisterns' is likely to have been inspired by details of 'The Capitol' (3).

(1) Roman Villa (SP 053135), Chedworth Woods, was discovered by chance in 1864. It stands, 100 ft. below the level where Fuller's Earth overlies Inferior Oolite, at a point where three gullies join to form the head of a single valley which falls gently E. towards the R. Coln, 250 yds. distant. Soon after discovery the extensive remains were exposed by James Farrer; later they were partly built up and roofed over by Farrer's nephew, Lord Eldon. Much patched, they are now in the care of the National Trust. Plan, p. 26.

Certain features shown on the plan are no longer visible on the ground. Some rebuilt walls differ slightly in position or structure from the original elements, and ground-levels, notably in the courtyard, are in places very different from those of Roman times. The 'fulling establishment' which once was thought to occupy the W. of the N. range is now interpreted as a muchaltered bath house.

The N. and W. ranges of the villa stand on artificial terraces partly recessed into the hillside; the S. wing is set lower, near the valley floor. The complete form of the S. wing remains uncertain, as does the E. limit of the N. range, in an area of slumping. Debris and indications of further structures have been noted in pits dug for electricity poles at points up to 255 ft. E. of room 1a in the S. wing. There has been confirmation that much debris occurs around the villa, and dark earth, coins and pottery have been found on rough terraces above the steep artificial scarps which rise from the central part of the W. range and from the N. range. In the latter area copious finds and an alleged 'chamber' were noted in 1864, before there could be any possibility of confusion with excavators' spoil heaps. Occupational debris lies against the walls of the S. wing.

Small excavations carried out in recent years by Sir Ian Richmond and others have shown development from the early 2nd century to the late 4th century, with evidence of fire in early and late phases. Richmond's excavations in 1958–65 (upon which much of our information as to sequence and function depends) consisted of sections or selective clearance in the following areas: the corridor between rooms 1 and 2; in and near rooms 1b, 3, 6, 7, 8 and 12; the nymphaeum (17); rooms 19, 21, 21a, 22, 24, 25, and 25a; the corridor S.W. of room 30 and S. of room 32; E. of room 33, and the terrace N. of room 30.

Monuments in Chedworth and Yanworth.

The early structures, three separate blocks, formed three sides of a rectangle open on the E. The W. end of the S. block, a timber-framed gable, suffered from fire before reconstruction and enlargement, as did much of the W. range. The S. wall of room 3 is built up from footings that are slightly offset. These could be part of an early precinct wall, otherwise traceable to an angle about 45 ft. W. of room 5a and thence in fragmentary exposures N. to a possible N.W. angle under the nymphaeum; its upper courses, when formerly exposed, partly revetting the W. scarp, were seen to be of herring-bone construction. The early details of the W. range are largely obscured by later development; an entrance 4½ ft. wide occurred in the W. wall of the narrow compartment N. of room 7. The N. range was largely taken up by a bath suite of normal Roman type. The lower portions of two lathe-turned columns (Plate 26) remain in situ on a wall beside the N. range (20). The main water supply to the villa was from a spring, S, at the N.W. corner of the site; it was channelled into a cistern just outside the suggested N.W. angle of the precinct wall. A roughly circular sinking 6 ft. across, possibly the site of a well, occurs at H, some 30 yds. E. of room 12.

Chedworth. (1) Roman Villa. Chedworth Woods.

Subsequently the villa was enlarged and the earlier buildings were integrated into a single complex. Corridors, interrupted on the E. side by a gateway, defined the inner courtyard and linked the N. and S. ranges. Room 7 was enlarged and the doorway on the W. was blocked. The bath suite in the W. wing (rooms 8–14) was established. The damp-heat baths in the N. wing (rooms 19–24) were converted by stages to a dryheat establishment of sauna type (Plate 25). Lead pipes connect the immersion baths (23). An entrance 7½ ft. wide into room 21 from the S., and another 4¼ ft. wide in the W. side of the western bath of 23 were blocked. A floor built over the stoke-hole S. of room 24 sealed a fill containing Oxfordshire red colour-coated ware, conventionally datable after 270. A porticoed nymphaeum (17) was constructed, having at its centre an octagonal cistern holding 1,100 gallons into which the main water supply was funnelled; from it the water was allegedly distributed through a stone-lined junction box. The W. wall of the nymphaeum crosses the original cistern and the south stylobate overlies the N. wall of the precinct, possibly at its former N.W. corner. A column 5½ ft. high, now loose in the bath suite near room 23, probably comes from the nymphaeum portico. The nymphaeum was associated with Christianity when chi-rho inscriptions were carved on at least three of the trapezoidal stones edging the cistern. These stones, recognisable by their shape, were removed late in the Roman period and one was built into the footings of steps to room 10.

The wall which extends W. from the S.W. corner of the nymphaeum is probably an addition, designed to overlap rather than to join the surviving part of the presumed precinct wall on the W., thus providing a ramped approach.

In its final form the villa covered at least 2 acres, but the E. limits have not as yet been determined. It was arranged in tiers, with steps joining the different levels and leading to rooms with floors raised above ground-level hypocausts. The corridor beside the final E. extension of the N. range represents the widening of an earlier version; at the E. end it is built above earlier Roman levels. The later masonry incorporates much reused stone; a baluster hypocaust pillar, for instance, is built into an apse of room 22. Drains led from the baths into the courtyard areas; one served the latrine (4); another is still visible N. of room 32. There were two dining rooms (5 and 32) and two kitchens (3 and 30) as well as bath suites in both W. and N. ranges. The piers of the channelled hypocaust in room 33 are about a foot higher than those surviving in room 32, suggesting the possibility that it was a dais at the E. end of dining room 32, rather than separate. The S. wing, as so far explored, had no heated living rooms or mosaic pavements. The latrine 4, S. of this wing, was finally altered for some different use. Outside it, in the angle with room 1b, an infant burial was found in a Blue Lias slab cist, inserted in rubbish spanning the 2nd to 4th centuries (dated by excavation in 1954).

Fire in the last Roman phase is indicated by the discovery of 67 lb. of melted lead in a room of the N. range. The villa has suffered from extensive stone robbing, and much stone, including calcined fragments of sculpture, has been found in and around a lime-kiln on the terrace 20 yds. N. of the N. range.

Tessellated floors of 4th-century date exist or existed in at least fifteen rooms and over most of the W. and N. corridors (Plates 2–7). All the mosaics except those of rooms 5 and 10, which are recognisably of the Corinian school of mosaicists (Plates 2–5), have geometric patterns. Two mosaics in the W. range were superseded: a pavement was crossed by the N. wall of room 6, and a pavement in room 14 (Plate 6) had a floor of Lias flags built over it. The 'hall' (25a) is said to have had tesserae over its entire length, some 56 ft. Repairs which resulted in changes of detail are seen in rooms 5 and 10. Patches of ancient burning in these rooms include one on the tessellated threshold of room 10, unlikely to be the result of scorching from a brazier. Tesserae vary in size from about ¼ in. in fine detail to 1½ in. for borders. Colours derive from the materials used: red from tile or Old Red Sandstone; blue from Lias; white and brown from limestones. Pennant flags, concrete and opus signinum were used for floors in all phases.

Heating was by means of hypocausts with pillars of tile (rooms 5, 11 and 12), of stone (rooms 24, 26 and the S. part of room 32), or channelled (rooms 6, 10, 24a, the N. of 25, and parts of rooms 32 and 33). The stone pillars, 108 originally in room 26, are of limestone, up to 2½ ft. high, square in section and expanded, balusterlike, at top and bottom to squares of 7 in. to 10 in. (Plate 25). Rearrangement involved the blocking or filling-in of hypocausts in rooms 6, 8, 21, 22, S.E. of 24 and 26. Fragments of stone pillars are built into the piers of the channelled system in the N. of room 32 and whole pillars into the S. apse of room 22.

Architectural details (Plate 26) and small finds contribute to the picture of a rich villa, with some poor objects possibly from the final phase. Painted plaster was extensively used inside, and there was also some marble facing; externally there was cement rendering. Room 21 had a quoin of tiles. The apses of room 22 were jacketed with flue tiles. Hollow baked clay voussoirs were noted from the roof of compartment 12. Other roofing materials include tegulae, hexagonal stone-slates and lead. Two stone bases with the surviving feet of two small figures probably once stood in a recess in the W. wall of room 5b. Fragments of fretted limestone balustrade embodying an 'S' motif probably derive from a corridor or a veranda. Other architectural stonework includes numerous column fragments, part of a cornice, and ridging stones. Huge iron beams weighing 256, 356 and 484 lbs., found in the corner of room 19, had probably been used to support a massive hot-water tank of copper or lead.

Inscribed stones include the three mentioned above, with chi-rho monograms, and one with ruled lines, all removed from the nymphaeum, and a building stone which had PRASINA carved on it.

Four altars were found. One, barbarously carved, is dedicated to Mars Lenus. One with a crude pilaster-like body and with scribed or drilled detail probably represents a Celtic god. The third, uninscribed, was found buried in the nymphaeum. The fourth, very crude and with saltires on the sides, was found just outside the S. wing.

The abundant finds include 360 coins, mostly from positions in or near room 2, suggesting that it was a steward's office. At least one-third of the coins belong to the period 364–78 and only 2 per cent. date from before 240; the three latest recognisable are of Gratian (issue of 378–83). Pottery includes samian ware of the early 2nd century (some with potters' stamps), an amphora with potter's stamp, a Rhenish beaker, large quantities of Oxfordshire late colour-coated and some painted ware, New Forest beakers, Nene Valley and a little Severn Valley ware and black-burnished ware of the 2nd to 4th century. A silver spoon inscribed CENSORINE GAUDEAS, now lost, was found in 'rubbish' on the W. of the W. range. Bronze finds include spoons, rings, a stylus, two prick-spurs of rivet type and the cast finger of a large effigy. Other metal finds include a pewter jug, iron saws, chisels, knives, shears, spadeirons, alleged hunting arrows, a pair of small shackles, horseshoes, hinges and locks. Glass was found from windows, bottles and bowls. Stone finds include a quern-stone 2½ ft. in diameter from room 30, a rectangular basin from room 4, and moulded and carved table-tops. There was a moulded disc of Kimmeridge shale, and bone pins, needles and handles. Other bones included two pieces of a human skull, and bones of pigs and sheep. There were antlers of red deer and oyster shells.

A late Roman zoomorphic buckle (Hawkes and Dunning type IIA) was found roughly 25 yds. E.N.E. of the museum, in the outer court of the villa. The museum contains at least two finds of uncertain origin. St. Clair Baddeley suggests that a unique brooch with seven human heads embossed on a bronze disc could have come from a quarry (perhaps that adjacent to the temple (4)) in the neighbourhood. A small limestone Christian cross is unprovenanced. Most surviving finds are in the museum, but some architectural fragments are dispersed in the rooms of the villa.

JBAA, XXIV (1868), 129–35; XXV (1869), 215–25; XXVI (1870), 251–2. PSAS, VI (1868), 278–83. Arch, LIX (1905), 210–14 (fulling theory). TBGAS, 76 (1957), 160–4 (room 4 and adjacent); 78 (1959), 5–23 (reinterpretation of laconicum), and 162–5 (coins); 86 (1967), 102–6 (mosaic). Ant J, XXXIX (1959), 66 (spurs). Med. Archaeol., V (1961), 51, No. 5 (buckle). Arch J, XLIV (1887), 322–36; CXXII (1965), 203. Toynbee (1964), passim. RIB, Nos. 126–8. Rivet (ed.), The Roman Villa in Britain (1969), passim. Britannia, II (1971), 200–2 (iron beams).

G. E. Fox collection in Soc. Ants., London, Box I, sheets 8–17; Box III, sheets 25, 29–31. Site drawings and notes by Sir Ian Richmond in Ashmolean Museum. Information and personal observations from Mr. A. N. Irvine (warden) and Mr. R. Goodburn.

(2) Roman Building (SP 05121358) in Chedworth Woods, 170 yds. N.W. of (1), is probably the structure called 'The Capitol' by its 19th-century discoverers; it was destroyed in the construction of a railway. The building stood precisely at the head of the narrow gully taking 'Dark Lane' N.W. from (1). Several small rooms are said to have been partly cleared, and in 1869 H. M. Scarth refers to the building as a 'circular temple'. The area immediately W. of the railway is flat enough for buildings, but it shows no sign of disturbance.

Finds include coins, hexagonal tiles, fragments of pillars, part of a shell-headed niche, and glass tesserae. The stone relief of a 'hunter god' with hare, dog and stag, sometimes ascribed to (4), might have come from this site. Surviving finds are in the museum at (1).

PSAS, VI (1868), 283. JBAA, XXV (1869), 222. Arch J, XLIV (1887), 323. Toynbee (1964), 179.

(3) Roman Building (SP 05571346), about 230 yds. E. of (1), is now a low knoll, perhaps partly artificial; cf. Duntisbourne Rouse (4), (a). The knoll extends 50 ft. N. from the road and has a nearly level top, about 40 ft. wide. Building debris can be seen on it when it is under the plough.

(4) Roman Temple (SP 06111329) in Chedworth Woods, ½ mile E. of (1), is now a massive but much disturbed artificial platform some 35 yds. from, and 50 ft. above, the flood-plain of the R. Coin. Excavations in 1864–5 and in 1930 show that the platform is a relic of a stone temple (Lewis's Romano-Celtic type IA) about 50 ft. square, with a colonnaded portico and cella set on a podium of very large hewn limestone blocks. The downhill scarp of the platform is 12 ft. high and a scarp rises uphill from it to a height of about 8 ft. The orientation of the platform is roughly 10°. In 1931 it was suggested that there might be an extension or annex at the N.W. angle of the platform.

A round-bottomed stone-lined hole, 7 ft. across and 5 ft. deep, midway along the ambulatory on the E. side, contained bones of red deer. Other finds in the platform area included the drums of stone columns 1½ ft. in diameter, and a fragment of a capital; also pieces of moulded stone architrave, sandstone hexagonal slates, tiles, opus signinum and concrete. Similar debris just outside the platform included 'hypocaust tiles', stone troughs and blocks of tufa. Much stone debris is said to have been taken for the repair of buildings in the area. Undated fragments of human skull were found in the ambulatory. Coins ranged from the mid 2nd to the 4th century. Two slabs roughly scribed with armed figures are said to have come from the site. The relief of a 'hunter god' with animals (see (2)) is sometimes ascribed to this site. Surviving small finds are in the museum at (1).

PSAS, VI (1868), 262. JRS, XIV (1924), 231. TBGAS, LII (1931), 255–64. Lewis (1966), passim.

(5) Romano-British Settlement (SP 06121307) in Chedworth Woods, 250 yds. S. of (4), was noted by James Farrer in 1865. The site is a natural shelf on a steep hillside at the junction of Inferior Oolite and Fuller's Earth. Traces of walls occur at the head of a rise in the shelf (about 06101307). Black earth and Romano-British potsherds have been found in an area extending about 100 yds. E.

PSAS, VI (1868), 283. Information from Mr. A. N. Irvine.

(6) Roman Villa (SP 07011175) at Listercombe, discovered by chance c. 1760, stood almost on the valley bottom, near the junction of Inferior Oolite and Fuller's Earth, on ground sloping gently to an adjacent stream. Other remains occur a little uphill to the W., below the steep (perhaps scarped) valley-side. The site has been explored several times without publication of results. A mosaic pavement is said to have been seen in 1892 during the rebuilding of a field wall immediately W. of the remains shown on the plan (information from Mr. J. Scotford). In 1930 Mr. C. E. Key uncovered a hypocaust (at least 10 ft. by 5 ft., internally) with tile pilae on Oolite bases; traces of walls, possible floors and a paved 'corridor' lay to S.E. Wall footings 4½ ft. wide extended S.W. for at least 90 ft. from the side of the hypocaust. Fifty yards to the W., the corner of another Roman room was discovered in an area where earthen platforms survive.

Building material included tegulae, Cotswold stone roofing tiles, box-tiles and painted wall-plaster. Former reports note a bath or cistern, tiles stamped Arveri from a hypocaust, and tesserae. It is likely, though not certain, that a hypocaust uncovered before 1865, and perhaps again by St. Clair Baddeley in 1926, was the same as that dug out by Mr. Key. Finds, in Cheltenham Museum, include six coins (untraced) one of them allegedly 'Constantinian'; also pottery including Oxfordshire colour-coated ware, a 4th-century mortarium and calcite-gritted wares, iron strapping and nails, a stone cosmetic palette, and bones and oyster shells.

Rudder (1779), 334. Bigland (1791), I, 305. JBAA, XXV (1889), 222–3. Chedworth Villa Guide (1926). JRS, XXI (1931), 239. Cheltenham Echo, 11 Apr., and Cheltenham Chronicle and Gloucestershire Graphic, 23 Aug. 1930. Information provided by Mr. C. E. Key, whose plan is in Cheltenham Museum. Plan by St. Clair Baddeley in Gloucester City Library.

R.A.F., VAP CPE/UK 1913: 4094–6.


  • 1. Information from Mr. J. Scotford.
  • 2. Sherd in Corinium Museum; other finds in private possession. JRS, XXXVII (1947), 182; XXXVIII (1948), 104. Other information from private sources.
  • 3. W. F. Grimes, Excavations on Defence Sites 1939–45 (1958), 133.
  • 4. TBGAS, LII (1930), 259, n. 6.