Tewkesbury lies at the northern tip of Gloucestershire, adjacent to the county border with Worcestershire. It is around 15 miles south of Worcester and 10 miles north of Gloucester. The town itself is located on the east bank of the River Avon, at the point where the Avon meets the River Severn. Tewkesbury Abbey stands south of the town centre on a quasi-peninsula formed by the confluence of the River Swilgate, which runs from the east, with the Avon. The area has been prone to flooding over the centuries and especially in recent years. Fortunately the abbey stands on a slightly elevated point, although this has not fully protected the building from flood damage.

Tewkesbury Abbey is regarded as one of the most important Romanesque buildings in the British Isles due to its architectural features and the state of preservation. The church has a cruciform plan comprising an 8-bay aisled nave (with additional 9th bays at the west end of the aisles), aisleless transepts with east-projecting chapels, and a 2-bay chancel flanked by an ambulatory with radial chapels. All of the ambulatory chapels are later medieval constructions. Many alterations were made to the E arm and vaulting throughout the church during the 14thc. There was originally a cloister to the S of the nave as well as other claustral buildings, however these were demolished after the Dissolution.

The abbey was restored by Sir George Gilbert Scott between 1865 and 1879. During this period, the stone screen dividing the nave from the crossing and E arm was removed. One bay of the cloister was reconstructed by Thomas Collins at the end of the 19thc and the W front was repaired in 1906. Various projects were carried out between 1932 and 1938, including repairs to the tower and the strengthening of pier foundations. The site of the E ambulatory Lady Chapel was excavated in 1940 by Sir Charles Peers and Thomas Overbury (the footprint of this structure is marked out on the grass at the E end of the abbey). Major repairs were made to the abbey roofs between 1978 and 1985. The roofs of the eastern chapels were repaired between 1994 and 1996. Repairs were made to the W nave turrets between 1999 and 2002.

There is evidence of Romano-British settlement around Tewkesbury from the 1st to 3rd centuries. In the late Anglo-Saxon period, Tewkesbury was a wealthy fortified settlement with a church and probably a market. The remains of a late Anglo-Saxon timber hall were excavated at nearby Holm Hill in 1974 (Heighway 2003, 1–4; Hannan 1997). After the Norman Conquest, Tewkesbury became a royal manor (Moore 1982, 163b—d). Later, in or shortly after 1087, the manor was granted to Robert fitz Haimon, also known as ‘Fitzhamon’ or ‘fitz Hamo’ (d. 1107). He was a prominent member of the Norman elite who supported William Rufus in the rebellion of 1088. William Rufus granted him a rich swathe of land in Gloucestershire, which included Tewkesbury and Bristol, and Robert used his new economic and territorial position to make advances into Glamorgan, Wales. Initially, he made benefactions to St Peter’s Abbey in Gloucester (now Gloucester Cathedral), which included churches in recently conquered areas of Wales (Green 2004).

Robert founded the new abbey of Tewkesbury at some point after 1087 (see ‘Comments/Opinions’ for a more detailed discussion). The eastern arm of the church was probably completed by 1102, at which point it was populated by monks from Cranborne Abbey (Dorset), and the building was consecrated in 1121, although building work is thought to have continued into the second quarter of the 12thc (Luxford 2003, 57; Thurlby 2003a, 89). After Robert fitz Haimon’s death in 1107, Tewkesbury passed to his son-in-law, Robert, an illegitimate son of King Henry I who was later appointed earl of Gloucester. By extension, Robert became the new patron of the abbey and presumably oversaw its completion (Crouch 2004).

The area suffered some disruption during the succession dispute of Stephen’s reign (1135—54). Tewkesbury was raided by Waleran of Meulan, Earl of Worcester, in 1140 which resulted in Robert Earl of Gloucester’s ‘magnificent residence’ being burned to the ground (McGurk 1998, 282–3). The residence is thought to have been located at nearby Holm Hill and has been identified with a fire-damaged structure excavated in 1974 (Hannan 1997, 95). In 1178, the abbey itself was ravaged by fire (Thurlby 2003a, 89).

Original plan and appearance of the Romanesque abbey

There have been many alterations to the abbey since its initial construction, although the present floor plan substantially corresponds with that of the Romanesque church. The original ambulatory probably possessed only three chapels, projecting NE, E and SE, respectively (Thurlby 2003a, 89). The lost cloister (of five bays) and claustral buildings were located on the S side of the nave.

There has been extensive scholarly debate on the original form of the chancel and ambulatory. The current consensus is that the chancel possessed a giant order; in other words, the six cylindrical piers were originally about twice as tall and rose to gallery level. These were shortened when the chancel and ambulatory were remodelled in the 14thc. There is evidence that the chancel possessed a high barrel vault, probably constructed from tufa, and the ambulatory was originally groin vaulted (Fernie 1985, 3; Thurlby 1985a, 36–8; Thurlby 2003a, 95–6; Birdsall and Morris 2003, 258–61). The S transept chapel possesses a rib vault. If this is an original late 11thc feature, it may be one of the earliest examples in the British Isles (Fernie 2000, 165).

The suggestion that Tewkesbury Abbey originally had a four-storey elevation has been rejected in recent years. Breaks in the masonry of the chancel and nave, and fire-reddened stone in the S transept indicate that the church was probably barrel-vaulted throughout and such an arrangement would have precluded a clerestory. Only in the 14thc (when the nave vault was replaced) were clerestory windows inserted into the nave (Fernie 1985, 4; Thurlby 1985b, 5, 17; Thurlby 2003a, 100—1; Birdsall and Morris 2003, 261).

External features of the church have also been altered over the course of successive rebuilding and restoration campaigns. The blind arcading on the north side of the nave has been extensively altered and renewed (see discussion above), and the external capitals of the N porch opening are modern replacements. It should also be reiterated that the W doorway originally had seven orders and the inner order is now missing. The baluster-shaped shafts in the openings of the W turrets were inspected between 1999 and 2002 and deemed to be 18thc replacements (Birdsall and Morris 2003, 264). It is unclear whether the heavily weathered and damaged turret fragments held in the Parvis Chamber are examples of this 18thc work or original parts of the turret fabric.

Chronology of construction

The construction of Tewkesbury Abbey is thought to have followed the conventional pattern with work beginning at the E end of the church and gradually progressing westwards.

It is clear from Domesday Book that Robert fitz Haimon acquired Tewkesbury manor no earlier than 1087 (Moore 1982, 163b—d). An early account of the abbey’s foundation history is preserved in the the 14th/15thc ‘Founders’ Book’ (Oxford Bodleian Library MS. Top. Glouc. d.2) and this indicates that an unspecified number of years elapsed between Robert acquiring the manor and founding the new abbey. The narrative also records that monks from Cranborne, led by their abbot, Gerald, settled at the pre-Conquest church of Tewkesbury three years before the new abbey was founded. On the basis of this evidence and the probability that the eastern arm of the new abbey church could have been substantially completed and functioning within five years, Luxford (2003, 57) has argued that Gerald arrived c. 1094 and building work was initiated c. 1097. A more conservative estimate, allowing a decade for the construction of the eastern arm, would place the foundation c. 1092.

In Thurlby’s opinion (2003a, 89), the chancel, ambulatory, crossing, transepts and first two bays of the nave were all completed by 1102, the year that the monks from Cranborne moved to the new abbey. With the possible exception of the transition from square to circular plinths after the 2nd bay of the nave (which could mark a building break), there is no physical or documentary evidence to confirm this interpretation. The Worcester Chronicle records that the abbey church was consecrated in 1121. This has been taken as evidence that the nave was completed, although work on the upper parts of the tower and the claustral buildings is thought to have continued beyond this date (Thurlby 2003a, 89).

The earliest fabric of Tewkesbury Abbey is sparsely decorated with sculpture like most large churches built in the late 11thc. However, there was evidently a shift c. 1115 when more elaborate geometric and figure designs were introduced to the capitals of the nave triforium. There can be little doubt that the builders, designer(s) and/or patron were influenced by emerging regional and national trends for more richly carved architecture. At Tewkesbury, this culminated with the sumptuously decorated crossing tower and cloister, probably constructed in the 1120s and 1130s.

Geology

There were efforts to geologically analyse stone fragments from the abbey at the start of the 21stc. Oolitic limestone was the primary building and sculpting material in the 12thc and could have been quarried from a number of nearby locations, including Stanway, Cleeve (Gloucestershire) and Bredon Hill (Worcestershire). The abbey may have owned a quarry at Stanway which would make it the prime candidate (Birdsall and Morris 2003, 272—3).

Parvis Chamber fragments

The Parvis Chamber fragments deserve additional discussion regarding date, iconography and function. Nos. 79/1A, 79/1B, 79/1C and 82/183 appear to be contemporary and perhaps derive from the same scheme. In his catalogue, Richard K. Morris dated them all to the mid-12thc and compared the richly decorated shaft fragments to the W front of Lincoln Cathedral. In Thurlby’s opinion (2003a, 103), 79/1A—C originate from the 12thc cloister. The base fragments (nos. 79/4 and 80/3) may also derive from this scheme.

No. 79/1C is of special interest because of the quality of the figure sculpture. In the present writer’s opinion, the best-preserved face of the capital depicts a man blowing a horn and could have been part of a hunting scene. Thurlby (2003a, 103), on the other hand, has interpreted the cylindrical object on the right-hand side as the man’s arm and suggests that the capital depicted a man riding a horse.

No. 13/40 has been identified as a bull’s head (Birdsall and Morris 260—1). It is so badly damaged that other interpretations cannot be ruled out. As a piece of architectural sculpture, it is likely to have functioned as a label stop. In terms of style, the fragment can be assigned to the second quarter of the 12thc.

The damaged cross head (no. 79/16) appears to date from the early or mid-12thc judging from the diagonal tooling and the application of beading.

Richard Morris was uncertain whether no. 82/252 was 12thc or Gothic. Based on the rigidity of the carved leaves and the depth of the carving, the present writer suspects the latter. There appear to be vestiges of red pigment on the surface of the stone.

Semi-circular relief set in the W wall of the N transept

The date, iconography and function of this relief is uncertain. It is bonded into the original masonry of the N transept which suggests that it cannot have been carved any later than the beginning of the 12thc. The shape of the relief and the arrangement of the figures vaguely recalls the 12thc tympanum at nearby Tredington church (Gloucestershire). If it was originally created to function as a tympanum, it is unclear why the relief ended up in its present location. Another possibility is that it pre-dates the foundation of the Norman abbey and is a rare example of an Anglo-Saxon tympanum, perhaps originating from the pre-Conquest monastic church at Tewkesbury.

Relationship to pre-Conquest architecture and sculpture

Double-splay (or twin aperture) windows, like those on the central tower and turrets, are typically associated with Anglo-Saxon architecture and continued to be used after the Norman Conquest (Thurlby 2003b, 128). Many elements of the turrets suggest pre-Conquest influence, including the tapering baluster-style shafts, the herringbone-style cable moulding on the stringcourses, and perhaps even the billet-enriched labels. The issue of whether the 18thc baluster-style shafts reflect original 12thc features is a moot point (Birdsall and Morris 2003, 264).

Fragment no. 13/40 in the Parvis Chamber has been identified as an animal-head label stop. Sculpted label stops of this kind are found in several regional churches dating from the second quarter of the 12thc, however famous pre-Conquest examples can be seen at the nearby church of St Mary, Deerhurst, Gloucestershire (Thurlby 2003b, 131; idem, 2015, 346). This raises the possibility that an Anglo-Saxon architectural feature was consciously revived in the 12thc.

External decoration of the abbey church is characterised by the extensive application of blind arcading. Similar use of the ornament can be seen at St Laurence’s church, Bradford-on-Avon (Wiltshire), which is believed to date from the first half of the 11thc (Fernie 1983, 145—51).

Relationship to other sculpture

The architectural affiliations of Tewkesbury Abbey are discussed at length by Fernie (2000 160—65) and Thurlby 2003a, 103—8). Evidently the builders, sculptors and patrons of Tewkesbury were influenced by a number of regional and international models. The grand W entrance at Tewkesbury is thought to have been inspired by the W portal of the late 11thc bishop’s palace at Hereford, and, in turn, this has been traced to imperial churches in Germany (Fernie 2000, 164; Thurlby 2003a, 104). In France, it is common to find church exteriors decorated with blind arcading like at Tewkesbury (Thurlby 2003a, 98). The volute capitals with upright curling leaves and sunken stars in bays 1 and 6 of the N nave triforium may have derived from Normandy since similar capital designs can be seen at the chapel of Sainte-Paix and the abbey of Sainte-Trinité, Caen.

There can be little doubt that Tewkesbury Abbey was closely modelled on St Peter’s Abbey, Gloucester (now Gloucester Cathedral), which was rebuilt under Abbot Serlo beginning in 1089 (Fernie 2000, 157, 164). As well as similarities in plan, the large cylindrical piers at Gloucester have the same impost and base profiles. Also akin to Tewkesbury is the profusion of mitred cushion and scallop capitals. The N nave aisle of Gloucester Cathedral features volute capitals with foliate decoration that are comparable to the volute capitals in bays 1 and 6 of the N nave triforium at Tewkesbury. Blind arcading can also be found at Gloucester in the chapter house. Based on these similarities and the close proximity of Gloucester to Tewkesbury, it is possible that masons and sculptors moved between the two sites.

Another possible model for Tewkesbury Abbey was Worcester Cathedral which Bishop Wulfstan had begun to rebuild in 1084. Besides similarities in plan, cushion capitals with mitred angles and billet ornament can be found in the late 11thc crypt, and blind arcading, albeit of a more complex intersecting variety, can be seen in the chapter house. Other Tewkesbury capitals, namely the grooved volute capital in the 8th bay of the S nave triforium, share affinities with early 12thc capitals at Hereford Cathedral (Thurlby 1980, 89).

Thurlby (1980, 93; 2013, 69) has compared the figure carvings on the Tewkesbury N nave triforium capitals and Parvis Chamber fragment no. 79/1C to the sculpture of the renowned Herefordshire School, going as far to argue that members of the school were employed at Tewkesbury from c. 1115. The bearded male in the 8th bay of the N nave triforium has the same cap-like hair, ribbed clothing and bulbous eyes as many examples of figure sculpture associated with the Herefordshire School, but particularly the fighting men depicted on the font at Eardisley. Similar stooped male figures can ben seen on the N side (3rd order) of the W doorway at Leominster Priory (Thurlby 1980, 92). In the 7th bay of the Tewkesbury N nave triforium, the eastern capital depicts a grotesque with a human head in its mouth. The same motif can be seen on corbel NN10 (T69) at Kilpeck church (Thurlby 2013, 69). To this comparative analysis can be added decorated shaft 79/1A in the Parvis Chamber. Very similar interlacing tendrils and splayed foliage can be seen on an abacus in the N nave arcade of Hereford Cathedral and on the inner E jamb of the Kilpeck S nave doorway (illustrated in Thurlby 2013, 59, figs. 12 and 13).

The stepped and radiating tympanum of the Tewkesbury N nave doorway is very unusual, although similar constructions can be seen above doorways at the nearby churches of Twyning and Lower Swell, Gloucestershire (Thurlby 2003a, 92).