Volume IV | Chapter 7 | The Development of Sculpture c. 950–1100Next Back to catalogue index
by D. Tweddle

The last century of Anglo-Saxon England witnessed an extraordinary increase in the survival and, therefore, probably in the production of stone sculpture in the south-east. Even excluding the material from the excavations at Winchester (see Chap. VIII), nearly two hundred sculptures survive which can be dated to this period, as opposed to about three dozen surviving from the period between c. 600 and 950. This contrast must reflect more than differential rates of survival. In addition there appears to have been a radical change in the types of sculpture produced. In the earlier period almost half the pieces were from cross-shafts or cross-heads, about one-third were architectural sculptures, and only five were monumental sculptures. In the late tenth and eleventh centuries there seems to have been almost a complete reversal. Free-standing crosses account for less than one tenth of the surviving examples, whereas monumental sculptures account for about a quarter. More than half, however, are architectural. The remaining pieces are unclassified. The period also seems to have witnessed the growth of a number of local workshops, apparently divorced from contemporary shifts in fashion, such as those producing the grave-covers of Surrey and Sussex, discussed in more detail below. The bulk and diversity of the material has prompted a sub-division by function. Within each group the dating of the individual pieces is then tackled.


Apart from the Winchester fragments (Chap. VIII), the crucifixion panels and figural scenes, and the sundials, which are all discussed separately below, more than 50 of the late Anglo-Saxon architectural sculptures from south-east England are in situ and another three sozen or so are ex situ. In dating the in situ sculptures, one approach is to attempt to date the fabric of which the sculpture forms a part, that is to use primary dating evidence. Apart from the problems of identifying reused pieces, and pieces which are later embellishments to a fabric, discussed in Chap. IV, the fundamental difficulty in employing this approach is that none of the fabrics is dated by inscription or has any surviving contemporary documentation, and the close stylistic dating of pre-Conquest fabrics still presents almost insuperable difficulties. The best that can yet be achieved is to separate those fabrics which are of early, middle, and late Anglo-Saxon date (Taylor and Taylor 1965–78, iii , 1068–70). This is of help in at least identifying some of the sculptures which might be assigned to the late tenth or eleventh centuries.

Taylor and Taylor have isolated several criteria which serve to distinguish late Anglo-Saxon fabrics. These include the use of double-splayed windows, long and short work quoins used in conjunction with pilaster strips, hoodmoulds and stripwork, and belfry towers. Using these criteria the following churches from south-east England which employ architectural sculpture can be identified. Hadstock, Essex; Breamore, Boarhunt, and Corhampton, all in Hampshire; North Leigh, St Michael in Oxford and Langford, Oxfordshire; and Sompting, Sussex. The church at Headbourne Worthy, Hampshire, also exhibits some of these features, but the only sculpture is the crucifixion group, discussed in detail below. Less survives of the church of Botolphs, Sussex, but its chancel arch shares a number of features with the fabrics at Sompting, Langford, and Hadstock. A late pre-Conquest date can also be advanced for parts of the fabric at Walkern and Little Munden, Hertfordshire, but on different grounds from those proposed by Taylor and Taylor.

In considering these eleven late Anglo-Saxon churches three architectural groups can be identified. The first group, comprising the Hampshire churches of Boarhunt, Breamore, and Corhampton, employs almost all of the features noted by Taylor, but belfry towers do not occur. At Boarhunt there are external pilaster strips, stripwork, and a hoodmould around the chancel arch, and there is a double-splayed window (Taylor and Taylor 1965–78, i , 76–8, fig. 34, ii , pl. 394). At Breamore there are long-and-short work quoins, external pilaster strips, and double-splayed windows (ibid., i , 94–6, fig. 42, ii , pls. 405–6). At Corhampton, in addition to these features, there is stripwork and a hoodmould outlining the north door (Ills. 442–3) and the west face of the chancel arch (ibid., i , 176–9, figs. 74–5, ii , pl. 440). The second group consists of the church at North Leigh and Oxford (St Michael). At both places only the plain, unbuttressed, west tower survives, with baluster shafts employed for the belfry windows. There are long-and-short work quoins on the north side at St Michael's, but the southern quoins are of rubble (ibid., i , 481); at North Leigh the quoins are side-alternate (ibid., i , 464), there are also single-splay windows at first floor level (ibid., i , 464, fig. 222). The third group of buildings includes the fabrics at Botolphs and Sompting, Sussex, Hadstock, Essex, and Langford, Oxfordshire. These employ some of the features noted by Taylor and Taylor, long-and-short work quoins and double-splayed windows at Sompting (ibid., ii , 558–62, figs. 271–3), Langford (ibid., i , 367–72, figs. 164–8, ii , pls. 511–13), and Hadstock (ibid., i , 272–5, fig. 120, ii , pls. 479–80), and pilasters and belfry towers at Sompting and Langford. More distinctively, all the churches in the third group employ features characteristic of Romanesque architecture, in particular, engaged angle shafts at Hadstock and Langford, and soffit shafts and rolls at Botolphs, Langford (Ill. 297), and Sompting (Ills. 183, 186). In addition there is an angle roll on the arch head of the north door at Hadstock (Ill. 277).

Churches in the first group and second groups are difficult to date more closely than simply to the late Anglo-Saxon period, although attempts have been made to do so. Taylor and Taylor, for example, place Breamore, Hampshire (Taylor and Taylor 1965–78, i , 94), and North Leigh, Oxfordshire (ibid., i , 464), in the period c. 950–1000, and Corhampton, Hampshire (ibid., i , 176), and Oxford (St Michael) (ibid., i , 481) in the period 1050–1100, but they are unable to refine the dating of the church at Boarhunt, Hampshire. However, the evidence on which these assertions are based is not always adequately argued. In contrast, the dating of the third group of buildings can be narrowed considerably as the Romanesque features which they embody only appeared on the continent in the course of the eleventh century, providing a terminus post quem for the English examples. Half-round pilasters like those at Sompting (Ill. 192) first appear at St Maria im Kapitol, Cologne, where the Holy Cross altar was dedicated in 1049 followed by another dedication in 1065 (Gem 1973, ii , 494; Conant 1974, ill. 337); at St Remi, Rheims dedicated in 1049 (Gem 1973, ii , 494); and in the mid eleventh-century Abbot's chapel at Vendôme (loc. cit.). Soffit shafts used at Botolphs, Langford (no. 4), and Sompting (Ill. 183) occur first on the continent from the second quarter of the eleventh century onwards, as at Speyer in Germany of c. 1030-61 (ibid.; Conant 1974, ill. 90), and at Bernay, under construction between 1017–55 (Gem 1973, ii , 494; Gem 1983, 126; Musset 1974, pls. 1–4), at Le Mont St Michel in work of 1034–64 (Gem 1973, ii , 494), and at Jumièges, 1037–66 (ibid.; Conant 1974, ills. 357–8), all in Normandy. Soffit rolls, used in conjunction with the shafts, are an equally late feature occurring first in the crypts of Auxerre and Nevers cathedrals, dated to c. 1030 and 1029 respectively (Bony 1967, 75; Conant 1974, 244, ill. 112) and at St Martin, Tours, dated to c. 1050 (Bony 1967, 75; Conant 1974, 162, ill. 115). Angle shafts and rolls, used at Hadstock, make their earliest appearance on the continent in the mid eleventh century, for example, at Le Mont St Michel in work dated to 1048–58 (Bony 1967, 76).

The late dating of the continental analogues for the Romanesque features, which appear in churches of the third group, must place them in the middle of the eleventh century or later. This in turn has fired a debate over whether the churches of this group are of pre- or post-Conquest date. Anglo-Saxonists such as H. M. and J. Taylor, naturally, regard them as of pre-Conquest date, and scholars of the Romanesque treat them as of post-Conquest date (Bony 1967; Gem 1983, 128). In defence of the protagonists of a pre-Conquest date it is clear that Romanesque buildings were constructed in England before the Conquest, most notably at Westminster abbey which was begun c. 1050 (Gem 1980, 34). Also a single well-dated pre-Conquest building of Romanesque style survives, at Kirkdale, Yorkshire. Here the two-celled building has a west door and chancel arch with angle shafts, but is dated by inscription to 1055–65 (Taylor and Taylor 1965–78, i , 357–61, figs. 158–60, ii , pls. 504–6). It is arguable that the inscription, on a sundial over the south door, has been reused from an earlier building, but there is no evidence to support such a suggestion. Moreover, the form of the bases to the angle shafts is not one encountered elsewhere in demonstrably post-Conquest Romanesque architecture (D. A. Stocker, pers. comm.). The evidence, then, suggests that some of the churches of the third group may be of pre-Conquest date, although the elaborate forms of some of them, most notably at Langford, must place them firmly after the Conquest. Precise dating, however, is relatively unimportant in terms of the sculptures which adorn them. These are rather different in type and approach from those which decorate the earliest incontrovertibly Romanesque buildings in south-east England, such as the capitals in Canterbury cathedral crypt of c. 1075 (Zarnecki 1951b, 26, pl. 11). The sculptures adorning the third group of churches are thus arguably of pre-Conquest style or tradition, if not necessarily of pre-Conquest date.

Turning from the buildings to the sculptures which adorn them, it is noticeable that architectural sculpture is used more sparingly on churches of the first and second groups than on buildings of the third group. At Boarhunt, Hampshire, decoration is confined to the cabled ornament on the mid-wall slab of the north window of the chancel (Ills. 423–4). Cabled ornament is used in south-east England exclusively on sculptures of the late pre-Conquest period, as on the cross-shaft from London All Hallows by the Tower (no. 1), a fact which apparently confirms the primary dating evidence. However, the motif is so simple that dating by this feature alone is unsafe. At Breamore, Hampshire, the decoration is slightly more elaborate. Apart from the crucifixion (no. 1), discussed below, which is ex situ, again there is cabled ornament, this time on the impost blocks of the opening into the south porticus (nos. 2a–b). On the arch head is an inscription (no. 2c) and there are two other fragmentary inscriptions (nos. 3–4); unfortunately, these are difficult to date precisely (see below).

At Corhampton, Hampshire, the decoration is equally elaborate. Apart from the sundial discussed below, four of the pilasters probably once had foliage-decorated bases. Only one of the bases (no. 2) survives more or less intact, and this is decorated with a broad bulbous leaf, flanked by a pair of out-turned volutes (Ill. 440). This is an arrangement which occurs frequently in tenth- and eleventh-century art, particularly in manuscripts (Fig. 16a–i). It is employed in initials as on the initial D on fol. 118r of the Junius Psalter, a work of the second quarter of the tenth century (Temple 1976, no. 7, ill. 26), and on the initial B on fol. 4r of BL MS Harley 2904, a work dated to the last quarter of the tenth century (ibid., no. 41, ill. 141). The motif occurs also on the corners of frames to figural scenes, as on fol. 21r of the Aethelstan Psalter, dated to 924–39 (ibid., no. 5, ill. 33), or fol. 10r of Cambridge, Pembroke College MS 301, dated to c. 1020 (ibid., no. 73, ill. 233). In some cases the majority of the lush acanthus work making up the frame is composed of this motif endlessly repeated to form a simple bush scroll, as in the case of fol. 2v of the New Minster Foundation Charter, dated to after 966 (ibid., no. 16, ill. 84), and on fol. 2v of New York, Pierpont Morgan MS M. 709, dated to the second quarter of the eleventh century (ibid., no. 93, ill. 285). This three-leaved motif occurs also in ivory, as on an eleventh-century seal matrix from Lincoln, and in metal, for example on the lead-alloy disc brooches from the Cheapside hoard (Clark 1980, 14, pl. on 13), and on two disc brooches from 16–22 Coppergate, York (Tweddle 1982, 26–9, fig. p. 26; Hall 1984, ill. 60). On each of these brooches the design in the central field is made up from interlinked three-leaved motifs of the form used at Corhampton. The two Coppergate brooches derive from tenth-century contexts (Hall 1984, 60), but the Cheapside hoard can probably be dated to the eleventh century (Clark 1980, 14). These parallels for the Corhampton pilaster bases serve to underline the tenth- or eleventh-century date suggested by the primary evidence, but do little to refine it further.

On churches of the second group, sculpture is confined to the use of baluster shafts. At North Leigh these are plain with undecorated capitals and bases (Ills. 400–3). At Oxford (St Michael 1) there is a simple roll moulding encircling each baluster at the mid-point (Ills. 364–70). As Wright pointed out in his seminal article on Anglo-Saxon architecture (Wright 1844), such turned balusters are depicted in late Anglo-Saxon manuscripts, as in Aelfric's Pentateuch, on fols. 37v and 74r. This is a work of the second quarter of the eleventh century, probably made at St Augustine's abbey, Canterbury (Brown 1991, 52), a site which has yielded two ex situ balusters (Ills. 41–9) of very similar form to those depicted in the manuscript, discussed more fully below. An eleventh-century date would, therefore, be plausible for examples from North Leigh and Oxford.

Sculpture is more abundant on fabrics of the third group of churches, those with Romanesque architectural features, with 32 out of the 52 in situ architectural sculptures occurring in churches of this group. Of these 32 there are two sculptures at Botolphs, Sussex; five at Hadstock, Essex (no. 1), thirteen at Langford, Oxfordshire (no. 4); and twelve at Sompting, Sussex (nos. 14–23). In addition, the six sculptures on the south porticus arch at Hadstock (no. 2) clearly derive from the same building as the in situ doorway, although there is some argument over whether they have been re-arranged (Rodwell 1976, 62–3; Fernie 1983a). As noted above, the primary, architectural, evidence suggests that all of these sculptures should be dated to the middle or later part of the eleventh century, and limited support for this dating can be derived from the form and style of the individual sculptures.

At Sompting, Sussex, the sculpture consists of an external string-course (no. 15), capitals and flanking impost-like strips on the internal tower arch (no. 14), capitals on the half-round pilaster strips (nos. 17–19), two leaf-decorated belfry capitals (nos. 20–1) and two voluted bases (no. 22). The string-course is of complex and unparalleled form and cannot be independently dated, and the voluted bases are so simple that comparative dating is impossible. However, some additional dating evidence may be gleaned from the remaining sculptures. On the tower arch the layout of the decoration, with soffit shaft capitals flanked by impost-like strips (Ills. 183–91), is unusual and is paralleled in England only at Langford, Oxfordshire (no. 4; Ills. 297–312), and on the continent at Bernay in Normandy, where similar strips are found flanking a number of capitals in the south transept in work dated to 1017–55 (Gem 1973, ii , 495, 522; idem 1983, 126). As noted above, the capitals of the tower arch and of the pilaster strips are of debased composite or Corinthian form in which either the zone of upright leaves is suppressed leaving only the volutes, or the volutes are suppressed leaving only the zone of upright leaves. This particular type of debasement is paralleled elsewhere in England only in Lincolnshire where, at Bracebridge and Glentworth, there are volute capitals with the zone of upright leaves suppressed (Brown 1925, fig. 192, VII, XVI), and at Great Hale there are upright leaf capitals with the volutes suppressed (ibid., fig. 192, XII, XIV). At all three sites the capitals adorn belfry towers of the Lincolnshire type, for which a firm early post-Conquest date is increasingly being argued. A number of the towers, for example, use standard Romanesque capital forms, such as the cushion capital, while at Branston the tower has a Romanesque blind arcade around the base (Taylor and Taylor 1965–78, i , 93–4, ii , pl. 404). This is clearly primary despite the Taylor's valiant attempt to argue that it is secondary (ibid.; D. A. Stocker, pers. comm.).

The belfry capitals at Sompting (nos. 20–1) are of the corbelled form, and have each of the main faces decorated with two pairs of volutes, one above the other, linked at the base. The volutes take the form of long, narrow tendrils with clubbed or lobed ends (Ills. 205–12). This attenuated leaf type occurs elsewhere in Anglo-Saxon art in the latest phase of the Winchester style (Kendrick 1949, 102–3), for example in Rheims, Bibl. Mun. MS 9. fol. 23, dated to c. 1062 (Temple 1976, no. 105, ill. 299). Here on fol. 88r the corner rosettes of the frame have the acanthus drawn out to form long, attenuated strands with the tips curled over (ibid., ill. 299), and a similar treatment is evident in the upper right-hand corner rosette on fol. 1v of Oxford, Bodleian Library MS Tanner 3, dated to the second quarter of the eleventh century (ibid., no. 89, ill. 298). Similar attenuated tapering leaves with lobed ends are employed also in the Ringerike style, as, for example, on the grave-marker from St Paul's, London (Ill. 351). The Ringerike style was current in England from the late tenth century to the middle of the eleventh century (Wilson and Klindt-Jensen 1966, 145–6). Whether the Sompting capitals draw on one or both of these artistic strands, the parallels for the form of the leaves suggest a date towards the middle of the eleventh century, a date for the tower which, in broad terms, is also suggested by the form of the other capitals, and by the disposition of the decoration on the tower arch.

At Botolphs, Sussex, the only architectural sculptures are the capitals of the chancel arch (Ills. 4–5). They are upright leaf types related to those on the tower arch at Sompting (Ills. 186, 190). Again, they appear to be debased versions of the composite or Corinthian form (Ills. 4–5), and a similar date can be argued for them.

At Langford, Oxfordshire, the sculptured decoration consists of nine capitals to the archivolt and soffit roll shafts of the belfry windows of the central tower (no. 4), and a sundial on the south face of the tower (no. 3). The narrow capitals are continued by flanking impost-like strips (Ills. 297–312), a feature which, as noted above, is paralleled elsewhere only at Sompting, Sussex, and Bernay in Normandy. The decoration of both capitals and friezes consists of sprays of acanthus developing from the lower edge moulding; each spray is contained within a semicircular field. This type of decoration is widely paralleled in late pre-Conquest art as, for example, in a highly conventionalised form on fol. 4v of the Sherborne Pontifical, dated to the last quarter of the tenth century (Temple 1976, no. 35, ill. 13); in the upper border of the scene of the Miracle of the Tribute Money in the Gospel Lectionary fragment formerly in the Musée van Maerlant, Damme, Belgium, dated to c. 1000 (ibid., no. 53, ill. 176); on the lower border of the scene on fol. 171r of the Winchcombe Psalter, dated to c. 1030–50 (ibid., no. 80, ill. 253); and in the borders of fol. 2r of Aelfric's Pentateuch, a work of the second quarter of the eleventh century (ibid., no. 86, ill. 265). In metalwork a similar composition is employed on the inlaid silver plates around the base of the Canterbury censer cover, dated to the mid tenth century (Fig. 17c; Wilson 1964, no. 9, pls. XII–XIV; Backhouse et al. 1984, no. 73, pl. XXII). These parallels confirm the late pre-Conquest or early post-Conquest date suggested for the sculpture by the primary, architectural, evidence, but do nothing to advance the dating further.

The sundial at Langford (no. 3) is equally of little help in further refining the dating. The dial is set very high up and is not easy to read and it is possible that it has been reused from an earlier building. Even if the dial is contemporary with the tower, the decoration, which consists of two secular figures, half twisted and holding aloft the semicircular dial (Ill. 293), is not closely datable. Secular figures clad in knee-length garments and wearing cloaks are commonplace in manuscript art in both the tenth and eleventh centuries. They occur, for example, in the late tenth-century illustrations to Prudentius's Psychomachia in Cambridge, Corpus Christi College MS 23, as on fols. 2r and 37v (Temple 1976, no. 48, ills. 155–6), and in the Presentation scene in the Foundation Charter of the New Minster, Winchester, dated to after 966 (ibid., no. 16, ill. 84). In the eleventh century similar figures occur in the illustrations of the labours of the months in BL MS Cotton Tiberius B. V, dated to the second quarter of the eleventh century (ibid., no. 87, ills. 273–4). The best parallel for the posture of the figures at Langford, half-turned and looking upwards with their arms raised, is provided by the figure of king Edgar in the Presentation scene on fol. 2v of the Foundation Charter of the New Minster. However, the solid naturalistic style of the draperies at Langford is closer in style to eleventh-century works, such as BL MS Cotton Tiberius B. V, than to the fluttering, agitated draperies of the New Minster Foundation Charter.

The church at Hadstock, Essex, has two sculptured capitals, two impost blocks, and a hoodmould to the north doorway (no. 1; Ills. 277–84), as well as four sculptured capitals and two impost blocks on the arch into the south porticus (no. 2; Ills. 285–91), although, as noted above, the present south porticus arch may be a re-arrangement of older material (see above). In every case the decoration consists of palmette-like leaf ornament made up of interlocking, sub-triangular groups of radiating wedge-shaped leaves. This type of decoration is extremely difficult to parallel elsewhere, but something similar occurs in the upper right-hand border of the crucifixion scene in the Winchcombe Psalter, a work of c. 1030–50 (Temple 1976, no. 80). There is also comparable decoration on the mid twelfth-century eastern impost of the south doorway of Great Canfield church, Essex. Here the motif has developed into a series of rigidly interlocking triangles of palmette ornament, but the form and grouping of the leaves is close to that at Hadstock (Cobbett 1937, 45–6, pl. I). On the continent similar ornament occurs in conjunction with early Romanesque features on a capital in the crypt of Nevers cathedral, dated to c. 1030 (Bony 1967, 75). Presumably the Hadstock ornament lies in date somewhere between these various works, between c. 1030–50 and c. 1150, a date which fits comfortably with that provided by the primary evidence.

Apart from these two groups of churches which can be recognised as late Anglo-Saxon by reference to the criteria advanced by the Taylors, it is clear that parts of the fabrics at Walkern and Little Munden, Hertfordshire, must also be of pre-Conquest date. At Walkern the south arcade of the nave is of early twelfth-century date (Page 1912c, 224). This cuts through an earlier wall, reusing one jamb and an impost from an earlier south doorway (no. 2) and also leaving in situ a sculptured crucifixion (no. 1) which was positioned over the doorway (Taylor and Taylor 1965–78, ii , 628–30, figs. 318–19). The date of the Romanesque arcade suggests that the earlier wall with its south doorway and crucifixion might be of pre-Conquest date. The more exact placing of this fabric within the pre-Conquest period is more difficult, but the iconography and style of the crucifixion group, discussed below, argues for a tenth-, or more probably an eleventh-century date, and this can be used to date the cable-decorated impost surviving from the original south doorway. The impost at Walkern (no. 2; Ills. 398–9) is made up of four bands of cabling, each projecting beyond the one below. The twists of the rope are modelled, and along the middle of each strand is a V-shaped notch. This distinctive type of cabled decoration occurs on three other impost blocks from south-east England, two at Little Munden, and one at Dartford, Kent. At Little Munden, about five miles from Walkern, the western bay of the north nave arcade has jambs of square section which carry imposts of precisely the same form as those at Walkern (Ills. 318–19). It seems likely that they also are of eleventh-century date. At Dartford an impost on the north side of the late eleventh-century tower arch (Ill. 61) has identical cabling to that employed at Walkern and Little Munden, but the bands of cabling do not project. It is clear that this block must have been reused from an earlier fabric, as only the western face is decorated; the south face has been cut back flush with the jamb of the arch. Moreover, the block does not extend through the full thickness of the wall. Like the imposts from Walkern and Little Munden this appears to be a late pre-Conquest piece, probably of eleventh-century date.


In addition to the in situ architectural sculptures discussed above there is also a number of late pre-Conquest architectural sculptures which are now ex situ. These originally served a variety of purposes. At Sompting, Sussex, there are eight pieces from an internal string-course (nos. 1–8) and three frieze or blind arcade fragments (nos. 9–11). There are single fragments of screens from Sonning, Berkshire, and St Mary in Castro at Dover in Kent (no. 4); a gable cross from West Wittering, Sussex, and possibly another from Walberton, Sussex; and balusters or baluster fragments from St Augustine's, Canterbury, where there are two, St Mary in Castro, Dover, where there are five, and Jevington, Sussex, (no. 2) where there are two, and St Albans, Hertfordshire, where there are sixteen. The fragment from St Aldate's, Oxford, may also originally have been architectural. In addition, the fragmentary roundel from Abingdon, Berkshire, appears to have served an architectural function. The two fragmentary ex situ inscriptions from Breamore, Hampshire (nos. 3–4) and the ex situ impost block from Dartford, Kent, are discussed above. There are also sundials from Warnford and Hannington, Hampshire; St Maurice's and St Michael's, Winchester; Orpington, Kent; Marsh Baldon, Oxfordshire and Stoke D'Abernon, Surrey, discussed below.

The most elaborately decorated of these ex situ fragments are the lengths of string-course from Sompting (Fig. 18a–e). Six of these (nos. 1–5 and 7; Ills. 162–3, 167–8, 170–1) are decorated with foliage, each element of which consists of a long stem ending in a pair of out-turned leaves, with an axial bud or third leaf. The stalk is longitudinally grooved and the leaves are hollowed. The stems are angled and interlace where they cross. On two other fragments (nos. 6 and 8; Ills. 166, 164) the stems rise vertically from the border and do not interlace; some stems develop into five or six leaves. On both of these fragments the curved leaves from pairs of adjacent stems interlock to form semicircular fields, each containing a small semicircular-ended grooved leaf rising from the border.

These types of foliage can be closely paralleled in the acanthus-decorated borders of manuscripts in the Winchester style. The main leaf type with its long grooved stems and three-element tip can be paralleled in a number of manuscripts, most notably in the Arenberg Gospels of c. 990–1000, as on fols. 17v and 126v, where some of the leaves have precisely the same form as those at Sompting (Temple 1976, no. 56, ill. 169; Backhouse et al. 1984, no. 47, pl. XI). Similar leaves occur in the borders of the St Mark portrait added c. 1000 to BL MS Royal I. E. VI (Temple 1976, no. 55, ill. 172), and in the corner rosettes on fol. 1v of Oxford, Bodleian Library MS Tanner 3, dated to the second quarter of the eleventh century (ibid., no. 89, ill. 298). Here the stalks are attenuated and angled so that they cross over each other, a feature characteristic of the plant ornament at Sompting. The combination of this leaf type with interlacing stems occurs also in the Arenberg Gospels (ibid., no. 56, ill. 169; Backhouse et al. 1984, pl. XI), and angled, interlocking stems with different forms of acanthus foliage are also abundant in manuscript art as, for example, on fol. 53r of BL MS Arundel 155, dated to 1012–23 (Temple 1976, no. 66, ill. 217); on fol. 11v of BL MS Cotton Vitellius C. III, dated to the early eleventh century (ibid., no. 63, ill. 187); and on fol. 62r of Rome, Vatican Biblioteca Apostolica MS Reg. lat. 12, dating to the second quarter of the eleventh century (ibid., no. 84, pl. 262).

Both the leaf type at Sompting, with its three-element tip, and the interlocking of the stems can also be paralleled in works in other materials in late Anglo-Saxon art. In metalwork three-element leaves are commonly encountered, as in the centres of the disc brooches from the Cheapside hoard, London (Clark 1980, 14, pl. p. 13) and from 16–22 Coppergate, York (Tweddle 1982, fig. p. 26; Hall 1984, ill. 60). The use of interlocking stems is more unusual, but is paralleled in ivory on the early eleventh-century Tau-shaped crosier from Alcester, Warwickshire (Beckwith 1972, no. 29, pls. 65–6). Here some of the acanthus sprays have long grooved stems and three-element tips, but the length of the stem varies greatly and the acanthus is inhabited. Apart from these differences the parallel with the Sompting frieze is close.

The parallels which can be drawn between the main leaf type at Sompting and these Anglo-Saxon works, principally of eleventh-century date, might suggest a similar date for the majority of the Sompting frieze fragments. However, both this leaf type and the use of interlocking stems persisted into the post-Conquest period and were assimilated into Romanesque art. At Canterbury one of the crypt capitals, dated to c. 1120, employs leaves very similar in form to those at Sompting, although the organisation of the leaves, to fill the faces of a cushion capital, is very different (Zarnecki 1955, 215, pl. 161). At Steyning on a capital in the south aisle, also dated to c. 1120, the foliate stems have voluted, not tripartite, ends, but interlock and interlace in a manner evocative of the arrangement of the stems at Sompting (ibid., 212, pl. 156). The foliate decoration at Sompting must lie in date between the pre-Conquest works of the first half of the eleventh century and these Romanesque capitals of the early twelfth century. The parallels with the Anglo-Saxon works are more numerous and more convincing than with the Romanesque, and it is probable that the Sompting pieces lie closest in date to them, perhaps towards the middle of the eleventh century.

The remaining two fragments of string-course from Sompting (nos. 6 and 8) have leaf decoration which is of the same general character as that on the other six pieces, if coarser and more formalised (Ills. 166, 164). From their dimensions and material it is clear that they originally formed part of the same architectural feature as the other five fragments, and must, therefore, be of the same date. Independent confirmation of this derives from the form of the leaf decoration, as the leaves form semicircular fields each containing a short vertical stem with a semicircular tip. This very closely parallels the decoration of the belfry windows at Langford, Oxfordshire (no. 4; Ills. 297–312), for which a mid eleventh-century date is argued above on primary evidence. As noted above, this form of decoration is employed on a number of late tenth- and eleventh-century manuscripts, and in metalwork on the mid tenth-century censer cover from Canterbury.

The three frieze or blind arcade fragments from Sompting (nos. 9–11) are decorated with reeded arch heads whose tympana and spandrels are filled with plant ornament (Ills. 172–80). Some of the plant stems are held together with collars. The leaves are lozenge shaped, or take the form of short tight volutes and in addition there are small trefoil leaves or berry bunches. One of the plant stems ends in an animal head with a domed forehead, open mouth with short jaws, and a reversed lentoid eye. In the reconstruction of the frieze proposed above (Fig. 19a–c) the arches are viewed as the heads of niches or decorated panels, or as a blind arcade, a layout which is reflected most closely in manuscript Canon Tables, which often employ series of arch heads supported by piers or columns. Surviving pre-Conquest Canon Tables range in date from the eighth to the eleventh century, but only one set, in the Trinity Gospels, dated to the second quarter of the eleventh century (Temple 1976, no. 65), offers a number of close parallels to the features of the Sompting arch heads (Fig. 19d). Like them it employs plant ornament filling the tympana or spandrels. The foliage is usually Winchester-style acanthus, but on occasions it is simplified to single volutes or pairs of tight volutes, as at Sompting. In addition there are cinquefoil berry bunches. The most precise parallel to the layout and treatment of the ornament on the Sompting fragments occurs on fol. 11r, where the spandrels are filled with pairs of out-turned plant sprays whose stems cross beneath the arch heads from the adjacent pairs of tympana and are held together by collars. The sprays are of acanthus, but the edges of the leaves are voluted to produce an effect very similar to that employed at Sompting (Fuglesang 1980, pl. 85A). Outside the pages of the Trinity Gospels, the only close parallel to the decoration on the Sompting friezes is provided by Warsaw, Bibl. Narodowa MS I. 3311 (Fig. 20a), a manuscript dating to c. 1000 (Temple 1976, no. 92, figs. 51–5, 59, ills. 281–4). On fol. 69r (ibid., ill. 282) the initial A is decorated with plant ornament having a number of features in common with the Sompting friezes, including the use of tightly voluted and trefoil leaves, the use of stems which are leafless for most of their length, and the use of stems ending in animal heads. As at Sompting the animal heads are small, barely wider than the stem of which they form a part. They have domed foreheads, open jaws and lentoid eyes, although not reversed as at Sompting.

Neither of these two manuscripts provides an exact parallel for the decoration of Sompting 9–11, although some of the individual features compare closely, but together they suggest a date in the eleventh century for the sculptures. This is reinforced by one distinctive individual feature, the use of the reversed lentoid eye on the single animal head (Ill. 179). This form of eye is widely used in Viking art, particularly in the Ringerike and Urnes styles. The Ringerike style was current in England from the late tenth century to the mid eleventh century (Wilson and Klindt-Jensen 1966, 145–6). The Urnes style was probably introduced after the Conquest, and persisted until the end of the eleventh century, and conceivably even later (ibid., 160).

There are two possible late pre-Conquest screen fragments from south-east England, from St Mary in Castro, Dover, Kent (no. 4), and Sonning, Berkshire. The fragment from Dover is decorated on face C with the junction of two reeded arch heads, with in the spandrel a series of stems held together by a collar; one of the stems ends in a tight volute (Ill. 70). The decoration very closely resembles that on the frieze fragments from Sompting, and must be given a similar, eleventh-century, date. This is reinforced by a consideration of the decoration on face A, which consists of a cross with parallel-sided arms and simple interlace in the re-entrant angles (Ill. 69). This type of decoration is very similar to that employed on a number of late tenth- and eleventh-century grave-covers, such as those from Cambridge castle (Fox 1920–1, 20–1, pls. III–IV, VII), and Cardington, Bedfordshire (Ill. 264), and suggests a similar date for the Dover fragment.

The layout of the decoration on face A of the Dover slab (Ill. 69) is echoed on the Sonning fragment, which is also decorated with a cross having interlace in each of the re-entrant angles (Fig. 37). In this case the cross is of Anglian form and the interlace is more complex, consisting of ring-knots each of which incorporates two closed loops (Ill. 454). The use of closed-circuit loops suggests a late date for the piece as this feature does not normally occur before the tenth century (Collingwood 1927, 65, 68). This late date may be confirmed by the uneven laying-out of the decoration, a feature which is again more characteristic of late Anglo-Saxon interlaces than of earlier material; the interlaces from Reculver 1e (Ills. 119–20), Preston by Faversham (Ills. 101–4), and Rochester (no. 1; Ill. 141) in Kent, for example, are all well worked out, and an early date can be suggested for all of them. Such evidence can, however, only be regarded as tenuous and unreliable.

Also of architectural function are the ex situ balusters and baluster fragments from St Albans (Ills. 376–96), St Augustine's, Canterbury (nos. 6–7; Ills. 41–9), St Mary in Castro, Dover (nos. 2–3; Ills. 64–7, 71–5), and Jevington (no. 2; Ills. 229–30). These balusters are of the type decorated with groups of projecting mouldings, and associated with belfry openings. in situ examples are encountered within the region at Oxford (St Michael 1) and North Leigh, Oxfordshire (Ills. 400–3). Such openings appear to be a late Anglo-Saxon phenomenon; Taylor and Taylor place the tower at North Leigh in the late tenth or eleventh century (Taylor and Taylor 1965–78, ii , 464–5), and that of St Michael at Oxford in the late eleventh century (ibid., 481–2).

At Abingdon, Berkshire, is a fragmentary roundel which may have served an architectural function (Ill. 413). The decoration can be reconstructed as a Greek cross with slightly pointed ends to the arms and an interlace triquetra in each of the re-entrant angles (Fig. 36). The decoration has obvious affinities with that on face A of the slab from St Mary in Castro, Dover (no. 4; Ill. 69) and on the fragment from Sonning, Berkshire (Ill. 454), and these parallels suggest a late pre-Conquest date, as does the use of rather thick median-incised strands; as noted elsewhere, there is no example of a median-incised interlace from south-east England dating to before the tenth century. This late dating is confirmed by the form of the piece; the disc decorated with a cross has obvious affinities with the plate-headed crosses for which a tenth- and eleventh-century date is argued below. The form of the cross at Abingdon, with its pointed ends, is paralleled elsewhere in south-eastern England only on the grave-marker or -cover from Arundel, Sussex (Ill. 1; Fig. 28), for which a tenth- or eleventh-century date is argued below, and in ivory on a pectoral cross in the Victoria and Albert Museum, for which a mid eleventh-century date has been argued (Beckwith 1972, no. 45, pls. 99–102; Backhouse et al. 1984, no. 125, pl. 125). The only parallel to the interlace decoration on the Abingdon roundel is provided on fol. 2v of Cambridge, Pembroke College MS 301, dated to c. 1020 (Temple 1976, no. 73, ill. 233). Here, on the top of pier three of the Canon Table is a roundel containing four interlace triquetras arranged in the same formation employed at Abingdon, but without the cross. All these parallels argue for a tenth- or eleventh-century date for the Abingdon fragment, but none is compelling.

The gable cross at West Wittering, Sussex (Ills. 217–18), is decorated on both faces with a simple incised Greek cross within a circle, a form related to that used at Abingdon (Fig. 36), and to the ring-, circle-, and plate-heads discussed in more detail below. On this basis a tenth- or eleventh-century date for the piece can be argued.


Apart from the crucifixion scenes and figural panels discussed below, the only other architectural sculptures from south-east England are sundials. There are nine of them (see Ch. III). The dial from Langford, Oxfordshire, is discussed separately above. All, with the exception of the dials from Corhampton and Hannington in Hampshire (and perhaps also that from Langford in Oxfordshire), are ex situ.

The dials from Warnford, Hampshire, and St Michael's, Winchester, must be grouped with the in situ example from Corhampton, Hampshire, as they share the same form, a circular dial sculpted on a square stone; the same calibration; and very similar decoration, consisting of a stem ending in three leaves in each corner of the slab on which the dial is carved. On primary evidence the dial from Corhampton can be placed in the eleventh century, and a late pre-Conquest date is confirmed by the form of the leaf ornament (Ill. 438) which, as noted above, occurs widely in manuscripts, ivory and metalwork of the tenth and eleventh centuries. This dating can be extended to the other dials of the group. Of similar date must be the dial from Hannington. Although apparently sculpted on a circular stone (Ill. 441), the calibration matches that of the other Hampshire dials. Architecturally, what remains of the Anglo-Saxon church at Hannington suggests a late tenth- or eleventh-century date for the fabric (Taylor and Taylor 1965–78, iii , 282–3), and, by extension, for the dial.

At Orpington, Kent, the dial is carved from a single circular stone (Ills. 105–7) and can be identified as of pre-Conquest date because of the Old English inscription around the edge, and the runic characters on the face of the dial. Unfortunately the epigraphy of the inscription is of little help in dating (see the catalogue entry), and the only decorative feature on which dating can be based is the cabled moulding encircling the dial. This has the cable modelled with an incised median line along each strand. The form is paralleled elsewhere in south-east England at Dartford, Kent, Walkern 2 and Little Munden, Hertfordshire, and All Hallows by the Tower, London, all probably works of the eleventh century. A similar date can be suggested for the Orpington dial.

The dial at Marsh Baldon, Oxfordshire (Ill. 358), like those from Corhampton, Hampshire (no. 1), Warnford, and St Michael's, Winchester, is sculpted on a square stone, and, like Orpington, has a cabled edge moulding, but without the incised median line of the latter. The gnomon hole is displaced towards the upper edge of the dial, but the calibration is of pre-Conquest type with the crossed lines representing the mid tides at 9 am, 12 noon, and 3 pm. Dating of the dial is difficult, but the use of a cabled moulding may suggest a late pre-Conquest date as there is no occurrence of this type of moulding in south-east England before the tenth century.

The sundials at Stoke D'Abernon, Surrey, and Winchester St Maurice 2, are even more difficult to date as they lack any form of decoration. The Stoke D'Abernon dial (Ill. 216) was carved on a square block, and that from St Maurice's (Ill. 672) on a circular stone, both types encountered among the demonstrably pre-Conquest dials. Both are calibrated in the Anglo-Saxon manner, again indicating a pre-Conquest date. Given the paucity of early pre-Conquest dials from south-east England (there is only one, from Bishopstone, Sussex) it may seem more likely that the St Maurice's and Stoke D'Abernon dials are of late pre-Conquest date, but the suggestion is not susceptible of proof.


Of all the architectural sculptures from south-east England the most splendid and imposing are the large-scale crucifixion groups. Six of these survive. Three are from sites in Hampshire, at Breamore (no. 1; Ills. 425–8), Headbourne Worthy (no. 1; Ills. 448–50), and Romsey (no. 1; Ills. 451–2); the others are at Walkern (no. 1) in Hertfordshire, and the two from Langford, Oxfordshire (nos. 1–2). All except the examples from Headbourne Worthy and Walkern are ex situ. A pre-Conquest date has also been suggested for the two smaller-scale crucifixions built into the west tower at New Alresford, Hampshire (Green and Green 1951, 40–1, pl. XII), but these have been excluded from consideration here as the disturbed walling around them suggests that they are later insertions into the fabric of the twelfth-century tower. This conclusion is supported by the small-scale, plastic, three-dimensional nature of the sculpture which is more appropriate to the thirteenth or fourteenth centuries, and by the chamfers visible along the edges of the cross-arms. This is a feature not encountered on the pre-Conquest examples, but which is typical of later medieval crosses, such as that in the churchyard at Terwick, Sussex. Given their relatively small size, it is probable that both crosses are reused later medieval gable crosses.

Despite the doubts raised over the pre-Conquest dating of the examples from Langford and Romsey by Clapham (Clapham 1951), there is clear primary evidence for the use of such large-scale crucifixion scenes in the pre-Conquest period. The scene at Headbourne Worthy, Hampshire (no. 1), is in situ above the west doorway of the pre-Conquest church. The Hand of God above the cross (Ill. 448) forms an integral part of the string-course spanning the west gable, and the foot of the cross forms one of the voussoirs of the pre-Conquest west doorway (Taylor and Taylor 1965–78, i , 290, fig. 128). The crucifixion at Walkern 1, Hertfordshire, is also in situ, over the original south doorway of the pre-Conquest church (see above). Outside the area of the present study parts of other similar scenes also remain in situ. At Bitton, Gloucestershire, the feet of a crucifixion are in situ over the chancel arch, although the rest of the scene has vanished (Taylor and Taylor 1966, 6–8, fig. 2), and two angels at Bradford on Avon, Wiltshire, discovered in situ over the chancel arch may also represent a lost crucifixion scene in this position (ibid., 29), although other scenes than the crucifixion are possible. The Anglo-Saxon wall painting at Nether Wallop, Hampshire, is also positioned over the chancel arch, but depicts the Ascension with Christ's mandorla supported by very similar angels (Gem and Tudor-Craig 1981). This accumulation of primary evidence for the use of large-scale crucifixion scenes in the pre-Conquest period suggests that the similar ex situ scenes at Romsey 1 and Langford 2 may also be regarded as of pre-Conquest date. In contrast, from the greater volume of architectural sculpture surviving from after the Conquest only a single scene of this type is known, from Barking abbey, Essex, a piece which can be dated to the twelfth century (Kendrick 1949, 53–4). The Barking crucifixion scene differs in style from the crucifixions for which a pre-Conquest date can be advanced based upon primary evidence, and it also differs radically in constructional technique. It is composed of nine fairly small, coursed stones on which the sculpture has been carved. In contrast, at Headbourne Worthy 1 the figure of Christ is composed of three massive stones, one forming the vertical limb of the cross and the other two the arms. The figures of the Virgin and St John are also each carved from a single massive block of stone (Ills. 448–50). A similar constructional technique is used at Walkern 1, although here any accompanying figures have been lost. The ex situ crucifixions all share the constructional technique of the Headbourne Worthy and Walkern pieces, although on the larger of the two crucifixions at Langford (no. 2; Ills. 294–5) the vertical limb of the cross is composed of two separate stones. Each of the arms is made from a single massive stone and tenons on their inner ends fit into a gap between the two stones making up the vertical limb. The use of tenoned ends to the arms of the cross can only be paralleled in the scene of the crucifixion of St Peter over the north-west portal of the twelfth-century church at Aulnay, France (Porter 1923, 7, 983). But here, as at Barking, the rest of the scene is composed of small, rectangular, coursed stone blocks, in contrast to the monolithic construction used at Langford.

One distinctive iconographic feature also suggests a pre-Conquest date for the crucifixions. Four of the six, at Walkern 1, Romsey 1, Headbourne Worthy 1, and Breamore 1, employ the Hand of God developing from a cloud above the head of Christ (Ills. 397, 451–2, 448, 427). The use of this feature is a strong indicator of a pre-Conquest date as it was commonly used in pre-Conquest crucifixion scenes, but only comparatively rarely after the Conquest. Of the fourteen pre-Conquest ivory crucifixion scenes catalogued by Beckwith, eleven employ this feature (Beckwith 1972, nos. 17, 17a, 30–8). Similarly, of the eleven Anglo-Saxon manuscript crucifixion scenes listed by Ohlgren, seven employ the Hand of God (Ohlgren 1983, nos. 140, 161, 182, 185, 187, 189, 198); it is difficult to find a single Romanesque manuscript illustration of the crucifixion which employs this feature, and of the seven Romanesque ivory crucifixions catalogued by Beckwith, only two employ the Hand of God (Beckwith 1972, nos. 78, 108). The number of surviving works is too small for compelling conclusions to be drawn, but it appears that the use of the Hand of God is a possible indicator of pre-Conquest date. The primary evidence, that of the constructional technique, and the use of the Hand of God, all suggest that the ex situ roods at Breamore 1, Romsey 1 and Langford 1–2 are of pre-Conquest date. This conclusion can be supported by comparisons in iconography and style between the stone sculptures and crucifixion scenes in other materials.

The iconographical study of the crucifixions is rendered difficult by the poor condition of most of the sculptures. Those at Breamore 1 and Headbourne Worthy 1 have been hacked back flush with the walls into which they are built. The crucifixions at Romsey and the larger of the two examples from Langford (no. 2) are ex situ, and it is likely that subsidiary figures have been lost. Only the smaller crucifixion at Langford (no. 1) appears to be intact. Despite these difficulties, as noted above, three iconographical groups can be identified, coinciding (with one minor variation) with Coatsworth's classification (Coatsworth 1988, n. 29). The crucifixions from Walkern (no. 1; Ill. 397) and the larger of the two from Langford (no. 2; Ills. 294–5) depict Christ as clad in a sleeved, ankle-length robe. The smaller of the two Langford crucifixions (no. 1; Ill. 293) and that from Breamore 1 (Ill. 428) depict the dead Christ with his arms and legs contorted. The examples from Romsey 1 (Ills. 451–2) and Headbourne Worthy 1 (Ill. 448) depict Christ as in repose, with the legs and arms straight or only slightly bent, and his head facing straight ahead or only slightly bent to the left.

To the robed crucifixions at Walkern 1 and Langford 2 can be added the example from Bitton, Gloucestershire, outside the area of the present study. This shares with the Langford crucifixion the use of transverse bars behind the hands of the crucified Christ, but in addition has a serpent below his feet. Such a feature may have existed at Langford as this area has been cut away. Although there is clear primary evidence for a late tenth- or eleventh-century date for the Walkern and Bitton crucifixions, closer dating is hampered by the fact that only two other Anglo-Saxon robed crucifixions survive in any medium (both of them much earlier than the fabrics at Walkern and Bitton), in Durham Cathedral Library MS A. II. 17, dated to the late seventh or early eighth century (Alexander 1978, no. 10, ills. 47, 202) and on the putative eighth-century ivory diptych now in the Musée de Cluny, Paris (Beckwith 1972, no. 6, pl. 19).

Because of this paucity of surviving Anglo-Saxon robed crucifixions, undue emphasis has been placed on the Liber Albus of Bury St Edmunds which records the presence there of a robed crucifixion, a copy of the Volto Santo of Lucca, ordered by Abbot Leofstan on his return from Rome in 1049/50 (Clapham 1951; Raw 1990, 92; Dodwell 1982, 318). The present Volto Santo (probably a thirteenth-century copy of the original which Leofstan saw) depicts a robed Christ. Hence it has been argued that the introduction of the robed crucifixion into English sculpture must be dated to after 1050. In fact there need be no connection between the Bury St Edmunds crucifixion and those at Langford 2, Walkern 1, and Bitton, and this reference to the Volto Santo is of importance only in that it demonstrates that examples of robed crucifixions were being made in England in the eleventh century, copying continental exemplars.

During the late tenth and eleventh centuries the use of the robed crucifixion, predominant from the sixth to the ninth century, but not common thereafter (Coatsworth 1979, i , 108), enjoyed a modest revival, at least on the continent. The type occurs, for example, in manuscript art in the Egbert Codex, a work of c. 980 (Holländer 1974, pl. 104), in Paris, Bibl. Nat. MS lat. 9453, a work of the second half of the tenth century (Hubert et al. 1969, pl. 144), and in the Uta Gospels of 1000–1025 (Dodwell 1971, pl. 84). Robed crucifixions occur also in ivory, as on a late tenth-century plaque in the Rheinisches Landesmuseum, Bonn (Goldschmidt 1914–23, ii , no. 37, pl. 37), and on an eleventh-century book cover in the Musée des Arts decoratifs, Brussels (ibid., no. 55, pl. 55). Similar, readily portable objects may have been copied in England in the late tenth or eleventh centuries. A continental derivation for the robed rood is supported by the occurrence of a serpent below the feet of Christ at Bitton, Gloucestershire. This feature occurs elsewhere only twice in Anglo-Saxon art, in Cambridge, Corpus Christi College MS 421, p. 1, a manuscript made in the west country in the second quarter of the eleventh century (Temple 1976, no. 82, ill. 254), and on fol. 35r of the Bury Psalter, made at the same time at Christ Church, Canterbury (ibid., no. 84, fig. 26, ills. 262–4). However, the serpent beneath the feet of Christ occurs frequently in Carolingian art as, for example, on the ivory on the front cover of the Psalter of Charles the Bald of c. 850–60 (Lasko 1972, pl. 30), on an ivory panel of c. 870 in the Victoria and Albert Museum (ibid., pl. 61), and on a crystal from the abbey of St Denis made c. 867 (ibid., pl. 56). It occurs also in Carolingian manuscript art, as on fol. 6v of a sacramentary from Metz (Mütherich and Gaehde 1977, pl. 34).

To the second group of crucifixions, those depicting the dead Christ, belong the examples at Breamore 1 (Ill. 428) and Langford 1 (Ill. 293). Like the robed crucifixions this group is well defined, the weight of Christ's sagging body being carried on the arms which slope down to the shoulders. The legs are bent and the head bowed. As noted above, there is no primary evidence for the dating of either of these crucifixions, but the use of the Hand of God at Breamore suggests a pre-Conquest date. At both Breamore 1 (Ill. 425) and Langford 1 (Ill. 293) the crucified Christ is accompanied by the figures of the Virgin and St John, as is the case in nine out of the eleven pre-Conquest manuscript crucifixions catalogued by Ohlgren (Ohlgren 1983, nos. 140, 146, 161, 177, 182, 185, 187, 198, 208), but in only four out of the fourteen ivory crucifixions catalogued by Beckwith (Beckwith 1972, nos. 17, 17a, 30, 38). At Breamore, disks over the heads of the Virgin and St John, at the ends of the arms of the cross (Ill. 425), probably represent the sun and moon, symbols often associated with the crowned Christ (Raw 1990, 129). These occur on five out of the ten manuscript crucifixions, always in association with the figures of the Virgin and St John (Ohlgren 1983, nos. 177, 182, 185, 198, 208), but on only three of the ivory crucifixions, in two cases with the Virgin and St John (Beckwith 1972, nos. 30, 34, 38). It is possible that the disks at Breamore have been moved from their original positions over the cross-arms, as there is sound evidence for the whole scene having been brought from elsewhere and inserted over the south door (Rodwell and Rouse 1984, 309–12). In all the manuscript scenes and on most of the ivories the sun and moon occupy positions over the cross-arms, but on one, the Heribert crosier (ibid., no. 30, pl. 80), the figures are at the ends of the cross-arms, an unusual arrangement paralleled otherwise only at Breamore 1. If it could be proved that at Breamore the sun and moon have not been misplaced then this feature might place the sculpture at about the same date as the crosier, in the early eleventh century.

The type of Christ used at both Breamore 1 (Ill. 428) and Langford 1 (Ill. 293) has both the arms and legs bent. This feature occurs only three times in pre-Conquest manuscript art: in the Arenberg Gospels, dated to c. 990–1000 (Temple 1976, no. 56, ill. 171); in the Judith of Flanders Gospels, dated to c. 1025–50 (ibid., no. 93, ill. 298); and in BL MS Cotton Tiberius C. VI, dated to c. 1050 (ibid., no. 98, ill. 311). This type is also found in an addition of c. 1080 to BL MS Arundel 60 (Kendrick 1949, pl. XXI.2). The Arenberg Gospels and the Judith of Flanders Gospels also employ the Virgin and St John, but only in the Judith of Flanders Gospels do the figures of the sun and moon appear as at Breamore. In BL MS Cotton Tiberius C. VI the Virgin and St John are replaced by Stephaton and Longinus, and there is a titulus. In BL MS Arundel 60 only two bushes flank the figure of Christ. Iconographically, the Breamore crucifixion, therefore, most closely corresponds to the scene in the Judith of Flanders Gospels. The Langford 1 crucifixion, lacking any other accompanying figures than the Virgin and St John, finds only one close parallel in the manuscripts, in BL MS Harley 2904, dating to c. 975–1000, where Christ is also accompanied only by the figures of the Virgin and St John, but there is also a titulus, absent at Langford (Temple 1976, no. 41, ill. 142). In BL MS Harley 2904, however, although Christ is clearly depicted as dead with the head bowed and arms bent, the legs are straight.

These manuscript parallels for the iconography at Breamore 1 and Langford 1 serve to place the crucifixions in the late tenth or eleventh centuries. A consideration of the ivories of the period does little to clarify this dating. Only two, a crucifixion plaque in the Victoria and Albert Museum dated to the late tenth or early eleventh century (Beckwith 1972, no. 32, pl. 68), and the Tau-shaped crosier from Alcester, dated to the early eleventh century (ibid., no. 29, pl. 66), use a figure of Christ with the legs bent. On the Alcester crosier there are no accompanying figures. The Victoria and Albert Museum plaque has roundels containing the symbols of the Evangelists in the re-entrant angles of the cross.

To the third iconographical group of crucifixions, those having Christ's legs and arms straight and his head slightly bowed to the left, belong the examples at Headbourne Worthy 1 (Ill. 448) and Romsey 1 (Ills. 451–2), Hampshire. As noted above, primary dating evidence places the Headbourne Worthy crucifixion in the late tenth or eleventh centuries, and a pre-Conquest date may be confirmed by the use of the Hand of God, employed also at Romsey. At Headbourne Worthy the figure of Christ is accompanied by the Virgin and St John. As noted above, this feature occurs in nine out of the eleven pre-Conquest manuscript crucifixion scenes catalogued by Ohlgren, and on four of the fourteen ivories of this date catalogued by Beckwith. An area of chiselling below the feet of Christ suggests that something has been cut away, probably either a chalice or a serpent, the only two symbols which occupy this position in pre-Conquest art. Of the two the serpent, symbolising Christ's triumph over evil (Raw 1990, 147), occurs in sculpture at Bitton, and in manuscript art in Cambridge, Corpus Christi College MS 421, dated to c. 1025–50 (Temple 1976, no. 82, ill. 254), and the Bury St Edmunds Psalter which is of similar date (ibid., no. 84). In Corpus Christi MS 421, as at Headbourne Worthy, the figures of the Virgin and St John accompany the crucified Christ. The chalice at the foot of the cross, a eucharistic symbol (Raw 1990, 147), occurs in combination with the serpent in the Bury St Edmunds Psalter (ibid., no. 84), and by itself in the Sherborne Pontifical of c. 975–1000 (ibid., no. 35, ill. 134), and in the Arenberg Gospels of c. 990–1000 (ibid., no. 56, ill. 171), in the two latter cases in combination with the Virgin and St John, however, both scenes have angels above the cross-arms and these are absent at Headbourne Worthy. These iconographical parallels suggest an eleventh-century date for the Headbourne Worthy crucifixion scene, but none is exact.

Little is added to the discussion by a consideration of the form of the figure of Christ. As noted above, both the legs are straight, a feature which occurs in at least four out of the eleven manuscript crucifixion scenes catalogued by Ohlgren (Ohlgren 1983, nos. 140, 177, 182, 187). It occurs also in ivory in nine out of the fourteen crucifixion scenes catalogued by Beckwith (Beckwith 1972, nos. 6, 17, 17a, 20, 31, 33, 35, 37-8). The vicious damage sustained by the Headbourne Worthy crucifixion renders closer comparisons impossible.

At Romsey 1, in contrast, all that survives is the figure of Christ, any accompanying figures having been lost when the figure was incorporated into the twelfth-century fabric of Romsey abbey (Ills. 451–2). Here, as at Headbourne Worthy 1 (Ill. 448), Christ's legs are straight and his arms are outstretched. His head is slightly bowed to the left, but his eyes are open. As noted above, this pose occurs in at least four pre-Conquest manuscript crucifixion scenes, but only in two, BL MS Cotton Titus D. XXVII (Temple 1976, no. 77, ill. 246) and the Sacramentary of Robert of Jumièges (ibid., no. 72; Rice 1952, pl. 53b), are Christ's eyes open. The Sacramentary of Robert of Jumièges dates to c. 1020 (ibid., 89–90) and Cotton Titus D. XXVII to c. 1025–35 (ibid., 94–5). The open-eyed Christ of this form is more commonly found in ivory, and at least five out of the fourteen ivory crucifixions catalogued by Beckwith exhibit this feature, all of them of late tenth- or eleventh-century date (Beckwith 1972, nos. 17, 17a, 31, 33, 35).

One peculiar feature of the Romsey crucifixion is the flattening of the top of the head and the suppression of the vertical limb of the cross of Christ's cruciform nimbus (Ill. 456). This is not accidental damage and suggests that something, probably in another material, was originally placed on top of Christ's head; something tall enough to obscure the vertical limb of the cruciform nimbus, rendering its carving superfluous. This feature can only have been a fairly tall crown, probably of metal, and emphasising Christ's kingship (Raw 1990, 129–46). This iconographic feature is only paralleled in Anglo-Saxon art on fol. 12v of BL MS Arundel 60, dating to c. 1060 (Temple 1976, no. 103, ill. 312), on fol. 13r of BL MS Cotton Tiberius C. VI, dating to c. 1050 (ibid., no. 98, ill. 311), and on p. 1 of Cambridge, Corpus Christi College MS 421, dating to the second quarter of the eleventh century (ibid., no. 82, ill. 254). It also occurs, if infrequently, in works produced on the continent, as in the Uta Gospels of the first quarter of the eleventh century and in a Psalter executed at St Germain des Prés, Paris, in the middle of the eleventh century. The Christ in the Uta Gospels is robed but open-eyed (Dodwell 1971, pl. 84); the Christ in the St Germain des Prés manuscript, as at Romsey and in the English manuscripts, is clad in a loin-cloth, has the legs and arms straight out, and is open-eyed. The figure is, however, accompanied by the Virgin and St John and the sun and moon, with a serpent at the base of the cross. The scene is surrounded by the symbols of the four Evangelists (Beckwith 1969, pl. 172).

That similar crucifixion scenes with a crowned Christ existed in other materials in Anglo-Saxon England, at least in the eleventh century, is confirmed by documentary sources. At Waltham abbey, Essex, a stone crucifixion was given a gold and jewelled crown by Glitha, wife of Earl Tovi the Proud (d. soon after 1042) (Dodwell 1982, 119, n. 145), and at Winchester king Cnut (1016–35) donated his own crown to the Old Minster; it was placed on or above the head of the crucified Christ (Dodwell 1982, 212, n. 227). The Anglo-Saxon Chronicle (version E) records that in 1070 'Hereward and his band' looted the monastery at Peterborough, even taking the gold diadem from the head of the crucified Christ (ibid., 212–13; Garmonsway 1967, 205), an incident which is also reported in the Peterborough Chronicle of Hugh Candidus (Dodwell 1982, n. 238).

The written sources and continental manuscript parallels may suggest an eleventh-century date for the Romsey crucifixion; it is possible, however, that the flattening of the top of the head of Christ is a twelfth-century re-working, done at the time when the crucifixion was reused in its present position, c. 1150–90. Numerous twelfth-century crucifixions survive which employ the crowned Christ, as, for example, on the Monmouth crucifix of c. 1170–80 (Zarnecki et al. 1984, no. 241, pl. 241) and two cast crucifix figures in the University Museum of Archaeology and Anthropology, Cambridge, both dating to the second quarter of the twelfth century (ibid., nos. 236–7, pls. 236–7).

Comparatively little is added to a discussion of the dating of most of these crucifixion scenes by a discussion of their style, particularly since the Headbourne Worthy 1 and Breamore 1 crucifixions are virtually destroyed, and the remaining examples, except that at Walkern 1, are heavily weathered. However, some observations can be made.

Of the two robed crucifixions that at Langford (no. 2) has a calm, massive presence. The robes are stiff and formalised without the elaborate folds and fluttering draperies characteristic of figure drawing in the Winchester style; the only folds in the robe are vertical and the hem of the garment is straight (Ills. 294–5). This formality has little in common with much of late pre-Conquest art, but in spirit lies closest to the rounded, solid figure style employed in some very late pre-Conquest works, notably in BL MS Cotton Tiberius B. V, a work of the second quarter of the eleventh century (Temple 1976, no. 87, ills. 273–6). Here, in a calendar, in illustrations accompanying Cicero's Aratea, and in the Marvels of the East, are figural illustrations probably executed by the same artist. The figures are solid and well-rounded with much of the weighty presence of the Langford Christ. The draperies are usually restrained and naturalistic, with only some of the hems betraying the fluttering folds of the Winchester style, features displayed at their best in scenes accompanying May and August in the calendar (ibid., ills. 273–4). There is really no equivalent to this style in the ivories, where the small scale of the crucifixions makes the figures rather delicate or even emaciated. Draperies are generally calmer and more naturalistic than in the manuscripts, but this may owe more to the nature of the material than to the intention of the artist. If these manuscript parallels can be accepted then it suggests a date for the Langford 2 crucifixion in the middle of the eleventh century, a date which is not at variance with the iconographic parallels.

At Walkern 1 (Ill. 397) the calm, expressionless face of Christ creates the same stillness and serenity as at Langford 2 and, although different in technique, the folds of the garment are linear and without the fluttering folds of the Winchester style. The effect which is sought appears to be the same as that at Langford. It is clear, however, that the flat robed body was intended to be painted and if the paint had survived the effect might have been quite different from that presented now. As for Langford, an eleventh-century date seems most appropriate and does not conflict with the architectural or iconographic evidence.

The other Langford crucifixion (no. 1) shares the solidity of these two robed crucifixions. The figures are naturalistic, without the agitated gestures characteristic of so much of Winchester-style art. As on Langford 2, the draperies lack the fluttering folds of the Winchester style; instead the folds fall vertically. Only on the figure of Christ is the hem of his loin-cloth folded in a manner typical of the Winchester style, if rather subdued (Ill. 293). Heavy parallel lines indicate the upper edge of the loin-cloth, probably a conventionalised version of the folds at the waist often employed in manuscript crucifixion scenes, for example, in the Judith of Flanders Gospels (Temple 1976, no. 93, ill. 289) or the Sherborne Pontifical (ibid., no. 56, ill. 171). A similar pair of hard horizontal lines indicates the hem of the garment of the figure of St John. Again this appears to be a conventionalisation of the folded hems employed in manuscript crucifixion scenes, as for example on the figures of the Virgin in the Arenberg Gospels (ibid., no. 56, ill. 171) and in the Sherborne Pontifical (ibid., no. 35, ill. 134). The heavy, solid form of the figures and the restrained nature of the draperies may suggest a date close to the robed Langford crucifixion, in the mid eleventh century.

In terms of stylistic analysis the Breamore 1 crucifixion presents a formidable problem since it is almost completely defaced (Ills. 425–6, 428); only the cloud from which issues the Hand of God survives (Ill. 427). This takes the form of three undulating, parallel mouldings from which emerge the Hand of God. Two further pairs of undulating mouldings link the outer ends of the cloud with the Hand of God to form a rough triangle. The use of a series of parallel, undulating lines to indicate the cloud occurs in a number of manuscript crucifixion scenes, most notably in the Arenberg Gospels (Temple 1976, no. 56, ill. 171). It occurs also in non-crucifixion scenes, as in the Evangelist portraits in the York Gospels of the late tenth or early eleventh centuries (ibid., no. 61, ills. 183–4). The triangular form of the cloud cannot be paralleled in crucifixion scenes, but in the translation of Enoch in the Caedmon Genesis of c. 1000, Enoch disappears into a triangular cloud which is composed of three, roughly parallel, undulating lines, as at Breamore (ibid., no. 58, ill. 189). This parallel may suggest an eleventh-century date for the Breamore crucifixion, but it is slender evidence.

Like the Breamore example the crucifixion at Headbourne Worthy is also destroyed leaving only the Hand of God, which is of similar form (Ills. 427, 448). At Romsey 1 the cloud from which the hand issues is composed of a number of tightly-packed volutes or curls (Ills. 451–2), a form unparalleled elsewhere, but which is presumably merely an elaboration of the single volute used on the cloud on fol. 22v of the York Gospels (Temple 1976, no. 61, ill. 181). The figure of Christ is calm and monumental, and the musculature is carved in detail. The loin-cloth is folded at the waist, and falls between the legs in an 'apron' fold paralleled only in BL MS Harley 2904, a work dated to c. 975–1000 (ibid., no. 41, ill. 142), and in BL MS Arundel 60, dated to c. 1060 (ibid., no. 103, ill. 312). However, on the Romsey crucifixion the cloth is pulled tight over the thighs, something not seen in BL MS Arundel 60, but which does occur in BL MS Harley 2904. The Romsey draperies exhibit few of the agitated folds employed in BL MS Harley 2904, but the manuscript figure in its calm monumentality, with all the musculature well drawn, is otherwise close in feel to the Romsey figure, for which on this evidence a similar late tenth-century date can be proposed, although this is slightly at variance with the iconographic evidence discussed above.

The primary, iconographical and stylistic evidence suggest a date in the late tenth or eleventh century for these large-scale crucifixion scenes, with the accent perhaps on the eleventh century. This is confirmed by the historical sources. These contain numerous mentions of crosses, but both Old English and Latin use the same word, rōd and crux respectively, to describe both a cross and a crucifix, so that crucifixion scenes can only be recognised where figures are mentioned (Dodwell 1982, 211). In addition to the crucifixions at Waltham, Essex, Winchester, Peterborough, and Bury St Edmunds, noted above, examples are recorded at Evesham, Winchester, Bury St Edmunds, Suffolk, Durham, and Ely. At Evesham, Leofric, earl of Mercia (d. 1057) and his wife Godiva gave a crucifixion accompanied by the figures of the Virgin and St John (ibid., 212, 318, n. 224). At Durham, where the donors were also distinguished secular figures (Judith and her husband Tostig, earl of Northumbria (1055–65)), again the crucifixion was accompanied by the figures of the Virgin and St John (ibid., 119, 281, n. 147). Stigand, bishop of Elmham (1043–47), Winchester (1047–1066), and archbishop of Canterbury (1052–70), was a particularly active benefactor of Ely (ibid., 211, 318, n. 221), Bury St Edmunds (ibid., 213, 319, n. 241) and, in 1057, of Winchester (ibid., 212, 319, no. 231).

All these crucifixions were large, probably life-sized (Dodwell 1982, 212) and made of precious metals, presumably over wooden cores, with the exception of the stone crucifixion at Waltham and perhaps also the one from Durham (ibid., 119, 213). All of the crucifixions, with the possible exceptions of the examples from Waltham and Peterborough, were of eleventh-century date. This may reflect an imbalance in the historical record; for example the great cross of gold given by Archbishop Albert of York to York Minster in the eighth century may have been a crucifix (ibid., 210, 317 n. 211), but it is also possible that the fashion for near life-sized figure groups was confined to the very late Anglo-Saxon period. This view may be supported by the fact that other life-sized figures are recorded in the late tenth and eleventh centuries, such as the four gold and silver statues of virgins given by Abbot Brihtnoth to Ely in the late tenth century (ibid., 214–15, 319 n. 244), and the Madonna and Child made for Abbot Aelfsige of Ely (981–c. 1019) (ibid., 215, 319, n. 246).


Besides the large-sale crucifixions there are five other, smaller, figure-decorated panels from south-east England: at Stepney in London; Oxford (St Michael 2); and in Sussex at Jevington 1, Sompting 13, and Tangmere.

Of these the crucifixion panel at Stepney is clearly related to the large-scale crucifixions, belonging iconographically to the second group discussed above in which Christ is depicted as dead with his arms and legs bent (Ill. 354), and for which a late tenth- or eleventh-century date has been proposed. The crucified Christ at Stepney is accompanied by the figures of the Virgin and St John. Over the horizontal arms of the cross are roundels containing busts personifying the sun and moon, symbols often associated with Christ as King (Raw 1990, 129). This feature can be paralleled in pre-Conquest manuscript art only in the eleventh century, in the Sacramentary of Robert of Jumièges dated to c. 1000 (Temple 1976, no. 72; Rice 1952, pl. 33b), BL MS Cotton Titus D. XXVII, dated to c. 1023–35 (Temple 1976, no. 77, ill. 246), the Winchcombe Psalter of c. 1030–50 (ibid., no. 80), the Judith of Flanders Gospels of 1025–50 (ibid., no. 93, ill. 289), and the psalter BL MS Arundel 60, dating to c. 1060 (ibid., no. 103, ill. 312). Of these, only the Winchcombe and Arundel Psalters have personifications of the sun and moon of precisely the same form used at Stepney, that is, with the face and hands of the half-length figures veiled. In ivory only two crucifixions have personifications of the sun and moon, a pectoral cross of the tenth or eleventh century in the Victoria and Albert Museum (Beckwith 1972, no. 34, pl. 71) and the Heribert crosier dated to the early eleventh century (ibid., no. 30, pl. 80). In neither case do the half-figures have their hands and faces veiled as on the Stepney slab. The iconographic evidence, therefore, confirms a very late tenth- or eleventh-century date for the slab, with the accent on the eleventh century. This dating is also supported by the form of the cross. This has elaborately moulded ends to the arms; the head and foot are also moulded, but the mouldings differ from each other and from those at the ends of the arms. This arrangement is paralleled elsewhere only in the crucifixion scene in the Arenberg Gospels (Fig. 21c), dated to c. 990–1000 (Temple 1976, no. 78, ill. 244). The cross in the presentation scene of the Liber Vitae of the New Minster, dated to 1031, also has elaborately moulded ends to the arms, but all the mouldings are identical (loc. cit.) and, as noted below, there is a similar cross on the grave-cover from Weyhill, Hampshire (Ill. 473).

The crucifixion scene is contained within a broad frame which is decorated with encircled lobed leaves separated by pairs of transverse bars. This is clearly a variant of the acanthus ornament used widely in both Carolingian and Ottonian art, and in Anglo-Saxon England, for framing narrative scenes in both manuscripts and ivory. This relationship is demonstrated in the borders of fol. 23r of Rheims, Bibl. Mun. MS 9, dated to c. 1062 (Temple 1976, no. 105, ill. 299). Here the vertical borders are enriched with acanthus ornament of the normal Winchester type; the horizontal borders are, however, filled with a conventionalised version in which encircled, fleshy trefoil leaves develop alternately from each side. The Stepney panel merely takes this conventionalisation one step further in a manner which is common in Romanesque art, as in the borders of the Chichester panels (Zarnecki 1951b, pl. 80) (Fig. 22), and those of the Tree of Jesse illustration in the Winchester Psalter (BL MS Cotton Nero C. IV, fol. 9r.), dating to c. 1140–60 (Turner 1971, pl. 1). This form of leaf decoration appears, then, to have been a late pre-Conquest development and would support a date for the Stepney panel in the eleventh century, and probably towards the middle of the century.

The panel at Sompting, Sussex (no. 13; Ill. 181), is decorated with a single figure of an ecclesiastic, probably St Benedict or St Gregory, beneath an arch. To the left is a crosier, to the right a reading desk. The panel has been dated as late as the twelfth century presumably on the basis of resemblances between the handling of the robes and the 'damp fold' style of twelfth-century art (Clapham 1935a, 408–9), although, more commonly, a late Anglo-Saxon date has been preferred (Rice 1952, 107, pl. 15b; Nairn and Pevsner 1965, 331). Any connection with the damp fold style must be rejected. In that style the standing figures normally have the weight of the body thrown on to one leg, usually the figure's right leg, while the body itself is curved back in that direction. The garment is pulled tight over the right leg, but narrow transverse swags sub-divide the tensed garment. The left leg is not normally emphasised in this way. The garment is also pulled tight over the torso of the figure with folds outlining an area of tightness over the stomach, and over the chest. The overgarment may fall in a fold between the legs, and the undergarment stops at ankle level. Figures of this type from a southern English atelier are perhaps best seen in the work of the 'master of the leaping figures' in the Winchester Bible (Oakeshott 1945, pls. VI, XXV). Few of these features are paralleled at Sompting. The weight of the figure is carried on both legs and there is no curve to the body. The robes are drawn equally tightly over both legs and are without the subdivisions normal in the damp fold style. There is no tightening of the garment over the stomach, and it is pulled tight over the shoulders rather than the chest. The clothing falls down in a fold between the legs, as in the damp fold style, but it also falls to each side of the figure and completely obscures the feet. It is conceivable that all these aberrant features might be accounted for if the piece were the product of a local workshop, but nothing else remotely like this figure is known from among the substantial corpus of Romanesque sculpture in south-east England.

Elements resembling both the damp fold style and the Sompting figure can be traced in pre-Conquest manuscript art. For example, in the illustrations to Prudentius's Psychomachia, BL MS Cotton Cleopatra C. VIII, dated to the late tenth century (Temple 1976, no. 49, ills. 159–62), a number of scenes have figures with their draperies pulled tightly over both legs, as on fols. 10v (ibid., ill. 159), 11r (ibid., ill. 160), and 17r (ibid., ill. 163). As on the Sompting figure the robes often fall in a fold between the legs, as on fol. 17r, and sometimes also to one or both sides, as on fol. 11r. Here the figure of Patience also has her draperies gathered in a swirling mass at the hem, but at ankle level, and not obscuring the feet, as at Sompting. A number of figures also have swirls or spirals on the shoulders, as on fols. 17v and 29r, a feature paralleled also at Sompting. The Prudentius scenes are drawings, but these features can be paralleled on fully painted figures. The pulling tight of the garment over the leg, for example, is encountered on fol. 88r of the Rheims, Bibl. Mun. MS 9, dated to c. 1062 (ibid., no. 105, ill. 299); there is also a swirl on the shoulder, a feature paralleled also on fol. 15r of Warsaw, Bibl. Narodowa MS I. 3311 (ibid., no. 92, ill. 283). In no case is the emphasis on the tightening of the garments as pronounced as it is at Sompting, but nevertheless the Sompting figure could be plausibly interpreted as a local variant of this pre-Conquest type.

Support for the pre-Conquest dating of Sompting 13 derives also from the form of the arch enclosing the figure (Ill. 181). Such arches commonly occur both in Romanesque and in pre-Conquest art. In Romanesque works, however, such arches are usually solid and architectural, as in the Albani Psalter of 1119–46 (Rickert 1954, pl. 61), or Boethius's De Consolatione Philosophiae of the second quarter of the twelfth century (ibid., pl. 68A). In both cases the arch is supported on piers as at Sompting, but the arch head is much more substantial than the flimsy structure employed at Sompting. Above it, filling the spandrels, are depictions of roofs and towers in place of the palmettes used at Sompting. Similar arches can be found in pre-Conquest art, as on fols. 21v and 30v of the St Margaret Gospels (Temple 1976, no. 91, ills. 279-80), but in tenth- and eleventh-century manuscripts the spandrels of the arches are more often filled with small foliate motifs (Fig. 20a, d–g), as at Sompting. Such plant motifs also occur in the spandrels of the arches of Canon Table frames, as on fols. 9v and 10r of the Kestner Museum, Hanover, MS WM XXIa, 36 (ibid., no. 67, ills. 224–5). This feature is not easily paralleled in post-Conquest manuscript art, and must suggest a pre-Conquest date for the Sompting panel.

A late pre-Conquest date, therefore, seems most plausible for Sompting 13, but it must be assigned to the very end of the period as the upright leaf capitals of the arch are paralleled elsewhere only on the chancel arch at Botolphs, Sussex (Ill. 5), for which a mid eleventh-century date is argued above, and to a lesser degree on the tower arch at Sompting itself (Ill. 183), for which a similar date might be suggested (but see above). Such a date would be consonant both with the rather solid, static style of the figure, without the agitated gestures and elaborate fluttering draperies of the Winchester style, and with the stone type. The fine-grained limestone of which the figure is made has the same slight polish as the lengths of acanthus-decorated frieze at Sompting (nos. 1–8); this may suggest that the pieces derive from the same feature or decorative scheme and are of similar date. A mid eleventh-century date is argued for the frieze fragments above.

Much more readily datable is the figure-decorated panel at Jevington, Sussex, depicting Christ trampling on the beasts (Ill. 232). The animal to the right of the figure (Ill. 234) is in the Urnes style, which developed in Scandinavia c. 1025–50 (Wilson and Klindt-Jensen 1966, 160). The style is characterised by the use of ribbon-like animals which form a pattern of loops and curves, and interlace with narrow, filiform elements. The loops are often arranged in figure-of-eight or multi-loop schemes and frequently the animals are engaged in combat (Fuglesang 1978, 207-8; Wilson and Klindt-Jensen 1966, 147). The animal's eye is usually of lentoid shape. At Jevington the animal is ribbon-like and is engaged in combat with another, smaller, ribbon-like animal. Its body is arranged in loops and there are narrower, looping elements, but coarser than the usual filiform elements of Scandinavian Urnes-style works. The eyes are lentoid. The second animal, to the left of Christ's feet (Ill. 233), is a semi-naturalistic quadruped, but it employs the same lentoid eye and has the hind legs developing into loops which resemble those of the Urnes-style animal.

There is little comparative material from south-east England where, apart from the Jevington panel, only four other Urnes-style pieces survive: a silver mount from London (Kendrick 1949, 118–19, fig. 20b); a manuscript sketch in BL MS Royal I. E. VI, probably made at St Augustine's abbey, Canterbury (Budny 1984, 255, 257–63); a mount from Colchester, Essex (Wilson 1981, 78, fig. 66), and a stirrup iron from Mottisfont, Hampshire (Seaby and Woodfield 1980, 108, no. 9, fig. 4). Of these, the London mount is decorated in the classic Urnes style and is probably an import from Scandinavia, but (like the Jevington sculpture) the remaining pieces are Urnes-style objects made in England rather than English Urnes-style pieces in the classification proposed by Owen (Owen 1979). In any event, a date after the Conquest must be proposed for all of them, since it seems unlikely that the Urnes style can have reached England before that date (Wilson and Klindt-Jensen 1966, 153).

Although there is no other sculpture from south-east England exhibiting Urnes-style features with which the Jevington piece can be compared, it has been suggested that the sculptures from Southwell (Kendrick 1949, 121–2, pl. LXXXVI) and Hoveringham (ibid., 122; Clapham 1930, pl. 59b) in Nottinghamshire, West Marton (Kendrick 1949, 123–4, pl. LXXXVIII) and Kirkburn (ibid., 120, pl. LXXXIV), Yorkshire, Norwich, Norfolk (Zarnecki 1951b, 38, pl. 76; Zarnecki et al. 1984, no. 126, pl. 126) and York (Solloway 1910, 56, pl. facing; Moulden and Tweddle 1986, no. 40, pl. 1b) are all decorated in the Urnes style. This argument is difficult to sustain. The West Marton cross-shaft has nothing in common with the features of the style outlined above. The York sculpture has an animal head with a lentoid eye and lappets, but too little of the body survives to be sure that this was in the Urnes style. At Kirkburn the capitals do have looped, ribbon-like animal bodies interlacing with filiform elements, but they are not in other respects related to the Urnes style. At Southwell and Hoveringham there are dragons with looped ends to their bodies interlacing with filiform elements, but in other respects the sculptures are of purely Romanesque style. At Norwich there is merely an echo of this arrangement; the filiform elements have become attenuated plant sprays and the animal's bodies are beaded. In no case can an Urnes-style ancestry be convincingly argued, but if these twelfth-century pieces retain only echoes or resonances of the Urnes style, then the Jevington piece, with its recognisably Urnes-style animals, must be placed earlier in date, probably in the late eleventh century.

This dating is supported by the handling of the sculpture of the figure itself, with its heavily-rounded limbs, a degree of undercutting, and an attempt to show the musculature of the body. This is related in style, if not in accomplishment, to the large-scale crucifixion scenes, and in particular to that at Romsey 1, Hampshire. Here there is the same use of undercutting and again an attempt to show the musculature of the body (Ills. 451–2). This same solid treatment, but in lower relief, occurs on the eleventh-century figural panel from the Old Minster, Winchester (no. 88; Ill. 646), and is reflected also in mid eleventh-century manuscript art. For example, the figures in the Canterbury manuscript BL MS Cotton Tiberius B. V, of c. 1025–50 (Temple 1976, no. 87, ills. 273–6), share none of the exaggerated gestures and fluttering draperies of the Winchester style; instead the figures are given a solid, monumental form, as in Aelfric's Pentateuch, a Canterbury manuscript of broadly similar date (ibid., no. 86, ills. 265–72, fig. 34). The Jevington sculpture, then, on the evidence of both the figural and animal style, should be placed in the second half of the eleventh century, continuing pre-Conquest sculptural traditions into the post-Conquest period in a geographically remote region.

The figure sculptures from Oxford (St Michael 2) and Tangmere, Sussex are completely divorced from contemporary trends in the ornament of sculpture, manuscripts, and metalwork, and belong firmly to the tradition of folk art. The apotropaic figure at St Michael's was discovered built into the belfry stage of the eleventh-century tower, and must, therefore, be of a similar or earlier date. The Tangmere panel (Ill. 219) has been reused as the head of a late eleventh- or early twelfth-century window and is presumably of pre-Conquest date. In both cases the simplicity of the work and the lack of comparative material renders closer dating impossible.


There are 53 funerary sculptures from south-east England of late pre-Conquest date, of these 41 are grave-covers, one is a coffin lid, ten are grave-markers, and one is possibly the end or side of a box sarcophagus.

Among the grave-covers, three geographical groups can be discerned. At Cardington and Milton Bryan, Bedfordshire, Great Maplestead, Essex, and St Benet Fink, London, there are grave-covers of a type common in the east midlands. There are eight covers at Oxted (two), Tandridge (one), and Titsey (five), all within a three-mile radius on the Greensand belt of Surrey; and on a ten-mile stretch of the Greensand belt in Sussex the churches at Chithurst, Cocking, Stedham, and Steyning, have between them produced nineteen covers; the one at Headbourne Worthy, Hampshire (no. 2; Ill. 685), is related to some of the Sussex examples. The grave-covers from Stratfield Mortimer, Berkshire (Ills. 695–709); Weyhill and Winchester (Old Minster 6 and St Maurice 1), Hampshire; St Augustine's Canterbury 2 and Dover St Peter, Kent; London (City 1); Oxford (cathedral); and Arundel and Bexhill in Sussex, are unrelated to these groups or to each other.

Grave-covers of the east midlands type were first described and classified by Fox (Fox 1920–1) who defined two groups and six types. In group A the sculpture is all in relief and all types have a median ridge. Covers of type 1 have the ridge crossed by short bars, but the cross-bars and ridge are filled with interlace and there are panels of interlace flanking the ridge (ibid., 25, pl. III). In type 2 there is a central cross-bar and the ridge has U- or V-shaped ends; interlace is confined to four panels flanking the median ridge (loc. cit.). Type 3 has a median ridge and cross-bars of half-round section instead of the more usual flat relief, and again there are flanking panels of interlace (ibid., 25, pls. II, IV). In group B the median ridges are incised although the flanking interlace is still in relief (ibid., 25, pls. IV, VII). Type 4 employs a cross at either end of the median ridge. Type 5 has a median ridge with U- or V-shaped ends (ibid., 25, pl. V), and type 6 employs a ringed cross at the ends of the median ridge, but no cross-bars (loc. cit.). The cover from Milton Bryan, Bedfordshire (Ill. 361), belongs to type 2 and the one from St Benet Fink, London (Ills. 345–6) to type 5. Of the grave-cover at Cardington, Bedfordshire (Ill. 264), only the central cross-bar and parts of the flanking interlace panel survive, but the median ridge is incised and the interlace panels have a broad borders around them, both features of type 4. The piece from Great Maplestead is fragmentary, but has a median ridge flanked by interlace (Ill. 274) and may, therefore, be part of a cover of the east midlands type.

Fox dated grave-covers of this general type to the late tenth and eleventh centuries on the basis of the archaeological evidence from two sites, Peterborough cathedral and Cambridge castle. At Peterborough during restoration work in 1888 two covers were found in situ immediately outside the walls of the pre-Conquest church which was assigned to the refoundation of the monastery by St Aethelwold in 970 (Fox 1920–1, 23–4, 31–2). At Cambridge castle in 1810 part of a cemetery employing eight such grave-covers was found in situ under the rampart which was constructed in 1068 (ibid., 20–1). Fox argued that these two sites served to fix the dates between which the grave-covers were used. Unfortunately, Fox's assumption that no burials can have taken place at Peterborough during the period between the destruction of the monastery in 867 and its refoundation in 970 cannot be relied upon. So deficient is the documentation that a church may well have continued to function on the site throughout this period. Even if there was no church presumably burials continued in the churchyard. Similarly, covers of this type may have continued to be used for some time after 1068; nevertheless, the archaeological evidence does point towards a late Anglo-Saxon date for the majority of them.

The grave-covers from Oxted, Titsey, and Tandridge, all in Surrey, form a more homogeneous group, all but one being decorated with a simple Latin cross in relief. The exception, Titsey 4 (Ill. 255), has a median ridge in relief with a cross-bar and part of a V-shaped terminal, a form closely related to type 2 of the east midlands group, although it is without the interlace. This single weak link may suggest a late pre-Conquest date for the whole group. Unfortunately, there is no confirmatory evidence from the circumstances of discovery.

Grave-covers similar to those encountered in Surrey occur also in Sussex. Here four general types can be identified: those decorated with a Latin cross (type 1); those decorated with a median ridge having a cross-bar near each end (type 2); those having a median ridge with V-shaped ends (type 3), and those decorated with a Greek or Latin cross near each end (type 4). Examples of types 1 and 2 occur at Chithurst, for example, no. 6 (type 1), no. 2 (type 2) and Stedham, for example, nos. 7–11 (type 1) and 4 (type 2), while covers of type 3 are found at Cocking, Stedham (no. 6), and Steyning (no. 2), and examples of type 4 occur at Stedham (no. 2) and Steyning (no. 1). A single cover from Chithurst (no. 1, Ill. 220) is a type 2/3 hybrid. The dating of most of the Sussex grave-covers is based on architectural evidence. The one from Cocking was discovered built into the foundations of the chancel, dated to the late eleventh century (Johnston 1921, 182, fig. 1), while one of the covers from Steyning (no. 1) was found incorporated into the foundations of a part of the nave dated to c. 1170–80 (Bloxham 1864, 238; Nairn and Pevsner 1973, 538). The grave-covers and -markers from Stedham were reused as building material in the base of the nave walls, dated to the twelfth century, a dating confirmed by the surviving twelfth-century window in the nave (Page 1907, 365). This data suggests that covers of the Sussex type had fallen out of use by the early twelfth century, although it is hardly decisive evidence. Certainly the in situ slabs of twelfth-century date from other southern English sites, such as Old Sarum (Shortt 1976, pl. p. 33), and Trowbridge, Wiltshire (Goddard 1904), are very different in their decoration. For example, at Trowbridge one of the covers has a Latin cross, but of far more elaborate form than those employed in Surrey and Sussex and without the stem extended to form a median ridge (ibid., pl. on 6). A second cover has a median ridge, but has a border of star pattern, and semicircles of star pattern and triple-reeded mouldings project from the border to touch the median ridge (loc. cit.).

Of the same general date must be the grave-cover from Headbourne Worthy, Hampshire (no. 2). It is coped, like the example from Tandridge (Ill. 231), and has a median ridge with, at the surviving end, subsidiary ridges developing from it and running into the corners: a variant of the bifurcating median ridge type (Ill. 685). This type of decoration, combined with interlace, is also employed among the East Anglian covers, as at Milton Bryan, Bedfordshire (Ill. 361).

Among the remaining grave-covers, three, from Dover (St Peter), Kent (no. 1; Ill. 77), Weyhill, Hampshire (Ill. 473), and Arundel, Sussex (Ill. 1), are decorated with Latin crosses, but the crosses are very different in form from each other and from the crosses employed on the Surrey and Sussex slabs. Unusually, the example from Dover St Peter has a rounded head and foot and is decorated with a relief Latin cross which has an exaggerated fan-shaped head. On the cross-head is a runic inscription; unfortunately, this is of little help in dating the carving (see Parsons in the catalogue entry). Dating therefore depends on the form of the relief cross (Ill. 77). There are no parallels for this form in southern or midland England, but in Northumbria the type appears to have been current in the tenth and eleventh centuries, as, for example, on the standing crosses from Kirkby Wharfe (Collingwood 1927, 88, fig. 107) and Collingham, Yorkshire (ibid., 88). The closest parallel to the form of the cross on the Dover cover is provided by a grave-cover from Spennithorne, Yorkshire, which is decorated with a closely comparable cross (ibid., 90, fig. 101). The Spennithorne cover is also decorated with bifurcating interlace typical of Anglo-Scandinavian sculpture in the north, and can, therefore, be dated to the tenth or eleventh century. A similar date can be suggested for the Dover monument, but such a geographically remote parallel is, however, an unsatisfactory basis for reliable dating.

At Weyhill, Hampshire (Ill. 473), the face of the grave-cover is edged and divided into two roughly equal faces by pairs of crude roll mouldings. The upper field is undecorated, although areas of rough chiselling suggest that sculpture may have been cut away in the manner of the crucifixions from Headbourne Worthy 1 and Breamore 1, Hampshire. The lower field is decorated with a Latin cross with the arms having concave edges and standing on an elaborately-moulded base. Pairs of transverse mouldings separate the arms of the cross from the centre which is decorated with a rosette. There is no close parallel for the shape of the cross, but the cross at Stepney, London, has a similar elaborately-moulded base (Ill. 354) and there are similar bases in the crucifixion scene on fol. 9v of the Arenberg Gospels, dated to c. 990–1000 (Temple 1976, no. 56, ill. 171) (Fig. 21), and in the presentation scene on fol. 6r of the Liber Vitae of the New Minster, dated to c. 1031 (ibid., no. 78, ill. 244). In each case, however, the ends of the arms of the cross are also elaborately moulded. There is no obvious parallel for the transverse mouldings on the cross-head and -arms, although something similar is seen on a tenth- or eleventh-century Anglo-Saxon or Danish ivory crucifixion now in the National Museum, Copenhagen (Beckwith 1972, no. 36, ill. 73). These parallels suggest a tenth- or eleventh-century date for the cover, but do not allow the dating to be further refined.

The Latin cross decorating the stone from Arundel, Sussex, is also of a distinctive, but different, form, having the lower limb of the cross expanding before ending in a point. In addition the cross is portrayed as if suspended from a triangular loop. The only adequate parallel for the shape of the cross is provided by an ivory pectoral cross of c. 1050 in the Victoria and Albert Museum (Beckwith 1972, no. 45, pls. 99–102: Backhouse et al. 1984, no. 125, pl. 125). This has a circular field in the centre, but otherwise compares closely in shape with the Arundel cross. Like the latter, the pectoral has a triangular suspension loop, but with the apex pointing upwards, not downwards as at Arundel. This parallel suggests a late pre-Conquest or early post-Conquest date for the Arundel piece.

The decoration of the remaining grave-covers from south-east England, from Stratfield Mortimer, Berkshire, Canterbury St Augustine's 2 and Dover St Mary in Castro 1, Kent, Oxford cathedral, and Bexhill, Sussex, is more diverse. At Stratfield Mortimer (Ills. 695–709) the decoration consists of an inscription around the edges of the upper surface of the stone. This records that it covered the grave of one Aethelweard son of Kyppingus. The name Aethelweard is so common in the surviving documentary sources that it is impossible to identify the person commemorated. The name Kyppingus is much less common and it is possible that he was related to, or can even be identified with, the Cheping who is recorded as holding the church of Stratfield Mortimer in the time of the Edward the Confessor (Cameron 1901–2, 73; but see the catalogue entry). This would place the cover around the time of the Conquest.

The grave-cover from St Augustine's Canterbury, Kent (no. 2), is also decorated with an inscription, this time arranged in eleven horizontal lines (Ills. 24–8). On the basis of the letter forms and the late form of the names, Higgitt dates the slab to the mid to late eleventh centuries (see the catalogue entry). The animal ornament, comprising an incised quadruped and bird between lines two and three, may suggest a rather earlier date, however. The quadruped has a triangular body and rear legs; the offside front leg is raised and the animal looks upwards. Similar animals occur fairly widely in the manuscripts, perhaps the best parallel being provided by Durham Cathedral Library MS A. IV. 19. On fol. 57r, again fitted into the text, is a closely-comparable animal, although here backward looking. This manuscript dates to the early tenth century (Temple 1976, no. 3, ill. 8), but similar quadrupeds occur in later works as, for example, in the borders of the presentation scene, fol. 1v, in the Vita Cuthberti, dated to c. 937 (ibid., no. 6, ill. 29). Later versions, however, usually have the animal's tail brought up between the legs and across the body as, for example, on p. 127 of Monte Cassino, Archivio della Badia MS BB. 437, 439, dated to c. 1050 (ibid., no. 95, ill. 288). This feature occurs also in metalwork, as on the cast bronze censer cover from Canterbury, dated to the tenth century (Wilson 1964, no. 9, pl. XII). The bird on the Canterbury grave-cover also finds its best parallels in the tenth century. It has a plump, rounded body, a parallel-sided tail with a square end, and large curved feet. The head is outstretched and looks slightly upwards. Birds of this type are found in numerous tenth-century manuscripts as, for example, on fol. 115v of Oxford, Bodleian Library MS Tanner 10, dating to the first half of the tenth century (Temple 1976, no. 9, ill. 37). Here the birds have precisely the same form as on the Canterbury example, although they form part of an initial. Similar birds are also encountered in metalwork, as on the Canterbury censer cover (Wilson 1964, no. 9, pl. XII) and on the small unprovenanced gilt-bronze jug in the British Museum (ibid., no. 147, pl. XLIII), both works of the tenth century. Good eleventh-century parallels are not so easy to find, but the type recurs on the Bayeux Tapestry, usually with the wings extended, but occasionally they are folded, as on the present carving. The birds on the tapestry are bigger and more elongated, however, and the wing and tail merge into each other (e.g. Wilson 1985, pls. 69–73).

The best parallels for the animal ornament on the Canterbury grave-cover are, therefore, tenth-century ones, whereas the inscription is unlikely to be earlier than the mid eleventh century (see Chap. IX). This apparent discrepancy might support Okasha's suggestion that the inscription is secondary (Okasha 1983, no. 161). Art-historical, epigraphical and linguistic dating methods are crude tools for such fine dating, however, and this apparent conflict in the dates derived from different methods may do no more than demonstrate that fact.

At Dover the grave-cover from St Mary in Castro (no. 1) is decorated with crude incised foliate decoration within cable or foliate-decorated borders (Ills. 62–3, 68). Its reuse as paving, probably in the late twelfth century (see the catalogue entry), suggests a date in the early part of the twelfth century at the latest, and, allowing for a reasonable period of primary use, probably in the eleventh century. The decoration itself is so crude that it is of little help in dating. Equally difficult to date is the grave-cover from Oxford cathedral (Ill. 362). This was recovered from the core of twelfth-century walling in the chancel (BL Add. MS 27765G, fol. 32r), perhaps suggesting a date in the eleventh century or earlier if a reasonable period of primary use is allowed. The decoration consists of a series of groups of concentric semicircles developing from the long edges. This arrangement is paralleled on the cross-shaft fragments from Saffron Walden (Ills. 371–3). A number of grave-covers in the east midlands, possibly produced at Barnack (Butler 1956, 90), also employ nested geometrical shapes, usually triangles (as at Waterbeach, Cambridgeshire (ibid., 90, fig. 1.1) and Oxford, where they develop from the long edges of the cover), or lozenges (as at Waterbeach and Wood Walton, Cambridgeshire (ibid., 90, figs. 1.1, 1.2)). These nested geometrical shapes are more widely spaced than at Oxford and are normally used with a central rib and cross patée, features also lacking at Oxford. Butler has suggested an eleventh-century date for the east midland covers (ibid., 90); if this is acceptable then the Oxford example, with its rather different decoration, may be regarded as a regional variant of similar date, or it may represent a late pre-Conquest prototype out of which covers of the east midlands type developed.

The grave-cover from Bexhill, Sussex, is unique in south-east England, and takes the form of a truncated pyramid on a rectangular base (Ills. 10–19). The faces are panelled, and the fields are decorated with interlace, animal, plant, and geometrical ornament. As noted above, this combination of different motifs is characteristic of the late eighth and ninth centuries, but the individual motifs employed at Bexhill cannot be paralleled in works of that period. Of the four major panels of interlace at Bexhill, three (Ills. 12–14) are filled with flaccid, disorganised patterns which lack any logical structure, in stark contrast to the carefully considered and rigidly organised interlace employed, for example, at Elstow, Bedfordshire (Ill. 272), or Rochester, Kent (no. 1; Ill. 141). Such incoherent interlaces are not paralleled on stone sculpture or metalwork, or in manuscripts in the late pre-Conquest period, but do occur on leather sheaths. A scramasax sheath from Gloucester, dated to the eleventh century, employs a closely comparable interlace, although with zoomorphic elements, and there is an almost identical sheath from Parliament Street, York, probably also datable to the eleventh century (Tweddle 1986a, 237–8, fig. 107, pl. XI).

The animal ornament on the Bexhill cover consists not of the winged or wingless bipeds characteristic of the late eighth and ninth centuries, as at Elstow, Bedfordshire (Ills. 269–71), or Godalming 1, Surrey (Ill. 84), but instead of ribbon-like animals with narrow, shapeless heads and fan-shaped tails (Ills. 18–19). The animals' bodies undulate and in the fields created by the undulations are triquetras linked to each other across the bodies. This type of animal is difficult to parallel on southern English stone sculpture except, perhaps, on the tenth-century shaft from Bishops Waltham, Hampshire, where there are similar ribbon-like animals, but tightly spiralled and without the interlace (Ill. 421). Looking further afield, however, an almost exact parallel is provided by a cross-shaft from St Mary Bishophill Senior, York (no. 1; Lang 1991, Ill. 249). Here face D is decorated with a similar ribbon-like animal, although the head and tail are missing, enmeshed in identical interlace. This shaft was found incorporated into eleventh- or early twelfth-century walling and probably dates to the tenth or eleventh century. There is no evidence for an earlier ecclesiastical occupation of the site despite thorough excavation (Ramm 1976). A closely comparable animal also occurs on face A of Aycliffe 2, a work of the second half of the tenth century (Cramp 1984, i , 43-4, pl. 9 (30)), and there is a fragment of a similar design at Chester-le-Street (no. 8), which is dated to the tenth century (ibid., i , 57, ii , pl. 24 (126)). The tenth- or eleventh-century date suggested by the nature of the interlace and animal ornament is further reinforced by the form of the cross in the lower of the two cross-decorated panels. This has expanding arms with convex ends and concave sides; the arms are linked together by short, slightly-curved bars. This cross is clearly related to the ring-head, a type which, as noted below, first appears in Anglo-Saxon sculpture in the tenth century and which was probably introduced by Scandinavian settlers.

The derivation of the ornament of the Bexhill cover, which combines late pre-Conquest motifs in a manner characteristic of earlier pre-Conquest art is puzzling and presents a combination not paralleled elsewhere in the pre-Conquest art of south-east England. One clue as to its origins is, perhaps, provided by the interlace in the central panel (Ill. 15). This pattern, with its interlacing pairs of diagonals linked at the ends and interlacing with plain and looped circles, is difficult to parallel in southern Anglo-Saxon art. The closest approach occurs in Cambridge, Corpus Christi College MS 411, where on fol. 40r the corners of a rectangular frame enclosing the initial Q prefacing psalm 51 have interlaces of similar form, combining free rings or nearly closed circles with diagonals or diagonal loops (Temple 1976, no. 40, ill. 128). This manuscript, dating to the last quarter of the tenth century, draws upon a Franco-Saxon exemplar (ibid., 63), and it is to Franco-Saxon manuscripts that it is necessary to turn to find a close parallel to the central interlace on the Bexhill cover. The Leofric Missal incorporates part of a ninth-century Sacramentary made in the region of Arras and Cambrai, and was in England by the second quarter of the tenth century (ibid., no. 17; Deshman 1977, 145–8). On fol. 61v the frame of the initial T has enlarged square corners, as in Cambridge, Corpus Christi College MS 411 (Deshman 1977, pl. 1a). One of the corner squares is filled with pairs of interlacing diagonals linked together at the ends, as on Bexhill. Another square employs a looped circle, as at Bexhill. It appears, therefore, that the central motif at Bexhill must be derived from an imported Franco-Saxon exemplar, such as the Sacramentary incorporated into the Leofric Missal. This would explain the anachronistic nature of the decoration on Bexhill 1, as Franco-Saxon works, drawing upon earlier Insular exemplars, employ the same range of motifs. For example, on fol. 127r of the Gospels produced at St Vaast in the late ninth century are panels of interlace, animal ornament, and plant ornament (Mütherich and Gaehde 1977, pl. 41). The plant ornament is of a type which is conventionalised, but recognisable, at Bexhill, and the animals are ribbon-like and enmeshed in interlace, although they have spindly legs. All that the sculptor at Bexhill has done is to copy and conventionalise some motifs and to substitute for others types with which he was more familiar.

Despite this great diversity of form and decoration, other types of grave-cover were probably used in south-east England. For example, the Old Minster, Winchester, has produced a coped cover (no. 6) with hipped ends and an axial inscription, dated to the early eleventh century (Ills. 509–13, 521), and the runic inscription from Winchester St Maurice 1, appears to have derived from a parallel-sided monument with a convex upper surface (Fig. 42; Ills. 667–70). The use of Old Norse runes probably serves to place it in the period of Scandinavian supremacy in southern England, c. 1016–42.

Closely related to some of the pre-Conquest grave-covers is the single surviving pre-Conquest coffin lid from south-east England, from Westminster abbey. The lid is decorated with a Latin cross having an expanding head and arms (type B6), but with a parallel-sided stem and foliated foot (Ills. 355, 357). There is no archaeological evidence for the dating of the lid, but the decoration resembles that of some of the covers from Surrey and Sussex, for which an eleventh-century date has been suggested above. The foliated foot employs a central leaf flanked by narrow, out-turned leaves, a form closely paralleled by the pilaster bases from Corhampton, Hampshire (no. 2; Ill. 440), for which a tenth- or eleventh-century date has been argued above; a similar date can be proposed for the coffin lid, although the evidence is clearly tenuous. Unfortunately, the building of Westminster abbey, begun c. 1050, does not provide a terminus post quem for the stone, as it is clear that there was some sort of monastic community here at an earlier date (Hunting 1981, 14–19).

Apart from the grave-covers and coffin lid, there are eleven probable grave-markers from south-east England, five from Winchester (Old Minster 2, 4–5, New Minster 2, and St Pancras 1), five from Stedham, Sussex (nos. 7–11; Ills. 240, 243–4, 246–8), and one from White Notley, Essex. Most of the examples from Stedham were recovered from the same twelfth-century footings as the grave-covers, and like them were moss-covered when found so may have been of considerable age when buried (Butler 1851, 20–1, fig. 2). An eleventh-century date, therefore, seems most probable; certainly they are similar to some of the small head-stones from Winchester, dated to the late tenth or early eleventh century, although with rather more elaborate decoration (e.g. New Minster 2; Ill. 661). The fragment from White Notley, Essex, is reused as the frame of a single-splay window (Ill. 375). The single-splay window is an eleventh- and twelfth-century type, and this suggests that the grave-marker belongs either to the tenth or eleventh century according to the date of the window. This is confirmed by the form of the cross decorating it. This was probably a disk-head, a tenth- and eleventh-century type which, as noted below, is related to the ring-head. Confirmation of this date also derives from the fact that there is a grave-marker of identical form from the graveyard at Cambridge castle, sealed by the defences built in 1068 (Fox 1920–1, pl. VII)

Among the grave-markers and -covers one group stands out, as the pieces are decorated with an accomplished version of the Ringerike style. This group comprises three carvings from London (City 1, St Paul's 1, and All Hallows 3), two examples from Rochester, Kent (nos. 2–3), and the grave-marker or box sarcophagus fragment from Great Canfield, Essex. On the grave-cover from the City and the Great Canfield fragment the decoration consists purely of plant ornament (Ill. 353; Fig. 32). On the grave-markers from St Paul's and All Hallows the design consists of a single large animal (Fig. 23; Ills. 351–2, 341). At Rochester the design on one of the pieces (no. 2) is too fragmentary for it to be reconstructed, but there is a runic inscription on one edge (Ill. 144) as there is on the piece from St Paul's (Ill. 350). On the other Rochester fragment (no. 3), the decoration consists of tendrils in Ringerike style on face A, a cross on face C, and a Latin inscription on the narrow edge of the rounded head (Ills. 147–50). In every case the sculpture is executed in low, flat relief against a flat background. Details of the design are picked out with incised lines having a V-shaped profile and often stopping at drilled holes. The similarities of sculptural technique as well as the repetitive designs, allied to the geographically restricted distribution of the group, suggests that they are the products of a single workshop, and possibly even of a single hand.

The nature of the decoration makes dating comparatively straightforward. The Ringerike style was current in Scandinavia from the late tenth century to the third quarter of the eleventh century (Wilson and Klindt-Jensen 1966, 145–6), and this provides the chronological limits for the English pieces, although it is possible that these can be more closely dated. Ringerike-style elements occur in four pre-Conquest manuscripts: in Aelfric's Pentateuch, dated to the second quarter of the eleventh century (Temple 1976, no. 86; Fuglesang 1980, no. 108, pls. 68–9); the Winchcombe Psalter, dated to c. 1030-50 (Temple 1976, no. 80; Fuglesang 1980, no. 109, pl. 70); the Caedmon Genesis, dated to c. 1000 (Temple 1976, no. 58; Fuglesang 1980, no. 110, pl. 71); and in the Bury St Edmunds Psalter, dated to the second quarter of the eleventh century (Temple 1976, no. 84; Fuglesang 1980, no. 108, pls. 68–9). This suggests that in England at least the high point of the style was in the first half of the eleventh century. Confirmation of this suggestion derives from the distribution of objects decorated in the Ringerike style which is firmly south-eastern, centred on the political heart of the kingdom, on the axis from Winchester to London. This suggests that the Ringerike style in England was a court style and, as such, probably relates to the period of Danish supremacy under king Cnut and his immediate successors, from 1016–1042. As Wilson has pointed out, it is unlikely that the style would have found favour with the Francophile court of Edward the Confessor (Wilson and Klindt-Jensen 1966, 145).


Most of the cross-shafts from south-east England of tenth- and eleventh-century date are of square or rectangular section, with the exceptions of those from Canterbury, Kent (St Augustine's abbey no. 1) and Wantage, Berkshire, which are round.

The most elaborate of the square-sectioned shafts is that from All Hallows by the Tower, London (no. 1; Ills. 320–40). This is decorated with a mixture of figural ornament, interlace, animal, and plant decoration. Figures occur on three of the four faces, but extensive damage has made the iconography difficult to interpret. On faces A and C are single, large standing figures. On face A only the shoulders and the lower part of the robes and legs survive (Ills. 323–4). On face C more of the figure survives, but it is so eroded that the features are difficult to decipher. Radford has referred to animals at the figure's feet and has thus identified the scene as Christ trampling on the beasts (Kendrick and Radford 1943, 15), but this appears to be a misreading of the figure's legs which emerge from its tube-like robe (Ill. 339). The figures on face D are more complete and in much better condition. Two standing or sitting figures are placed side by side (Ills. 336–8). The left-hand figure holds a key and can be identified as St Peter. The right-hand figure also holds something vertically; Radford has identified it as a sword and the figure, therefore, as St Paul (ibid., 15–16). This is unlikely as a sword is only depicted once in Anglo-Saxon art in association with St Paul, on fol. 95v of the Benedictional of St Aethelwold (Temple 1976, no. 23). Here it occurs in the scene of the execution of St Paul and is carried by the executioner, not by the saint. Elsewhere, as on fol. 31r of the Hereford Troper (ibid., no. 97; Rice 1952, pl. 67b) and on fol. 29v of the Benedictional of Robert of Jumièges (Temple 1976, no. 24, frontispiece), the saint is depicted as bearded and he holds a book. It is equally possible that the feature held by the figure on the All Hallows shaft is a sceptre or scroll, not a sword.

The robes of the left-hand figure are arranged in a pronounced V-fold between the legs, and the robe appears to be gathered up from left to right (Ill. 338). This disposition of the robes is more characteristic of seated than standing figures, as are the figure's out-turned feet. It is, therefore, possible that this pair of figures derives from a model in which two figures were seated side by side. Such scenes are uncommon in Anglo-Saxon art, but ranges of seated figures do occur in some eleventh-century manuscripts, as, for example, on fol. 2v of BL MS Cotton Tiberius A. III, a work of c. 1050 (Temple 1976, no. 100, ill. 313), and on fol. 56v of Durham Cathedral Library MS B. III. 32, a work of similar date copied from it (ibid., no. 101, ill. 315). In the latter, the three figures in Tiberius A. III (king Edgar and saints Dunstan and Aethelwold) have, with the omission of the king, been reduced to two, as on the All Hallows shaft. In both manuscripts the figures' garments fall in V-folds between the legs and in most cases the feet are out-turned. One unusual detail which links the scene in the Durham manuscript with that on the All Hallows shaft is the use of horizontally-ribbed hose with Classical garb, worn by both of the figures on the shaft, and by the figure to the left in the manuscript. King Edgar in Tiberius A. III has his legs clad in similar hose, as is often the case with figures wearing secular dress in tenth- and eleventh-century art. Something similar is seen, for example, on fol. 17v of BL MS Cotton Tiberius C. VI, dated to c. 1050 (ibid., no. 98, ill. 306), and in BL MS Stowe 944, a work of c. 1031 (ibid., no. 78, ill. 244), where the figure of king Cnut on fol. 6r wears comparable hose.

The manuscript parallels for the iconography of the scene on face D suggest a date for the shaft in the eleventh century and supporting evidence derives from the style of the figures and their draperies. The figures on face D are unnaturally elongated, as apparently are the figures on faces A and C, although the extent of the damage to these faces makes it difficult to judge. Such thin and elongated figures are paralleled in late pre-Conquest art, most notably in the manuscripts, as in the Judith of Flanders Gospels, a work of the second quarter of the eleventh century (Temple 1976, no. 93, ills. 285, 289), in Oxford, Bodleian Library MS Douce 296, a work of similar date (ibid., no. 79, ill. 259), and in Monte Casino, Archivio della Badia MS BB. 437, 439, made in c. 1050 (ibid., no. 95, ill. 287). The figures in these three works are all characterised by their very narrow shoulders, long torsos in proportion to the overall length of the body, and by the lack of a waist; indeed so similar are they that it has been suggested that these manuscripts derive from the same workshop (ibid., 112). All these features are paralleled on the All Hallows shaft, particularly on face D.

The handling of the robes on face D also points to an eleventh-century date for the shaft. The robes have the folds indicated by narrow, closely-spaced grooves, often arranged as a series of nested U-shapes, as over the legs of the figures, or in apron-like folds as at the waist of the left-hand figure (Ills. 226–8). The effect is rigid and non-naturalistic, concentrating on surface patterning not on the shape of the body beneath. In manuscript art this treatment finds its closest parallel in Aelfric's Pentateuch, a product of St Augustine's abbey, Canterbury and dated to the second quarter of the eleventh century (Temple 1976, no. 86, ills. 265–72). Here also the folds of the robes are indicated by closely-spaced lines which are rigid and formalised and have begun to loose contact with the shape of the body beneath. Frequently the lines are grouped in nested U-shapes, as on the cloaks of the figures on fol. 139v (ibid., ill. 266). Certainly this effect of surface patterning with line is as closely comparable to the effect on the All Hallows shaft as the difference in medium will permit. A sculptural example of the same phenomenon occurs on the eleventh-century cross-shaft from Shelford, Nottinghamshire (Kendrick 1949, 78–9, pl. LI). Here the figures of the Virgin and Child on face A and a seraph on face C have the folds of the garments indicated by narrow, closely-spaced grooves as on the All Hallows shaft. The grooves are placed in groups at angles to each other, but paralleling different edges of the garments, a patterning which is more formalised than that of the All Hallows shaft. The emphasis on the knees and the patterning of the lower parts of the garments at Shelford suggests that, as at All Hallows, the standing figures were derived from a model where the figures were seated.

The figure of the Virgin at Shelford has the overgarment indicated by narrow closely-spaced grooves, and the undergarment, visible below the knees, indicated by widely spaced vertical grooves. This difference in treatment provides a crucial clue for the understanding of the treatment of the robes of the figure on face A of the All Hallows shaft. Here the upper part of the robe, above knee level, has narrow vertical grooving; the lower part splays out and is formed from a mass of folded vertical edges. The two elements are separated by a pair of transverse bars (Ills. 323–4). This appears to be a further conventionalisation of the arrangement of the overgarment and undergarment seen at Shelford. A similar conventionalisation, albeit geographically distant, occurs on an eleventh-century cross-shaft from Aycliffe, co. Durham, where on face D the crucified figure of St Peter has the overgarment and undergarment arranged in precisely this manner. The Aycliffe shaft can be dated to the last quarter of the tenth century or first quarter of the eleventh century (Cramp 1984, i , 41–3, ii , pl. 8 (28)).

An analysis of the iconography and figure style of the All Hallows shaft suggests a date for the piece in the eleventh century, and possibly in the middle years of the century. Little is added to such a conclusion by a consideration of the interlace, animal and plant ornament on the shaft. The simple forms of the interlace on face B exhibit no feature which would permit close dating. The paired animals on face A find no adequate parallel in south-east England in late pre-Conquest art, but the reversed lentoid eye and the elongated ears and lips with clubbed ends may be derived from Scandinavian art, as on the grave-marker from St Paul's, London (Ill. 351). The body of the animal, however, is very different. It is almost parallel sided and the legs are exceptionally long; a pair of transverse incised lines separates the hindquarters from the rest of its body. These are not usual features in late pre-Conquest art where the animal bodies usually have a triangular shape, as on Canterbury St Augustine's 2. The transverse lines across the hindquarters may, however, be derived from the tail brought up across the hindquarters, a feature found widely in late pre-Conquest art, as, for example, in Monte Cassino, Archivio della Badia MS BB. 437, 439 on p. 127, or in metal on the censer cover from Canterbury (Wilson 1964, no. 9, pl. XII).

The plant ornament on face D is, however, slightly more informative. The spindly, ridged stalks set at angles to each other and interlacing where they cross (Ill. 336) seem to derive from the acanthus borders of Winchester-style manuscripts. As Kendrick has noted (Kendrick 1949, 103, fig. 15), in several of these the acanthus leaves cross and intertwine in precisely the fashion seen here, notably on fol. 11v of BL MS Cotton Vitellius C. III (Temple 1976, no. 63, ill. 187), on fol. 10v of Cambridge, Pembroke College MS 301 (ibid., no. 73, ill. 234), and on fol. 62r of Rome, Biblioteca Apostolica MS Reg. lat. 12 (ibid., no. 84, ill. 262), all manuscripts of the first half of the eleventh century. On the All Hallows shaft the fleshy acanthus leaves with their lobed ends employed in the manuscripts have been reduced to rigid stalks, but their ancestry is not in doubt. In sculpture the closest parallel is provided by the acanthus string-courses or friezes from Sompting, Sussex (nos. 1–3, 5, and 7), which employ similar interlacing ridged stems (Ills. 162–3, 167–9, 170–1). The Sompting friezes can be dated to the mid eleventh century, a date which on all the available evidence would suit the All Hallows shaft.

Such a mid eleventh-century date conflicts with the dating of the inscription based on its letter forms; Higgitt suggests in the catalogue entry that this is best placed in the ninth century. This is a difficult problem to resolve. As Higgitt points out, the use of early letter forms on a late monument cannot be paralleled elsewhere, although it is possible. Equally, the decoration on the shaft is very difficult to parallel among the surviving material before the eleventh century. There is nothing similar in the eighth or ninth centuries in south-east England (where early sculptures are in any event rare) or, indeed, from elsewhere in the country.

Like the All Hallows shaft the larger of the two cross-shaft fragments from Southampton, Hampshire (Ills. 459–64), employs figural ornament, but its fragmentary nature makes both iconographical and stylistic analysis, and therefore dating, difficult. A pre-Conquest date is suggested by the type of monument; standing crosses are rare in south-east England after the Conquest, although examples do exist as at Castle Hedingham, Essex (Brown 1903–37, vi (2), 147–8, pl. XL.2). Such a date is supported also by the layout and nature of the ornament.

Face A of the Southampton shaft (Ill. 459) is divided into two fields, the upper larger than the lower, by a horizontal bar which at each end protrudes into a pier with a stepped base flanking the edge moulding. This type of layout is paralleled in a number of late tenth- and eleventh-century manuscripts, for example, in the Nativity scene on fol. 85r of the Aethelstan Psalter, dated to the second quarter of the tenth century (Temple 1976, no. 5, ill. 30); on fol. 2v of BL MS Cotton Tiberius A. III, a manuscript dating to c. 1050 (ibid., no. 100, ill. 313); and in the illustration prefacing Aelfric's Grammar on fol. 56v of Durham Cathedral Library MS B. III. 32, a Canterbury manuscript of c. 1050 (ibid., no. 101, ill. 315). The closest parallel is, however, provided by the St Luke Evangelist portrait on fol. 21v of Queen Margaret's Gospels (Oxford, Bodleian Library MS lat. lit. F. 5 (ibid., no. 91, ill. 279)). Here the figure of the Evangelist is contained within an arch whose supporting piers abut the frame and have stepped bases. These rest on a cross-bar dividing the page into two unequal parts. If the piers were extended to stand on the lower border then this layout would provide a precise parallel for that on the Southampton shaft. The area beneath the Evangelist is filled with a representation of the ground, but figures related to the scenes in the upper fields are found in analogous positions in other manuscripts, as possibly on the Southampton shaft. In both BL MS Cotton Tiberius A. III (ibid., no. 100, ill. 313) and Durham Cathedral Library MS B. III. 32 (ibid., no. 101, ill. 315) there is a monk with a scroll, and in the Aethelstan Psalter the scene in the lower field represents the midwives washing the infant Christ (ibid., no. 5, ill. 30).

The identification of the scene or scenes on the Southampton shaft (Ill. 459) remains entirely speculative. In the upper field only the feet of a figure survive, projecting from a robe and standing on a T-shaped support. A drapery fold falls to the left. This might be interpreted as a robed crucifixion with Christ's feet standing on a suppedaneum; as an Evangelist portrait with an unusually-shaped footstool; or as the standing figure of a saint. For example, the figures of the Virgin and St John flanking a cross on a fragment from Halton, Lancashire, stand on chalice-like features which may be related to the support on the Southampton shaft (Collingwood 1927, 159–60, fig. 191). No surviving example of any of these three scenes depicts animals below the feet of the figure, however. The animals on face A of fragment a can be identified as a sheep and a predator of some kind, probably a lion. On fragment b, face A depicts an eagle holding a book (Ill. 458), the symbol of St John the Evangelist, with on face B an unidentified animal lying down (Ill. 457).

The iconographical evidence, therefore, advances little the dating of the Southampton shaft, and the stylistic evidence is almost equally uninformative. The falling fold of drapery to the left of the figure on face A of fragment a suggests a relationship with works of the Winchester style which often employ similar narrow folds of drapery developing extravagantly and non-naturalistically from the robes of the figures, as for example, on fol. 1r of the St Dunstan Classbook (Oxford, Bodleian Library MS Auct. F. 4. 32), dating to the mid tenth century (Temple 1976, no. 11, ill. 41), and on fol. 16r of BL MS Cotton Tiberius C. VI, dated to c. 1050 (ibid., no. 98, ill. 10). The folds on the Southampton shaft lack the characteristic broken profile of folds of the Winchester style and are harder and more mechanical (Ill. 459). This may reflect a difference in style dictated by the change in medium, or, alternatively, that the Southampton shaft belongs in the latest period of pre-Conquest art when in the manuscripts also draperies tend to be more subdued and regular without the exaggeration characteristic of earlier works as, for example, in Aelfric's Pentateuch (ibid., no. 86, ills. 269–72) and BL MS Cotton Tiberius B. V (ibid., no. 87, ills. 273–6), both works of the second quarter of the eleventh century. Both of these manuscripts also contain depictions of naturalistic sheep closely comparable with that on face A of fragment a at Southampton, for example, in the scene of the disembarkation of Noah on fol. 15v of the Aelfric Pentateuch (ibid., ill. 269) and on fol. 5r of Tiberius B. V (ibid., ill. 273). In both scenes the sheep have the same head shape and curved horns depicted on the Southampton shaft.

A late pre-Conquest date for the Southampton shaft is supported by the form of the leaf decoration on face D of fragment a (Ill. 464) and face A of fragment b (Ill. 458). In both cases this consists of a series of closely-spaced, expanding, rounded-ended leaves forming a palmette. A leaf form incorporating this feature is encountered in manuscript art only in the eighth-century Vespasian Psalter, where on fol. 30v the bush-scrolls in the spandrels of the arch head have similar leaves developing from their undersides (Alexander 1978, no. 29, ill. 146). Something similar is encountered on Reculver 1e, Kent (Ills. 119–20), dated above to the early ninth century. Apart from this the motif occurs most commonly in metalwork of the Trewhiddle style (Fig. 24), for example, on the ninth- or tenth-century sword-pommel from the River Seine, Paris (Wilson 1964, no. 66, pl. XXIX); on an early tenth-century disc brooch found in Stockholm (Backhouse et al. 1984, no. 17, pl. 17); on the hilt of the late ninth or early tenth-century Abingdon sword (Hinton 1974, no. 1, figs. pp. 2-3, pls. I–III); and on the ninth-century Fuller brooch where the sprays held by the personification of Sight have a similar form (Wilson 1964, no. 110, pl. XLIV). A cruder version of this ornament occurs on a disc brooch from Winchester which copies a penny of Edward the Elder, 899–924 (Okasha 1971, no. 141, pl. 141). All that has occurred on the Southampton shaft is that the stem of the palmette has been assimilated into the border.

The iconographic and stylistic parallels for the Southampton fragments, therefore, suggest a date in the late pre-Conquest period. The style of the draperies and the handling of the animals points to an eleventh-century date, but the form of leaf decoration to the late ninth or tenth century. So little of the decoration survives, however, that no convincing conclusion can be drawn.

Of the eight remaining shafts of square section, from Bishops Waltham and Wherwell, Hampshire; Barking (no. 1) and Saffron Walden, Essex; Kingston upon Thames and Reigate, Surrey; Lavendon (no. 1), Buckinghamshire; and Stanbridge, Bedfordshire, only that from Bishops Waltham is relatively intact, the piece forming one complete section of a shaft constructed from separate pieces. The decoration consists predominantly of panels of interlace. The patterns are well worked out and elegantly drawn, but lack the fine strand and mechanical perfection of eighth- and ninth-century sculptured interlace from south-east England. This, together with the use of a median incision which does not occur on the earlier material from south-east England, may suggest a late date for the shaft, but such flimsy evidence is not a sound basis for dating. Of much more importance in dating is the animal ornament on face C of the shaft (Ill. 421). This consists of a pair of tightly coiled, confronted, ribbon-like animals, one placed upside-down with respect to the other and with their appendages forming interlacing diagonals. There is no parallel for this arrangement among the sculptures of south-east England, but a tenth-century shaft from Gloucester has in one panel a single ribbon-like animal with the body in tight coils (Fig. 25b). Its appendages form interlacing diagonals and end in feet having a fan-like shape, with the toes clearly differentiated; a treatment which compares closely with that of the tails of the animals on the Bishops Waltham shaft, but at Bishops Waltham the single animal at Gloucester has been split into two (Ill. 421). An analogous arrangement of two tightly curled animals with one placed upside-down can be traced in manuscript art where in a number of manuscripts the decorated initial S is formed in this way. The closest parallel to the Bishops Waltham shaft is provided by the initial on fol. 2r of Cambridge, Corpus Christi College MS 23, a work of the late tenth century, probably made at Christ Church, Canterbury (Temple 1976, no. 48, ill. 50). A similar arrangement occurs on fol. 80v of Lambeth Palace Library MS 200, a work of the late tenth century, produced at St Augustine's abbey, Canterbury (ibid., no. 39, ill. 131). In both cases the detail is very different from that on the Bishops Waltham shaft, but the overall arrangement of the design is strikingly similar.

The tenth-century date suggested for the Bishops Waltham shaft by these manuscript parallels and by comparison with the Gloucester shaft is reinforced by comparison with the animals on the shaft and those on the cover from Bexhill, Sussex, for which a tenth-century date is argued above. The animals on Bexhill (Ills. 18–19) are not coiled, but they have the same ribbon-like bodies as the animals on the Bishops Waltham shaft (Ill. 421) and divided, fan-like tails allied with a similar head shape, with a domed forehead and open, elongated, narrow jaws.

The proposed tenth-century date of the Bishops Waltham shaft is of assistance in dating the shafts from Wherwell, Hampshire, and Barking 1, Essex, both of which, although damaged, are decorated only with interlace. At Wherwell (Ills. 479–81) the patterns are well drawn, but the strands are thick and, in one case, median-incised. At Barking (Ills. 256–9) the strands are finer, although the designs are not well worked out and often become incoherent. The thickness of the strand at Wherwell and the incoherence of the pattern at Barking may indicate a late date. These features are not sufficient in themselves for secure dating, but comparison with the Bishops Waltham shaft at least confirms that interlace-decorated shafts of square section were being made in the tenth century and using the same types of interlaces as at Wherwell and Barking (no. 1).

The shaft from Kingston upon Thames, Surrey, may also have been decorated purely with interlace, but, as only a small part of the shaft survives, with decoration on only two faces, it is impossible to be certain. On face D there are at least two elements of an encircled pattern C apparently incorporating free rings (Ill. 95). The pattern on face B was apparently similar (Ill. 96). The use of free rings in the interlace suggests a date in the tenth or eleventh century, when this feature appears to have originated (Collingwood 1927, 65, 68), and such a late date is supported by the flaccid and disorganised laying-out of the pattern.

Of the shaft from Reigate, Surrey, only a small fragment from one corner of a face survives, decorated with part of what was probably a plain plait with broad, low-relief strands (Ill. 140). Interlace of this type is not encountered among Anglo-Saxon sculptures of the seventh to ninth centuries, but does occur in the tenth and eleventh centuries, especially in northern England, for example, at Sockburn, co. Durham (Cramp 1984, i , 138–9, ii , pl. 136 (734–6)), Aycliffe, co. Durham (ibid., i , 46–7, ii , pl. 14 (58, 60)), and Middleton by Pickering, Yorkshire (Lang 1991, ills. 676, 678). On this (admittedly slender) evidence a tenth- to eleventh-century date can be suggested for the Reigate fragment.

The fragment from Lavendon, Buckinghamshire (no. 1), can be identified as part of a cross-shaft because of its taper. The decoration on the only visible face consists of a pair of interlaced closed-circuit loops. The strands are median-incised (Ill. 313). The simplicity of the decoration and the use of closed-circuit elements suggest a late pre-Conquest date. The second fragment from Lavendon (no. 2) may have been part of the same or a different shaft, but too little survives for its function to be firmly established. Like Lavendon 1, it is decorated with a closed-circuit pattern (Ill. 314), which may suggest a late date. The two shaft fragments from Saffron Walden, Essex, may well have formed part of the same monument. They are simply decorated, but this time with geometrical ornament consisting of concentric semicircles along each of the long edges and touching along the vertical axis of the face (Ills. 371–3). This type of decoration very closely resembles that on the grave-cover from Oxford cathedral (Ill. 362) and a similar mid to late eleventh-century date can be suggested. The decoration on the shaft from Stanbridge, Bedfordshire, is even simpler, consisting of an incised Greek cross (Ill. 409). In view of its great size and of the paucity of post-Conquest standing crosses in south-east England, a pre-Conquest date can be suggested, but there is no decisive evidence.

The two remaining cross-shafts from south-east England, from Wantage, Berkshire, and Canterbury, St Augustine's abbey 1, Kent, are different varieties of round shafts. The shaft from Wantage (Ills. 474–7) is of circular section, a form which, as noted above, appears to have been a ninth-century innovation. Like the shafts from Bishops Waltham and Wherwell in Hampshire, and Barking 1 in Essex, it is decorated purely with interlace; as at Barking and Kingston upon Thames, Surrey, an encircled pattern is employed. The strands are thick and median-incised, a form encountered at Bishops Waltham and Wherwell. The parallels which can be drawn between the Wantage shaft and the similarly-decorated shafts of square section serve to suggest a similar tenth- or eleventh-century date. The shaft from St Augustine's abbey, Canterbury, is of the type where the upper part of the shaft is of square section and the lower part of circular section (Ills. 20–3). As noted by Kendrick, this type is most common in north-west Mercia, although outliers of the group are known from Cumbria, Denbighshire, Yorkshire, and Dorset (Kendrick 1949, 68). Kendrick has proposed a late Anglo-Saxon date for the whole series, deriving the type from ninth-century monuments, such as the shaft from Dewsbury, Yorkshire, which also appears to have had an upper part of square section and a lower part of circular section (ibid., 72–3). His late dating of the group is underlined by the distinctive decoration of the Gosforth cross, Cumberland, the most important and imposing surviving monument of this type. This employs, among other motifs, scenes from Norse pagan myths and a Borre-style ring-chain. Taken together these would suggest a tenth-century date for the shaft and this provides a cross-check for the dating proposed by Kendrick for the group as a whole (Bailey 1980, 125–31, fig. 23; Wilson 1976, 502). A tenth/eleventh-century date would be equally acceptable for the St Augustine's shaft which, like other shafts of this date from south-east England, is purely interlace decorated, although the patterns are too fragmentary for reconstructions to be attempted.


Four cross-heads of late Anglo-Saxon date survive from south-east England, at All Hallows by the Tower 2 and St John Walbrook in London; South Leigh, Oxfordshire; and Pagham, Sussex. Of these the heads from Pagham and South Leigh are ring-heads. That from All Hallows 2 is a circle-head, and the head from St John's, Walbrook, is a plate-head (Cramp 1991, p. xiv, fig. 3). As noted above, both the plate- and circle-heads are types closely related to the ring-head. Based upon the distribution of the ring-heads which is predominantly northern and eastern, Collingwood has suggested that the type was introduced by Scandinavian settlers in the early tenth century and has pointed to the Isle of Man or Ireland as a source (Collingwood 1927, 137–9, fig. 153). As Bailey has observed, the dating of ring-heads depends not only on their distribution, but also on the fact that no cross-head of this type in England occurs with ornament of pre-Viking type, and that examples do occur which employ animal, figural, or interlace ornament deriving from Scandinavian or Scandinavian-derived motifs (Bailey 1978, 178–9). All these strands of evidence serve to place the four south-eastern cross-heads in the tenth or eleventh centuries.

Confirmation of the dating for the South Leigh cross-head derives from the archaeological evidence for the dating of an almost identical cross-head from Glastonbury Tor, Somerset. This came from a context which contained pottery of a type current in Somerset from c. 1000 to c. 1200 (Rahtz 1971, 31, 48, fig. 21). This evidence accords well with the angular form of the heads from both Glastonbury and South Leigh (Ill. 406), a feature which Collingwood has suggested is an indicator of a late date. He has, for example, suggested a post-Conquest date for the cross from St Crux, York, which is of similar angular form (Collingwood 1927, 93–4, fig. 15). This is confirmed by the occurrence of crosses of comparable angular form carved in relief on some of the head- and foot-stones of late twelfth-century date from the Canon's cemetery at Old Sarum, Wiltshire.

Little is contributed to a dating of these four cross-heads by a discussion of their decoration. The head from South Leigh is undecorated (Ill. 406), while that from Pagham (Ills. 98–100) is ornamented only with flaccid and disorganised interlace. This may confirm a tenth/eleventh-century date for the cross-head, but does not refine the dating further. The All Hallows head has a memorial inscription around the edge of face A (Ills. 343–4), but the majority of the decoration was painted and cannot now be reconstructed from the surviving colour. The cross-head from St John Walbrook has a simple mixture of pellets and roll mouldings which is not distinctive enough for reliable comparisons to be drawn (Ills. 347–8).


There are eleven fragments which are so small or incomplete as to defy even tentative classification. Four from Selsey, Sussex; two from Lewknor, Oxfordshire; one from Oxford (St Aldate's); two from St Augustine's abbey, Canterbury (nos. 9 and 11); one from Ford, Sussex, and one from Rochester, Kent (no. 4).

One of the Selsey fragments (no. 3) is decorated with incoherent, disorganised interlace for which a late date might reasonably be suggested (Ill. 159). A second (no. 4) is decorated with a six-strand median-incised plain plait (Ill. 158). The strands are fairly rounded and the closest parallel is provided by the interlace on the chancel arch from Selham, Sussex, for which a twelfth-century date might be suggested. However, the parallel is not exact, as the strands at Selham are finer and even more rounded, so an earlier date for this fragment is possible. The third fragment (no. 1) is decorated with interlace based on a closed-circuit pattern with some affinities to the Borre-style ring-chain (Ill. 161). Again a tenth- or eleventh-century date seems plausible. The last fragment (no. 2) is decorated with incoherent interlace which develops into a three-element leaf (Ill. 160). The latter is best paralleled in the tenth century; similar leaves occur, for example, on the shaft from East Stour, Dorset (Cramp 1975, fig. 19j). However, the Selsey piece is too weathered for detailed comparisons to be drawn.

The two fragments from Lewknor, Oxfordshire, are both decorated with figure-of-eight interlaces (simple pattern F) (Ills. 316–17). This decoration suggests, but does not prove, that they derive from a grave-cover of a type common in Lincolnshire, where the upper surface is covered in similar interlace. Examples of this type in Lincolnshire are known from Cammeringham (Davies 1926, 9, pl. VII), Northorpe (ibid., 17–18, fig. 3) and Stow (ibid., 19–20, fig. 4), where they are dated to the late pre-Conquest period.

A slab from St Augustine's abbey, Canterbury (no. 9), has interlace decoration, apparently consisting of surrounded pattern D (Ill. 53). The piece has no taper and a convex upper face, but it may have formed part of a grave-cover. The thick strand of the interlace may suggest a late date, but this is tenuous evidence indeed for dating. The runic inscription from St Augustine's (no. 11; Ill. 58) may have formed part of a monumental sculpture. Enough of the runes survive to demonstrate that they are of Old Norse character and the piece can probably be dated to the period of the Scandinavian supremacy in southern England, c. 1016–42.

The fragment from Ford, Sussex, is decorated with two groups of nested ovoids at right angles to each other and interlacing where they cross (Ill. 78). The pattern resembles that used at Dover, Kent (St Mary in Castro no. 4; Ill. 69) and Lavendon 1, Buckinghamshire (Ill. 313), for which tenth- or eleventh-century dates have been suggested, and a similar date would be appropriate for this piece. The piece from Rochester, Kent (no. 4) is also extremely difficult to date. It is decorated with an animal flanked by interlace, but very little of the animal survives, only part of the hindquarters and body. The body is thin with bulbous, exaggerated thighs, and the tail is looped around the body (Ill. 146). In southern English art, perhaps closely comparable animals are on fol. 11r of the mid eighth-century Stockholm Codex Aureus. These have similar, exaggerated thighs and thin bodies, although they are enmeshed in interlace, a feature not found on the Rochester piece (Alexander 1978, no. 30, ill. 152). Similar animals are, however, encountered in late Anglo-Saxon art, as, for example, on fol. 1v of the Vita Cuthberti, dated to c. 937 (Temple 1976, no. 6, ill. 29) and on fol. 57r of Durham Cathedral Library MS A. IV. 19, dated to the early eleventh century (ibid., no. 3, ill. 8). It is true that these later animals tend to have more triangular bodies, with the tail brought up between the legs, but these are not consistent features, and it must be frankly admitted that not enough evidence remains for the Rochester carving to be accurately placed.

Next Back to catalogue index