THE SEVENTH CENTURY
There are seventeen pieces of sculpture from south-east England which may belong to the seventh century or earlier, excluding some excavated fragments from Winchester which may be as early as this (see Chap. VIII): two memorial stones from Sandwich, Kent, two columns from Reculver, Kent (no. 4), a column base from St Pancras's, Canterbury; nine baluster-decorated fragments from the excavations at the Old Minster, Winchester; a similarly-decorated fragment from Barking abbey, Essex (no. 2); a baluster fragment from St Augustine's, Canterbury (no. 8); and an inscription from St Martin's, Canterbury.4
Possibly the earliest of these five sculptures are the two tapering memorial stones of square section found between Sandwich and Richborough in Kent (Ills. 151–7). On no. 2, at the upper end, two faces are decorated with framed rectangular fields (Ills. 152, 154). On no. 1, at the upper end, two adjacent faces have their common long edge paralleled by pairs of incised lines (Ills. 156–7). On one of these faces (Ill. 156) is a runic inscription, commonly supposed to be a personal name, though the analysis proposed by Dr Parsons in the catalogue entry casts some doubt on this. The linguistic forms allegedly recorded in the inscription have previously been considered crucial for the dating of the pieces (Elliott 1959, 81; Evison 1960, 244). The revised reading by Parsons has cast doubt on some of these features, but an early date still seems probable on other grounds.
If, as is still possible, the inscription can be read as a personal name, then it is likely that both this stone and the closely related piece found with it performed a memorial function. In that case the absence of any Christian symbol or allusion is remarkable and unparalleled in any other complete memorial inscription from south-east England. It is possible that a cross at the beginning of the inscription has been obliterated by the weathering which also makes the first character difficult to read, but if not then this absence of Christian symbolism or allusion may suggest a date for the piece before the introduction of Christianity to Kent in 597, or shortly thereafter.
The form of the two Sandwich stones also provides support for this early dating. Memorial sculpture in the ninth, tenth, and eleventh centuries in south-east England falls into clearly defined categories: coffin lids, grave-covers, grave-markers, and small memorial crosses, discussed in detail in Chap. VII. The Sandwich stones bear little resemblance in form to any of these, although they were clearly used partially sunken into the ground in the same way as the grave-markers (Ills. 151–7). It is, therefore, probable that they belong to the period before the emergence of these well-defined groups, i.e., to the eighth century or earlier. Certainly the stones appear to copy wooden prototypes, a fact demonstrated by the tapering of their bases, a feature which can have had no functional purpose. The stones could not have been driven into the ground without shattering, and the tapering of the bases would not have eased the sinking of the stones into the ground; the labour saved in digging the holes would have been more than offset by the labour involved in shaping the stones. This tapering can best be explained if the prototype for these monuments was a squared wooden post with a sharpened end which was driven into the ground. A number of pagan Anglo-Saxon cemeteries in Kent do have graves with either a single post hole in association, as at Finglesham (Hogarth 1973, 113) or post holes in association with another structural feature, as at St Peter's, Broadstairs (ibid.). Some of these post holes may have accommodated the wooden prototypes of the Sandwich stones.
With the coming of Christianity and the consequent re-introduction of the tradition of stone building into south-east England, it might be expected that stone sculpture would also be re-introduced and widely used. This, apparently, was not the case. A large number of early Christian sites in south-east England is known from historical sources: Dorchester, Berkshire (Mayr-Harting 1972, 100, 117); Barking abbey and Bradwell-juxta-mare, Essex (Taylor and Taylor 1965–78,
At Reculver (no. 4) the surviving sculptures consist of a pair of columns with decorated bases and capitals (Ills. 123, 126, 128–38), originally supporting the triple arcade separating the nave from the chancel of the church (Ills. 124–5, 127). Although the church stands in the centre of a Roman fortress it is unlikely that the columns are directly reused Roman material as the form of both the capitals and the bases is difficult to parallel in the Classical world. The capitals are decorated with inclined fasciae above a necking with a projecting half-round moulding. This form is unparalleled among the surviving Romano-British capitals, and appears to be a version of the impost block placed above the capital proper (Blagg 1981, 52–3). This was an east Mediterranean innovation, first occurring in the fifth century, for example, in Ravenna at St Giovanni Evangelista of 423–34 and St Apollinare Nuovo of 490, and in Salonika at St Demetrios of c. 470 (ibid., 53). In Anglo-Saxon England there are similar capitals or bases from Ripon in Yorkshire, possibly from the seventh-century church built by St Wilfrid (Taylor and Taylor 1965–78,
The use of ornamented column bases is much more readily paralleled in the Roman world, and some Roman bases even have cabled mouldings as at Reculver, particularly on the lower torus. There are bases of this type in the Capitoline Museum and in the Palazzo dei Conservatori in Rome (Blagg 1981, 52–3). Few such bases survive in Britain and none of them closely resembles the Reculver bases (ibid., 53). Their profile, with collars and a square-cut torus, can, however, be paralleled on bases at St Apollinare in Classe, Ravenna of 532/6–549 (loc. cit.). As with the capitals, this suggests an east Mediterranean and post-Roman prototype for the Reculver columns (Blagg 1981, 53).
If the Reculver columns are not reused Roman material, then their use in a late seventh-century building might be assumed to provide a terminus ante quem for them. Unfortunately, the demolition of Reculver church in 1805 involved the complete destruction of the fabric virtually without record. It is thus now impossible to be certain that the columns were not a secondary insertion, something which has been suggested for the similar triple arcade carried on columns at St Pancras, Canterbury (Jenkins 1975–6, 4). Controlled excavation in the area of the chancel arch might have elucidated this problem, but the area was heavily disturbed in 1878 in excavations by Dowker (Dowker 1878, 259) and little, if any, of the stratigraphy can now remain intact.
In default of a stone-by-stone survey of the fabric, or of archaeological evidence, it is necessary to use the few drawings and sketches made during the demolition of the church to assess whether the triple arcade was a primary feature or an insertion. A drawing by Gandy (Society of Antiquaries MS, Red Portfolios, Kent, L–R, fol. 23r (Ill. 125)) shows that the columns supported arches turned in Roman tile or brick, and that the outer ends of the outer arches rested on walls made of coursed, small, rectangular stone blocks. Each of these had three bonding courses of Roman tile or brick, on the topmost of which rested the arch heads. The accompanying annotated ground plan notes that these walls returned to the east to form the north and south walls of the chancel. There is no indication of the materials used for the gable-end supported by the triple arcade. A drawing by Baynes (BL Add. MS 32370, fol. 106r (Ill. 127)) confirms that the north and south walls of the chancel were of the same build as the wall supporting the outer ends of the triple arcade, since the tile bonding courses run through both walls at the same level. However, in the chancel there was a fourth tile bonding course above the others and just below the eaves. If it could be demonstrated that this returned across the wall supported by the triple arcade and immediately above the arch heads, it would prove that there was no disturbance in the walling and hence that the triple arcade was a primary feature. Unfortunately, in the Baynes drawing this area is largely obscured by plaster and all that can be seen through the gaps in it are the same small, rectangular dressed stones seen elsewhere in the earliest walling, and at a level below the possible position of the tile bonding course. This suggests, but cannot prove, that the triple arcade was a primary feature as it is unlikely that it could have been inserted without leaving some trace. Combined with the other evidence noted above, this would date the decorated columns to the late seventh century.
The single surviving column base from the original four columns supporting the triple arcade at St Pancras's, Canterbury (Ills. 59–60), can also be assigned to the seventh century. As Blagg has pointed out, although this base is decorated only with mouldings, their handling does not correspond with that observed on Romano-British column bases; it is, therefore, very unlikely that they are reused Roman material (Blagg 1981, 50). The best parallels for column bases with low torus mouldings placed on tall plinths, the form employed at St Pancras's, are to be found in the east Mediterranean, where the type developed in the fifth and sixth centuries, as at St Demetrios, Salonika of c. 470, St Apollinare in Classe, Ravenna of 532/6–549, and Grado cathedral of 571–9 (loc. cit.). Excavation at St Pancras's suggests that the triple arcade was a secondary insertion, probably of seventh-century date (Jenkins 1975–6, 4). It seems likely that the columns were made specifically for this purpose at that time, imitating late Roman and early Byzantine prototypes from the east Mediterranean.
Other architectural sculptures of possible seventh-century date from south-east England derive from the excavations at Barking abbey, the Old Minster, Winchester, and St Augustine's abbey, Canterbury. From Barking abbey comes a corner fragment, probably from an impost, decorated on both surviving faces with plain balusters of rounded section (no. 2; Ills. 260–2). The material from Winchester Old Minster (nos. 37–40, 42, 44, 47–9) is similar, consisting of architectural fragments decorated with baluster friezes. The balusters (e.g. no. 37; Ill. 561) are plain, of semicircular section, and are widely spaced; in some cases (e.g. no. 47; Ill. 576) a pair of such friezes flanks a broad zone decorated with two undulating guilloche strands. All of these pieces derive from the demolition levels of the Old Minster, and are probably (though not certainly) as early as the seventh century in date. They have, moreover, been compared by the Biddles with the baluster-decorated friezes and fragments from Hexham, Northumberland and Jarrow, co. Durham, an analogy which tends to support an early date (and which would also apply to the Barking fragment). The comparison certainly appears sound, for a date in the late seventh or early eighth century has been suggested for these Northumbrian baluster-decorated friezes on the grounds that some of the balusters form parts of decorative schemes on readily datable monuments. In particular, the cross-head from Jarrow with a baluster-decorated edge (no. 9) has been dated by Cramp on the basis of its form and decoration to the first half of the eighth century (Cramp 1984,
This type of large, stone, lathe-turned baluster appears to have been derived from Gaul where there is a surviving fifth-century example from Nouaillé (Cramp 1965, 4, pl. 3). This accords well with the historical evidence for the building of the churches at Monkwearmouth and Jarrow recorded in the anonymous Life of Ceolfrith ((——) 1896, 390), and in Bede's Lives of the Holy Abbots (Bede 1896, 368), which confirm that Benedict Biscop imported masons from Gaul to oversee the works. There are no small-scale baluster-decorated friezes surviving from Gaul earlier than the ninth-century impost from Germigny-des-Prés (Cramp 1965, 4; Hubert et al. 1970, pl. 251), but there are small-scale balusters in wood, for example, on the seventh-century reading desk of Ste Radegonde (Hubert et al. 1969, pl. 23) and on the bed and chair from the Frankish prince's grave under Cologne cathedral (Werner 1964, 206, figs. 6–7).
The dating of the Northumbrian baluster-decorated fragments suggests a comparable late seventh-century or early eighth-century date for the Winchester examples, and in this context it may be significant that there was a large measure of Frankish influence on the West Saxon church in the seventh century. Agilbert, the second bishop of the West Saxons (650–60 or 664) was a Frank (Stenton 1971, 122), and Wine, his successor, had been consecrated in Gaul (loc. cit.). Less is known about the origins of Birinus, the first bishop of the West Saxons (634–650), but it is likely that he was of continental stock as he was consecrated by bishop Asterius of Milan (Mayr-Harting 1972, 100) and there is a surviving letter to him from Pope Honorius I urging him to preach in Britain (Stenton 1971, 117). This would make little sense if he were a native Anglo-Saxon.
To judge from the surviving fragments from Winchester Old Minster, large lathe-turned balusters of the Monkwearmouth/Jarrow type did not form part of its early decorative scheme, but at St Augustine's abbey, Canterbury, there is a single fragment of a similar baluster (no. 8; Ill. 50). This was discovered during the demolition of the nineteenth-century wall overlying the south side of the abbey church (Wood, pers. comm.), and can thus only be dated typologically. It is without taper and decorated with a roll moulding, now largely broken away, flanked by pairs of narrow grooves. This use of narrow grooving and the lack of taper are not features typical of later Anglo-Saxon balusters from the region. Those from this site (nos. 6–7; Ills. 41–9) and from St Mary in Castro, Dover (nos. 2–3; Ills. 64–7, 71–5), for example, all lack them. These features are, however, found consistently on balusters of the Monkwearmouth-Jarrow type. It seems reasonable, therefore, to assign the St Augustine's piece to the same date as the Northumbrian balusters, that is, to the late seventh or early eighth century.
Of possible early date is the dedication inscription from St Martin's church, Canterbury (Ill. 57), which is discussed in Chapter IX by John Higgitt. This is incorporated into a pre-Conquest doorway in the south wall of the chancel, a doorway which Taylor and Taylor have suggested is contemporary with the nave of the church, dated by them to the period 600–50 (Taylor and Taylor 1965–78,
The pattern of use of sculpture in south-east England in the seventh century, therefore, seems relatively clear. The bulk of the material was clearly architectural. It draws on both eastern Mediterranean traditions, as with the column bases from Reculver and St Pancras, Canterbury, and on Frankish traditions in the case of the baluster friezes from Winchester Old Minster and the baluster from St Augustine's Canterbury. In addition there are the two memorial sculptures from Sandwich. This mix of architectural and memorial sculpture is one which is exactly paralleled in early Northumbria (Cramp 1984,
THE EIGHTH CENTURY
Apart from the baluster friezes from the Old Minster and the lathe-turned baluster from St Augustine's abbey, which may all belong equally to the late seventh or early eighth century, there is only a small quantity of eighth-century sculpture from south-east England. The only possible candidates are: the cross-shafts from Bedford St Peter and Elstow, Bedfordshire; the shaft from Steventon, the fragments from Little Somborne, and the base from South Hayling, all in Hampshire; and in Winchester itself, some possible examples from the Old Minster (e.g. no. 75), and Upper Brook Street. Even with these pieces it is uncertain whether they belong to the late eighth or the early ninth century.5
The use of bipeds on faces A, B, and D of the Elstow shaft (Ills. 269–71) suggests a date for the piece after the middle of the eighth century, when this animal became the standard type. It is not employed in works of the seventh or first half of the eighth century, such as the Lindisfarne Gospels (Alexander 1978, no. 9, ills. 28–46) or the Vespasian Psalter (ibid., no. 29, ills. 143–6). In these works the animals are quadrupeds or birds, often conventionalised and contorted, but still recognisable. Bipeds predominate in the works of the mid to late eighth century, however, as in the Leningrad Gospels (ibid., no. 39, ills. 188–95) and the Barberini Gospels (ibid., no. 36, ills. 169–78), and this remained the case into the early ninth century. BL MS Royal I. E. VI, for example, makes extensive use of bipeds (ibid., no. 32, ills. 160–4; see Fig. 8).
The late eighth- or early ninth-century date suggested for the Elstow shaft by the use of bipeds is supported by the fact that on faces A and D the hindquarters of the animals develop into interlace (Ills. 269, 271). This feature appears to have been a mid eighth-century development, first observed in the Stockholm Codex Aureus, as, for example, on fol. 11r (Alexander 1978, no. 30, ill. 152), and one which became very popular in the late eighth and early ninth centuries in works in almost every medium: in manuscripts, as on fol. 16r of the late eighth-century Leningrad Gospels; in ivory, as on the rear of the Gandersheim casket (Beckwith 1972, 18, no. 2, pls. 10–13); and in metal, as on the Witham pins (Wilson 1964, no. 19. pl. XVIII) and on the nasal of the Coppergate helmet (Tweddle 1983a, 110–11, fig. 2; 1984, pls. on 14). Bipeds developing into interlace also occur in the repertoire of the late eighth or early ninth century Maaseik embroideries (Budny and Tweddle 1984, pl. Vb). In all of these examples the animals not only develop into interlace, but are enmeshed in it. The reduction and disciplining of the interlace, seen on the Elstow shaft, was apparently a development of the early ninth century. For example, in the early ninth-century Bible fragment BL MS Royal I. E. VI, on fol. 4r, the fields above the capitals of piers two and four of the Canon Table arcade are filled with paired bipeds very similar in form to those at Elstow, and also developing into interlace; the animal's bodies are interlocked, but they are rigidly separated from the interlace which, as at Elstow, fills the lower part of the field (Alexander 1978, no. 32, ill. 162). This more highly-disciplined disposition of the interlace persisted throughout the ninth century, as, for example, on works decorated in the Trewhiddle style. On them paired bipeds occur relatively infrequently, but single animals with their hindquarters developing into interlace are common. Again the separation of the animal from the interlace is usual, as, for example, on the largest of the brooches from Pentney, Norfolk, where the outer border of the brooch is decorated with eight fields each containing ornament of this nature (Wilson 1984, pl. 120).
This disciplining of the interlace suggests an early ninth- century date for the Elstow shaft, but is contradicted by the occurrence of spiraliform elements within the interlace itself. These argue for an earlier, eighth-century, date. Spiral ornament was popular throughout the eighth century in works in all materials, as, for example on fol. 30v of the Vespasian Psalter (Wilson 1984, no. 29, pl. 146), on fol. 150v of the Stockholm Codex Aureus (ibid., no. 30, pl. 147), on the rear of the Ganersheim casket (Beckwith 1972, no. 2, pl. 13), and on a fragmentary copper-alloy disc from Ixworth, Suffolk (Hinton 1974, no. 18, pl. VIII.18). It had, however, fallen out of fashion by the early ninth century (Budny and Graham-Campbell 1981, 11); spiral ornament is absent, for example, from both BL MS Royal I. E. VI and from almost the entire corpus of Trewhiddle-style metalwork. The only exception is the pommel from Fetter Lane in London, but this arguably belongs just before or at the very beginning of the Trewhiddle style, around 800 (Wilson 1964, no. 41, pl. XXIII). Certainly the best parallel to the spiraliform interlace on face D at Elstow (Ill. 271) is provided on fols. 78r and 119r of the mid to late eighth-century Leningrad Gospels (Alexander 1978, no. 39, ills. 193–4).
The general repertoire of the decoration of the Elstow shaft therefore, consisting of animal ornament and interlace, without plant ornament, suggests a date in the late eighth century, at a time before plant ornament, with the possible exception of vine- or plant-scroll, became popular in Anglo-Saxon art (Budny and Graham-Campbell 1981, 11). This suggestion is supported both by the form of the animals and by the use of spiraliform elements, as spirals seem to have dropped out of use in Anglo-Saxon art by the early years of the ninth century. Only the rigid separation of the interlace from the animal bodies on faces A and B argues in favour of an early ninth-century date, and this feature alone is not distinctive enough for much weight to be placed upon it.
Closely related to the decoration of the Elstow shaft is that on the shaft from Bedford (Ills. 265–7). Only two of the faces are now visible. One of them is decorated with interlace, but the other is ornamented with a pair of confronted winged bipeds with their hindquarters developing into interlace. This decoration is of the same type as that used on face A of the Elstow shaft (Ill. 269), although the animals are simplified and lacking in detail, and the workmanship is of much poorer quality. As with the animals on the Elstow shaft the interlace is highly disciplined, occupying only the lower third of the field; the animals themselves are not enmeshed or embedded in it. As argued above, all these features are indicative of a late eighth-century date, with only the highly disciplined nature of the interlace perhaps arguing for a slightly later date.
In addition to the cross-shafts from Elstow and St Peter's, Bedford, there are other possible eighth-century cross-shaft fragments from the region, all from Hampshire: one from Steventon, and one from Winchester (Upper Brook Street). In addition, the closely-related fragments from Little Somborne probably derive from a cross-shaft.
All of these sculptures share the same basic decorative repertoire; they are ornamented with paired or single ribbon-like animals with their contoured and hatched or pelleted bodies developing into, and embedded in, interlace. There are differences in detail. On the Steventon shaft (Fig. 9; Ills. 471–2) the animals on face A are confronted, with their necks and undulating bodies crossing symmetrically. The limbs are suppressed and the interlace develops from the animals' tongues. On face B each field is filled with a single animal body, which lacks both head and limbs and is enmeshed in interlace. The shaft from Upper Brook Street, Winchester, is heavily weathered (Ill. 683), but the single visible face is apparently decorated with two looped animal bodies, one developing from each end (Fig. 10a), and enmeshed in interlace. The fragments from Little Somborne are too fragmentary for the form or disposition of the animals to be established (Ill. 447), but their presence is clear, as is their relationship with the other Hampshire pieces noted above.
These pieces form part of a well-defined group of sculpture, first recognised by F. Cottrill (Cottrill 1935), whose distribution is firmly south-western. The principal examples are from Ramsbury and Colerne, Wiltshire; Shaftesbury, Dorset; Rowberrrow and West Camel, Somerset; Dolton, Devon (Figs. 10 and 11); and Tenbury, Worcestershire. These additional pieces greatly extend the repertoire of ornament employed on the Hampshire examples. At Colerne there are confronted animals with spiral hips (ibid., pl. XV); at Dolton there are addorsed winged bipeds (Reed 1935, 286–7, pl. XXVI, fig. 2); and at Dolton (loc. cit.) and West Camel (Cottrill 1935, pl. XVIII) there are paired animals viewed from above with their bodies crossing symmetrically. Animals viewed from above appear again at Tenbury (ibid. 1935, pl. XIV), and there is a leafless tree-scroll with a segmented stem at West Camel (ibid., pl. XVIII).
Wilson has suggested that the animals on the sculptures from Ramsbury, West Camel, and Colerne, with their ribbon-like, contoured, and textured bodies, are related to those of the Viking Jellinge style (Wilson 1984, 106). This view divorces these pieces from the closely-related examples discussed by Cottrill without argument, and fails to provide an historical context for the introduction of the Jellinge style into the heartland of Wessex. In fact, all of the features of the decoration of these shafts, and of the rest of the group defined by Cottrill, can be paralleled within the native Anglo-Saxon artistic tradition current in southern England. As noted above, the use of animals embedded in interlace was particularly popular in the eighth century, with the use of animals with their hindquarters actually developing into interlace, as at Steventon, Ramsbury, and Tenbury, being an innovation of the mid eighth century, employed for the first time in the Stockholm Codex Aureus. In contrast, by the early ninth century the animal and interlace were often rigorously separated, the interlace reduced, and the emphasis was placed upon the animal. The late eighth-century date which can be suggested for the group on these grounds is confirmed by the use of spirals on the joints and on the neck of the animals at Colerne. As noted above, the spiral was popular in the seventh and eighth centuries, but had virtually dropped out of use in Anglo-Saxon art by the early ninth century (Budny and Graham-Campbell 1981, 11). Equally suggestive is the appearance of plant ornament at West Camel, as this was an innovation of the mid to late eighth century (loc. cit.).
The late eighth-century dating for this group suggested by the overall type and disposition of the ornament can be confirmed by specific comparisons with works in other materials, particularly with the mid to late eighth-century Leningrad Gospels (Alexander 1978, no. 39, ills. 188–95, frontispiece). This uses the same general repertoire of ornament as the south-western group of sculptures: a small amount of plant decoration, as on fols. 12r, 12v and 13v (ibid., no. 39, frontispiece, ill. 188), and a mixture of single or paired winged bipeds or ribbon-like animals developing into, and embedded in, interlace. As at Dolton and Tenbury some of these animals are viewed from above, for example on fols. 12v, 16v and 16r (Fig. 12; ibid., no. 39, ills. 188 and 190). Almost all of the animals in the Leningrad Gospels have contoured bodies, but the effect is achieved by the use of colour and not line. Hatching of the bodies is rarer, but does occur, as on the bodies of the animals forming the arch heads and pier bases of the Canon Table on fol. 16r (ibid., no. 39, pl. 190), and on the body of the quadruped enclosed by the letters L and I of the Liber Generationis page, fol. 18r. As with the Colerne shaft, a number of the animals have spiral joints, either to the wing or legs, as on fols. 12r and 18r (Wilson 1984, pl. 110). In their disposition some of the animals on the Leningrad Gospels provide very close parallels to those employed on the south-western group of sculptures. For example, the paired winged bipeds on fols. 12r and 16r are relatively close in form to those on the Dolton fragments (Alexander 1978, no. 39, frontispiece, ill. 190), whilst the looping of the paired animal bodies from either end of a narrow field on fols. 12r and 12v (ibid., no. 39, frontispiece, ill. 188) provides a good parallel for the ornament on face B of the Steventon shaft (Ill. 472), and for the decoration on the Upper Brook Street shaft from Winchester (Ill. 683). In one field on fol. 12r the loop of the animal's body is clasped in interlace of almost identical form to that deployed in similar situations on the Steventon shaft. The Leningrad Gospels also make extensive use of paired animals whose bodies undulate regularly, crossing and recrossing, as on fols. 12v, 17r, and 17v (ibid., ills. 188, 191), a feature paralleled at Steventon, West Camel, and Dolton. All these comparisons suggest a broad contemporaneity between the Leningrad Gospels and the south-western group of sculptures.
By contrast, it is difficult to find an adequate parallel either in metal or ivory for the decorative schemes deployed on the south-western group of sculptures. They are clearly generically similar to those on a number of eighth-century works, such as the Gandersheim casket (Beckwith 1972, 118, no. 2, pls. 10-13), the Larling plaque (Wilson 1984, pl. 97), and the Witham pins (Wilson 1964, no. 19, pl. XXIII), but the animals on the sculptures are usually ribbon-like, rather than the bipeds which are the norm on all of these works, although one of the Witham pins does employ a ribbon-like animal viewed from above, as on the Dolton and Tenbury shafts. The only close parallel to the sculptures in either metal or ivory is provided by the decoration on the nasal of the late eighth-century Coppergate helmet (Tweddle 1983a, 110–11, fig. 2; 1984, pls. 14). This consists of a pair of confronted bipeds with undulating ribbon-like bodies regularly crossing and recrossing before developing into interlace, a form employed at Steventon, Colerne, West Camel, and Dolton. In addition the Coppergate animals have a spiral on the neck, a feature paralleled at Glastonbury, Somerset, and Colerne. The bodies of the Coppergate animals are contoured and hatched, a treatment which is difficult to parallel on other Anglo-Saxon metalwork, but which, as noted above, consistently occurs in works of the south-western group of sculptures. The only major difference between the decoration employed on the nasal and that of the sculpture lies in the tight disciplining of the helmet's decoration, something normally lacking on the sculptures. Unfortunately, this comparison does little to refine or support the late eighth to early ninth-century date advanced for this group of sculptures on the basis of the manuscript parallels, as the helmet itself is dated primarily by art-historical methods.
Another possible eighth-century piece from Hampshire is the cross-base from South Hayling, Hampshire (Ills. 465–9). Unfortunately, it has been savagely damaged by weathering and the decoration is difficult to decipher (see catalogue entry and Fig. 38). Given the condition of the piece, detailed comparative study and hence close dating is impossible; however, the ornament permits dating to within broad limits. In particular, the combination of spiral-based ornament and plant ornament suggests a late eighth- or early ninth-century date, a period, as noted above, when spiral ornament was dropping out of use (Budny and Graham-Campbell 1981, 11) and when plant ornament was making its first consistent appearance in Anglo-Saxon art (loc. cit.). This dating is reinforced by a consideration of the decoration on face C (Ill. 467). If this can be interpreted as a double-spiral animal then the form is both distinctive and unusual. It occurs in the Canon Table arch heads on fol. 16r of the Leningrad Gospels (Alexander 1978, no. 39, ill. 190), and in the late eighth-century Hereford Cathedral Library MS P. I. 2 on fol. 102r, (ibid., no. 38, ill. 199). Here the motif appears twice on the monogram INP, but in each case the elements of the double spiral are separated by the downstrokes of the letters. Nevertheless, the comparison with the South Hayling example is reasonably close.
THE NINTH AND EARLY TENTH CENTURIES
In the ninth to early tenth centuries, there was a modest increase in sculptural production, with nearly three times as many pieces attributable to this period compared with the eighth century.6<\sup> Moreover, whereas the eighth-century pieces from south-east England appear to belong to the periphery of sculptural traditions centred outside the region, in the ninth century there seems to be a resurgence of sculptural innovation and production centred in the region. Four of the sculptures probably from this period are from Hampshire, the round shafts from Priors Barton and High Street, Winchester, a grave-marker from Whitchurch, and a crucifixion panel from Romsey (no. 2). There are also twelve pieces from Kent: seven fragments of a round shaft from Reculver (no. 1) discussed in detail in Chap. VI, fragments of cross-shafts of square section from Preston by Faversham and Rochester 1, and three fragmentary capitals from St Augustine's abbey, Canterbury (nos. 3–5). In addition there is a stoup rim from Godalming, Surrey (no. 1), a grave-cover from Oxford (New Examination Schools 1) and a sundial from Bishopstone, Sussex.
Amongst the Hampshire sculptures the most varied in its decoration is the grave-marker from Whitchurch, Hampshire. This has a semicircular head which on face C is filled with an incised leafless tree-scroll with only a single pair of branches and confined within a semicircular border (Ill. 484). On face A is a recessed field containing a high-relief, half-length figure identified by its cruciform nimbus as Christ (Ill. 483). On the edge, running over the semicircular head of the stone is a framed memorial inscription (Ills. 482, 485–9). Originally the sculpture must have been accomplished, but it is now heavily weathered. With such a restricted repertoire of ornament, combined with extensive damage, dating is difficult, but, as noted above, the use of plant ornament suggests a late eighth-century date at the earliest; plant ornament was not used consistently in Anglo-Saxon art until that date, becoming common only in the ninth century (Budny and Graham-Campbell 1981, 11). The precise form of plant ornament used here (Ill. 484) occurs in manuscript art only in one of the Canon Table column bases on fol. 12v of the mid to late eighth-century Leningrad gospels (Alexander 1978, no. 39, ill. 188), although there is something similar within the upper element of the initial B on fol. 3v of the Leningrad Bede, dated to c. 746 (ibid., no. 19, ill. 83). However, neither of these provides a particularly close parallel to the Whitchurch scroll. To find something much more closely comparable it is necessary to turn to metalwork (Fig. 13). A firm ninth-century date for the Whitchurch piece is supported by a ninth-century pressblech disc from Hedeby (Fig. 13d; Capelle 1968, no. 72, taf. 25.1). This disc is decorated with a free-armed Anglian cross having in each of the re-entrant angles a semicircular field filled with interlace. The interlace is, however, almost identical in form with the scroll at Whitchurch and must originally have been derived from a similar leafless tree-scroll.7
The motif of the leafless tree-scroll used on the Whitchurch piece occurs again on a fragment of a round shaft from Winchester, High Street (Ills. 679–82). Here the plant has a stepped base and a single pair of median-incised branches, and as the shaft has been broken away above, there must originally have been several more pairs of branches. This type of continuous tree-scroll has a very wide date range, from the late eighth or early ninth century, as in the Barberini Gospels (Alexander 1978, no. 36, ill. 170) or on the Alstad mount (Bakka 1963, 37–40, figs. 32–3), to the late ninth or early tenth century as on the shaft from East Stour, Dorset (Backhouse et al. 1984, no. 23, pl. II). After this date the form persisted but was reclothed in acanthus foliage, as in the borders of the Presentation Scene of the New Minster foundation charter, for example (Temple 1976, no. 16, ill. 84). Leafless tree-scrolls are, however, relatively uncommon in southern England and, apart from the specialised version employed on the Whitchurch piece, the closest parallel to the scroll on Winchester High Street 1 is on the late eighth- or early ninth-century shaft from West Camel, Somerset (Cottrill 1935, pl. XVIII.2). Here the scroll is similarly without leaves and has a segmented stem. If the decoration of the shaft from High Street is ambiguous in terms of dating, then its round-shaft form is suggestive, for that appears to have been a ninth-century introduction (Cramp 1978, 9); it does not, for example, occur among the eighth-century pieces discussed above.
From Priors Barton on the outskirts of Winchester comes another fragmentary round shaft, in this case decorated with four vertical fields separated by plain relief mouldings on stepped bases, possibly originally supporting arch heads (Ills. 686–90). The use of acanthus foliage (Ill. 688) is significant and would, at first sight, suggest a tenth-century date for the shaft, since acanthus foliage does not appear in manuscript art until that date. However, the combination of animal, plant, and interlace ornament, with the different types of ornament alternating with each other, is more reminiscent of late eighth- and early ninth-century art. The embroideries of this date from the church of St Catherine at Maaseik, Belgium, for example, have a series of arches filled alternately with plant and animal ornament. The arch heads are decorated with alternating interlace and plant ornament, while the columns supporting the arch heads and the spandrels are filled with animal, plant, or interlace ornament varying in complex rhythm. These embroideries were probably made in southern England (Budny and Tweddle 1984). Similarly, the arcades framing the Canon Tables in the early ninth-century manuscript BL MS Royal I. E. VI are decorated with rectangular fields filled alternately with different types of ornament (Alexander 1978, no. 32, ills. 162–4). As on the Priors Barton shaft the plant ornament develops animal heads, although this feature is not confined to the ninth century, and is seen on the eighth-century Bjørke mount, for example (Bakka 1963, 12–15, fig. 8). This rhythmical alternation of different types of ornament occurs also in metal as, for example, in the borders of one of the pairs of early ninth-century disc brooches from Pentney, Norfolk. Here animal and plant ornament alternate as does the technique of the decoration. The plant ornament is in openwork and the animal ornament has the background inlaid with niello (Wilson 1984, pl. 120).
Disregarding the leaf form, the organisation of the partially defaced plant on the Priors Barton shaft, with the branches ending in a leaf and berry bunch, is also best paralleled in ninth-century art; in particular on the Mosnaes brooch, a Scandinavian type of trefoil brooch with the ornament based on Anglo-Saxon exemplar (Graham-Campbell 1980b, 438), although a tenth-century date is arguable for this piece (Wilson, pers. comm.). There is no precise equivalent to this arrangement which can be securely dated to the tenth century, although there is something similar on the cross-shaft from Littleton Drew, Wiltshire (Kendrick 1938, pl. LXXXIV 2–3). However, the combination of bush scrolls with berry bunches seen at Priors Barton can equally be paralleled in the tenth century, for example, in the margins of the presentation scene of the Vita Cuthberti, made in Winchester c. 935–9. Here as at Priors Barton the scrolls also have acanthus leaves (Temple 1976, no. 6, ill. 29). It is also at this period that acanthus appears elsewhere on sculpture. A grave-cover from St Oswald's Priory, Gloucester, is decorated with a bush scroll with acanthus foliage very similar to that on the Priors Barton shaft (West 1983) and this piece has been dated to the first half of the tenth century, possibly to the 930s (ibid., 50) as has a similar cover, decorated with an acanthus tree-scroll, from Wells, Somerset (Rodwell 1980, pl. 7). It is, therefore, possible that the Priors Barton shaft is not a ninth-century piece, as the types of ornament and their combination suggests, but represents the survival into the tenth century of these earlier decorative traditions which were then combined with newer elements such as acanthus foliage.
Related to the ornament on the Priors Barton shaft is the decoration of a grave-cover from the New Examination Schools, Oxford (Ill. 363). This has a median ridge on either side of which is an inhabited simple scroll with a segmented stem, large ragged acanthus leaves, and berry bunches each with three fruit, the same general type of ornament as on the Priors Barton shaft. The details of the cover are obscured by heavy weathering, but the ornament appears to be closely paralleled in the borders of the presentation scene in the Vita Cuthberti where the same simple scroll with a segmented stem is also employed (Fig. 14). As on the Oxford piece, the scroll is inhabited by birds and has acanthus foliage and berry bunches with three fruit. This close parallel may suggest a date in the first half of the tenth century for the Oxford cover, a date supported by comparison with the similar inhabited simple scroll with acanthus foliage on a shaft from Colyton, Devon. This shares with the Oxford carving the use of the simple scroll with large, flat, rather shapeless acanthus leaves inhabited by birds. Again there are berry bunches composed of three fruit (Kendrick 1949, pl. XXXIV). A later date for the Oxford piece is, perhaps, unlikely; although inhabited acanthus scrolls do occur in the late tenth and eleventh centuries, as on the bronze openwork strap-end from Winchester (Roesdahl et al. 1981, J6, pl. p. 168; Biddle 1990a, no. 1057), or the ivory pen case from London (Kendrick 1949, pl. XXXVI.1), they are usually bush-scrolls. Moreover, the later examples employ full-blown Winchester-style acanthus ornament, absent both from the Vita Cuthberti and the Oxford cover.
Also linking closely to the manuscript art are the acanthus sprays on the crucifixion panel from Romsey (no. 2). This depicts the crucified Christ accompanied by the Virgin and St John, the figures of Stephaton and Longinus, and an angel on either side of the cross-head. Acanthus sprays develop from the lower limb of the cross and the lower edge of the panel (Ills. 453, 455). Two of these sprays consist of a stem from which emerges a pair of down-turned leaves, a form which echoes the tree-scroll used at Whitchurch (Ill. 484). Acanthus of very similar form is used to decorate the lower corners of the border of the second Christ in Majesty page of the Aethelstan Psalter (Temple 1976, no. 5, ill. 33) and to decorate the borders containing single figures in the calendrical material prefacing the Psalter, as on fol. 9v (ibid., ill. 15). Similar acanthus sprays are used as scene separators on the St Cuthbert stole (Battiscombe 1956, pl. XXXIV). The stole is dated by inscription to c. 909–16, and there is little reason to doubt the traditional association of the Psalter with king Aethelstan (924–39). These parallels for the foliate ornament would, therefore, place the Romsey panel in the first half of the tenth century, although the possibility of an earlier date should not be discounted, particularly since Cramp has pointed out the strong stylistic links between the piece and west midland styles of the ninth century (Cramp 1972, 146). These links are particularly noticeable in the round staring eyes and the flat treatment of the drapery with the details picked out by shallow incised lines.
In summary, dealing with the group of Hampshire material, with its outlier in Oxfordshire, it is possible to indicate some stylistic links, including the preference for tree or bush scrolls, sometimes within a semicircular border, and sometimes with acanthus foliage. Other traits include the preference for small-scale figure sculpture and the use of round shafts. These links are clearly far too weak and the sculptures too few to suggest the existence of a single school of sculpture, but, combined with both the restricted geographical distribution and date range of the pieces, they are enough to suggest that the sculptures belong within a single evolving tradition. The distribution of the pieces suggests that this tradition was based on Winchester, a conclusion supported by the close links between the sculptures and works in other materials which were probably produced there, such as the Vita Cuthberti, the Anglo-Saxon additions to the Aethelstan Psalter, and the St Cuthbert stole and maniple.
As well as the Hampshire material, there is also important ninth- and early tenth-century sculpture from Kent. Of the thirteen fragments from the county dating to this period, by far the most important are the carvings from Reculver (nos. 1–3). So multifarious and complex are the problems which these highly classicising pieces raise that they are discussed in detail in Chap. VI, where it is argued that they should be dated to the early part of the ninth century and set alongside BL MS Royal I. E. VI. This is another such highly classicising work, which was probably made at St Augustine's abbey, Canterbury, in the period c. 820–40. Both this luxurious manuscript and the Reculver fragments seem to derive from, and closely reflect, contemporary developments in Carolingian art.
The presence in Kent of these two strongly classicising works, reflecting Carolingian artistic trends, may provide a context for the three classicising capitals from St Augustine's abbey, Canterbury (nos. 3–5). These were discovered built into the foundations of the screen of the Romanesque abbey church during excavations in 1915–16, and consequently have no good archaeological context. They have, therefore, proved difficult to date. The capitals are of a debased composite form with a zone of upright leaves at the base separated from the corner volutes by a narrow recessed zone of interlocking triangles. In the centre of each face in the upper zone of the most intact example is a rosette (Ills. 29–40). There are no closely comparable pieces from Anglo-Saxon England, but debased composite capitals were produced in France at least from the seventh century right up to the eleventh century (Fossard 1947). The problem with the Canterbury capitals has been to find a context into which the introduction of this type of capital into England would fit. A seventh-century date, contemporary with the Merovingian examples, is possible and has been argued by Gem (pers. comm.) and West (pers. comm.). More often the capitals have been connected with the building of the octagon at St Augustine's abbey by Abbot Wulfric (1047–59), since the design of the building is known to have been influenced by William de Volpiano's St Benigné at Dijon, begun in the year 1001 (Peers 1927a, 215; Conant 1974, 149–53). There, debased Corinthian capitals survive, although none closely resembles the Canterbury capitals, but there are no surviving composite capitals. H. M. and J. Taylor have, however, pointed out the very close resemblance between the Canterbury capitals and a Corinthian capital in the crypt of St Germain, Auxerre, which can be dated to the period 841–65 (Taylor and Taylor 1966, 47–9, fig. 20). Here the lower zone of upright leaves is similar to those on the Canterbury capitals (though less stylized), but the zone with the volutes is very similar in form and, as at Canterbury, has a rosette in the centre of each face. There are related capitals in the ninth-century crypt at Flavigny (loc. cit.), where the upright leaves more closely resemble those at Canterbury, but the volutes on the upper zone are suppressed and the rosettes greatly enlarged. No ninth-century building campaign is documented at St Augustine's abbey, but the production there in the mid ninth century of BL MS Royal I. E. VI argues for very strong Carolingian connections, and provides circumstantial evidence for the dating of the capitals proposed by Taylor. Supporting evidence for this proposed ninth-century date derives from the analysis of the blue paint used on one of the capitals (see the catalogue entry). This proves to be Egyptian blue, a colourant current in the late Roman period, but which then fell out of use until the period of the Carolingian renaissance. Given the lack of knowledge of colourants used on stone in the Anglo-Saxon period, too much weight should not be given to this evidence, but it is nonetheless suggestive.
The Reculver fragments (nos. 1–3) and the Canterbury capitals (nos. 3–5) belong to a highly classicising tradition which can be paralleled in part in the decoration of BL MS Royal I. E. VI. The native Anglo-Saxon element in the decoration of this manuscript is less in evidence on stone sculpture from Kent, but is paralleled on a piece of sculpture from Godalming, Surrey (no. 1). This consists of a stone ring, the outer face of which is divided by four equally-spaced animal masks into rectangular fields, each with a broad, plain border. Two of the fields (Ills. 85, 91) are filled with median-incised interlace, one (Ill. 84) with a biped whose tail develops into interlace, and the fourth (Ill. 86) with a simple scroll with elongated triangular leaves. The layout of the ornament on Godalming 1 very closely resembles that of the Canon Table frames of Royal I. E. VI, which also have the decoration split into a series of rectangular panels within broad, plain borders (Alexander 1978, no. 32, ills. 162–4). This panelling of the decoration within broad borders developed in the eighth century, as in the Canon Table arcades of the Leningrad Gospels (ibid., no. 39, frontispiece, ills. 188–91) and Maaseik Gospels (ibid., no. 23, ills. 96–107) and in the initials of the Barberini Gospels (ibid., no. 36, ills. 169–72), but persisted into the ninth century, as in BL MS Royal I. E. VI (ibid., no. 32, ills. 162–4) and Cotton Tiberius C. II (ibid., no. 33, ill. 165). In the eighth-century examples the fields are filled almost exclusively with animals enmeshed in interlace, but, as noted above), in the late eighth and early ninth century the alternation of animal, interlace, and plant ornament used at Godalming 1 became the norm. In Royal I. E. VI, for example, the fields are filled with interlace, animal or plant ornament. The animals are bipeds, as at Godalming 1, and develop at their extremities into different forms of decoration. This is usually plant ornament in Royal I. E. VI, but on Godalming 1 it is interlace. The Royal manuscript also employs the simple scroll with triangular leaves tucked under the stems, as on the Godalming piece, and the fields of ornament in the manuscript are often separated from each other by small square or circular fields; these serve the same purpose as the animal masks at Godalming (Ills. 82, 87–90). Animal masks are used as separators in the Royal manuscript, as on fol. 4r, but there they divide the ornament within a single field (ibid., no. 32, ill. 162). These parallels are enough to suggest an early ninth-century date for the Godalming piece, close to that of Royal I. E. VI.
A similar or earlier date can be suggested for a second, fret-decorated, fragment from Godalming (no. 2), as the fret (Ills. 92–3), although popular in the seventh to early ninth centuries (see, respectively, fols. 17v and 94v of the Lindisfarne Gospels (Backhouse 1981, pls. 22 and 28), fol. 30v of the Vespasian Psalter (Alexander 1978, no. 29, ill. 146), and fol. 43r of BL MS Royal I. E. VI (ibid., no. 32, ill. 161)), fell out of common use, at least in manuscript art, thereafter. Simple frets continued to be used in other materials in the tenth and eleventh centuries, as on the shanks of bone and metal dress pins from York, but none approaches in complexity the fret on the Godalming piece.
There remain two fragments of sculpture in Kent which possibly date to before c. 950; the interlace-decorated fragmentary cross-shafts from Rochester (no. 1) and Preston by Faversham. The Rochester piece was discovered during a controlled excavation, but was residual in a later level and, therefore, not closely datable archaeologically. The interlace on the single surviving decorated face (Ill. 141) has been reconstructed by Swanton as an encircled pattern with bifurcating strands (Harrison and Williams 1979, 34–5, fig. 7), but the pattern can be reconstructed more convincingly without them. Whatever its precise reconstruction, the fine strand of the interlace can be compared with that on Reculver 1e (Ills. 119–20), and the piece may, therefore, be tentatively placed in the ninth century. Similarly, the fragmentary cross-shaft of square section from Preston by Faversham (Ills. 101–4) is decorated exclusively with interlace, including half patterns A and C. The use of half patterns is unusual in south-east England, being paralleled on Reculver 1e, where half pattern D is used. The interlace of the Preston shaft has a thicker strand than that used at Reculver, but the design is well worked out and executed, and may, on analogy with the Reculver fragment, be assigned to the ninth century.
Despite the diversity of the Kentish sculpture there are links between some of the pieces, and in particular, between the Reculver fragments (nos. 1–3), the Canterbury capitals (nos. 3–5), and the carvings from Godalming; the Rochester 1 and Preston by Faversham fragments are more distantly related. As with the Hampshire material these relationships are not strong enough to suggest that the pieces belong to a single school, but they do appear to belong to the same artistic milieu. As in Hampshire, the sculpture of the region is focused on a major ecclesiastical centre, in this case Canterbury, with the manuscript art of which the sculpture has close relationships. Again the material has a classicising element, apparently derived from Carolingian art, although that element is very much stronger in Kent than in Hampshire. The parallels and links between the two regions are significant, but many more discoveries will need to be made before these conclusions can be either substantiated or disproved.
One piece which is geographically isolated from both these groups of ninth-century sculpture is the sundial from Bishopstone, Sussex (Ills. 6–7). This is apparently in situ in a fabric which the Taylors place before 950 (Taylor and Taylor 1965–78,
NOTES
4. A seventh-century date has also been argued for the cross-shaft fragments from Reculver, Kent (no. 1); the problem of their date is discussed separately in Chap. VI.
5. The dating of the fragments from Reculver, Kent (no. 1), for which an eighth-century date has been proposed (Wilson 1984, 72) is discussed separately in Chap. VI.
6. This figure excludes the grave-marker, Old Minster 1, and several minor fragments from Old Minster (nos. 31, 59–61, 63, and 65) which are probably also of ninth-century date.
7. The ninth century date argued for the Whitchurch piece on art-historical grounds is supported by a study of its epigraphy. As Higgitt points out in the catalogue entry, the letter forms of the inscription suggest a date not much later than the end of the ninth century.
