Volume I | Chapter 2 | Sequence of FormsNext Back to catalogue index
by Rosemary Cramp, University of Durham

As an aid to the discussion below, and as a summary of information in the Catalogue, a Form and Motif Table can be found on p. 253 ff.


Although the foundation dates of some stone churches in this area are historically attested, and although we know the date for the introduction of the wooden cross,5 there is no such clear date for the introduction of the stone cross. Baldwin Brown noted that St Cuthbert's cross on Fame, erected before his death in 687, could have been of stone (Brown 1921, 164). However the medium is not definitely stated. No transitional or skeuomorphic forms have survived, such as one finds in the St John's cross on Iona, Argyllshire, where the ring-head is mortised and tenoned together (R.C.H.A.M. 1982, 200–1). It seems plausible to think of a variety of influences which contributed towards the development of the Northumbrian free-armed cross: crosses in wood, or in metal (such as are mentioned in the literature: Bede 1969, ii , 20, 204) and the stone triumphal pillars or columns of the Roman world. The specific reference to the signum — the cross — on the Jarrow inscription (no. 16), or the term sigbecn — sign of Victory — to describe the cross in the Bewcastle, Cumberland, inscription, clearly indicate that the Northumbrians interpreted the cross as the vexillum regis and could draw on antique models for such victorious signs. Since there are free-standing crosses which are not carved on every side (at Jarrow and Hexham), one must speculate on their relationship with the slabs bearing a cross in high relief, such as are found at Monkwearmouth and Jarrow. Such slabs remain the major monument of the Picts, either emerging subsequent to the appearance of free-standing crosses or contributing to their development. Both free-standing crosses and monumental slabs could have originally been part of the interior arrangements of churches, whether as foci for prayer or as important funerary monuments.10 By c. 740 there is well attested evidence for a commemorative stone cross (presumably in the cemetery) at Lindisfarne,9 and by tradition, Acca's grave (died 740) was marked by a cross. At some sites, such as Edlingham, Nunnykirk, Falstone and Simonburn, there is now no evidence for the use of the churchyard later in the Anglo-Saxon period, but in most instances the pattern of development could have been of one outstanding and large-scale cross followed by later, usually smaller, monuments. This seems to be the position at Hart or Hexham. However, any site from which one piece of cross sculpture is known is likely to produce more examples when archaeological investigation takes place.

How early the idea of putting crosses on the limits of monastic territory was developed we do not know. It could have been an Irish custom taken over by the early eighth century. The Tynemouth cross (no. 1) has been claimed as such, but other crosses, such as the Legs Cross, were possibly wayside crosses or marked the boundaries of secular estates.

GRAVE MARKERS AND GRAVE COVERS (Grammar of Ornament, pp. xiv, xxi, and figs. 4–7)

It has been conjectured (Mercer 1964) that there could have been a direct influence from the funerary monuments of sub-Roman Britain on those of Anglo-Saxon Northumbria and, since the area covered in this volume touches on the British, Pictish, and Scottish kingdoms, this theory should be explored.

Simple, erect slabs, which can bear the name of the deceased, a Christian symbol, and sometimes a request for prayer, were well established as grave-markers and commemorative monuments in sub-Roman Britain. In Wales and parts of southern Scotland they continued into the seventh and eighth centuries (Nash-Williams 1950). In a changed form the slab is taken over in Pictland, and in Wales there seems to be an unbroken tradition of upright, cross-incised, or cross-decorated stones from the fifth to the eleventh century. The position in Ireland is more complicated, since here also the rough, monolithic pillar-stone with cross and sometimes inscriptions in Latin or Ogham characters was adopted in the fifth to sixth century (Henry 1965, 54–7, figs. 14–16). However, the forms became more regular and upright cross-slabs became much rarer. By far the greatest number of grave-markers in Ireland from the seventh to the twelfth century were recumbent cross decorated slabs. In this the situation is similar to England.11 Northumbrian artists did not apparently adopt the upright pillar-stone or slab as a Christian grave-marker at the same time as Christianity was adopted, although the short head- or foot-stones were used in the later period.

The difference between the slabs from Clonmacnois, co. Offaly, or Inis Cealtra, co. Clare, and those from such Northumbrian sites as Lindisfarne, Hartlepool, Billingham, Monkwearmouth or Birtley, is that the English slab is carefully shaped, as in Gaul (Salin 1952, 88–90; Cramp 1965b, pl. 1).

Moreover at Hartlepool (the only site where slabs were discovered in direct association with graves) some seem to have been found in the graves (nos. 0, 2–5, 7–8). This presumably accounts for their unworn condition.12 Hilda, abbess of Hartlepool, could have had knowledge of Gaul through her sister, who was a religious at Chelles, and the rectangular shapes of the Hartlepool stones, as well as their use of the Alpha and Omega, can be paralleled in Gaul. A similar grave-marker has been found at Monkwearmouth (no. 4), a centre with strong Gaulish influence. On the other hand, Hartlepool and Lindisfarne are foundations of the Aidanic church, and the varied cross shapes, the development from incised technique to relief, as well as decorated relief crosses found at Lindisfarne, are paralleled at Iona and in Ireland. Nevertheless the chronology of the Irish series does not support a date which is significantly earlier than the English group. What seems to be the earliest form of the recumbent slab — incised with crosses and lettered in runes and Anglo-Saxon capitals — is a Bernician phenomenon.

There is no clear chronological division in the use of upright and recumbent monuments. It seems that the large slab, used either as a recumbent grave-cover (Monkwearmouth 5) or as an upright memorial (Jarrow 13–14), could have existed by c. 700 and that it co-existed with the smaller inscribed and incised monuments such as occur at Lindisfarne, Hartlepool, Monkwearmouth and Billingham. The round-headed type of slab found at Lindisfarne could perhaps have developed directly into an upright monument by the late eighth century. However, an upright type with round head and simple relief cross is popular in this area from the ninth to the eleventh century (at Bothal, Tynemouth, Woodhorn, Escomb, Lindisfarne, Warkworth, Aycliffe, Elwick Hall, Gainford and Haughton-le-Skerne). Warkworth 2 is associated with Scandinavian ring-chain ornament.

Slabs in relief seem, in general, to be later than those which are incised, but there was perhaps a period of overlap when both types could have co-existed. Large inscribed slabs could have covered graves or stone receptacles, or have been placed against interior walls of churches, but what may be their ninth-century successors, such as Lindisfarne 38 or Norham 16, are uninscribed. None of the later slabs, including grave-covers like Hexham 17 or Durham 11, have inscriptions, and only a tentative dating by the forms of the crosses can be attempted for the rectangular recumbent slabs like Gainford 24.

No corner-post shrines occur in this area unless the slab from Monkwearmouth (no. 22) is part of one. Moreover other monuments which are derived from house-shaped forms, whether Roman sarcophagi or Insular metal or wooden shrines, are rare. The tiny, inscribed memorial from Falstone (no. 2) could be a skeuomorphic rendering of a metal shrine, but coped monuments such as Durham 11 are rare in comparison with parts of the country farther south. The idea of a mortuary structure which provides a house for the dead is widely dispersed among the Germanic peoples and it seems clear that from the migration period onwards important burials could be covered by a wooden structure, which would have provided a major focus in the cemetery.13

On the continent production of stone sarcophagi, whether of the hollow or block type (Grammar of Ornament, fig. 4, type d or e), continued without a break from the Roman to the Carolingian period and this fashion seems to have affected Mercia and southern Yorkshire in the ninth century (Cramp 1977, 224). This was a fashion which does not seem to have made inroads farther north, but it could be that the joint influence of the old wooden monuments and the Classically-derived stone shrines produced the distinctive hogback, which occurs in this area in the tenth century at sites usually with strong evidence, from other sources, for Scandinavian influence. The Durham/Northumberland groups are on the fringe of the main distribution of this form, but at Sockburn there is an important centre of carving where hogbacks and crosses are inventive and influential. Eight hogbacks survive from Sockburn as opposed to single examples at other centres (Fig. 3).

CROSS SHAPES ON GRAVE-MARKERS ETC. (Grammar of Ornament, fig. 2)

In some parts of the British Isles the shapes of crosses, incised or carved on slabs, with or without inscriptions, have been used as criteria of date. For slabs in this area it is difficult to use cross shapes as an indication of date. Only one linear type, Al, has been found, on Coquet Island (no. 1); if this is compared with the same type on Iona (R.C.H.A.M. 1982, 182, nos. 18, 20–1), its date could be seventh to eighth century. On one of the Hartlepool stones (no. 1) and at Birtley (no. 2) the 'cross potent' or A3 is found. Nash-Williams calls this form 'with barred terminals' and places it in his Group 1 (for example, on a pillar-stone from Llanlleonfell, Brecon, which he dates seventh to ninth century: Nash-Williams 1950, fig. 3, 13). However the early Welsh forms are linear crosses, and in the broad-outline form he dates slabs with such crosses twelfth century (ibid., pl. 62). Often this type is combined with crosses of type E8 which in the English series seem also to be late. In Ireland such crosses are dated eighth century (Lionard 1961, fig. 10, 11). The cross type with expanded terminals, F1, G1, is a type very common in Ireland where F1 develops as a highly decorated form (ibid., figs. 23–5). In Ireland this cross type, without ornament, can range in date, if inscriptions are any guide, from the ninth to the eleventh century, but with interlace and fret fillings, all dated examples are of the tenth century. However, as Scott has convincingly argued, slabs, such as those from Lindisfarne or Hartlepool which use this type of cross, can hardly be dated so late, so that we seem to have an example of parallel development (Scott 1956). There is a good case for seeing crosses with expanded terminals as derived from the traditions of Germanic art rather than from Byzantine crosses with splayed terminals as Lionard does (Lionard 1961, 129–30). The composite brooches from seventh-century graves, such as Dover, Kent (Avent 1975, pl. 64), have cruciform devices of the shape G1. Such a cross shape is also known on early eighth-century sceattas (Brown 1921, pl. 10, 6).

The cross type F1, common on the inscribed, recumbent stones at Lindisfarne, is also the form found on the silver casing of St Cuthbert's altar, although the flourishes on the terminals may be a later development (Radford 1956, 33, fig. 3). The English distribution of this type of incised cross — Lindisfarne, Hartlepool, Billingham and York — coincides with centres founded in the time of the Celtic supremacy of the north. Moreover the type is common and has a long popularity in the Celtic west, occurring in Ireland and Scotland (Lionard 1961, figs. 21, 23–6), and the Isle of Man (Webster and Cherry 1977; 168–9, fig. 49). Type A2 which is found on the Monkwearmouth/Jarrow relief slabs, both large and small, dies out after the ninth century (Bothal 5 and 6). One of the latest, on a Monkwearmouth slab (no. 30), has a small cross type, A12. This and type B12 are commonest on late recumbent pieces, whilst B8 and E8 are the commonest forms on small foot- or head-stones which seem to span the Conquest period.

In general then the types with angular arm-pits, 1–4, are commonest before the ninth century, and some, no doubt, overlap with those with circular arm-pits, 11–12, but these continue up to the Conquest. The types which seem most popular in the period c. 1000–1100 in Northumbria are, however, those with narrow V-shaped arm-pits, 6–8.

FREE STANDING CROSSES (Grammar of Ornament, p. xiv, and figs. 1–3)


Better quality carving, which in this area tends to be earlier than the mid ninth century, is often associated with shafts which are square or nearly square in section. However, slab-like shafts can appear early, as at Hexham (no. 3), where the depth was about half the width, while squarish shafts, such as Great Stainton 1 and 2, Chester-le-Street 1 or Gainford 4, are later than the mid ninth century. From this area there are few shouldered shafts or round-shaft derivatives which are characteristic of the Danelaw, although some occur at sites in the southern part of co. Durham such as Coniscliffe (no. 6). The most common shaft type of the tenth and eleventh centuries is slab-like.


The free-armed head seems to have a significant geographical and chronological distribution, although it is evident that different types could coexist, and it is difficult to say whether period or local fashions are more important. A type which is associated with early contexts and ornament is B9, as at Jarrow (no. 9), Hexham (no. 1) or Northallerton, Yorkshire. Chevrons, plates, pellets and filigree ornament are all copied from metalwork on these slim cross-heads. The use on Jarrow 9 of baluster ornament, such as is found in architectural decoration of late seventh-century churches, supports an early date for this type (pp. 23–7). All such heads seem to have a bossed centre. The type continues in southern Northumbria into the eighth to ninth century, and later crosses bearing Anglo-Scandinavian ornament, such as Aycliffe 2, preserve the head type and central boss with jewel-like indentations. Perhaps the Wooler head (no. 1) is also a late development of this type. An example of a simple central boss surrounded by interlace ornament has been found at Tynemouth (no. 6), in what may be a tenth-century context. On the whole, however, the slender elegant outline of the early B9 crosses gives way by the ninth century to the more stunted forms of B8 and B10.

Hexham and its dependencies favoured another type of cross-head in addition to B9. This, A10, is found on monuments less pretentious than Hexham 1, such as the very plain cross-heads Hexham 8 and 10, with either no decoration or some simple motif such as a rosette centre. This type is found outside the region associated with inscribed crosses, at, for example, Whitby, Yorkshire, and Carlisle, Cumberland (Okasha 1971, pl. 23, a–b).

The double-curved form does not, as Baldwin Brown first pointed out (Brown 1921, 167, fig. 11), occur in other media at this period. It is possible that all the forms of 9 with widely curving arm-pits could have developed together, and there are sometimes difficulties with worn cross-heads in deciding whether the type is A9 or D9. It is possible also that Lindisfarne was the original centre for the double-curved form, since it is found at that site associated with fine double-stranded interlace on early cross-heads (nos. 15 and 16), and on a grave-cover (no. 38). The type with the curve nearer the end of the arm assumes massive proportions in Yorkshire at such centres as Masham and Lastingham. In our area it occurs on a smaller scale at Norham (nos. 10, 11, 14), where like the Yorkshire heads, it is decorated with a flowing plant-scroll.

The type A9 could well have developed at Monkwearmouth/Jarrow, since this centre favoured straight-sided cross-heads on slabs, and the ornament of heads, such as Ruthwell or Auckland St Andrew 1, with their Evangelists' symbols could have derived from that centre. By the time Rothbury 1 was carved, it seems as though types A9 and D9 were merging, although Rothbury (like the eleventh-century cross-heads from Durham) does retain something of the double curve. The Durham cross-heads tend to have a more circular arm-pit and shorter length of arm than occurs with the flowing curves of earlier types. These head types are often the most elaborately decorated of all cross shapes (for example, the Crucifixion and Passion scenes at Rothbury or the packed liturgical scenes at Durham, nos. 5–7).

In general, cross-heads associated with demonstrably later ornament are smaller and clumsier, with shorter arms, as befits the increased number and provincialization of monuments. Types B10, B11 and C11 are all found on small, slab-like crosses which seem to belong to the tenth and eleventh centuries. The decoration of such heads consists most often of interlace or joined knots (Gainford). However, the tenacity of the fashion for the free-armed head in late centres, such as Aycliffe or Durham, Gainford and Chester-le-Street, is remarkable in comparison with the vogue for Anglo-Scandinavian forms with rings, wheels or plates manifested by the Bernicians' neighbours in Cumbria and Deira. The ring-head is a tenth century phenomenon (Bailey 1978, 177–9) and appears at that period in the Tees valley (Collingwood 1927, fig. 153), sharing the same distribution pattern as the hogbacks; this has prompted the suggestion that both types were originally developed in places settled by the Norwegian Vikings (Lang 1967). A later ringed type, such as is found at Lindisfarne or Coldstream, Berwickshire, became popular in northern Bernicia by the late eleventh century (Cramp and Douglas-Home 1980, 230, pl. 14).


Despite the conscious attempts of such churchmen as Wilfrid and Benedict Biscop to found and equip churches in the continental manner, we can hardly assess their success, since few of the furnishings have survived. It is possible that much of the surrounding of the altar (pp. 24–5) was in wood, and that much of the furniture was portable, such as metal stools of the folding type. One might have expected fragments of ambos or cancelli such as appear so conspicuously in the continental record; only Monkwearmouth 9 and Hexham 21 can be thought of as parts of cancelli. However Durham and Northumberland are comparatively rich in seats, which show an interesting range of types.

The Hexham 'frith stool' (no. 41) is a well known piece, and its simple design and restrained, incised ornament of twists and triquetras are clearly in the Classical tradition. A fragment from Escomb (no. 4) with a similar simple twist could well be an upright for a piece of furniture, although it is quite possible that it could be part of the altar fitting. The lion arm-rests from Monkwearmouth (nos. 15a–b) also maintain the traditions of Classical furniture, both in form and style. However the animal head from Monkwearmouth (no. 16), if it is the terminal of the arm of a chair as I have proposed, reflects the same Insular tradition as furniture depicted in manuscripts such as Durham B. II. 30, fol. 81v (Kendrick 1938, pl. 54) or the Lichfield Gospels, p. 218 (Henry 1965, pl. 103). It could also be compared with wooden animal-headed posts at Oseberg (Wilson and Klindt-Jensen 1966, pls. 11–12, 14). Perhaps the most important new discovery is the arm of the Bamburgh chair (no. 1), which in form can be paralleled in episcopal seats, or Charlemagne's imperial throne at Aachen (Schramm 1954, pl. 44). Although it is fragmentary and not associated with an early structure, it seems to reflect the traditions of such Early Christian cathedrae as Maximian's throne at Ravenna (ibid., 316–44, pl. 34). However the ornament which richly covers it has departed entirely from the Classical tradition. Its plaits and interlaced beasts are clearly Insular.

The pieces which have been interpreted as supports for reading desks or tables are more tentatively assigned. The Egglescliffe column (no. 2) with its flared top and base is also very clearly in the Classical tradition (cf. Liversidge 1955, 18–19), and it could well be that the drums and colonettes at Monkwearmouth (no. 14) and Jarrow (no. 30), which are Insular in form, not only make up balustrades but also function as supports; examples of their wooden counterparts in furniture are St Radegund's reading desk, or the bed and chair from the prince's grave under Cologne Cathedral (Werner 1964, figs. 6–7).

The Jarrow octagon (no. 22) with vine-scroll decoration interestingly emphasizes its architectural form in the column decoration. It has a formal equivalent at Melsonby, Yorkshire (Cramp and Lang 1977, nos. 7–8), but both, like Hexham 7 or the half-column fragment from Hexham (no. 42a), cannot be proved to fulfil the function assigned to them. Another possible item of furniture may be no. 4 from Lindisfarne, but any interpretation of its function is complicated by its reuse.


5. See under Chapter 1.

9. See under Chapter 1.

10. The Reculver cross was described by Leland as standing in front of the altar in the church; and in the north some important crosses, such as Ruthwell, or in this area, Rothbury and Auckland St Andrew, are unweathered and could well have been under cover for much of their life, presumably in a church.

11. The origin of the recumbent slab in England and Ireland is, however, very difficult to determine. It has been said for Ireland: 'It is not clear either what relation it bears to erect cross bearing pillars or slabs with which it may often have been associated over the same tomb.  The use of a recumbent slab over the tomb must have been dictated in many cases by the desire to protect the dead body from wild beasts.... But this does not explain the existence of the very small slabs which are so common. They may parallel the stone tablets of Roman tombs, placed over some part of the body. Many graves of Merovingian Gaul show the same custom' (Lionard 1961, 99).

12. The position of the stones in the graves and their relationship to the Irish series have been most fully discussed by Baldwin Brown (Brown 1921, 71–101) and Scott (1956, 197).

13. For a recently discussed example see Hills 1977.

Next Back to catalogue index