Standing loose on the cill of the east window of the south aisle. It has stood here probably since at least 1888 (Nottinghamshire Archives Office, DD/TS/6/4/7, p. 57) and certainly since 1903 (Standish 1903).

'Early in 1902 we were shown a well-preserved section of an early shaft, stated to have been found under the floor of the Chesterfield chapel' (Stapleton 1903, 90). However, Stapleton's second and subsequent accounts report that the stone was 'stated by one writer to have been found during the restoration of 1877–8 built up in a brick buttress that used to support the south aisle' (Stapleton 1911, 140–1; id. 1912, 30; see Standish 1903; Fellows 1907, 20; Godfrey 1907, 428). As Stapleton revised his account between 1903 and 1911, we can presume that he thought that his earlier research had been incorrect, and that Standish's and Godfrey's accounts of its discovery outside, built into a brick-built buttress against the south wall in 1877–8 during restorations by Ewan Christian, were to be preferred. Indeed, the stone contains evidence for its reuse in fabric (below). The date of the brick buttress is difficult to estimate, but in the light of the discussion below, we should note that there was substantial repair work at the church in 1578 (Robertson 1910, 85) and 'every part of the building bears marks of [that] age' ([Barker] 1835, 138–46).

St Peter and St Paul

It is likely that the earliest damage to the monument occurred in the sixteenth century, when the faces of Mary and the Christ-child were carefully and delicately planed off without damaging the surrounding sculpture. This careful piece of iconoclasm is interesting in itself, as few pre-Conquest monuments show this feature; but it also tells us that the monument was still in use in the sixteenth century, probably as an intact cross-shaft. It is even possible that this very specific iconoclastic act might date from the 1570s when Archbishop Grindal of York ordered that churchyard crosses throughout the diocese of York should be defaced as 'monuments of superstition' (Walsham 2011, 102, 117–21). In addition, on the opposite face, a similar though cruder piece of iconoclasm has affected the area of the Angel's raised left hand, where a raised feature in the sculpture has been smashed by direct targeted blows. This may belong to the same round of activity. Apart from these very specific pieces of iconoclastic damage, the stone has also been re-cut at a later date, almost certainly when it was reused in the buttress. What was originally a substantial rebate has been cut at the base of face D using a crudely wielded broad boaster, whose marks remain visible. This is likely to have been created to provide a seating for the adjacent fabric within which the stone was re-set. Unfortunately, above this rebate there is a substantial area of mechanical damage extending right across face D which removed more of faces A and C. Face C has also suffered similar mechanical damage at the top right that extends onto face D. Both of these areas of mechanical damage look more recent. Apart from all this informative deliberate damage, three of the sculpted faces (A, C and D) survive in remarkably good condition. They are somewhat weathered, though perhaps surprisingly little for a cross which seems likely to have stood in an exterior position for six centuries or more. However, face B is so badly weathered that only the last vestiges of the interlace or plant-scroll pattern with which it was decorated can be made out in strong raking light. Face B, then, is likely to have faced outwards during the stone's reuse in the brick buttress during the early modern period.

The stone is a section from the central part of a major standing shaft of tapered profile and has a cross-section that is approaching square. All four faces were originally intricately carved in a deep low relief. The angles of the shaft are elaborately worked, and are square in section. On faces B, C and D the arris is flanked by a single run of pellet-moulding in turn flanked by an inner fillet bordering the sculpted panel. On face A, the main surviving panel (though not that above it) is flanked by a double run of pellets. The decoration on each of the four surviving faces is contained within a panel under an arched head, which is outlined with the same pellet moulding, flanked by fillets at the angles. The point at which the arch-head springs from the angle moulding on all four faces is marked by a simply delineated transverse roll moulding in imitation of an impost. These 'imposts' have miniature hollows (scotia) between two rolls (torus). We cannot now tell how the shaft below the surviving fragment was treated, or indeed whether the surviving sculpture represents the lowermost order of figural decoration on the monument. The evidence does survive, however, for at least one further order of arched panels (presumably containing sculpture) above those surviving. Unfortunately nothing comprehensible survives of the decoration on any of these four upper panels.

A (broad): This face is fully occupied by a magnificent seated figure of the Virgin Mary with the Christ-child sitting on her knee (Ills. 89, 92). The Virgin's head is surrounded by a deeply dished 'halo-cum-hairdo' (Pattison 1973, 222–3; Lang 1993, 265–7), which has tightly curled terminals sitting above her shoulders. As with the Angel on face C, the clothing of both the Virgin and the Christ-child is depicted with neat parallel lines representing the drapery's folds. The Virgin wears a full-length gown, which appears to lie beneath a long 'shawl' which sits over her shoulders and crosses on her chest, where it is evidently fixed with a large ring or brooch or a knot. The two ends of this shawl extend beneath the Child and fly out symmetrically across the Virgin's shins. Its left-hand end has been lost to the area of damage created by the substantial rebate, but its right-hand end terminates in an accurately depicted zig-zag hem. The Virgin seems to wear a waistband of parallel lines above her knees, which in turn are clearly delineated with transverse folds, and the gown falls to just above the ground in a series of straight folds with a tubular appearance. The hem itself is depicted in fine detail, with each fold being shown almost in three dimensions. The hem of the gown arches upwards to reveal two simply depicted feet in pointed slippers facing downwards and forwards, and there is a suggestion that they stand on a slight roll. Her arms are clothed in tubular drapery depicted by parallel horizontal folds, and her small hands, with delicately depicted fingers, emerge beneath the Child's bottom.

The Christ-child sits on the Virgin's right knee, his body facing to the right (as we view the image) but head turned to face the viewer. He is clad in a simple one-piece garment wholly depicted by precisely inscribed parallel lines, pulled in at the waist by a simple belt. The garment's hem is depicted in similar detail to the other garments on this monument and it terminates in a row of accurately delineated, three-dimensional, zig-zag folds. Only Christ's right arm appears to be shown and not the left, and it emerges from a short sleeve in his over-garment below the shoulder. The lower part of his arm is also draped, with the folds running in the opposite direction to those in his over-garment. In his right hand, Christ holds a large rectangular wallet or, more probably, a book upright across his mother's midriff. The hand and book are entirely lacking surviving detail, and this is partly because this part of the sculpture has suffered from slight local abrasion, as though this point on the sculpture has been repeatedly touched. As has already been mentioned, the faces of both Virgin and Child have been carefully planed away and are now flat and featureless discs. The surfaces of these discs still show some signs of slight weathering, however, suggesting that the shaft continued to stand for some time after the careful iconoclasm to which this detail bears witness (see below).

B (narrow): Although this face is extremely weathered, the bare outlines of the original decoration can be understood in strong raking light (Ill. 90). It is framed in the same manner as are panels on the other three faces, and its decoration seems to start below the current base of the stone. This indicates that decoration continued some way further down the original monument than the section represented here, at least on this face. Towards the top of the surviving section, the remains of the arched head, with the same pellet moulding as on the other faces can be discerned. Within the panel, an arrangement of thick and thin interlacing strands is evident, and it is also clear that they formed a coherent, complex pattern, organized in a series of circular elements one above the other. This was certainly not a panel of running plaitwork; the elements seem rather to be the medallions of a developed plant-scroll, with some form of leaf filling the spandrels. The alternative of some type of 'ribbon beast' or 'beasts', comparable with the shafts at Newgate and St Mary Bishophill Senior 1 in York, and at Sutton-upon-Derwent in the nearby county (Lang 1991, 88–9, 105–7, 220–1), would be unlikely to exhibit such a regular and rounded residual pattern.

C (broad): This side of the shaft is dominated by a striking standing figure of an Angel, unusual for its stress on male features like facial hair (Ills. 91, 93). He has four wings, two held upwards behind his head, and two falling behind him like a great cloak. The wings are delineated with strong parallel lines for the main pinions and, towards the shoulders, small circular motifs imitating shorter, more downy feathers. The tips of the upper pair of wings nearly meet in two symmetrical, inward-curving tips. The clothing that covers the Angel's body is also defined in a series of precisely cut parallel lines, giving, in this deployment, an approximation of drapery folds. His upper body is covered in two garments: an upper 'jacket' and a lower 'skirt'. The 'skirt' has a scalloped hem that is raised at its foot to reveal two forward- and downward-facing feet. The drapery over the knees is treated in distinctive and elaborate detail with a swirl of overlapping folds that suggests the Angel is seated, whilst the drapery on his arms is represented by horizontal folds that might suggest he wears an undergarment beneath a sleeveless jacket. The right arm is slightly bent at the elbow. Held beneath it is a wallet or a book, which is without detail. This hand is nicely carved with great delicacy. The left arm is raised high above the Angel's head in what might appear to be a gesture of benediction, but the area of the hand has suffered from an impact and the detail is obscured. What is clear, however, is that the arm narrows noticeably just as the right arm does — since the curving top of the wing can be seen clearly on both sides of it — but this left hand is of a different, larger, scale than the other. This difference of scale is so marked that it appears that the figure was holding aloft some quite substantial object. The Angel's face is beautifully and carefully depicted with rounded cheeks and a detailed nose and eyebrows. The eyes are lightly inscribed and lentoid. A long moustache with tightly curled ends curves downwards beneath his nose, whilst a forked beard falls from his carefully delineated chin, again ending in two tight curls. There are some signs that the Angel had a relatively short hair-cut, with a curly fringe, but the top of the head is too weathered for certainty on this point. There is a diagonal moulded fillet across the Angel's lower wing on his left side (right as we view the image). It is not clear what this might have represented, and it now disappears into a broken area.

D (narrow): As described above, the original sculpted surfaces of the lower part of this face are missing, partly through re-cutting during reuse and partly through more recent mechanical damage, as is a smaller area in the top right of the face affecting the frame of the panel above. Where the original surface has escaped this damage, however, the sculpture survives in good condition and is distinguished by a deep low-relief technique (Ill. 88). The panel contained a complex pattern of symmetrical interlace, with the strands enhanced by an incised medial line. Within the surviving area, the interlace forms complete pattern A with irregular V-bends (Cramp 1991, figs. 14 and 16). The interlace is not very accurately laid-out, however, and the individual units quickly fall out of step with each other as they progress up the shaft.

The Shelford shaft offers by far the finest early figure sculpture in Nottinghamshire; indeed it has amongst the most interesting and best-preserved figure sculpture of its date in England. Seeing it in May 1904 when visiting for his contributions to the first Nottinghamshire VCH volume caused J. C. Cox to dash off an enthusiastic note from the Victoria Station Hotel in Nottingham to Romilly Allen, commending its interest and quality; and the latter was immediately able to point out the relevance of a comparison with the shaft at Nunburnholme in Yorkshire (BL, Add. MS 37552, ff. 219–220, 224). The great scholar of Anglo- Scandinavian art, T. D. Kendrick, said of the Shelford monument, 'It is clear that this Shelford cross must have been a magnificent carving' (1949, 79). Despite this accolade there has been no evaluation of it in its own right, only in relation to other — principally Yorkshire — comparanda; and that assessment has been rather simplistic, most obviously in relation to its iconography.

Iconography

The images on the two principal surviving faces of this shaft have perhaps seemed so easily identifiable as to require little further exploration. One side carries the depiction of the Virgin and Child, with the Christ-child turned to look directly out at the observer and holding out a book between his two hands. The other side depicts an angel, who also — superfluously? — carries a book and is (as Jim Lang pointed out, but had no explanation for: 1988, 22) canonically incorrect, both in being bearded when angels are supposed to be androgynous and in blessing with the wrong hand. Nevertheless — if one can accept such unexplained incompetencies — the angel might, at a pinch, be thought from its loose association with the Virgin to be the archangel Gabriel and that the two images taken together amount to an Annunciation scene.

The main face of the shaft is occupied by the Virgin and Child (Ills. 89, 92). It is further picked out as the main face — and probably the principal image of the shaft — by a doubling of the pelleted pilasters on either side of the architectural frame, compared with those flanking the Angel image, and their more substantial blocky capitals. This additional elaboration is not carried over the head of the semi-circular arch, however, or in the pilasters as they continue upwards.

A more demanding agenda has been set by modern scholarship — led by Raw (1966), Kelly (1995) and especially Hawkes (1997), supported by Mary Clayton's fundamental study of the Virgin's cult (1990) — in the analysis of iconic Virgin and Child depictions on Insular monuments, especially ones with such clear detailing as Shelford displays. Jane Hawkes emphasizes particularly that the pre-Conquest Insular portrayal of the Virgin and Child, whose surviving examples she reviews, does not lend itself to ready or convincing categorization in any one — or any small number — of the standard iconic 'types' of this theme into which products of the Byzantine and later medieval worlds have been fitted. Instead, they exhibit minor variations and idiosyncrasies, which cut across those 'types' (Hawkes 1997, 109–13). This was the experience of Kitzinger, too, in his seminal study of the early Virgin and Child on Cuthbert's wooden coffin (1956). He emphasized instead the 'complementary' pose, whereby the Child is held in such a way that his body crosses that of the Virgin, as a diagnostic feature. Hawkes finds that useful, but only as a start to commentary, demonstrating that other compositional details can vary considerably and indicate distinctive intentions or model-types being drawn on. In the Shelford carving, then, the Virgin and Child utilize the 'complementary' pose, as do the majority of Insular instances. But the Virgin's seated pose is firmly forward-facing, while the Child's head — twisted at the neck — also presents a full-faced, front-facing aspect. This already contrasts with the poses of two of the most detailed comparative images available: the Cuthbert coffin — which has full-frontal faces but differently posed bodies — and the Virgin and Child of the Book of Kells — where the Child turns to his mother and there is an interplay between them (Hawkes 1997, figs. 1 and 2). At Shelford, the Child is grown, stiff-backed and stable on the lap: with his feet depicted hanging below the hem of his garment, he is neither the swaddled baby of the cross-slab from Brechin, Angus, or the Kildalton cross on Islay (ibid., figs. 5 and 6), nor even receiving the support of one of the Virgin's hands behind the back and the other steadying and restraining across his knees, as seen on Cuthbert's coffin or the Book of Kells. Rather, the Virgin's hands are tucked away under his bottom and his knees, effectively holding the Child out to the observer in a gesture of presentation. And the Child in turn both presents himself with his direct view and also holds out the book with both hands. This, then, is the Hodegetria image in full spirit and effect (The One Who Shows the Way) but without the range of conventional compositional details that define that icon-type (ibid., 109–11); and in two senses, since 'The Way' here is both the presence of God made man in the form of a child and the good news he brings about and presents as proof and guidance. The closest Insular parallels to this imagery, according to Hawkes, are those at Drumcliff in Ireland, St Oran's cross on Iona and at Kildalton in the Western Isles, and on the Sandbach south cross in Cheshire (ibid., 116–18); but the carving in each of those cases makes a fine comparison difficult to establish. A regional link may exist rather with the Virgin and Child at Dewsbury in Yorkshire, whose more legible detailing reveals differences such as the Child's hand raised in blessing but also a relationship in the Virgin's posture to Nunburnholme (Ill. 188, which stands as almost a pair to Shelford in several other respects; below). Dewsbury is widely accepted as an early ninth-century sculpture, and available as a potential model for these other Virgin and Child images in the northern Danelaw (Hawkes 1997, fig. 7; Coatsworth 2008, 133–5, ill. 198).

Striking as this iconography is, there is a further very distinctive detail in the Shelford Virgin's portrayal. That is the long, narrow shawl that passes round her neck, crosses in a knot or fastening at her breast, passes behind and beneath the Child and emerges with a symmetrical flourish in the lower foreground. This cannot be a casual or meaningless feature, since it is fully and carefully articulated — the crossing knot is very prominently located, filling the space between the book and the two principals' iconic faces, which form the focus of the image, and it emerges with a notable showy flourish of hem below. Painted in contrasting bold colours its articulation would have been even clearer than it is now in bare stone. There seems no single convincing explanation of its significance, however. Arranged as it is between the Virgin's arms and the Child, it is reminiscent of the way in which saints and evangelists are commonly shown holding a book with a fold of their garments intervening between them and the holy thing. Here the Virgin makes explicit the holy nature of the Child she holds. In form, this shawl resembles a liturgical stole in its narrow length; and the stole is traditionally worn crossed across the chest of ecclesiastics for certain ritual purposes and occasions (Cross 1957, 1294). While not literally appropriate for the Virgin, such an allusion might emphasize the reverence due to the Child who became the body and blood of the Eucharist, and might give an added significance to this Marian image for a local priesthood. It seems unlikely, perhaps, that the shawl actually represents a traditional Jewish prayer shawl — the tallit — if only because there seems no iconographic purpose or parallel, and because in orthodox traditions the tallit seems to have been a male prerogative.

By contrast, it may be more plausible that the garment could sustain an allusion to Mary's famous girdle/sash/stole (Gk. ξϖυη). This might remind a viewer: of Mary's childbirth, with its assurance of Christ's humanity; of the yoke of suffering Mary bore through Christ's life and death (through the symbolism of the liturgical stole looped round the neck of its wearer); of Mary's own Assumption, because it was at that juncture that she traditionally unknotted the girdle and handed it to Thomas the Apostle.

None of these address the circular knot or fastening that is central to the depiction, however. For a literate religious audience, versed in the early Church Fathers, this might rather have brought to mind the words of Irenaeus of Lyons in the second century, that Mary had 'untied the knot of sin bound up by Eve', thereby emphasizing her active participation in the redemption of mankind (Warner 1976, 50–67, especially 59–60; Clayton 1990, 5). This idea of Mary as the second Eve was importantly bound up in patristic theology with the parallel one of Christ as the second Adam, developing a Pauline view (Romans 5.14; 1 Corinthians 15.45; Cross 1957, 14–15 'Adam', 479 'Eve'): both, by their obedience at the Annunciation and at Calvary, redeemed the original Fall and brought salvation to mankind. While the origin of this Mary–Eve relationship lay with the Church Fathers, there is ample evidence of its currency in the Anglo-Saxon Church from Bede onwards (e.g. Clayton 1990, 15, 186–7, 203, 223, 230, 256–7).

None of these readings is mutually exclusive, of course; and it would be wholly characteristic of the ecclesiastical context in which our Shelford monument was evidently produced (below) that imagery should be multivalent, as it certainly was for Bede or in the long-established Christian communities of the Celtic west.

Of the Angel on the opposite side of the shaft, too, much more can be made (Ills. 91, 93). The role of angels in general as objects for contemplation when carved into standing crosses has recently been explored by Jane Hawkes (2005, especially 268–9). But in this case, the Shelford angel is so finely carved, with a high level of detailing which is well preserved, that we may look for a less general understanding. It is a fair presumption, from this quality, that the sculptor's choices were purposeful and iconographically significant. The principal of these is that the Angel has four wings: not two as is standard for angels and archangels, however important their iconographic function; and not six, which distinguishes the seraphim in accordance with their description in the celestial vision of Isaiah (Isaiah 6.2–7), and occurs commonly in the finely detailed imagery of manuscripts and wall painting (e.g. Ohlgren 1992, 102, index under 'Angel: seraph'). The 'living creatures' that distinctively had four wings were cherubim. Occurring several times in the Old Testament from the earliest times of the Garden of Eden, which they guard after the Fall (Genesis 3.24), they figure most prominently and are described most fully in the prophecies of Ezekiel. 'In appearance their form was that of a man, but each of them had four faces and four wings. Their legs were straight; their feet were like those of a calf and gleamed like burnished bronze. Under their wings ... they had the hands of a man' (Ezekiel 1.5–8). A later, closely similar version of this description in Ezekiel 10.12 adds the detail that 'their entire bodies, including their backs, their hands and their wings, were completely full of eyes'. These amazing creatures were tetramorphs; their four faces or aspects, facing in different directions, showed the face of a man, of a lion, of an ox, and of an eagle. Each aspect had four wings, which touched at the corners: two raised and two down by or over the body (Ezekiel 1.10–11). Their name seems to have meant 'the near ones', familiars, personal servants, or bodyguards. Their function was to guard the presence of God: guarding the Ark of the Covenant in its desert tent (Exodus 25.18–21) and guarding the presence of Yahweh in the Holy of Holies in Solomon's temple in Jerusalem, in the form (variously) of two massive figures of olive-wood overlain with gold (1 Kings 6.23–8; 2 Chronicles 3.10–14), embroidered images on the Veil of the Tabernacle (Exodus 26.31), and other ornamentation (1 Kings 6.29–35). When the presence of God abandoned the Temple in Ezekiel's vision, it was the cherubim who bore him (Ezekiel 10). This was clearly understood throughout the Old Testament as the most important and characteristic of their functions: 'who sits upon the cherubim' is a much-repeated epithet (1 Samuel 4.4; 2 Samuel 6.2; 2 Kings 19.15; Isaiah 37.16; Psalms 80.2; Psalms 99.1), relating to the concept of the Ark as the 'mercy seat'. And, more vividly, David sang that the Lord 'mounted the cherubim and flew; he soared on the wings of the wind' (2 Samuel 22.11; Psalms 18.10). This was the circumstance of Ezekiel's first encounter with cherubim near the Kebar river: they bore on high a throne of sapphire, and 'on the throne was a figure like that of a man ... from the waist up he looked like glowing metal, as if full of fire, and from there down he looked like fire; and brilliant light surrounded him ... This was the appearance of the likeness of the glory of the Lord' (Ezekiel 1.25–8).

A number of features on the Angel face of the Shelford shaft clearly pick up on these characteristics of cherubim, and, previously anomalous, rather give its iconography a specific purpose. The creature has four wings, and the tall upper pair curve distinctively inwards in a way that evokes the cherubim's long-established function as overhanging and protecting the Ark of the Covenant throughout its long journeyings. The treatment of the upper parts of both pairs of wings, though a conventional differentiation of feather types, may in its round pellets embody a reference to the all-pervasive eyes ascribed to the creatures, just as eyes fill the equivalent zones of the wings of the elegantly drawn seraphim from the Arenberg Gospels in New York and Caedmon Genesis in Oxford (Pierpont Morgan Library, M. 869, f. 13r; Bodeian Library, MS Junius 11, pp. 1, 2, 17, 66: Ohlgren 1992, ills. 2.16, 6.8, 12.1, 16.1, 16.2, 16.12, 16.38). But eyes would be a possible addition in paint anyway. The creature duly exhibits human hands, which are detailed with notable delicacy, as if their identification was important. Little can be made of the plain legs or feet that protrude from the garment's hem: no doubt they are straight but their colouration would be key. If the creature were standing on a circular or elliptical surface, then that would represent the 'wheels' which feature so prominently in the cherubims' form (Ezekiel 1 and 10). Just possibly the puzzling diagonal fillet to the lower right, which Hill thought was the tail of a beast below the Angel's feet (Hill 1916a, 203), could have been a creature's tail and a reminder of the cherubims' tetramorphic form, which is so difficult to convey sculpturally or even graphically.

A rare attempt to overcome this difficulty — in this case in respect of the symbols of the evangelists (which are the New Testament manifestation of the cherubims' four aspects; see below) — is found in a later eighth-century gospel book from Echternach, signed by an English scribe (Trier, Domschatz Codex 61: Alexander 1978, cat. 26). Here, the miniature features the Man symbol, but in front of him, revealed by parted curtains, are the legs and paws of the Lion, the talons of the Eagle and the legs and hooves of the Calf (ibid., ill. 110). This recasting of the Old Testament's awesome cherubic tetramorph as four separate 'living creatures', and subsequently identifying them as symbols of the four evangelists, was set in train in the Revelation of St John (Revelation 4.6–9; 5.8–14; 7.11) and latterly enthusiastically developed in patristic writings and in the Christian art of the first millennium. The imagery was certainly popular in Insular art and perhaps especially within Ireland and in Irish-influenced works; but means other than the oddity tried in the Trier Gospels were typically sought in order to stress the vital unity of the four gospels, which a single tetramorph and its origin symbolized. These means were more to do with layout and association (Cronin 1995). Nevertheless, the link of the cherubic tetramorph with the New Testament and the later separated version of the evangelists' symbols, and vice versa, must always have been implicit — especially for literate clerics — and was sometimes made explicit, as in the four-winged Calf and Eagle symbols in the Book of Armagh, which deploy medallions of the other three symbols on those wings (Alexander 1978, cat. 53, ff. 68v, 91) and four-winged symbols on its four-symbol page (ibid., f. 32v). The explicit link between the Old Testament tetramorph and the Evangelist symbols carries through into the twelfth century, for example, in imagery drawn from Ezekiel that is found in the Floreffe Bible (BL MS Add. 17738) and contributes to that manuscript's distinctive linkage of New and Old Testament themes and images (Dodwell 1971, 162–3).

The aspect of the tetramorph displayed at Shelford, however, is that which took the precedence for Ezekiel — the Man, boldly bearded to identify it unequivocally. As Lang noted, it is difficult to find images of bearded angels or archangels (Lang 1988, 22); they should not exist. A rare pre-Conquest exception is the angel-like figure of Death — paired with Christ in a diagram within a mid eleventh-century psalter in the British Library (BL MS Cotton Tiberius C. VI, f. 6v: Temple 1976, ill. 303) — which does little more than prove the rule. One strand of traditional depiction of the Man symbol for Matthew in early manuscripts is a bearded type, and it seems to be principally Irish (e.g. Alexander 1978, ills. 14, 54(?), 104, 110, 226, 241), though the man who emerges from the curtain in the Matthew portrait of the Lindisfarne Gospels, in contrast to the 'imago hominis' of this picture, is certainly bearded and with the lively curly type of combined beard and moustache deployed at Shelford (ibid., ill. 28). Matthew himself is more commonly bearded, but that particular type of beard appears in the Cutbercht Gospel Book, with an Anglo-Saxon scribe, and at St Gall with an Irish pedigree, both in the later eighth century (ibid., ills. 181, 207), as well as in the bust at the top of Canon Table IV (f. 4) of the Book of Kells (ibid., 239). The fashion for forked beards and moustaches is also one that is prevalent amongst Hiberno-Norse figure sculpture of the early tenth century, as in examples from Old Malton, Yorkshire ER (Lang 1991, 196–7, ill. 736), and it is seen on several of the Irish high crosses, such as the Cross of the Scriptures at Clonmacnoise, a monument which is likely to have been erected around the time of the death of King Flann Sinna in 916 (Harbison 1992, i, 48–53, ii, fig. 133).

Under its right arm the creature holds a book, presumably the Old Testament or old Covenant, though the shape might just also be taken to allude to the Ark, traditionally guarded by cherubim. The raised left hand seems to narrow to a wrist and (damaged) hand in a similar way to the right. Then, rather than expanding to an unnaturally large hand raised in benediction as previously supposed (Lang 1988, 22), this normally small hand holds up or supports an object whose surviving outline and limited background detail suggest that this was a torch or flaming ball or area of incandescence. It is tempting, in view of the primary role of the cherubim, to identify it as an image of the glory of Yahweh's presence, as a man-like enthroned form emitting flames and light, as described by Ezekiel. This proposal would explain why this specific area of the sculpture was the subject of selective iconoclastic attack, like the faces of the Virgin and the Child on the opposite side of the shaft.

Date

A search for four-winged angels in Anglo-Saxon art as the potential basis for dating the Shelford shaft is unpromising. Angels with more than one pair of wings are rare. The fine images of such angels in the Gospels in the Bodleian Library (MS Bodley 155, ff. 93v, 146v) dating from around the year 1000 are seraphim (Backhouse et al. 1984, 55, no. 35; Temple 1976, cat. 59, ills. 177, 178). Other depictions of seraphim in later Anglo-Saxon manuscripts and metalwork typically have a more standard configuration of their six wings, matching Isaiah's famous description (Isaiah 6.2; Ohlgren 1992, ills. 2.16, 6.8, 12.1, 16.1, 16.2, 16.12, 16.38). In sculpture, the angel who stands at the top of cross-head no. 5 from Durham Cathedral Chapter House distinctively has two pairs of wings, and his identification as a cherubim might link with the series of evangelist symbols identified on the opposite face by Coatsworth and an overall apocalyptic iconography. The sculpture is dated variously to about the year 1000 and to the second quarter of the eleventh century (Coatsworth 1978, 85–96, pl. 1a; Cramp 1984, 68–9, pl. 43.205; see Ill. 187).

Although Hawkes was prepared to consider the Shelford shaft, alongside iconic Virgin and Child images at Nunburnholme and Sutton-upon-Derwent in the East Riding of Yorkshire, as related eleventh-century sculptures (Hawkes 1997, 132–3 and appendix, 134–5), such a late date seems improbable and Hawkes herself subsequently modified it to 'tenth- or eleventh-century' (2002, 36 n.12).

Jim Lang always linked Shelford with the group of cross-shafts to which he devoted particular study in Yorkshire, comprising the shafts with Virgin and Child at Nunburnholme and Sutton-upon-Derwent and the shaft section from St Mary Bishophill Senior in York (Lang 1991, 88–9, 189–93, 220–1, ills. 246, 711, 723, 868). Closely linked to the Nunburnholme monument is the shaft fragment from Newgate in York, which preserves only the halo of a hieratic figure, apparently a figure of Christ, beneath the encompassing arch (ibid., 105–7, ill. 342). This Newgate shaft is, according to Lang, one of the principal works of the sculptor he called the 'York Master'. All these monuments are clearly linked together stylistically through their use of a technique of sculpture of which Talbot Rice was contemptuous: '... it is poor and debased, and shows no signs of the new continental influence which is at the basis of most tenth-century figural work south of the Humber' (Rice 1952, 142). By continental influences, of course, he meant that it was free of the classicizing tendencies of Carolingian and Byzantine art. Scholars of Viking art, by contrast, have tended to see this style as a triumph of the barbarian aesthetic and Kendrick, at least, saw Shelford as the greatest achievement of this style of figure sculpture: 'The two surviving panels represent a completely successful translation of the once plastic figure-subject into an abstract ornamental style, and their uncompromising indifference to natural proportions and values ... enhance rather than detract from the power of the achieved pattern' (Kendrick 1949, 79). Of course, insofar as this view of Shelford's 'unnatural proportions' was based on the supposed difference in size of the Angel's hands or the appearance of the faces of Virgin and Child, its aesthetic judgment is undermined by our understanding of the iconoclasm the sculpture has suffered, as we have set out above.

While Kendrick's opinion has been widely shared amongst the specialists in Viking art, one cannot say there has been a similar agreement about the date of the Shelford monument. Its date has been such a puzzle, indeed, that Richard Bailey used it as the prime example of the 'flexibility' in dating that bedevils the whole subject, citing a label that apparently used to stand by the shaft in Shelford church, which assigned the carving to 'c.1100 (700)' (Bailey 1996, 13). Kendrick thought the monument was of early or mid eleventh-century date (1949, 78), and his dating has been followed by many subsequent scholars (e.g. Stone, L. 1972, 37). But Kendrick made the same links between the Shelford shaft and some of the Yorkshire monuments that Lang identified. He specifically noted the similarity with the Nunburnholme cross, which — perhaps as a consequence — he was inclined to date later than accepted opinion today would have it (Lang 1977, 86–8; 1991, 193). Lang's considered dating of these monuments suffers a little from the 'flexibility' of which Bailey complained; but he was inclined to place all of these monuments considerably earlier in date than Kendrick, with the Nunburnholme shaft (which was worked in at least two phases) being the earliest, perhaps originating in the late ninth century, though with panels such as the Virgin and Child belonging to the early tenth. Lang dates the remainder of these shafts either generally to the tenth century, or more specifically to the first half of that century (1991, 89, 107, 193). All, that is, with the exception of the shaft from Sutton-upon-Derwent, to which he gives a broad date bracket extending from the later tenth century to the mid eleventh. Partly, no doubt, because he can find parallels across many parts of England for some of the other details on the shaft, he sees this shaft as a 'revival' of the style of the earlier monuments rather than as a member of the original group (Lang 1991, 220–1). The Shelford Virgin and Child shares some similarities with that image at Sutton, such as the basic 'complementary' pose of the Child, the small cushion on which the Virgin's feet are placed, and the rigid parallel folds with which the drapery is depicted. Nevertheless, it is enormously more skilfully executed than that at Sutton, and, as Hawkes has spelled out (1997, 133), the Child at Sutton exhibits very different and distinctive gestures in raising his face and arms towards his mother, in the manner of later medieval Eleousa-type icons, rather than looking directly out at the spectator. As part of this major typological difference, the Christ-child at Sutton holds no book, which is a feature common to Shelford and Nunburnholme. There is no reason, then, to think that Shelford represents a 'revival' of an earlier tradition of figure sculpture. In fact, the Shelford Virgin has the same stylized halo-cum-hairdo with the curled ends that is seen surrounding the Virgin's head at Nunburnholme, and indeed on a number of other Yorkshire monuments that are or might be dated to the first half of the tenth century, such as the Leeds 1 cross or Ilkley cross-shaft 3 (Coatsworth 2008, 171, 198–202, ills. 361, 482–3; Coatsworth dates Ilkley 3 to the 'mid to late ninth century').

But the most significant link between the Shelford and Nunburnholme shafts does not lie in the similarity in detail between the sculpted figures, close though the Virgin and Child images are, but in the placement of the hieratic figures on both monuments beneath arched canopies decorated with precisely similar pellet and fillet design, and with the same three-piece respond moulding (Ills. 92–3; Lang 1991, ills. 709–24). The similarity between these two details is so close, indeed, that we might use it to argue that the Shelford shaft is not missing another hieratic figure above those surviving, but only a quarter-storey containing a lunette inhabited by a beast, immediately beneath the cross-head, as at Nunburnholme. This line of argument would reconstruct the complete Shelford cross-shaft as a monument standing about two metres high above the base but not including the cross-head (which might add a further half metre). Furthermore, this close comparison with the frame at Nunburnholme offers us a simple stratigraphic date for the carving of the Shelford shaft, as this carefully moulded pelleted frame was the work of the First Sculptor, on face D at Nunburnholme (Lang 1977, 76, 79, pl. 5a). This First Sculptor at Nunburnholme also created the most sophisticated figure sculpture on the monument (Lang 1991, 192), though it is in a quite different, continentally-inspired, style to the work at Shelford. However, this sculptor did not cut the Nunburnholme Virgin and Child image. That, according to Lang, was the work of the Second Sculptor, and his work can be shown to be stratigraphically later than the first master's. A third hand's work is later still. Lang then uses the different stylistic and iconographic associations of the work of these three sculptors to offer approximate dates for their respective working and re-workings (1977, 86–8; 1991, 191–3). Critically, the work of the Second Sculptor, though nominally Christian, plays syncretic iconographic games, paralleling Christian and Norse myth. Such work, Lang argues, is likely to date from the period between the capture of York by the Hiberno-Norse Vikings in 914 and the expulsion of Eric Bloodaxe in 954. Taken at face value, then, the work of the First Sculptor, who laid his shaft out in such a similar manner to Shelford, ought to date from prior to about 914, though how much earlier is hard to say: possibly not long. This might imply a date for the Shelford shaft around the year 900, or even a little earlier. Its architectural imposts — if legitimately termed 'stepped' or 'slab' imposts — are found in early Insular contexts such as the Franks Casket or the Codex Amiatinus and Lindisfarne Gospels (Wilson 1984, ills. 34, 37; Bruce-Mitford 1969, pls. VIII (1), XVI, XVII), and the classicizing Dewsbury shaft (Coatsworth 2008, ills. 196–7), which may support this proposition.

A final important Yorkshire parallel for the Shelford shaft should be mentioned. It is the stone of uncertain function — it might be a 'plaque' but is perhaps most likely to be a section from another shaft — at Newburgh Priory in the North Riding (Lang 2001, 179–80, ill. 653). This stone depicts two seated figures under a single arch but divided by a projection suspended from the frame above. Lang interprets these figures as two evangelists holding their own symbols in their laps, and he suggests that the small figure in the lap of that on the left is the 'winged man' of Matthew, whilst the figure to the right would be either Mark or Luke, depending on whether the quadruped in the charger was a bull or a lion. As Lang himself admits, however, this iconography is completely unparalleled in England, and the only even generally similar examples to which he can point in medieval art come from twelfth-century Languedoc. Is it possible that the figure on the left is actually a Virgin and Child and the 'winged man' actually the Christ-child? The figure lying across the lap does not have any wings; but it does hold a book. The combination of a recumbent, babyish Child supported across her lap by the Virgin and holding a book is rare, but paralleled in the ninth-century cross-slab from Brechin, Angus, where the Child is tightly swaddled (Hawkes 1997, 115, fig. 6; Hawkes calls this 'unique in an Insular context ... and apparently in Christian art generally'). The larger figure at Newburgh has a lightly inscribed zig-zag line, for example, which (if not secondary) might be interpreted as the coronet of Mary, but is difficult to explain if the figure is considered to be an evangelist. This suggested re-identification perhaps gains some support from the fact that Mary can very appropriately be paired with St John Baptist, as both are instrumental in the Incarnation; the one in physical, the other in spiritual terms. In the period with which we are concerned, this specific pairing was of such significance that it was given its own name: the Deesis (Pelikan 1996, 103–4). The Christological message it contained was of such importance that it was recurring image in Byzantine and Byzantine-inspired art (ibid.). There are also stylistic similarities between the two sculptures.

Early tenth century, perhaps about 900

Standish 1903, plate; Stapleton 1903, 90; (—) 1904, 151; Fellows 1907, 20–1; Godfrey 1907, 428; Stapleton 1911, 140–1; Cox 1912a, 182–3, pl. II; Stapleton 1912, 30; Hill 1916a, 203, pl. II; Guilford 1927, 167–8; Mee 1938, 255; Wood 1947, 18; Kendrick 1949, 78–9, pl. LI; Pevsner 1951, 157, pl. 34 (b); Rice 1952, 107, 142; Jope 1964, 108; Stenton 1970, 368; Stone, L. 1972, 37, 39 & plate; Pattison 1973, 218–19, 221, 223–5, 227–9, pl. XLIVa–b; Lang 1977, 85–6, 91; Pevsner and Williamson 1979, 308–9 & plate; Thorold 1984, 152; Kaye 1987, 28; Lang 1988, 22, 56; Clayton 1990, 176; Lang 1991, 37, 88, 192, 221; Lang 1993, 266; Billson 1994, 3 & plate; Sidebottom 1994, 133, 149, 268 & plates; Tweddle et al. 1995, 89, 220; Bailey 1996, 13; Hawkes 1997, 109, 116–17, 132–3, 135; Priestland 2000, 31; Lang 2001, 179; Hawkes 2002, 36 n.12

[1] The following are unpublished manuscript references to Shelford 1: Nottinghamshire Archives Office, DD/TS/6/4/7, p. 57 ('Notes on churches visited No. II' by Arthur Barratt of Lambley); BL, Add. MS 37552, f. 218–227, illustrations; Add. MS 37604, f. 386 (Romilly Allen collection).

[2] We owe thanks to Dr David Parsons for interim guidance on this.