Cross-head, type A9/B9, with ring, type 1(a); cross-shaft, type g, set in its (presumably original) stepped socket. The lower part of the cylinder is undecorated; the upper part is covered with multiple ring-chain.

A (west): The ring on the head is decorated with a three-strand plain plait bordered by a roll moulding. At the centre is a boss enclosed within a frame formed by mouldings which break from the borders of each arm to cross and link to the borders of the adjacent arm. Within each arm is a triquetra formed by a strand with median-incised line. The arms are framed by a roll moulding but there is no border separating the lower arm of the head and the upper part of the shaft.

The rectangular section of the shaft below is decorated with a single full-length panel bordered by a roll moulding. At the top is an animal's head with outlined jaws, round eye and single hollowed ear; each jaw carried a large tooth. To this head is attached a form of ring-chain (G.I., fig. 26 cvii) executed in double strands. This links below to two parallel strips of three-strand plain plait which each terminate at the bottom in downward-biting zoomorphic heads of identical type to the animal at the top of the panel. Immediately below these two heads is a staff held horizontally across the panel by a human figure who is depicted as standing on the left-hand moulding. He wears a belted kirtle and holds a horn in his outstretched left hand; his pear-shaped head is sunk deep into his shoulders. Below him is a horseman shown upside down. This figure also wears a belted kirtle and is drawn full-face with his head sunk into the shoulders. His left hand rests on the horse's neck whilst his outstretched right arm holds a spear reversed over its rump. Within the curved scallop at the bottom of the panel lies a pigtailed human figure, with arms and legs manacled. Around his neck is a cord which is knotted with the body of a snake whose head is set above that of the captive. Over the bound man is the kneeling figure of a pigtailed woman holding out a bowl.

B (south): There are traces of a cabled border moulding on the ring of the head whilst the end of the arm carries a panel of framed four-strand plain plait. At the top of the full-length shaft panel is a three-strand plain plait ending below in a spiral. Its upper end terminates in a beast head with rounded contoured eye, single hollowed ear, and fanged jaws similar to those on the west face. Below is a run of ring-chain, (G.I., fig. 26 cvii) topped by an animal's head with contoured, gaping jaws, round eye and a single hollowed ear. The head and toothless jaws are bound by a ring. Underneath is a horned quadruped (stag/hart). Below, set sideways in relation to the shaft, is a wolf/dog with ear and tail whose legs are caught in a tangle of interlace formed by thin strands. Below this is a belted horseman shown full-face with pear-shaped head sunk into his shoulders. His left hand holds a bridle and his outstretched right arm grasps a spear reversed over the horse's rump. The feet of the horse are set over a horizontal strip of three-strand plain plait whilst at the base of the panel is a creature, seen in profile, with pointed, open jaws, round eye and peaked forehead whose lower parts are interlaced with his arm.

C (east): The ring is decorated with three-strand plain plait bordered by a roll moulding. At the centre of the head is a boss enclosed within a frame formed by mouldings which break from the borders of each arm, to cross and link to the borders of the adjacent arm. Within each arm is a triquetra. There is no division between the ornament on the head and the decoration on the full-length shaft panel below.

At the top of the shaft is a four-strand plain plait terminating at both ends in near-identical beasts' heads each with round eye, gaping outlined jaw and single hollowed ear. The upper beast has a tooth in each jaw whilst the lower animal has a forked tongue. This tongue is split around one leg of a human figure which is placed on the lower jaw of the beast. The man's other leg is set on the right-hand border so that he stands sideways in relation to the shaft. The figure wears a belted kirtle, presses with his left hand against the animal's upper jaw and holds a staff/spear in his outstretched right arm. Below is a run of ring-chain (G.I., fig. 26 cvii), executed in double strands.

The scene below is framed by a rectangular cabled moulding and contains the figure of the crucified Christ with arms outstretched. He is shown full-face with pear-shaped head sunk deep into his shoulders and is dressed in a short belted kirtle which dips to a point at the corners. A moulding, representing the stream of blood, runs from the figure's right side down to the point of the kirtle. Beneath this point is the head of a spear whose shaft passes under the lower frame. This spear is grasped by a profile figure, wearing a short belted kirtle which dips to points at the corners. Facing, to the right, is a female figure also shown in profile, with a trailing dress and knotted pigtail. She carries an horn-like object with bulbous base.

Underneath these two figures, within the curved scallop at the base of the frame, are two ribbon beasts, their bodies knotted together, each with open contoured jaws, round eye and single hollowed ear.

D (north): The ornament on the head is identical to that on side B. At the top of the single shaft panel is a triquetra attached to a vertical rod, which terminates below in an animal's head with a round outlined eye, outlined fanged jaws, and single pointed, hollowed ear. Eight wing-like features are attached to this rod by free rings. Below the beast's jaws are two horsemen set one above the other with the lower one depicted upside down. Both are belted and are shown full-face with pear-shaped heads sunk into the shoulders. Each has the left hand on his horse's neck whilst the outstretched right grasps a spear reversed over the animal's rump. The rest of the panel is filled by a four-strand plain plait with median-incised lines; drill-holes mark the interstices of the strands.

In many ways this is an unique monument, standing alone among English Viking-age crosses not only in its size and complete survival but also in the quality of its carving and its iconographic inventiveness. Though it is conveniently grouped with round-shaft derivative crosses like those at Penrith or the Peak District group (Pape 1945–6) it must be emphasized that in such details as its slender proportions, the organization of its ornament, and its handling of the junction of the cylindrical and squared sections, it stands apart from all other surviving carvings.

There are inevitable difficulties in interpreting the iconography of this carving. There is, first, the problem of establishing whether features which might seem purely ornamental carry significance for the total meaning of the cross. Secondly, the lack of panelling often makes the boundaries between scenes, and the relationship between figures, difficult to establish. Thirdly, the Scandinavian literary sources which give details of the episodes apparently depicted on the cross are far distanced both geographically and temporally from Viking-age Cumbria (Davidson 1950, 130 ff.; Berg 1958, 34; Bailey 1980, 103). In spite of all this, however, it is possible to discern the basic iconographic scheme.

The only clear Christian scene on the cross is the Crucifixion depiction on the east face, though even its iconography does not fit easily into conventional classifications. Certain of its details are only rarely found elsewhere among pre-Norman sculptures in Britain. The flow of blood, for example, is shown on a cross from Maghera in Ireland (Porter 1931, 58) but English representations of this motif are limited to manuscript occurrences (e.g. Wormald 1952, pl. 21). The snakes beneath the attendant figures, if indeed they are part of the passion scene, may be a double-headed rendering of the defeated devil familiar in Carolingian and later art (Beckwith 1964, pls. 56, 189, 126; Ferber 1966, 324). If so, there are only two other comparable English examples in sculpture, at Bitton in Gloucestershire and Kirkdale in Yorkshire (Taylor and Taylor 1965, i, fig. 33; Collingwood 1907a, 345, fig. a), but Gosforth alone has the double-headed form of serpent. Rare also is the crossless Crucifixion for, apart from its repetition on Gosforth 5 by the same artist, it is otherwise only found at Bothal, Northumberland (Cramp 1984, pl. 159, 824) and (significantly, as will be seen below) on the newly discovered Penrith plaque (no. 11).

Perhaps more important than the use of such rarely paralleled details is the unexpected combination of the subsidiary figures. Longinus with his spear is usually partnered by Stephaton, yet here is matched with a female figure. Berg (1958, 31) suggested that she was intended as Ecclesia with a chalice and that the Gosforth artist had selected his participants from the fuller groupings familiar in Carolingian art. Yet Ecclesia should properly be placed so as to catch Christ's blood and at Gosforth is on the wrong side to do so. Nor does she hold a chalice. I have argued elsewhere that the object is an alabastron, the symbol of Mary Magdalene, and that both attendant figures are types of the converted heathen and the establishment of the Church (Bailey 1974a, I, 320–1; idem 1980, 130). This pairing would thus link with the 'heathen' nature of the rest of the ornament on the cross.

At least three other scenes are certainly non-Christian and it has long been recognized that they can all be associated with the Scandinavian mythological concept of Ragnarök (Olrik 1902). The figure with arm and foot in the beast's jaw on the east face is most convincingly identified as Viðar, for the scene exactly matches the manner in which Óðinn's son avenged the death of his father at Ragnarök as described in Snorri and Vafƿrúðnismál (Neckel 1962, 55; Jónsson 1907, 100; de Vries 1956–7, ii, 275–6). This depiction is unparalleled in surviving art. Parker was also surely correct in identifying the figure with the horn on the west face as Heimdall, the watchman god, with his Gjallarhorn (Parker 1896, 49; see Pering 1941; de Vries 1955; idem 1956–7, ii, 238–44; Turville-Petre 1964, 147–55). Gosforth provides the only clear depiction of this god in Insular sculpture, though Icelandic portrayals are recorded (Jónsson 1907, 133–6). Literary sources do not, admittedly, refer to any specific encounters which might explain Heimdall's confrontation with two beasts here at Gosforth, but Snorri certainly saw him as engaged in the final battle of Ragnarök (Jónsson 1907, 101).

At the bottom of the west face is a scene first identified by Calverley as showing the bound Loki with Sigyn alongside him (Calverley 1883b). The only other illustration of this episode occurs on a Gotland stone from Ardre (Wilson and Klindt-Jensen 1980, 79, fig. 42). The figure of Loki is one with a complex literary evolution (de Vries 1933; de Vries 1956–7, ii, 255–67; Dumézil 1959a, 94ff; idem 1959b; Holtsmark 1962) but what is significant about him for present purposes is that he is described in Völuspá as leading the forces of evil on the day of Ragnarök. Moreover, the onset of the final battle is signalled by his escape in Baldrs Draumar, whilst Snorri narrates the story of his binding immediately before his account of Ragnarök. The bound Loki is thus closely linked to the Ragnarök theme in literary sources.

At least three scenes on the cross are thus associated with the Doom of the the Gods. Further episodes from the literature have been invoked to explain other parts of the ornament: the winged beast on the north face as Surt; the horseman with bridle on the south face as Óðinn with Mímir below and the wolf Garm above. Such explanations are a credit to scholarly ingenuity (and may be right), but they do not carry universal approval. It might, however, be noted that the presence of so many riders chimes with Snorri's account of the combatants riding to the battle (Jónsson 1907, 97 ff.), whilst zoomorphic ornament is a highly appropriate means of representing the monstrous forces of evil whom the gods engaged.

It is reasonable to claim that the positioning and relationship of the scenes are deliberate, even if now only partly explicable. Thus it is unlikely to be sheer coincidence that Heimdall and Loki appear on the same face of the shaft for there seems to be evidence that they were seen as enemies of long standing (Jónsson 1907, 101, 133–6). More significant perhaps is the larger patterning: the Crucifixion (the end of one world and the beginning of a new) is set alongside – and by implication compared and contrasted with – the end of another world drawn from Scandinavian mythology. Further, the ease with which Christian liturgy and teaching moved from contemplation of the Crucifixion to evoking Christ's second coming (Bailey 1980, 129, 163–70) suggests that the concept of the Christian Doomsday is present also in the theological patterning of the cross.

This granted, a whole series of parallels and contrasts can be brought into play. Earthquake, fire and the summoning horn all play their part in the end of the world in both Christian and Scandinavian traditions. Christ lived on after the Crucifixion; is it coincidence that Viðar, who is shown on the same cross-face, also survived the holocaust? Is it not also significant that Mary Magdalene, depicted with pigtailed hair, holds a curved alabastron whilst Sigyn, with her pigtailed hair, holds a curved bowl?

The patterning of the cross, if correctly interpreted here, thus shows an original mind at work, exploiting links and contrasts in a manner which reflects a radical theological approach which would otherwise never have been suspected in Viking-age Cumbria.

The date of the cross is dependent upon analysis of the style of its carving and clearly this owes much to Scandinavia. The strength of that non-English tradition can be measured first by the lack of Insular panelling; in this respect Gosforth 1 is much nearer to Scandinavian tastes as represented in the perishable media of wood and tapestries (Shetelig 1920, figs. 19, 21, 22, pl. vi; Krafft 1956), and indeed the whole manner of carving is evocative of wooden techniques. It can be seen also in Mary Magdalene, whose trailing dress, pigtail and proffered horn-like object repeat exactly familiar details of Scandinavian presentations (e.g. Ills. 692–3) which are in turn reflected in a more stylized form on the Isle of Man (Kermode 1907, nos. 99, 105). Such Scandinavian links do not of themselves necessarily imply a dating of the cross to the early Viking period, though they might seem more likely at this time. There are, however, other hints that this is the appropriate context. The use of the Borre-style ring-chain motif must imply a tenth-century date (Wilson 1978, 141) whilst the same conclusion can be drawn from the fact that the partial outlining of the jaws occurs at Oseberg (Shetelig 1920, figs. 19, 21–4, 26–7) and in the Cumbrian circle-head school. A tenth-century date thus seems the most likely context for this carving and to the same period must belong those other Gosforth sculptures (nos. 2, 3, 5, 6, and 7) which share such features as the crossless Crucifixion, figural styles, and interlace types (Bailey and Lang 1975).

Whilst a Scandinavian background, and models in fabric, wood and metal, contributed much to the art of the Gosforth cross, another possible source has recently been suggested by the discovery of the stone plaque, Penrith 11. It is argued below (pp. 140–2) that this small slab depends upon an Irish openwork plaque, possibly attached to a book or relic, circulating in Cumbria in the tenth century. The Penrith stone presents the crucified Christ dressed in a kirtle which dips to points at the corners. It also shows him without any supporting cross but with his hands overlapping the frame of the scene. These features are all represented on Gosforth 1 and it seems highly likely that the eclectic Gosforth master has adopted them from the same Irish metalwork model on which the Penrith plaque is so clearly dependent.

'Carbo' 1799, 833, pl. facing 833; Lysons and Lysons 1816, cci–ccii, pl. facing cci; Jefferson 1842, 302–3; Whellan 1860, 391; Stuart 1867, 18, 124, pls. xxiv, xxv; Allen 1878, 356; Calverley 1883b, pl. facing 143, figs.; Calverley 1883c, pl. facing 373, figs. on 376, 380, 382, 387, 388, 391, 394; Black 1883–6, 125; Stephens 1884–9, 7–26, figs. 13–18, 20–1; Allen 1884a, 158, 280, 394; Browne 1885a, figs. ii, iii; Parker 1886; Allen 1886, 333; Honeyman 1886; Allen 1887b, 250; March 1891, 86, 89, pl. xii, fig. 5; March 1894; Allen 1892–5, 146; Kermode 1892–6, 356, 363; Parker 1896, 1–68, pl. facing 6, figs. on 27 and facing 28, 41, 55; Bugge 1899, xlviii–xlix, lxiv–lxvi; Calverley 1899a, 139–67, 291, 294, figs. on 138, 144, 146, 149, 154, 156, 157; Collingwood 1901a, 263, 266–70, figs. on 266, 267, 268, 269; Collingwood 1901e, fig. on 312; Parker 1902, 85; Olrik 1902, 161–7, 243–4, figs. 1 and 3; Collingwood 1903a, 387; Collingwood 1903b, 266; Stevens 1904, 47, 88, 90; Kermode 1904, 12, 29; Bugge 1905, 376; Taylor 1906, 352–4; Collingwood 1906–7a, 128–36, figs. on 131, 137, 138; Collingwood 1907a, 275; Collingwood 1907b, 154; Kermode 1907, 40; Collingwood 1908, 219; Collingwood 1911a, 273; Collingwood 1912b, 160; Prior and Gardner 1912, 116, 117, 127, figs. 96, 106; Collingwood 1915a, 203, 214, 242; Parker and Collingwood 1917, figs. on 98, 100, 101, 103, 104; Collingwood 1918, 49, fig. 33; Brøndsted 1920, 221; Scott 1920, 104, pl. facing 104; Collingwood 1923b, 5–6, pl. 1; Brøndsted 1924, 79; Reitzenstein 1924, 169–77, figs. 2–6; Cook 1925, 305, figs. 195–6; Parker 1926, 64–9, figs. on 65 and 67; Collingwood 1927a, 6, 104, 155–7, figs. 12, 184; Collingwood 1927b, 175–6; Collingwood 1928b, 408; Collingwood 1928c, 326; Clapham 1930, 133–4, pl. 57; Porter 1931, 112; Saunders 1932, 17; Collingwood 1932b, 85–6, 312; Shetelig 1933, 223; de Vries 1933, 180–1; Routh 1937, 9–10; Baum 1937, 106, pl. LVI; Brown 1937, 97, 122, 212; Gardner 1937, 39, figs. 27–8; Kendrick 1941b, 10–11, 13, pl. facing 9; Pape 1945–6, 43; Shetelig 1948, 90; Kendrick 1949, 59, 63, 68–70, 72, 91, pl. XLIV; Davidson 1950, 130; Fair 1950, 93–4; Gardner 1951, 34, figs. 45, 46; Rice 1952, 133, pl. 24a; Shetelig 1954b, 128, 131; Stone 1955, 32, 239, pl. 19a; de Vries 1956–7, ii, 142, 276, 332, 397, pl. XVII; Radford 1958b; Berg 1958, pls., figs.; Fisher 1959, 73, pl. 31a; Høst 1960, 449; Davidson 1964, 173, 179, 207, 208, 221; Radford 1967b, 176; Pevsner 1967, 16, 128, 130, fig. on 29; Wilson 1967, 46; Wilson 1968, 301; Davidson 1967, i, 28–9; Davidson 1969, 104, 123; Anker 1970, 417; Lang 1971, 156; Foote and Wilson 1970, 302; Lang 1972, 236, 243, 247; Martin 1972, 72–3; Cramp 1972, 147; Coatsworth 1973, 235; Lang 1973, 17, 23; Pattison 1973, 231; Bailey 1974a, I, 295, 316–35, 337, 341, 347, 355, 358–9, 364, 369, 370, 373, 375, 379–80, II, 118–25, pls.; Johansen 1974, 115; Bailey and Lang 1975, pls. xxviii(a). xxix(a), fig. 1; Lang 1976a, 90; Wilson 1976a, 502; Wilson 1976c, 399; Cramp and Lang 1977, 3, 25, pl. facing 1; Loyn 1977, 65; Wilson 1978, 141; Bailey 1978, 177; Lang 1978c, 13; Smyth 1979, 265, 270, 271, 283; Coatsworth 1979, I, 261–5, II, 21–2, pl. 125; Graham-Campbell 1980a, 75, 141, pl. on 74; Graham-Campbell 1980c, 158; Wilson 1980a, pl. facing 65; Fell 1980b, 187; Bailey 1980, 21, 54, 79, 125–31, 132, 133, 134, 136, 138, 146, 176, 215, 230, 234, 236, 242, 255, figs. 3, 23, 60, pl. 32; Wilson and Klindt-Jensen 1980, 106, 107, 109, pls. XL–XLI; Owen 1981, 170–1, fig. 32; Firby and Lang 1981, 21; Bailey 1981, 83, 84, 87, 94, pl. on 136; Lang 1982, 60; Wormald 1982, 163, pl. 163; Cramp 1982, 13, 17, pls. 14, 17; Lang 1983, 186; Wilson 1983, 185; Bailey 1984, 13; Cramp 1984, 20, 30, 31, 141, 143–4; Wilson 1984, 150, ill. 186; Bailey forthcoming a; Bailey forthcoming b

==R.N.B.