Set in a modern stone base under west arch of nave north arcade

First recorded in the mid nineteenth century, reused as the lintel of the twelfth-century doorway between the west tower and nave of St Oswald's church (Stonehouse 1839, 405–6; Harding 1940, 28). The shaft had been set with the principal decorated face (A below) visible, facing east. Limewash that obscured its decoration was removed by Fowler in 1867 with brush, water and a wooden scraper, revealing part of a runic inscription, and adjacent stones were subsequently removed to expose the full length for study (Fowler 1867–70a, 187–8). In August 1869 the shaft was taken out for recording by Fowler, and reset so that faces A and C could be seen (Fowler 1867–70b, 378; Stephens 1884a, 187). The shaft had been moved to its present position by 1926, evidently after Cox's first edition in 1916 (Cox 1916, 113; Davies 1926, 11–12).

St Oswald

Generally good with the exception of the figural scenes of the lower three-quarters of face A, from which all the upper surface has been removed by diagonal tooling. A matching length of the narrow face D has been similarly removed, and the central part of the arris between faces D and C has been recut in a narrow chamfer that has removed the cable moulding. (Note that this feature and the relationships between the shaft's faces and sides have hitherto been incorrectly depicted in all drawn illustrations, cf. description below.) D would have been the exposed downward-facing side when the stone

The stone comprises the greater part of a slightly tapering cross-shaft of rectangular section. The break at the present summit cuts across decoration on all four faces and represents the original shaft's reduction for secondary use. There is no sign of mortise or tenon. Fowler recorded that the same is true of the base and assumed that it too represents a reduction from the original (quoted in Stephens 1884a, 187), but presumably by little since the interlace run on face C is complete. The shaft is decorated in low relief on all four faces. The four arrises were each worked in a narrow cabled border, which though damaged survives almost complete on the angles between faces A and B and faces C and D. The vertical fields were not otherwise formally divided into panels. The rather loose style of interlace that typifies the shaft may partly be a consequence of the coarseness of the stone.

The form of the cross-head is uncertain. At the same height on both narrow faces there is a trimmed-off scar approximately 15cm (6 in) deep that corresponds to the interruption of the cable-moulded border by symmetrical loops or scrolls on both broad faces: the cable border apparently turns outwards to go round these features. Whether these scars mark the location of a collar, the abutment of the ring of a ring-headed cross or even the junction of cross-arms is unclear from direct evidence. In either of the latter two cases the decoration of the uppermost part of the shaft would have been part of the cross-head.

A (broad): This face carries the most complex and the only figural decoration on the shaft. (i) At the top, a pair of confronted creatures, possibly birds, truncated by the cut end of the shaft. They are characterised by short, pointed, swept-back wings, fan (or fish) tails with the feathering marked by incised lines that survive more clearly on the right-hand figure, and necks or upper bodies that arch back in a distinctively exaggerated manner. Between them is a small circular disc with concentric incised inner ring and visible centrepoint. (ii) Below, and between the pair of indented loops that break up the cabled border, is a closed-circuit twist arranged to form a circle around a small plain disc at its centre. (iii) Below this, a pair of confronted human figures, shown in profile. Despite damage to the surface of the carving, it is evident that both figures are male and secular: both are prominently bearded, and their active stance, each with a forward leg raised and bent at the knee while the weight is taken on the straight rear leg, shows that their dress is some form of knee-length tunic or cloak with leggings or close-fitting trousers below. For the right-hand man, facial details of nose profile, an incised lenticular eye, a narrow band of hair across the top and down the back of the head, and incised downward curve marking the mouth and/or limit of beard can be discerned. His right hand is forward, the hand grasping the hilt of a sword as if to draw it from a scabbard steadied by the other hand. In contrast, the only clear details of the left-hand figure lie in the profiles of nose and beard. (iv) Beneath these, the figure of a horseman riding to the right and again portrayed in profile. This shows the rider to be bearded and with prominent nose like the figures above, but with no other surviving detail except one foot shown in profile below the horse's flanks. The horse is depicted in lively motion, not stiff-legged, but with two pricked ears, an incised eye, muzzle down, somewhat square-ended but with its mouth part open, and with a long tail. A deeply incised channel cuts diagonally down towards the mouth, perhaps depicting a harness that might also have been represented by a metal inlay. Beneath or across its fore-feet is a roughly V-shaped or forked feature that curves in sympathy with the inscribed ribbon in the field below. Behind the head and shoulders of the rider and below the feet of the left-hand figure in the scene above, a large irregularly oval feature almost fills the available space. No distinguishing details are visible. (v) The lowest portion of the face is occupied by a broad raised band or ribbon that is now extremely indistinct. It survives only down the right-hand side of the shaft and appears to curve over like an arch beneath the horseman. The left-hand side of this field has been completely removed, and all early drawings show it so. On the ribbon the runic inscription first recorded by Fowler is not readily visible, both because of its faintness and because it is usually hidden behind a church pew.

(The Committee is most grateful to the Rev. D. Schofield for allowing the pew to be temporarily moved in March 1994 so as to permit inspection and photography of the south face of the shaft.)

Inscription A fragmentary runic inscription runs left to right down the raised band towards the base of the stone. It reads:

5 10

Much of the original surface of this part of the stone has been cut away, leaving rough diagonal score-marks. This accounts for the loss of much of the inscription, and the scoring intrudes into the extant part, obliterating the bases of several characters, notably 1, 2, 5, 6 and 9. Of these 1, 6 and 9 cannot clearly be identified. 1 has a squarish bow at the top, which might belong to 'r', 'b' or 'w'. Although faint, the bow seems to be curving out towards the base, making 'r' the more likely reading. 6 also has a bow, this time more angular, at its top. Lower down there may be faint traces of another bow; they are obscured by the scoring, but must certainly be supplied, for 'b' is demanded by the context, giving 'bæcun', 'a monument'. This context likewise dictates that 9 is 'u' and not damaged 'y'. After 11 there may be a trace of the base of a stave, but no further identifiable runes or rune-fragments can be seen.

Early drawings by Fowler (1867–70a, 188) and Dodds (1869–70, 1, 6) indicate more at the beginning of the inscription than can now be seen. The drawings, however, agree neither with each other nor with what remains. Fowler shows a text beginning '-æælic', where 'æ' is today invisible and 'p' is not consistent with the fragments of 1, while Dodds draws '-nunlic', where 'n' is now absent, and 'un' inconsistent with 1 and 2. At the end of the inscription Fowler shows 'o' where I read 'æ', a relatively minor discrepancy (though the reading 'æ' is reasonably clear and secure). Stephens (1884a, 185), however, follows 10 'n' with 'bêa[...]', where 'b' conflicts with, and the rest expands upon, what can now be seen. It appears sensible to discount all of these earlier drawings.

The form of the 'c'-runes demonstrates that this is the Anglo-Saxon, not the Scandinavian, variety of runic script. The word 'bæcun' shows that the language is Old English.

B (narrow): (i) Decorated above the scar with a simple figure-of-eight twist pattern, truncated at the top. The strand has a deep median groove and a squared terminal. (ii) Below the scar, the whole field is occupied by a run of two-strand interlace in which a pair of U-bends is created on each strand alternately at their crossing and separated by glides to make an extended and rounded form of half pattern B (Cramp 1991, fig. 21; cf. Allen 1904, 266, 268, knot no. 8). The strands are neatly finished in a pattern E terminal below the scar, but are truncated at the base.

C (broad): (i) Decorated at the top with a complex and irregular interlacement truncated by the cut end of the shaft. Its lower limit coincides with the lower edge of the lateral scars. One free end of the strands terminates in an inverted animal head, with sub-triangular lappet and traces of an incised almond-shaped eye. Another disappears into the animal's open mouth. A third is bent sharply back across the shaft beneath the animal's head to delimit the lower horizontal edge of this motif. It ends in a triangular terminal or tail. (ii) The field below the lateral scars is filled with a run of four-cord interlace in a changing pattern, forming three large knots with thick heavy strands. At the top, one unit of turned half pattern F, finishing in a cross-joined terminal; below this, a unit of simple pattern F; and at the base a patternE terminal.

D (narrow): Severely defaced, but what can be discerned of the patterns suggests that it was decorated in a manner identical to face B, both above and below the scar.

Inscription The word bæcun, 'beacon, monument', is immediately recognisable: in the form bēcun it is well-attested on northern English memorial stones. In most of the other instances it appears as part of a formula, followed by æfter, 'after, in memory of', and this is probably the word introduced by 11 'æ' here. The Crowle inscription, however, departs from attested formula in what precedes bæcun. 'lic' probably represents Old English līc, 'body, corpse', producing a satisfactory compound 'corpse-monument' or 'grave-stone'. The compound is not attested elsewhere, but is not unlikely, cf. OE līcþrūh 'corpse-trough, sarcophagus', līctūn 'corpse-enclosure, cemetery'. On the other hand, -lic might represent the end of a preceding adjective, a possibility which is demonstrated by the existence of ælic, 'lawful', though a reading 'lawful (? legal) monument' is without parallel and unconvincing. It is hard to find attested alternatives that fit the possible patterns -rælic, -bælic or -wælic (though the reverse-alphabetical list of the Old English concordance (Healey and Venezky 1980) makes it possible to look), but -lic was a very productive suffix, and suitable words may have existed. There is no parallel in extant inscriptions for the use of an adjective preceding bēcun, but a compound noun is not usual in this context either (though the Bewcastle Cross may perhaps have had 'sigbæacn', 'victory monument'). On balance, a compound līcbæcun is the preferable reading, but the alternative, involving an unidentified adjective, cannot be excluded.

A form bæcun may represent an earlier linguistic stage than the usual bēcun (Campbell 1959, §§ 222, 225), but the difference between the spellings may be primarily of dialectal rather than chronological importance (Hogg 1992, § 5.99). Consequently it is difficult to draw conclusions from comparison with, for instance, runic 'bekun', and roman-script BEC- and BECUN, respectively on two stones from Thornhill and one from Dewsbury, both places about thirty miles west of Crowle in the West Riding of Yorkshire (Page 1973, 153–6; Okasha 1971, nos. 116, 30). Unfortunately, no other Old English texts survive for comparison from north Lincolnshire. The first 'æ' in the Crowle inscription is similarly inconclusive: it is likely to be an unstressed vowel and, if so, should be informative, for æ seems to have become e before about the middle of the ninth century everywhere except, perhaps, the north of Northumbria (Dahl 1938, 191–6; cf. Hogg 1992, §§ 6.48–9). This hardly tallies, however, with the clear influence on the Crowle stone of Viking art style, which presumably prohibits a date earlier than the beginning of Scandinavian settlement in the last quarter of the ninth century. Evidently a single rune in a damaged context cannot outweigh these indications, and the 'æ', if it is unstressed, should perhaps be regarded either as a peculiarity of an otherwise unrecorded dialect or as an archaism. Nonetheless, the two potentially early linguistic features might be noted in conjunction with the likelihood that the Crowle text is related to other inscriptions that record the Old English 'bēcun formula', in either runes or roman script, a group which is usually attributed to the eighth and ninth centuries. There may be some reason here cautiously to favour a date early rather than late in the period of Scandinavian influence on eastern England.

A complication is introduced, however, by the layout of the inscription. Memorial texts on Anglo-Saxon stones are often cut in rectangular panels, and the Crowle stone is unusual in setting its runes within a raised band. What survives shows that this band curved round at the top before running vertically down the side of the stone, and this immediately suggests parallels in the Danish area, especially Jutland and Skåne, where the arch-shaped curving band is a common arrangement on rune-stones (Jacobsen and Moltke 1941–2, I, cols. 823–4; II, passim). Yet if this parallel is relevant it might have surprising chronological consequences, for such curved bands are usual only from the late tenth century, and have been considered diagnostic of the period c. 1000 (Jacobsen and Moltke 1941–2, I, cols. 1024–5; Moltke 1985, 229–30).

There is insufficient evidence in the inscription to reconcile these faint and contradictory indications: no date can be suggested on epigraphical grounds.

The standard assessment of this shaft has been that it is the most purely Scandinavian sculpture that survives in Lincolnshire. At that general level its decoration finds plentiful parallels in crosses of tenth-century date and Scandinavian influence throughout Northumbria, discussed in detail below.

One aspect of its form nevertheless is problematic and perhaps surprising. That is the evidence of the broken borders, and the corresponding paired loops on faces A and C and trimmed-off scars on B and D, which together indicate that lateral projections have been removed in its reuse. It seems highly improbable, either in respect of the decorative layout or the stone technology, that these were the cross-arms of a crosshead. More likely they were minimal protrusions whose analogy, and perhaps source, lies with the Mercian tradition of rectangular collared shafts seen, for example, in a pre-Viking form on the churchyard cross at Bakewell, Derbyshire (Routh 1937, 5–7, pl. IIa–c), and in the later pre-Conquest period on members of the South Kesteven shaft group in Lincolnshire such as Creeton 1 (Ills. 124–7) and the Elloe Stone (Ill. 171–2, 176–8). Untypically of this tradition, the projection is not carried around the shaft as a collar: but the break is marked by the angular termination to the interlace on face C, and by the focal position of the closed-circuit knot on face A. A fragmentary Viking-period shaft in Derby Museum also has traces of analogous loops but the surviving broad decorated face is uninterrupted by any raised collar. In this example too, precisely as at Crowle, the border moulding turns out at right angles, presumably to pass round the protrusion. A number of shafts in Cumberland, Bromfield 2, Penrith 4, Rockcliffe 1 (Bailey and Cramp 1988, ills. 173–6, 489–99, 539–42), have collars or off-set panels, but typically carried around the shaft. It may therefore be that this feature suggests a non-Yorkshire, but not a non-Scandinavian connection.

More specifically, the occurrence of the runic inscription on the southern fringe of the principal distribution of post-650 rune-stones (Page 1973, 29 and fig. 7) confirms Crowle's northern affiliations. The inscription appears to have been without word division and was set in the Norse manner within a curved ribbon, reading clockwise, in a form better paralleled on Scandinavian rune-stones than Anglo-Saxon and with possible chronological implications that Dr Parsons points out above. The runes, however, are Anglo-Saxon not Scandinavian, and the language Old English. Parsons concludes that līcbæcun, 'corpse-monument, memorial stone' is a likely interpretation of what remains legible, but that -lic could alternatively represent the end of an adjective (see above). In the latter case the cross would thereby still be designated a 'monument', bæcun, carrying a commemorative message, but not necessarily a funerary monument by its own description.

The details of the shaft's decoration also confirm its northern and Scandinavian affiliations. The run of simple repetitive interlace and combination of motifs on the lower part of face C might be akin to what is found on Lincolnshire monuments such as the Brattleby shaft (Ills. 60–4, 66–7) or the mid-Kesteven covers (Fig. 9), and the figure-of-eight twist at the top of faces B and D similarly; but rather, this is a case of common motifs of wide contemporary currency. Unlike Brattleby 1 and the covers, Crowle is not formally divided into panels. The half pattern interlace of 'slipped knot' form on faces B and D with the looseness imparted by its glides finds some analogy in Cumbrian pieces such as Penrith 1, St Bees 1 and Waberthwaite 1 (Bailey and Cramp 1988, ills. 477, 545, 578). The figure carving and the more complex interlace at Crowle perhaps suggest that this does not simply reflect difficulty in carving an intractable stone. The complex interlace terminating in a dragon's head at the top of face C has its best analogy in Jellinge-derived Yorkshire material of the later ninth and early tenth century, not least the York city grave-covers (Lang 1978b; id. 1991, especially 39–40), one Lincolnshire example of which exists at Holton le Clay as a coastal import (Ill. 203).The figural elements of face A point to similar connections. The paired confronted ?birds at the top may be compared quite locally with those on the shaft at Nunburnholme, Yorkshire ER (Lang 1977b; id. 1991, ills. 721–4): but more widely, too, as a pre-Viking sculptural motif – at Nunnykirk, Northumberland (Cramp 1984, pl. 207, 1192) and Knells, Cumberland (Bailey and Cramp 1988, ills. 425–6) – that persisted into the Viking period in the north and Scotland (Bailey 1974, I, 41–2). Whether a Christian iconography as souls or angels was intended is doubtful. Certainly, in contrast to their earlier suggested interpretation as Adam and Eve, or as the eremitic saints Paul and Anthony, the confronted figures below are clearly secular. They are reminiscent in stance and relationship of the pre-Viking secular pair from St Mary Bishophill Junior in York (Cramp 1982, 12, pl. 7; Lang 1991, ill. 216), but in contrast to their half-turned or three-quarters pose, are strictly in profile in the manner of Viking-period depictions, and are heavily bearded and with stylised hair. Paired confronted secular figures portrayed in profile occur on two crosses (nos. 3 and 11), both of tenth-century date, among the large collection of pre-Conquest material at Sockburn, co. Durham (see Ills. 484–5; Cramp 1984, pls. 130, 710, 712, and 132, 720).

A Christian interpretation has also been sought for the rider below, either Christ's entry into Jerusalem or the flight into Egypt. Although the latter scene is known in Romanesque sculpture with a star prominently occupying the field above the rear of the horse in the position filled by the unexplained oval lump at Crowle (e.g. at Fleury and at Rucqueville in Normandy – Zarnecki 1966, pl. IX), the figure of Mary is typically shown in a frontal pose as if side-saddle on the beast, and cradling the child. The Crowle horseman is in strict profile and bearded, and like the scene above, this image is evidently secular/non-Christian: its analogy is the figure of the warrior horseman, whether in procession or alone, that is a recurring image in the Germanic and Viking world (Cramp 1982). Again, specific tenth-century analogies are best illustrated in stone at Sockburn (Ill. 484), but also in contemporary metalwork like the brooch from Fulmodestone in Norfolk (Margeson 1997, 12, fig. 12). The feature behind the rider's head may have been a symbol or space-filler like that in this position relative to the horseman on the picture stone from Lillbyärs, Stenkyrka, Gotland (Cramp 1982, pl. 8b).

The general secular Scandinavian and lordly milieu of this decorative scheme seems clear. It may be a further possibility that the series of scenes formed a coherent theme recognisable to contemporaries though lacking enough strikingly distinctive features to be unequivocally intelligible to us now. If so, it is most likely to be part of the story of Sigurd from the Vǫlsunga Saga (cf. Bailey 1980b, 116–25; Finch 1965), which includes the elements of birds, the encounter of Sigurd and Mimir, and Sigurd's journey on his horse Grani to kill Fafnir. Might the ribbon below the rider itself have been the dragon Fafnir, with runic inscription on its body as, for example, on Rasmus rock, Jäder, Södermanland (Bailey 1980b, 117, pl. 30)? The use of Sigurd as the pre-eminent image of Scandinavian lordship and Norse identity has been emphasised by Bailey (ibid., 122–4) and by Lang (1976). The latter has also argued that a period of Norse colonial expansion is directly reflected in the stone sculpture of Yorkshire, mediated through the city of York and limited in time by the creation and collapse of the Norse kingdom of York to the early tenth century before c. 950 (Lang 1978c; 1989; 1991). In date and iconography, the Crowle shaft provides an independent association of this sort. It, like the cross-shaft and -head from North Frodingham in eastern Yorkshire that figures prominently in Lang's arguments (1991, 187–9), is carved from Millstone Grit. This stone type is alien to Lincolnshire; and its pre-Conquest use in the hinterland of York is assumed to result entirely from redistribution of Roman masonry from the city rather than from new quarrying in the Knaresborough area (Morris 1988; also Buckland 1988). Such Roman monumental masonry in Millstone Grit is known to have been available and reused in churches and sculpture along the south bank of the Humber estuary (as at Barton (q.v.), Broughton and Winteringham), and a specific link to York within its Lincolnshire distribution is provided by the Metropolitan school grave-cover, Holton le Clay 1 (p. 178). Crowle's location in Axholme gives it equally ready access to the Humber basin's water transport network and makes plausible the possibility that the origin and production of the cross lay in York (Fig. 22).

Early to mid tenth century

Stonehouse 1839, 405–6, fig. on 406; Fowler 1867–70a, 187–90, figs. on 188, 189; Fowler 1867–70b, 378–9; (—) 1869; Dodds 1869–70, 1–8, figs. on 1 and 4, pl. I; Haigh 1869–70, 197; Keyser 1883, 175; Stephens 1884a, 185–8, fig. on 185; Stephens 1884b, 125–7, fig. on 126; Allen and Browne 1885,356; (—) 1885–6b, viii–ix; Allen 1889, 207; Tyack 1891, 72–8, fig. on 75; Stephens 1894, 36; Chadwick 1894–9, 103; Fowler 1896, 3; (—) 1905–6b, xiii; Davies 1914–15, 143–5; Fowler 1914–15, 180–1; Davies 1915, 53; Stevenson 1917, 80; Cox 1924, 113; Hunt 1924, 112–14, 117, pl. facing 107; Clapham 1926, 5; Davies 1926, 11–12, fig. 1; Collingwood 1927, 135; Dudley 1931, 62–3; Arntz 1938, 89; Harding 1940, 28, and fig.; Rice 1952, 141; Butler 1961, 18, fig. 4; Marquardt 1961, 38–9; Pevsner and Harris 1964, 26, 224; (—) 1971, 8; Page 1971, 176; Page 1973, 29, 134, 135–6, 140, 145–6, 157, 219, 221, 226; Stafford 1985, 187, fig. 69; Higgitt 1986b, 131, 134, 147; Pevsner et al. 1989, 43, 242; Stocker with Everson 1990, 88