Loose at west end of nave

The Rev. John Wild, writing in 1871 and recording activity in the previous decade, found, not the grave-stone mentioned by Gough (1789, 249), '... but another of perhaps greater interest. It was built into a barn-wall, and on its outer surfaces were traces of an inscription. By the kind permission of R. Heathcote, Esq., I removed it, and found that it was part of the shaft of a cross. The letters are probably – DE NAPNO GRAP... Further search brought to light two other and smaller pieces of the same cross. On one of them was an incised star or flower...' (Wild 1871, 134). The stone from Mr Heathcote's barn is clearly 2c; the remaining two must be 2a and b. The restoration account of 1898–99 records the discovery of portions of 'a richly carved Norman cross' ((—) 1898–1900, xlviii–ix), but the Victorian restorers had presumably merely rediscovered these stones, and the following year all three fragments were in the sacristy, and were thought to be 'Robyn Hudde's Cros' as recorded by Leland (Cotton-Smith 1900, 5; Smith 1907, 23).

2c is in good condition with only slight weathering although it is abraded at the upper end. 2b is very badly abraded on every face. 2a is also badly abraded, though not as damaged as 2b.

There can be no doubt that these three sections came from the same original monument. Stones 2a and b fit together and there are dowel holes in the single section formed by 2a and b which match dowel holes in the end of stone 2c. Both sections also exhibit a similar pattern of recutting. When laid out the three pieces fit together with only a small part missing between 2b and c. All four angles of the shaft thus reconstructed were originally decorated with a finely wrought tight cable moulding.

The shaft was originally composed of two stones (at least) which were held together by two metal dowels sunk in lead between stones 2b and c. The shaft was evidently first reused as a single item before being broken into two fragments and then into three. This is clear from the rough chamfer which runs down the length of one arris of the shaft. The shaft was then subsequently split into its two original component stones and the still intact stone 2a/b was reused again. This reuse probably involved the evident cutting back of faces A, B and C. Stones 2a and b must have been broken apart subsequent to this second reuse.

A (broad): Undecorated, but with an incomplete bold vertical inscription in a combination of capital and uncial letters, which is probably indecipherable and whose language is uncertain. John Higgitt suggests the following tentative transcription (reading from bottom to top):

(stone 2c)

(stone 2a)

The surface of this face on stone 2b has been almost entirely cut away, although isolated incisions suggest that the inscription was originally continuous between its two surviving fragments. John Higgitt comments: 'The lettering is formed of thin, unmodelled strokes. There is no clear serifing, although it is possible that there were modest serifs at the top of the verticals of the N and perhaps at the foot of the P on the larger stone. The right foot of the second A seems to terminate in a decorative curl. The language and meaning of the inscription are uncertain.'

B (narrow): Undecorated except for the cable-moulded borders.

C (broad): Decorated with two circular motifs in low relief. The upper, on stone 2a, is a small eleven-lobed flower with rounded petals set around a central boss. The lower motif (stone 2c) comprises an interlaced cross pattée with 'ears'. At the centre of the cross is a similar flower, this time with ten rounded petals. Both motifs are cut with great skill and accuracy.

D (narrow): This face carries decoration only on stone 2c. It is uncertain whether or not this face on 2a and b was undecorated or whether decoration of a similar type was removed during a period of reuse. Face D on stone 2b preserves a deep square socket of unknown function. The decoration on stone 2c consists of a finely carved foliage trail with small, tightly-carved leaves grouped in fronds of two and three with spiralling tendrils filling in the gaps. The trail emerges out of a blank area at the base of the face.

Appendix G item (the continuing tradition).

The two circular motifs on face C allow this monument to be dated securely to the central part of the twelfth century, and so it belongs to the interesting regional group of Romanesque cross-shafts (Chapter IX). The lower, cross pattée motif (Ill. 452) is found on several other monuments in Lincolnshire (for example at Keddington, Ills. 458–9) all of which have been assigned twelfth-century dates. The small upper flower (Ill. 450), likewise, is to be associated with monuments such as the cross-shaft at Aunsby (Ills. 434, 436) and with a number of items of architectural decoration. Most significant amongst these are the presence of rather similar flowers in the soffit of the central door at the Cathedral and in the hood-moulding of St Mary's Guildhall, Lincoln which are dated to the 1150s (Stocker 1991, 33–41). The similarity between the examples at Castle Bytham and in Lincoln is quite strong, especially in the way that the flowers are dished into the surface of the stone.

The style of the acanthus trail on face D of Castle Bytham 2c (Ill. 455), however, is more difficult to account for, and has been called ninth-century (Okasha 1983, 112, citing the opinion of Dr S. Plunkett). It is quite different in appearance from the trails on the other late cross-shafts at Digby, Minting and Revesby where the leaves are large and thick (Ills. 460–1, 462, 470–1). At Castle Bytham, the leaves, although reminiscent of Romanesque acanthus, are small and often pointed, and the stalk and tendrils of the trail are integral to the complex design. Indeed the way in which the tendrils spiral back on themselves to fill the gaps left by the undulations of the stalk is very reminiscent of Anglo-Scandinavian examples such as those on Creeton 1 (Ill. 124) or, more particularly, of the pre-Viking foliage trails on the shaft at Edenham (no. 1, Ills. 164, 166), which, like that at Bytham, emerge from curving fields at the base of the shaft. It is possible that the sculptor of the Castle Bytham monument was looking to one or other of these local monuments (both less than five miles away) to provide guidance for how such a foliage trail should be cut.

The Castle Bytham shaft, with its sparingly placed decoration and its major inscription, is a relatively rare type of monument. There is, however, a monument from Gainford, co. Durham, which also appears to be a late cross-shaft with a cable-moulded border decorated only with a bold vertical inscription (Cramp 1984, 87, no. 21, pls. 69–70). Furthermore, by the fourteenth century, a cross had been erected at Edenham churchyard which was very similarly decorated with a long inscription in French (Stukeley 1726).

The inscription on Castle Bytham 2 is probably of the later twelfth or thirteenth century, based on the closed form of the E, of which there are no examples on English seals before the 1170s (Kingsford 1929), although, of course, letter cutting in stone need not have kept in exact step with that on seals. The combination of capital and uncial letters would also fit a twelfth- to thirteenth-century date. There is thus a possibility, impossible to confirm now, that the inscription was a later addition to a slightly earlier shaft.

John Higgitt thinks that the inscription as it now stands is indecipherable. He did, however, offer what he felt was a very speculative and in the end unconvincing interpretation of the surviving letters as a fragment of a Latin inscription. 'The letter third from the end of the group of letters on stone 2c could be a very narrow and rectangular C, although a very narrow L or an I are also just possible. The letters transcribed here as N might conceivably represent H and the letter transcribed as Q could just possibly have been intended as closed C with its serif protracted downwards. The G might have been intended as a C, although G is what it looks like. If we take these rather less likely readings and take the third letter from the end as L, again not the most obvious reading, we would have:

This might just be Latin: –E HAC HOC LAP[IDE]–. Lapis is used in medieval funerary inscriptions in France and no doubt in England. This is, however, based on rather forced readings of individual letters. The inscription on the smaller stone is even less promising.'

If this speculative reading can be accepted, however, then it might suggest that the story, apparently first suggested by Wild (1871, 134), that this was a boundary marker, was indeed true. Inscriptions on crosses which have formulae declaring that 'this stone marks [the boundary]' are indeed known from the twelfth century, with perhaps the most famous example also being in Lincolnshire, on the St Guthlac Stone less than twenty miles away (see Crowland 2), although the twelfth-century Fletton cross, near Peterborough, may be another example. If Castle Bytham 2 is a boundary marker, however, this would support the local tradition (Cotton-Smith 1900, 5) that this is indeed the famous Robin Hood's Cross recorded by Leland (Smith 1907, 23), which marked the boundary of Kesteven on the open heath near the hamlet of Aunby a few miles to the south, since that cross was last recorded in this location as recently as the first quarter of the eighteenth century (Owen, D. 1971, 2). If this is Robin Hood's Cross, it must have been taken down in the late eighteenth or early nineteenth century to be recut for use in a building prior to the building in which it was discovered by Wild in 1871. Although Gough referred to that cross in 1789 and 1806, he was merely repeating an observation made by earlier antiquaries (1789, 249; 1806, 357).

Wild 1871, 5, 134–5; Allen and Browne 1885, 356; Fowler 1896, 3; (—) 1899–1900, xlviii–ix; Cotton-Smith 1900, 5; Davies 1912–13, 130, 133; Davies 1915, 53; Cox 1924, 97; Clapham 1926, 5; Davies 1926, 9, pls. II, III; Butler 1964, 116; Pevsner and Harris 1964, 493; Taylor and Taylor 1965, 716; Okasha 1971, 149; Okasha 1983, 112; Pevsner et al. 1989, 215; Stocker with Everson 1990, 88