A fragment with horizontally disposed ornament, with the remains of four decorative elements, separated, and probably framed on all sides, by narrow roll mouldings. There appears to be a flat moulding along the lower edge of the stone. (i) On the left is a fragment of a panel in which only the right-hand edge of a circular element survives. (ii) The second panel, which is incomplete at the top, specially the top left, has a decorative element composed from four compass-drawn half circles, their bases forming the sides of the square, drawn so that they overlap and enclose the centre of the square. The effect is of flower with open petals, or a closedcircuit simple pattern E interlace (otherwise unknown). (iii) The third panel is the most complete. It consists of a 'marigold' pattern with four solid petals arranged in the form of a cross, and four hollow petals filling the spandrels, all with pointed ends arranged round a solid centre, and enclosed in a circle. Four curls protrude from the circle to fill each corner of the square. (iv) The fragmentary panel on the right is again composed of fine strands in an open pattern, possibly a repeat of the pattern in the second panel.

The height is not complete but the form of panel iii suggests it was not much more than the present maximum of 15 cm. The marigold pattern has an association with early sites, such as Hexham, Northumberland (Cramp 1984, pls. 172.910, 178.944, 182.972; see Ill. 859). The encircled leaf and pellet design at Ledsham (no. 5, Ills. 471–5) and a related pattern at Ripon (no. 8, Ill. 667), also compass-drawn patterns, seem to belong to the same early phase. Examples of similar roundels with compass-drawn patterns, some including formal flowers with both solid and open leaves, can be found on sixth- to seventh-century Visigothic architectural pieces and furnishings, including pillars, string-courses and items of church furniture (Palol and Hirmer 1967, pls. 6, 13, 17). This fragment would seem to belong to an early church, although the earliest date suggested for parts of the present building is Norman (Pevsner 1967, 173–4). No church is mentioned in Domesday Book, although the absence of an early church cannot be argued from this silence. Ryder (1982, 109) has suggested this may be part of either a grave slab, or, his preferred solution, a closure slab from a screen around an altar or sanctuary. It is perhaps more likely to be part of a string-course and at 15 cm in height this does not seem unreasonable.