An octagonal shaft with bold angle-roll mouldings ending in splayed hollow-cut bases. The shaft base is plain and octagonal at the top, but roughly shaped and dressed below, where squared-off wedge-shaped holes have been cut beneath faces A and E.

A: On the lower fragment is the terminal of an unpinned loop enclosed in a twist (see G.I., fig. 24, Dii), below a plant element composed of a pair of inturned split leaves.

B: On the lower fragment are two pairs of split leaves linked by long stems and disposed in a cross, and traces of twist or plait.

C: Missing.

D: On the lower fragment, springing from a horizontal binding or root, a double-stranded plant trail with a loose pellet filling in the curve. On the upper fragment, two plant stems and a single pointed and veined leaf.

E: On the lower fragment, traces of strands. On the upper, what may be a closed circuit simple or half pattern F with a long glide.

F: On the upper fragment, a closed circuit pattern, possibly as G. I., fig. 26, Bii.

G: Missing.

H: On the lower fragment, a plant motif which springs from a single stem whose roots span the space between the bases of the two columns. The stem divides into two strands from which fall two plain half-moon leaves. Above, the stems appear to cross, enclosing a small anonymous leaf, and at the left of the carving is a veined half leaf.

This is an important carving, since its archaeological context (deeply bedded in a pit which had then been covered by the opus signinum floor of building A) renders it unlikely to be later than about A.D. 700. The object had been buried below the floor at a level just below the carved face, and the sockets in its uncarved base, if they be not a survival of an original Roman piece, imply that it had been lifted into place by some sort of mechanical tackle. Although there are a few other carved fragments which could fill in some interlace, there is not enough to complete another pattern. There are also, however, many uncarved fragments from the same deposit. I originally interpreted these fragments as part of a carved column central in the building (Cramp 1969, 45). However, although it is impossible to calculate the original layout, there does not seem enough surviving for a column, nor is a span on either side of such a column of about 8.5 feet likely for an Anglo-Saxon opening. However, the central position at one end of a building would be appropriate also for a stone reading-desk such as are a feature of monastic refectories in Egypt. Comparable pieces which could likewise be part of such reading-desks survive also at Kirby Moorside, Yorkshire (Collingwood 1907, 342, d); Kirkby Stephen, Cumbria (Collingwood 1927, fig. 15), and there is possibly a more massive and later version from Melsonby, Yorkshire (Cramp and Lang 1977, nos. 7-8 and fig.). The architectural forms are difficult to parallel, but their form is identical with the legs of the throne on which the man symbol is seated in the Echternach Gospels (Bibl. Nat. ms. lat. 9389, fol. 18v; Hubert, Porcher and Volbach 1969, pl. 174). The half veined leaf type, though more boldly cut, is the same as on Hexham 1 (Acca's cross). The closest parallel for both the exploded scroll and the trail is however a cross at Lancaster (Collingwood 1903, 259, fig. 3). The interlace patterns can also be compared with Hexham (nos. 36-38). The archaeological context of this piece supports a date contemporary with these imposts at Hexham, which should be part of the work of Wilfrid or Acca.