(Fig. 2) repeated in a double vertical row; (west face) the Figureof-Eight knot (Fig. 3) repeated twice in a vertical row, and terminating in a Stafford knot at top and bottom, double-beaded.

On the Cap of the Pedestal.- Four horizontal bands running right round all the faces: (1) a twist; (2) a three-cord plait; (3) a twist; and (4) a twist. Below this there is a cable moulding.

On the Body of the Pedestal.-At the four angles are columns treated architecturally with capitals and bases, the whole being ornamented with plaitwork. On the north face the plaitwork on the column on the right side has a border of three loops at the bottom. The north and south faces are narrower than the east and west faces, so that the space between the columns at the angles in the former case is only about 4 ins. or 5 ins. wide, as against 10 ins. or 11 ins. in the latter case. The space between the columns on the north and south faces is therefore left plain for want of room to display the ornament. The space between the columns on the east face has a double row of Stafford knots (Fig. 2) upon it, and on the west face a four-cord plait, with two horizontal breaks bearing a Figure-of-Eight knot in the middle.

On the Base of the Pedestal (north face).-The bust of a man in relief on a background of plaitwork; (south face) a similar bust on a background of twisted bands; (east face) four three-quarter length figures-the one on the right wears a crown with threepoints, and the three figures in the middle are carrying crosses; and (west face) a man on horseback on a background of plaitwork. Between the legs of the horse is the not uncommon device of a pair of oval rings, crossed and interlaced. There also appears to have been some ornament on the roll-mouldings round the four panels of the base, but it is now so much weathered that the patterns cannot be properly made out.

It appears from the foregoing description that the ornament on the Llandough cross is of three kinds, namely, (1) interlaced work; (2) key-patterns; and (3) figure subjects.

The knots used in the interlaced work are shown below (figs. 4 to 7).

x X W X

Fig. 4.

Fig. 5.

Fig. 6.

Fig. 7.

and also the device formed of two oval rings crossed and interlaced thus:

Only one kind of key pattern used, as shown on Fig. 9.

AVA

Fig. 9.

The other localities where these patterns occur are given in my Paper on "Early Christian Art in Wales" in the Archeologia Cambrensis (5th Ser., vol. xvi, p. 51.)

The figure subjects are confined to the four faces of the base of the pedestal of the cross. There is nothing in the figures themselves or their grouping to give any

clue to their meaning. The man on horseback on the west face is the most remarkable, and the only thing of a similar kind which occurs in Wales is on the base of the great wheel-cross of Conbelin at Margam,' Glamorganshire. Both on the cross at Llandough and at Margam, the device, composed of two oval rings crossed and interlaced, is introduced in the background, but whether as a symbol or as mere ornament to fill a vacant space, it is impossible to say. At Margam, this device is repeated twice, and associated with the triquetra, or three-cornered knot.

Dr. H. Colley March, F.S.A., has brought together much curious and interesting information relating to the use of the triquetra and the crossed oval rings (or "duplex," as he terms the device thus formed), in a Paper on "Two Examples of Symbolism," in the Proceedings of the Dorset Natural History and Antiquarian Field Club (vol. xxv, p. 17). In his opinionif I understand him rightly-the triquetra and duplex were, in the first instance, Pagan sun symbols, associated with the worship of Odin and Frey, and became in Christian times symbols of the Trinity and of Christ. If this view be correct, the horseman on the crosses at Llandough and Margam should be intended for Christ. I am quite prepared to admit that the triquetra and duplex were probably Pagan symbols connected with sun worship. However, the difficulty is to determine whether the Christian artists who adopted these Pagan devices were aware of their primary significance, and gave them a new symbolical meaning; or whether, after being copied over and over again, the devices degenerated into meaningless pieces of decoration. Who can say, indeed, where symbolism ends and ornament begins?

At the top of the middle part of the pedestal of the Llandough cross, just below the projecting cap, is a

1 Arch. Camb., 5th Ser., vol. xvi, p. 17.

minuscule inscription in one horizontal line, which

reads

irbici

"or (The Cross) of Irbic."

The Plates illustrating this article are from photographs of the cast in the Cardiff Museum, taken specially by Mr. Alfred Freke, of 12, Duke Street, Cardiff. We have to thank the Museum Committee for permission to have the photographs taken, and Mr. John Ward, F.S.A., the Curator, for superintending the operation.

The Llandough cross has been previously described and illustrated by Prof. J. O. Westwood, in his Lapidarium Wallia, but not in the Archæologia Cambrensis. This cross has been taken as a model for the memorial at Haverfordwest to the Pembrokeshire men killed in the South African war. Mr. Arthur G. Langdon, F.S.A., is the architect.

Archaeologia Cambrensis.

SIXTH SERIES.-VOL. IV, PART IV.

OCTOBER, 1904.

GLIMPSES OF ELIZABETHAN PEMBROKESHIRE.

MORE ODDS AND ENDS ABOUT HAVERFORDWEST AND OTHER PLACES.

BY THE REV. JAMES PHILLIPS.

THIS number of the Archeologia Cambrensis should have contained the third and last Paper on the "Oldest Parish Registers of Pembrokeshire." Unfortunately, the Editor's request for "copy" found me at a little distance from Haverfordwest. Now, a careful re-examination of the old manuscripts is necessary, and possibly a cautious application of "restorative fluid" to some of the faded pages, before I venture to commit my notes to print. If therefore, I am not to disappoint our courteous Editor, I must postpone the "Parish Registers" to the January number, and send him a few extracts from my notebooks, in the hope that they will not be found uninteresting.

The Archæologia Cambrensis for 1899 contained a Paper on the "Oldest Municipal Balance Sheet. This document, consisting of two four-page sheets, is the oldest business paper (as distinguished from conveyances, leases, etc.) that I have been able to find in the lumber-room which has proved so rich in archæological treasures. There was, indeed, in one of the drawers,

6TH SER., VOL. IV.

18