tracery. It is surmounted by a Tudor cresting, and has over the centre a Calvary Cross. It was designed by Mr. Douglas of Chester.

When we turn from screens and rood-lofts to the cognate subject of churchyard and wayside crosses, it is remarkable that there is not, as far as I know, a single instance of the survival of either the one or the other in the county; and this notwithstanding the far more durable material of which they were made; nor can I recall to mind more than one place-name that seems to hand them down: that of "Gungrog" (Cefn Grog) near Welshpool. The adjoining Abbey of Strata Marcella, with its township of Tirymynech (Monksland), would readily explain the name, were it not that it appears to be of much earlier date than the Abbey. It is not, I believe, because they never existed; the bases of some of them may still be doing duty for sundials yet we have no record of their demolition. Their nonexistence now, however, is the more noticeable by way of contrast to the adjoining counties. Thus in Merionethshire we have the extremely early Cadfan Stone at Towyn, and all but the head of the cross at Corwen. In Flintshire we find the Celtic crosses of Maen Achwyfan and Dyserth, with the medieval crosses of Hanmer and Newmarket, and in Denbighshire Eliseg's Pillar (the head of the cross is lost), of the ninth century, and the fourteenth-century cross in the Churchyard of Derwen. This last is the more significant, because it controverts the plea that where there was a rood within the church a cross outside would be superfluous, and that vice versa, a churchyard cross would render an inside rood unnecessary; for here at Derwen both evidently co-existed. The four faces of the cross bear sculptured representations of the Holy Trinity, the Judgment, the Virgin and Child, and the Crucifixion with the Virgin and St. John at the foot; but this last is the rood. In the church, and in

excellent preservation, is a fine roodloft, with sixteen panels ornamented with tracery, and having a band of the vine pattern as a cornice. In the top of the western beam and at its central point is a socket, or mortise, to receive the foot of the rood, which would face the congregation. This, indeed, is no longer there, but its witness remains. It is worthy of mention that this parish adjoins that of Bettws; and their two churches, which are only about five miles apart, are both of them noteworthy for their rare ecclesiastical remains.1

Mem.-This article was written in the first instance for, and read in part before, the Newtown Clerical Association; and has subsequently been enlarged and illustrated for its present use.

1 Supra, p. 94.

THE HERMITAGE OF THEODORIC, AND THE SITE OF PENDAR.

BY THOMAS GRAY, ESQ., M. INST. C.E.

I.

BEFORE January, 1894, I was unaware of a hermitage having existed in these parts. At this date, Miss Talbot kindly sent me volume i of the Margam and Penrice MSS., by Dr. W. de Gray Birch; and in it I found that a mile or so from where I live in the parish of Margam, there existed as far back, and probably before the year A.D. 1147, the Hermitage of Theodoricus; but where was it situated? No ruins existed to mark its site, no tradition survived about it, and the building had disappeared completely.

In the earliest charters of Margam Abbey we find mentioned as a landmark the Hermitage of Theodoric ; but as no ruins indicated its position, it was not possible to fix its site. The original charter founding the Abbey of Margam is not extant; but its text is found in an Inspeximus by Edward le Despenser, Lord of Glamorgan and Morgan, dated July 13th, 1358, of an Inspeximus by Hugh le Despenser, dated Oct. 9th, 1338. In this document the Earl William notifies to the Bishop Nicholas' and others concerned, that he has confirmed the gift which Robert2 his father gave to the monks of Clairvaux: "That is to say all the lands which extend between Kenfig and the further bank

1 A.D. 1149-1183.

2 Robert of Caen, natural son of Henry I, King of England, Consul or Earl of Gloucester. He became possessed of these lands by his marriage with Mabilia, the heiress of Robert Fitzhamon, the leader of the Norman knights, who retained Kenfig and district in addition to Cardiff as his share of the conquest.

of the water of the further Afan, which is to the west of the Hermitage of Theodoricus as the water aforesaid descends from the mountains. All this land I grant to the monks as it goes through the mountains,

Fig. 1.-Map of Lands given to the Monks of Clairvaux, and Grants of Land by Caradoc Uerbeis to Pendar, etc.

namely, from the source of Kenefeg water between the source of Rudelf (Ffrwdwyllt) and Gelli-fret (Gellivrith) on to Red-Kewelthi (Rhyd Gyfylchi), that is the ford of Kewelthi, into Aven (Afan river) in wood and in

ท

plain, in fields and in pastures and waters, in moors and marshes, also all the fisheries of Aven, that no one may interfere with them on the other side, nor put their hand to fishing in the whole of Aven except by their consent."

This certainly points to the site of the Hermitage as being near the river Afan, where it falls into the sea, and just to the east of it.

In the midst of the lonely sand-dunes near the old mouth of the River Afan (in 1836-38 it was diverted, and is now further west), some fifteen or sixteen years

ago, I picked up a tile-stone having a neatly-made nailhole at the top part; and later I discovered part of a wall; still later I found some green glazed earthenware tiles, ridge and flat,' and several pennant-stone tiles, similar to the first one I found. Three years ago I had the sand cleared off around a pile of stones, and found building about 85 ft. in length, which is here shown in elevation and plan: water then prevented further clearing of the ruins. I have recently I have recently (in this year)

1 See p. 149, No. 66, Arch. Camb., April, 1900, illustrations of similar old ridge-tiles found in Llantwit Major Church.