On

The tiles that formed the top floor seem to have been made of a different clay and some vitreous substance, and are much harder than the others. some of them there is the same circular pattern already noticed, only here it sometimes takes the form of three circles. One fragment had a raised moulding round the edge. The whole of the ground round the place excavated is covered with bricks and fragments of the tiles that were dug up, and, although I made a careful search, I cannot pretend to have made an exhaustive examination; but I think I have mentioned all the prevailing marks. Unfortunately, the place was left without any protection or fencing, and the result is that, what with cows, visitors, and boys, by October the excavated portion was nearly destroyed.

I went again carefully over the room in October, but found nothing more to notice; but about halfway along the west wall I dug up a large quantity of soot and a few fragments of bone. In the south-west corner I began a small excavation, to see if the south wall was continuous; it appears to go on in a westerly direction. I found fragments of broken bricks and tiles arranged in the same order as those above described; a large piece of concrete, two small fragments of whitish pottery, some iron T-nails, a piece of glass, and some fragments of bone. The wall appears to be continuous; but I had not time to carry my excavation very far.

The day before 1 left, as a man was ploughing in a field to the right rather deeper than usual, he struck the stones of the Roman road. I say this because the stones were obviously paving-stones, and placed as part of a pavement about 15 inches below the surface, and, on their being removed, no trace of building was to be found underneath. The man also came upon a

Wall. Mr. Wright adds, the tiles are always scored in patterns of great variety, apparently for the purpose of being fixed more tenaciously by the mortar.

fragment of a wall built with very large stones. I had it excavated some depth down, but only found pieces of charcoal, bone, and fragments of oak board, very thin, and a nail or two; there was no brick or pottery, and I was unable to trace the wall in any direction. This building would be a few yards from where the Roman road passed on its way to the Teifi.

The specimens of bricks, etc., which I produce are fairly illustrative of the bricks and tiles found. There are some bricks very much larger, 20 by 17 inches; but the majority of the fragments are such as I have brought.

I shall hope to continue the excavations in a more systematic manner another year.

Before concluding this paper I must say a word as to the inscribed stones. At present there are three, all figured by Meyrick, and also by Westwood, Lap. Wall., part iv, pl. 71, fig. 3; pl. 78, figs. 1 and 2.

The first is the Ennius stone; it is 11 inches high, and 6 inches wide; it consists of the following threeline inscription, with the ordinary border:—

> ARTISM)

ENNIVS

PRIMVS.

It is figured in Gibson's Camden by Lhwyd, who says that he reads it " Caij Artij Manibus (aut fortè memoria) Ennius Primus". Meyrick (1810) also figures it at pl. v, fig. 7, and speaks of its being built in the wall by the side of the door of a cottage. It was removed thence, and disappeared for some years, but, at the meeting of the Cambrian Archæological Association at Lampeter, in 1878, Mr. J. N. Davies, of Aberystwith, sent it to the local museum then formed at Lampeter, and after the meeting it was deposited in the library of St. David's College, where it still is. This stone is also figured by Hübner, Inscriptiones Britannia Latina, Berlin, 1873, p. 44, as "No. 148 intra tabellam ansatam". The inscription is given incorrectly as

As Hübner's book was published at the time the stone had disappeared, his. account is taken from Meyrick, whose plates and accounts of inscriptions are most inaccurate. Hübner says of No. 148: "Latet fortasse (centuria) martialis? Ennius Primus (fecit)." Westwood, Lapidarium Walliæ, p. 142, describes this stone, which he figures, pl. lxxviii, fig. 2.

The second stone is also figured by Lhwyd in Gibson's Camden, and by Meyrick, pl. v, fig. 8, who says: "Another stone, on a chimney of another cottage, is to be read oVERIONI." This stone is now built into the wall of the farmhouse, near the back door; it is about 14 inches long, and 4 inches wide. Lhwyd and Meyrick both give the inscription incorrectly as OVERIONI, as will be seen from the rubbing I produce, which I made in October; it is—

IO VERIONI.

An account of this stone, with an engraving, with the incorrect inscription, is published by Professor Westwood in Archeologia Cambrensis (4th Ser., vol. ii, p. 263), the figures being taken from rubbings supplied to him by the Rev. H. L. Jones, who made them on July 16, 1861. Mr. Jones, in a paper in the Archaologia Cambrensis (3rd Ser., vol. vii, 1861, p. 312), says the stone was on the east wall of the house, above the horse-block, having the rudely-executed name of OVERIONI.

In the Archæologia Cambrensis, Professor Westwood says: "Amongst the many Roman inscriptions found at Llanio i Sav, close to Llandewi Brefi, Cardiganshire, is one of which an engraving is here presented, representing the name OVERIONI, inscribed within an oblong space, defined by incised lines, about 13 inches long by 3 inches high. The letters are thin, tall, and ill-formed."

The stone is also figured by Hübner as No. 149. He gives OVERIONI, giving Lhwyd and Meyrick as his authorities; he adds No. 149, "est (centuria) Verioni (?)"

Westwood, in his Lapidarium Wallia, describes the stone, and figures it pl. lxxi, fig. 3 (the figure is not quite correct, the R and I being conjoined, as well as the v and E), and gives an account of it at p. 142. He says the stone "is now built into the east wall of one of the farm-buildings, about 15 feet from the ground above the horse-block." To obviate any mistake in the future, it may be pointed out that it is into the wall of the house, near the back door, not that of the farm-buildings, that the stone is built, and it has been there for years. After remarking that in his paper in the Arch. Camb. the inscription is given as OVERIONI, he says: "The stone is, however, injured at the left end, and, on examining it carefully during the Lampeter Meeting in August 1878, we adopted the conclusion suggested by Mr. Robinson (one of the Secretaries of the Cambrian Archæological Association), that the first supposed letter was incomplete, and that its supposed right side indicated a centurial mark, leaving the real name VERIONI."

As above stated, the interpretation of Mr. Robinson was really that suggested by Hübner, without seeing the stone. From the rubbing it will be seen that the so-called o does not exist, that the first letter has been injured, and that the stone appears to be merely a fragment; that what has been taken for the end of the border seems to be part of a letter, and it is doubtful whether the semicircle is the centurial mark or the fragment of some letter, such as D. It is not a matter of much importance; but none of the drawings of this stone are correct, as they do not give both the VE and the RI as conjoined. Until the plate in the Lapidarium Wallia all the letters were given sepaThe plate there gives the VE conjoined, but not

rate.

the RI.

The plate in the Lapidarium Walliæ repre

sents the stone as far too perfect, especially at the left side. It has every appearance of having been broken off at the end, and not being complete, as shown in the plate.

The next stone, which Professor Westwood calls the legionary stone, is the most interesting. It was, I believe, first mentioned by Sir R. C. Hoare, who, in his introduction to Giraldus Cambrensis, vol. i, p. clii, says: "I had the good fortune to decipher another (inscription), far more interesting than the two former (he is alluding to the two stones already described), which stands before the threshold of the farm-house. If I read it rightly, it appears to record some work done at this place by a cohort of the second legion, COH. II. A. - - G. FVP, Cohors secunda (legionis) Augustæ fecit quinque passus." This interpretation of Sir R. C. Hoare has been adopted by all or nearly all subsequent writers until Mr. Thompson Watkin. Meyrick, who figures the stone in pl. v, fig. 9, thus describes it: "In the porch of the house is a very large one, now serving for a seat, and much obliterated, has on itCohors secundæ Augusta (sic) fecit quinque passus

which shows that a cohort of the second legion of Augustus was stationed here, and built a part of the walls of the city." This statement of Meyrick's has been quoted over and over again, but unfortunately it is difficult, if not impossible, to make out Meyrick's inscription from the stone itself, and even his plate is difficult to understand.

The Rev. J. L. Jones, in his visit in 1861, thus speaks of the stone: "The other (is) in the lower part of the stable wall, thither removed from the horseblock, not many years back, with traces of two lines of words on it, but of which con is almost the only portion now legible." If in 1809 the stone was in the porch, and then in the interval to 1861 removed first to the horse-block, and then to the stable, it is not

1 Arch. Camb., 3rd Ser., vol. vii, p. 312.