Roman work. Where the valley of the Dulas narrows, about two miles from Lampeter, are two forts, one on the right bank of the valley called Gaer, close to where the Derry Ormond column stands the one on the opposite bank, called Castell Goytre, a large and fairly perfect fort; while guarding the Teifi valley are two forts, one on each bank, that on the right bank known as Castell Allt Goch, and that on the left as Caernau. All of these are marked on the Ordnance Map. On the south, above Llanfairclydogau, just where the Sarn Helen turns off over the mountain, at a place not marked on the Ordnance Map called Panteg,1 is a small square fort or camp, in good preservation, about 36 yards long by 28 yards wide; the banks have been partly cultivated away, but enough still remains to show very plainly its extent, and the four entrances opposite each other are evident. It will thus be seen that on each side the approach to Llanio was carefully guarded; so it may fairly be inferred it was a station of some importance. It is difficult to trace the Sarn Helen from Llanfairclydogau to Llanio; local tradition says the road crossed the Teifi by a bridge near a farm called Godregarth, and that when the river is very low the foundations of the bridge can still be seen. I have, however, looked in vain for them. In a dry summer the line of the road is said to be very plain between Llanio and the river. This summer (1887) the site of the road could be clearly traced from the grass burning up across a pasture field on it sooner than in other places. This field adjoined the railway, and the burnt part of the field went in a straight line towards the river for the reputed site of the bridge. In a field between the two points, but also in this line, traces of the road, i.e., paving-stones, were found in October 1887, when ploughing. To be able to fix the line of the road is important, as showing the route the Sarn Helen took between the two portions that now remain, and also as showing that

1 Arch. Camb., 4th Ser., vol. ix, p. 326; vol. x, p. 56.

the station was a far larger one than has been usually supposed; for Caer Castell, where the inscribed stones were said to have been found, and the site of the buildings where the excavations have taken place, are at least some two or three hundred yards away from the road, and from some buildings found this autumn and from the road it is nearly a quarter of a mile to the other side of the station. Caer Castell, which is always pointed out as the site of the station, was probably that of the camp. On one side of it are some faint traces of embankment, and in it stones have been constantly found. It is an arable field of some five acres, higher than the rest of the surrounding ground. I am told that this year the corn withered up in two broad lines across the field, the lines crossing at right angles, a statement which, if true, would go to show the existence of two paved streets crossing each other at right angles. Adjoining Caer Castell on the flat towards the river the foundations of buildings are clearly to be seen. Here it was that the excavations of this year (1887) were made. The fact of there being a Roman station at this spot is, I believe, first noticed in Lhwyd's additions to Gibson's edition of Camden's Britannia (1695). On col. 645 he figures two of the inscribed stones that have been found here, and states:

"A Country-man told me there was another [inscription] at a house called Lhanio-fav, in this parish, distant about a mile from the Church. Being come thither, I found these two Inscriptions, and was inform'd that several others had been discover'd by digging, but that the stones were applied to some uses, and the Inscriptions not regarded." He adds: "Besides Roman Inscriptions, they find here sometimes their coyns, and frequently dig up brick and large free-stone neatly wrought. The place where these Antiquities are found, is called Kae'r Keftilh, which signifies Castle-field, or to speak more distinctly, the Field of Castles; tho' at present there remains

not above ground the least sign of any building: nor were there any (for what I could learn) within the memory of any person now living in the neighbourhood, or of their Fathers or Grandfathers. However, seeing it is thus call'd, and that it affords also such manifest marks of its being once inhabited by the Romans, we have little or no reason to doubt, but that they had a Fort or Garison, if not a considerable Town at this place. And that being granted, it will also appear highly probable, that what we now call Lhanio, was the very same with that which Ptolemy places in the Country of the Dimetae, by the name of Lovantinum, or (as Mr. Camden reads it) Lovantium.

If any shall urge, that to suppose it only a Castle, and not a City or Town of note, is to grant it not to have been the old Lovantium; I answer, that perhaps we do but commit a vulgar Error, when we take all the Stations in the Itinerary, and Burroughs of Ptolemy, for considerable Towns or Cities; it being not improbable, but that many of them might have been only Forts or Castles with the addition of a few Houses, as occasion requir'd."

Meyrick, in his History of Cardiganshire (1810), p. 272, gives the following account of the place: "Llanioissa was formerly the ancient Loventium of the Romans, and a considerable station on the great western road called Sarn Ellen, between Maridunum, or Caermarthen, and Penallt, near Machynlleth. Several coins

and culinary utensils have been dug up here, and three Roman inscribed stones are built up in the walls of two cottages on this spot. . . . Almost the whole of this place is covered with the fragments of the finest brick, which the Romans must have brought with them. There are also some small remains of pieces of brickwork and lime mixed with common stone still to be seen; and one entire piece, having its surface smooth and polished, was taken up not long ago, and placed at the bottom of an oven then making in a neighbouring mill, where it still remains. In one of

the grounds of this farm a large piece of unshapen lead was dug up, which, when melted, weighed sixteen pounds. There is a piece of ground to the southeast of the farmhouse called 'Cae'r Castell', or the 'field of the Castle', in which are still the remains of the foundations of buildings."

1

All subsequent writers have practically adopted this inaccurate description of Meyrick's in their account of Llanio; it is the one that is found in the South Wales guide-books of the present day. It is obviously the basis of the following description by the Rev. H. L. Jones, written in July 1861, and which appears in the Archæologia Cambrensis for that year.' He says: Any casual observer might visit this spot without perceiving that he was on the site of a Roman town at least as large as Lampeter of the present day. Some faint traces of embankment may be observed on and about Cae'r Castell; but it is on the flat towards the river that you must look for foundations of houses. Here the tenant of the farm, a person of intelligence and courtesy, pointed out to us the sites of several buildings. Here they dug up for us stones and mortar of walls, still in their courses, under ground; here they showed us how the soil of the surrounding fields was filled with bricks, and where lumps or weights of lead had been discovered."

In 1878, at their Lampeter Meeting, the Cambrian Archæological Society visited Llanio, and give this description of it: "At Llanio traces could be seen of portions of the Roman camp, Loventium, and in all directions pieces of Roman brick and mortar; but much excavation will have to be done before any satisfactory account can be given of it."

Both these accounts are incorrect in describing pieces of Roman bricks and mortar as being found in all directions. They are only found, as far as I can make out, in one place, the flat towards the river, where the

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

2 Ibid., 4th Ser., vol. ix, p. 353.

recent excavations have been made. In the other fields stones are often found, but no bricks.

Before describing the excavations it will be as well to mention some of the things that have been found at Llanio from time to time. As far as I can ascertain, very little record remains of what has hitherto been found, and the things themselves are all dispersed or lost. I leave the inscribed stones to a later part of the paper.

The most interesting and most curious find is a wooden female head (which, by the courtesy of the owner, Mr. S. Jones of Llanio Fawr, I am able to exhibit here to-night), found some years ago, when digging peat in a field called Caer Gwyrfil, which adjoins Caer Castell. The head is fully described and figured in a paper in Archæologia Cambrensis. It is said to be of birch, and, notwithstanding it is in a most wonderful state of preservation, it is suggested it is of Roman origin. "The careful and artistic braiding of the hair, from the forehead to the back of the head, with the cavities in the place of eyes, suggested that the head was not of modern workmanship, and led to the inference, when the place of its find was taken into account, that it may be Roman. A socket-hole extends from the collar upwards into the neck, which apparently served to fix the head on the body of the figure or statuette to which it belonged; but there are no rivet-holes or signs of any other mode of attachment. On examination the right side of the head appears to be smooth and perfect, while the surface of the left side is slightly abraded. This may be accounted for by the supposition that the left side was that exposed to the atmosphere on its deposit. Mr. Jones said that there were hands with part of

1 4th Ser., vol. x, p. 81.

2 In a discussion that took place upon this paper, the President, Dr. Evans, suggested the head was of yew, alluding to the fact that yew in a fairly preserved state has been found in the Swiss lake. dwellings.