tedious, and could not be accomplished with any profit in the space at my command.

The picture of the social and domestic life of the Welsh in the days of their independence afforded by their law-books, can to some extent be filled in by means of the information handed down to us in the works of a celebrated Welshman of the twelfth century. Gerald de Barri (usually called Giraldus Cambrensis) was born in 1147 in the castle of Manorbier, which still stands on the rocks of the South Pembrokeshire coast. He came of a Welsh family which had a Norman strain, and his grandmother was the Nest-the "Helen of Wales "-who had been the mistress of Henry I, and afterwards wife of William de Londres, lord of Pembroke. His father, William de Barri, and other members of his family, had joined in warfare in Ireland. We must not linger over the details of his life or of his persistent struggle to secure for St. David's archiepiscopal status, or in other words the independence of the Welsh Church. In that effort he failed, but he has left for us valuable books, of which the most relevant for our present purpose are the Itinerarium Cambria and the Descriptio Cambria.'

In 1188 Baldwin, archbishop of Canterbury, journeyed through Wales to preach a crusade. He was accompanied by Giraldus, who recorded their experiences in the Itinerary. The second work is, as its name implies, a description of

The works of Giraldus are to be found in the Rolls series, vols. i, ii, iii, iv (ed. by Professor Brewer), vols. v, vi, vii (ed. by the Rev. J. F. Dymock). The Topography and History of the Conquest of Ireland (translated by Thomas Forester), and the Itinerary through Wales, and the Description of Wales (translated by Sir R. Colt Hoare, Bart.) are published in vol. vii, Bohn's Antiquarian Library (ed. by Thomas Wright, F.S.A.). For his life, see Dict. Nat. Biog., sub nom.; the introduction to vol. i in the Rolls series; and Gerald the Welshman, by Henry Owen, B.C.L., F.S.A.

the country and the people. Notwithstanding some attempt at fine writing which may have led to undue emphasis on particular points, we have no doubt that in these books we have a true record of the characteristics of the mediæval Cymry from the pen of an able and honest observer. These and the laws being our principal anthorities, we find that the condition of society in Wales was removed by very many degrees from a barbaric or nomadic stage, but it was backward as compared with the south-eastern Britain of that time. It may be that the economic progress of the scanty population of Wales had been checked by the war with Harold, the collapse of Gruffudd ab Llewelyn's power, and the subsequent course of events. Gerald deals with a people who had sustained many reverses, and who had been driven from the most fertile portions of their country by bands of Norman adventurers; and it is obviously likely that these things told for a time against any great social advance, though it may be noted as a curious fact that it was in the eleventh century that modern Welsh poetry has its beginning, and that in that region of culture contact, whether friendly or inimical, with the Norman lords, it had a stimulating effect. Neither Howel Dda nor Gruffydd ab Llewelyn, the only two chieftains of the Cymry who, after Rhodri Mawr, had played any really considerable part in the affairs of the island, were celebrated by contemporary bards whose works have come down to our time; but from the end of the eleventh century we find many poems devoted to the praise (often in extravagant language) of princes, some of whom were hardly of a position higher than that of a petty lordmarcher.

In the centuries with which we are dealing Wales presented a physical aspect very different from that which it does to-day. The greater part was waste land on which

the foot of man rarely trod, mere boulder-strewn moorland, or boggy tracts; and large portions of the estates now divided into farm holdings and highly cultivated were covered with trees that have disappeared. The roads (if we exclude the few which seem to derive their origin from the time of Roman occupation) were mere mountain tracks. There were practically no enclosures apart from the mounds or wooden fences which were made around the houses of the more important families.'

When Giraldus wrote, towns were beginning to arise under the shelter of some of the Norman castles, but there were no truly Cymric towns. Caerleon on Usk was in ruins, and Chester was in Norman hands. The social and domestic life of the Welsh centred round the timber-built houses of the kings, princes, lords or uchelwyr, which were scattered in the valleys and the lower slopes of the hills.

1 Rice Merrick, in his Booke of Glamorganshire Antiquities (1578), referring to the Vale of Glamorgan, says it was "a champyon and open country without great store of inclosures," and that the old men reported that "their ffore-fathers told them that great part of th' enclosures was made in their daies." (Cambrian Register (1796), pp. 96-8; Report of the Welsh Land Commission (Lond. 1896), p. 663.) 2 Giraldus says, "this city (Caerleon) was of undoubted antiquity and handsomely built of masonry, with courses of bricks, by the Romans. Many vestiges of its former splendour may still be seen; immense palaces formerly ornamented with gilded roofs in imitation of Roman magnificence, inasmuch as they were first raised by Roman princes, and embellished with splendid buildings; a tower of prodigious size, remarkable hot baths, relics of temples, and theatres all inclosed within fine walls, part of which remain standing," etc. (Desc., i, c. 5). The castle of Cardiff was surrounded by high walls, and Giraldus refers to the city as containing many soldiers. The Brut, in one of its versions, says, under the year 1080, "the building of Cardiff began." This is not in the Brut reproduced in the Oxford series. It occurs in the MS. called D, by Ab Ithel (see preface to Rolls ed., p. xlvi). The MS. is in the B. M. Cottonian collection, marked Cleopatra, B. v.” Whether this entry means that the building of Cardiff castle, or that

Except, perhaps, in some of the villein-trefs, there were no villages or clusters of dwelling-houses close adjoining one another, though the principal hall of men of higher position had accessory buildings. The dwellings of some families were duplicated; in the summer they lived in a house on the higher part of their property called the havod-ty (literally, "summer-house "), and in winter returned to the hendref (literally, "the old stead"), that is, the principal residence set up in more sheltered places below.

One of the most interesting texts of this Book of the Law is that on Briodolion Leoedd (appropriate places). It is what in modern times we should call a "table of precedence", and though nominally it only applies to the arrangement of the household at the meals in the king's hall, it really determines and indicates the order of the different officers. The arrangement cannot be understood without stating the character of the house of a Welsh chieftain. Fortunately Giraldus Cambrensis has given us a fairly minute description of the typical Welsh house of his time, and further material for its reconstruction is also furnished by the laws we are considering, so that we can ascertain what it was like in the later period of the tribal system. The evidence of these two authorities has been summarised by Mr. Seebohm, and we cannot do better than quote his description: "The tribal house was built

of the town, began, the date is too early. This MS. D. is of the fifteenth century. Giraldus calls Carmarthen an "ancient city", and notices that it was strongly inclosed with walls of bricks, part of which were still standing (Desc. i, c. 10). It is only with the building of the stone castle that Carmarthen begins to be noticed in authentic history, at any rate, after Roman times. Dinevwr, higher up the Towy, was the seat of the South Welsh princes.

1

See English Village Community, pp. 239-40; Report, p. 691.

of trees newly cut from the forest. A long straight pole is selected for the roof-tree. Six well-grown trees with suitable branches, apparently reaching over to meet one another, and of about the same size as the roof-tree, are stuck upright in the ground at even distances in two parallel rows, three in each row. Their extremities bending over make a Gothic arch, and crossing one another at the top each pair makes a fork, upon which the roof-tree is fixed. These trees supporting the roof-tree are called gavaels, forks, or columns, and they form the nave of the tribal house. Then, at some distance back from these rows of columns or forks, low walls of stakes and wattle shut in the aisles of the house, and over all is the roof of branches and rough thatch, while at the aisles behind the pillars are placed beds of rushes, called gwely (lecti), on which the inmates sleep. The footboards of the beds. between the columns form their seats in the daytime. The fire is lighted on an open hearth in the centre of the nave between the two middle columns."" This tribal house was the living and the sleeping-place of the household. The kitchen and buildings for cattle and horses were separate and detached, and it seems that, if not the whole set of buildings, yet the set of buildings with more or less completeness was duplicated for summer purposes on the higher grazing grounds. The house of persons of smaller importance was not, of course, so extensive. Giraldus describes the ordinary house as circular, with the fireplace in the centre and beds of rushes all round it, on which the inmates slept with their feet towards the fire."

In the king's house screens extending from each middle.

1 See also Arch. Camb., 3rd Ser., vol. iv (1858), p. 195; and 4th Ser., vol. x (1893), p. 172.

X,

2

Report of Welsh Land Commission (Lond. 1896), p. 692. Desc., i, and xvii.