75. Blank. Entrance facing north-west; wall 4 ft. to 5 ft. high.

86. Yielded an ox tooth and some tiny fragments of red pottery. Entrance facing north; wall 4 ft. high.

The exploration lasted ten working days, and 32 huts were examined. Of these, 23 yielded charcoal or objects indicative of human occupancy, and nine were blank. Only two huts yielded a considerable amount of charcoal, viz., Nos. 4 and 50. The pottery found was wheelmade, and Late Celtic in character, excepting the tiny fragments of thin red pottery found in 13 and 86. These were undoubtedly Roman.

The bronze triskele found in No. 8, together with the largest of the porcellanic beads, is a spill 3 ins. long, looped at one end, and at the other carrying three curved arms springing from a plate, one perfect, the other two broken, the perfect arm terminating with an ornamental boss. The holed plate carrying the arms now revolves on a spill, and is retained on same by a head. Thus far we have found no clue as to the use of this object.1

The fibula may be described as of bronze, with ornaments in relief, the whole thickly plated with gold. One end is in the form of a half-tube, which has contained the spiral spring of the pin. From this rises a C-shaped curve formed of two members, both of them somewhat crescent-shaped. On that nearest the halftube is a symmetrical design consisting of two curves, each enclosing a rosette formed of seven pellets. The second member, which is much broader, is not in such good preservation. The ornament is in very slight relief, and certain parts of it are accentuated by being covered with dots. The prominent feature is, however, on each side a boss, now much decayed, but apparently the original idea of it was a spiral or helix

1 There is a similar object from Berkshire in the British Museum, and others have been found at Hunsbury, near Northampton, and Kingsholm, Gloucester.-ED.

like design. The rest of the bow narrows down to where it joined the shaft, now lost, and is modelled in converging curves, in the same way as the half-tube on the other end. The back has apparently not been gold-plated.

We may be allowed to give those conclusions at which we have arrived in consequence of the preliminary researches made in Tre'r Ceiri, subject to modification from the results of further exploration.

1. That the fortifications were probably erected and occupied by that people to whom the finds appertained. There was no evidence of any earlier occupation: not a trace of flint tools or weapons, not a fragment of pottery of the Bronze Age; nor were there any indications (with a possible exception of the bone comb) of a later occupation.

2. That Tre'r Ceiri was only temporarily-and that for a short time-occupied in the summer season alone, as the amount of charcoal found was remarkably small, and some of the cytiau seem not to have been occupied at all. The exposed position of Tre'r Ceiri and excessive dampness of the site would render it impossible of occupation, unless under dire necessity, during the winter.

3. That the race which erected the walls and constructed the huts was Celtic, probably British, and that the period to which they belonged was the first or second century of the Christian era.

Our reasons for coming to this conclusion are as follows:-(a) The pottery is wheel-turned and distinctively Celtic, but along with this was found a small amount of what is certainly Roman pottery. (b) The amount of iron found proves that the tenants belonged to the Iron Age, but there was nothing in the character of the tools and weapons found to determine the precise period in that age. (c) However, the fibula is unmistakably Celtic, and resembles one found at Æsica on the Roman Wall. (d) The two porcelain beads are of Egyptian manufacture. One of them, we are in

formed, is the finest that has been found and recorded in the United Kingdom, and by its shape, paste, and glaze reveals its origin as either Alexandria or the basin of the Nile. These must have been imported, and probably were so during, or even slightly preceding, the Roman occupation. (e) The distinguishing Celtic ornament of the fibula practically ceased, or became degraded, after the second century in Britain. During the first two hundred years of the Roman occupation, the originality and elegance of Celtic ornament influenced decoration in Britain, and even in Rome, being so totally different to the stiff and conventional character of Roman ornamentation.

4. The extraordinary rudeness and clumsiness of construction of the walls and huts seems to show that the builders had not been influenced by the Roman art of wall building; and this, in our opinion, points to the erection of the fortress at an early period of the first century.1

Finally, we would urge most strongly on the Cambrian Archæological Association the importance of making a thorough and exhaustive investigation of Tre'r Ceiri, of which our exploration was but preliminary, and limited to one purpose. True British relics of this period are rare, and till recently have been little studied and separated into a class by themselves. Furthur exploration should be made at Tre'r Ceiri, and especially of the entrances, which should be cleared of débris, so that they could be measured and accurately planned. The outer defences and the enclosures within them should also be examined.

Tre'r Ceiri is a difficult place to explore, for it is exposed to all weathers. It is peculiarly liable to be enveloped in fogs, and time must be allowed for excavation. The weather cannot be depended upon for continuous digging; days may pass when the workmen

1 Segontium (Carnarvon) was founded by the Romans towards the end of the first century.

employed on the research are condemned to compulsory inactivity. We would also earnestly recommend Mr. Wood, the owner of Tre'r Ceiri, to decide on some efficient public museum, under the charge of a competent curator, where the relics found or to be found may be carefully preserved and exhibited, for the sake of their great importance as a contribution to the knowledge of the arts of the British at the dawn of the history of our island.

17

THE ARCHITECTURAL HISTORY

OF THE

CATHEDRAL CHURCH OF ST. DEINIOL, BANGOR.

BY HAROLD HUGHES, ESQ., A.R.I.B.A.

(Continued from 6th Ser., vol. ii, p. 276.)

THOMAS SKEFFINGTON, Skevynton, or Pace, was appointed Bishop by Papal provision in 1508, and consecrated in June, 1509. To him is generally given the credit of building the nave and western tower. An inscription over the western doorway, indeed, states that the tower and church are his work. The reconstruction, commenced by Bishop Dean at the east end of the church, may have been gradually carried on westward during the few years intervening between his translation to Salisbury and the appointment of Bishop Skeffington. There is nothing definite to indicate where the one work ended and the other began. Probably, Skeffington's work commenced with the rebuilding of the nave arcades. These were inserted between the fourteenth-century aisle walls. Whereas the central line of the fourteenth-century nave, if continued, would have struck the east window considerably to the south, that of the fifteenth-century nave strikes the window to the north of its centre. The arcades consist of six bays, and have been built without reference to the aisle walls. The latter are divided into seven bays. The arches are four-centred, and are of two orders of hollow chamfers, contained

1 Browne Willis, p. 96.

2 Bishop Dean was translated to Salisbury in 1500, and succeeded by Thomas Pigot, who died in 1504. John Penny, the next bishop, was translated to Carlisle in 1508.

6TH SER., VOL. IV.

2