in the west end (though the building is not strictly oriented, but stands more nearly N.E. and S.W), is 4 feet 10 inches high, with inclining instead of perpendicular sides. There is one window in the east end of the building 18 inches by 15 externally, and splayed to the interior both vertically and horizontally, but so much vertically that in the interior it is broader than it is high. The second of these quadrangular buildings stands a little apart from the general group. It is of somewhat smaller dimensions, rudely built, and its side walls begin to converge almost from the floor, and thus gradually inclining towards each other with a gentle curve give the structure somewhat the appearance of an inverted boat. The roof is finished at the height of 8 feet by slabs laid across the narrow space between the converging walls. The door is in the west end, and is 3 feet 6 inches high, with inclining instead of perpendicular sides. There is one window, placed in the east end, 2 feet wide and 1 foot high externally, and internally 20 inches wide and 10 inches high. It is thus broader than it is high, and splayed from the interior externally. The features in which these two structures differ from the beehive cells, with which we find them associated, are thus-(1) their quadrangular form on the ground-plan both externally and internally, (2) their doorways being placed in the west end, (3) their having a small window in the east end, and (4) the remains of an altar platform under the east window. There is thus no difficulty in concluding that notwithstanding their small size and the rudeness of their construction, they were edifices constructed for worship and not for ordinary habitation.1 Close by the still ruder cells, which were the dwelling-places

1 Besides these there is the larger church of St. Michael, which, though mostly lime built, has part of its walls of uncemented stone. It has its doorway in the south wall, with jambs of dressed stone, and an east window 3 feet 7 inches high by 11 inches wide, with a round-arched head cut out of a single

of the monastic family, are two stone enclosures, which contain their graves. A great cross stands in the centre of one, and round two sides of the other is a line of pillar-stones cross-graven or rudely cross-shaped. Such are the salient features of this most characteristic group of early Christian remains. There is no historic evidence by which a precise date can be assigned to any portion of them. There are incidental notices in the Annals from which the existence of an ecclesiastical settlement on the Skellig may be inferred from the year 812 to the year 1044. But it is not a necessary part of the function of archæology to determine the dates of its specimens. What it does is simply to classify them according to their several types, and to determine the relative sequence of these types. Whatever may be the precise dates of the different members of this group of ecclesiastical remains, it is clear from the characteristics which have been described that they belong to the class consisting of a church or churches (that is, a form of structure which is not indigenous), associated with a cluster of dwellings constructed in the native manner, and surrounded by a rath or cashel, thus forming a composite group of a special type, of which it may be concluded that no earlier is likely to be discovered, because it presents a mixture of forms and characteristics which partly belong to Christian and partly to Pagan times, and thus marks the transition from the one system to the other.

The second group which I have selected is that on Ardoilean or High Island, lying six miles off the coast of Connemara, and equally wild and inaccessible. The group of structures composing the monastic settlement is surrounded by a cashel or uncemented stone wall, nearly circular, and enclosing an area of 108 feet in diameter. The entrance is on the south-east

stone. In the order of time it must be placed after the two dry-built structures previously described.

side of the cashel, and on either side of it, and outside the enclosing wall, were circular buildings, probably intended for strangers and pilgrims not belonging to the monastic family, as we are told was the arrangement in St. Cuthbert's monastery in the island of Farne. The principal structure within the cashel is the church, one of the rudest of all the early churches of Ireland. Internally it measures but 12 feet by 10, and 10 feet in height. The doorway is 2 feet wide and 4 feet 6 inches high, and its horizontal lintel is decorated with a cross exactly similar to that on the lintel of St. Fechin's Church at Fore. It has one window in the east end 1 foot high and 6 inches wide, with a semicircular head. In 1820, when Dr. Petrie visited it, the altar still remained, and was covered with rude offerings such as nails, buttons, and shells, but chiefly with fish-hooks, the most characteristic tributes of the calling of the votaries. On the east of the church is a large cist of slabs, with a slab for a cover. The stones at the ends are carved with crosses. The church is surrounded by a wall about 15 feet distant, and from this wall a covered passage, about 15 feet long by 3 feet wide, leads to a cell or circular house of uncemented masonry, dome-roofed in the usual way, and measuring 7 feet by 6 internally, and 8 feet high. On the east side there is another cell of similar construction 9 feet square and 7 feet 6 inches high. In both these cells the doorways are only 3 feet 6 inches high and 2 feet 4 inches wide. On the other side of the chapel are a number of smaller cells about 6 feet long, 3 feet wide, and 4 feet high. Within the enclosure there are also a number of stone crosses and flags sculptured with memorial crosses, probably sepulchral monuments. The foundation of this settlement on Ardoilean is attributed to St. Fechin, abbot of Fore,1

1 O'Flaherty's Iar Connaught, p. 114. Ardoilean is also celebrated for the eremitical retirement of St. Gormgall, who died 5th August 1017, together with divers holy hermits that lived with him. Ten of them are named by

who died in 664, and to whom, under the Latin form of Vigeanus, we find a church dedicated at St. Vigeans in Forfarshire-a site of special interest in connection with its group of sculptured stones, to be described in a subsequent lecture.

The third group which I have selected for description is the ecclesiastical settlement of St. Molaise on Innismurry,1

Fig. 31.-Ground-plan of the Cashel and its included structures on Innismurry. (From Lord Dunraven's Notes on Irish Architecture.)

an island in the Bay of Sligo. It is surrounded by a stone wall or cashel (Fig. 31) enclosing an irregularly circular space Colgan-Acta Sanctorum, p. 715. Mr. Kinahan, in a communication to the Royal Irish Academy in 1869, gives some further details of the characteristics of these structures. He found that the ruins had been greatly destroyed since Petrie's time. The cashel wall surrounding the church has a chamber in its thickness 32 feet long, 4 feet wide at the bottom, and coving in to 3 feet at the top, where it is roofed by flags laid across, after the manner of the chambers in the walls of the stone forts of the Pagan time. This chamber is entered from the interior of the cashel by a doorway at the south end 3 feet high and 24 feet wide. Diagrammatic representations of some of the cells (one of which closely resembles that on North Rona) are given as illustrations to his paper. -See Proceedings R. I. A., vol. x. p. 551, and Plates 45-48.

1 Beranger, in his Tour in Connaught, 1779, gives an amusing account of

of 200 feet in diameter. This wall is built of rough uncemented stones, and varies in thickness from 11 to 13 feet on the north side, and on the south from 7 to 8 feet. It is still in some parts as much as 13 feet in height. The gateway on the north-east side is quadrangular, but with inclining instead of perpendicular jambs. Thus we have traced this peculiar feature of construction back through all the pre-Norman varieties of ecclesiastical structure, till we find it also characterising the cashel which surrounds the church and forms the link between the architecture of the Christian period and that of Pagan times. This gateway measures 6 feet 3 inches in height, 3 feet 5 inches in the width of the

his visit to Innismurry in company with Mr. Irwin, the proprietor of the island. The first thing that attracted his attention was a curragh, or boat, made of basket-work, and covered with a horse's or cow's skin. "As the members were six or eight inches asunder, and the sun was shining bright, and the skin transparent, it seemed to me to be a vessel of glass, as I could see the water through it." These boats, he says, were then common in the province. After describing the ruins and the wooden statue of St. Molaise, which the islanders had daubed over with red paint, "to make him look handsome," he gives the following account of a remarkable relic called "The cursing altar.' It is a kind of altar stone, "about two feet high, covered with globular stones, somewhat flattened, of different sizes, very like the Dutch cheeses; the tradition is that if any one is wronged by another, he goes to this altar, curses the one who wronged him, wishing such evil may befall him, and turns one of the stones; and if he was really wronged, the specified evil fell on his enemy; but if not, on himself, which makes them so precautionate that the altar is become useless." There was another turning or cursing stone at the well of St. Fechin, near Cong. Mr. Wakeman describes and figures the "cursing stone of St. Brigid," situated on the shore of Loch Macnean, near the church of Killinagh. It is a boulder of red sandstone 5 feet 9 inches by 5 feet 2 inches, flat on the upper surface, in which there are nine cavities, one near the centre of the stone, and the others placed irregularly round it. Each of these cavities contains a rounded or ovalshaped boulder smoothly water-worn. The use of this "cursing stone," by turning the stones in the cavities, with the expression of maledictions on the person to whom evil was wished, is yet dimly remembered.-Memoir of Gabriel Beranger, by Sir W. R. Wilde; Journal of the Royal Hist. and Arch. Association of Ireland, 4th series, vol. i. p. 135, and vol. iii. p. 459.