the ruins which formerly existed on St. Kilda,' it may have

Fig. 49.-Tigh Beannachadh, Gallon Head. Ground-plan.

(From Muir's Characteristics.)

1 Christ Church in St. Kilda, as described by Macaulay, was a little thatched temple, built of stone without any cement. No trace of it remains. The churches of St. Columba and St. Brendan are also gone. The "well of many virtues" to which pilgrims came from Harris, has a low massy square-shaped stone cell built over it, with a stone roof, and near it, says Macaulay, there stood an altar on which the pilgrims deposited such offerings as shells and pebbles, rags of clothing, pins, needles, rusty nails, and copper coins. Near it are the remains of a beehive-roofed cell, with small chambers in the thickness of the wall, known as "The Amazon's House."

tells, the saints of early times were wont to retire for a season of uninterrupted solitude.1

Of the same type, though somewhat better built, is the Tigh Beannachadh or Blessing House on Gallon Head in the

Fig. 50.-Teampull Beannachadh. West end. (From Muir's Characteristics.)

Island of Lewis. It is a rectangular oblong on the groundplan (Fig. 49), 18 feet 2 inches by 10 feet 4 inches in its internal measurements. The walls, which are about 3 feet thick, are built of undressed stones without mortar. The roof

1 "I have not been able," says Mr. Muir, "to find any legend regarding the holy man who founded the Teampull on Sula-Sgeir and kept ward at its altar. Whoever he was, he was surely a hero. That the conies, who are but a feeble folk, should yet make their houses in the rocks, is indeed a fact to be thought of; but that a still feebler nature, burdened with fears, and longings for ease, should even in its most transcendent flights of devotional zeal, have had the hardihood to seek out a home on a spot so morselled and wild, is beyond all understanding."-Characteristics, p. 206.

is gone. It has two doorways at the west ends of the north and south walls, and two ambries near the east end. The east window is splayed internally and underneath it is the altar.

Still smaller and equally rude in character is the Teampull Beannachadh (Figs. 50, 51), on Eilean Mor, the largest of the

Fig. 51.-Teampull Beannachadh. East end. (From Muir's Characteristics.)

Flannan Isles. Muir describes it as a primitive-looking thing, composed of rough stones joggled compactly together without mortar, built in the form of a squared oblong, but irregular on the ground-plan, the lengths of the side walls externally being respectively 11 feet 11 inches and 12 feet 2 inches, and the lengths of the end walls 10 feet 3 inches and 9 feet 2 inches. The walls vary in thickness from 2 feet 5 to 2 feet 11 inches. The roof which is formed internally by transverse slabs laid across from wall to wall takes externally the form of the bottom of a boat. The chamber measures about 7 feet

long, by 5 feet wide, and 5 feet 9 inches high. The doorway in the west end is but 3 feet high, and there is no window or other opening of any kind in the building. Its exceptionally small size, irregular construction, and the want of the usual east window, are quite uncommon features, and wholly inconsistent with any attribution of an ecclesiastical purpose. Yet it bears such a striking resemblance in its form and its features of construction to the class of structures whose characteristics connect them only with ecclesiastical associations, that we cannot on this account deny the possibility of its ecclesiastical connection. There is no known type of Pagan construction to which it can be affiliated, and if it be not the penitential habitation of some island hermit, who, like St. Cormac, had sought and found in this sequestered isle a desert in the ocean, it is impossible to draw more definite conclusions from the singularity of its characteristics.2

1

In this and the previous lecture I have thus endeavoured, by tracing back the sequence of types, to demonstrate the `character of the earliest type of Christian structures now remaining in Scotland, and to do so as clearly and fully as

1 In the times of the early monks we often find that towards the close of their lives they left the monastery which had been the scene of their labours, to seek some lonely island or mountain solitude, there to pass their latter days in undisturbed communion with God, and resting from all worldly care. Thus Caencomrac, Bishop and Abbot of Louth, died in Inis-enagh in 898, having "left Cluain in consequence of the veneration in which he was held there, for the neighbours worshipped him as a prophet, so that he went to seek for solitude in Lough Ree." Miss Stokes's Christian Inscriptions of Ireland, vol. i. p. 9.

2 On the western side of the island are some dry-built beehive cells, vernacularly called Bothien Clann Igphail, the largest containing two apartments, one of which is about 8 feet square, the other smaller and of an irregularly oval shape, both roofed with conical vaults rudely constructed of thin stones gradually converging into a dome, but leaving in the summit a circular aperture serving for smoke-hole and window.—Muir's Characteristics, p. 181.

the scanty nature of the materials will permit. It will be observed that of the actual age of any of the specimens which have been described I have said nothing. And I now repeat that it does not concern me as an archæologist to determine of any one of them whether it belongs to this century or to that, or whether it was built by this saint or by that. Facts and dates like these are the exclusive property of record. In the very nature of things they are unattainable, except through the medium of record. In its absence, the only alternative is that the investigation shall proceed on archæological methods; but from these it would be manifestly absurd to expect historical results. Archæology proceeds by the classification of types and determination of distinctive features, which give to these types an early character, but do not. necessarily attribute this earliness to the several specimens of the types. Primitive types are often the most persistent. The stone axe, for instance, one of the earliest of human instruments, is still in use. The older type has thus outlived all the types of the bronze instruments which successively appeared, and in certain cases it is absolutely impossible to say, without the assistance of record, whether a particular specimen may belong to the nineteenth century, or to the earliest times. Let me illustrate this more fully. Again I take the illustration from the process of spinning, because it is the oldest of the industrial arts of which we possess evidence in the remains of its special implements. I show you three such implements made of stone. They are discs of small size, pierced with a hole in the centre, by which they may be tightly jammed on the spindle, so as to give it the requisite momentum in twisting the thread. The three specimens―or whorls as they are termed-differ in size, in weight, in form, and in finish, but they do not differ so much as to prevent the conclusion that they are examples of the same type of implement. This conclusion is indeed obvious and inevitable, but