the floor for the principal apartments. The wall-plates of broad flagstones, and the apertures for the discharge of the water, show where the roof was set on; but, what is chiefly remarkable is, that the parapet-wall of the tower rises to so great a height as must have quite masked the roof, even though of a very high pitch. Perhaps the design was to protect the sedge or other light materials with which it may have been covered; for the tower stands exposed to every blast from the Atlantic. I afterwards observed the same disproportion between the height of the roof and parapet at the church of MacDuagh on Arran More.

Re-entering among the rocks, we passed through another village, the pathway to which runs between enclosures of apparently a very unprofit able kind; for, in several cases, the only thing enclosed is the bare surface of limestone, no earth having yet been laid down; and, when earth does cccur, it is wholly adventitious, having been carried from a distance and spread on the rock. Yet these patches of fictitious soil yield very good crops of oats and potatoes. To see the careful way in which the most has been made of every spot available for the growth of produce, might correct the impression so generally entertained and so studiously encouraged, that the native Irish are a thriftless people. Here, where they have been left to themselves, notwithstanding the natural sterility of their islands, they are certainly a very superior populationphysically, morally, and even economically-to those of many of the mixed and planted districts.

This practice of forming artificial fields by the transport of earth, recalls the old tradition of the Fir-Volgic origin of the early inhabitants of Arran. It is stated by Duald MacFirbis, on the authority of an ancient tract preserved in the book of Leacan, and the statement is corroborated by very evident remains of which I shall speak by-andbye, that after the overthrow of the Fir-Volg, by the Tuatha-de-Danaan, at the battles of Traigh Eochaille and northern Moy-Tuire, the remains of these people crossed over to the Isles of Arran, and inhabited them at the beginning of the Christian era. These Fir-Volg, according to their own account, were Thracians, who had been enslaved in Greece, and there employed

in carrying earth in leathern bags, to form the artificial terrace-gardens of Boeotia. If any portion of the existing population of Ireland can, with propriety, be termed Celts, they are of this race; and, certainly, those who now represent them here, whether of Welsh or Gaelic descent, do the name no discredit.

Passing beyond the village and its rugged, diminutive gardens, the track conducts to the little old cell of St. Gobinet, seated under the shelter of a great limestone crag, and backed by a stunted thorn-bush, the one tree of the island. St. Gobinet's is of the primitive type, so fully illustrated by Petrie ; and may vie in diminutiveness with any of those described in the Round Tower Essay, measuring, internally, no more than thirteen feet by nine. A corpu lent person would find some difficulty in entering by the narrow Egyptianformed doorway. This Gobinet is the same who has given her name to the church of Kilgobbin, in the county of Dublin; and there seems no reason to doubt that her cell in South Arran is of the period at which she lived, the early part of the seventh century. South Island seems to have been a favourite resort of female ascetics. At the western extremity of the little valley, running inland from St. Kevin's, are the graves of seven holy ladies, called the Seven Sisters, and now converted into penitential stations. The path leading to the fountain here is the only part of the whole island where a person can walk a distance of twenty yards on tolerably level ground. All the rest is rock, reticulated with dry stone walls enclosing scattered patches of cultivation and pasture.

An hour's rowing brought us from Inishere over Gregory's Sound to Middle Island, or Innis Maen. On approaching our landing-place on the southern extremity of the island, my eye was attracted by two edifices, the only ones in sight, in remarkable contrast with one another. Close by the sea, under the shelter of a similar wall of limestone to that which encircles the inner platform of Inishere, is a church, to all appearance older, as it is still smaller, than that of St. Gobinet. On the eminence above stands one of the circular stone fortresses of the Firvolg, enclosing, perhaps, half a rood of land, and its walls, of twelve feet in thickness, still standing, to the

height, in some places, of fifteen feet. The little church is called Teampul Cinerigi; the fort, Dun-Farvagh. Who Kinerigy, or Cennanach, may have been I know not; but an elder bush of moderate size suffices to fill his church to overgrowing: its whole interior space might be scooped out of the thickness of the wall of the Gentile stone fortress that overlooks it. Its dimensions, internally, are twelve feet by eight; and when roofed, it received its modicum of light through a triangular-headed cast window of some ten inches high. The doorway, not more than two-andtwenty inches across, is of equal width at top and bottom. In the north-east angle the walls are joined by the mere apposition of the stones, without any tying of the masonry. I know not whether it was from the idea of a portable shrine, or arc, that these little churches were designed; but in this, and several other primitive churches, I observe a peculiar feature, not irreconcilable with this idea. At each end, two of the quoin stones of the side wall are made to project, like handles, by which the little structure might seemingly be lifted, as one would lift a sedanchair. I vouch for nothing but the fancifulness of the idea, and the fact, that the primitive churches of Macduach on Arran More, and of Oughtmama, in Burren, are provided with similar appendages. These cellule have no chancels; as I shall also have to observe respecting the larger churches of the same epoch.

The Christian remains on Innis Maen yield in interest, as well as importance, to the pagan. The fort on the hill over Teampul Cennanach, imposing as it appears in contrast with that little oratory, is of moderate dimensions when compared with the really magnificent stone fortress of Dun-Conor, which crowns the central summit of the island. A square barbican, containing about a rood of ground, stands in front of the entrance, towards the east. The walls of this outpost are about six feet in thickness, built, as all the rest of these Gentile works are, without mortar; but the size of the stones is not sufficient to justify the appellation of Cyclopean. In all the structures of this kind which I shall have to mention, the stones employed in the main work of the walls are such as could be carried by, at most, two or three men of ordinary strength.

In this respect, the Fir-Volgic remains in Arran correspond, I believe, with the Picts' towers of the Scottish isles, which Irish tradition assigns to the same people for it is said, that on the dispersion of the Fir-Volg, after their last stand at Moy-Tuire, they took refuge not only in the isles of the west and north-west of Ireland, but also in Rathlin, Isla, and the Innis-Gaul, or Hebrides. Judging from the nearer approximation of Dun-Dornadišla, and other like structures in the Scottish isles, to the form of the feudal round castle in its early development, as at Coucy and Donagore, it seems probable that these are of more recent date than the wider, lower, and less artificial stone fortresses of the west of Ireland. A peculiar Greek feature, however, which strikingly corroborates the tradition associating these structures with the Cyclopean architecture of Greece, the triangular aperture over the doorway, present in the Hebridean duns, is wanting in these Arranese stone fortresses. The doorway of DunConor is broken down, and although the walls which remain are of a considerable height externally (twenty to twenty-five feet in some places), the internal arrangements can only be guessed at through masses of debris, the ruins of the upper portion of the fort. From the remains, however, of several flights of stairs, still visible on the inner face of the rampart, corresponding to similar remains in DunFarvagh, and other fortresses of like construction, it appears pretty clearly, that the construction was much the same as that of Staigue fort in Kerry, a model of which may be seen in the vestibule of the Dublin Society House, and which, as it is the most perfect of these remains, though not comparable in extent to the Arran fortresses, I may here shortly describe :Externally, the circular enclosure presents the appearance of a low, round tower of wide diameter, bulging above the base, and thence receding to the summit-a form observable in several of the Hebridean duns, and apparently designed to prevent escalade. A single low, broad doorway admits to the interior. Within, the thickness of the wall, at about six feet from the surface, is diminished by one-third, so as to leave a circular ledge, or terreplein, of five or six feet in width, projecting all round. This ledge is reached

by flights of stairs, on the inner face of the wall. At a height of five or six feet higher, another contraction of the thickness of the rampart takes place, leaving a like ridge, or shelf of masonry, approached in like manner, by steps from the former, and serving as a kind of banquette to the parapet formed by the remaining height of the rampart. At Staigue, the flights of stairs are carried up in symmetrical lines, each lower pair of stair-flights converging to the point where each upper pair diverge, and so surrounding the internal face of the rampart with a reticulation of pyramidally-arranged stone steps. In the Arran fortresses the same distribution of the rampart into three successive thicknesses, forming successive platforms, or banquettes, on the interior face, is still quite traceable, but the stairs which led from one to the other do not appear to have been symmetrically arranged, or so numerous as at Staigue. The communication with the first platform at DunFarvagh appears to have been by a stair ascending laterally on the right of the entrance. At Dun-Conor there are tiers of several ascents, not laterally stepped in the plane of the wall, but carried perpendicularly into its thickness, giving access to both banquettes. An arrangement in the building, exhibiting a good deal of military contrivance, is made subservient to the formation of these internal stages. Instead of building the rampart in bulk, and starting with a fresh face of masonry above each ledge, the Fir-Volg builders have, in every case, built their rampart from the foundation in as many concentric independent walls as they designed to have banquettes; so that if an enemy should succeed in breaching the external envelope, he would find immediately behind it a new face of masonry, instead of the easily-disturbed loose interior of a dry stone wall.

The outer envelope, as rising higher than either of the others, and having only its own thickness to oppose to the elements, has fallen all round DunConor to the level of the second, and in some places below it; so that what formerly constituted the upper banquette behind the parapet, now forms the top of the rampart its independent face of regular stone-work being visible in some places as low as the

foundation, through breaches of the external rind of masonry that formerly overlapped and overtopped it. The dimensions of these walls is something surprising. Making allowance for the disruption and spreading of the masonry, each envelope appears to have been nine feet thick at the base, giving an aggregate breadth for the composite triple rampart, of twenty-seven feet. The original height may have been about twenty-five feet. Cavities are discernible, which seem to indicate the site of chambers in the wall; but the dislocation of the loose materials renders this a very uncertain speculation. Exposed as the fabric is to tempests from the Atlantic, and dependent for its cohesion on the weight of its materials only, it is surprising that it should have stood, even in ruins, for so many centuries. The walls, however, are built with considerable art, long stones being employed on both faces, and carefully laid with their ends outward. Dun-Conor covers a large space of ground; the area is an irregular oval, the greater axis measuring seventy yards, and the lesser forty. Looking at the enclosure as it now stands, one is led to speculate on the sort of habitations its tenants may have had within it. Traces of minor buildings appear over the area, but too indistinct to afford material for any tangible conjecture. A sloping roof might easily have been erected against the inner face of the wall; and from the occurrence in other buildings of the class, of recesses and cells round the internal area, it seems not improbable that such may have been the arrangement. It is interesting to trace the gradation from the single enceinte of stone, behind which the warrior could stand, and throw his dart at his enemy, to the round castle-tower of feudal civilisation. First, we have the means of access to the summit of a loftier rampart, provided by means of stairs cut perpendicularly up the inner surface, as here at Arran; next, we have these stairs and passages, the extent of which before was limited by the thickness of the wall, carried spirally up its plane, and included within. it, as at Dun-Dornadilla and the Scottish Picts' houses; next, the height being thus capable of indefinite increase, the diameter is narrowed, and, while a covering, thrown over the top,

excludes the weather, light is admitted by windows, pierced through the upper side walls; but the stairs and chambers of greater security are still retained within their thickness, as at Dunagore; and, finally, the admission of light by windows diminishing too much the strength of the walls, to admit of other cavities within them, the stairs, and all other accomodations of the dwelling, are brought under the area of the roof, as in the finished castle or modern dwelling-house.

Leaving Dun-Conor with feelings of admiration strongly excited, I pursued my way to the beach by Killmurry, a little old church, which, with the addition of a modern transept, suffices for the island congregation, and past the grave and holy well of St. Canannach. Leaba Cinerigi, as the grave is called, serves as a penitential station; it consists of an oblong pile of rough stones, of no great size, and appears as puny a work, compared with a Gentile sepulchre, called by its generic name of Leeba Diarmuid as Graine, on the beach hard-by, as the patched, dwarfish chapel does in comparison with Dun-Conor. I were ungrateful, having recorded the name of my guide, on Inishere, if I left Martin Faherty, my companion on Innis Maen, uncommemorated. He was waiting for the evening tide to return to Kilronan with his hooker, and would have been happy to accommodate me with a passage, or to tow us at his stern; but time pressed, and I put off for Arran More in my coracle.

A spit of sand, terminating in what constitutes an island at high water, forms the southern boundary of the harbour of Killaney. Here we landed, through a considerable surf; and having drawn our canoe across the isthmus, launched again on the still waters of the inner basin, crossing which, we reached the creek and pier of Kilronan.

After a two days' dependence on the hospitality of poor villagers, I found the change to Mrs. Costello's comfortable, albeit deal-furnished and carpetless, apartments extremely grateful. Her house immediately adjoins the inner pier; and should her two rooms be occupied, the visitor will find accommodation of the same description in the house of Mr. Patrick Dillon, a little farther up the street of the village. Patrick Mullen is the guide and

antiquary of the island. Hitherto his intellectual pursuits have not added much to his worldly wealth; but as the island becomes better known, Mullen, I should hope, will be able to make an appearance more suitable to the dignity of his calling. Ponies can easily be procured, and I believe Mrs. Costello can, if necessary, furnish a side-saddle; but there are no wheeled vehicles of any description on this or either of the other islands. It is, however, a great relief, after the toil of stepping from rock to rock, or scrambling up and down the rugged trackways of Inishere and Innis Maen, to find here, all along the eastern coast of Arran More, a smooth and tolerably level road. The island faces the east and north, and rises to the south-west. All the higher portions are bare rock, although divided by innumerable dry stone enclosures. In the crevices, which everywhere occur through the limestone, there is found a sweet winter herbage; that none of these paddocks can be said to be absolutely barren, but the quantity of pasture fenced off by such an enclosure is extremely small. The fertile soil of the island lies at foot of the rocky ascent, along the eastern shore; and in this tract all the religious houses have here been located; but the great Gentile stone fortresses of Dun-Angus, Dun-Eochaill, and DooCathair are erected on the summit and Atlantic verge of the rocky desert above; Dun-Angus on the north, Doo-Cathair on the south, and DunEochaill, on the highest and central point of the island. The first speculation that suggests itself on sight of these immense keeps is, as to whence supplies were procured for so large a force as they must have needed for their occupation. In the end of the fifth and beginning of the sixth century, when the first Christian recluses settled here, the difficulty of obtaining supplies of food for even a few persons was so great as to furnish material for the best part of their chroniclers' miracles. Was it by plunder, or by those arts of rock-cultivation which their ancestors practised in Greece, that Angus and Muirbheac Mil, and Conor, and Farvagh victualled their stone palaces? Of the Picts, their cousins, or, more probably, their brethren, who remained in Ireland after

the first Cuithneac emigration, recorded in the Irish Nennius, we are told(Irish Nennius, p. 145)—

Plundering in ships

By them was taught

Hills and rocks they prepared for the plough ;"

so that, they may have been utrobique parati; but, I should suppose the former method of livelihood the more probable. King Angus must have possessed more treasures than the rocks of Arran could well have afforded, when he employed such labour as has been expended at DunAngus, in fortifying the approaches to his stronghold.

I noticed the Pictish form given to the name as pronounced by the Arran people, "Doon-Ungust," and "DoonUnguish." It is the same name as the Scandinavian Hengist, the Roman Ancus, and the Trojan Anchises. Angus, the founder of this fortress, was son of Uaidhmore, and built it in the first century of our era (Mac Firbis's Account of the Firbolgs, Book of Lecan, fo. 2776). As the most extensively fortified and best authenticated of these remains, it will probably be the first object sought by visiters. The road to it from Kilronan abounds in objects of the greatest interest.

THE expectation which the announcement of Lord John Russell's "Memoirs of Moore" was calculated to create, has not been disappointed. The first two volumes have appeared, and are among the most interesting books we have read. A graceful preface tells the share which Lord John Russell has in the work, which as yet seems to have been little more than selection. The work opens with a memoir commenced by Moore in the year 1833, in which he gives an account of his early school and college life, but which is continued no farther than to the period of his becoming a law student at the Middle Temple, in order to his being called to the Bar. The memoir is followed by nearly 400 letters of Moore-the first bearing date April 3, 1793; the last, November 8, 1818. At the close of this part of the work is the following note of the edi

tor:

"These letters are, many of them-most of them, I may say without a full date, and I fear several have been wrongly placed."

Then follows a very full diary of Moore's, extending from the 18th of August, 1818, to the 30th of August, 1819 this being, as might be antici

pated, the most interesting part of the book.

In Moore's will, written in 1828, there was a request that Lord John Russell should, from such papers or letters as Moore might leave, form some publication that might afford the means of making some provision for his wife and family. This publication is undertaken in fulfilment of that request. The business of selection from a mass of letters written on private business, must always be accompanied with difficulties. Had there not been a sort of custom, creating almost an indisputable law of society on such subjects, we should have imagined it more than doubtful whether the kind of confidence in which letters giving domestic details, intended for friends and not for the public, are written, is not violated by such exposure; and we rather think that such things, after being read to the fireside circle, for whom they are intended, should be thrown into the fire. Such was, we believe, Sydney Smith's habit—a man not likely to err in a question of the minor morals of life. However, the question seems to have been settled the other way. It was regarded as a moot point in the days of "Mason's Gray." Hayley, in his "Cowper," followed the example; and there seems now no feeling on the

VOL. XLI.-NO. CCXLI.

4