[1] The word Crannoge, by which the artificial island fort was designated in the Irish Annals (modified by Drs. Robertson and Stuart into Crannog), is derived from the Gaelic crann, a mast or tree; but as it is doubtful whether this etymology applies to the timber of which the island was constructed, or to the wooden huts erected over it, its use as a precise term to indicate the scope of this work would be equally doubtful. Hence I have preferred the word Lake-Dwelling.

[2] Journal of the Royal Historical and ArchÆological Association of Ireland, vol. ii. p. 305.

[3] Proceedings of the Royal Irish Academy, vol. i. p. 425.

[4] Proceedings of the Royal Irish Academy, vol. vii. p. 155.

[5] Ibid. vol. viii. p. 417.

[6] Journal of the Royal Historical and ArchÆological Association of Ireland, vol. i. p. 362.

[7] Journal of the Royal Historical and ArchÆological Association of Ireland, vol. i. p. 583.

[8] Arch. Journal, vol. iii. p. 47.

[9] Proceedings Soc. Antiq. Scot. vol. vi. p. 114.

[10] Proceedings Soc. Antiq. Scot. vol. iii. p. 43.

[11] "Instructions to Andro bischop of the Yllis, Andro lord Steuart of Vchiltrie, and James lord of Bewlie, comptroller, conteining suche overturis and articles as they sall propone, to Angus M?Coneill of Dunnyvaig and Hector M?Clayne of Dowart for the obedyence of thame and thair clanis. 14 Aprilis 1608.... That the haill houssis of defence strong-holdis and cranokis in the Yllis perteining to thame and their foirsaidis sal be delyverit to his Maiestie and sic as his Heynes sall appoint to ressave the same to be vsit at his Maiesty's pleasour.... That they sall forbeir the vse and weiring of all kynd of armour outwith thair houssis especiallie gunis bowis and twa handit swordis, except onlie ane handit swordis and targeis."—(Regist. Secreti Concilii: Acta penes Marchiarum et Insularum Ordinem 1608-1623, pp. 4, 5. Robertson's Notes.)

[12] The island in the Loch of Forfar, known as Queen Margaret's Inch, was discovered to be artificial on the partial drainage of the loch in 1781. It is thus referred to in the Old Stat. Account of Scot., vol. vi. p. 528:—

"Before this loch was drained, and near the north side of it, there was an artificial island composed of large piles of oak and loose stones, with a stratum of earth above, on which are planted some aspen and sloe trees, supposed to have been a place of religious retirement for Queen Margaret. This now forms a very curious peninsula. The vestiges of a building, probably a place of worship, are still to be seen.... It appears that the loch has at some period surrounded the rising ground, called the Manor, and the adjacent hill, on which the Castle of Forfar stood; which hill is not, as the authors of the EncyclopÆdia Britannica suppose, artificial, but a congestum of sand and fat clay, evidently disposed in various irregular strata by the hand of nature."

Dr. Stuart says:[13]—"The drought of 1864 brought to light a sort of causeway, leading from the west end of the island. It was traced for about 100 yards; and it is supposed that it turned to the shore on one side, the popular belief being that it formed a way of escape in former times. As, however, it must have formerly been under a great depth of water, it seems doubtful for what purpose it may have been designed." (For historical notices of this Inch, see Proc. Soc. Antiq. Scot. vol. vi. p. 310.) Subsequent excavations (autumn of 1868) prove that St. Margaret's Inch is "the highest part of a narrow ridge of natural gravel which runs out into the loch, and the so-called causeway is the continuation of this ridge as it dips into the deep water." Dr. Stuart, who was present during the operations, remarks that the results obtained "afford another instance of the little reliance which can be placed on the descriptions of early remains given by the observers of last century, so far as relates to details." This is in allusion to Dr. Jamieson (ArchÆologia Scotica, vol. ii. p. 14) and others, who describe the Inch as being wholly artificial. (For Dr. Stuart's report, see Proc. Soc. Antiq. Scot. vol. x. p. 31.)

[13] Proceedings Soc. Antiq. Scot. vol. vi. p. 125.

[14] "The only antiquity in the parish is the fortalix at Lochindorb, where a thick wall of mason-work (20 feet high even at this period, and supposed to have been much higher) surrounds an acre of land within the loch, with watch-towers at every corner, all entire.... Great rafts or planks of oak, by the beating of the waters against the old walls, occasionally make their appearance; which confirms an opinion entertained of this place, that it had been a national business, originally built on an artificial island."—(Old Stat. Account, vol. vii. p. 259.)

[15] Before the level of Loch Canmor was reduced in 1858, it appears that it contained four islands—only one of which was found to be artificial. Of the three natural islands, the largest has an area of about a Scotch acre, and is known as the Castle Island, because the traditional castle of Malcolm Canmore was placed on it. It is supposed, from the occasional fishing up of great oak beams between it and the shore, that it was connected with the mainland by two projecting piers and a drawbridge. Fordun's expression "in turre sua de Canmore," and other historical references to the Isle of Loch Canmor apply therefore to this island with its castle, and not to the artificial island. The following extracts regarding the artificial island are from a paper on Loch Canmor, prepared by the Rev. James Wattie, Bellastraid, at the suggestion of Mr. Robertson, who intended to use it in his article on Scottish Crannogs:—

"The Prison Island is about the middle of the loch, and about 250 yards from its north shore. It is something of an oval shape. It is 25 yards long and 21 yards broad. It is evidently artificial, and seems to have been formed by oak piles driven into the loch, the space within the piling being filled up with stones, and crossed with horizontal beams or pieces of wood to keep all secure. The piles seem to have been driven or ranged in a rectangular form. They are quite distinct and apart from one another. The upright ones are generally round, though some of them have been splitted. The horizontal beams are mostly arms of trees, from 4 to 6 inches thick; but there is one horizontal beam, squared evidently with an iron tool, about 8 inches on the side. There are not many horizontal beams now to be seen. I remember having seen more (the ends of trees) a good many years ago. My recollection of them is, that they had been splitted. There seems to have been upright piles on all sides of the island, but least distinct at the east end, and most numerous at the west. At the west end thirty upright piles are visible. On the south side, outside the regular row of piles, is a kind of out-fencing of upright and horizontal beams, seemingly for protection against the force of the water. At the west end there are two rectangular corners, and there may have been the same at the east end, though now overgrown with grass. Outside the piles is what may be called a rough, loose causewaying of stones sloping outwards into the water; while inside is what may be called a heap of stones, arising, no doubt, from the putting into the water of whatever building had been on it. At the west end the piles stand 18 inches above the present level of the stones, and from 12 to 15 inches apart. They are 4 inches thick at the top, and 6 inches thick where they had been under water. Scarcely any of the upright piles are perpendicular; they slope to the north on the west side of the island, and to the west on the south side. Round the heap of stones now forming this island a clump of trees has sprung up. There is no appearance of a pier or jetty about the island, nor any mark of communication between it and the shore, or any of the other islands. The present depth of the loch near the island is 7 feet; half-way between it and the Castle Island, 10 feet."

"On the 16th June 1859 there was fished up from the bottom of the loch, near the north shore, opposite to the Prison Island, a canoe (Fig. 1), hollowed out of a single oak-tree, 22¹/₂ feet long, 3 feet 2 inches wide over the top at the stern, 2 feet 10 inches in the middle, and 2 feet 9 inches at 6 feet from the bow, which ended nearly in a point. The edges are thin and sharp, the depth irregular—in one place 5 inches, the greatest 9 inches. There are no seats nor rollocks or places for oars; but there may have been seats along the sides, secured by pins through holes still in the bottom. There are two rents in the bottom, alongside of each other, about 18 feet long each; to remedy these, five bars across had been mortised into the bottom outside, from 22 to 27 inches long and 3 inches broad, except at the ends, where they were a kind of dovetailed, and 4 inches broad. One of these bars still remains, and is of very neat workmanship, and neatly mortised in. The other bars are lost, but their places are quite distinct. They have been fastened with pins, for which there are five pairs of holes through the bottom of the canoe, at the opposite side, at a distance of from 18 to 20 inches, the bottom being flattish. There are also five pairs of larger holes through the bottom, and also at the opposite sides, which may have been for fastening seats with pins along the sides of the canoe. There are two bars mortised longitudinally into the bottom of the boat outside, above the seats before spoken of, 2¹/₄ inches broad, one at the stern 5 feet long, and the other beginning 5 feet from the stern, and extending 7¹/₂ feet towards the bow. The canoe looks as it had been partly scooped out with fire. The bottom is 2 feet 8 inches wide at the stern, and 28 inches wide at the middle. The stern is 18 inches thick, and somewhat worn down at the top.

"M?Pherson, the turner, says that twenty years ago a boat was taken up from the loch 26 feet long, sharp at both ends, otherwise coble-built, 8 feet broad in the bottom, which was flat, made of oak planks, overlapping one another, and lined under the overlapping with wool and tar."—(Proc. Soc. Antiq. Scot. vol. vi. pp. 167-171.)

[16] In Maitland's History of Scotland I find the curious statement made that Boece states that in Loch Lomond there were fish without fins, waves without wind, and a floating island.—(Boet. Scot. Reg. Descript. fol. 7.)

[17] Proceedings, 8th December 1865.

[18] See also Proceedings of Dumfries and Galloway N. H. Soc. for 1865.

[19] Proceedings Soc. Antiq. Scot. vol. vi. pp. 114 et seq.

[20] Proceedings Soc. Antiq. Scot. vol. iii., new series, p. 155.

[21] Proceedings Soc. Antiq. Scot. vol. ii., new series, pp. 127, 128.

[22] Proceedings Soc. Antiq. Scot. vol. iii., new series, p. 266.

[23] These relics were sent as donations to the Museum of the Society of Antiquaries of Scotland, by the Right Hon. the Earl of Stair, in 1872.

[24] The existence of these coins in the crannog is thus explained by Mr. Dalrymple:—"It is known that the island has been planted two or three times, and that considerable quantities of soil and stones have been added to it. The two feet of soil which covered the uppermost remains, and which so much raised the centre of the crannog, was probably added, in great part, about 1720, when Field-Marshal the Earl of Stair laid out the grounds of Castle Kennedy. Some of these operations may, to some extent, have disturbed the remains. They would, at all events, account for the modern coins found so far below the surface."

[25] Before this sketch was taken some of the horizontal beams were removed.

[26] In the Museum of the Royal Irish Academy are several shears similar to the above, which were found on the crannogs of Dunshaughlin, Clonfinlough, and Strokestown; and in the York Museum there is also a collection of articles of the Early Iron Age and late Celtic period, amongst which I noticed five shears made on the same principle as that from Lochlee.

[27] Colonel Gould Weston, F.S.A., has pointed out that one of these devices is a fylfot (croix gammÉe or swastika), an ancient symbol which in modern times has called forth a considerable amount of speculative writing. Its occurrence on four Irish monumental stones of the early Christian period has been the occasion of a recent article by the Bishop of Limerick (see Proceed. of Royal Irish Acad. vol. xxvii. part 3). The following extract from a paper, by M. Oscar Montelius, on the Sculptured Rocks of Sweden, is of interest as bearing on this point:—

"La frÉquence de la roue ou du cercle crucifÈre (Fig. 11) et l'absence totale de la croix gammÉe (Fig. 12). Toutes deux sont, sans doute, des symboles religieux. La premiÈre (Fig. 11) qui se trouve trÈs-souvent sur les monuments de l'Âge du bronze, est presque totalement inconnue pendant l'Âge du fer. La croix gammÉe (Fig. 12), au contraire, est trÈs-frÉquente pendant ce dernier Âge; je ne l'ai jamais vue sur les rochers sculptÉs dont nous parlons À prÉsent."—Compte-Rendu, Congres Inter. d'Anthrop. et d'Arch. PrÉhistorique, 7^me Session, 1874, Tom. i. pp. 459, 460.

See also Dr. Schliemann's works on the excavations at Troy and MycenÆ, where both these symbols are referred to as of frequent occurrence. In Dr. Schliemann's more recent work on Troy or "Ilios," an interesting account of the meaning and prevalence of this symbol among all nations is given.

It is found on some of the sculptured stones of Scotland. On a slab of greywacke from Craignarget, Gillespie, Glenluce, now deposited in the National Museum, Edinburgh, there is a cross on the upper part, with the sun and moon in the usual position above the arms, and two small crosses underneath, and below them a fylfot or swastika, together with cup-marks and concentric circles and various other devices. (See woodcut page 251, vol. iii. Proceed. Soc. Antiq. Scot., new series.)

It also occurs on the famous Newton stone, along with two inscriptions, one of which is written in "Roman minuscular letters of an exceedingly debased form," and the other in Ogham characters, as well as on several other monuments of Christian time.

According to Dr. Joseph Anderson, although of Pagan origin, the fylfot has become a Christian symbol from the fourth to the fourteenth or fifteenth century.—(Scotland in Early Christian Times, vol. ii. p. 218.)

It is seen also in a mosaic pavement in the recently explored Roman villa in the Isle of Wight. (See Report by Cornelius Nicholson, F.G.S.)

[28] See Nathusius, SchweineschÄdel, 1864, p. 147; RÜtimeyer, Basel. Gesell. Naturforscher, 1864, p. 161; Naumann, Archiv fÜr Anthrop. viii. p. 23, 1875; StÜder, ZÜrich. Mittheilung., Pfahlbauten, 1876, xix. 3, p. 67.

[29] For reference to the bibliography of prehistoric Swine, see Linnean Soc. Trans., ser. ii., Zool. vol. i. 1877, p. 272.

[30] For reference to the history of this variety of sheep, see British Barrows, p. 740.

[31] As the articles found on the occasion here referred to, as well as the bronze armlet returned by Mr. MacFadzean, have been misplaced, and, in the absence of Sir James Fergusson, could not be found so as to be described and figured in the general notice of Lochspouts, I may state that they were as follows, viz., a hammer-stone, a chisel, two bronze armlets, two pointed implements of deer's horn, a granite quern-stone, several bruising-stones, together with a large quantity of bones.

[32] Regarding the suggestion that these polished stone discs might have been used as mirrors, see Notes by Mr. Joseph Anderson, Proceed. Soc. Antiq. Scot. vol. x. p. 717.

[33] In case K, in the York Museum, which was constructed in 1872-3, to hold the specimens of Roman Metal Work, Implements, and Ornaments of Bone, Jet, etc., are:—Four keys almost identical with that figured above; several articles of bronze of a similar type with the object represented by Fig. 175; a small circular bronze brooch with transverse grooves like Fig. 241; three small bifurcated objects like Fig. 240; harp-shaped fibulÆ, like those from Lochlee; besides many bone pins and combs, jet rings, beads, etc., all of which are wonderfully like the corresponding articles found on the crannogs.

[34] Since writing the above I understand that the natural basin of Lochspouts is about to be converted into a reservoir for supplying the town of Maybole with water, and that, in order to make it suitable for this purpose, according to the engineer's report, it will be necessary to clear away the whole of the lake sediment, including the crannog, at an expense of some £900. As no explorations directed from an archÆological point of view could be more satisfactory than these contemplated operations, we may expect, in the course of their execution, to find not only additional relics that may have dropped into the surrounding lake, but to secure absolute accuracy regarding several doubtful points, such as the dimensions and mode of structure of the island, etc. See Appendix.

[35] In the Museum of the Royal Irish Academy I noticed several knives precisely similar to those here figured, which were found on the crannog of Ballinderry.

[36] J. Romilly Allen, Esq., C.E., F.S.A. Scot., who first recognised this object as part of an ancient padlock, is presently preparing a paper on barbed locks, with special reference to their presence in crannogs, which will be read at an early meeting of the Society of Antiquaries of Scotland. Meantime he writes as follows:—"Padlocks with barbed bolts are almost the only kind used all over China and India at the present day, and it is only reasonable to suppose that they are of Eastern origin. They have been found in England in connection with Roman remains (see paper on Locks found at Great Chesterfield, Essex, by the Hon. R. Cornwallis Neville, ArchÆolog. Journal, vol. xiii. p. 7); and I think it probable they were introduced into this country by the Romans. Their use in this country continued to mediÆval times, after which they disappeared before improved locks of more modern construction."

[37] In Prehistoric Annals of Scotland, vol. i. p. 474, is figured a spiral bronze ring, of three twists, found during the construction of a new road leading from Granton Pier to Edinburgh, in a small stone cist, distant only about twenty yards from the seashore, which has called forth the following remarks from its learned author: "Examples of the spiral finger-ring have been repeatedly found in Britain with remains of different periods. They are also known to northern antiquaries among the older relics of Denmark and Sweden. This may indeed be regarded as among the earliest forms of the ring, since it is only at a comparatively late period that traces of any knowledge of the art of soldering among native metallurgists became apparent. A silver ring of the same early type formed one of the celebrated Norrie's Law hoard, found on the opposite shore of the Firth of Forth." Dr. Schliemann, in giving an account of the discovery of a treasure in a tomb at MycenÆ, writes as follows: "There were further found four spirals of wire, five plain gold rings, and a similar one of silver, of which a selection is represented under No. 529. I remind the reader that similar spirals and rings of thick gold wire occur in the wall paintings of the Egyptian tombs. They are supposed to have served as presents, or perhaps as a medium of exchange."—(MycenÆ and Tiryns, p. 354.) Judging from the paucity of gold spiral finger-rings in our Museums they appear to have been rare. Among the collection of antiquities in the Museum of the Royal Irish Academy, so rich in gold articles, I find only one, thus referred to in Wilde's Catalogue, "a five-sided bar of gold, flat on the inside next the finger, and angular externally, weighing 1 oz. 12 dwt. 6 grs. It may be denominated a torque-ring" (see page 82, fig. 610, Catalogue). In the Belfast collection I also noticed a gold finger-ring with five twists, and having the two ends flattened.

[38] Mr. Dunlop, the finder of the coin, and Mr. John Borland, F.C.S., F.R.M.S., Kilmarnock, analysed this substance, and both pronounced it to be a salt of copper.

[39] Along with a few other relics here exhibited (most of which, I believe, were taken from the Buston crannog) were—the bone pin represented by Fig. 212, a small bronze ring, an iron knife-blade, and a fragment of pottery which was found to fit exactly into that represented by Fig. 251.

[40] Caledonia, Book iii. chap. v. pp. 358 et seq.

[41] Amongst the donations to the Museum of the Society of Antiquaries of Scotland, I find various vessels of brass found in marshy ground near Balgone, East Lothian (also deer's horns and bones of animals.—Proceed. vol. vi. p. 174), by Sir George Grant Suttie, Bart., which are at least suggestive of the former existence of a lake and a crannog in the locality.

These vessels consist of—

A large bronze (brass) tripod pot, with loops at the neck for handle, 13 inches across the mouth, 15¹/₂ inches high, and circumference round the middle 45 inches.

Three other bronze (brass) pots of similar type, varying in size.
A bronze (brass) pot (measuring, etc.).
Shallow bronze (brass) basin, etc.
Portion of a larger bronze (brass) basin.
Bronze (brass) tripod vessel, with spout and looped handles, etc.
Bronze (brass) tripod vessel with straight spout, etc.

The following extract of a letter from Sir G. Grant Suttie gives the details of the discovery, dated 16th Feby. 1849:—"Last autumn my labourers were trenching amongst some rhododendrons in a piece of mossy ground under a peculiar ledge of grey rocks, in my park at Balgone, near my house, and about a mile and a half due south from North Berwick Law, when they found a number of camp-kettles of various sizes, one very large, and in this, one of the goblets was found. They were close to each other, and about 8 feet from the surface. The meadow, extending to about 20 acres, where they were found, was generally under water till imperfectly drained by me; since then the level has sunk from 3 to 4 feet. I have little doubt that when these kettles were deposited here the meadow was a lake, or at all events a morass."—(Proc. Soc. Antiq. Scot. vol. iii. p. 251.)

[42] See Proc. Soc. Antiq. Scot. vol. vi. p. 91.

[43] Lake-Dwellings, by Keller, Second Edition, p. 657.

[44] A Treatise on the Origin, Qualities, and Cultivation of Moss Earth, 1811.

[45] Sketches of Early Scottish History, p. 100.

[46] Portion of the trunk of an oak-tree, recently dug out of the moss at the margin of the lake basin, still lies on the surrounding hedge bank.

[47] Notes of a Tour in the East, 1875.

[48] "It (the comb, Fig. 255) is 2¹/₂ inches long, and 1³/₄ deep, and the three pectinated portions are held together by flat sides, decorated with scrolls and circles. The top or handle shows a triple open-work decoration, and the side-pieces are grooved at one end for receiving the clasp of a metal tooth, which replaced one of the lost bone ones."

Double-edged combs "vary from 3 to 4¹/₂ inches in length, and from 1¹/₄ to 2¹/₄ across, the teeth portions being double, and passing through and through the sides to which they were riveted."

"The crannogs of Dunshaughlin, Ardakillen, and Cloonfinlough, and the street cuttings in the city of Dublin, have afforded nearly all the specimens of which the localities have been recorded. The total number of combs at present in the collection, including those on the 'Find Trays,' is eighty. Many of these combs are but fragmentary; yet in each a sufficiency has been preserved to enable us to judge of the original size, and also of its style of ornamentation, which generally consists of transverse or oblique grooves, diced-work, interlacings, dotted lines, and circles surrounding a central indented spot."—(Wilde's Catalogue.)

[49] "The comb drawn in the sketch (Fig. 259) is an elaborate affair, consisting of central piece or pieces, with teeth; the upper portion being strengthened in front and behind by transverse pieces of ornamented bone, riveted together by now much corroded iron rivets; three of the rivets are reproduced."—(Roman City of Uriconium, p. 86.)

[50] Collect. Antiq. and Journal of Anthrop. Inst. vol. i.

[51] Proceed. of Soc. of Antiq. Scot. vol. x. p. 476; xi. p. 305; and xii. p. 669.

[52] The Past in the Present, by Arthur Mitchell, LL.D.

[53] Quarterly Journal of the Geological Society, vol. xii. p. 255, May 7, 1856.

[54] Lake-Dwellings, by Keller (2d ed., trans. by Lee), p. 660.

[55] See article on Inscribed Stones at Kirkmadrine, in the parish of Stonykirk, county of Wigtown, by Dr. Arthur Mitchell.—Proc. Soc. Antiq. Scot. vol. ix. p. 568. Also, Scotland in Early Christian Times (Second Series), p. 252.