2. The whole mass was kept firmly together by a peculiar arrangement of upright and horizontal beams, forming a united series of circular stockades.

3. The outer circle was intended more for protection than for giving stability to the island, and in some parts, as at the east side of refuse-heap, was neatly constructed after the manner of a stair-railing, while the inner one not only gave stability to the island, but was used as a fence, or in connection with the superstructural buildings.

4. The central portion was rudely paved with wooden beams, many of which were firmly fixed to the lower wood-work by stout wooden pegs as well as to the encircling stockades, thus affording here and there, as it were, points d'appui.

5. While there was one general fireplace situated near the centre, evidence of occasional fires elsewhere was quite conclusive, one of which appeared to have been a smelting-furnace.

6. The entrance to the central area was looking south-east, and in front of it there was a well-constructed wooden platform, made of large oak planks, supported on solid layers of wood to which they were pinned down.

7. Beyond the platform, but separated from it by a massive wooden railing, was the refuse-heap; and to the right of it a pathway, also protected on its outer side by a railing, led downwards and westwards to the line of the outer circle, where there appeared to have been an opening towards a rude landing-stage at the water edge.

8. As to the kind of dwelling-house that no doubt once occupied this site, whether one large pagoda-like building or a series of small huts, the evidence is inconclusive, but so far as it goes it appears to me to be indicative of the former. In addition to what has already been stated, there remains to notice only a few broken pieces of wood containing round holes, together with a variety of large and small pins similar to those described and figured in my notice of the Lochlee crannog.

Fig. 191.—Appearance of Canoe in situ, after exposure. (From a Drawing by Mrs. Anstruther.)

Discovery and Description of Canoe.—The experience derived from the investigations of the crannogs at Lochlee and Lochspouts, in both of which a submerged gangway was found running to the nearest shore, was sufficiently suggestive to keep me on the qui vive for any indications of a similar structure here. On the north side, where the shore was nearest, though the digging was carried considerably deeper and farther out from the margin of the crannog than elsewhere, not the slightest appearance of outlying wood-work was observed; and as there was no probability of an approach from the more distant ends of the lake, the situation of a gangway, if such existed at all, was limited to the south-west side, where the shore would be about 150 yards distant. To determine this, the men were set to cut a trench about 12 yards distant from the crannog, across the most likely line, so as to intercept it, and after going down 4 feet they came upon a layer of brushwood, along with one or two beams, below which there seemed to be the usual lake mud. Upon forcing the spade downwards, however, a hard beam was encountered, which at first I took to be the discovery of part of the gangway we were in search of, and to satisfy myself on the point I took an iron rod, and, by carefully probing all over the bottom of the trench, ascertained that instead of a gangway we had come upon portion of a canoe. Guided by the direction of the supposed side of the canoe, which looked like a thin oak beam running along the edge of the trench, a suitable clearance was made, which revealed to the wondering gaze of the bystanders the front half of a large canoe. Upon being subsequently exposed in its entirety it was found to have the following dimensions and peculiarities. Its sides were supported by a series of well-shaped ribs, which extended from the rim to near the mesial line, and sometimes a little beyond it. This, at first sight, gave the canoe the appearance of a boat, but after careful inspection it became apparent that these ribs were no part of the original vessel, but subsequent additions made for repairing and strengthening purposes. Nearly the whole of one side was lined with broad thin boards made of soft wood, external to which was the thin oak side of the canoe, having its cracks as well as the intermediate spaces between it and the strengthening boards actually stuffed with a species of moss. Moreover, the ribs on this side were more numerous than on the other side, no less than ten having been observed on the former, and their arrangement on both sides was totally devoid of regularity. Of the whole series of ribs only two were made of oak, the rest being of birch or some perishable wood, and so decayed that it was with great care they were prevented from being entirely destroyed by the workmen, as they offered no resistance to their spades. They were fastened to the canoe by wooden pins, arranged generally in couples forming two rows along the rib, and so closely were they placed that not less than sixteen were counted in one rib. In two places the canoe had been repaired by inserting a nicely fitting piece of oak planking instead of the original portion of the side. One of these patches measured 2 feet 3 inches long, and 10 inches broad, and was kept in position by two ribs, one near each end. The stem, which was symmetrically shaped and pointed, was pierced horizontally by a large hole, and about 3 feet from its tip each side had also an elongated hole near the rim, sufficiently large to admit of being easily grasped by the hand. Externally, and on both sides, there was fastened to the edge of the canoe, by means of wooden pins, a sort of gunwale, which extended from within a few feet of the stem till it projected a little beyond the stern. Close to the stern, two slender bars of wood, a few inches apart, stretched across, and after passing through the edges of the canoe terminated in being tightly mortised into the gunwale. These transverses contained two round holes similarly arranged as to position, being near the right side, and between them was inserted a moveable sternpiece which was shaped to the curve of the canoe, i.e. approximately a semicircle, and made to fit into a shallow groove cut out of the solid wood. This sternpiece was strongly constructed, being 312 inches thick, 3 feet 6 inches long, and 1 foot 412 inches deep about the middle. About 15 inches in advance of the sternpiece there was a ridge across the bottom and sides of the canoe which looked like a rib, but was really part of the solid oak, evidently left for a special purpose. I also noticed one or two round holes in the floor, as well as others along its upper edge as if for thole-pins. In two places equidistant from the ends, and about 4 feet apart, the gunwale had short pieces of wood fastened to it by vertical pins, as if intended for the use of oars. Amongst the decayed brushwood lying across the canoe was an oak beam, having one end projecting so much beyond the edge into the clay bank that the workmen in endeavouring to pull it out broke off the free end. This portion was rectangularly shaped, 5 inches by 312, and had its narrow side pierced with three round holes 1 foot 10 inches apart, which still contained the remains of broken pins. The shell of the canoe was oak, made by scooping out the interior of a large trunk, but all its attachments, such as gunwale, sternpiece, cross spars at stern, and all the ribs except two, were made of a much less durable wood.