Chapter XVII: Part 17

The site of the crannog at Friar's Carse was a small pear-shaped basin situated behind a wooded knoll, close to the Parliamentary road to Dumfries, and in the midst of a well-cultivated but singularly undulating district. By deepening the outlet of this lake to the extent of two feet, a partial drainage was effected, which reduced its area from 10 to 3 acres. It was only then (1878) that it became generally known that a small bushy island near the middle of the loch had been artificially constructed of oak-planks and trunks of trees. As the weather was dry for some weeks previous to our visit, and the water particularly low, we readily stepped on to the island, over what appeared to have been the old bed of the lake, then presenting a hard, crisp, and dried-up surface of aquatic plants. The island was nearly circular in shape, measuring 80 by 70 feet, strongly built, and surrounded by piles, some of which, however, were only visible through the water. The log-pavement, which by this time had been completely bared, was composed of parallel beams of oak, arranged in groups, lying in various directions, and firmly united together by the overlapping and sometimes mortising of their ends. At the margin of the island there was a large quantity of stones, especially on its north side--_i.e._ the side towards the deepest portion of the lake.

Through these stones, which shelved under the water, a few heads of the surrounding piles projected, some above and some below the water. Mortised holes were here and there to be seen in the horizontal beams, but there was no appearance of a breastwork surrounding the wooden pavement--thus differing from the Lochlee crannog. In the centre were a few ends of uprights, in rectangular rows, seemingly the remains of partitions, one of which I traced for 40 feet in a straight line.

Upon inquiring where the rubbish removed from the island was located, we were informed that it had been wheeled to the west side of the crannog, and heaped up close to where we had stepped on to the island. Here it lay for some days; but one morning, to the great astonishment of the workmen, it was nowhere to be seen. Upon examination, it turned out that the apparently dry bed of the lake was a matted crust of mud and the roots of aquatic plants, which, virtually floating over the water, suddenly gave way under the accumulated weight and so the entire mass of the crannog rubbish disappeared in the water beneath. With this singular, but unfortunate, catastrophe terminated all prospects of finding any more relics. It appears that there was not a great depth of _débris_ on the island, its maximum thickness being only 2 to 3 feet in the centre, where it formed a heap of ashes, charcoal, and some broken bones. Here a few fragments of pottery were found.

A circular portion of the log-pavement, near its centre, was covered with small stones, as if to protect it from fire; some remains of clay-flooring were observed in other parts of the island.

Regarding the deeper structures little can be said. Mr. Nelson attempted to cut a hole through the timber, and, as far as the water allowed the men to penetrate, he saw nothing but layer upon layer of oak-beams lying transversely to each other. Judging, however, from the solidity and firmness of the island, the great size of some of the logs, and the depth of the surrounding water (still about twelve feet a little to the west of the island), the total thickness of this mass of timber cannot be less than 12 or 16 feet.

In Grose's "Antiquities of Scotland"[101] the following reference to this island occurs:--

"Here was a cell dependent on the rich abbey of Melrose, which, at the Reformation, was granted by the Commendator to the Laird of Elliesland, a cadet of the Kirkpatricks of Closeburne. From whom it passed to the Maxwells of Tinwald, and from them to the Barncleugh family, also cadets of the Lords of Maxwell. From whom it went to the Riddells, of Glenriddell, the present possessors. The old refectory, or dining-room, had walls 8 feet thick, and the chimney was 12 feet wide. This old building having become ruinous, was pulled down in 1773, to make way for the present house.

"Near the house was the Lough, which was the fishpond of the friary. In the middle of which is a very curious artificial island, founded upon large piles and planks of oak, where the monks lodged their valuable effects when the English made an inroad into Strathnith."

The relics collected during the operations above recorded are very few. A canoe 22 feet long, and a ponderous axe-hammer head of whinstone (=Fig. 170=) were found at some distance from the crannog. Two handles of jars with traces of a yellowish glaze, some fragments of pottery ornamented with rows of pitted impressions (=Fig. 171=), a circular stone polisher, and an oval-shaped mass of vitreous paste, are all that were found on the crannog itself.

STONE LAKE-DWELLINGS AND OTHER ARTIFICIAL ISLANDS.

But besides the wooden islands there are others, still extant in several of our Scottish lakes, which appear to be composed entirely of stones and earth irregularly heaped together. In the absence of any historical knowledge as to their age there is no _prima facie_ reason why some of these should not be contemporary with the former, as it cannot be assumed that the crannog-builders made wood a _sine quâ non_ in the structure of islands. There were, no doubt, certain stagnant marshes and small lochs in which a wooden foundation was essential for the construction of an island, owing to the softness and yielding nature of the mud; but, on the other hand, there were others with compact rocky or gravelly beds, in which any solid materials, as stones, earth, turf, etc., would be equally applicable. The outlets of the larger lakes, more especially such as were formed in glacial and rock-cut basins, were more adapted for the latter, and as far as my observations have enabled me to form an opinion, these are the very situations in which the lake stone-dwellings abound. Some of them are mere shapeless cairns, without any indications of having been formerly inhabited, while on others some remains of stone buildings are to be seen. As to wooden huts or houses, had such structures ever been erected on them, it is not likely that they could, for any length of time, have resisted the decaying tendencies of atmospheric agencies, so that all traces of them would have disappeared long ago.

The social or military exigencies that led people to construct artificial islands would also lead them to take advantage of such natural ones as would be found most suitable, and we may reasonably infer that it is in the absence of the latter that the former would be resorted to. The great and primary object of the island-builder was the protection afforded by the surrounding lake or morass, the securing of which has continued to be the ruling principle in the erection of defensive works down to the Middle Ages, long after the wooden islands ceased to be constructed. The transition from the crannog to the massive mediæval castle, with its moat and drawbridge, is but a stage in the progressive march of civilisation.

GEOGRAPHICAL DISTRIBUTION OF SCOTTISH LAKE-DWELLINGS.

To these remarks on the structural details of a few typical crannogs, as disclosed by systematic research, I subjoin a tabulated list of all the sites hitherto recognised in Scotland, comprising not only the artificial islands, whether of wood or other materials, but also some natural ones known to have been artificially strengthened, as well as a few examples of castles, etc., now or formerly located in bogs or drained marshes.

LIST OF SCOTTISH CRANNOGS ALPHABETICALLY ARRANGED,
WITH NOTES AND REFERENCES.

N.B.--An obelisk (†) before a name in this text indicates that the
island is, in whole or in part, constructed of wood. N.S.A. or O.S.A.
stand for New or Old Statistical Account of Scotland.

†Achilty L., co. Ross. N. S. A., vol. xiv. p. 238.
Achray L., co. Perth. B. 94, pp. 172-7.
†Airrieoulland, co. Wigtown. B. 426.
Ard L., co. Perth. O. S. A., vol. x. p. 130.
[102]†Arisaig L., co. Inverness. B. 150, p. 576.
†Banchory (L. of the Leys), co. Aberdeen.
_Proc. S. A. Scot._, vol. i. p. 26; vol. vi. p. 126.
†Barean L., co. Kirkcudbright. B. 373, p. 37, and
_Dumf. and G. N. H. Soc._, 1865.
†Barhapple L., co. Wigtown. "Ayr and Wig. Col.,"
vols. iii. and v.; B. 373; B. 192.
[103]†Barlockhart L., co. Wigtown. _Proc. S. A. Scot._,
vol. xi. p. 583; vol. xv. p. 267.
†Barnsallzie L., co. Wigtown. _Ibid._, vol. ix. p. 377.
Battleknowes, co. Berwick. N. S. A., vol. ii. p. 171.
[104]†Black Cairn, Beauly Firth, co. Ross. "Hill Forts and
Stone Circles of Scotland," p. 89; N. S. A.,
vol. xvii. p. 350.
†Boghall (Beith), co. Ayr. N. S. A., vol. v. p. 580.
Borgue, co. Kirkcudbright. N. S. A., vol. iv. p. 54.
Brora L., co. Sutherland. O. S. A., vol. x. p. 303;
N. S. A., vol. xv. p. 151.
[105]†Bruich L. (Beauly), co. Ross. B. 442.
†Buston, co. Ayr. "Ayr and Wig. Col.," vol. iii.; B. 373.
†Canmor (Kinord) L., co. Aberdeen. B. 94, pp. 167-71.
[106]†Carlingwark L., two crannogs, co. Kirkcudbright. O. S. A.,
vol. viii. p. 304; B. 94, p. 126; _Proc. S. A. Scot._,
vol. vii. p. 7, and x. p. 286.
†Castle Loch, co. Wigtown. Rev. G. Wilson's "Notes."
Castletown, co. Roxburgh. N. S. A., vol. iii. p. 164.

Closeburn, co. Dumfries. _Phil. Trans._, 1756,
p. 521; Grose, "Ant. of Scot.," vol. i. p. 150.
Clunie L., co. Perth. O. S. A., vol. ix. p. 231.
Collessie, co. Fife. O. S. A., vol. ii. p. 418.
†Corncockle (Applegarth), co. Dumfries. B. 94, p. 163.
†Cot L., co. Linlithgow. _Ibid._, p. 159.
[107]†Croy, co. Inverness. N. S. A., vol. xiv. p. 448.
†Dhu Loch, co. Bute. B. 21, p. 43.
Dolay L., co. Sutherland. B. 94, pp. 172-7.
Doon L., co. Ayr. N. S. A., vol. v. p. 337.
†Dowalton, five crannogs, co. Wigtown. B. 56, 94, 373, and 426.
Earn L., co. Perth. O. S. A., vol. xi. p. 180.
Eldrig L., three crannogs, co. Wigtown.
Rev. G. Wilson's "Notes."
[108]†Eriska, co. Argyll. B. 427, p. 192.
Fasnacloich (Appin), co. Argyll. B. 94, p. 175.
Federatt, co. Aberdeen. O. S. A., vol. ix. p. 191.
Fell L., co. Wigtown. B. 192, vol. ix. p. 378.
[109]Fergus L., co. Kirkcudbright. O. S. A., vol. xi. p. 25.
†Flemington, L., co. Nairn. B. 55, p. 118.
Forfar, Loch of, co. Forfar. B. 1; O. S. A.,
vol. vi. p. 528; B. 94, p. 125; B. 216, p. 31.
Freuchie L., co. Perth. B. 94, p. 173.
†Friar's Carse, co. Dumfries. B. 373, p. 152, and B. 374, p. 73.
Fullah L., co. Perth. B. 94, p. 172.
Glass L., co. Ross. O. S. A., vol. i. p. 282.
Granech L., co. Perth. B. 94, p. 177.
†Green Knowe, co. Lanark. N. S. A., vol. vi. p. 346;
_Proc. S. A. S._, vol. vi. p. 160, and vol. viii. p. 19.
Gynag L., co. Inverness. N. S. A., vol. xiv. p. 65.
Heron L., two islands, co. Wigtown. B. 192, vol. ix. p. 378.
Hogsetter L., Shetland. _Proc. S. A. Scot._ vol. xv. p. 303.
[110]†Kielziebar L., co. Argyll. B. 134, pp. 332 and 516.
†Kilbirnie L., co. Ayr. B. 268, p. 284.
Kilchonan, co. Argyll. O. S. A., vol. xi. p. 281.
[111]†Kinder L., co. Kirkcudbright. _Old. S. A._, vol. ii. p. 139.
†Kinellan L., co. Ross. N. S. A., vol. xiv. p. 238; B. 94, p. 126.
Laggan L., co. Perth. O. S. A., vol. xviii. p. 327.
[112]†Ledaig, co. Argyll. B. 190.
[113]†Leven L., co. Kinross. B. 460, p. 118.
†Loch-of-the-Clans, two crannogs, co. Nairn. B. 55,
pp. 116 and 332.
†Loch-in-Dunty, co. Nairn. B. 55, p. 118.
[114]†Loch-inch-Cryndil, co. Wigtown. B. 212, pp. 381 and 388.

†Lochindorb, co. Moray. O. S. A., vol. vii. p. 259.
†Lochlee, co. Ayr. "Ayr and Wig. Col.," vol. ii.; B. 331 and 373.
†Lochmaben, co. Dumfries. B. 94, p. 160; _Arch. Scot._,
vol. iii. p. 77.
†Loch-na-Mial, Island of Mull. B. 172, p. 465.
†Lochnell, co. Argyll. B. 190, vol. ix. p. 105.
†Lochore, co. Fife. B. 94, p. 160.
†Lochrutton, co. Kirkcudbright. O. S. A., vol. ii. p. 37.
†Lochspouts, co. Ayr. "Ayr and Wig. Col.," vol. iii. p. 18;
iv. p. 9; B. 373, pp. 158 and 305.
Lochwood, co. Dumfries. O. S. A., vol. iv. p. 224.
[115]†Lochy L., co. Inverness. B. 94, p. 160.
[116]†Lomond L., co. Sterling. _Ibid._, p. 131.
†Lotus L., co. Kirkcudbright. _Proc. S. A. Scot._,
vol. xi. p. 21.
[117] Machermore L., several crannogs, co. Wigtown. B. 192.
†Merton L., co. Wigtown. B. 94, p. 123.
Mochrum L., co. Wigtown. B. 192.
Monivaird L., co. Perth. O. S. A., vol. viii. p. 570.
Morall L., co. Perth. B. 94, p. 176.

Morton, co. Dumfries. N. S. A., vol. iv. p. 96.
Moulin L., drained, co. Perth. O. S. A., vol. v. p. 69.
Mountblairy, co. Moray. O. S. A., vol. iv. p. 399.
†Moy L., Ellan-na-Glack, co. Inverness. N. S. A.,
vol. xiv. p. 100; B. 94, p. 129.
[118]†Oban (Lochavoullin), co. Argyll.
Orr L., co. Dumfries. O. S. A., vol. ii. p. 342.
Peel Bog, co. Aberdeen. N. S. A., vol. xii. p. 1089.
†Quien Loch, co. Bute. B. 21, p. 45.
†Rannoch L., co. Perth. N. S. A., vol. x. p. 539; B. 94, p. 129.
†Ravenstone L., co. Wigtown. B. 426, p. 121.
Rescobie L., co. Forfar. B. 94, p. 176.
†Rothiemurchus, Loch-an-Eilan, co. Moray. N. S. A.,
vol. xiii. p. 137; B. 94, p. 145.

[119]†Sanquhar, Black Loch of, co. Dumfries. _Proc. Dumf. and
Gal. N. H. Soc._, 1863-4, p. 12, and B. 373, p. 36.
Shin L., co. Sutherland. B. 94, pp. 172-7.
†Spinie L., co. Moray. O. S. A., vol. x. p. 625.
Stravithy, co. Fife. N. S. A., vol. ix. p. 365.
†Sunonness L., co. Wigtown. B. 192, p. 738.
Tay L., co. Perth. B. 94, p. 173; O. S. A.,
vol. xvii. p. 465; N. S. A., vol. x. p. 465.
†Tolsta, Lewis, co. Ross. _Proc. S. A. S._, vol. x. p. 741.
Torlundie, drained loch at, co. Inverness.
_Proc. S. A. Scot._, vol. vii. p. 519.
†Tullah L., co. Perth. B. 94, p. 172.
Tummell L., co. Perth. O. S. A., vol. ii. p. 475; B. 94, p. 129.
Urr L., co. Dumfries. B. 94, p. 160.
Vennachar L., co. Dumfries. B. 94, p. 177.
Weyoch L., co. Wigtown. B. 192.
Yetholm L., co. Rosburgh. N. S. A., vol. iii. p. 164.

CRITICAL EXAMINATION OF RELICS.

The great value, however, of the investigations of the lake-dwellings, especially in the south-west of Scotland, depends on the quantity and variety of the remains of human industry discovered in and around their sites. It is from such fragmentary remains as food refuse, stray ornaments, broken weapons, useless and worn-out implements, and such-like waifs and strays of human occupancy, that archæologists attempt to reconstruct the outlines of the social life and organisation of the prehistoric past. To those who may wish to occupy themselves with this problem these explorations have furnished, as we have just seen, a vast collection of objects made of stone, bone, horn, wood, bronze, iron, and gold.

Among the stone objects are--querns, hammer-stones, whetstones, so-called sling-stones, a few cup-marked stones (one surrounded by concentric circles), spindle-whorls, flint flakes, and scrapers, a polished celt, a perforated axe-hammer head, portions of two polished circular discs, and some oval implements with a wrought hollowed surface on each side.

Bones and horns of deer were utilised in various ways and manufactured into pins, needles, bodkins, awls, picks, toilet-combs, knife-handles, etc. The combs are neatly formed of three or four flat pieces kept in position by two transverse slips, one on each side, and riveted together by iron rivets. They are frequently ornamented by a series of incised circles, which are sometimes connected by a running scroll, as in =Fig. 174=.

The wooden articles consist of bowls, ladles, mallets, hoes, clubs, etc., together with a variety of other objects apparently intended for agricultural purposes.

Implements and weapons of iron are numerous. Amongst the former are gouges, chisels, knives, shears, saws, hatchets, awls, hammers, a bridle-bit, the bolt of a padlock, and other objects of unknown use. The weapons consist of leaf-shaped spear-heads, both socketed and tanged, daggers, and arrow-heads resembling those of the crossbow bolt.

The objects made of bronze are mostly of an ornamental character, comprising:--harp-shaped fibulæ, circular and penannular brooches, finger-rings, a spiral ornament, ornamented pins, one with a ring top and another with a glass setting, a small key, and some other articles of an indeterminate character. From Dowalton there are basins or cauldrons of beaten bronze, some clouted and riveted; one, presumably a Roman saucepan, has the name of the maker on the handle.

On the Buston crannog were found two handsome and massive spiral finger-rings made of gold. One is plain with five and a half twists; the other, besides an additional twist, has both ends ornamented by a series of circular grooves. From the same place there is a curious gold coin, of Saxon origin, and a forgery of the sixth or seventh century.

Pottery is represented by numerous fragments, some of which are of so-called Samian ware, but the most of them are of vessels of a glazed ware, while a few are of an archaic type. Several neatly formed crucibles, containing traces of gold and slag, are also in the collection.

Among miscellaneous objects are bracelets and beads made of coloured and of variegated glass or vitreous paste; also some jet ornaments, one of which is a handsome pendant in the form of an equal-armed cross, inscribed in a circle and having one surface ornamented by a series of incised circles which contained the remains of a yellow enamel. Dr. Joseph Anderson considers this a Christian relic of a very early type. A smooth and flat piece of ashwood, with peculiar spiral carvings on both sides, and a fringe-like apparatus made of the long stems of a moss, are among the objects which have excited the greatest curiosity. Regarding a finely polished conical object made of rock-crystal found at Lochspouts, a reviewer in the _Academy_, October 14th, writes:--"Is it a charm or can it have formed the centre knob or boss in the binding of some richly decorated breviary or gospel book? Crystals very similar, but oblong in form--like a Brazil nut--may be seen in some of the rich covers of books of early date, and a few that have been detached are preserved in collections. One such object forms part of a crystal necklace in the Ashmolean Museum, and another in private hands was employed, not so very many years ago, in the West Riding of Yorkshire, for the purpose of seeing spirits. If this relic be, indeed, a book-boss, it makes it probable that the crannog was at one time inhabited, or at least visited, by Christian missionaries." Dr. Joseph Anderson has also pointed out that this object is extremely like a "large circular rock crystal which forms the central ornament on the inferior surface of the foot of the famous silver chalice, dug up at the Rath of Reerasta, near Ardagh, county Limerick, Ireland, in 1868, and now in the Museum of the Royal Irish Academy, Dublin. According to the Earl of Dunraven, this most beautiful example of our ancient art was executed either in the ninth or tenth century." (See _Proc. Soc. Antiq. Scot._, December 4th, 1882.)

From the respective reports of Professors Owen, Rolleston, and Cleland, on a selection of osseous remains taken from the lake-dwellings at Dowalton, Lochlee, and Buston, we can form a fair idea of the food of the occupiers. The Celtic shorthorn ox, the so-called goat-horned sheep, and a domestic breed of pigs were largely consumed. The horse was only scantily used. The number of bones and horns of the red-deer and roebuck showed that venison was by no means a rare addition to the list of their dietary. Among birds, only the goose has been identified, but this is no criterion of the extent of their encroachment on the feathered tribe, as only the larger bones were collected and reported on. To this bill of fare the occupiers of Lochspouts crannog, being comparatively near the sea, added several kinds of shell-fish. In all the lake-dwellings that have come under my own observation the broken shells of hazel-nuts were in profuse abundance.

From the number of querns, and the great preponderance of the bones of domestic over those of wild animals, it may be inferred that, for subsistence, they depended more on the cultivation of the soil and the rearing of cattle, sheep, and pigs, than on the ordinary produce of the chase.

Proofs of a prolonged but occasionally interrupted occupancy are also manifested by the great accumulation of _débris_ over the wooden pavements, the size and contents of the kitchen-middens, and the superimposed hearths.

Let us now look at the remarkable series of implements, weapons, ornaments, and nondescript objects here presented to us, with the view of abstracting from them some scraps of information regarding their original owners. The fragments of Samian ware, bronze dishes, harp-shaped fibulæ, and the large assortment of beads, bronze and bone pins, bone combs, jet ornaments, etc., are so similar to the class of remains found on the excavated sites of Romano-British towns, that there can hardly be any doubt that Roman civilisation had come in contact with the lake-dwellers. The Celtic element is, however, strongly developed, not only in the general character of many of the industrial implements of stone, bone, and iron, but also in the style of art manifested in some of the ornamental objects included in the collection. Thus the piece of ashwood with its carved spiral patterns (=Figs. 144= and =145=), the combs, especially the one showing a series of concentric circles connected by a running scroll design (=Fig. 174=), and some of the bronze brooches and ornaments (=Fig. 130=) present a style of ornamentation which is considered peculiar to Celtic art. The spiral finger-rings seem also to have been of native origin, and the probability is that they were manufactured where they were found, as several crucibles are amongst the relics from the same lake-dwelling, one of which, from the fact that it still contains particles of gold, proves that it had been used in melting this metal. (B. 373, p. 236.)

On the other hand, the forged gold coin is the only relic that can with certainty be said to have emanated from a Saxon source--at least, that cannot otherwise be accounted for.

But if from internal evidence a presumptive case is made out in favour of the Celtic origin and occupation of these lake-dwellings, it is greatly strengthened when we consider that the neighbouring Celtic races, especially in Ireland, were in the habit of erecting similar island abodes, while there is not a particle of evidence in favour of the idea that such structures originated with the Roman conquerors of Britain or its Saxon invaders.

Comb from the Knowe of Saverough, Orkney (½).

Two Combs from the Broch of Burrian, Orkney (½).

=Fig. 176.=--Bone Combs, for comparison with those from the Lake-Dwellings.]

The resemblance between the remains found in the Scottish and Irish lake-dwellings, as well as other antiquarian finds of Celtic character, must also not be overlooked. Combs, similar in structure and ornamentation to those from Buston, have been found in several of the Irish crannogs, in the brochs and other antiquities of the north of Scotland, and in many of the ruins of the Romano-British towns in England. (See =Figs. 105=, =108=, and =176=.) Iron knives and shears, variegated beads of impure glass with grooves and spiral marks, ornaments of jet and bronze, implements of stone, bone, and horn, besides querns, whetstones, etc., are all common to Celtic antiquities, wherever found.

That many of these relics were the products of a refined civilisation is not more remarkable than the unexpected and strangely discordant circumstances in which they have been found. For this reason it might be supposed that the crannogs were the headquarters of thieves and robbers, where the proceeds of their marauding excursions among the surrounding Roman provincials were stored up. The inferences derived from a careful consideration of all the facts do not appear to me to support this view, nor do they uphold another view, sometimes propounded, viz. that they were fortified islands occupied by the guardian soldiers of the people. Indeed, amongst the relics military remains are only feebly represented by a few iron daggers and spear-heads, one or two doubtful arrow-points, and a quantity of round pebbles and so-called sling-stones. On the other hand, a very large percentage of the articles consists of querns, implements and tools, crucibles, various domestic utensils, etc., from which, not to mention the great variety of ornaments, there can be no ambiguity as to the testimony they afford of the peaceful prosecution of various arts and industries by the lake-dwellers.

There is, in my opinion, only one hypothesis that can satisfactorily account for all the facts and phenomena here adduced, viz. that the lake-dwellings in the south-west of Scotland were resorted to by the Celtic inhabitants as a means of protecting their lives and movable property when, upon the frequent withdrawal of the Roman soldiers from the district, they were left, single-handed, to contend against the Angles on the east and the Picts and Scots on the north. It is not likely that these provincials, so long accustomed to the luxury and comforts of Roman civilisation, or their descendants in the subsequent kingdom of Strathclyde, would become the assailants of such fierce and lawless enemies, from whom, even if conquered, they could derive no benefit. Hence their military tactics and operations would assume more the character of defence than aggression, and in order to defeat the object of the frequent and sudden inroads of the northern tribes, which was to plunder the inhabitants rather than to conquer the country, experience taught them the necessity of being prepared for emergencies by having certain places of more than ordinary security where they could deposit their wealth, or to which they could retire as a last resource when hard pressed. These retreats might be caves, fortified camps, or inaccessible islands, but in localities where no such natural strongholds existed the military genius of the Celtic inhabitants, prompted perhaps by inherited notions, led them to construct these wooden islands. From the final departure of the Romans to the conquest of the kingdom of Strathclyde by the Northumbrian Angles, a period of several centuries, this unfortunate people had few intervals of peace, and with their complete subjugation ended the special functions of the lake-dwellings as a national system of protection. No doubt some of them, as well as caves and such hiding-places, would continue to afford refuge to straggling remnants of natives, rendered desperate by the relentless persecution of their enemies; but ultimately all of them would fall into the hands of their Saxon conquerors, when henceforth they would be allowed to subside into mud or crumble into decay.

III.--ENGLISH LAKE-DWELLINGS.

The discovery of lacustrine abodes south of the Scottish border, though the examples are by no means so numerous or so prolific in industrial remains as those of Scotland and Ireland, is, nevertheless, of special interest on account of the intermediary position in which England stands geographically to the areas of their earliest and latest development in Europe. It will be noticed that some of the recorded observations here reproduced were actually made before antiquaries realised the importance of the subject; otherwise it is impossible to conceive how such highly suggestive facts did not at once lead to more definite information.

THE MERES OF NORFOLK AND SUFFOLK.

WRETHAM MERE.--Sir Charles F. Bunbury, as early as 1856, noticed some appearances in a drained _mere_ near Wretham Hall which clearly point to being the remains of a lake-dwelling. In a communication on the subject to the Geological Society he says:--

"Wretham Hall, the seat of W. Birch, Esq., is situated about six miles north of Thetford, in that extensive tract of open sandy plains which may be called upland in comparison with the fens, but of very moderate elevation above the sea-level as is shown by the slow course of the streams flowing from it. About Wretham there are several _meres_ or small natural sheets of water without any outlet. The one to which my attention was particularly directed by Mr. Birch occupied about forty-eight acres, and was situated in a slight natural depression, the ground sloping gently to it from all sides. The water has been drawn off by machinery, for the purpose of making use, as manure, of the black peaty mud which formed the bottom. This black mud, which is in parts above twenty feet deep, is nothing else than a soft, rotten, unconsolidated peat; or perhaps it should be described as vegetable matter in a more complete state of decomposition than ordinary peat, showing no distinct trace of vegetable structure. Numerous horns of red deer have been found in this peaty mud, generally (as I was informed) at 5 or 6 feet below the surface, seldom deeper; many attached to the skull, others separate, and with the appearance of having been shed naturally. What is most remarkable, several of those which were found with the skulls attached had been _sawn off_ just above the brow antlers--not broken, but cut off clean and smoothly, evidently by human agency. Some of the horns are of large size, measuring 9 inches round immediately below the brow antler....

"Numerous posts of oak-wood, shaped and pointed by human art, were found standing erect, entirely buried in the peat."

It appears that in 1851 a more remarkable "find" became visible on draining another _mere_ on this same estate, though the events remained unrecorded till the years 1858 and 1862. The following notice is compiled mainly and almost verbatim from Mr. Newton's observations, which he states were directly obtained from Mr. Birch, the proprietor:--

In this mere (West Mere) there was ordinarily about four feet of water, and beneath it, about eight feet of soft black mud, partly held in suspension and requiring to be removed in scoops. Near the centre of the _mere_, lying below the black mud, was found a ring or circular bank of fine white earth, sufficiently solid to allow Mr. Birch to ride upon it without yielding to the weight of his pony. Outside this ring the bottom of the _mere_ was so soft and deep as to be almost impassable until the mud was cleared away. The ring was some twenty or thirty feet across, a foot wide at the top, and about four feet in height. Not far from its inner circumference was a circular hole, about four feet and a half in diameter and some six feet deeper than the bottom of the _mere_. It was marked out by a circle of stout stakes or small piles, apparently of alder, and it bore traces of having been wattled. Between these two circles were the remains of a wall, about two feet high and consequently lower than the top of the ring, composed of flints packed together with marl or soft chalk. In the same place was some earth of a bright blue colour, which, when dried, crumbled to powder, and was not preserved, though traces were still to be seen on the bones. In this interspace a still greater number of bones was found, and also the remains of a much decayed ladder, the sides and rounds of which were 15 inches apart. The stakes were about four inches in diameter, very hard, as heavy as stone, and of a dark grey colour. The fragments of the ladder, on the contrary, were very rotten and light, but the remains of both, after being kept some time, exfoliated and crumbled entirely to dust. In and around this ring there lay a vast number of bones, of which no small portion were the upper parts of the skulls of _Bos longifrons_, with the horn cores attached, and many antlers of the red deer, either entire or in fragments. All the former, excepting one unusually large example, had a fracture the size of half-a-crown in the forehead (Babington). Of the deer's antlers, some have certainly been shed in the due course of nature; but others, on the contrary, have been separated from the head by sawing. Of the other bones found in West Mere, and I am told there were hundreds of them, most of the larger ones have been fractured at one or other extremity, doubtless in order to extract the marrow they contained. Another bone, and, as far as I can make out, the only one found which presents this peculiarity, has been polished on one side; but the reason why is not very obvious, unless it has served, as I before suggested in the case of a similar specimen, for a skate. I must add that no weapons or implements of metal which can be referred to a period at all remote were brought to light in this or any of the adjoining _meres_, but a great number of flint discs were found, which, according to the description I have received (for unfortunately none of them seem to have been preserved), must have closely resembled those known to the Danish antiquaries as "Sling-Stones," from the probable use made of them. (B. 46, p. 17.)

BARTON MERE.--In 1869 the Rev. Harry Jones communicated a paper to the Suffolk Institute of Archæology and Natural History "on the discovery of some supposed vestiges of a pile-dwelling in Barton Mere, near Bury St. Edmund's," of which the following is an abstract:--

Barton Mere is situated in a natural depression, about four miles east of Bury St. Edmund's, and is mainly supplied by springs, but at some seasons water flows into it from the high land on the south, west, and north. When full it consists of about ten acres, and averages 7 feet in depth. On the north side of the _mere_ there is a marly chalk, which, indeed, forms the main bottom of the _mere_, being overlaid with a dark clay deposit from 1 to 5 feet deep. The bottom layer of this deposit consists of a peaty coloured clay, so tenacious as to keep its shape upon the potter's wheel. Most of the bones and some fragments of pottery were found in this lower layer, which varies in thickness from a few inches to about a foot and a half. The _mere_ is subject to occasional droughts. It has been dry at least four times in the last forty years. About thirty-eight years ago (1830), the _mere_ being then dry, his grandfather, Mr. Quayle, who lived at Barton Mere, dug out a quantity of stuff for the purpose of laying it on the land. His digging resulted in a hole, which on two succeeding occasions when the water was low, saved enough to keep some of the fish alive, and provide a pond for the cattle. Bones and horns of deer, and several spear-heads and rings of bronze, were reported to have been found amidst six or seven stakes of wood _sticking up out of the bottom and about as thick as the thin part of a man's leg_.

The excavations conducted by Mr. Jones in 1867 were made by digging several holes about three feet square. In the first two holes nothing was found, but in the third an ox skull, broken bones, portions of pointed implements of bone, and a bronze socketed spear-head were disinterred. The latter, which was only 18 inches below the surface and above the peaty clay, measured 13 inches long and two inches at its widest part. The bones were of _Bos longifrons_, stag, pig, sheep or goat, large dog or wolf, urus (_Bos primigenius_), and hare. These were all in the peaty stratum. Beside, and along with the bones, were found two or three flint flakes, cores, and rude flint implements. There were several pieces of sandstone, burnt, with the mark of fire plainly upon them, and divers calcined flints. Also a fragment of a thin hand-made vessel. Besides the bones were several stags' antlers, one or two of which were gnawed, probably by dogs, and another had marks of some small-toothed animal, such as a rat. Others were cut by human hands. One antler had a hole rudely worked in it at its broadest part. There were also divers horns of the _Bos longifrons_, and, curiously enough, one of the vertebræ of a Saurian. The latter was a short distance off from the chief "find," and it was suggested that it might have been used as a hammer by some of the natives who brought it to the spot.

The portion of the "find" which caused most conjecture was, however, a fabric of stake and wattle. "I found one stake 2½ inches thick, and 2 feet long, lying close over the spot where we found most of the bones, but the fabric to which I now allude occurred some twenty-eight inches below the surface of the deepest part of the _mere_. The soil in the neighbourhood of it had been disturbed, so I took a spud and trowel and worked the thing out with my own hand. It resulted in an oval or egg-shaped structure of wattle, 5 feet 7 inches long, and 3 feet 10 inches wide. There were 14 uprights, varying from 2 to 2¾ inches in thickness, at nearly equal distances apart. Twigs and sticks were worked in these like the side of a very rough basket. At first I thought it might have been a sunken coracle, but on scooping out the clay with which it was filled, I found that the wattle ceased about 14 inches down, and that the uprights were merely stakes, from 21 to 27 inches long, driven originally into the chalk marl. The bottom of this fabric was filled with broken flints which were also found outside the lower part of the uprights and between them. The flints must have been put in, the points and edges of the points of the stakes being so sharp and clean that they could not have been driven through the bed of flints."

"The top of the wattle was on the level of the chalk marl, on which most of the bones, fragments of pottery, etc., were strewn, and which had been covered over to a depth of from 2 to 4½ feet of dark clay. No more stakes were found, but there occurred divers holes in the chalk marl, some of them nearly in line, in which we could not help thinking they might have once stood. Yet we found no remains of wood in these holes." (B. 161, p. 31.)

Professor Boyd Dawkins, under the heading "Habitations in Britain in the Bronze Age," writes as follows:--

"Sometimes, for the sake of protection, houses were built upon piles driven into a morass or bottom of a lake, as for example in Barton Mere, near Bury St. Edmund's, where bronze spear-heads have been discovered, one 13 inches long, among piles and large blocks of stone, as in some of the lakes in Switzerland. Along with them were vast quantities of the broken bones of the stag, roe, wild boar, and hare, to which must also be added the urus, an animal proved to be wild by its large bones, with strongly-marked ridges for the attachment of muscles. The inhabitants also fed upon domestic animals--the horse, short-horned ox, and domestic hog, and in all probability the dog, the bones of the last-named animal being in the same fractured state as those of the rest. Fragments of pottery were also found. The accumulation may be inferred to belong to the late, rather than the early, Bronze Age, from the discovery of a socketed spear-head. This discovery is of considerable zoological value, since it proves that the urus was living in Britain in a wild state as late as the Bronze Age. It must, however, have been very rare, since this is the only case of its occurrence at this period in Britain with which I am acquainted." ("Early Man in Britain," p. 352.)

LAKE-DWELLINGS IN THE FENLAND.

The discovery of so many submarine dwellings in Holland and the adjacent coasts of Germany which I have already described suggests that similar remains might be found in the Fens and other low-lying districts in Britain. The only reference, however, to such dwellings with which I am acquainted is the following short notice by Mr. Skertchly:--

"I detected the remains of one (lake-dwelling) at Crowland in the year 1870, during some excavations. The piles were of sallow planted very closely together, upon these was laid brushwood, and over this a layer of gravel. Immense quantities of bones, chiefly of the Keltic shorthorn, were found, together with a few bone implements, and a curious ornament of jet. Near Ely, stakes have been found in the peat, but they do not seem to belong to a lake-dwelling." ("The Fenland Past and Present," by Miller and Skertchly. 1878.)

PILE-STRUCTURES IN LONDON.

On December 18th, 1866, Col. Lane Fox (now General Fox-Pitt-Rivers) read a paper at the Anthropological Society entitled, "A Description of certain Piles found near London Wall and Southwark, possibly the Remains of Pile-Buildings."

The author commenced by observing that his attention was directed to this locality by a short paragraph in the _Times_ of the 20th October, stating that upwards of twenty cart-loads of bones had been dug out of the excavations which were being made for the foundations of a wool warehouse near London Wall. The excavation commenced at 40 yards south of the street pavement: therefore, in all probability, at about 70 or 80 yards from the site of the old wall. The area then excavated was of an irregular oblong form, 61 yards in length, running north and south, and 23 yards wide.

A section of the soil consisted of--

"1. Gravel similar to Thames ballast at a depth of 17 feet towards
the north, inclining to 22 feet towards the south end.
"2. Above this, peat of unequal thickness, varying from 7 to 9 feet.
"3. Modern remains of London earth composed of the accumulated
rubbish of the city."

Between the bottom of the peat and the highest spring tide water-mark, as at present existing, there is a margin of 5 feet; but, of course, this might have been different in Roman times.

Regarding the remains of piles in this locality the author makes the following observations:--

"Upon looking over the ground, my attention was at once attracted by a number of piles, the decayed tops of which appeared above the unexcavated portions of the peat, dotted here and there over the whole of the space cleared. I noted down the positions of all that were above ground at the time; and as the excavations continued during the last two months, I have marked from time to time the positions of all the others as they became exposed to view.

"Commencing on the south, a row of them ran north and south on the west side, to the right of these a curved row, as if forming part of a ring. Higher up and running obliquely across the ground was a row of piles, having a plank about an inch and a half thick and a foot broad placed along the south face, as if binding the piles together. To the left of these another row of piles ran east and west; to the north-east again were several circular clusters of piles; these were not in rings but grouped in clusters, and the piles were from eight to sixteen inches apart. To the left of this another row of piles and a plank two inches thick ran north and south. There were two other rows north of this and several detached piles, but no doubt several towards the north end had been removed before I arrived.

"The piles averaged 6 to 8 inches square; others of smaller size measured 4 inches by 3; and one or two were as much as a foot square. They appeared to be roughly cut, as if with an axe, and pointed square; there was no trace of iron-shoeing on any of them, nor was there any appearance of metal fastenings in its planks; they may have been tied to the piles, but if so, the binding material had decayed.[120] The grain of the wood was still visible in some of them, and they appear to be of oak. The planks averaged from one to two inches thick. The points of the piles were inserted from one to two feet in the gravel, and were, for the most part, well preserved, but all the tops had rotted off at about two feet above the gravel, which I conclude must have been the surface of the ground, or of the water, at the time these structures were in existence."

These relics were exclusively found in the peat or middle stratum (which varied from 7 to 9 feet in thickness), and "interspersed at different levels from top to bottom throughout it."

"Amongst the articles of human workmanship found in the peat the vast majority are undoubtedly of the Roman era. Amongst them are quantities of broken red Samian pottery, mostly plain, but some of it depicting men and animals in relief; one specimen is stamped with the name of Macrinus. All this pottery, in the opinion of Mr. Franks, to whom I showed it, is of foreign manufacture. Other samples are of the kind supposed to have been manufactured in the Upchurch Marshes in Kent, and upon the site of St. Paul's Churchyard. Bronze and copper pins, iron knives, iron and bronze stylus, tweezers, iron shears, a piece of polished metal mirror so bright that you may see your face in it (this Dr. Percy has pronounced to be of iron pyrites, white sulphuret of iron without alloy), an iron double-edged hatchet, an iron implement, apparently for dressing leather, a piece of bronze vessel, and other bronze and iron implements, which, thanks to the preserving properties of the peat, are all in excellent preservation. Amongst these were also a quantity of leather soles of shoes or sandals, some apparently much worn, and others, being thickly studded with hob-nails, may be recognised as the caliga of the Roman legions; also a piece of tile with the letters P. PR. BR. stamped upon it. Specimens of these are on the table. The coins found are those of Nerva, Vespasian, Trajan, Adrian, and Antoninus Pius....

"In addition to the Roman relics above mentioned, others of ruder construction remain to be described. They consist of what, in the absence of any evidence respecting their uses, may be called handles and points of bone. The former are composed of the metacarpal bones of the red-deer and _Bos longifrons_ cut through in the middle, and roughly squared at the small end; the others, which are called by the workmen spear-heads, are pointed at one end and hollowed out at the other, as if to receive a shaft. Both Professor Owen and Mr. Blake concur in thinking these implements may possibly have been formed with flint, but I cannot ascertain that they were found at a lower level than the Roman remains, nor have any flint implements, to my knowledge, been found in the place. With them were also found the two bone skates on the table; they are of the metacarpal bone of a small horse or ass, one of which has been much used on the ice. Exactly similar skates also of the metacarpal of the horse or ass have been found in a tumulus of the Stone Period at Oosterend in Friesland; a drawing of them is given in Lindenschmit's Catalogue of the Museum at Mayence, etc. Others have also been found in Zeeland, at Utrecht, and in Guelderland, and there is a specimen in the Museum at Hanover. Professor Lindenschmit attributes all these to the Stone Period, but the specimens on the table are evidently of the Iron Age, the holes in the back having been formed for the insertion of an iron staple. Similar skates have been found in the Thames, but they have not hitherto been considered to date so early in England as in Roman times."

Throughout the peat were several kitchen-middens. One, deposited a foot and a half above the gravel, is thus described:--"A layer of oyster and mussel shells about a foot thick, with a filtration of carbonate of lime permeating through the moss. In this kitchen-midden, Roman pottery and a Roman caliga were found. Close by, the point of a pile, part of which is exhibited, was found upright in the peat; it had been driven in in such a manner that the point descends to the level of the kitchen-midden and no farther. Now, as a pile, in order to obtain a holding, must have been driven at least two feet in the ground, it is evident the peat must have grown at least one foot above the summit of the kitchen-midden before this pile was driven in."

A second kitchen-midden is noted at a height of 3½ feet above the gravel, "composed of oyster, cockle, and mussel shells, and periwinkles, with Roman pottery and bones of the goat and _Bos longifrons_, etc., split lengthwise as if to extract the marrow, with the skulls broken and the horns cut off. It is about a foot and a half thick in the centre, thinning out towards the ends as a heap of refuse would naturally do, and from 12 to 14 feet long; above this is peat for about a foot or a foot and a half, and above the peat another kitchen-midden of the same kind as the preceding. Lastly, the soles of shoes and Roman pottery of the same kind as that found lower down have been taken out at the very top of the peat."

The distinguished investigator, being anxious to obtain further evidence as to the thickness of the stratum in which the Roman remains were found, states that he determined to watch the workmen for four or five hours together during several successive days, while they dug from top to bottom, commencing with the superficial earth, and passing through the peat to the gravel below. The result was as follows:--"Roman red Samian ware is found as high as 13 feet from the surface, but very rarely, and in small quantities. At 15 feet it is frequently found, and from that depth it increases in quantity till the gravel is reached at 18 to 21 feet. The chief region of Roman remains is within two or three feet of the gravel."

Amongst the animal remains were, according to Professor Owen, those "of the horse or ass, the red deer, the wild boar, the wild goat (_bouquetin_), the dog, the _Bos longifrons_, and the roebuck. The horns of the roebuck, I afterwards ascertained, were all found at a higher level. These, and also the horse and goat, entered the superficial earth, in which glazed pottery was also found; but the remainder, including the red deer, wild boar, and _Bos longifrons_, appeared, so far as my observations enabled me to judge, to be confined to the peat."

Subsequently Mr. Carter Blake identified amongst these osseous remains no less than four different kinds of the genus _Bos_--viz. _primigenius_, _trochoceros_, _longifrons_, and _frontosus_; as also a specimen of the ibex of the Pyrenees.

Some human skulls were found in the lowest formation of the peat, or immediately over the gravel. Along with these skulls only three other human bones were found; but this, according to the author, might not be the result of an oversight, as both the Celts and the Romans were known to have practised decapitation.

The piles at the south end were identified as elm, the remainder were oak (_Quercus robur_).

From the above carefully observed and recorded facts it will be observed that in addition to the primary piles which were inserted into the gravel there were others which did not penetrate so deeply, one having been carefully noted which terminated in the peat a foot and a half above the gravel. Facts precisely similar have been observed in almost all pile-dwellings whether on land or in water, showing that the elevations on which the platforms and huts were reared were successively renewed. Another conclusion which we are entitled to draw from the character of the relics and the conditions in which they were found is that in the low-lying districts of London the system of pile-dwellings was known in Britain in post-Roman times. Nor can it be said that this was a solitary instance, for similar remains were found in New Southwark Street, in regard to which the author writes as follows:--