Archaeologia/Volume 38 part 1/On Lake-Dwellings of the Early Periods

XII. On Lake-Dwellings of the Early Periods: by William Michael Wylie, Esq., M.A., F.S.A.

Read April 7, 1859.

Some time has now elapsed since we received the first tidings of the discoveries of very ancient remains in the lakes of Switzerland,[1] which have scarcely obtained the attention really due to their archæological importance. These accounts were followed by an earnest appeal from the Society of Antiquaries of Zürich, praying our Society to undertake an examination of the lake Prasias, in the hope of verifying the descriptions of Herodotus, with which these discoveries in Switzerland seem closely to correspond.

The earliest account we have of any people betaking themselves permanently to dwellings on sites of artificial construction among waters is that given by Herodotus of a Thracian tribe who thus dwelt in Prasias,[2] a small mountain lake of Pæonia now part of modern Roumelia. Their habitations, we learn, were constructed on platforms raised above the lake on piles, and connected with the shore by a narrow causeway of similar formation. These platforms must have been of considerable extent, for the Pæonians lived there, in a state of polygamy, with their families and horses; their chief food being the fish which the lake produced in great abundance.

Such an investigation, full of interest as it doubtless would have been, was of course beyond the powers of our Society. Inquiry was nevertheless attempted, and, pending the results of this, I think we may profitably take a rapid review of the circumstances which the Swiss Antiquaries have so deeply at heart. While the many recent discoveries of the vestiges of lake-dwellings in Switzerland allow of a more perfect generalisation than heretofore, we have also better means of attempting some comparison with the analogous, and no less remarkable, remains presented by the archæology of our own land.

In calm weather clustering collections of stems of trees may be observed in the clear waters of Swiss lakes, at depths varying from ten to twenty feet, which have given rise to many theories. Sometimes they have been conceived to be the remains of submerged forests. They are usually observed to run in a parallel direction with the shore, at a distance of about 300 feet from it. Little wots the fisherman, gliding in his skiff over the glassy surface of the lake, that these dark, mouldering stems are the monuments of the patient industry and independence of the earliest inhabitants of his country. How little he suspects that here his forefathers founded their dwellings in bold security among the floods, and that beneath may still be found the most irrefragable evidence of their industry, their arts, and daily occupations. But the veil which concealed 2,000 years of the past has born raised.

In the dry winter of 1853-4, the Swiss lakes and rivers sank lower than had ever been previously known, and the inhabitants of Meilen, on the Lake of Zürich, availed themselves of this favourable opportunity to recover a piece of ground from the lake.[3] Their excavations led to the discovery of the remains of a number of piles, deeply driven into the bed of the lake, formed of the stems of oaks, beech, birch, and fir trees. Among these piles lay a great mass of reliques which, with one single exception, belonged to the stone period. They consisted of hammers, corn-crushers, &.c., and especially a great variety of axes and celts of various kinds of stone, peculiar in some instances to the East. Many of these were fitted into hafts of stag-horn. Implements of flint also were numerous, which is the more remarkable as flint is rarely found in that country. Several ponderous slabs were also noticed which bad evidently done duty as hearth-stones. Mixed with these were numerous implements in bone, the teeth of bears, boars' tusks, and numerous skeletons of deer and wild boar. One amber bead was found, and an armlet of thin brass wire, which was the sole instance of any metal whatever. Pottery occurred in abundance, in a fragmentary state. It was of a rude, coarse description, fashioned by the hand. Masses of charred wood, which apparently were parts of the platform of the building, were abundant. Indeed it was evident that not only this settlement, but the great majority of those subsequently found, perished by fire. Not a trace of a saw was perceptible in the wood-work; the piles had all been pointed with stone axes, or by fire, and split by means of wedges. In fact these remains belong to a very early period, though others, since discovered, claim a yet more remote antiquity. Probably they belong to a race preceding the Kelts, whose very name is lost to us. Other similar discoveries have been made in the same lake.

Of a different class are the pile-remains in the Lake of Bienne, examined by Colonel Schwab in the spring of 1854. An artificial mound exists in this lake near Nidau. It consists of a mass of round stones collected with an immense amount of labour, and which, if the theory be correct of a subsequent rise in the waters of the lake, must at that period have formed an island. The base of this mound is encircled by piles driven vertically, and among them, at the bottom of the lake, planking is observed lying in a horizontal position, possibly for the purpose of retaining the stones in their place. Horizontal planking is not usual in these lakes, though common enough in the Irish crannoges. Remains of piles are to be seen extending across the lake, which narrows here considerably, to the opposite shore. A bridge may have existed here. The depth of water at present is about 20 feet.

The discoveries at Meilen seem to have prompted these researches at the Nidau Steinberg, which have resulted in obtaining one of the most beautiful collections of bronze reliques of the Keltic period perhaps ever made in one spot. They have sufficed to furnish several cabinets, and consist of swords, spear and javelin heads, numerous examples of sickles, celts in great variety, rings, armlets, &c. all of bronze, and often covered with ornamented designs. Implements of stone seem confined to those required for grinding and crushing corn. The pottery, which occurs abundantly, is precisely similar to that found in tombs of the same period in Switzerland. It will be seen that these remains belong to another and later people than those at Meilen.

One curious result of the examination of the Nidau Steinberg is that, among other débris dredged up from the surface of the stone mound, there are masses of the clay used for plastering the interior of the huts which stood either on the island or the pile-supported platforms. This clay in its natural state would have dissolved in the water; but it had been burnt quite hard, probably in the fire which consumed the whole building. The imprints on one side of these clay masses tell us that the sides of the huts must have been of wattled work, and their curves show the diameter of the huts to have been from 10 to 15 feet.[4] Several more pile-constructions have been discovered in the Bienne Lake, one of which, at Möringen, incloses another smaller mound, formed of stones, covering a surface of about half an acre. In the depths of this lake a large boat may be distinguished, which has evidently capsized when laden with stones for one of these artificial islands. It is of the extraordinary length of fifty feet, by three feet and a half in breadth, and is hollowed out of the stem of a single tree. Several smaller boats of the same build have also been found. Such vessels, termed einbäume, are said to have remained in common use in Switzerland down to the beginning of the present century, and I have seen them myself on the small mountain lakes of Bavaria.[5]

Such are the first investigated examples of the lake settlements of the early inhabitants of Helvetia. They belong entirely to the primæval and bronze periods, and may be taken as types of numerous later discoveries. Implements of iron have so rarely occurred, that their presence may be considered subsequent, and purely adventitious. The sword indeed found by Herr Müller at Möringen appears strongly to assimilate with those found in England, and ascribed to the late Romano-British period.[6] But to the subject of the use of iron we shall have occasion to recur.

The investigations of 1854 attracted general attention, and, as a natural consequence, a great number of fresh discoveries ensued. I am able to state, on the authority of our zealous colleague M. Frederic Troyon, who takes a deep interest in these researches, that precisely similar remains of lake-dwellings have been found in the lakes of Constance, Zürich, Bienne, Neuchâtel, Morat, and Geneva; in the small lakes also of Inkwyl and Moosseedorf (Canton de Berne); and in that of Annecy in Savoy. But greatly varying dates must be attributed to these establishments, so far at least as we may gather from the varying degrees of culture displayed in their respective reliques. Thus, for instance, the remains of Moosseedorf would appear far more ancient than those of Meilen, which approximate to the bronze period.

Constructions of the age of stone have been found in the lakes of Constance, Zürich, Inkwyl, and Moosseedorf; also in the fens of the Vallée de l'Orbe, above the Lake of Neuchâtel. Those of the age of bronze are far more numerous. M. Troyon's personal researches have ascertained the existence of thirty of this class in the Leman Lake alone, and of twelve in that of Neuchâtel; while Colonel Schwab has discovered no less than ten in the small lake of Bienne. Similar remains have also been met with in the deep peat-bogs which surround the hill of Chamblon, in the Vallée de l'Orbe. A very ingenious attempt has been made by M. Troyon to arrive at the probable period when the waters of the Lake of Yverdun extended thus far up the Vallée, and encircled the pile-constructions of Chamblon. The Lake of Yverdun, or Neuchâtel, is gradually silting up; the mud and detritus brought down by the tributary streams, together with the growth of the sub-aqueous peat, is gradually filling up the bed of the lake which yearly recedes from the town of Yverdun. The site of this town was still beneath the waters of the lake when the Roman city of Eburodunum was founded on its shore, from which the remains are now 2,500 feet distant. The chalk hill of Chamblon is now 2,800 feet from the site of Eburodunum, and 5,500 feet from the lake, in which it was formerly a chalk-island when those pile-buildings became grouped around it. We have seen that the lake has been about 1,500 years in sinking from Eburodunum to its present level; and M. Troyon argues that, if the silting up has always gone on in the same ratio, 3,300 years must have elapsed since the occupation of the pile-buildings in the Vallée de l'Orbe.

These lake-dwellings seem for the most part to have perished by fire at various times down to the termination of the bronze period. This at once accounts for the extraordinary mass of reliques which invariably appear to be found whenever an examination is made of the foundations. The inhabitants most probably contrived, on these occasions, to escape the conflagration in their boats, since human remains have so rarely been discovered; and their effects sank to the bottom of the lake, among the foundation-piles, where they are always found. Hence we are able to arrive at a tolerably precise knowledge of the culture and mode of life of the owners. Some few settlements, however, whether from a more secure position or other cause, seem to have enjoyed a longer immunity; as we find occasionally the iron and bronze ages in positive approximation. In a recent letter from M. Troyon, I learn that in such rare localities there have been lately found "iron celts, sickles, spear-heads, and swords. The swords arc very remark-able for their perfect conservation. They have great iron blades, broad, straight, two-edged, thin, and flexible; with iron scabbards ornamented with designs which do not belong to the bronze period, and are equally foreign to the Roman. These swords call to mind various passages of ancient authors relating to the arms of the Gauls. I must further add that I am continually finding iron in Helvetic tumuli prior to the time of Cæsar, and that they contain a great number of objects distinct from those of the bronze period. Further discoveries will show whether these lacustral abodes existed during the Roman sway in Helvetia."

It has been invariably noticed that the piles of the stone period are far more decayed than those of the bronze. In general the former barely show themselves above the surface of the mud, while the latter project several feet—a circumstance to be attributed to the greater antiquity of the former, and the wear and tear of the ever-moving water. Sometimes taller piles—that is newer—are found inserted between the old ones. Systematic examination would no doubt often show that Keltic establishments have frequently been perpetuated on the sites chosen by the former race. In such case the layer of stone reliques would be found below those of bronze, just as we but now had occasion to refer to the presence of iron implements with those of bronze. In the former case we may suppose the Keltic invaders to have ousted the early inhabitants; while, in the latter, they themselves underwent the same fate at the hands of some mightier, iron-armed foe.

The number of settlements now ascertained allows us to determine with some decree of accuracy their mode of construction. A site appears to have been selected in from eight to twenty feet of water, where the lake deepened gradually, at about 300 feet from the shore. There the first piles were driven. The diameter of the piles varies from four to ten inches. They stand from one to two feet apart—often at much greater distances, and no further order appears to have been observed in their arrangement than that they ran parallel with the shore. The most extensive of these settlements hitherto met with is that opposite Morges on the Lake of Geneva, where the piles extend 1,200 feet in length by 150 in width, giving a platform surface of 18,000 feet. On this M. Troyon calculates that about 316 cabins may easily have stood; which, only allowing four persons to a cabin, would give a population of 1,264.[7]

That, in betaking themselves to dwellings so difficult of construction and limited in space, these early races had in view a perfect freedom from dangers of every kind, can, I think, hardly be doubted. But we are not at liberty to suppose that these constructions were merely strongholds to which they retreated when pressed beyond their strength by invasion, for everything combines to tell of regular everyday life and permanent occupation. The situation would insure their being fishermen, as the Pæonians were, if even hooks and other fishing-gear did not prove it; and the numerous remains of the urus, the bear, wild boar, fox, beaver, and birds of various kinds,[8] prove the chase to have been a common pursuit. The dog, then as now, was the companion and guard of man; and the presence of the cat shews the fixed and domestic life of these dwellers on the lakes. Even in the remote period that we designate the age of stone, they seem to have had horses; and the adjacent shores supported their oxen, swine, sheep, and goats.[9] In the later bronze period the numerous remains of oxen found during the researches in the Lake of Bienne prove that these animals were then abundant.

In some places, as at Moosseedorf, the great accumulation of chippings of stone and flint, the polishing tools, &c., prove the rude material to have been brought to the platforms, and fashioned, by a long course of patient industry, into implements of daily requirement.[10] Dr. Uhlmann has collected at this one spot above one thousand examples of such implements of the earliest period. Among them are some rare instances of arrow-heads of rock crystal. Elsewhere the débris of the oven or the kiln tell us of the exercise of the potter's art. In the whorl-stones we have evidence that the women plied the spindle. Whether flax was known to them does not appear; hemp they certainly had, and of course wool. It is a matter, too, of some interest to find in this wild region a more widely extended agriculture than usually supposed. "We are now well assured," says M. Troyon, in a private letter, "of the cultivation even then of wheat and barley, grains of which, carbonised in the conflagrations, have remained in perfect preservation. I have also in my possession nuts, beech mast, the seeds of raspberries; also the remains of a kind of mat made from hemp, the cultivation of which must therefore have been known."

It only remains to allude to the general destruction of these dwellings by fire after so long a duration. Impregnable fortresses they no doubt appeared in times when the art of boat-building in this wild region did not extend further than the tedious process of hollowing out a log of timber. Herodotus has certainly placed on record that the Pæonians of the Prasias Lake preserved their independence on the Persian invasion, and defied the attacks of Xerxes: but their safety may probably be attributed to their peculiar position in the lake—ἐν μέση ἕστηκε τῇ λίμνῃ. The short distance of these Swiss villages from the shore, which the depth of the lakes would ordinarily compel, was probably the cause of their destruction. Within easy bow-shot of the shore, they were not out of reach of fiery projectiles, against which thatched roofs and wooden walls would present but a poor defence. Unless therefore the inhabitants were strong enough to defend the shore, their habitations would remain very much at the mercy of an attacking foe.

This brief consideration of the lake-buildings of Switzerland enables us to turn with increased interest to the very analogous remains of our own land—the Irish crannoges. These artificial islands, though long well known, never seem to have been examined till the year 1839, in the case of that of Lagore. A detailed account of the results of this research was communicated to the Royal Irish Academy in the following year by Mr. Wilde.[11] Since that period, in consequence of the works of the Commission for the Arterial Drainage and Inland Navigation of Ireland, no less than forty-six crannoges have been discovered in the lakes of Leitrim, Roscommon, Cavan, Down, Monaghan, Limerick, Meath, King's County, and Tyrone.[12] Many more are probably known, though we have at present no precise information; and a comprehensive work on this interesting portion of our national archæology is greatly to be desired.

We are indebted to Mr. Digby Wyatt for a lucid abstract[13] of the modes of constructing these buildings, which appear to have varied materially. One class, as Lagore, was formed by placing oak beams at the bottom of the lake, above which there are now sixteen feet of bog. Into these horizontal beams, oak posts from six to eight feet high were mortised, and held together by cross beams, till a circular inclosure of 520 feet was obtained. This was divided into sundry timbered compartments which were filled up with earth, and vast quantities of ancient animal remains.[14] Indeed the great demand for the latter as manure mainly led to the discovery in question. So this artificial islet was formed. A second inclosure of posts, based on the first, would show a subsequent rise in the waters of the lake. A great collection of antiquities was found, which, as Lord Talbot de Malahide states, belonged evidently to the iron age.[15] This crannoge, however, probably had its origin in far earlier times, and reliques of the stone and bronze periods might reasonably have been expected had circumstances allowed an earlier and more systematic examination.

The crannoge in Ardekillin Lough, Roscommon, is of an oval form, and may be taken as the type of a second class of construction. Here a stone wall, raised on oak piles, surrounds the island; and the inroads of the waters have also been further provided against by sheet piling, strengthened by oak stockades. The surface of this islet was formed by a layer of stones, resting on strata of ashes, bones, and logs of timber, which would again indicate a rising of the lake, or a subsidence of the island, subsequent to its original formation.

Two crannoges in Drumaleague Lough, Leitrim, present us with a third class. One of these islets, 60 feet in diameter, is encircled by oak piles, in some places doubled and trebled. Piles have also been sunk at various spots within the inclosure, perhaps as a foundation for some building. A sort of platform, of alder trees laid horizontally, surrounds a mass of stones in the centre of the island. Here there are traces of fire. A root of a large tree may also be noticed as having been used as a sort of table. It has been worked to a surface with the hatchet, and round it was a mass of animal remains.[16] Unfortunately, though there is a great amount of precise and very valuable information as to the construction of these early settlements, we know but very little of their positive archæology. Hence a short article by Mr. E. P. Shirley, in the Archæological Journal,[17] becomes of still greater interest and importance. Here we have a brief account of a crannoge island, found in 1844, in the lake of Monalty, co. Monaghan, and of another in the adjoining Lough-na-Glack. Mr. Shirley considers these islets as purely artificial, "from the remains of piles, and transverse portions of oak timber found there." Many reliques have been discovered on them, of various periods; and the very early date of their construction—and a similar attribution would probably be right in every case—is proved by the discovery there of stone celts and other remains of the first period. Besides these, were numerous examples of the usual early bronze implements, and among them some very rare ones of retractile weapons. Many other articles also were found there of various dates, down to the seventeenth century. This is just what might be expected from such permanent island formations, the security of which would at all periods render them favourite "cities of refuge." "It cannot be doubted," concludes Mr. Shirley, "that these islands, or crannoges, were for many years the resorts of petty chieftains, and afterwards, perhaps, of gangs of free-booters." Hearth-stones have generally been met with in the Irish crannoges, as they were also at Meilen on the Lake of Zürich. In a few instances causeways appear to have connected the islets with the mainland. The more usual access was by boats formed like the Swiss einbäume. Such vessels have almost always been found in the vicinity of each crannoge; and one in Drumaleague Lough presented the variety of having apertures cut in its sides, as if for row-locks.

We have ventured to infer from indirect evidence that the Swiss huts were of a circular form, and composed of wattled-work. So too a solitary but certain instance of an old Irish cabin may be supposed to portray the original huts that stood on the islands. In 1833, an ancient log-cabin was accidentally discovered in Drumkellin bog, in Donegal, at a depth of fourteen feet. A full account, with plans, &c., by Captain Mudge, R.N., exists in the Archæologia.[18] It was twelve feet square and nine feet high, being divided into two stories, each four feet high. It was constructed of rough oak planking, split evidently with wedges. The posts of the frame-work were mortised into beams laid horizontally, as noticed in the crannoge at Lagore; the roof was flat. The interstices between the planking of the floor and roof seemed to have been filled up with a composition of grease and sand. A staked inclosure apparently had been raised around it, and there was reason to believe that similar huts were close by. A stone celt, found in the very house, a piece of leather sandal, an arrow-head of flint, and a wooden sword, found very near the spot, give evidence of the remote antiquity of this building, which, no doubt, may be taken as a type of the early dwellings on the crannoge islands.

It will be seen that this subject of lake-buildings is full of interest, if only from the period of extreme antiquity to which it would seem reasonable to ascribe their primary construction and occupation. The reliques found at Moosseedorf clearly belong to that remote period we term generically the age of stone—that age when the rude races that preceded the vast outpouring of the Kelts in Europe were as yet unacquainted with the art of working metals. The pile-buildings of Switzerland then are to be attributed to a pre-historic race— so too perchance may the crannoges of Ireland. Yet no one can imagine that such constructions were confined to Swiss or Irish waters. It is probable enough that, when public attention is sufficiently attracted to keen research, every lake and every river of any magnitude will be found able to furnish evidence of a long-passed-away race that once dwelt on its waters.[19]

Thus far, I believe, we have not been able to recognise in our Irish lakes the Pæonian system of constructing these lake-cabins—that is, on platforms supported by deeply-driven piles—which we have seen was so generally followed in Helvetia. On the other hand, the crannoge system of Ireland seems well-nigh without a parallel in Swiss waters. We have spoken of the two Steinberge of the Bienne Lake, and there is a yet nearer approach to a crannoge in a small island of the stone period in the little Inkwyl Lake; but here, I believe, terminates the list of artificial islands in this region.

The purpose of all, however, was alike the same. From the "walled cities" of the Anakim to "the moated grange" of our own land, the habitations of man in every age will be found to testify to his jealousy of surprise, whether by a treacherous neighbour or an open foe. The yearning for security, and the love of independence that saved the Pæonians from the Persian yoke, and in after ages laid the foundations of Venice in the lagunes of the Adriatic, was no less powerful in the bosoms of the first inhabitants of Helvetia and Ierne.


U. University Club,
March 28, 1859.
  1. Proceedings of the Society of Antiquaries, vol. iii. p. 102.
  2. Herodotus, lib. v. cap. 16.
  3. Keller's Keltischen Pfahlbauten in den Schweizerseen. Zürich, 1854. This work of Dr. Keller's contains a most elaborate and valuable account of the details of the Meilen discovery.
  4. Habitations Lacustres de la Suisse—Statistique des Antiquités de la Suisse Occidentale, par F. Troyon.
  5. Pliny, lib. xvi. cap. 76. Archæologia, Vol. XXXVI. p. 142, note.
  6. Keltischen Pfahlbauten, Pl. iv. fig. 23. Collectanea Antiqua, vol. iii. pl. xvi. p. 67. Proceedings Soc. Ant. vol. iv. p. 166.
  7. Statistique des Antiquités, &c.
  8. Habitations Lacustres.
  9. Id.
  10. Troyon's Ossemens et Antiquités du Lac de Moosseedorf.
  11. Proceedings of the Royal Irish Academy, vol. i. p. 420.
  12. Some accounts of these will be found in Wilde's Catalogue of Antiquities, &c. in the Museum of the Royal Irish Academy; and also in vols. i. and v. of the Proceedings of the Royal Irish Academy. In the Appendix to vol. v. p. xliii. is a valuable report by one of the engineers of the Commission, Mr. T. J. Mulvany, on the "Artificial or Stockaded Islands in the Counties of Leitrim, Cavan, and Monaghan."
  13. Observations on the Early Habitations of the Irish, &c.
  14. These bones consisted of those of several varieties of oxen; also of swine, deer, goats, sheep, dogs, foxes, horses, and asses. Specimens will be found in the Museum of the Royal Irish Academy. Another such crannoge yielded no less than 150 tons of animal remains.
  15. Archæol. Journal, vol. vi. p. 101.
  16. Wilde's Catalogue of Antiquities, &c. in the Museum of the Royal Irish Academy, 1857.
  17. Vol. iii. p. 44.
  18. Vol. XXVI. pl. 361.
  19. To such probably belong the remains found on draining a mere near Wretham Hall, Thetford, Norfolk. Here, in a deposit of peaty mud, twenty feet in depth, "numerous posts of oak wood, shaped and pointed by human art, were found standing erect, entirely buried in the peat." At a depth of from five to six feet from the surface were found some very large antlers of the red deer, which had evidently been sawn off. It is to be regretted that no further investigation seems to have been made.— Quarterly Geological Journal, vol. xii. p. 355.

    It is said too that these lake-buildings have been noticed as still existing in several parts of Asia. In a series of bas-reliefs found at Kouyunjik in the palace of Sennacherib, are represented the conquests of the Assyrians over a tribe who inhabited a marshy region; in one of these slabs (engraved in " The Monuments of Nineveh," second series, pl. 25), we see represented several small artificial islands, formed apparently by wattling together the tall reeds which grew in the marshes, and erecting a platform in which are sheltered five or six people. It has been conjectured by Mr. Layard, that these slabs represent the conquests over the inhabitants of the lower part of the Euphrates. See Nineveh and Babylon, 1853, p. 584. The notice of the habitations of the Papoos, in New Guinea, would seem a complete parallel of the Helvetian and Pæonian constructions.—Dumont d'Urville's Voyages, tom. iv. p. 607. Rawlinson's Herodotus, lib. v. cap. 16, note.