Chapter XVIII: Section I

_Classification and Geographical Distribution of Ancient Scottish Lake-Dwellings._

The notices of artificial islands in Chapter II. are confined to such as were found to be constructed on timber or surrounded by stockades. There are, however, many others still extant in several of our Scottish lakes, which appear to be entirely composed 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 at least should not be contemporary with the former, as it cannot be assumed that the crannog-builders made wood a _sine qua 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 formation 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 suitable. The outlets of the larger lakes, more especially of 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 stone islands are now found to prevail. Some of them are mere shapeless cairns, without any indications of having been formerly inhabited, while on others are to be found some remains of stone buildings. As to wooden huts or houses, had such structures been erected over them, it is not likely that they could, for any length of time, have resisted the decaying tendencies of a Scottish climate, 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. We have therefore no _prima facie_ grounds for dis-associating chronologically the artificially-formed islands of wood or stone, either from each other or from such natural islands as may furnish evidence of early occupancy. 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 a 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 an island fort to the massive mediæval castle, with its moat and drawbridge, is but another step in the progressive march of civilisation.

When the greater advantages of stone buildings became generally recognised, the old wooden refuges, so liable to decay, so easily destroyed by fire, and so unsuitable for supporting heavy buildings of masonry, would be gradually superseded. It would then be found easier and better to conduct the water to the stronghold than to construct the stronghold in a natural basin of water, however convenient its locality might be. To the transitional period preceding this great change, which culminated in the almost impregnable moated and mediæval castle, may be assigned many of the remains of stone forts, castles, etc., still abundantly found in bogs, drained marshes, and natural or, maybe, artificially built islands.

While it is therefore possible to assign the wooden islands to a fairly well-defined period, which, speaking generally, precedes that when stone and lime were used for building purposes in this country, the claims of all other island homes to great antiquity must be judged of by their special peculiarities.

The annexed tabular statement comprises not only all the known artificial islands, whether constructed of wood or other materials, but also some natural ones known to have been artificially strengthened or fortified, as well as a few examples of other structural remains, such as camps, castles, etc., now or formerly located in bogs or lakes. The first column of numbers in this table contains only the crannogs proper, _i.e._ islands constructed on wood and surrounded by piles; while the second includes all the remaining classes. The characteristic or differential features of all these examples, when not referred to in the text, will be found in the marginal notes or references, so that a mere glance gives a general idea of their number, character, and geographical distribution.

TABLE showing GEOGRAPHICAL DISTRIBUTION OF LAKE-DWELLINGS, ARTIFICIAL ISLANDS, ETC., IN SCOTLAND, with Notes and References. Those to which no reference is given will be found described in the text.

+---------------+-----------------------------+-----------+------------+
|COUNTY. | NAME. |Constructed|Constructed |
| | |with wood, |with stones,|
| | |etc. |earth, etc. |
+---------------+-----------------------------+-----------+------------+
|Ayrshire, |Loch of Kilbirnie, | 1 | ... |
| " |Lochlee, | 1 | ... |
| " |Lochspouts, | 1 | ... |
| " |Buston, | 1 | ... |
| " |Loch Doon,[1] | ... | 1 |
|(Renfrewshire),|Loch Winnoch (Pail),[2] | ... | 1 |
|Aberdeenshire, |Loch Canmore, | 1 | 1 |
| " |Banchory, | 1 | ... |
| " |Federatt,[3] | ... | 1 |
| " |Peel Bog,[4] | ... | 1 |
|Buteshire, |Loch Quien, | 1 | ... |
| " |Dhu Loch, | 1 | ... |
|Berwickshire, |Battleknowes,[5] | ... | 1 |
|Argyllshire, |Kielziebar | 1 | ... |
| " |Loch na Mial | | |
| | (island of Mull), | 1 | ... |
| " |Ledaig, | 1 | ... |
| " |Lochnell, | 1 | ... |
| " |Parish of Kilchoman,[6] | ... | 1 |
| " |Fasnacloich (Appin),[7] | ... | 1 |
|Dumfriesshire, |Lochmaben, | 1 | ... |
| " |Black Loch of Sanquhar, | 1 | ... |
| " |Friars' Carse, | 1 | ... |
| " |Loch Orr,[8] | ... | 1 |
| " |Lochwood,[9] | ... | 1 |
| " |Closeburn,[10] | ... | 1 |
| " |Corncockle (Applegarth),[11] | 1 | ... |
| " |Morton (parish of),[12] | ... | 1 |
|Fifeshire, |Collessie,[13] | ... | 1 |
| " |Stravithy,[14] | ... | 1 |
+---------------+-----------------------------+-----------+------------+

Notes and References.

[1] Island, with castle of Saxon and Gothic architecture. Several
canoes found near it in the loch.--_New Stat. Account_, vol. v. p.
337.

[2] Old castle on an island, near Castle Semple (_ibid._ vol.
xv. p. 69). Canoes also found in the loch.--_New Stat. Account_,
Renfrew, p. 97.

[3] Old castle surrounded by a fosse and morass, with access by a
stone causeway and a drawbridge.--_Old Stat. Account_, vol. ix. p.
191.

[4] Circular earthen mound, having formerly a wooden castle.--_New
Stat. Account_, vol. xii. p. 1089.

[5] Square camp 42 yards each side.--_New Stat. Account_, vol. ii.
p. 171.

[6] Small island strongly fortified.--_Old Stat. Account_, vol.
xi. p. 281.

[7] Artificial island formed of stones and earth.--_Proc. Soc.
Antiq. Scot._ vol. vi. p. 175.

[8] Small island with remains of stone walls.--_Old Stat.
Account_, vol. ii. p. 342.

[9] Strong castle in impassable bogs.--_Ibid._ vol. iv. p. 224.

[10] Old castle formerly surrounded by a lake; canoe and bronze
tripod found in bed of lake.--_Phil. Trans._ 1756, p. 521; also
_Antiq. of Scotland_, Grose, vol. i. p. 150.

[11] Curious wooden structures in moss.--_Proc. Soc. Antiq. Scot._
vol. vi. p. 163.

[12] Old castle, near which canoe was dug up; also a small copper
camp kettle and copper teapot.--_New Stat. Account_, vol. iv. p. 96.

[13] Castle in marshy ground.--_Old Stat. Account_, vol. ii. p. 418.

[14] Regular fortalice situated in a bog, with ditch and
drawbridge.--_New Stat. Account_, vol. ix. p. 365.

TABLE showing GEOGRAPHICAL DISTRIBUTION OF LAKE-DWELLINGS, ARTIFICIAL ISLANDS, ETC., IN SCOTLAND,--_Continued_.

+-------------------+---------------------------+-----------+-----------+
| | |Constructed|Constructed|
| COUNTY. | NAME. | with wood,|with stone,|
| | | etc. |earth, etc.|
+-------------------+---------------------------+-----------+-----------+
|Forfarshire, | Loch of Forfar, | 1 | 1 |
| " | Loch of Rescobie,[15] | ... | 1 |
|Inverness-shire, | Loch Lochy, | 1 | ... |
| " | Loch in Croy (drained), | 1 | ... |
| " | Loch Gynag,[16] | ... | 1 |
| " | Loch Moy,[17] | ... | 1 |
|Kirkcudbrightshire,| Lochrutton, | 1 | ... |
| " | Loch Kinder, | 1 | ... |
| " | Carlingwark, | 2 | ... |
| " | Loch Lotus, | 1 | ... |
| " | Barean, | 1 | ... |
| " | Borgue (parish of),[18] | ... | 1 |
| " | Loch Fergus,[19] | ... | 1 |
|Lanarkshire, | Green Knowe, | 1 | ... |
|Linlithgowshire, | Loch Cot, | 1 | ... |
|Moray, Nairn, and | | | |
|Elgin, | Lochindorb, | 1 | 1 |
| " | Loch Spinie,[20] | 1 | 1 |
| " | Loch of the Clans, | 1 | ... |
| " | Loch Flemington, | 1 | ... |
| " | Loch in Dunty, | 1 | ... |
| " | Lake of Rothiemurchus,[21]| ... | 1 |
| " | Mountblairy,[22] | ... | 1 |
|Perthshire, | Loch Rannoch, | 1 | ... |
| " | Loch Clunie,[23] | ... | 1 |
| " | Loch Earn,[24] | ... | 2 |
| " | Loch Ard,[25] | ... | 1 |
| " | Loch Laggan, Kippen,[26] | ... | 1 |
| " | Loch Morall,[27] | ... | 1 |
| " | Loch Tummell,[28] | ... | 1 |
| " | Loch Tay,[29] | ... | 3 |
| " | Loch Freuchie,[30] | ... | 1 |
| " | Lake in Blairgowrie,[31] | ... | 1 |
| " | Moulin (drained),[32] | ... | 1 |
| " | Loch Granech,[33] | ... | 1 |
| " | Loch Fullah,[34] | ... | 1 |
| " | Loch of Monivaird,[35] | ... | 1 |
| " | Loch Achray,[36] | ... | 1 |
| " | Loch Vennachar,[36] | ... | 1 |
| " | Loch Kinnard,[36] | ... | 1 |
|Stirlingshire, | Loch Lomond, | 1 | ... |
|Sutherlandshire, | Loch Brora,[37] | ... | 1 |
| " | Loch Shin,[36] | ... | 1 |
| " | Loch Dolay,[36] | ... | 1 |
|Ross-shire, | Loch of Kinellan, | 1 | ... |
| " | Loch Achilty, | 1 | ... |
| " | Loch Glass,[38] | ... | 1 |
|Roxburghshire, | Castletown,[39] | ... | 1 |
| " | Loch of Yetholm,[39] | ... | 1 |
|Wigtownshire, | Dowalton, | 5 | ... |
| " | Loch Inch Crindil, | 1 | ... |
| " | Castle Loch,[40] | 1 | ... |
| " | Barlockhart Loch,[41] | 1 | ... |
| " | Sunonness Loch,[41] | 1 | ... |
| " | Barneallzie Loch,[42] | 1 | ... |
| " | Machermore Loch,[43] | Several | ... |
| " | Barhapple Loch, | 1 | ... |
| " | Loch Heron,[44] | ... | 2 |
| " | Mochrum Loch, | ... | 1 |
| " | Fell Loch,[45] | ... | 1 |
| " | Merton Loch, | 1 | ... |
| " | Eldrig Loch,[46] | 3 (?) | ... |
+-------------------+---------------------------+-----------+-----------+

Notes and References.

[15] _Proc. Soc. Antiq. Scot._ vol. vi. 176.]

[16] Small island, with traces of stone castle.--_New Stat.
Account_, vol. xiv. p. 65.

[17] In Loch Moy are two islands, on one of which stands the old
residence of the family of Mackintosh. The other is merely a heap
of stones, probably artificial, and was used by the Lairds of
Mackintosh as a prison. It is called _Ellan-na-glach_.--_Ibid._
vol. xiv. p. 100.

[18] Fort surrounded by water. In the drained lake fragments of
spears and a silver coin found.--_New Stat. Account_, vol. iv. p.
54.

[19] Artificial lake, with two islands, said to be seats of Fergus
Lord of Galloway.--_Old Stat. Account_, vol. xi. p. 25.

[20] On the north-west border of Loch Spinie there are standing
on an artificial mound, surrounded by a fosse and drawbridge, the
walls of a strong castle called Old Duffus.--_Old Stat. Account_,
vol. viii. p. 395.

[21] Contains an island said to be one of the strongholds of the
Wolf of Badenoch; also called Loch-an-Eilean.--_New Stat. Account_,
vol. xiii. p. 137.

[22] Castle situated in a swamp.--_Old Stat. Account_, vol. iv. p.
399.

[23] A small island, mostly artificial, with ruins of an old
castle.--_Old Stat. Account_, vol. ix. p. 231.

[24] Near each end there is a small artificial island with
ruins.--_Ibid._ vol. xi. p. 180.

[25] Small island, with ruins of castle.--_Ibid._ vol. x. p. 130.

[26] Middle of loch a cairn of stones.--_Ibid._ vol. xviii. p. 327.

[27] _Proc. Soc. Antiq. Scot._ vol. vi. p. 176.

[28] Island, partly artificial.--_Old Stat. Account_, vol. ii. p.
475.

[29] Several islands.--_Old Stat. Account_, vol. xvii. p. 465, and
_Proc. Soc. Antiq. Scot._ pp. 173, 175, and 176; also _New Stat.
Account_, vol. x. p. 465.

[30] _Proc. Soc. Antiq. Scot._ vol. vi. p. 173.

[31] In the middle of one of the lakes is a small island, with the
remains of an old building.--_Old Stat. Account_, vol. xvii. p. 195.

[32] Castle stood in lake, now drained, with vestiges of a
causeway.--_Old Stat. Account_, vol. v. p. 69.

[33] Mr. Robertson's notes.--_Proc. Soc. Antiq. Scot._ vol. vi. p.
177.

[34] _Ibid._ p. 172.

[35] Castle anciently surrounded by water.--_Old Stat. Account_,
vol. viii. p. 570.

[36] _Proc. Soc. Antiq. Scot._ vol. vi. pp. 172 to 177.

[37] Small island near the lower end artificially constructed of
stones, with ruins.--_Old Stat. Account_, vol. x. p. 303, and _New
Stat. Account_, vol. xv. p. 151.

[38] Small island near lower end artificially formed of
stones.--_Old Stat. Account_, vol. i. p. 282.

[39] Immense cairn of stones in the midst of an extensive and deep
morass. Old castle of Yetholm Loch.--_New Stat. Account_, vol. iii.
p. 164.

[40] Contains an island of stones and oak stakes, and mossy bogs on
south shore, and a peninsula at north-west, with a double row of
stakes.--_Rev. W. Wilson, Glenluce._

[41] See _Proc. Soc. Antiq. Scot._ vol. x. pp. 737-8.

[42] _Ibid._ vol. ix. p. 397.

[43] _Ibid._ vol. ix. p. 368.

[44] Mr. Faed examined its two islands and found them
artificial.--_Ibid._ p. 378.

[45] "On the east shore, opposite Fern Island, I found an oak in
the peat, with an axe-mark. My companion waded to the island and
reported the remains of a paved ford for 20 or 25 feet next the
island."--_Ibid._ p. 378.

[46] "Three crannogs, one with ford to one shore and annular stone
heap, the others with a ford to each shore.--_Rev. W. Wilson._

It is manifest, however, that a table of this kind can have no permanent value, beyond giving a full and accurate statement of discoveries up to date, as further researches may not only change its numerical data, but give a totally different aspect to inferences based on the existence or absence of these remains in certain districts. Thus it is only within the last few years that Ayrshire could be included in the lake-dwelling area, so that, previously, the conflicting statement made by Chalmers,[40] that Galloway was colonised by the Irish about the eighth century, derived some countenance from the archæological discoveries in Loch Dowalton, and other lakes in the neighbourhood, when taken in conjunction with the prevalence of analogous remains in Ireland. Though we cannot, therefore, argue definitely from the present geographical distribution of Scottish Lake-Dwellings, the indications are so clearly suggestive of their having been peculiar to those districts formerly occupied by Celtic races that the significance of this generalisation cannot be overlooked. Thus, adopting Skene's division of the four kingdoms into which Scotland was ultimately divided by the contending nationalities of Picts, Scots, Angles, and Strathclyde Britons, after the final withdrawal of the Romans, we see that of all the crannogs proper, none have been found within the territories of the Angles;[41] ten and six are respectively within the confines of the Picts and Scots; while no less than twenty-eight are situated in the Scottish portion of the ancient kingdom of Strathclyde. Nor is this generalisation much affected by an extension of the list, so as to include those stony islets so frequently met with in the Highland lakes. On the other hand, that they have not been found in the south-eastern parts of Scotland may suggest the theory that these districts had been occupied by the Angles before Celtic civilisation--or rather the warlike necessities of the times--gave birth to the island dwellings. In that case we would suppose that their development dates back to the unsettled events which immediately followed the withdrawal of the Roman soldiers, to whose protection the Romano-British population in the south-west of Scotland had been so long accustomed. But this leads me to notice some of the historical phenomena associated with the localities thus referred to.