476 THE FRESH-WATER LOCHS OF SCOTLAND 
vi. Lakes resulting from the irregular distribution of the drift : (1) those 
lying in boulder clay ; (2) those resting on morainic deposits ; (3) those 
situated partly on drift and partly on solid rock ; (4) kettle-holes caused 
by the accumulation of fluvio-glacial sand and gravel round isolated 
masses of ice during retreat, 
vii. Lakes occupying rock-basins, which may be thus classified : (1) plateau 
rock-basins, (2) valley rock-basins, (3) corrie rock-basins, (4) those 
lying along shatter belts due to faults. 
By far the largest number of Scottish lakes is included under the 
last two groups of the above table. There is little room for controversy 
regarding the origin of the various lakes in Scotland, except those 
lying in true rock-basins. We will now proceed to consider the 
probable origin of the latter series in the light of the evidence which 
has already been presented regarding the geological structure, the 
topography, and the glaciation of the country, with the aid of the 
fresh data obtained by the Lake Survey. 
PLATEAU ROCK-BASINS 
The plateau basins are extremely abundant in the coastal belt 
occupied by the Lewisian gneiss on the western seaboard of Sutherland 
and Ross, and also in the Outer Hebrides, where the rocks are re- 
markably bare of drift. They may, however, occur at any elevation. 
Contrasted with the valley rock-basins they are comparatively small 
and shallow. Their distribution is very irregular, and altogether 
independent of drainage. The soundings show that their floors are 
uneven, and that in some cases, as in the Outer Hebrides, four or five 
separate basins occur in one lake. 
To account for them by differential movement would not only 
necessitate a special subsidence in each case, but several irregular 
movements for each lake containing several distinct basins. It is 
no doubt true, as described in the section relating to geological 
structure, that the Lewisian rocks are traversed by shear planes and 
disruption lines which modified the structures of these rocks in pre- 
Torridonian time ; but such movements cannot possibly account for 
these shallow, irregular depressions. This theory seems to us so 
improbable as to be quite untenable. 
On the other hand, evidence has been adduced in the section deal- 
ing with glaciation to show that this coastal belt was crossed by an 
ice-sheet that filled the Minch and overtopped the Outer Hebrides, 
wdiose thickness could not have been less than several thousand feet. 
Throughout the Lewisian Gneiss plateau there are clear proofs of 
the moulding of the surface features by glacial action, and of the 
differential erosion of the rocks by ice. The lake soundings show 
that weak structures have there undergone the greatest modification, 
which may be reasonably attributed to the action of this agent. In 
