150 FORESTRY OF NORWAY. 
But it is not thus that these rock lakes have been formed. 
Of the theory of Professor Ramsay the following illustra- 
tion is supplied by Professor Geikic:—‘ A river of ice is 
not bound by the same restraints as those which determine 
the action of a river of water. When a glacier is, as it 
were, choked by the narrowing of its valley, the ice 
actually rises. In such places there is necessarily an 
enormous amount of pressure, the ice is broken into 
yawning crevices, and the solid rocks suffer a propor- 
tionate abrasion. The increased thickness of the mass of 
ice at these points must augment the vertical pressure, 
and give rise to a greater scooping of the bed of the 
glacier. If this state of things lasts, it is plain that a 
hollow or basin will be here ground out of the rock, and 
that once formed, there will always be a tendency to pre- 
serve it during the general lowering of the bottom 
of the valley. On the retreat of the ice, owing to 
climatal changes, this hollow, unless previously choked up 
with sand and stones, will be filled with water, and form a 
lake. It will be a true rock basin, with ice-worn surfaces 
around its lip, and over its sides and bottom. 
‘ And such is the appearance presented by many a lake 
and tarn in the Highlands of Scotland. One of the largest 
and noblest of the whole—Loch Awe—may be taken as 
an illustrative example.’ 
Nor isit only the formation of single Jakes which can be thus 
accounted for; a continuous succession of lakes in the direc- 
tion of the movement of the glacier may be thus produced. 
As popular illustrations of the mode of operation I may 
cite the following:— Young boys, and girls too,amuse them- 
selves making what they call ‘ducks and drakes’ by 
throwing flat stones across a placid sheet of water, as 
nearly parallel to the surface of it as they can, causing 
them to skim along and above the water, touching and 
rising again and again, rebounding in ever-diminishing 
bounds till they sink, The same phenomenon may be 
seen on a larger scale in the recochetting of a cannon ball 
