140 Proceedings oj the Itoyal Society 
It is not difficult to understand bow boulders, originally cubical 
in shape, may undergo a change by the action of a stream of such 
a nature as to grind or smooth them. If fig. YIII represents a 
boulder cubical in shape when deposited, a stream coming against 
it in the direction of the arrow might break off or grind down 
the portion a, b , c, and leave the rest a, c, /, e, d, in the form of a 
boulder, smoothed and pointed at the windward, and rough at the 
leeward side. 
That many of these boulders, after being brought to their present 
position, have been subjected to great attrition, is further proved 
by the markings on their surface. Scratches and sometimes 
deep ruts occur, not only on their upper surfaces, but also occa- 
sionally on their sides, as if indicating the passage over and along 
them, of stones harder than themselves, and pressed against them 
by some powerful agent or body. These scratches and ruts are 
most frequently in a direction coincident with the longer axis of 
the boulder, and show that the movement has been towards the 
boulder at its smooth and narrow end. 
2. Another class of facts of some interest is connected with the 
positions of the boulders. 
The rounded boulders, though frequently on the surface, are 
also, and perhaps more frequently, buried in mud, gravel, and sand- 
beds. 
The angular boulders are occasionally found in these deposits ; 
but they are much more frequently on the surface. 
Angular boulders are very frequently on knolls or low hills, 
perhaps even more so than on lower levels. They are occasionally 
seen in clusters upon or round these knolls, as if the agent, whatever 
it was, which transported the boulders, had been obstructed in its 
further progress by the knolls, and had dropped them there. 
As an example of this class of cases, reference may be made to 
the hill of Craigiebarns, about 1| mile north of Dunkeld. This hill 
is about 1100 feet above the sea, and about 800 feet above the river 
Tay, which flows along its base. Four or five large boulders, mostly 
angular, lie on the top, or very near the top, of several rocky knolls 
which form the ridge of that hill. In like manner, there is a hill 
to the south of Dunkeld, also on the east bank of the Tay, with 
an angular boulder on the top of a rocky knoll. 
