610 D. MILNE HOME ON THE PARALLEL ROADS OF LOCHABER. 
streams which now flow down the sides of the hills. These streams must 
have existed during the period of the lakes. They may have flowed not 
only into the lake, but also upon the detritus near the lake in Glen Roy at its 
south end, and have cut out channels through the detritus, which filled that 
Glen near the present mouth of Glen Glaster. The distance between the 
highest blockage in Glen Roy (viz. for Shelf 2) and Glen Glaster was small, 
probably not more than a few hundred yards. The intervening mass of 
detritus would, by the action of the hill streams, after a time be worn down, 
and then the outflow of the lake would take the course of Glen Glaster, leaving 
untouched the remainder of the detritus between Craig Dhu and Bohuntine. 
If this was what happened, it is natural that the new line of blockage should 
be (as shown by the terminations of Shelf 3) right across Glen Roy, and parallel 
with the new outlet through Glen Glaster. 















































































































































VSIA 








. (See p. 609.) 
Section to explain why the upper shelf in Glen Gluoy CC extends farther down the glen 
than the lowertshelf D D. 
It is important next to notice what happened in Glen Collarig. It will be 
seen from the maps (Plates XLI. and XLII.) that Shelf 3, when the lake 
subsided to it, made its mark along the west side of Bohuntine Hill, and also on 
the opposite hill. The scooping out of the Collarig blockage need not have 
happened at exactly the same time as the change in the Glen Roy blockage. 
But the change was of the same nature, and was probably produced by a similar 
cause. Strong streams rush down on each side of Glen Collarig, and at or near 
the very place where the blockage must have originally existed. These streams, 
descending on the detritus, would wear down and remove a large part of it, 
and so allow of the extension of this arm of the lake farther south in the glen. 
The next change which took place was the entire removal of the blockage 
between Craig Dhu and Bohuntine, whereby the lake reached the lowest 
level, viz., that marked by Shelf 4. This change might be effected in the course 
of the general erosion which had long been going on. In the first place, 

