OE THE OUTER HEBRIDES. 841 



caying and receiving tributary glaciers from the mountain-glens of 

 Harris. To this time also must belong many of the large erratics 

 that are scattered over the mountain-slopes of the island. During this 

 stage we may readily conceive that the direction of ice-flow in some 

 valleys would be very different from what it had been when the land 

 lay buried at its greatest depth. Each local glacier would now 

 follow the course of its valley, which might or might not coincide 

 in direction with the flow of the mer de glace. That such was the 

 case we have evidence to show. I have already pointed out that the 

 region between Loch Seaforth and West Loch Tarbert is glaciated 

 from south-east to north-west. This is proved by the trend of striae, 

 by the Stoss-seite of roches moutonnees, and by the phenomena of the 

 till. The glaciation of Beinn Chulish and the hills above the Tarbert 

 points to the north-west ; but in the valley occupied by Loch Lacas- 

 dail the striae run down the valley towards East Loch Tarbert — that 

 is, in almost exactly the opposite direction. As we follow the striae 

 and roches moutonnees down the valley, however, we find them 

 gradually turning away to the west, until the lines of glaciation 

 sweep round the hill above the Tarbert and assume the normal 

 north-westerly course. This remarkable deflection proves that the 

 local glacier of Loch Lacasdail coalesced with and was forced out of 

 its south-east direction by the great mer de glace, which, although 

 at this time considerably reduced in thickness, was yet able to stream 

 through the Tarbert into the Atlantic. Whether, even when the ice- 

 sheet was at its thickest, there may not have been in the bottoms of 

 such deep glens an under- current setting down the valleys while 

 the upper portion of the ice-flow streamed away in the general 

 north-westerly direction, is a question which future investigations 

 will probably answer in the affirmative. 



Long after the mer de glace of the Minch had ceased to impinge 

 upon Harris, the higher valleys continued to hold glaciers, and some 

 of these were of considerable size, comparatively speaking. Ter- 

 minal and lateral moraines, more or less well developed, occur in all 

 the glens of Forth Harris visited by us ; and the striae in such glens 

 were, as a rule, fresher than those on the hill-tops and in the more 

 open parts of South Harris which have never supported local glaciers. 

 I may refer specially to the striae in the valley of Loch Lacasdail, 

 which are wonderfully well preserved, contrasting strongly with the 

 smoothed faces of rock on the higher slopes above the Tarbert, from 

 which the striae have in most cases entirely disappeared. The con- 

 trast between the general and local glaciation is perhaps seen to 

 best advantage in the valleys that come down from the Cliseam and 

 the Langa. In Gleann Abhuinn-eadar, for example, there are not 

 only well-marked moraines, but the rocks in the glen itself give evi- 

 dence of having been ground and rubbed by a local glacier flowing 

 south, while the hill- tops overlooking the valley from west and east 

 are all finely glaciated from south-east to north-west, the ice-worn 

 surface extending up to a height of 1C00 feet at least. 



I have mentioned that moraines and morainic debris occur 

 in greatest profusion in the more mountainous districts of Harris. 



