246 MR ALFRED HARKER ON 



sheets intersecting the gabbro invariably play an unduly prominent part as compared 

 with the boulders of gabbro. The composition of the drift in the lower corries is, 

 indeed, very remarkable. Tairneilear is a good example, since here the question is not 

 complicated by patches of basaltic lavas enclosed in the gabbro. The bare surface of 

 the corrie is composed of gabbro intersected by dykes and sheets of basalt and dolerite. 

 These minor intrusions, though very numerous, make up but a small part of the whole : 

 probably one-twentieth would be an over-estimate. When, however, towards the 

 mouth of the corrie, we come upon the drift, we find that boulders of these rocks make 

 up, not one-twentieth, but from one-third to one-half of the total boulders. Clearly 

 some selective influence has operated. Since the average constitution of the drift must 

 be the same as that of the rocks in place, we conclude that the jointed and brittle 

 rocks of the minor intrusions have mostly broken away and formed boulders, while the 

 more massive gabbro has in much greater measure been ground down and gone into the 

 matrix. 



Hitherto we have made no classification of the drift accumulations. The more or 

 less continuous deposits with which we have been dealing, seem to belong wholly to 

 the time when the flow of the Skye ice outside the mountain-area was diverted in the 

 manner already described by the pressure of the Scottish ice-sheet. Among them we 

 may, however, distinguish two types : one corresponding with the phase of maximum 

 glaciation, when central Skye was covered by a continuous ice-cap ; the other connected 

 with the waning phase of this principal glaciation, when only fragmentary relics of the 

 ice-cap remained. The most widely spread type of drift in our area has the ordinary 

 characters of a ground-moraine. In the valleys and on the lower slopes of the basalt 

 country, where it is best displayed, it imparts to the landscape the familiar gently 

 undulating appearance with smooth flowing outlines. Here it consists of a reddish 

 sandy clay enclosing numerous small boulders and some large ones, often planed and 

 scratched. Elsewhere its composition varies to some extent, the local element being 

 always important. The second type is what we have called in mapping the country the 

 ' hummocky ' drift, and it seems to answer to the ' kettle-moraine ' of some American 

 geologists. The ground is closely studded with circular mounds, like tumuli, usually 

 from 10 or 15 feet to 50 or 60 feet in height, only rarely showing any linear or other 

 arrangement. The finer material, which may be regarded as a matrix, is commonly 

 reduced to a minimum, the bulk of the accumulation consisting of boulders, mostly 

 subangular but rarely scratched. It is often noticeable that the larger boulders occur 

 towards the summit of the mound. The hollows of the irregular surface frequently 

 hold tarns, from 300 or 400 yards long down to mere pools, those in the neighbourhood 

 of Sligachan being good examples. Loch an Fhir-bhallaich, near Glen Brittle, illus- 

 trates another kind of tarn, occurring above and outside the margin of an area of 

 kettle-moraine, but held in by it. 



The hummocky drift has a much more restricted distribution than the smooth, and 

 this distribution is a significant one. The patches of hummocky drift lie constantly 



