Vol. 59.] TORRIDON1AN, ETC. OF THE ISLE OF RUM. 205 



addition to these plutonic rocks, there is the peculiar porphy- 

 ritic quartz-felsite already mentioned, of somewhat later date, 

 confined to the overthrust belt and generally intruded among or in 

 contact with the crush-breccias. Since the prevalent rock among 

 the gneisses is always of acid composition, it is with the third 

 group, if any, that we must correlate it. The rocks which we have 

 mapped as gneiss and granite, respectively, come together in the 

 central part of the island, about Loch Sgathaig. The ground here 

 is much obscured by peat, and the boundaries laid down are merely 

 empirical, being drawn to divide as simply as possible the outcrops 

 which show gneissic banding from those which do not. There is 

 nothing inconsistent with the supposition that the two are parts of 

 one and the same mass, but the exposures are not sufficient to 

 warrant a conclusion either in that sense or the opposite. Farther 

 west, on Beinn a' Bharr-shaibh, the evidence is much clearer. Here 

 a patch of well-characterized gneiss, mainly of granitic composition, 

 but with some basic material, forms the summit of the hill, and 

 extends for some distance down the south-eastern slope, and, as 

 stated above, this patch is wholly surrounded by granite. On the 

 bare upper part of the hill the relations between the two rocks are 

 easily examined, and it is seen that no sharp divisional line can be 

 drawn between them. The one appears to graduate into the other, 

 often through a transitional zone of a rather coarse-looking or 

 pegmatoid rock, a type common in other localities as an integral 

 part of the banded gneissic complex. Thus, at the only place 

 where the relations are clearly displayed, the gneiss is to all appear- 

 ance inseparable from the granite. The locality and the situation 

 of the patch of gneiss on the shoulder of the hill are consistent with 

 the supposition that the original boundary of the large acid intru- 

 sion was not far above this place. The porphyritic quartz-felsite of 

 the mountain-border belongs to a later phase of Tertiary igneous 

 activity, and its posteriority to the gneiss is easily demonstrated. 

 On the hill east of Loch Gainmhich, a strip of gneiss nearly 100 

 yards long, w r ith metamorphosed Torridon Sandstone adherent on 

 one side, is enveloped in the felsite. The lower intrusion of felsite 

 on Beinn nan Stac (fig. 4, p. 202) seems to be of dyke-like habit, 

 forced in along the border of the lenticle of gneiss. 



The relation of the gneiss to the ultrabasic and basic intrusions 

 is yet to be considered. It is to this that we should look for a 

 final test of the suggested Tertiary age of the gneissic rocks ; for, 

 on this hypothesis, the gneiss, or at least the acid rock which is its 

 dominant element, should be newer than the massive intrusions. 

 In the central part of Rum, near Loch Sgathaig and Priomh-loch, 

 the gneiss is bounded in several places by peridotite andgabbro ; and 

 one patch, west of Priomh-loch Mor, is entirely surrounded by those 

 rocks. I have not, however, detected here any exposure of the 

 actual junction, from which one might draw conclusions. The 

 gneiss in this district has a high dip to the north, irrespective of 

 the form of its boundary ; but this does not imply that the gneissic 



