208 MR. A. HARKER ON THE OVERTHRUST [May 1903, 



detailed structures in all these banded rocks, though the larger 

 alternations seem rather to represent distinct and successive intru- 

 sions, with a stratiform disposition following the same direction. 

 We may infer that the rarity of banding in the granites and 

 granophyres of the same districts is connected with the greater 

 homogeneity of the magmas from which these rocks have been 

 formed, and a study of the several groups amply confirms this con- 

 clusion. Throughout the Inner Hebrides — not to go farther afield — 

 the acid plutonic rocks are much more uniform, on a small as well 

 as on a large scale, than the basic, just as these latter rarely 

 approximate to the extreme variability of the ultrabasic. 



At certain places in Skye, however, the acid intrusions assume 

 locally a strongly gneissic structure, and the circumstances of these 

 exceptional occurrences are very significant. They may be examined 

 at more than one ]ocality on Marsco, where considerable bodies of 

 gabbro have been enveloped in, and attacked by, the granite-magma. 

 The acid rock in these places is modified by the inclusion of a 

 certain amount of gabbro-debris, which is found in various stages of 

 digestion down to complete absorption, and the banded rocks form 

 offshoots from the main body traversing the altered gabbro-mass. 

 They compare closely with some of the more acid portions of the 

 Rum gneisses, and are also, like these, intimately bound up with a 

 purer pegmatoid rock, which suggests an effort of the acid magma 

 to free itself from foreign contamination. These exceptional 

 occurrences, in places where the relationships of the rocks are un- 

 equivocal, are especially instructive from the point of view to which 

 we have been led. Primary gneissic banding in an igneous rock 

 may be regarded as the result of flowing-movement in a hetero- 

 geneous magma. In these stratiform intrusions we may reasonably 

 postulate a considerable degree of flow in acid and basic magmas 

 alike, and the development of any pronounced banding will thus 

 depend upon the greater or less heterogeneity of the magma. But 

 heterogeneity may arise in two ways, illustrated on the one hand 

 by the gabbros of Druim an Eidhne and the peridotites of Allival, 

 on the other by the veins on Marsco and the Rum gneisses. The 

 requisite variability must be attributed in the former case to imper- 

 fect segregation of the differentiates from one parent magma, and 

 in the latter case to imperfect commingling of distinct rocks (includ- 

 ing rock-magmas), the presumable consanguinity of which is of a 

 more remote degree. In a due recognition of this principle is 

 contained the clue to the peculiarities presented by the gneissic 

 rocks under discussion. 



I proceed to consider, though without going deeply into petro- 

 graphical details, the characters of the gneisses themselves. Most 

 of them are more or less markedly-acid rocks, of dominant quartzo- 

 felspathic composition ; but of these prevalent acid rocks only a 

 part are of a pure type, representing an unmixed granite-magma. 

 This type is found, associated with other types, basic and hybrid, 

 in all the localities mentioned, though it does not constitute the 



