In the Strath district of Skye, Mr. Harker has recently observed that many of the basic dykes, both those older and those younger than the granophyre protrusions, are double, triple or multiple. Thus in a conspicuous dyke, more than 100 feet wide, to the south-east of Loch Kilchrist, belonging to the older series, he has detected at least six contiguous dykes which as they are traced south-eastward, in spite of their interruption by the Beinn an Dubhaich granite, can be seen to separate and take different courses, or successively die out. He remarks, further, that "many cases of apparent bifurcation of dykes are really due to the separation of distinct dykes which have run for some distance in one fissure. Sometimes apparent variations in the width of a dyke are to be explained by this dying out of one member of a double dyke. These multiple dykes are less easily detected in the newer than the older set, owing to greater uniformity of lithological type in the prevalent kinds and to the frequent absence of chilled selvages."[198] An example of a compound basic dyke cutting the crest of the gabbro-mass of the Cuillin Hills is shown in Fig. 333. [Footnote 198: MS. notes supplied by Mr. Harker.] Instances of the second type of compound dykes are less common. Here, instead of being re-opened along one of the walls, the fissure has been ruptured along the centre of the dyke, and a second injection of molten material has then taken place. This structure may be observed where the materials of the compound dyke are on the whole similar, such as varieties of dolerite, basalt, diabase or andesite. In these cases the rock of the central dyke is generally rather fine-grained, sometimes decidedly porphyritic, and often a true basalt. Where broad enough to show the difference of texture between margin and centre, it exhibits the usual close grain along its edges, indicative of quicker cooling. The older dyke presenting no such change at its junction with the younger, was obviously already cooled and consolidated before its rupture. Whilst the centre of a dyke has occasionally proved to be a line of weakness which has given way under intense strains in the terrestrial crust, this rupture and the accompanying or subsequent ascent of molten material in the re-opened fissure may sometimes have been included as phases of one connected volcanic episode. In those instances, for example, which have been above described, where a central vitreous band has risen along the heart of a dyke, the petrographical affinities of the rocks may be so close as to suggest that although the main dyke had consolidated and had subsequently been ruptured along its centre by powerful earth-movements, these changes all belonged to the same period of dyke-making, and the subsequent uprise of glassy material was merely a later phase in the movements of the same subterranean magma.