These characters, Mr. Harker observes, are those of crystals derived from some foreign source, and it can scarcely be doubted that this is the explanation of their presence. He noticed that the dykes in question frequently enclose fragments, varying up to several inches in diameter, of gabbro, granite or granophyre, bedded lava, quartzite, etc., which show clear evidence of having been rounded and corroded by an enveloping magma, and recognizable crystals from some of the fragments may be observed in the surrounding parts of the matrix of the dykes. Most of the felspar and augite crystals disseminated through these porphyritic basic dykes may be referred to the partial reabsorption of enclosed fragments of gabbro. The same observer has found that many of the dykes which rise through the basalt-plateau of Strathaird are crowded with gabbro fragments. Another megascopic character of the material composing the dykes is the frequent presence of amygdales. It has sometimes been supposed that amygdaloidal structure may be relied upon as a test to distinguish a mass of molten rock which has reached the surface from one which has consolidated under considerable pressure below ground. That this supposition, however, is erroneous is demonstrated by hundreds of dykes in the great system which I am now describing. But the amygdales of a dyke offer certain peculiarities which serve in a general way to mark them off from those of an outflowing lava. They are usually smaller and more uniform in size than in the latter rock. They are also more regularly spherical and less frequently elongated in the direction of flow. Moreover, they are not usually distributed through the whole breadth of a dyke, but tend to arrange themselves in lines especially towards its centre (Fig. 236). In these central bands the cavities are largest and depart farthest from the regular spherical form, so that for short spaces they may equal in bulk the mass of enclosing rock. In some rare instances, a whole dyke is composed of cellular basalt, like one of the lava-sheets in the plateaux, as may be seen on the north flank of Beinn Suardal, Skye. Mr. Harker has observed that an amygdaloidal structure is more common among the earlier than among the later dykes of that district. [Illustration: Fig. 236.--Arrangement of lines of amygdales in a dyke, Strathmore, Skye.]