often be easy to suppose the latter intrusive in the former. Careful examination, however, will show that there is no doubt as to their true relations. The serpentine is generally much decomposed.
On the western side of the headland the gabbro extends for some distance along the shore. I was unable to examine closely its junction with the main mass of the serpentine; it appeared, however, to be, as usual, intrusive. A small shallow gully just on this side of the actual headland affords the best study of the gabbro, which might be mistaken for a metamorphic rock. The schistose structure strikes about N. and S., extending in considerable perfection over a space about 5–6 yards broad, and dips to the eastern side at an angle of about 80°. It is, however, quite impossible to draw any line of demarcation between the foliated and the ordinary (rather coarsely) crystalline gabbro. This consists of a purplish-grey plagioclase felspar, probably labradorite, often mixed up with a dead-yellowish-white felspar (the saussuritic variety already mentioned), which of course predominates on exposed surfaces, crystals of brownish diallage, often about 12 inch across, and having a metalloidal lustre, and a considerable, but variable, quantity of the minute rather dark green hornblende already described. In short, the process of alteration from an augitic to a hornblendic rock has taken place here as at the Balk; and specimens may be found in almost every stage. Not seldom the diallage seems to be entirely replaced by these pseudomorphs. In some of the most schistose varieties the dark "eyes" of this hornblende remind one in appearance of the spots in the Knotenschiefer.
The two minerals (the felspar and diallage, or hornblende) are often quite separated in alternating bands, those of felspar being from nearly 14 inch downwards to mere lines. Not seldom the diallage predominates, felspar only occurring in very thin threads, with occasional "eyes" as described above. From such specimens we pass to normal coarse gabbro—a variety in which the plates of the diallagoid mineral are wavy in outline, and tend to be parallel, being very common.
Among the most schistose varieties lenticular pieces and long slab-like masses of included serpentine are very abundant, and may not improbably have contributed to the development of the structure, as at the Balk.
I have had two sections cut from the gabbro of Karak Clews. The normal rock consists now chiefly of short broad rather irregular crystals of partly altered plagioclase, with numerous microliths and aggregated small crystals of pseudomorphous actinolite. Here and there, however, portions of plagioclase crystals still remain but little altered, as well as crystals of diallage in which the change to hornblende has not been completed. Every stage of the change from plagioclase to the saussuritic mineral can be traced in various parts of the slide. There are occasional microliths and larger grains of magnetite; and one or two of the diallage crystals are filled with an opaque black dust, the result of decomposition. These
