28 AUSTKALASIAN ANTAKCTIC EXPEDITION. 



is granoblastic, which, in the massive types, might be called gabbro structure. Some- 

 times there is a tendency to a nematoblastic (fibrous) structure and poikiloblastic and 

 relic structures are also present. 



Mineral Composition. 



Hornblende, biotite, and felspar are the commonest mineral constituents that can be 

 recognised in the hand specimen. Sphene, ilmenite, pyrite, epidote, and lawsonite 

 are occasionally visible to the naked eye. Quartz, chlorite, calcite, apatite, fluorite, 

 allanite, and rutile have been determined microscopically. In general, the mineral 

 composition appears uniform along any one band, but this has not been completely 

 tested microscopically. The variation, however, in the mineral content of the different 

 bands is so marked and interesting that it has been considered advisable to give 

 quantitative expression to it. One cannot apply the Rosiwal method and determine 

 the relative mineral volumes in a schistose rock with the same ease that is possible in 

 igneous rocks. We can fairly assume in igneous rocks a similarity along three directions 

 at right-angles, but such assumption is far from correct in schists. Lamellar constituents, 

 like biotites, have relatively large surface area in planes parallel to the schistosity. Hence 

 we have to make the Rosiwal measurements in three sections cut in three planes at right- 

 angles. This involves the preparation of three thin rock sections of each specimen 

 and three times the labor of measurement that is necessary for an igneous rock. In 

 order to obtain some definite idea of the variation in different planes, two sections of 

 specimen No. 153 were prepared, one at right-angles to the plane of schistosity and the 

 other parallel to it. Rosiwal counts were then made on each slide and the figures are 

 given in Table 1. Biotite and hornblende increase their volume by one-third, the epidote 

 increases in smaller proportion, while felspar decreases its volume by one-third. Not- 

 withstanding these differences, it was considered that a relative quantitative expression 

 of the mineral volumes would be obtained from a single section provided that each 

 section of the rocks to be compared is cut in the same direction. The sections studied 

 in each case are cut at right-angles to the plane of schistosity and typical examples 

 selected for treatment. The figures obtained are sufficiently accurate to be serviceable 

 in expressing the relative variation of the constituents in this suite of rocks of common 

 origin and of more or less common history. 



Later study of the crystalline schists from Stillwell Island and Cape Pigeon Rocks 

 shows that the estimation of the mineral content from a single slide may be absolutely 

 misleading in regard to the nature of the rock. Small specimens of gneiss apparently 

 uniform in the hand specimen may contain considerable garnet in one part and none 

 in another part. When, however, the rocks are of comparatively uniform mineral 

 composition, as in the case of this dyke series, the results are useful. 



The figures are given in Table 1, where each example represents a different band, 

 and these are arranged in the order of outcrops met in traversing the area from west 

 to east. No. 1 is the most westerly. The heading " felspar " includes all the colourless 

 constituents the clear albite, the cloudy felspar, quartz, and apatite, lawsonite, and 



