: 
. 
j 
NOTES ON ANTARCTIC ROCKS. 487 
which is probably augite. Also magnetite dust, a large crop of 
lath-shaped felspar microlites and many small brownish-red 
olivines. Porphyritic constituents absent. There is, however, in 
one of the sections a crystal of zoisite, about 0°75 mm. long by 
05 mm. broad (Plate 15, fig. 2). This is perfectly colourless and 
conspicuous by its very low double-refraction, high index of 
refraction, and rhombic characters. The sectién is six-sided, show- 
ing two prism faces and four pyramid faces. The cleavage 
parallel to 010 is well shown and a series of somewhat curved 
parting planes at right angles to the vertical axis. In this 
particular section the greater elasticity is parallel to the axis c. 
There are some smaller sections of this mineral giving rhombic 
and trapezoidal outlines. 
Basalt, (No. 8). 
This rock in the mass presents a curious foliated appearance. 
It seems to be made up of little irregular ovoid lumps, about 5 mm. 
by 2 to 3 mms., together with less regular fragments into which 
it easily breaks up. The ovoid pieces all lie with their longer 
axes parallel and a face at right angles to these axes was polished. 
It was then seen that the ovoid lumps gave rise to dull round 
spots fairly regularly distributed over this surface. They are of 
softer material than the portions surrounding them, and form 
centres from which irregular cracks radiate out through the sur- 
rounding layers (Plate 13, fig. 8). 
“y would thus appear that the rock has split up into a number 
of irregular ovoid lumps, the longer axes of which are all parallel, 
that most of these have a kernel of softer material the outer 
evelope being traversed by irregular radial cracks. We have not 
“een anything exactly like this structure before, and no definite 
*xplanation has up to the present suggested itself. Doubtless 
Contraction on cooling together with lateral pressures and subse- 
quent percolation of waters along the cracks to the kernels have 
® their share in its formation. In thin sections the slice splits 
- into a number of rounded and polygonal pieces, but no dis- 
Unction is observable between those forming the kernels and those 
