488 THE ANTARCTIC MANDAL, 
running in different directions, and varying from } to 1 inch in width. 
While the crevasses had vertical sides and were hung with icicles, these 
narrow cracks wére full of water. This mosaic of cracks occurs on the 
top of the boss of ice which forms the summit, while on the slopes the 
crevasses assume a transverse direction, being evidently produced by 
the tension of the ice as it creeps downwards on every side towards the 
sea, One main crevasse ran along the whole length of the back of the 
island. On the side where we landed the ice-covering was pierced by 
a few scarcely visible points of rock, while the whole southern shore 
was bordered by a cliff of ice. Another island, similar in every way to 
the one on which we landed, lay quite near, and on it also streams of 
FIG. 10.—ANTAROTIC ISLAND COVERED BY AN IOE-OAP. 
water were trickling from the ice. I made a sketch of this island, which 
is shown in Fig. 10. 
From the eastern side the sierra had a much more gentle slope. 
The cirques in the crest were occupied by very steep glaciers, but 
lower down the wide snow-fields reduced the slopes to a very gentle 
gradient. The rocky walls so characteristic of the north-west of the 
sierra, as seen from the other side of the island, were not represented 
at all on this. On the northern slopes of Graham Land, on the other 
side of the wide channel, I saw an immense glacier descending the 
gentle slopes from the snow-fields which lay about the heights situated 
in the south-west. It was really a majestic ice-stream filling two large 
valleys for three-quarters of their depth, while higher up it completely 
drowned the rocky spur that separated them. 
With Cook I walked round our islet, and at its northern end 
found several fragments of moraines plastered against the slope nearly 
80 feet above the level of the sea, and from 15 to 25 feet in height. 
They contained the same gneiss, granites, and other rock collected 
in the little bays of the shore. The predominant rock was granite with 
hornblende, in fragments which were often angular; the blocks of 
gneiss were often very large and perfectly polished. Since the granite 
with erthoclase only occurs in the form of well-rounded pebbles, it 
doubtless has come from a distance, and the same is true of other rocks. 
The moraine descends very slightly towards the west, and its direction 
