approached by a winding path with a gradual fall. As soon as the snow
was reached, the slope became very steep, and led promptly to an arch
in the rock, where the stream of ice began. The cave being shallow,
the stream soon came to an end, and, unlike that in the lower
glacière, it filled the cave down to the terminal wall, and did not
fill it up to the left wall. Here the ground of the cave was visible,
strewn with the remains of columns, and showing the thickness of the
bottom of the stream to be about 6 feet only. The arch of entrance had
evidently been almost closed by a succession of large columns, but
these had succumbed to the rain and heat to which they had been
exposed by their position.

The left side of the cave, in descending, that is the west side, was
comparatively light, being in the line from the arch; but the other side
was quite dark, and after a time we found that the ice-stream, instead
of terminating as we had supposed with the wall of rock at the end of
the cavern, turned off to the right, and was lost in the darkness. Of
course candles were brought out, though Louis assured us that he had
explored this part of the cave on his previous visit, and had found that
the right wall of the cave very soon stopped the stream: we, on the
contrary, by tying a candle to a long stick, and thrusting it down the
slope of ice, found that the stream passed down extremely steeply, and
poured under a narrow and low arch in the wall of the cave, beyond
which nothing could be seen. We despatched pieces of ice along the
slope, and could hear them whizzing on after they had passed the arch,
and landing apparently on stones far below; so I called for the cords,
and told Louis that we must cut our way down. But, alas! the cords had
been left at the other glacière! One long bag, with a hole in the middle
like an old-fashioned purse, had carried the luncheon at one end and the
ropes at the other; and when the luncheon was finished, the bag had been