some mysterious way, dependent on the junction of the gneissitic beds
with, or their transition into, the harder protogine of the aiguilles.

Look back to Plate +33+. The peak of the Bouchard, _a_, is of gneiss,
and its beds run down in lines originally straight, but more or less
hollowed by weathering, to the point _h_, where they plunge under
debris. But the point _b_ is, I believe, of protogine; and all the
opposed writhing of the waves of rock to the right appears to be in
consequence of the junction.

[Illustration: 34. Cleavages of Aiguille Bouchard.]

Sec. 17. The way in which these curves are produced cannot, however, be
guessed at until we examine the junction more closely. Ascending about
five hundred feet above the cabin of the Montanvert, the opposite crest
of the Bouchard, from _a_ to _c_, Plate +33+, is seen more in front,
expanded into the jagged line, _a_ to _c_, Plate +34+, and the beds,
with their fractures, are now seen clearly throughout the mass, namely:

1st. (See references on plate). The true gneiss beds dipping down in the
direction G H, the point H being the same as _h_ in Plate +33+. These
are the beds so notable for their accurate straightness and parallelism.

2nd. The smooth fractures which in the middle of the etching seem to
divide the column of rock into a kind of brickwork. They are very neat
and sharp, running nearly at right angles with the true beds.[72]

3rd. The curved fractures of the aiguilles (seen first under the letter
_b_, and seeming to push outwards against the gneiss beds[73])
continuing through _c_ and the spur below.

4th. An irregular cleavage, something like that of starch, showing
itself in broken vertical lines.

5th. Writhing lines, cut by water. These have the greatest possible
influence on the aspect of the precipice: they are not merely caused by
torrents, but by falls of winter snow, and stones from the glacier
moraines, so that the cliff being continually worn away at the foot of
it, is wrought into a great amphitheatre, of which the receding sweep
continually varies the apparent steepness of the crest, as already
explained. I believe in ancient times the great Glacier des Bois itself
used to fill this amphitheatre, and break right up against the base of
the Bouchard.

6th. Curvatures worn by water over the back of the crest towards the
valley, in the direction _g i_.

7th. A tendency (which I do not understand) to form horizontal masses
at the leve