more conspicuous than the three best-developed ones at Coquimbo. They give
to the landscape the most singular and formal aspect; and when the clouds
hung low, hiding the neighbouring mountains, the valley resembled in the
most striking manner that of Santa Cruz. The whole thickness of these
terraces or plains seems composed of gravel, rather firmly aggregated
together, with occasional parting seams of clay: the pebbles on the upper
plain are often whitewashed with an aluminous substance, as in Patagonia.
Near the coast I observed many sea-shells on the lower plains. At Freyrina
(twelve miles up the valley), there are six terraces beside the bottom-
surface of the valley: the two lower ones are here only from two hundred to
three hundred yards in width, but higher up the valley they expand into
plains; the third terrace is generally narrow; the fourth I saw only in one
place, but there it was distinct for the length of a mile; the fifth is
very broad; the sixth is the summit-plain, which expands inland into a
great basin. Not having a barometer with me, I did not ascertain the height
of these plains, but they appeared considerably higher than those at
Coquimbo. Their width varies much, sometimes being very broad, and
sometimes contracting into mere fringes of separate flat-topped
projections, and then quite disappearing: at the one spot, where the fourth
terrace was visible, the whole six terraces were cut off for a short space
by one single bold escarpment. Near Ballenar (thirty-seven miles from the
mouth of the river), the valley between the summit-edges of the highest
escarpments is several miles in width, and the five terraces on both sides
are broadly developed: the highest cannot be less than six hundred feet
above the bed of the river, which itself must, I conceive, be some hundred
feet above the sea.

A north and south section across the valley in this part is represented in
Figure 12.