XV 
TERRACES OF SHINGLE 
337 
deal of food, in case we should be snowed up, as the season was 
rather late for passing the Portillo. 
March igtk .—We rode during this day to the last, and 
therefore most elevated house in the valley. The number of 
inhabitants became scanty ; 
but wherever water could 
be brought on the land, it 
was very fertile. All the 
main valleys in the Cordillera 
are characterised by having, 
on both sides, a fringe or 
terrace of shingle and sand, 
rudely stratified, and gener¬ 
ally of considerable thickness. 
These fringes evidently once 
extended across the valleys, 
and were united ; and the 
bottoms of the valleys in 
northern Chile, where there 
are no streams, are thus 
smoothly filled up. On these 
fringes the roads are gener¬ 
ally carried, for their surfaces 
are even, and they rise with 
a very gentle slope up the 
valleys ; hence, also, they 
are easily cultivated by 
irrigation. They may be 
traced up to a height of between 7000 and 9000 feet, 
where they become hidden by the irregular piles of debris. 
At the lower end or mouths of the valleys, they are 
continuously united to those land-locked plains (also formed 
of shingle) at the foot of the main Cordillera, which I 
have described in a former chapter as characteristic of the 
scenery of Chile, and which were undoubtedly deposited when 
the sea penetrated Chile, as it now does the more southern 
coasts. No one fact in the geology of South America inter¬ 
ested me more than these terraces of rudely-stratified shingle. 
They precisely resemble in composition the matter which the 
torrents in each valley would deposit, if they were checked in 
