INLAND DUNES. 
483 
most intricate labyrinths. * * A plain often appears to be 
covered with a row of medanos, and some days afterward it 
is again restored to its level and uniform aspect. * * * 
“ The medanos with immovable bases are formed on the 
blocks of rocks which are scattered about the plain. The sand 
is driven against them by the wind, and as soon as it reaches 
the top point, it descends on the other side until that is likewise 
covered ; thus gradually arises a conical-formed hill. Entire 
hillock chains with acute crests are formed in a similar manner. 
* * * On their southern declivities are found vast masses 
of sand, drifted thither by the mid-day gales. The northern 
declivity, though not steeper than the southern, is only spar¬ 
ingly covered with sand. If a hillock chain somewhat distant 
from the sea extends in a line parallel with the Andes, namely, 
from S. S. E. to N. N. W., the western declivity is almost en¬ 
tirely free of sand, as it is driven to the plain below by the 
southeast wind, which constantly alternates with the wind from 
the south.” * 
It is difficult to reconcile this description with that of Meyen, 
but if confidence is to be reposed in the accuracy of either 
observer, the formation of the sand hills in question must be 
governed by very different laws from those which determine 
the structure of coast dunes. Captain Gilliss, of the American 
navy, found the sand hills of the Peruvian desert to be in gen¬ 
eral crescent shaped, as described by Meyen, and a similar 
structure is said to characterize the inland dunes of the Llano 
Estacado and other plateaus of the North American desert, 
though these latter are of greater height and other dimensions 
than those described by Meyen. There is no very obvious ex¬ 
planation of this difference in form between maritime and 
inland sand hills, and the subject merits investigation.! 
* Travels in Peru , New York, 1848, chap. ix. 
f Notwithstanding the general tendency of isolated coast dunes and 
of the peaks of the sand ridges to assume a conical form, Andresen states 
that the hills of the inner or landward rows are sometimes bow-shaped, 
and sometimes undulating in outline .—Om Klitformationen, p. 84. He 
says further that: “Before an obstruction, two or three feet high and con- 
