126 MOSTLY MAMMALS 
aqueous formations; and, secondly, even supposing that 
they had been so deposited, they would almost certainly 
have been washed away as the land rose from beneath 
the sea. Then, again, we do not meet with marine shells 
in the desert-sands, of which at least some traces ought 
to have been left had they been marine deposits of com- 
paratively modern age. 
Whether or no the subjacent strata have ever been 
beneath the ocean, it is absolutely certain that the sands 
of all the great deserts of the world have been formed zx 
situ by the disintegration of the solid racks on which they 
“Test, and have been blown about and rearranged by the 
action of wind alone. All deserts are situated in districts 
where the winds blowing from the ocean’s surface have 
to pass over mountains or extensive tracts of land, which 
drain them more or less completely of their load of 
moisture. Hence, in the desert itself, when of the typical 
kind, little or no rain falls, and there is consequently 
no flow of water to wash away the débris resulting from 
the action of the atmosphere on the rocks below. 
In other words, as has been well said, desert-sands 
correspond in all respects, so far as their mode of origin 
is concerned, to the dust and sand which accumulate on 
our high roads during a dry summer. On our highways, 
indeed, the summer’s dust and sand are removed by the 
rains of autumn and winter, only to be renewed the following 
season; but in a desert no such removal takes place, and 
the amount of sand increases year by year, owing to the 
disintegration of the solid rock here and there exposed. 
Only one degree less untrue than the idea of their 
submarine origin is the notion that deserts consist of 
unbroken tracts of sand. It is true that such tracts ‘in 
‘ certain districts may extend on every side as far as the 
