DAVIS : THE HARVARD GEOGRAPHICAL MODELS. 
95 
harcl-rock hills of the back country. The rivers on the plain are 
not separated by definite divides; a great part of the interstream 
surface is really undivided; and in this respect the drainage of the 
immature plain is very unlike that of the maturely dissected moun¬ 
tains, where sharp crested ridges divide the rainfall very definitely 
between the streams in the neighboring valleys. The implication 
often made in text-books, that the basins of adjoining rivers are 
always distinctly divided, is thus seen to be incorrect. As the lateral 
ravines gnaw further headward into the doabs, a better definition 
of the divides will be given, but much work must be done before 
that time arrives. 
Where a river swings to one side of its valley door or strath , and 
flows against the valley side, it undercuts the slope and lays bare 
the structure of the plain. It is thus seen to consist of layers or 
strata of gravel, sand, and clay for the most part. The smooth 
plains of the doabs are merely the surface of the uppermost of these 
strata. The pebbles in the gravels often resemble the harder rocks 
of the hilly or mountainous back country; the clays often contain 
shells, similar to those found in the neighboring sea. 
The rainfall that sinks beneath the surface is known as ground 
water. It moves slowly through the gravelly and sandy layers 
beneath the doab plains towards the sea or towards the valleys. It 
emerges in springs at the base of the valley sides, in the river 
channels, on the shore, or even off shore on the sea bottom. 
Ordinary wells on the doabs encounter the ground water at a 
moderate depth ; artesian wells near the shore find it at a greater 
depth, and fresh water may thus be obtained even on the offshore 
sand reefs. 
The little side ravines by which the margin of the doabs is more 
or less fringed also occasionally disclose the bedded or stratified 
structure of the plain. There is very little land waste washed from 
the fiat upland of the doabs, but the steep side slopes of the ravines 
and the edges of the larger valleys suffer rapid loss by the creeping, 
slipping, and washing of the loose sediments there exposed. It is 
manifest, from the amount of waste that the side streams carry in 
wet weather, that the ravines are continually encroaching at heads 
and sides on the low upland. Years ago, the ravines were smaller 
than they are now; ages ago, the ravines did not exist; in still 
earlier time, even the main valleys had not been cut out. All 
the material that once filled the ravines, as well as that which 
