DAVIS: TIIE HARVARD GEOGRAPHICAL MODELS. 
107 
easily apprehended. Among these peculiarities, none is more 
marked than the irregular distribution of falls on the extended 
streams of to-day. Normally, the falls should he found only where 
the streams leave the oldland rocks at the inner margin of the plain ; 
in Maine falls occur in the coastal plain wherever the streams happen 
to incise their channels upon a buried rocky spur or knob. It is for 
this reason that so many falls occur there close to the sea, while the 
inner margin of the coastal plain is thirty or more miles inland. Thus 
starting with an example that seems to correspond, as far as drowned 
valleys are concerned, with the third model, we are led to consider 
the effects of elevation, in which the second model is our aid. 
These complications in so familiar a region as the coast of Maine 
suggest the importance of adding new models to the small series 
here described, and thus systematically developing the most varied 
topography from the simplest elementary examples. It is therefore 
proposed to construct a successor to the third model which shall 
represent the same district after the lapse of a considerable period 
of time, during which the ridges shall have been perceptibly worn 
down to fainter relief, and the large stream to the north of the, main 
range shall have been captured by the headwaters of the south- 
flowing stream that lies on the contact between the older and 
younger rock series. At the same time the deltas will have greatly 
advanced, filling in most of the bays. A number of independent 
streams in the third model will thus be engrafted upon a single trunk 
again ; but this trunk ma}^ not run precisely above the course of its 
ancestor in the first model. While all this is going on, the waves 
will have consumed the outer islands, and the headlands will be 
much worn back. Sand reefs will be spread along certain parts of 
the shore, tangent to the cliffs from which much of their supply is 
derived; and some of the bays will be thus enclosed. The inner 
islands will be tied to the mainland, either by forward growth of the 
delta plain, or by backward growth of sand reefs. The area of easily 
occupied lowland will be greatly increased; a large and flourishing 
population might settle on such a plain. Still another model might 
represent the mountain peaks and ridges reduced by long continued 
denudation to a peneplain, above which a few monadnocks might 
stand as the last witnesses of the vanished highlands. This would 
then serve to explain the ancient peneplain worn on the older rocks 
of the first model before the newer strata were deposited. 
