DAVIS: THE HARVARD GEOGRAPHICAL MODELS. 
103 
are merely “ drowned valleys.” The difficulty lias not been in 
the lack of intelligence on the part of the scholars. Their 
natural powers have been long retarded by basing geography on 
artificial, empirical definitions, instead of on observation and rational 
explanation. 
An interesting consequence of the depression by which the first 
model is converted into the third is seen in the dismemberment 
of the chief river system in the original mountainous region, and 
the survival of only the head streams, each of which appears as 
an independent river system in the partly-drowned mountains. 
Streams thus separated from the main river to which they once 
belonged may be called betrunked. They are very common in 
many parts of the world. All the rivers that now flow independ¬ 
ently into Narragansett bay, or into the Baltic sea, are betrunked 
rivers; they once joined a trunk river whose drowned valley now 
forms the bay or sea. 
Just as the explanatory treatment of the dissected coastal plain 
suggested a general method for the description of land forms, so the 
rational view of the uplifted or partly-drowned regions leads to a 
natural way of describing shore lines. Their outline depends, first, on 
the form of the crust (land surface or sea bottom) upon which the sea 
came to edge at the time of the last movement of elevation or 
depression; and, secondly, on the amount of change suffered in the 
original outline by the action of streams and waves. Thus the 
smooth sea bottom of the first model, uplifted to form a coastal 
plain in the second model, had a very simple initial shore line; and 
the initial shore line has suffered small sequential changes in the 
way of delta and sand-reef building, whereby the existing littoral 
features of the second model are produced. The initial outline of 
the third model was very irregular, because the uneven land of the 
first model, with strongly eroded valleys and minutely carved 
ravines, was depressed, so that the sea rose upon the flanks of the 
ridges, entering far between them and following every sinuosity in 
their contour. Since then, sequential changes of appropriate order 
have been introduced in small amount. It is serviceable as well as 
interesting to describe actual shore lines in this way; and by means 
of the type forms represented on the three models much actual 
description may be accomplished. 
The terminology here suggested, initial, sequential, etc., is of 
relatively small importance, although I believe it to be of the same 
