40 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY. 
'ently have not had great effect in confining the swinging and slowly 
degrading river to more and more restricted belts. The narrowing 
of the valleys results rather from the confining influence of rock ledges 
discovered successively nearer and nearer the center of the valley. 
Another conclusion, moreover, is that the curves and scarps of river 
terraces may be most minutely accounted for, and that the action of a 
meandering and swinging river fully explains every detail of terrace 
forms. The West River terraces not onlv confirm the theorv but also 
the entire train of consequences as deduced by Davis. 
Further, and a far more important conclusion from this study of a 
meandering and swinging river, is a better knowledge of the controls 
which determine the behavior of such a river. A river of definite 
Tolume swinging freely in unconsolidated material tends to flow for¬ 
ward in symmetrical curves. During the approach and after the 
attainment of the symmetrical curves of this normal type, continuous 
deposits are formed on the inner and the lower side of each down¬ 
stream-sweeping meander causing the growth of an even plain slop¬ 
ing slightly toward the river — the meander-built plain. Such a per¬ 
fect series of river curves is rarely if ever developed because of the 
many accidents which by new deflection cause a departure of the 
current from the normal delicate adjustment of its symmetrical 
curves. Such a departure of the current may be caused by change of 
volume, cut-off, short-cut, rock barrier, accidental natural obstruc¬ 
tion, or artificial construction. Any one of these controls may 
so throw the river out of adjustment as to initiate the formation 
of a new series of meanders. The sudden deflection of the current 
forces the river to withdraw suddenly from some banks by the forma¬ 
tion of sand bars. The sand bars appear up and down stream from 
the cause of deflection. In many instances their formation can be 
definitely anticipated and located. The addition of successive sand 
Bars is at times of excessive floods. Other islands, portions of the flood 
plain, are often severed at flood season by new short-cut channels. 
The sand-bar and flood-plain islands have invariably been a source 
of trouble and expense in river navigation. It is possible that having 
in mind the controls which cause the departure of the current of a 
river, we may foresee where the islands will form along our rivers and 
thereby simplify navigation. 
This process of lateral swinging due to a deflection of the current 
from the normal curves is recognized by the formation of sand-bar 
