( ■ 
FISHER: RIVER TERRACES. 
37 
from the defended cusps, successive terraces may be consumed, until 
by chance a lower rock ledge may be discovered. Such a lower ledge 
would protect the last flood plain and the high escarpment back of it. 
Davis (’02, p. 320) has concluded “that low undefended, high-level ter¬ 
races of early swings are most likely to be preserved back of defended 
cusps of later swings; that the undefended terraces of early swings 
would probably be swept away in the production of a single high 
scarp terrace wherever broad swinging at low levels is not prevented; 
and that when high scarps occur in a flight of stepping terraces they 
are more likely to be found at or near the top than at the bottom 
of the flight.” 
It is therefore evident that the preservation of the many flights of 
stepping terraces has been made possible by rock ledges — an emi¬ 
nently important control — as was suggested by Miller, as Davis has 
clearly demonstrated and as is proclaimed by almost every river ter¬ 
race in New England. 
Relation of terrace 'pattern on two sides of valley .— It is a generally 
accepted fact that the terraces on the two sides of the valley,of an 
essentially graded river slowly degrading, do not necessarily agree in 
number or in height. The complete swing to and fro across the 
valley results in an appreciable amount of degradation. In the Brattle- 
boro region, twelve feet appears to be a characteristic one-swing scarp. 
The difference need not necessarily be always maintained, since by 
the cutting-off process the river is immediately withdrawn to a new 
position. Much evidence has been given to prove that the number of 
terraces remaining on opposite sides must be accidental as determined 
by constraining ledges, so that the same number or a widely differing 
number may exist on the two sides,— assuming always that they w^ere 
systematically formed on both sides, but that one or more may have 
been completely removed. Another fact with reference to the terrace 
plains is that they slope with the stream; if the grade of the river is a 
strong slope, the difference in the up-stream and the down-stream 
ends of the plain will be appreciable. The West River (see profiles, 
plate 2) at Brattleboro shows a fall of ten or twelve feet in a distance 
of a half mile, consequently the meadow shows a grade of probably 
ten or twelve feet within the same distance. Such a down-stream 
slope of the flood plain makes it difficult to correlate the terraces on 
opposite sides of a steeply graded river, because any change of grade 
would have directly affected the slope of the flood plain, Granted 
