TJic Poie plain . — Tarr, 357 
A similar method was followed in Connecticut; and in 
each sheet where it was possible only those high hills whose 
elevations were especially marked by the U. S. Geological 
Survey were chosen, an especial effort being made to select 
only those hills which could be considered to be a part of the 
supposed ancient peneplain. On the westernmost sheet se- 
lected, the Cornwall, the range is between 1,787 and 1,215 ^^^t. 
From this sheet eastward a strip five miles wide was followed ; 
ahd upon the Winsted sheet, next east from the Cornwall, the 
range is between 1,600 and 1,160 feet; on the Granby between 
1,240 and 720 feet; on the Hartford sheet the prevailing con- 
dition is lowland; on the Tollard sheet the range is between 
985 and 660 feet, and on the Woodstock between 761 and 540 
feet. The distance between the edges of the sheets, in an 
easterly direction, is 91 miles, in which space the total range of 
the elevation of the "peneplain" crests is 1,247 feet, while in 
each sheet the distance to the sea shore in the southern part 
of Connecticut is practically the same. Upon most of these 
sheets a range of several hundred feet between the most 
uniform of the crests ma}' be found at places not more than a 
mile or two apart. 
In Maine, New Hampshire, Vermont, western Massachu- 
setts and the Adirondack region, with similar structure to 
that of the region above mentioned, and so near them that 
they must have been subjected to the same general degrada- 
tion, the lack of uniformity of upland crests is very much more 
marked. One standing upon the crest of one of the moun- 
tains of central Maine would hardly find the evenness suffi- 
cient to give the appearance of levelness even to the eye,> 
unless he were upon the top of a mountain rising well above ' 
the lower peaks, so that the .differences in level disappear from 
view. 
This unevenness of crests in the ancient peneplain remnant '\ 
has not been overlooked by the advocates of this explanation ; 
but to account for it tvyo assumptions have been made. In 
New England there is an increasing elevation of the upland 
crests from south to north and from east to west, a difference 
which,' in the section of Connecticut above considered, 
amounts to over 1.200 feet in 91 miles, measured east and 
west. Near the sea coast the hill tops, the supposed renmants ^ 
