186 J. D. Dana—The Flood of the Connecticut River Valley 
all judgments. It is quite possible, however, that a more exact 
study of the terraces will place the flood-level higher than I have 
done ; certainly higher, I think, rather than lower. 
The section gives an impressive idea of the greatness of the 
ancient flood. The river had, as it shows, two great falls in 
its course, one, the Fifteen-mile Falls to the north, below 
Lancaster, the other a’ twenty-five mile fall on the way to 
the Sound from Middletown, Connecticut. It took its full 
magnitude west of the White Mountain region, after its junc- 
tion with the Passumpsic, one of its two chief sources, and after 
accessions from its other head waters, Stevens’ Brook and 
Wells River on the west, and the Ammonoosuc River ané 
Oliverian Brook on the east. At North Haverhill the height 
of the water above modern low-water level was at least 250 
feet, and its width exceeded two miles; and it was nearly or 
quite 200 feet above the same level at Middletown in Connect- 
icut, 172 miles south of Haverhill and 26 miles in a direct 
line along the valley, from the Sound. The diminution in 
depth southward was evidently a consequence of the inereasing 
of some prominent hills. Yet those of the hills that are within 
one to two miles of the river are mostly free from stratified 
drift and terraces, and have instead rounded surfaces and a 
deep covering of till; and it may be that this scoured condi- 
tion was produced by the violent rush of the waters at the fina 
destruction of the ice-dam. 
enable the reader to compare the ancient with the mod- 
ern floods of the Connecticut, a section of the latter from Hol- 
yoke to Long Island Sound is given on the general section. For 
the part from Holyoke to Hartford it is taken from the sec 
tion published in General Warren’s Report as the result of a 
series of careful measurements by General Ellis. For the 
* Report of Mr. Upham, p. 24. + This Journal, III, xi, 178, 1876. 
