Philadelphia Meeting of Geological Society .— Upham. 103 
ently came into force subsequent to the water distribution and is of 
minor importance. 
Mr. Gilbert, in discussion, mentioned that in eastern Colorado, along 
the Missouri Pacific railroad, loess has gathered on the leeward side of 
sand dunes; and Prof. Emerson spoke of the aqueous loess deposited 
by the annual floods of the Connecticut river on the meadows of Had¬ 
ley. Mass., and of the eolian loess on the neighboring hills. 
High-level Terraces of the Middle Ohio aud its Tributaries. G. Fred¬ 
erick Wright, Oberlin, Ohio. Between Steubenville and Marietta, 
Ohio, deposits of glacial gravel were observed in various places on the 
300-foot rock shelf, as on both sides of the river in the vicinity of Steu¬ 
benville, and on the West Virginia side at Wellsburgh. This is about 
twenty miles farther south than such gravel had been observed by Mr. 
Leverett, at Toronto, Ohio. 
Below Wheeling scattered granitic pel^bles of small size were found 
on rock shelves from 200 to 250 feet above the river near Moundsville, 
New Martinsville, and Sistersville. At New Martinsville they occur in 
considerable numbers over an area about a half mile long; and one 
was found a foot in diameter. It is jjossible that these have all been 
carried up from the river a mile away; but their distribution would in¬ 
dicate a deposition by natural means. At Sistersville the pebbles seem 
to have been thrown out from a deposit of river silt 190 feet above the 
river, in which an excavation had been recently made for an oil tank. 
Between St. Mary’s, West Virginia, and Newport, Ohio, there is a 
most singular abandoned channel of the Ohio river, about three miles 
long, which had been worn down to the level of the present bottom of 
the river previous to the deposition of the glacial gravel terrace : for 
both channels were filled up with the gravel to a depth of from 70 to 90 
feet, and the abandoned channel has been nearly oxiened by the small 
streams entering the Ohio at either end. . 
At the lower end of the island between the two channels, there is a 
deposit of sand 200 feet above the river and extending a half mile or 
more in its lee. This has an irregular surface near its upper end, and 
is several hundred feet wide. It could have been formed only when 
the water was flowing at that level. It seems improbable, if not im¬ 
possible, that such a deposit has retained its elevated position during 
all the period required for the Ohio to erode its rocky bed from that 
level to its present bottom. It is, therefore, one of the many residual 
phenomena which point to some such temporary flooding of the middle 
Ohio valley as would have been produced by the Cincinnati ice-dam. 
Its explanation will probably be similar to that of the accumulation of 
silt in Teazes valley, West Virginia, which is at nearly the same eleva¬ 
tion . 
In discussion, Prof. I. C. White attributed the island in large part 
to a preglacial channel of Middle Island creek, which debouches into 
the Ohio valley at St. Mary’s, directly athwart its course and through 
a gorge that is continued in the abandoned channel now forming the 
