KAMES. 347 



slope of the surface of the Connecticut would still have been 

 more than one foot per mile. This, in a torrent 2,500 feet 

 wide, with a depth of 140 feet, would produce a current of 

 eight miles per hour on the surface and of six miles on the 

 bottom. With this size of the flood, the rate of discharge 

 would be about four hundred cubic miles of water per an- 

 num ; whereas, at the present time, the total discharge of a 

 year is only about five cubic miles. To cause this enormous 

 rate, Professor Dana supposes that, for a short period, the 

 Connecticut glacier melted at the rate of more than a cubic 

 mile per day. As he estimates the area of this drainage-basin 

 to be about 8,500 square miles, this would imply that at times 

 as much as eight inches per day melted from this surface. 

 This rapid rate of removal in summer is not, however, sup- 

 posed to continue for a long period — probably less than five 

 years. Professor Dana supposes that, at that time, the long 

 tunnels worn in the glacier by the Connecticut and its tribu- 

 taries, when they existed as subglacial streams, had become 

 open channels in the ice. 



Later and fuller investigations of Professor Emerson, 

 give a different interpretation to many of the facts in the 

 Connecticut Valley. It would now appear that the ice dis- 

 appeared from the high lands on either side before it did in 

 the valley, so that there was a series of marginal lakes at 

 successively lower levels leaving accumulations of gravel, 

 and these were traversed occasionally by kame-like ridges. 

 This conforms to what was evidently the course of events in 

 the Champlain Valley parallel to it. There high-level gravel 

 deposits which were at first supposed to be indications of a 

 general depression of land, are explainable on the theory of 

 marginal glacial lakes. 



The shore lines of such marginal lakes are clearly marked 

 at Bakersfield, Franklin County, Vermont, and southward 

 in the valleys of the Lamoile and Winooski rivers. The 

 bowlder channel in Chazy, N. Y., described above, p. 320, 

 marks as already said a water weir between the decaying 

 ice-sheet and the Adirondack Mountains. 



