146 GEORGE C. MATSON 



drift-filled gorge. The stream which cut this small gorge may 

 have turned aside and entered the broad drift-filled gorge through 

 one of the channels which were mentioned earlier in this paper 

 as being tributary to this gorge. 



Origin of the rock isla?ids. — As already shown, there are three 

 rock islands on Map II. In considering the origin of the rock 

 islands, the first question is : Could they have been formed during 

 the normal stream development? A stream abandons its old 

 course and takes a new one when ox-blow lakes are formed ; but 

 this happens when the stream is flowing on a flood plain, where 

 the work of cutting across a spur is comparatively easy. In the 

 case which we have to consider the rock is hard enough to offer 

 considerable resistance to stream erosion. Moreover, the shape 

 of the rock island is a serious objection to this theory. If they 

 had been formed by a stream meandering, they should be rounded 

 on the upstream side. Reference to the map will show that all 

 of the rock islands are cut off squarely on the upstream side. 



A second way in which rock islands may be formed is by 

 lateral swinging of two streams until they cut through the divide 

 which separates them. If the rock islands had been formed in 

 this way by the uniting of the main stream and a tributary, they 

 should taper to a point on the upstream and downstream sides. 

 None of the rock islands have this form. In the case of the 

 southern rock island, this hypothesis would meet with another 

 objection from the fact that the rock-wall of the old gorge is 

 continuous for one-half mile above the area of Map II. 



From what has just been said it seems impossible to attribute 

 the rock islands to normal stream action. Another possible 

 hypothesis is that the rock islands have been formed as a result 

 of glaciation. In this connection, two possibilities arise. Some 

 of these gorges may have been overflow channels of glacial 

 lakes, or the stream may have been forced to cut new channels 

 around glacial deposits which obstructed its former course. 



Can any of the gorges be the overflow channel of the glacial 

 lakes? The gorges cannot be the overflow channels of a glacial 

 lake, for there is no place where a lake could form with an out- 

 let at this point. Any lake formed in front of the ice in the 



