534 JOHN L. RICH 



down, but no faster, for if melting were more rapid than downcutting, 

 a new channel farther down the hillside would at once be begun. 



Fig. 4. — Cayuta Gorge and vicinity. Note tributary streams A, B, and C, 

 referred to in the text. Scale 1/62,500. 



A channel of this type could be formed only in connection with a 

 retreating or possibly stationary phase of ice-movement; never in 

 connection with an advancing movement, for, since the initial stage 



