DRUMLINS OF CENTRAL WESTERN NEW YORK 427 



the finger lakes, and chiefly in the north and south depression of 

 Sodus bay and Cayuga lake. The nondrumlin spaces can not be 

 regarded as having been occupied by stagnant ice during the 

 drumlin-shaping episode since they are surrounded by drumlins, 

 and are comparatively free from moraine. They can not represent 

 areas of ice movement too vigorous for driimlin accretion or shaping 

 as the ice along the line of flow must have had a practical equality 

 of motion. There is no reason for supposing that there was any 

 lack of drift, since an immense quantity is piled in drumlins imme- 

 diately southward. The location and distribution of the spaces, as 

 well as the drumlins themselves, are such as to oppose the idea that 

 the drumlins represent an original morainal distribution of earlier 

 drift. We have to recognize the probable equality of the drumlin 

 and nondrumlin loci in the elements of depth and pressure of the 

 ice, in its impact and velocity of motion and in its burden of drift 



The following suggestions are offered toward the explanation of 

 these puzzling features. In the region of deep valley filling it is 

 possible that some depressions were below the average level and 

 consequently below the plane of the more vigorous thrustal motion, 

 and it is conceivable that a plane of shearing might have been estab- 

 lished above the depressions. Shearing once established would 

 probably be unfavorable to the initiation of drumlins, as the drum- 

 lins imply some degree of local drag in the bottom ice during the 

 time of accretion or shaping of the forms. The lowest of the open 

 spaces have been partly filled with lake silts and stream detritus and 

 vegetal accumulation, and some are still partly under water, as the 

 Montezuma marshes; but the spaces north of Clyde do not appear 

 to have been leveled by postglacial agents. The existence of well 

 developed drumlins within or on the borders of open spaces might 

 be due to accretion on existing obstructions, while the shearing 

 tendency discouraged initiation of new masses. 



A second suggestion is based on the idea of a complex glacial 

 history. An earlier ice invasion may have localized and heaped 

 the drift in part, while the interglacial stream work carved broad 

 channels through the area, which the latest ice work has not wholly 

 obscured. 



Channels among" the drumlins. These are connected with or 

 blend into the open spaces discussed above and are part of the 

 same problem. They are specially developed between Fairport and 



