426 NEW YORK STATE MUSEUM 



The isolation of these groups can not be entirely due to channel- 

 ing by the later drainage because the drumlins which cap the island- 

 like rock masses are not themselves eroded, but lie wholly above the 

 plane of the river work. It appears as if the ice rubbing did not 

 touch the lower levels but was confined to higher planes. The forms 

 do not seem explicable on the postulate of a single ice invasion with 

 one episode of correlating ice border drainage. The relation of the 

 drumlins to the erosion, and the character and direction of the stream 

 courses [pi. 9. 10] suggest a complicated history. We have here only 

 one of several groups of phenomena which argue for more than one 

 ice epoch with their correlated stream work. 



Montezuma island groups. Plate 12 shows the largest of sev- 

 eral groups of drumlins which rise out of the marshes that occupy 

 the low ground north of Cayuga lake. These are not conical or 

 cumulative masses like those described above, but isolated groups, 

 of irregular forms and sizes, even down to single drumlins. In the 

 Montezuma district these groups or individuals rise out of the broad 

 marshes as if half drowned. A few small knolls are mapped about 

 the borders of the marshes, like the summits of nearly buried drum- 

 lins, but it does not seem likely that the absence of drumlins over 

 wide tracts could be due to entire burial of drumlins under lake and 

 vegetal accumulations. More likely the marshes are only the low 

 areas similar to others at higher levels .that are destitute of drumlins. 

 This leads to the next topic. 



Nondrumlin areas : open spaces. The broad swamp tracts, 

 like the Montezuma marshes, belong in this category as well as more 

 elevated and drier areas. An example of the latter lies north of 

 Clyde, where a large tract in the center of the Clyde quadrangle 

 shows white on the topographic sheet. The east edge of this tract 

 is shown in plate 3, figure 3. This surface was under Iroquois waters 

 but the lack of drumlins is certainly not due to their destruction. 

 Evidently they were not formed in this tract. The reason is obscure, 

 since the area is irregular in shape with scattering drumlins on all 

 meridians and close set on the west, east and south. This absence 

 of drumlins over considerable' tracts in the midst of heavy develop- 

 ment is more difficult of explanation than the formation of the 

 drumlins themselves 



These puzzling features lie in the region of deep drift filling of 

 ancient valleys, the northward continuation of those now holding 



