MARGINAL GLACIAL DRAINAGE FEATURES 547 



an even slope eastward toward the ice. There is evidence that this 

 delta was built against the ice. Its eastern margin shows several 

 deep kettles and other signs of the melting out of buried ice. 



The top of the delta is one hundred feet higher than the Watkins 

 Lake outlet channel, which began only three-fourths of a mile distant. 

 This indicates clearly the deposition of the delta material in a marginal 

 lake rather than in Lake Watkins. There is good evidence that 

 while a large stream was still coming down Johnson Hollow, the ice 

 melted out sufficiently from Seneca valley to cause the draining of this 

 marginal lake and to bring about the initial stage of Lake Watkins 

 at a level one hundred feet lower. At this level, just opposite the 

 beginning of the outlet channel is a second delta built into Lake 

 Watkins by a stream from Johnson Hollow. While forming the 

 lower delta the stream cut a broad channel though the one first formed. 



The Johnson Hollow series of channels presents many features of 

 interest. It shows that while an ice-tongue lay in Seneca valley, at 

 least as far as Millport, a lobe pushed westward into the Moreland 

 valley and ended approximately at the head of Johnson Hollow. It 

 also illustrates the case of a stream leaving the ice at one point and 

 returning to it at a lower level several miles distant. This gives a 

 means of estimating the ice slope, which seems to have been about 

 sixty-six feet per mile. There is a possibility, however, though from 

 the evidence it seems very improbable, that, while the upper delta was 

 building in the marginal lake, the ice-lobe from which the stream came 

 pushed for considerable distance into Johnson Hollow ; and that it 

 was only during the building of the later delta in the early stage of 

 Lake Watkins that the channels at the head of Johnson Hollow v^ere 

 formed. If this were the case, the ice margin might have had a con- 

 siderably greater slope, for the ice-tongue would not then necessarily 

 reach as far south as the delta. 



SUMMARY 



Our study of the Finger Lake channels has brought out some 

 features of interest and importance. It has shown, in the first place, 

 the large number of these channels. It has shown also that, although 

 from their nature the greater number of the channels now visible 

 were formed during the final retreat of the ice, a considerable per- 



