678 



SCIENCE. 



[N. S. Vol. II. No. 47. 



water was turned eastward through the 

 narrow defiles across the Luzerne range to 

 Glens Falls, where it found its present 

 channel to the south. 



2. The gravel deposits bordering the river 

 east of the Luzerne range, and extending to 

 Sandy Hill, are a true delta deposit of the 

 Hudson when swollen by the torrents ac- 

 companying the melting of the ice over the 

 Adirondack region during the last stages of 

 the glacial period. The limitation of the 

 amount of debris and the brevity of the 

 period appear iu the fact that the channel 

 between Fort Ann and Fort Edward was 

 not filled by gravel. 



3. The gravel deposits extending through 

 Saratoga county were made at an eai-lier 

 stage of the recession, when ice occupied 

 not only the region to the north, but the 

 eastern part of the Hudson Valley to a con- 

 siderable distance farther south. This view 

 is supported not only by the line of eskers 

 referred to, but by the fact that throughout 

 this region the glacial striae are from north- 

 east to southwest. These are very pro- 

 nounced in the vicinity of Saratoga Springs 

 and at Fort Ann. It would seem that the 

 retreat of the ice was from the southwest, 

 and that the area about the mouth of the 

 Mohawk was earlier free from ice than were 

 the flanks of the Green Mountains north of 

 Troy; so that during the closing stages the 

 line of resistance for the movement of ice 

 was diagonally across the Hudson toward 

 the area just south of the Kayadarosseras 

 Mountains. 



4. The main line of the Champlain Val- 

 ley extends southward through South Bay, 

 while the main line of the Lake George 

 Valley extends southward through Dun- 

 ham Bay to the Hudson. 



6. The subsidence of the Champlain 

 epoch, which amounted to about 300 feet in 

 the vicinity of Ticonderoga, was probably 

 not much less in the vicinity of Fort Ed- 

 ward ; for it seems evident tliat the delta of 



the Hudson Eiver, which came down at 

 Sandy Hill to the border of the Fort Ed- 

 ward-Fort Ann channel, must there have 

 met still water nearly up to its level of 300 

 feet. The deposits of sand were sharply 

 limited by deep water, while the clay had 

 ample opportunity to settle over all the 

 areas along the Hudson up to a height of 

 from 200 to 250 feet above tide. 



6. There is nothing in this region which 

 indicates a post-glacial depression of more 

 than 300 feet, but everything to indicate 

 the opposite. All the gravel deposits above 

 that level are of the nature of eskers and 

 kames. 



7. The preglacial watershed between the 

 St. Lawrence and the Hudson was probably 

 near the middle of Lake George and at 

 Fort Ann. G. F. Weight. 



Obeelin, O. 



THE EARLY SEGREGATION OF FRESH- 

 WATER TYPES* 

 Dr. Gill prefaced his communication 

 with the statement that it was a familiar fact 

 that some of the most primitive types of ani- 

 mals were represented in the fresh-waters 

 and in them only; this is especially the 

 case with true fishes. It is also well known 

 that fresh-water animals show all degrees 

 of relationship to salt-water forms, rang- 

 ing from species that are anadromous or 

 catadromous to those that are representa- 

 tives of families or groups of families con- 

 fined to the fresh water. But it has not 

 been appreciated how radically a large pro- 

 portion of the fresh-water fauna has been 

 differentiated from the marine. The per- 

 ception of the extent of this differentiation 

 has been delayed by the false taxonomic 

 principles that have long prevailed. A 

 typical instance of the truth of this prop- 

 osition is furnished by the Ostariophysial 



*Al)stract of a jjaper presented by Dr. Theo. Gill 

 before the meeting of Tlie National Academy of Sci- 

 ences, Philadelphia, October 30th. 



