WEATHER AND OLACIATION REEDS 321 



Reeds (192C), calls attention to two correlation coefficients, one of 

 + 35, which suggests that the teleconnections appear to be undenia- 

 ble; the other of +24, whicii shows that nearly half the apparent 

 connections are due to a 2-year periodicity coiunion to botli. Brooks 

 makes further comments, but it does not seem desirable to feature 

 them, since Do Oeers correlations, in this instance arc premature and 

 misleading. Keeds informed l)e (Jeer of this fact in August, 1927. 



On Plate II (1) De Geer attempts to correlate: (1) The Swedish 

 curve, years 5540 to 5029, with portions of Reeds' Hackensack curve 

 and Antevs' Connecticut Valley curve; (2) through years 5029 to 

 5684 he arranges the Hackensack and Connecticut Valley curves 

 in juxtaposition; (3) for years 5000 to 5892 he places Hackensack 

 varves opposite Scanian (southern Swedish) varves; (4) for years 

 5736 to 5809 he diagrams varves from Dutchess Junction, N. Y., 

 opposite those from Hackensack and Scania. 



The objections to the aforementioned teleconnections are as 

 follows : 



A. The agreement between the various graphs is apparently not 

 close enougii to suggest a correlation. 



B. The direction of ice retreat from the terminal moraine on 

 Staten Island and Long Island was northward up the Hackensack, 

 Hudson, Connecticut, and smaller river valleys. As the ice border, 

 which extended in a general east-west direction, retreated northward 

 at a rate somewhat less than 100 feet per year, clays were de- 

 posited in marginal glacial lakes in the separate river basins. The 

 Hackensack Valley clays at Little Ferry, N. J., are the southernmost 

 exposures and consequently the oldest known in eastern North 

 America. Twenty-five miles to the north of Little Ferry, and in 

 direct line of retreat of the ice, is Haverstraw, N. Y., with the oldest 

 clays appearing in the Hudson River Basin. The Hackensack and 

 Hudson River Basins, although parallel for nearly 50 miles, are 

 separated by the Palisade Ridge and Verdriderger Hook. Fifteen 

 miles to the north of Haverstraw, in direct line of retreat of the ice 

 and on the north side of the Highlands of the Hudson, is Dutchess 

 Junction, N. Y. 



C. Antevs (1928) has definitely correlated the Haverstraw clays 

 with those in the Quinnipiac River Basin at New Haven, Conn. 

 Likewise, he correlates the Dutchess Junction clays with Hartford, 

 Conn., clays. The 2,500 varves in the Hackensack clays, being older 

 than the Haverstraw clays, are not considered in the correlations 

 which Antevs has established between Connecticut localities and the 

 Hudson River Valley. 



D. The geographical position of the localities in New Jersey, New 

 York, and Connecticut with reference to ice retreat and deposition, 

 as well as the correlations which Antevs has established in eastern 



