318 



ANNUAL REPORT SMITfiSONIAN iNStlTtJTlON, 19 31 



-~=Tt-lJ3tni) 2'/S"(6-3cm) 



Z■n■^6l.m) 8 "CZOcm) 



^ 



S 



M 



s 



5 



S 



ai ci 





wise favorable conditions, about the same everywhere in the area in 

 which the total summer heat underwent similar yearly fluctuations. 



Varve graphs from the same region, 

 therefore, show essentially the same 

 fluctuations and can be matched, and 

 the separate varves identified (fig. 4) . 

 Owing to the imbricated position of 

 the varves, the rate of recession of the 

 ice edge can be determined (fig. 2). 

 Thus, if the bottom varve at one local- 

 ity is found to correspond to varve 

 number 21 at another locality, the lat- 

 ter place was uncovered 20 years 

 earlier than the former. By a series 

 of varve measurements in the direction 

 of the ice retreat, the rate of uncover- 

 ing and the time involved can be de- 

 termined. By measurements distrib- 

 uted over an area the outline of the 

 ice border can be mapped. This is De 

 Geer's method of study. 



Graphs of clay varves may be cor- 

 related, if they derive from regions 

 that experienced similar yearly varia- 

 tions of the total summer heat and 

 were released from the ice at about 

 the same time. These conditions con- 

 fine the possibilities of varve correla- 

 tions within somewhat narrow limits. 

 They preclude correlations of varve 

 graphs from North America and Eu- 

 rope and from distant parts of the 

 same glaciated area. The conditions 

 evidentl}' do not prevent correlations 

 of varve graphs from adjacent valleys, 

 for instance, from the Hudson, the 

 Connecticut, and the Merrimac valleys. 

 The last ice sheet of North America 

 comprised almost entire Canada and 

 the States down to a line running 

 through New York City, south of the 

 Great Lakes, and a little below the in- 

 ternational boundary in the far West. 

 During the retreat of this ice small and 





4) -Jj 



B S 



o s 



o ^ 

 ■S 00 



