Woodworth.J 206 [Feb. 7, 
in the other portions of wide valleys, such as that east of Bristol, 
Conn., is further evidence that typical eskers were not formed by 
streams originating si)oradically on the margin of the glacier dur- 
ing the period of stagnation. That streams originated on the ice 
and became superimposed on the subjacent terrane in north- 
eastern Iowa has recently been shown by McGee.^ Similar drain- 
age lines arose on the margin of the ice in Plymouth County, 
Mass., and have determined the broad flat drift ridges and plains, 
lying between N. E. and S. W. lines of ponds and streams at the 
head of Buzzard's Bay,^ but here the i)resent drainage is in 
the depressions held open by the ice-divides of the original glacial 
drainage. In no part apparently of this glacial field did 
typical eskers originate independently of channels exhibiting a 
control exerted by an active ice-sheet. The comparative rates of 
construction in glacial sand plains made out by Professor Davis^ 
support this view, viz., that the period of stagnant ice in southern 
New England was brief. 
There are minor irregularities in the course of eskers which as 
yet baffle explanation. It remains to discuss a few examples of 
meander in which there appears something like a rule of occur- 
rence. Noting the Auburndale esker as a type, this deposit a 
short distance back from the head of the sand-plain described by 
Professor Davis, turns abruptly ; thence it curves gracefully, 
convexly to the east for about half a mile, and the remainder of 
its course is somewhat tortuous. These changes of direction 
may be diagnosed as follows. 
The Lines of an Esker. 
In the Auburndale esker, maxima of change in the elevation of 
the crest-line correspond with maxima of change in course. In 
the accompanying diagram, the line A B in horizontal projection 
Fig. 3. 
' Eleventh Annual report U. S. i^eol. si 
'■* See Plyinoiitli atlas sheet. 
^Bull. G. S. A., V. l,p. 195-20-2, 1890. 
