2o8 FREDERICK G. CLAPP 



by careful study of the individual cases their differences in origin can 

 be detected. 



4. The elevation of all the plains of the system is very nearly 

 or quite uniform, at about 1 50 feet above tide. Thus, notwithstanding 

 the supposition that there can have been no large open lake in the 

 region, there was certainly some connection between the lakelets 

 adequate to maintain the water at a common level. The remnants 

 of the decaying ice-sheet consisted of stagnant valley-blocks much 

 broken up, the whole mass saturated with water, which was main- 

 tained at a uniform level consequent on some factor outside the Need- 

 ham area. Into this basin were discharged torrents of water, which 

 carried the gravels and deposited them upon, beneath, and between 

 the ice-masses. This view is satisfied both by the superglacial and 

 subglacial theories of eskers, and further corroborates the supposi- 

 tion that in this area there are examples of both classes. 



As the difference in elevation between the different plains of the 

 series can not be more than a few feet, it is evident that in the waters 

 of the lakelets themselves or below their level there can have been 

 but little current. Notwithstanding this, it is known that water- 

 worn bowlders of considerable size are abundant in many of the 

 eskers and associated deposits, indicating a very powerful current. 

 The most probable explanation, already alluded to, is that to the 

 main superglacial or subglacial rivers were tributary many side 

 streams which discharged into them from higher levels on the ice 

 or surrounding land. The extensive kames and moraine-terraces 

 are best explained, as shown by Fuller in the case of Lake Nepon- 

 set, by supposing them to have been deposited upon masses of ice 

 sometimes of considerable thickness, having a front sloping out- 

 ward beneath the lacustrine waters. 



The probable conditions existing during this stage are indicated 

 in Fig. 3, in which the lakelets are represented by the dotted por- 

 tions and the stagnant remnants of the ice-sheet by shading. The 

 regions covered by a combination of these patterns usually corre- 

 spond in position with large areas of moraine-terrace, and were at 

 this time ice-tongues overlain by considerable depths of water, in 

 which the superglacial and subglacial gravels were deposited. Gla- 

 cial streams, so far as traceable, are shown by heavy black lines. 



