GLACIAL LAMS. LAKES, AXD WATERFALLS. 397 



of the lakes can not well be doubted, for they are not related 

 to any great natural lines of drainage, but follow the wind- 

 ings of a definite level, receding from the lake wherever 

 there is a transverse valley, and forming in some cases parallel 

 embankments on either side of such valley, running inland 

 as far as to the general level of the series, and then return- 

 ing on itself upon the other side, to strike off again par- 

 allel with the shore at the same level. Their relation to the 

 lake is also shown by the local character of the material. It is 

 usually such as would wash up on the shore out of the rock in 

 place. In the sandstone region the ridges are largely made 

 up of sand, mingled with fragments from the general glacial 

 deposit. Over the regions of outcropping shales the ridges 

 are composed largely of the harder nodules which have suc- 

 cessfully resisted the attrition of the waves, Other evidences 

 that they are shore-deposits are their stratification, the rela- 

 tive steepness of their sides toward the lake, and the frequent 

 occurrence of fragments of wood buried at greater or less 

 depths on their outer margin. 



It need not be said that there has been much speculation 

 concerning the cause which maintained the waters of the 

 lakes at the levels indicated by these ridges, and permitted 

 them to fall from the level of one to that of another in suc- 

 cessive stages, so suddenly as they seem to have done ; for, 

 from the absence of intermediate deposits, it is evident that 

 the formation of one ridge had no sooner been completed, 

 than the one at the next lower level began to form, In the 

 earlier stages of glacial investigation, before the full power 

 and flexibility of glacial ice were appreciated, and before 

 the exact course of the southern boundary of the ice-sheet 

 was known, the elevation of the water to produce these 

 ridges was supposed to have resulted either from a general 

 subsidence of the whole region to the ocean-level, or from 

 the elevation of a rocky barrier across the outlet. Both 

 these theories were attended with insuperable difficulties. 

 In the first place, there is no such amount of collateral evi- 

 dence to support the theory of general subsidence as there 



