242 THE HISTORY OF THE NIAGARA RIVER. 



other debris brouglit by the current; but it was not completely tilled, 

 and its reniarkable pn^sent depth is one of the surviving: witnesses of 

 the shifting- drama of the Ontario. Near Fort Xiagara 12 fathoms of 

 water are shown on the charts. 



Mr. Warren lJi)ham has made a similar discovery in the basin of the 

 Red liiver of the North. That basin held a large lake, draining south- 

 ward to theMississii)pi — a lake whose association with the great glacier 

 Upham ap[)ropriately signalized b}' naming it after the aposile of" the 

 glacial theory," Louis Agassiz. The height of the old Agassiz shore 

 has been carefully measured by J\Ir. Upham, through long distances, 

 and it is found to rise continuously, though not quite uniformly, toward 

 the north. Similar discoveries have been made in the basins of Erie, 

 Huron, and Michigan, and the phenomena all belong ai)proximat<'ly to 

 the same epoch. So, while the details remain to be worked out, the 

 general fact is already established that during the epoch of the ice 

 retreat the great plain constituting the Laurentiaji basin was more 

 inclined to the north wanl than at present. 



It was shown, first in the case of Lake Agassiz, and afterward, as 

 already stated, in the case of Lake Ontario, that the change from the 

 old attitude of the land to the present attitude was in progress during 

 the epoch of the ice retreat. The land was gradually rising to the 

 north or northeast. In each lake basin the water either retreated from 

 its northern margin, so as to lay bare more land, or encroached on its 

 southern margin, or else both these changes occurred together ; and in 

 some cases we have reason to believe that the changes were so exten- 

 sive that the outlets of lakes were shifted from northerly passes to more 

 southerly passes. 



To illustrate the ett'ect of the earlier system of land slopes upon the 

 distril)ution of water in the region of the (xreat Lakes I have con- 

 structed the map in PI. iv. It does not postulate the system of levels 

 most divergent from the present system, but a system such as may 

 have existed at the point of time when the last glacial ice was melted 

 from the region. The modern system of drainage is drawn in broken 

 lines; the hypothetic system in full lines, with shading for the lake 

 areas; and a heavier broken line toward the bottom of the map marks 

 the position of the present water-parting at the southern edge of the 

 Laurentian basin. 



In the ancient system of drainage, Georgian Bay, instead of being a 

 <lependency of Lake Huron, is itself the ])rincii)al hike, and receives 

 the overtlow from IJuron. It expands toward the northeast so as to 

 include the basin of Lake Nipissing, and its discharge is across a some- 

 what low i»ass at the east end of Lake Nipissing, and thence down the 

 Ottawa River to tlie St. Lawrence. Lake Midiigan, instead of com- 

 municating with Lake Huron by a strait, forms a tributary lake, dis- 

 charging its surplus through a river. Lake Superior has tiie same 

 relations as now, but its overtlow traverses a greater distance before 



