420 GEOLOGY. 



which it is discharged determines its erosive effects. If it is poured 

 rapidly through its outlet, the effects are proportionately much greater 

 than if it be discharged equably throughout the whole season of pre- 

 cipitation. The headwater area of the Mississippi is particularly 

 affected by lakes, ponds, marshes, ill-drained flats, tortuous streams, 

 and other topographic features that even now greatly interfere with 

 the rapidity of discharge of the precipitation of the region. Since 

 the cutting began, the drainage lines have been deepened, widened, and 

 extended in the natural course of things, and the facilities for dis- 

 charge have been constantly improved. Presumably, therefore, there 

 has been a very appreciable increase in the rate of discharge of the 

 waters since the ice retreated, even without such aid as recent settle- 

 ment has brought. It follows that the effectiveness of erosion has 

 increased. It is the very latest rate of erosion that was determined 

 and used in the above calculation. The 8000 years should perhaps 

 be increased to 12,000 or 16,000 years. 



It will be seen therefore that even in these cases of best data, there 

 are serious sources of qualification, and that these qualifications may, 

 in the judgment of experienced geologists, affect the results to the 

 extent of several hundred per cent. If the range of the estimates 

 of the Niagara be placed at 10,000 to 30,000 years, and if this be 

 added to the range of estimates for the time of retreat of the ice 

 before the falls came into existence, also 10,000 to 30,000, the result 

 is 20,000 to 60,000 years for the time since the Late Wisconsin ice- 

 sheet began to retreat. If the estimates for the St. Anthony gorge-cut- 

 ting be placed at 8000 to 16,000 years, and the estimates for retreat be 

 added, the range of estimates for the time since the beginning of the 

 Late Wisconsin ice retreat is 20,000 to 56,000 years. These may be 

 taken for a rough, wide-ranging estimate, such as it is, of the time 

 since the climax of the Late Wisconsin ice invasion. Now, using the 

 estimates in the table of relative duration above, and remembering 

 that we are multiplying the errors of the previous estimates, we reach 

 the following dates for the climaxes of the several ice invasions: 



Climax of the Late Wisconsin. 20,000 to 60,000 years ago. 



1 ' " Early Wisconsin 40,000 to 150,000 " 



" " Iowan 60,000 to 300,000 " 



" " Ulinoian 140,000 to 540,000 " " 



" " Kansan 300,000 to 1,020,000 " " 



" " Sub-Aftonian y to z " " 



