THE PLEISTOCENE OR GLACIAL PERIOD. 419 



over the Niagara Falls. Now the ordinary rate of erosion is measured 

 by a high power of the volume, when it induces an accelerated velocity 

 (Vol. I, pp. 115 to 123). Precisely how this general law is modified 

 in the case of falls is not known by direct experiment, but it may be 

 inferred from the phenomena of the falls under consideration. Since 

 the Horseshoe and American Falls separated, the latter has retired but 

 slightly from the position it occupied at the time of separation, while 

 the Horseshoe Fall has retired about ten times as far. With little 

 doubt this is due almost wholly to the superior volume of water poured 

 over the latter. This is further indicated by the form of the Horse- 

 shoe, since the volume per unit breadth is greater in the center than 

 on the sides. It is also shown by the recent extraordinarily rapid 

 recession at a point where the volume is exceptional. 



In view of these considerations, Gilbert, Taylor, and Spencer have 

 urged that the cutting of the narrower portions of the gorge was prob- 

 ably the work of the relatively limited volume of water from the Erie 

 basin, and that the recession proceeded at a relatively slow rate on 

 this account, while the recession has been much accelerated since the 

 upper lakes joined their greater volume to that issuing from Lake Erie. 

 It is this accelerated rate that is used as the divisor in the simple com- 

 putation that gives 7000 years. In view of the probable rate of in- 

 crease of recession of the fall, due to increase in the volume of the 

 river after the drainage of the upper lakes was diverted to it, it is 

 thought that the simple quotient 7000 is to be multiplied several times 

 to give the true time-estimate. Spencer places the period at 31,000 or 

 32,000 years, and Taylor at 50,000 years as an approximate maximum. 

 There are, however, those who do not accept these qualifications and 

 who take appeal to other phenomena that cannot here be discussed. 

 The estimate of Upham, 7000 years, and that of Wright, 10,000 years, 

 are representative of this class. The mean of all the above estimates 

 is about 25,000 years. 



From a comparison of the earlier and later surveys of St. Anthony 

 Falls, N. H. Winchell estimates the time of recession from the mouth of 

 the gorge to be about 8000 years. The chief qualification that affects the 

 rate of recession in this case seems to be the rapidity with which the 

 precipitation upon the catchment area above the falls was discharged. 

 This is but another application of the principle involved in the pre- 

 ceding case, for, given a certain amount of precipitation, the rate at 



