226 GROVE KARL GILBERT— DAVIS !MeM0IKS [y^xxi l 



able." These are: (1) That the whole region moves together without internal warping and 

 (2) that the direction of its present tilting is identical with the direction of the total change 

 since the epoch of the Nipissing outlet of the upper lakes. What we know of the general char- 

 acter of earth movements gives no warrant for such assumptions of uniformity, but no better 

 assumptions as to this region are now available. On under the law of probabilities the close 

 agreement of four measurements, three of which are wholly independent, gives a good status 

 to their mean, but there are other considerations tending to weaken this status. The probable 

 errors of the individual measurements are rather high, ranging from 14 to 50 per cent, and this 

 suggests the possibility that the closeness of their correspondence may be accidental. Several 

 causes of small errors arc then enumerated. 



For all these reasons I am disposed to ascribe only a low order of precision to the deduced rate of change, 

 and to regard it as indicating the order of magnitude rather than the actual magnitude of the differential 

 movement. 



On reading these passages and seeing how clearly the admirable quality of scientific candor 

 is expressed in them, one must wonder that it is ever absent from geological essays; and yet 

 how few geologists there are who, on reviewing their own writings, will not find that various 

 half-conceived uncertainties have occasionally been left in the background instead of being 

 frankly brought into the foreground of their conclusions. 



THE FUTURE DISCHARGE OF THE GREAT LAKES AT CHICAGO 



No element of Gilbert's study is neater than one that is graphically presented on an out- 

 line map of the Great Lakes, in which lines parallel to the general course of the isobases are drawn 

 through each of the lake outlets; for on the northeast of these lines the continuation of the 

 tilting should cause the lake shores to emerge; whde on the southwest it should cause them to 

 be submerged. Certain facts recorded in the paper regarding the drowning of the mouths of 

 certain creeks where submergence is inferred give support to this interpretation; but its full 

 confirmation rests with the future. Then follows, "under tho stated assumption as to rate of 

 tilting," a specific prediction of expectable future changes which has gained wide currency: 



Eventually, unless a dam is erected to prevent, Lake Michigan will again overflow to the Illinois River, its 

 discharge occupying the channel carved t>3' the outlet of a Pleistocene glacial lake. The summit in that channel 

 is now 8 feet above the mean level of the lake, and the time before it will be overtopped . . . may be computed. 

 Evidently the first water to overflow will be that of some high stage of the lake, and the discharge may at first 

 be intermittent. Such high-water discharge will occur in 500 or 600 years. For the mean lake stage such dis- 

 charge will begin in about 1,000 years, and after 1,500 years there will be no interruption. In about 2,000 years 

 the Illinois River and the Niagara will carry equal portions of the surplus water of the Great Lakes. In 2,500 

 years the discharge of Niagara will be intermittent, failing at low stages of the lake, and in 3,500 years there will 

 be no Niagara. The basin of Lake Erie will then be tributary to Lake Huron, the current being reversed in 

 the Detroit and St. Clair channels. 



This statement has been often quoted, and deservedly so; but the careful statement in 

 the original essay that the prediction is based on " the stated assumption as to the rate of tilting" 

 has been very commonly overlooked. The prediction is plainly the most interesting part of 

 the whole study; but the most valuable part, much the most valuable part, is the specific 

 statement of the assumption on which the prediction is made. 



It should be here noted that in the popular statement of recent changes in the lake-shore 

 lines, above cited, as published in 1897, Gilbert ascribed the prediction of the possible future 

 outflow of tho Great Lakes at Chicago and the resulting extinction of Niagara to another 

 observer in 1894, although he had himself very clearly indicated the possibility of this geo- 

 graphical revolution in an article of his own in 1888, referred to in an earlier section. His 

 statement then was: 



Had the oscillation [of the land] received no check, our hydrography and avenues of commerce might have 

 been very different; a further tilting of the laud to the extent of three inches in each mile would send a great 

 river from Chicago to the Mississippi, reverse the current in the Detroit, stop Niagara Falls, and rob the upper 

 St. Lawrence of seven-eighths of its water. 



