JuiT 31, 1908] 



SCIENCE 



151 



ficulty. In its proper application there would 

 be need to take account of various qualifying 

 conditions, not all of which are readily evalu- 

 ated. Among them are: (1) The width of 

 the gorge as aifecting quantity of erosion. 

 From a gorge 1,000 feet wide twice as much 

 rock must be excavated as from a gorge 500 

 feet wide in producing a recession of the falls 

 of one foot. (2) The depth of the gorge, 

 from crest of fall to bottom of pool, as afEect- 

 ing quantity of erosion. (3) Concentration 

 of flow as affecting efficiency. For the same 

 discharge and the same hight of cataract, a 

 narrow, deep river is a different engin from a 

 broad, shallow river, and probably has a 

 higher efficiency. If, for example, the cata- 

 ract were now so broad that the depth of 

 water on its crest was nowhere greater than 

 in the American fall, the rate of recession 

 would be only that of the American falL 



(4) Thickness of the capping limestone as 

 affecting efficiency. Where the cap was rela- 

 tivly thick, the quantity of fallen fragments, 

 by serving as pestles for grinding, may have 

 promoted erosion; or, when the river was 

 small and feeble, the fragments may have 

 cumbered the way and interfered with erosion. 



(5) The relation of the Medina sandrock to 

 efficiency. When the pool hollowed by the 

 cataract reacht only to the sandrock the 

 primary erosiv attack was on the upper 

 shales ; when the cataract penetrated the sand- 

 rock the primary attack was on the lower 

 shale, and the upper shales were partly pro- 

 tected by the sandrock. The change in 

 method of erosion may have materially 

 affected the rate. Spencer's computations do 

 not include data dependent on these variables. 



The determination of the volume of the 

 river at various times involves the correlation 

 of parts of the gorge history with stages of 

 lake history in the Huron and associated 

 basins, so that the lake history constitutes an 

 essential factor. F. B. Taylor, from a study 

 of certain features about the strait connect- 

 ing Huron and Erie, inferred that the pass, 

 after having once been crost and eroded by a 

 great river, was for a time laid bare. 

 Spencer, from an independent study, infers 

 that from the time when Huron water first 



overtopt the divide it has been continuously 

 tributary to Niagara. The facts adduced by 

 Taylor (Proc. A. A. A. S., 1897, 201-2) ap- 

 pear to the reviewer demonstrativ, but space 

 can not be taken to discuss the matter. Quite 

 recently Taylor, in summarizing the results of 

 extensiv studies made in later years (Science, 

 XXVII., 725), states that the St. Clair-De- 

 troit channel is now occupied by a great river 

 for the third time instead of the second. 

 Pending the publication of his data the ques- 

 tion may be regarded as open, but if his an- 

 nouncement is sustained — that Niagara has 

 thrice instead of once carried large volume, 

 and twice instead of once run small — 

 Spencer's computation will need still further 

 reconstruction. 



The Niagara problem resembles other sci- 

 entific problems in that the enlargement of 

 knowledge leads to the recognition of com- 

 plexity. It differs from many geologic prob- 

 lems in the great extent of its available data. 

 In all the regions covered by the lakes with 

 whose changes it is concerned, those changes 

 were the latest geologic events, so that their 

 evidences overlie all earlier records. They 

 may not be so plain that "he who runs may 

 read," but they are so clear and full that the 

 patient observer can bring together a com- 

 plete, coherent, demonstrativ body of data. 

 As the facts are gradually assembled and in- 

 terpreted an intricate history is developt, a 

 history interwoven on one side with that of 

 the oscillating and waning ice-sheet, and on 

 the other with that of Niagara. The complete 

 correlation of Niagara and the establishment 

 of its chronology promis not only to tell us its 

 age, but to give fairly definit dates to various 

 events in the later Pleistocene history of 

 eastern North America, and to assist the 

 imagination in its broader conceptions of 

 geologic time. Truly the problem is not an 

 unworthy one. G. K. Gilbert 



The Dancing Mouse: A Study in Animal 

 Behavior. By Egbert M. Yerkes. New 

 York, The Macmillan Co. 1907. Pp. xxi + 

 290. 

 The comparative anatomist, the zoologist 



and the human psychologist are rafidly ac- 



