Januaey 29, 1904.] 



SCIENCE. 



183 



of land forms, is studying a phase of 

 physiography. The two sciences are mani- 

 festly related, but they need not be con- 

 fused. For, as has been shown for sciences 

 in general, geology and geography are best 

 characterized by the relations in which 

 their topics are studied, and not by the 

 topics themselves. Both are concerned 

 with the earth and life. The whole content 

 of knowledge concerning the earth and life 

 might be shown by a cube, in which ver- 

 tical lines represented the passage of time, 

 and horizontal planes represented phe- 

 nomena considered in their areal exten- 

 sion; then if the whole mass of the cube 

 were conceived as made up of vertical lines, 

 that would suggest the geological concep- 

 tion of the whole problem; while if the 

 cube were made up of horizontal planes, 

 that would suggest its geographical aspect ; 

 and the whole series of paleogeographies, 

 horizontally stratified with respect to the 

 vertical time line, would culminate in the 

 geography of to-day. 



Objection is sometimes made to the plan 

 of geography, as here set forth, that it in- 

 volves hypotheses and theories, instead of 

 being content with matters of fact, as the 

 advocates of a more conservative method 

 in geography suppose themselves to be. 

 There is no doubt that geographical inves- 

 tigation of the kind here exposed does in- 

 volve abundant theorizing, but that is one 

 of its chief merits, for therein it adopts 

 the methods of all indiictive sciences. 

 Furthermore, as between the progressive 

 geographer, who candidly recognizes that 

 he must theorize, and the conserva- 

 tive geographer, who thinks that he 

 observes facts only and lets theories alone, 

 the chief difference is not that the first 

 one theorizes and the second does not, 

 but that the first one knows when he is 

 theorizing and takes care to separate his 

 facts and his inferences, to theorize log- 

 ically, to evaluate his results, while the 



second one theorizes unconsciously and 

 hence uncritically, and, therefore, fails to 

 separate his inferences sharply from his 

 facts, and gives little attention to the eval- 

 uation of his results. Geography has, in- 

 deed, suffered so long and so seriously 

 from the failure of geographers to culti- 

 vate the habit of theorizing as critically as 

 the habit of observing— studies of the at- 

 mosphere and the ocean still excepted, as 

 above — that a strong recommendation 

 must be given to the acquisition of the 

 methods of theoretical investigation, in 

 which deduction is an essential part, by 

 every one who proposes to call himself a 

 scientific geographer. Let me give an ex- 

 ample of the less of time that has resulted 

 from the failure of geographers to develop 

 the habit of theorizing. 



For forty years past there has been ac- 

 tive discussion as to how far land forms 

 in glaciated regions had been shaped by 

 glacial erosion, but not till within five 

 years has any geographer clearly defined 

 the deductive side of this problem. In or- 

 der to determine whether land forms are 

 carved by glacial erosion or not, two meth- 

 ods have been open: one is to observe the 

 action of existing glaciers and thus deter- 

 mine whether they are competent or not 

 to carve land forms; but this is difficult, 

 because the beds on which glaciers lie can 

 not be well examined. The other method is 

 to deduce the appropriate consequences of 

 both the affirmative and the negative sup- 

 positions, and then to confront these con- 

 sequences with the facts found in regions 

 once glaciated, and see which set of con- 

 sequences is best supported. This deduct- 

 ive method is very simple. Its application 

 involves no principle that was not per- 

 fectly well known fifty years ago, though it 

 does involve a facility in theorizing that 

 does not seem to have been familiar or 

 habitual with geographers until more re- 

 cent times. On the supposition 'that gla- 



