306 FKOF. AV. MOEEIS DAVIS ON [Aug. I909, 



(/) Deduction of the consequences of glaciation, 

 on the assumption that glaciers are eroding agents. — 

 In order to learn whether the features peculiar to glaciated 

 mountains may he reasonably ascribed to glacial erosion, it is 

 legitimate to make the provisional assumptions that glaciers do 

 erode and that glaciers do not erode; then to deduce the conse- 

 quences of these assumptions as carefully as possible, and finally, 

 to confront the two sets of consequences with the appropriate facts 

 of observation. If an essential agreement is found between one 

 set of consequences and the facts, while the other set of conse- 

 quences correspondingly fails to meet the facts, it may be fairly 

 concluded that the assumption which led to the successful con- 

 sequences is correct, particularly if the two sets of consequences 

 involve peculiar and unlike features. This method is employed in 

 the following pages. 



(g) Search for other explanations of features that 

 are ascribed to glacial erosion. — Caution requires that all 

 additional explanations, which have been suggested to account for 

 the peculiar features of glaciated mountains, should be critically 

 examined, to see whether they also are capable of producing such 

 features. Among these explanations are : warping of normal 

 valleys to produce lakes ; faulting to produce hanging lateral 

 valleys ; revival of normal erosion by tilting, to produce hanging 

 valleys (already considered) ; subglacial stream-erosion, or ordinary 

 stream-erosion during interglacial periods, to produce over-deepened 

 valleys ; and so on. Several of these suggested explanations are 

 • considered on later pages. 



The correct application of the foregoing methods ought to lead to 

 accordant results : yet so incomplete are our observations and so 

 divergent are our lines of discussion, that the results obtained by 

 different investigators are notoriously unlike. The best safeguards 

 against error are : a broad basis of observation, a careful develop- 

 ment of inductions, a friendly consideration of every hypothesis, a 

 logical deduction of all possible consequences from each hypothesis, 

 a critical comparison of these consequences with the facts of obser- 

 vation, and a judicial decision in view of all the evidence presented. 

 The impartial consideration of rival hypotheses is of especial im- 

 portance. For this reason an effort is made in the following pages 

 to discuss with equal thoroughness the opposed hypotheses that 

 glaciers are effective eroding agents, and that glaciers are chiefly 

 protective agents. The essential consequences of each hypothesis 

 are deduced in some detail, and then with as little prejudice as 

 possible confronted with the facts. There appears to be no safer 

 method of procedure than this : it is the soundest method that 

 can be employed by anyone who wishes to reach an independent 

 opinion on this mooted question. 



The first three of the above-named methods will not be further 

 considered here. The fourth and fifth methods — restoration of 



