348 H. H. ROBINSON 



more probable. The result will show that both hypotheses are 

 equally probable, but as that of differential warping involves simpler 

 assumptions, it is considered preferable in our present state of knowl- 

 edge concerning the subject. 



As ideas in regard to the character of crustal movements in the 

 Great Lakes region are based upon the water-planes reconstructed 

 from observations upon the beaches, terraces, and other shore-line 

 features of the ancient Great Lakes, it seems advisable first to consider 

 the several errors that may enter into the determination of the posi- 

 tion of such water-planes. These errors have been recognized for 

 many years, and statements respecting most of them may be found in 

 various reports dealing with both the ancient and modern Great Lakes. 

 They have not, however, been collected in any one article, so far as 

 the reviewer knows. It seems appropriate, therefore, to bring them 

 together and point out the limitations they impose upon the correla- 

 tion of old shore-lines and the deductions that may be drawn from 

 reconstructed water-planes in regard to the character of the deforma- 

 tion of the Great Lakes region, especially as work in this field is now 

 being placed upon a more accurate and refined basis. 



The errors that enter into the determination of the height of the 

 ancient beaches, etc., above the present lake levels — the datum planes — 

 need only brief consideration, since with proper care they may be 

 practically eliminated. The principal ones are, perhaps, those con- 

 nected with the measurements of heights, seasonal fluctuations of the 

 lake level, and changes in level by winds and waves. The first one 

 can only be reduced to a negligible quantity by the use of a wye level, 

 the instrument adopted by Spencer, Lawson, Upham, and Goldthwait 

 in their work. The second may be eliminated by applying to the 

 altitude determinations a correction obtained from a continuous- 

 record of the fluctuation of the lake level, a correction that was applied 

 by Goldthwait. The last two errors appear rather troublesome, though 

 they may be largely eliminated by restricting work to calm days. To 

 do this, however, might often prove inconvenient, and it may be 

 assumed that an error of from i to 2 feet"^ may arise from these causes. 



The more important errors are connected with the location of 

 water-planes with respect to ancient beaches and terraces, with the 



I Goldthwait, op. cit., p. loo. 



