354 H. H. ROBINSON 



tion which must be taken into account, namely, the gradual lowering 

 of the water-level from one stage to another. The condition is similar 

 to that described under (3), but is of greater magnitude. It necessi- 

 tates representing the water-planes as zones of such a width as will 

 include these variations in water-level. 



7. Variations in the elevations of the old shore-lines would result 

 from deformation of the water-level induced by local attraction of the 

 ice sheet. The error introduced is small, though it would have to be 

 taken into account in any consideration of the old shore-lines as a 

 whole. For small areas, presumably remote from the ice sheet, it may 

 be neglected. 



It will be seen from the foregoing that there are several noticeable 

 errors connected with the location of a water-plane with respect to 

 ancient shore-lines. A calculation, which takes account of only the 

 more important errors, gives the probable error of observation as ± i . 5 

 feet, the limit of error 10 feet. 



While these errors affect to some extent the correlation of a series 

 of ancient beaches, terraces, and related shore forms, it is desired to 

 point out here the influence they have upon conclusions, based on the 

 present attitude of the reconstructed water-planes, as to the character 

 of the crustal movements in the Great Lakes region. This may be 

 illustrated by fitting curves of different kinds to Goldthwait's very 

 complete set of observations as they are plotted on Plate XXXVII of 

 his report. 



We may consider first that portion of the profile extending from 

 Rock Island to Jacksonport, a distance of 35 miles, since this is the 

 stretch in which the several ancient water-planes have been individu- 

 ally distinguished. For the sake of simplicity a straight line may be 

 drawn through each of the series of observations representing the four 

 water-planes A,A',B, and C. The deviations of all points of observa- 

 tion in each series from its respective straight line may then be 

 measured, and Hkewise from the broken lines as drawn by Gold- 

 thwait. The results are as follows : 



Average deviation of all sure points from straight lines = 1 .4 feet. 



" " " " " broken " =1.1 " 

 Maximum deviation of ajiy sure point from straight " =4.0 



" " " " " broken " =4.0 " 



