104 EFFECTS OF WINDS AND OF 



In table No. 25 the observed elevation for any day was obtained by 

 merely taking a mean of the three observed elevations for that day as re- 

 corded in table No. 21 for Milwaukee, in table No. 22 for Harbor Beach, 

 and in table No. 23 for Mackinaw. 



In table No. 25 the corrected elevation for any day is the weighted mean of 

 the three corrected elevations as given in tables No. 21, 22, and 23, the rela- 

 tive weights assigned being 1.9 for Milwaukee, 2.9 for Harbor Beach, and 5.1 

 for Mackinaw. 



EXPLANATION OF PLATES 7 to 13. 



Plates 7 to 13, inclusive, show graphically a considerable part of the in- 

 formation which is given in numerical form in tables Nos. 19 to 25, inclusive. 

 These graphs are especially valuable as a means of showing the increase in ac- 

 curacy which has been secured by applying corrections for wind effects and 

 barometric effects. The desired values are the elevations of the mean sur- 

 face of the whole of a lake from day to day. Consider the evidence in these 

 graphs that the corrected elevations are much closer approximations to the 

 desired elevations of the mean surface of the whole lake from day to day 

 than are the directly observed elevations. 



Plates 7, 8, 9, and 10 combined show continuous graphs for Lake Erie 

 from June 1, 1910, to October 31, 1910. The graphs on the lower half of 

 each of the plates 7, 8, and 9 are continuations of those on the upper half of 

 the same plate, and are continued, in turn, by those on the upper half of 

 the next plate. Plate 10 is a continuation from plate 9 on the same scale. 

 On account of the extreme fluctuations in observed elevations in the last 

 10 days of October, it was necessary to use the whole of plate 10 for a 

 single set of graphs. 



Similarly, plates 11 to 13 combined show continuous graphs for Lake 

 Michigan-Huron from June 1, 1911, to September 30, 1911. 



Throughout these plates 7 to 13 the elevations directly observed at each 

 gage, or means of such elevations from the two or three gages on each lake, 

 are indicated by a dash-and-dot line a long dash and a dot alternately. 



The corrected elevations from each gage, from the weighted mean of the 

 two or three gages on a lake, are indicated by a continuous solid line. 



The difference between the dash-and-dot line and the continuous line has 

 been produced by the application of corrections for wind effects and baro- 

 metric effects. The question on which it is desired to concentrate attention 

 at present is, How much more accurately in each case does the continuous 

 line represent the fluctuation in elevation of the mean lake surface than the 

 dash-and-dot line? 



The figures showing elevations, at the right-hand margin of each plate, 

 are feet and refer to mean sea-level as a datum. 



The computed fluctuation in elevation of the mean lake surface, in each 

 case, as produced by the rainfall on the lake surface, the inflow into the lake 

 from the next lake above in the chain of Great Lakes, and the outflow to the 



