BAROMETRIC PRESSURES ON THE GREAT LAKES 113 



general. The final decision was to use groups of five days each as being the 

 best available compromise between the two difficulties. The end residuals 

 of each five-day group include, in general, the actual fluctuation in mean 

 lake elevation for two days only. 



With the five-day groups, and on the basis of the estimate stated in the 

 second paragraph before this, it appears that of the end residuals of each 

 group 0.01 foot is usually due to actual fluctuation in the elevation of the 

 mean surface of the lake. In a few cases as much as 0.04 foot or more of 

 such end residuals may be due to such fluctuation. 



From the considerations which have been indicated briefly above, the 

 writer estimates that the probable errors shown in the last column of table 

 No. 28 are appreciably too large to represent the true degree of accuracy of 

 the corrected elevations. It is possible that the probable error of a cor- 

 rected elevation for one day for the mean surface of the whole of Lake 

 Michigan-Huron, as derived from observations at the Mackinaw gage, is as 

 small as 0.010 foot, instead of the value 0.016 foot, as shown in table 

 No. 28. Much better determinations of these probable errors will become 

 available when the investigation of evaporation is made from this data. 

 Until that time, one must be content with the approximate estimate of this 

 paragraph and the positive knowledge that the values in the last column of 

 table No. 28 are too large to represent the truth. 



TIDES. 



The question may properly be raised, Do the true tides produced in the 

 Great Lakes by the moon and sun have any appreciable effect upon this 

 investigation? No account has been taken of such tides anywhere in this 

 investigation. The true tide at Chicago and at Milwaukee produces a total 

 range of oscillation of 0.14 foot or less. (See Coast and Geodetic Survey 

 Report for 1907, pages 483-486, in the Manual of Tides, by R. A. Harris.) 

 This tidal oscillation is made up of several components, each approximately 

 a sine wave, and each with a period which is either approximately 12 solar 

 hours or 12 lunar hours, or 24 solar hours or 24 lunar hours. Such an oscil- 

 lation has practically no effect on the mean for each day. Certainly it does 

 not affect it by as much as 0.01 foot in any case. At other stations on the 

 Great Lakes the tide is, as a rule, smaller than at Chicago and Milwaukee, 

 and is nowhere much larger. The true tides produce an appreciable effect 

 on the hourly elevations of the water surface such as were used in determin- 

 ing the wind effects, reaching a maximum probably not greater than 0.10 

 foot for any hour. But the determination of wind effects is essentially based 

 on the rate of change of elevation from hour to hour. That rate of change 

 is so slightly affected by the true tides and enters so nearly as an accidental 

 error, positive and negative with almost exactly the same frequency, that 

 the final conclusion as to wind effects is certainly not affected appreciably 

 by the tides. 



