BAROMETRIC PRESSURES ON THE GREAT LAKES 25 



For Lake Erie: 



^=(5-7) at 8 p.m. on the day before the day preceding the current 



clay minus (5-7) at 8 a.m. on that preceding day. 

 b ni = (5-7) at 8 a.m. of the preceding day minus (5-7) at 8 p.m. of that 



day. 

 b nj = (5-7) at 8 p.m. of the preceding day minus (5-7) at 8 a.m. of the 



current day. 

 5 w = (5-7) at 8 a.m. of the current day minus (5-7) at 8 p.m. of the 



current day. 



The quantity (5-7) is the barometric pressure at the point marked 5 on 

 plate 2 minus the barometric pressure at the point marked 7 on that plate. 



For Lake Michigan-Huron, use (3-6) instead of (5-7) ; see plate 2. 



The appropriate modification for any other lake is to use in place of point 

 5 (for Lake Erie) a point to the northward of the lake, and a point to the 

 southward of the lake and in the same meridian in the place of point 7. 



BW BWV B m , B m , B no , B ni , B m , and B n3 are constants of which the values 

 are to be determined from the least-square solution. 



/ is the observed rise in the water surface, at the gage station from the 

 preceding day to the current day, corrected for wind effects if such correc- 

 tions are available, and corrected for inflow from the next lake above, out- 

 flow to the next lake below, and rainfall on the lake surface. In other words, 

 ignoring the various corrections for the moment, I is the mean elevation of 

 the water surface on the current day minus the mean elevation of the water 

 surface on the preceding day. 



The residual v in the second member of equation (40) is to be derived 

 by substitution of the computed values of B m , B m , . . . B ni , B H3 in the 

 observation equations after the solution is complete. The residual v, one for 

 any observation equation, is the discrepancy between the theory and the 

 observed facts for the particular day. 



The following set of observation equations for Milwaukee for the month 

 of September 1910 serve as a typical illustration. They are a part of the 

 equations for solution K, which included in all 186 such equations covering 

 the months June to September 1910 and June to September 1911 8 months 

 in all. 



No record of the elevation of the water surface was obtained from the 

 gage at Milwaukee on September 2, 3, and 4. 



The following possible equations were rejected: for September 2-5 (as one 

 equation), for September 7, for September 8, for September 17, for Septem- 

 ber 24, for September 25, for September 26, and for September 28. It is 

 unusual for so many rejections to occur in any one month. As shown, the 

 equations for certain days were combined in pairs to make one equation in 

 each case, namely, for August 31 and September 1 and for September 12 and 

 13. Each combined equation was obtained by adding corresponding terms 

 of the separate equations. 



