26 EFFECTS OF WINDS AND OF 



OBSERVATION EQUATIONS. 

 Date 

 1910 

 Sept. 



(Aug.) 31-1 



6....+ 2Bw- 3Bwi- 



9....+ 3Bteo-36fiwi- 

 10.. 



ll....+ll-Bw> + 9Bm+ 3fiu< 2 - 3Brw+ 45- 

 12-13 ____ -225>o- 6fiwi -225^2+ Ifiws- 9Bm-12Bm- 6Bn*+llBn 3 + 44=?; 

 14 

 15 

 16 

 18 

 19 



20 ____ - 3Bw a - 3B W1 - QBw*- 4Bw+ 5Bno+ SBm + llBn t - 9fin 3 -163=ir 12 



21 ____ - 9-Btw- 4B Wl - lBwi+ QBm+llBno- 9Bm-14Bn 2 + 4Bns+118 = t;ij 



22 ____ - 1B W + 9Bwi + 16Bwi+ GBw,-14Bn a + 45n,+ 5Bn 2 + 8B 3 + 68=014 

 23... 



27 



29 ____ + 7Bm + lQBwi+ 7Bwt+ 2Bw 3 + 3Bno-l2B ni + 



30 ____ + 7Bm+ 2Bwi+ 



The basis on which rejections and combinations were made is shown later 

 in connection with the discussion of the accuracy of the computed baro- 

 metric effects. 



BAROMETRIC TERMS IN OBSERVATION EQUATIONS. 



For Lake Michigan the barometric pressures used in the computations 

 were those read at points 3, 4, 5, and 6 as shown on plate 2. The values of 

 the barometric pressures at these points at 8 a.m. and 8 p.m. (75th meridian 

 time) of each day for a part of September 1910 are shown in table No. 3. 

 The table also shows the values of the differences of barometric pressure 

 (4-5) and (3-6) which were used in the preparation of a part of the above 

 observation equations. 



In the above equations the values of b wi , b W2 , . . . b nt and b m are ex- 

 pressed in units of 0.01 inch, and the absolute term 7, the rise of the water 

 surface, is expressed in units of 0.001 foot. This arbitrary selection of 

 units was made in order that the average magnitude of quantities in the 

 various columns, as the equations are arranged in the example given, might 

 be about the same. To have them about the same increases the accuracy 

 of certain checks against error in the computations. 



The derivation of the values of b m , b m , . . . b n , and b n , as given in the 

 observation equations may be verified from the table for the period which it 

 covers. For example, note that the exceptionally large value of b wl in the 

 equation for September 9 is (4-5) at 8 a.m. of September 8, ( 17), minus 

 (4-5) at 8 p.m. of September 8, ( + 19), or 36. This is the most rapid 

 change of barometric difference shown anywhere in table No. 3. Note that 



