BAROMETRIC PRESSURES ON THE GREAT LAKES 71 



Huron under the influence of barometric changes is about its own center of 

 gravity as a nodal point, the rise and fall at Mackinaw gage would be much 

 greater under fluctuation of E-W barometric gradients than if the oscilla- 

 tion were about the center of gravity of the combined lakes Lake Michigan- 

 Huron as assumed in the computations. The extremely large value of 

 P v at the Mackinaw gage, +21.17, is believed, therefore, to be a strong 

 indication of a throttling effect in the Strait of Mackinac. 



The negative value of P n at Harbor Beach, 0.32, is believed to be due to 

 a peculiar effect of the Strait of Mackinac arising from the fact that it is the 

 only connection between Lake Michigan and Lake Huron, and that said 

 connection is almost at the extreme northern end of each lake. The fact 

 that B n0 and B n \ at Harbor Beach (see table No. 6, page 32) have apparently 

 abnormal negative signs, while B n 2 and Z? B3 have the positive signs, which 

 would be normal according to the derivation of the formula used in this in- 

 vestigation, is also believed to be due to the same cause. 



No adequate explanation is offered for the apparent^ abnormal value of 

 P v at Cleveland, 2.83. It is surmised that it is related in some way to the 

 following three facts: (1) the center of gravity of Lake Erie, which normally 

 should be the nodal point for barometric effects, is not far to the eastward of 

 Cleveland; (2) the nodal line for east or west winds on Lake Erie is to the 

 westward of Cleveland ; and (3) there is a decided bend in the south shore of 

 Lake Erie at Cleveland. Note that the minus sign for P w at Cleveland 

 indicates that the actual nodal point for barometric effects on Lake Erie is 

 to the westward of Cleveland. 



The remaining seven values of P w and P n on which no special comments 

 have been made are all positive, their mean value is +3.23, and the largest 

 two values are +9.08 and +4.89 for P n at Buffalo and Cleveland, respec- 

 tively. Inertia tends strongly to make P v and P n greater than +1.00. 

 Other evidence, set forth later in this publication, indicates that inertia 

 effects upon the elevation of the water surface are probably much greater 

 on Lake Erie than on Lake Michigan-Huron. The effects of errors in 

 assumption No. 2 (see page 16) are probably greater for N-S effects on Lake 

 Erie than in any other case in this investigation, since the ratio of the dis- 

 tance from point 5 to point 7 (see plate 2) to the maximum extent of Lake 

 Erie in the N-S direction is greater than any other similar ratio involved in 

 assumption No. 2. Note that for the reason stated on page 16 the de- 

 parture of the facts from assumption No. 2 is likely to be such as to make the 

 values of P w and P n considerably greater than +1.00. For the reasons 

 indicated briefly in this paragraph, it is believed that the seven values of P w 

 and P n here commented upon have in them no indications of inaccuracy of 

 observation or computation or unreliability of the theory on which the 

 computations were made. 



Taking into account all ten values of P w and P n , the comments made 

 upon them above, and other consideration of details not here set forth, the 

 general conclusions reached are as follows: 



