(1922) gave a value of 13. 1 hours. Since Hayford's 

 value was derived from 28 cases and included over- 

 lapping periods under new wind influences, it is be- 

 lieved that 14. 1 hours is nearer the true period for 

 the lake. Table 10 gives the number of complete os- 

 cillations and the corresponding period for that num- 

 ber of observations. The mean period of 14. 1 hours 

 was derived from this table. 



Table 10. --Observed periods of the primary 

 east -west oscQlation on Lake Erie 



When a prolonged westerly wind blows acceler- 

 atedly, the lake level at Buffalo tends to rise until 

 the maximum wind velocity is reached. In this sit- 

 uation the time interval of the initial rise at Buffalo 

 may greatly exceed 7 hours (half the period of os- 

 cillation); but after the wind reaches a maximum 

 velocity and begins to subside, a half period of ap- 

 proximately 7 hours soon obtains and continues until 

 distorted by a new wind impulse. 



Examination of the Port Stanley graphs (figs. 19 

 and 20), reveals that two maimKcillations are taking 



place. One is the north and south, and the other 

 the 14-hour, east -west oscillation. The computed 

 period of the north -south oscillation is about 2. 5 

 hours, and the mean of 90 observed periods from the 

 Port Stanley gauge (May and August 1929) is 2. 7 

 hours. Hayford gives a period of 2. 6 hours for the 

 north-south oscillation at Cleveland. 



The east -west 14-hour oscillation is clearly evi- 

 denced on the Port Stanley graphs and has an ampli- 

 tude of approximately one -third that of the Buffalo 

 curve. The most peculiar feature of the curves at 

 Port Stanley is the apparent lag of 3 hours in this os- 

 cillation from the Buffalo graph. The lag is not only 

 apparent in the two figures submitted, but is found to 

 be constantly present in comparisons of the available 

 graphs for the two places. 



If the oscillation is "stationary, " as the graphs 

 of Buffalo and Put -in-Bay most assuredly indicate, 

 one might expect the maxima and minima at Port 

 Stanley to occur simultaneously with the maxima and 

 minima at either Buffalo or Put -in-Bay, depending 

 on whether the nodal line of this oscillation is east or 

 west of Port Stanley. Additional graphs of the lake 

 level at stations some 30 miles on either side of Port 

 Stanley would undoubtedly enable us to fix the posi- 

 tion of the nodal line and to arrive at the source of 

 the 3 -hour lag. In the absence of such observations, 

 three possible reasons for the lag are suggested. First, 

 the 14-hour period at Port Stanley may be from a dif- 

 ferent seiche area which happens to have the same 

 period as the primary east -west oscillation. An ex- 

 amination of the chart, however, fails to disclose the 

 presence of any such area. Furthermore, the lag is 

 always approximately 3 hours, regardless of whether 

 the time interval between maxima happens to be 

 greater or less than the mean period of 14. 1 hours. 

 We may assume, then, that the Port Stanley oscil- 

 lation is definitely associated if not identical with the 

 main east -west oscillation. Second, since the vol- 

 ume of water is alternately increased and decreased 

 on each side of the nodal line, there is an actual 

 periodic east and west movement of water back and 

 forth across the nodal line. During the half period 

 when the level is rising at Buffalo, the water is in 

 eastward movement across the nodal line. Now if 

 inertia causes a lag in this water movement, or, in 

 other words, if it continues eastward for some time 

 after the maximum elevation is reached at Buffalo, 

 then the average level of the waters east of the nodal 



67 



