118 EFFECTS OF WINDS AND OF 



1897. This table is based on the well-known laws of the rate of propagation 

 of a tidal wave, or, in general, of a wave in water of which the depth is small 

 in comparison with the length of the wave. In that case, as shown in the 

 table, the rate of propagation of the wave is purely a function of the depth 

 of the water. For a statement of the theory involved, especially as applied 

 to seiches, consult page 348, in the volume referred to, in the Manual of 

 Tides, part 7, by R. A. Harris. 



It was known to the investigator, also on the basis of Dr. Harris's tidal 

 theory, that in a lengthwise oscillation of the lake as a whole the end limits 

 between which the oscillation primarily takes place would probably lie not 

 at the very ends of the lake, but at the point near each end at which the 

 depths begin to change rapidly as the shore is approached. 



The mean depth at each part of the lake to be used in computing the rate 

 of propagation of the wave was easily derivable from the estimates of depths 

 already made in connection with the computation of S# for Lake Erie, as 

 described on pages 44-48. 



It was found that the theoretical period of oscillation of Lake Erie length- 

 wise is 13.1 hours, provided the two end limits of the primary oscillation are 

 at the west end at a meridian about 7,000 feet east of Cedar Point (near 

 Sandusky), and at the east end about 49,000 feet to the westward of the 

 extreme eastern shore, in 60 feet of water, in the meridian which is about 

 midway between Sturgeon Point and Windmill Point. At the eastern 

 limit named the water begins to shoal rapidly. The western limit named is 

 at the entrance of the shallow western extension of the lake, in which the 

 depths are everywhere 6 fathoms or less. These eastern and western limits 

 are shown by three stars on plate 2. In the main portion of Lake Erie 

 (see plate 2) the depths are uniform over the greater portion at 12 to 14 

 fathoms. There is a relatively small portion in the eastern quarter of the 

 lake in which the depths are from 20 to 34 fathoms. 



The evidence is conclusive that the prevailing long-period seiche at both 

 Buffalo and Cleveland is a lengthwise oscillation of Lake Erie between the 

 limits stated in the preceding paragraph. This seiche is of the type which 

 might be called a wash-basin oscillation, in which there is a nodal line which 

 changes but little, if any, in elevation. The water surface falls on one side 

 of the nodal line simultaneously with the rise on the other side. 



By a study similar to that indicated above in connection with the 13.1 

 hour seiche, it appears to be probable that the primary oscillation of the 

 3.7 hour seiche observed at Buffalo is a lengthwise oscillation in that deep 

 part of Lake Erie already referred to as containing depths of 20 to 34 

 fathoms, in sharp contrast with the greater part of the lake, which has a very 

 flat bottom in depths of 12 to 14 fathoms. The computation indicated the 

 eastern limit of the 3.7 hour oscillation to be at the same point (marked by 

 three stars on plate 2) as the eastern limit of the 13.1 hour oscillation. The 

 western limit was indicated by the computation to be at the locality marked 

 by two stars on plate 2, in the most constricted part, as to width, of 



