•92 s. Nakamura and K. Honda. 



no means be considered to be rectangular. But as the position of 

 the nodal line was found by observation, we were able to calculate 

 iihe period from data for the west half of the lake, and obtained 

 7.78 minutes for the period, taking 

 L = 2x3.20x10' cm., 

 /i= 7.70x10* cm., 

 which agrees better with the observed value. 



For the verification of the above result, a model was made in 

 ■cement on the scale of 1: 18400 and 1: 2000 for the length and 

 the depth respectively. The uninodal oscillation reproduced in 

 Plate VII Fig. 3. could only be found for it with a period 



« = 1.12 sec, 

 which gave on reduction a period of 7.70 minutes for the actual 

 lake. 



Mr. Ishitani calculated the period and the position of the node 

 for this lake using his formula above cited with the result that, 

 the period was found to equal 7.0 minutes, and the node to lie at 

 -3.3 kilometres from the west end, both agreeing pretty well with 

 4;he observed values. 



