78 s. Nakamura and K. Honda. 



damped away very soon. The mode of oscillation of the first 

 motion is roughly reproduced in Fig. 20, in which we see a 

 principal motion accompanied with many secondary motions. On 

 reduction, this motion is found to correspond to a motion with a 

 period of 39.3 min. in the actual lake, which is pro])abl3'" identical 

 with the motion 37.3 min. once observed at Washizu. The second 

 motion of the model was principally a motion of the deep basin on 

 the north side of the lake, and is equivalent to a motion with a 

 period of 10-14 rain, iu the actual lake. It probably corresponds 

 to the observed seiche with a period of 12.34 min.. This view is 

 further supported by the fact that the phases at the stations Horié 

 and Osaki were found to be opposite. 



As to the effect of the tides on the lake, it produced, as already 

 mentioned, a double amplitude of about 10 cm.. In order to 

 compare the times of high and low water for the diurnal tides 

 within the lake with those in the open sea, we must calculate the 

 latter from the observations at the nearest tide-guage stations. 

 They are situated at Kushimoto and Misaki, each more than a 

 hundred miles distant from Hamana. As it is known that on the 

 Pacific coast of Japan, the co-tidal lines are nearly parallel to the 

 meildians, we may calculate the I'equired times by simple interpola- 

 tions with respect to longitude. In the following table, the times 

 of high (tl) and low (L) water iit the two stations are taken from 

 their mareograms. 



