74 s. Nakamura and K. Honda.. 



.:(=5.75xl0"'sq. cm., 



î =5.53x10'^ c. cm., 

 we get the mean depth /' 



//=963 cm, 

 Avhich combined with the length of the lake 



L=5.86xl0'cm. 

 gives, according to Mérian's formula, a periodof 20. 1 minuts for 

 the uninodal longitudinal motion. 



A model of the lake was constructed in cement and from its 

 oscillations (Plate XIII), we got three longitudinal motions with 

 periods 



^1=1.36 sec, 



^2=0.62,, 



«3=0.46,, 

 which are in the ratios 



tr.t^-h^'i. 00:0.45:0.35. 

 The first uninodal motion ^, gives on reduction a period of 23.86 

 minutes for the actual lake, and evidently corresponds to T^. The 

 second is a binodal oscillation of the whole lake, but the amplitude 

 is large only on the Funatsu side of tlie model, the level on the 

 Nagahama side apparently remaining quite still. As the ratio «i :?.2 

 is very nearly equal to the ratio T, : T^, and as T^ was not observed 

 at Nagahama, we may perhaps conclude that the motion ^3= 10.66 

 minutes is binodal. The third motion ^3 witli a period of 0.46 sec. 

 is trinodal, but as one of its nodal lines is very much curved, and 

 touches the shore, it may be called quadrinodal. From the ratio 

 ti : ^3, it is probable that /3 corresjDonds either to T^ or to T^, but 

 owing to the shortness of their periods we can not decide which is 

 right. We are inclined however to think that T^=8.5S minutes 

 corresponds to ^3 and is trinodal. As to the motion T2=11.50 



