36 



s. Nakamnra and K. ITonda : 



General results. 



If we collect al] the values of the periods obtained at the 

 several stations we get about 180 different values; but of course 

 they can not in reality be all different from one another. Plotting 

 the number of times these motions were observed as ordinates and 

 the corresponding periods as abscissae, we get a curve consisting 

 of many maxima much resembling so many probability curves. 

 If we calculate the positions of these maxima, we have the follow- 

 ing values, arranged in the order of the number of observed times. 



TABLE 22. 



(1) 9.4 lûinutes. 



(6) 31.9 minutes. 



(2) 30.0 



(7) 12.0 



(3) 8.5 



(8) 10.6 



(4) 25.2 



(9) 16.7 



• (5) 22.7 



(10) 72.6 



These are then the most frequently observed periods in Lake 

 Biwa. We say the most frequently observed periods and not the 

 most frequently occurring periods; because a particular motion is 

 excited only under a condition favorable to it; it may be present 

 one day, but not on the next day. Thus at Hikoné, we had not 

 the oO-minute motion during the first series, but on the contrary 

 only the 10-minute motion was present during it. In order to find 

 out all modes of motion peculiar to a given place, Ave must 

 sufficiently extend the interval of observation. In many places, 

 our records were taken for only one or two days, and we could not 

 have had the chance of recording some motions proper to the 

 places. We must also bear in mind that a seiche with a few 



