Seiches in Some Lakes of Japan. 'J 



periodic terms in a given function is not yet known. We are there- 

 fore compelled in such cases to use some tentative method. Now 

 when we compare many limnograms of a station Avith one another 

 and also with those of other stations, we find that it happens very 

 often that as only one kind of oscillation is developed, or as other 

 oscillations, though present, j^et have very small amplitudes, the 

 limnograms are extraordinarily simple. Collecting such simple 



^*-<"''''%W'*''''*'"X.^M^^ 



TJmijiri, Aug. 29, 1901. 



30 



40 



50 



lô>' 10 



Fig. 3 a. 



30 



40 



Hakoné, Aug. 23, 1901. 



10 



20 



30^ 



40 



50 l?h 



Fig. 3 b. 



10 



20 



30"- 



40 



Hakoné, Aug. 30, 1901. 



30^ 40 50 S'' 10 20 30m 



Fig. 3 c. 



i/o 



50 



curves, we can at once deduce pretty accurate values of the several 

 periods. Guided by this knowledge, when we examine other limno- 

 grams, where several oscillations are superposed, we can still discern 

 the existence or nonexistence of a certain motion Avith an approxi- 

 mately known period. An example will make clear what we 

 mean. The accompanying curves are some limnograms taken 

 at Hakoné lake in which a is a curve obtained at the Umijiri 



