46 S. Nakamnra and K. Honda : 



other shorter oscillations, there remains some uncertainty. We 

 must decide what U is. Plate IX shows that tliis third oscillation 

 ts is quite complex, one of the nodal lines, that lying nearest to 

 the wider end of the lake, being so much curved that we can 

 equally well say that this oscillation has three or four nodes. 

 Indeed, in this motion, w^e see a place where the figures made by 

 the grains of aluminum powder are points, and this place lies on a 

 loop for the vertical motion; but in reality this is not the case. 

 The figures made by the powder are points, not because they are 

 executing vertical motions like those placed on loops, but because 

 they are actually at rest. We see that a curved nodal line touches 

 the shore at that place, or rather that two nodal lines start in two 

 different directions from that point, which is in fact on a node. 

 We prefer to call the oscillation ts, not a trinodal, but a quaclrinodal 

 seiche, though we were unable to observe in the model any un- 

 ambiguous trinodal seiche. Now our question is, what oscillation 

 in the actual lake corresponds to t-,? If we take 15,38 minutes as 

 the value of tj in the actual oscillation, and calculate ts and U by 

 assuming the ratios for ti:t2 its given above, we get 6.38 minutes 

 for ts and 4.06 minutes for t.^, and ta seems to correspond rather 

 to T,i than to T3. Again if we calculate the actual periods corre- 

 sponding to ti, to and U by multiplying them by the scale for 

 length and dividing by the square root of the scale for depth, we 

 have 17.6, 7.3, and 4.6 minutes. Here U is exactly equal to 

 the observed period T3 and they seem to correspond to each other, 

 but at the same time we see that the values deduced from ti and 

 U much exceed T, and T2. If we, however, bear in mind that 

 'i\ was observed at Hakon<', Motohakoné and Umijiri but not at 

 Hyakkwan, Hotokegasaki and Takogawa, while T4 was observed 

 only at ITakoné, and moreover if we carefully study plate IX we 



