84 Art 9.-T. Terada: 



thanks to Prof. A. Tanakadate, under whose supervision the 

 entire work was carried out, for his kind guidance throughout 

 the course of the investigations. Most cordial thanks are also 

 due to Prof. H. Nagaoka and to Prof. S. Sano, for the interest 

 shown in the investigations and for many valuable suggestions 

 and instructions given. Last, but not least, we feel very much 

 indebted to Prof. I. Ijima, the Director of the Marine Biological 

 Laboratory, for his kindness and generosity in affording lodging 

 and many other conveniences to the observers resident in 

 Misaki, for which the best thanks of all the participators in the 

 present work are due. 



Summary of the Results of Investgations. 



1. The period of tlie magnetic pulsations has no sharply 

 defined value, varying from about 20 seconds to nearly 1 hour. 

 Nor is it exactly constant even in a coherent train. 



2. During the day time, waves of 0*5-1 minute periods 

 predominate, whereas during the night hours longer periods 

 l'D-2'5 minutes are most frequent. 



3. A periodicity of 25-30 days is suspected in the daily 

 frequency of the pulsations. 



4. The vertical component of the waves is a reduced 

 reproduction of the NS-component except the phase retardation. 

 The shorter the period, the more remarkable is the reduction 

 of the amplitude as well as the phase retardation. 



5. The azimuth of the linearly pulsating magnetic field 

 undergoes a remarkable diurnal variation, showing maximum 

 deviations from the meridian a few hours before noon and 

 midnight. 



6. The disturbing field is generally more or less rotatory. 

 The sense of rotation shows a semidiurnal variation. The clock- 

 wise rotation is most frequent duiing the hours between sunrise 

 and noon, as well as between sunset and midnight, wdiile the 

 opposite rotation falls most frequently in the remaining hours. 



