On Rapid Periodic Variations of Terrestrial Magnetism. ]^5 



tiens" to (listingnisli them from peculiar disturbances called 

 " spasms," and discovered a remarkable daily period of frequency 

 with a maximum near midnight. Afterwards'-" he compared the 

 materials from Zi-Ka-Wei and Kew, and found for the former 

 station a similar nightly maximum as in the case of Batavia, 

 while for the latter station, the maximum frequency was found in 

 the day time. Recent authorities seem to agree in the opinion 

 that the magnetic waves in question are chiefly due to some 

 fluctuations of the electric current existing in the upper at- 

 mosphere, though the actual modes of fluctuation still remain 

 obscure. 



Though the original purpose of our investigation was to 

 detect any abnormal disturbances associated with earthquakes, 

 it was in any case necessary to study tlie characteristic pulsatory 

 waves in some detail, even if these waves should have no direct 

 connection with earthquakes, in order to be able to distinguish 

 which were the normal and which the aljnorinal disturbances. 

 The present paper is chiefly confined to the study of these 

 characteristic pulsatory disturbances, since unfortunately no 

 positive results have yet been obtained with regard to earth- 

 quakes. 



In the following, we shall enumerate the most interesting 

 results obtained by the detailed study of the magnetograpiiic 

 records comprising observations extending over four years. 



5. «) Generally speaking, the magnetic waves in question 

 are decidedly more regular in the night than in day time when 

 waves of different periods seem to be very irregularly superposed. 

 In the great majority of cases remarkably lon^ continuous trains of 

 moderately regular waves with period of 30-60 sec. appear at 

 5^-7'' in the morning and continue up to 9^-11^, with occasional 

 interruptions (PI. IL, III., IV. and V). The number of hours, in 

 the course of which such trains occurred, was noted for successive 

 days, quite regardless of the length or the number of the trains. 

 The number shows occasionally an apparent periodicity of 25-30 

 days, though generally not so regular as to allow us to deduce any 



2) W. van Bemmeln, Natuurk. Tijdschr. v. Xederlandsch-Indie, 62, 1902, p. 71. 



