On Rapid Periodic Variations of Terrestrial Mafjnetism. 25 



8. Hourhj distribution of ivaves of different j)eriods. The 

 different periods were classified into the following arbitrary 

 groups: 30•0^-49•9^ 50•0^-69•9^ 70•0^-89•9^ 90•0^-129•9^ 

 ]30^-20(/. The occurrence of waves Ijelonging to the different 

 groups, in different hours of successive days, was plotted down in 

 diagrams similar to those drawn by Bidlingmeier'^ in his " Ue- 

 bersicht über die Tätigkeit des Erdmagnetismus." 



From these diagrams, it could at once be seen that waves 

 with periods shorter than about 50' are most frequently met with 

 during the day time, while the longer waves over 90^ periods are 

 generali}' most frequent during the night. This is naturally true 

 for each of the three components, though in Z-component the 

 shorter waves are generally of almost insignificant amplitudes. 

 To make this remarkable fact more apparent, the fohowing 

 procedure was adopted. Different days of a month, or of a 

 season, were taken together, and the number of those days in the 

 said interval, in which a certain given hour was disturbed by the 

 waves of a certain given period, was counted and then divided by 

 the total numljer or days in that interval, in which the ol)- 

 servation was actually made at the very hour in question; the 

 latter reduction was necessary, since especially in summer, the 

 records were often imperfect. The number tlius reduced, multi- 

 plied with 100, was called the "frequency" of that given group 

 of waves in that particular hour, for the month or season in 

 question. The following tables give the hourly frequency of the 

 different wave groups in X- and .Y-component, for different 

 seasons, all tlie records available being taken into account. Z- 

 component is omitted, since it is always a reduced facsimile of 

 X-component, as already mentioned. 



1) Bidlingmeier, Veröffentlichungen d. kais. Observatoriums in Wilhtlmshaven 



