( 268 ) 



In accordance with this description, during the earlier part of a 

 storm the horizontal force diminishes, the vertical force increases, 

 and during the latter part these forces resume slowly their original 

 values. The characteristic features sub 1 and 2 either do not neces- 

 sarily attend every storm, or if so, they do not show themselves 

 clearly enough to enable us to decide definitely whether a succession 

 of waves in a curve must be considered as a storm or not. 



On the contrary the increased agitation is an essential feature 

 and has therefore been adopted by me as a criterion. 



Unfortunately it is impossible to establish the lowest level above 

 which the never absent agitation may be called a storm, because 

 the agitation is not only determined by the amplitude of the waves, 

 but also by their steepness and frequency. 



To eliminate as much as possible the bad consequences which 

 necessarily attend a personal judgment, the list has been prepared: 



1 . by one person ; 



2. in as short a time as possible; 



3. from the aspect of the curves for one component only (in 

 casu the horizontal intensity, which in Batavia is most liable 

 to disturbance); 



4. for a period with nearly constant scale-value of the curves 

 (1 mm. = ± 0,00005 C.G.S.); 



For each storm has been noted: 



1. the hour of commencement; 



2. „ „ „ expiration; 



3. „ „ „ maximum; 



4. the intensity. 



Mr. Maunder calls a storm with a sudden start an S-storm; 

 analogously I will call one with a gradual beginning a G-storm. 

 In the case of a sudden impulse the time of beginning is given to 

 the tenth of an hour; in that of gradual increase of agitation only 

 by entire hours. 



The hour of beginning of a G-storm is not easy to fix. I have 

 chosen for it the time of the very beginning of the increased agita- 

 tion, and not the moment in which the agitation begins to show an 

 unmistakable disturbance character. 



Afterwards it became clear I had shown a decided preference for 

 the even hours, which may be accounted for by the fact that only 

 the even hours are marked on the diagrams. 



To eliminate this discordance I have added the numbers of G-storras 



