( 275 ) 



from the original recording-sheets on the contrary at 1'' p. m. The 

 cause of this discrepancy he finds in the manner" he looked for the 

 commencement. He writes : "the times when the phases of diurnal 

 disturbance are most strongly marked are naturally most often taken 

 as the times of commencement." 



At Greenwich these phases of agitation are most prominent at 

 1'' p. m. and 6^ p. m. 



As I assumed for the hour of beginning the first increase of 

 agitation it is clear my times of beginning are on an average much 

 earlier. 



Thus the difference shown by the hourly distribution of commen- 

 cement between the Greenwich- and the Batavia-list, may be ascribed 

 chiefly to difference of interpretation. 



As appears from the figures given above, compared with those for 

 Greenwich the annual periodicity is quite the same for both the 

 northerly and the equatorial stations, which diifer no less than 60 

 degrees in latitude. But the Greenwich dates, quoted from a complete 

 magnetic calendar, prepared by Mr. Ellis and extending from 1848 

 to 1902 give no separation of G- and S-storms. Thus it is not 

 possible to decide whether at Greenwich the G-storms lack an annual 

 periodicity in their frequency. 



TJie impulse at the start. 



The material at present at my disposal for investigating the features 

 of this phenomenon in other places on earth, is very small. 



Notwithstanding this I may conclude: that this phenomenon is of 

 great constancy in features all over the earth, and consequently a 

 phenomenon of great interest, tvhich might teach us much about the 

 manlier the S-storms reach the earth. 



Description of its features for some places. 



Greenwich. According to the reduced reproductions of disturbance- 

 curves published in the volumes of the Greenwich-Observations, 

 the impulse consists of a sudden movement in H, D and Z, instantly 

 followed by the reversed movement, the latter being considerably 

 greater. The direction of the movement is always the same. 



I have measured S-l cases and have found on an average : 

 HAD AH AZ 



25 V W + 77 r -f 39 y (1 7 = 0.00001 C.G.S.). 



Batavia. The preceding impulse is missing for H and Z, only for 

 D it is often present. 



