On Eapid Periodic Variaticns of Terrestrial Magnetism. 



31 



On closer examination of the diagrams mentioned in p. 25 

 above, it is frequently found that wave trains with a certain 

 period appear in almost the same isolated hours of two or three 

 successive days, as if to repeat the particular program of the 

 preceding daj^s. This is interesting as one of the facts strongly 

 suggesting that the agent producing these waves is intimately 

 connected with the position of the earth relative to the sun. In 

 some cases, it also happened that the hours of tlie successive days 

 in which a particular train appears seemed to shift gradually in 

 either direction. 



9. The ^' spectra^' of the magnetic waves. To find the most 

 frequent period, the o1;>served periods were at first classified second 

 by second and the total number of trains (not that of days or 

 hours in which they occurred) belonging to these groups were 

 plotted in a diagram with the periods as abscissa. The diagram 

 obtained showed a great number of maxima and minima with 



intervals of several 

 seconds. Most of 

 these maxima and 

 minima were, how- 

 ever, found to have 

 no real physical 

 meaning, being 



either due to 

 chance,^-' or due to 

 some involuntary 

 tendency of par- 

 ticular persons to 

 some par- 

 ticular fractions of the scale division when measuring the wave 

 trains on the records with a millimeter scale. ^^ Hence, finally the 

 following groups were taken: 20'-30', 30'-40' etc. up to 190'- 

 200'. The results are given in Table III. and also plotted in 

 Fig. 9. As will be seen, the maximum frequency is generally 



II T. Terad», Proc Tokyo Matb.-Phys. Soc, 8, 1916, p. 492. 



2) Note that a tenth of a millimeter corresponds to about I'S sec. 



20 40 60 50 



100 



120 140 IfcC 



T'' prefer 



