36 Art. 9. -T. Terada: 



which is shoAvn in Fig. 10,1). Tliis point decidedly deserves 

 further investigation. 



12. To carry out a more detailed quantitative investigation 

 regarding the nature of the magnetic pulsations under con- 

 sideration, the records of 1913 were specially chosen, since in 

 this year the determination of tlie sensihilities of the instruments 

 was most regularly made and lience most adequate for the 

 quantitative comparison of the different components. It is, 

 however, to be regretted that during the summer months the 

 records were too frequently defective, chiefly due to damage to 

 the photographic paper on account of the extreme dampness 

 notwithstanding tlie use of the desiccator, and also due to the 

 condensation of the atmospheric humidity in minute drops or 

 mist which was especially dense in tlie daytime and caused a 

 remarkable absorption of light. It must therefore be remarked 

 that in some statistical studies to be described later, tlie winter 

 season has a rather overweighing influence on the general results 

 though the statistics extend over a full year. As far as the 

 present investigations are concerned, no serious modification of 

 the general results will, however, l)e required on that account. 



13. Ratio of amplitudes of X- and Z-components. As already 

 mentioned, the waves appearing in the Z-component arc generally 

 the reduced facsimiles of those in the X-component, except that 

 the former always lags behind the latter in definite amounts 

 depending on the periods, but usually less than a quarter of a 

 period. In other words, the end of the magnetic vector repre- 

 senting the periodic disturbing field revolves on more or less 

 elliptic orbits, with their major axes mostly dipping towards N. 

 To find first a quantitative relation between the amplitudes of 

 the two components, the following reductions were made. 

 Specially regular portions of wave trains were carefully chosen 

 out from among all the records of the year, and for each train 

 the ratio of the mean amplitude of corresponding waves for the 

 two components w^as calculated, the daily value of the sensibilities 

 being duly taken into account. The results were tabulated 

 together with the corresponding periods and tlie hours of 



