48 Art. 9.-T. Terada: 



of the magnetic pulsations in question, tliough it is at the present 

 stage rather difficult to draw anything like a conclusive inference, 

 for even if the observations had been carried out for a much 

 longer period with more reliable instruments, the data were in any 

 case confined only to a single station. In the following, we wdll 

 try, instead of hastening to any premature conclusion, to consider, 

 merely by way of tentative suggestion, different possibilities 

 regarding the probable cause of the phenomena in question. 

 The considerations will inevitably be of a speculative cliaracter, 

 but may serve at least as useful hints for projecting future 

 investigations of allied phenomena, especially of the as yet very 

 obscure nature of the electrical as well as meclianical behaviour 

 of tlie upper atmosphere. 



19. A fact strongly impressing us in the first place, is 

 that the occurrence of the magnetic pulsations in question is 

 subjected in more than one respect to a remarkable diurnal 

 variation. This evidence alone is sufficient to infer the important 

 rôle played by the sun, wdiether its influence l;e direct or indirect. 

 The position of the sun, not only determines the length of the 

 periods of the most frequent waves in different hours, but also 

 affects the direction in wliich the periodic magnetic field 

 fluctuates. It seems quite plausible to assume from the outset 

 that the seat of the primary cause of the phenomena is chiefly 

 to be sought in our atmosphere subjected to solar radiation of 

 different kinds. The periodic heating of the superficial layer 

 of the earth crust, though it may possibl}^ cause a slow variation 

 of the terrestrial magnetism, may scarcely account for the periodic 

 nature of the disturbances in question. The direct magnetic 

 influence of the sun itself^^ seems also improbable, since if such 

 be the primary cause, the more or less complete screening off of 

 waves shorter than 50^ during night hours must be explained, 

 while 100^ waves are so conspicuous in these hours. Moreover, 

 the unsymmetrical distribution of the cliaracteristic waves at noon 

 is rather difficult to explain on this view. On the other hand, 

 the existence of remarkable electric currents in the upper region 



1) Bosler, Journal de Physique, [6] 2, 1912, p. 877. 



