ß8 Art. 9. — T. Terada : 



neccessary result or cause of the fluctuation of the intensity, p 

 must be equal to q, in which case the period of dY will be 

 octave to that of JX. This latter case was frequently observed, 

 though not always. 



The above consideration seems to explain the fact that 

 the cases where the period of the X-waves is longer than that 

 of Y are decidedly more frequent than the opposite cases, if we 

 remember that the hypothetical current is generally inclined 

 to the circle of latitude by an angle decidedly less than 45.° 

 It is however strange to observe that the hourly distribution 

 given in § 5 is rather conformable to the supposition of the 

 "difference tone" being more conspicuous. This latter point 

 requires further investigation. 



Different facts remain still to be considered in the light 

 of the above hypotheses, i.e. the dependency of the prevalent 

 periods on the hours of occurrence, the rotatory character of the 

 horizontal components, etc. We shall consider these points 

 later, after having tried to ascertain the nature of the supposed 

 atmospheric electric currents. 



26. Thus far, we have tried to find some plausible ex- 

 planations of the observed phenomena on tlie assumption of a 

 system of fluctuating currents existing in the upper atmosphere, 

 neither alluding to the possibility of such, nor giving hint as to 

 the cause of the supposed fluctuations. It will be by far the 

 more interesting physical problem to inquire into the origin of 

 these remarkable periodic variations. Tliough ^ve are not yet 

 in a position to answer the question in any definite sense, it will 

 not be quite out of place to say a few words on the subject, 

 merely by way of suggesting many interesting problems regarding 

 the electrical behaviour of the upper atmosphere. 



The existense of a permanent s.ystem of atmospheric currents 

 pru<kicing the remarkable diurnal variation of the terrestrial 

 magnetism, first effectively elucidated by Schuster, may in these 

 days be regarded as a universally accepted hypothesis and requires 

 scarcel}^ any further comment. It is the fluctuation of this 

 current that we are here chiefly concerned with, and the 



