On Rapid Periodic Variations of Terrestrial Magnetism. 



38 



it must be taken into account tliat tlie shorter the period, the 

 greater will be the chance of revealing isolated trains in a given 

 interval of time, even if the mean intensity and number of 

 maxima of the trains be independent of the periods. This 

 partially explains the falling off of tlie frequenc}^ curve tOAvard 

 the longer period in a form of h3"perbola. Hence the real 

 meaning of the above result can be simply interpreted as follows: 

 As far as the sensibilities of the present instruments reveal, every 

 period is rather evenly represented as in a continuous spectrum, 

 showing no very sharply defined maximum, except two rather 



Fig. 10. 



XI XI. I II ill ,v 



1^12 j;i ,x 



M XII 



I I 



II III 



(ßJ 



ce; 



0^ To 20 ctai^S 



100 



5" 10 cUv5 



flat relative maxima near 60^ and 90? The above holds good 

 when we take the different hours of day and night altogether 

 into account. AVhen the different hours arc taken separately, 

 the results are somewliat different as already remarked in §8. 

 This latter point will be referred to again later on, with respect 

 to difïerentty chosen materials. Here it may be remarked only 



