A.— MATHEMATICAL AND PHYSICAL SCIENCES. 33 



move the focus of the system westward, i.e. towards the a.m. meridian. 

 In practice the focus is near the 11 a.m. meridian. 



Another diurnal variation, namely, the lunar variation, is also 

 attributed to circulating electric currents, but in this case it appears 

 fairly certain that a dynamo effect is the cause, the conducting medium 

 being moved by direct tidal action. The results observed are in close 

 accordance with the theory. 



Variation due to Solar Eclipse. 



If, as we suppose, the sun's radiation is responsible for the conductivity 

 of the upper layers, it follows that the solar diurnal variation is not the 

 only magnetic change associated with the sun, for everything that affects 

 the radiation will produce corresponding effects in the ionisation of the 

 upper atmosphere, and hence changes in the magnetic field. An interesting 

 instance is the magnetic effect of an eclipse of the sun. Some effect should 

 ob\'iously arise since there is a cancellation of a portion of the solar 

 radiation and this will diminish the ionisation effects ; the solar diurnal 

 variation should, therefore, suffer a check. Kecent eclipse results are in 

 accordance with these views, though before 1900 doubt as to any measurable 

 effect being obtained was almost universal. In 1900 special observations 

 were made, and there have been others since at a number of stations, 

 princii)ally in the United States of America ; the records show that small 

 magnetic disturbances do result, the duration being roughly that of an 

 eclipse. Analyses of the changes indicate that the cause is external to the 

 earth's crust. The effect differs from that of an ordinary magnetic storm 

 inasmuch as it begins, progresses and ends gradually, and a definite con- 

 clusion to be drawn is that the effect is due to changes in the upper 

 atmosphere by the obliteration of the sun's rays due to the moon as an 

 obstruction. 



SUNSPOTS AND MAGNETIC StORMS. 



Any unevenness in the radiation from the sun as it rotates must also 

 affect the conductivity and hence produce variations. Examination of 

 magnetic records shows that many variations are related to the sun's 

 period and also to sunspot periods, and it appears not improbable that 

 there is overlapping of several periods probably intimately connected. 

 The results obtained show that with rise and fall of sunspot frequency 

 there are corresponding changes in the diurnal variation. Moreover, 

 the amplitude of the daily changes rises and falls with the intensity of the 

 magnetic disturbance. It follows, therefore, that changes in amplitude 

 of the diurnal variation in years of many sunspots is due to the same 

 ultimate cause, namely, solar radiation, as that causing magnetic dis- 

 turbance, but the existence of a 27-day period does not, as Chree pointed 

 out, justify the conclusion that sunspots are the only disturbing sources. 

 The same result would be obtained if the intensity were a function of the 

 solar longitude and did not vary too ra])idly with time. 



Magnetic storms are marked disturbances of solar origin, and to 



explain these many theories have been advanced, but the facts are not 



easy of explanation. The belief in the connection of solar activity with 



magnetic storms is old, but one of the earliest and most striking declarations 



19.30 D 



