1884.] and its Relations to the Sun-spot Period. 393 



greater in the maximum years of the eleven-yearly period than in the 

 minimum years of that period, the amplitude varying nearly in pro- 

 portion to the mean diurnal inequality -for the year. 



3. It is also subject to a variation whose duration is more than 

 thirty-three years. 



4. If any other periodical variations than those above mentioned 

 have any existence, their amplitudes must be small, or the duration of 

 their periods must be less than one year, or more than thirty-three 

 years. 



5. The maximum and minimum values of the eleven-yearly period, 

 and the range between them, have all increased in the successive 

 eleven-yearly periods of the years 1846 to 1879. 



6. The range of the eleven-yearly period is roughly proportional to 

 the mean diurnal inequality for the interval from which the range is 

 calculated. 



7. The principal features of the eleven-yearly period coincide 

 (except in point of time) with those of the well-known sun-spot period. 

 The duration of the period is the same in both cases, and both varia- 

 tions exhibit a relatively rapid rise from minimum to maximum in 

 about four and a half years, and a relatively slow fall from maximum 

 to minimum in about six and a half years. 



8. The magnetic curve precedes the sun-spot area curve by as much 

 as six months. 



9. The magnetic curve is more regular than the sun-spot area 

 curve. 



These or similar conclusions, with the possible exceptions of the 



last two, have either been enunciated by previous investigators, or 



are directly deducible from their labours ; but the evidence now 



advanced is probably more decisive than any that has hitherto been 



- published. 



On the whole, the conclusions arrived at seem to support Dr. Stewart's 

 views in a very unexpected manner, for not only does the eleven- 

 yearly magnetic variation precede the eleven-yearly -variations of 

 certain meteorological elements, but it actually precedes, by several 

 months, the eleven-yearly variation of the sun-spots themselves; and 

 the probable error of a single monthly determination of a point on 

 the magnetic curve is little more than 5 per cent, of the greatest range 

 of the eleven-yearly oscillation, a degree of accuracy which is pro- 

 bably unapproached by any other kind of observations. 



