150 Dr. C, Chree : Auroral and 



summer than in winter, i.e. H : exceeds B,/andR 2 exceeds E 2 '. 

 Tims (2) shows that relatively considered the diurnal range 

 is more variable throughout the year when sun-spots are 

 few than when they are many. In other words, if the 

 annual change of the diurnal range be illustrated by a curve 

 Avhose ordinates represent the ratios borne by the ranges in 

 individual months to their arithmetic mean for the twelve 

 months, the maximum and minimum ordinates differ more 

 when the year selected is one of few than when it is one of 

 many sun-spots. 



§ 2. Again, takingalist of the more considerable magnetic 

 disturbances recorded at Greenwich from 1848 to 1903, as 

 given by Mr. W. Maunder, I obtained the following figures * 

 for the relative frequency at different seasons of the year, 

 treating separately the fourteen years of largest (S max.) 

 and the fifteen years of smallest (S min.) sun-spot frequency: 



Winter. Equinox. Summer. 



November to February. March, April, Sept. Oct. May to August. 

 S ruax. S min. S max. S min. S max. S min. 



35 28 " 38 48 27 21 



The figures denote percentages of the totals for the Avhole 

 year. The average absolute numbers of storms per annum 

 were 



Whole period. S max. years. S min. years. 



13-0 18;4 &Z 



With increased sun-spot frequency, the absolute number 

 of storms increased in all three seasons of the year ; but 

 relatively considered the increase was least in the equinoctial 

 months — the season when magnetic storms are most numerous 

 at Greenwich — and the tendency obviously was towards a 

 more uniform distribution throughout the year. The pheno- 

 menon is thus analogous to that described above in the case 

 of the regular diurnal range. 



§ 3. In temperate latitudes, as is well known, magnetic 

 storms of any considerable intensity are usually associated 

 with auroras. It was thus of interest to determine whether 

 auroral frequency showed phenomena corresponding to those 

 just described in Terrestrial Magnetism. It has long been 

 known that auroral frequency, as observed in temperate 

 latitudes, varies in a general way with sun-spot frequency. 

 The results obtained, however, have not shown a very exact 

 correspondence between the years of maximum and mini- 

 mum in the two classes of phenomena. In the case of 



* Phil. Mag. Sept. 190-5, p. 314. 



