158 Dr. C. Chree : Auroral and 



Table IY .—-Auroral Observations in Scandinavia. 







District. 





All 

 Scandinavia. 



1 



I. 



II. 



III. IV. 



V. 



1 July to June. 

























1701 to 1794 



22 



293 



970 



1453 



555 



2481 



1794 to 1827 



52 



694 



408 



518 



56 



1316 



1827 tolSGO 



827 



873 



352 



1531 



473 



2811 



1st & 3rd periods combined. 



849 



1166 



1322 



2984 



1028 



5292 



i 



An aurora is often seen in more than one district, but 

 comparatively seldom in ail. Obviously the increase in the 

 auroral frequency in the latest as compared to the earliest of 

 the three periods arises from tbe large increase of observations 

 in the two most northern districts ; and this being so, the 

 tendency naturally is to bring the annual frequency nearer 

 to the Arctic type with a single maximum at midwinter. 



Coming now to the intermediate period in Table IV., we 

 see that the development of district I. was mainly subsequent 

 to 1827. Thus what w r e should a priori expect to observe in 

 the annual variation from this period in Table III. would be 

 a form intermediate between those from the other 33-year 

 periods, but approaching most closely that from the earlier 

 period. What we actually do find is a variation differing 

 from that of 17G1-93 in the expected direction, but to an 

 even greater extent than does the variation from the period 

 1827-59. Comparing the results from 1791-1826 in Table 

 III. with the means from the preceding and succeeding 

 33-year periods we largely eliminate the influence of the 

 change in observational conditions. The figures show that 

 in the 33-year period characterized by few sun-spots, summer 

 occurrences of aurora were relatively much fewer and winter 

 occurrences more numerous than in the adjacent 33-year 

 periods of high average sun-spot frequency. We thus have 

 low sun-spot frequency associated with an exaggeration in 

 the annual variation of auroras, the precise phenomenon already 

 described in connexion with Terrestrial Magnetism. 



§ 8. The mean sun-spot frequency for the two 33-year 

 periods 1761-93 and 1827-59 combined was 60*83, whilst 

 the corresponding auroral frequency for the year w r as 80*2. 



Comparing these with the corresponding figures for 1794- 

 1826, we have for a trebling of sun-spot frequency only a 

 doubling of auroral frequency. This would suggest that if 



