512 



NATURE 



[August 17, 1916 



turbed conditions, and every day may have a numerical 

 measure attached to its disturbance. International 

 "character" figures naturally suggest themselves for 

 the purpose. 



The "character" figures were entered in successive 

 columns, representing from so many days before to 

 so many days after the representative disturbed day. 

 The successive columns were summed, and the result- 

 ing means taken as a measure of the average disturb- 

 ance presented from so many days before to so many 

 days after the representative day. 



The days recognised by Maunder as magnetic storms 

 average only about one a month, and were much more 

 numerous in some years than others. If the 27-day 

 period had been a phenomenon confined to such highly 

 disturbed days, the procedure adopted here could 

 scarcely have brought it into evidence, except in dis- 

 turbed years. It proved, however, to be as much in 

 evidence in the less disturbed as in the more disturbed 

 years. This suggests that it is not peculiar to dis- 

 turbed conditions, a conclusion which is strongly sup- 

 ported by Fig. 4, which shows the results of apply- 



acter" figures on the days which are twenty-seven 

 days subsequent to the representative disturbed and 

 quiet days respectively. The total length of the ver- 

 tical line may be regarded as a measure of the primary 

 dilTerence pulse (disturbed less quiet), and the length 

 of the thickened portion as a measure of the corre- 

 sponding secondary pulse. The short horizontal line 

 shows the " character " level of the average day of the 

 year. The lengths of thickened line above and below 

 this level may thus be regarded as representing re- 

 spectively the amplitudes of the secondary pulses of 

 disturbed and quiet, conditions. Above the nine lines 

 are given Wolfer's mean sun-spot frequencies for the 

 respective years. 



The 27-day period is conspicuously shown in Fig. 4 

 in every year except 1914, where the secondary pulse 

 associated with the representative disturbed dav is 

 abnormal. The two years in which the 27-day period 

 is most in evidence are igii and 1913, both, especially 

 the latter, years of few sun-spots ; while 1907, the year 

 of sun-spot maximum, shows it less than any other 

 year except 1914. In 1912 the secondary disturbed 



Sunspot f requencj 53-8 62-0 4S-5 45-0 16-6 5-2 3-6 1-4! 9.7 



I-30 



I-20 

 I 10 

 I-OO 



<u 



^0-90 

 ho 

 <f-0-8o 



^0-70 



o 

 ^0-60 



50-50 



OvJ.O 



030 

 ozo 

 010'- 



- 1 



Normal 



26 27 



Yed.ri906 I907 I908 1909 I9IO I9U 1912 1913 1914 



Fig. 4. — 27-day period. International "character" figures 1906 to 1914. 



ing the procedure explained above to the international 

 quiet days as well as to the representative disturbed 

 days of the nine years 1906 to 1914. The representa- 

 tive days in each category were five a month. The 

 normal line in Fig. 4 represents the mean " character " 

 figure, o-6o<, of all days of the nine years. Above 

 this normal line we have the primary and secondary 

 pulses associated with the representative disturbed 

 day, the " character " figure of which was 1-26, and 

 below it are the primary and secondary pulses asso- 

 ciated with the representative quiet day, the " char- 

 acter " figure of which was o-ii. The secondary pulse 

 associated with the representative quiet day is not 

 quite so deep as that associated with the representative 

 disturbed day, but the same is true and to a like extent 

 of the primary pulses. 



The graphical representation of the results for the 

 individual years in Fig. 4 is confined to days o and 27. 

 The extreme top and bottom of the lines represent the 

 " character " figures on the representative disturbed 

 and quiet days, on the same scale that serves for the 

 nine years combined. The top and bottom of the 

 thickened portions of these lines represent the " char- 



NO. 2442, VOL. 97] 



pulse is much better developed than the secondary 

 quiet pulse, and 19 13 shows the same phenomenon to 

 a minor extent. In 1906, on the other hand, the 

 secondary quiet pulse is the more prominent. In the 

 years 1907 to 191 1 the development of the two 

 secondary pulses is very similar. 



A good deal probably remains to be done to unravel 

 the exact nature of the relationship between sun-spots 

 and magnetic phenomena. There can scarcely be any 

 doubt that the range of the mean diurnal variation 

 for the whole year varies from year to year in almost 

 exactly the same way as the mean sun-spot frequency 

 or the sun-spot area. Also the two phenomena exhibit 

 a 27-day period, and to approximately the same extent. 

 In the average year of an ii-year period, 1896 to 1900, 

 the daily range of H at Kew showed a decided tend- 

 ency to be above its mean value during several suc- 

 cessive days subsequent to the appearance of excep- 

 tionally large sun-spot area, the maximum in the 

 range following four days after the maximum in the 

 area. The phenomenon, however, did not seem to 

 arise from special disturbance, but rather to be a 

 variant of the phenomenon of large regular diurnal 



