MAGNETIC DISTURBANCES AT KEW. 273 



absolute range i.e., the excess of the absolute maximum over the absolute minimum 

 during the 24 hours was necessarily under-estimated. 



Non-Cyclic Changes. 



2. The principal object of the present paper is to consider the diurnal inequalities 

 derived from the disturbed days. The interpretation, however, of the phenomena, 

 especially in the case of H, depends so considerably on the treatment of the non-cyclic 

 changes, and these changes themselves are of so remarkable a character that their 

 discussion calls for some detail. A general discussion of the non-cyclic changes at 

 Kew on quiet days from 1890 to 1900 is given in (A), 13 to 17. The mean values 

 were for D + 0''044 ( = 0'23y), for H + 3'34y, for V -0'84y, and for I -0''245. 



When first describing the phenomenon, I remarked* that the non-cyclic change in 

 quiet days was " not unlikely to be only another phase of phenomena observed many 

 years ago by General SABINE and Dr. LLOYD." These gentlemen, dealing respectively 

 with Kew and Dublin results, described a marked tendency for disturbance to cause 

 diminution in H, while SABINE observed a less marked tendency for disturbance to 

 bring about increase of V. The diminution in the value of H during a magnetic 

 storm is sometimes so large as to at once catch the eye. Dr. VAN BEMMELEN,! who 

 discovered the phenomenon independently, regarded the non-cyclic effect apparently 

 as simply a reaction or recovery from the effects of a storm. 



When dealing in (C') Avith non-cyclic changes of declination I found, somewhat to 

 my surprise, that the mean value derived from the 209 disturbed days, viz., + 0''327 

 had the same sign as the mean for the average quiet day, and that it was the 

 " ordinary " days which supplied the requisite balance of negative non-cyclic changes. 

 Another unexpected phenomenon was a peculiar oscillation of sign in non-cyclic 

 changes. Out of 46 cases in which two, and only two, successive days were classified 

 as disturbed, there were no less than 29 in which westerly declination diminished on 

 the first day and rose on the second, whereas this particular sequence would have 

 been expected on only 11 or 1 2 of the occasions. These two phenomena alone would 

 suffice to show that the relation between the non-cyclic changes observed during quiet 

 days and the alterations accompanying magnetic disturbance is of a somewhat complex 

 character. One serious difficulty in the study of the connection is that magnetic 

 disturbance is in reality the rule rather than the exception. An unprejudiced idea of 

 the prevalence of disturbance will perhaps be best derived by reference to some data 

 from the international quarterly lists published of late years in the Netherlands. 

 Each day is classified at each contributory observatory as " " (quiet), " 1 " (moderately 

 disturbed), " 2 " (highly disturbed). It has not yet proved possible to secure a uniform 

 standard at all the observatories, but at Greenwich, I believe, "0" includes the days 



* 'Brit. Assoc. Report' for 1895, p. 212. 

 t 'Met. Zeit.,' vol. 12, 1895, p. 321. 

 YOL. CCX. A. 2 N 



