PART I. ON MAGNETIC STORMS. CHAP. III. 253 



possibly near Iceland, there should be a point of divergence for P k . At Val Joyeux and Pawlowsk, 

 there is a distinct vertical component directed downwards. 



It may further be stated that the current-arrows at Sitka, Baldwin, Toronto, and Cheltenham con- 

 verge towards a point in the vicinity of Prince Albert Land. 



Eastwards from Europe, the arrows turn off, but now towards the east. The directions of the 

 arrows, in connection with that at Sitka, indicate that somewhere in the north-east of Siberia, there is 

 a point of convergence for the perturbing force. 



The third section, from about i8 h 30 to o h 30, is characterised by a long polar storm. The 

 field of force of this storm is shown on the Charts V, VI, XI, XII, XIII and XIV, respectively for 



Ig h 30 m j 20 h gotn^ 2I h I5 m ( 2I h ^o, 2 2 h , and 23**. 



We see that the distribution of force is about the same in all of them, the strength of the field 

 alone showing any variation. 



At the arctic stations, the direction of the force is generally SSE and SE. 



There is a departure from this condition at ig h 30, when the force at Axeleen and Kaafjord is 

 almost westerly in direction. At 2o h 30 the force at Kaafjord is SSE in direction, but it is still west 

 at Axeleen. 



In the rest of Europe and in Asia, the direction of the force is ESE. At San Fernando it turns 

 a little more south, and in America the direction is south-west. This shows that in the North Atlantic 

 there must be a point of divergence of P h similar to that described at i8 h . At Sitka, the direction of 

 the perturbing force is WNW. 



During this period, however, there are several departures from these conditions, and it is evident 

 from the copies of the curves, that of these there are three principal ones, the first occurring at about 

 r8 h 34 m (see Chart IV), the second between 2o h 45 and 2i h 20 (see Charts VII to X), and the third 

 between about 23 h io m and o h 25 (see Chart XV). 



The fact that after these short interruptions the field once more assumes its original form, makes 

 it probable that the interruption is due to comparatively independent, brief current-systems, that occur 

 simultaneously with the long polar storm. The correctness of this view of the matter is also confirmed 

 by the fact that the differences do not occasion the same relative increase in strength at the various 

 stations. If we look at the curves, we shall see that these differences occur simultaneously all over the 

 world, even as far off as Christchurch. At the Norwegian stations also, sudden powerful perturbations 

 are observed, some of which have a different direction from that of the long storm. The three short 

 perturbations are thus polar storms, which intrude themselves upon the long storm. The latter we will 

 designate as the principal storm, and the three others as intermediate storms. 



The circumstances, as we see, are such as justify a decomposition of the perturbing forces into two 

 components, each of which is the effect of a separate current-system. This decomposition has been 

 effected in the case of the last two intermediate storms, but not of the first, as that storm commences 

 at the time of transition from the second to the third section. 



This is apparent in the curves, e. g. for Tiflis and the south-east Asiatic stations, where the 

 //-curve, at about i8 h 30, drops suddenly, showing that P, n from being positive, has become negative. 

 This circumstance makes it impossible to draw any exact normal line for the taking out of the partial 

 forces. 



We will now describe in detail the three intermediate storms. 



The first intermediate storm, at about i8 h 34. 



This perturbation appears in the curves as a great, but brief, deflection at about 18'' 34. At the 

 southern stations, such as Tiflis, Dehra Dun, etc., it appears to be the long perturbation only that is at 



