PART I. ON MAGNETIC STORMS. CHAP. II. 83 



perturbations under consideration are also due to the rising of new, independent systems, and do not 

 merely indicate a weakening of that which may already exist. 



On the other hand, it is our opinion that the positive equatorial perturbations find their natural 

 explanation in the second of the two systems mentioned. At the place in which the earth is found, the 

 system will have a force directed towards the north. If the system is far off in proportion to the 

 earth's dimensions, the force round the equator can be almost constant. If the system is nearer, there 

 will be a stronger effect upon the evening side. This is also what we find in reality, as the effect 

 about Dehra Dun is somewhat stronger than at Honolulu. It must be remembered, however, that the 

 observed force is also dependent upon the magnetic induction in the earth. 



It would be useless to attempt here a more detailed description of these current-systems. It seems 

 probable that at times they may have a somewhat different character, being at one time fairly symmet- 

 rical about the equator, and at another pushed out more towards the north or the south. 



The experiment shows that the system may extend considerably in directions north and south. 

 This, together with the effect of the magnetic induction of the earth, will account for the smallness of 

 the vertical components. 



We have observed certain impulses in the north that appear to be of a local character, as the 

 force about the auroral zone might diverge greatly in direction at two adjacent stations, and receive 

 a marked, opposite twist. The equatorial perturbation of the 22nd March, 1903, is an instance. 

 This agrees very well with our view, as at times radial impulses may come right down to the earth 

 about the poles. In the experiment, moreover, we see that the equatorial system finally unites with the 

 polar; and we shall often have great polar precipitations of corpuscles. For this reason, a number of 

 these perturbations will be found described under the polar storms. 



THE NEGATIVE EQUATORIAL STORMS. 



32. On several occasions in the course of our investigations of the -composite magnetic storms, we 

 shall meet with conditions in the field of force, which naturally lead to the assumption that the per- 

 turbing force in the polar regions, on account of its independence of the polar systems, must be due 

 to systems that have their greatest strength in the equatorial regions. They differ, however, distinctly 

 from the previously-described equatorial perturbations in two very important respects, namely: 



(1) The perturbing force is directed southwards, answering to a current-arrow towards the 

 west, and 



(2) The curve has not the characteristic, serrated appearance that marks the positive equatorial 

 perturbations. The latter generally appear very suddenly, whereas those now under consideration 

 appear more gradually. 



We have not succeeded, however, in finding in our material of this kind of perturbation, suffi- 

 ciently distinct types to enable us to class them under any elementary form. In the treatment of the 

 composite perturbations, we shall repeatedly have opportunities of examining more closely the reasons 

 that determine the assumption of such perturbations. We may here mention as instances the perturba- 

 tions of the 3ist October, 1902, and the 8th February, 1903. 



These, like the positive equatorial perturbations, have a very wide distribution, as the conditions 

 of perturbation alter slowly from place to place. This, together with the quiet character of the curve, 

 shows that the systems that are to condition the perturbation, must be sought at a considerable height 

 above the earth. While we are thus led to suppose them to be corpuscular currents, we shall naturally 

 be obliged to connect this perturbation with the circular systems, which, according to the theoretical 

 investigations of the trajectories of electric corpuscles, can exist, and the possibility of which we have 

 also proved experimentally by the previously-mentioned ring (see fig. 37). 



