666 BIRKKLAND. THE NORWEGIAN AURORA POLARIS EXPEDITION, IQO2 1903. 



In this case the origin of the sun-spots must be that the presumptive more or less insulating 

 photospheric envelope was sometimes pierced by disruptive discharges, thus forming great electric arcs. 

 That the tension necessary to pierce the photosphere would be very great would not be surprising 

 this alone being sufficient to explain the very great rigidity of the cathode-rays emitted. 



The temperature of the spots should, upon this hypothesis, be very high. This, it is said, does 

 not seem to be well confirmed by the measurements; but the temperature of a spot cannot be measured 

 by STEFAN'S law, because under high degrees of dispersion the spectrum of the spots is not continuous; 

 it contains nothing but lines. 



It may be imagined that under the action of these violent arcs the photosphere tends to 

 more insulating (thicker?), and that after the maximum of the spots, the discharges cannot penetrate the 

 photosphere as easily as after a certain cooling by radiation. The discharges then begin again in higl; 

 latitudes as long as the necessary tension is at its maximum. 



We do not know sufficiently how electric arcs move in gases, but it is at any rate not difficult, b 

 magnetic forces, to attain a transversal velocity of 200 metres per second for an electric arc in air. 



In order to be able to some extent to form an estimate of the manner in which the a 

 electric arcs in the sun would move, we ought to know how the sun's magnetism is distributed, 01 

 rather its cause. In my opinion it is the pencils of cathode rays appearing at indefinite intervals at tin- 

 outbreak and in the development of the sun-spots, that give rise to solar magnetism by creating almost 

 constant currents by induction in the conductive interior of the sun. 



I have several times begun the calculations that should serve to verify my hypothesis, but t 

 not yet completed. 



We know that the electric currents circulating in great spheres have a very great persistence (set 

 LORBERG, Crelles Journal, vol. 71, 1870, and LAMB, Phil. Trans., 1883). Lamb finds that in a copper 



sphere of the size of the earth, the time necessary for a current to fall to of its initial value is ten 



c 



million years. 



The induction impulses originating in the cathode-rays emitted at intervals from the sun, seem t< 

 be able, in the course of time, to create a perceptibly constant current. 



In support of my calculations, I am making experiments with a rotating sphere made ot the x 

 softest magnetisable steel. The diameter of the sphere is 70 cm. The results of these investigations! 

 be included in the next volume. 



If, to obtain a clearer conception, we assume a circular current round the centre of the sun in th 

 plane of the equator, and with a radius equal to half the solar radius, it becomes easy to calculate 

 magnetic effects in different latitudes of the photosphere. In assuming spherical currents, we obtain th 

 same degree of conformity with the currents circulating much nearer the solar surface. 



The table gives Fp divided by cos// for each ten degrees of latitude comprised between o and 50' 

 where Fp is the component of the magnetic force in an arbitrary unit, the length of the meridian, 

 purposes of comparison, cos 2 ,:? is given, which, according to FAYE, should be perceptibly proportional 

 to the variation of the angular diurnal motion of the spots. 



/3 o 10 20 30 40 50 



Fp sec/? 1.17 1. 10 0.88 0.69 0-54 0.41 



Cos 2 /? i. oo 0.97 0.88 0.75 0.59 0.41 



These figures have perhaps a certain interest, although, as we have said, we do not yi 

 well how electric arcs move in gases, under the action of magnetic forces. 



The second assumption may indeed, from a physical point of view, be possible, but it is scared 

 probable that any process of this nature will play a decisive part in these phenomena. It would imply 



