300 BIRKELAND. THE NORWEGIAN AURORA POLARIS EXPEDITION, IQOZ 1903. 



In investigations made at the same time, but not published until some months later (') I had shown, 

 however, that the phenomena were not so simple; it is certain, indeed, that no theoretically clear under- 

 standing has yet been arrived at with regard to the formation of the cones of light shown in fig. 135. 

 In the above-named paper, I have shown how the theory can explain a number of discontinuously occurring 

 luminous rings in the discharge-tube, even if we suppose the cathode to emit a whole sheaf of rays, and 

 not only separate bundles with definite angles of emanation for the rays. It may possibly also be shown 

 that the above-mentioned cones of light in space are formed by a maximal agglomeration of rays about 

 certain surfaces, thus making the density of the rays there so great that the rarefied air in the tube 

 becomes more luminous near these surfaces. 



I have here touched upon this matter because these cones of light and their attendant phenomena 

 will be found to play an important part in our theory of terrestrial-magnetic and auroral phenomena. 



My special reason for here reproducing the above photographs in fig. 134 and mentioning the experi- 

 ments, is my desire to indicate phenomena that may possibly afford a full explanation of a peculiar circum- 

 stance that has frequently been pointed out in the preceding pages. We have seen that during the 

 so-called positive polar storms on the post-meridian side of the earth, the current-arrows at Dyrafjord, 

 Kaafjord and Matotchkin Schar have often been directed eastwards, more or less along the auroral 

 zone, while at the same time the arrow at Axeleen pointed in the opposite direction, westwards along 

 the auroral zone (cf. the perturbations of the nth, 2jth and 3131 October, 23rd November, gth December, 

 and 8th and I5th February). 



The great spiral of light round the magnetic south pole of the terrella represents, in my opinion, 

 the precipitation of rays on the night-side of the globe during long magnetic storms. It represents the 

 "horizontal part" of the current generally passing between Kaafjord and Axeleen at about midnight, the 

 breadth of which I have estimated to be not more than 500 km. While discussing the long magnetic 

 storms, we have frequently pointed out that in the afternoon the negative storm at Axeleen seems to 

 be closely connected with storms farther east on the night-side of the earth; while at the same time a 

 positive storm is observed at Dyrafjord and Kaafjord. 



Our photographs in fig. 134 answering to 6 a. m. at the magnetic south pole, clearly show that 

 the spiral of light begins in a very high latitude on the post-meridian side, whence it passes round 

 the terrella in its descent to lower latitudes. When, for instance, the terrella is turned so that it 

 is noon at the pole the beginning of the spiral also moves down towards lower latitudes, its longitude, 

 however, changing only slightly, measured from the cathode. 



In this connection I will mention that during the observations of aurora at the Haldde observatory 

 in mid-winter, 1899 1900, the following phenomena were observed day after day. Early in the after- 

 noon, generally at about 5 or 6 p. m., local time, an arc would appear far to the north and close down 

 on the horizon, and would remain through the evening, moving farther and farther south, and higher 

 and higher in the sky. As it came nearer, it would sometimes divide into several separate arcs. At 

 about 9 h or io h it would disappear, generally rather suddenly. During these auroral displays, our magneto- 

 meters were generally disturbed; but the most powerful magnetic storms almost always occurred after 

 midnight, when there was generally no aurora to be seen. This seems to agree well with the conditions 

 on the terrella, where the first great precipitation begins on the post-meridian side far up near the pole, 

 and descends to lower latitudes before it ceases or becomes a faint band of light, which continues 

 round the terrella. This greatest precipitation consists of rays that descend almost perpendicularly upon 

 the terrella; while the slighter precipitation on the night-side must be produced by rays that rather 

 glance past the terrella. Corresponding rays that glanced past the earth on the night-side would generally 

 produce magnetic storms. 



(!) Archives des Sciences Physiques et Naturelles, Geneva, 4th period, vol. IV, 1898. 



