5 86 



BIRKELAND. THE NORWEGIAN AURORA POLARIS EXPEDITION, lgO2 1903. 



A new arrangement of screens for the terrella was now carried out, as the succeding photographs 

 distinctly show. A vertical annular plate, coated with tungstate of lime, was soldered to the ends 

 of the eight arms of the star-screen. 



Nos. 4 and 5 were taken from positions with hour-angles of respectively 235 with 15 declination 

 and 180 without incline. The magnetising current to the terrella was only 6 amperes, and the discharge- 

 current was 23 milliamperes with a tension of 3000 volts. 



In No. 4 are seen continuations of the positive precipitation on branch 2, and this continuation seems 

 to be formed from the same rays that formed the first line of precipitation on our earlier equatorial scnrn. 

 When the magnetising is increased, the precipitation spreads over the screen farther from the terrella. 



No. 5 shows one of the characteristic luminous triangles that we saw in fig. 68 of Section I; but 

 here there are also shadows of the suspending and current-conducting wires. 



The position in No. 6 answers to an hour-angle of 90. It will be observed that the polar ligh 

 has been reduced, and we see two peculiar lines of precipitation on the vertical screen to the left, 

 magnetising current was 10 amperes, the tension 2800 volts. The shadow of the suspending wire in 

 the polar ring of light seems to show that the rays forming the latter pass above the screen. 



The conditions in No. 7 are similar to those in No. 6, except that the magnetising current is 

 amperes. In this case, with the slighter magnetisation, the peculiar lines of precipitation on the vertical 

 screen have moved anti-clockwise, and the polar ring of light is even fainter than before. This shows, 

 as we have already seen, that with slight magnetisation the rays go closer to the terrella at the equator. 



No. 8 was taken during the same experiment as No. 7; but the hour-angle of the position is 270. 

 Here too we see, as in No. 4, the very remarkable continuation on the annular screen of the 

 positive precipitation on branch No. 2. 



Nos. 9 12 are all from one experiment, in which the magnetising current was 20 amperes, and 

 the tension in the discharge 2900 volts. The hour-angles of the positions were 90, 235 (with 15' 

 declination), 270, and 295 (with 20 declination). The polar ring of light on the night-side is fainter 

 in No. 9; but the shadow of the suspending wire is very clear. No. 10 shows the positive precipitation 

 upon branches i, 2 and 3; but there is no distinct negative polar ring. 



There is a faint negative polar ring in No. u. In this photograph, the great peculiarity is perhaps 

 the shadows behind branches 6 and 7. 



In No. 12 there is scarcely any of the usual negative precipitation on branches 4 and 5. 



Nos. 13 18 are from a very important experiment with a very small terrella of only 2.5 cm. 

 diameter. The iron core in it was cylindrical, and measured 10 mm. in diameter, and was wound round 

 with 240 turns of 0.4 mm. copper wire covered with silk. 



This terrella was placed in the middle of a flat screen, in such a manner that the magnetic axis 

 was at right angles to the screen. The object of the experiments made with this tiny terrella in the 

 vacuum-box of 22 litres, was to prove that the lines of precipitation that appeared on the screen had 

 nothing to do with the enclosing plates of the vacuum box. It was possible that our former terrellas 

 were too large in proportion to the vacuum-box; but it will be seen that the experiments with this little 

 terrella show our previous results to be unaffected as far as the distribution of the rays nearest the 

 terrella are concerned. 



No. 13 shows the terrella with screen seen edge-wise. The hour-angle of the position was 180. 

 The luminous ring outside the terrella is only from the cathode in the background. 



No. 14 shows discharge without magnetisation of the terrella, the hour-angle of the position being 

 270. There are shadows behind the terrella. The discharge took place with 2700 volts and 23 

 milliamperes. 



No. 15 shows the conditions with a magnetising current of 2 amperes, 3000 volts. 



