568 BIRKELAND. THE NORWEGIAN AURORA POLARIS EXPEDITION, I QO2 1903. 



and form a new line of intersection above the second slit, through which the rays also pass. We 

 see all three precipitations upon the terrella, and the third line of precipitation upon the horizontal 

 screen could also be distinguished, although faintly, during the experiment. It fell on the margin of the 

 third slit. 



Some interesting experiments were made for the purpose of throwing light upon the origin of 

 the secondary precipitation, of which frequent mention has been made in describing the experiments 

 with a vertical screen (see figs. 203 & 204), and which was found again here under different conditions. 



Nos. 7, 8 and 9 are of experiments with a pressure of 0.0014 mm - a discharge-current of 18.5 

 milliamperes, a tension of 3300 volts, and a magnetising current of 24 amperes. The photographs were 

 taken from places with hour-angles of 90, 180 and 270. 



Slit i was placed at 125, just so that a small pencil of rays fell through the screen at the end 

 nearest the terrella. This little pencil of rays which thus passed through the screen, at once gave rise 

 to a distinct, but faint, precipitation upon the terrella. Even the third precipitation was single, without 

 any secondary precipitation; but there is a strange precipitation upon the horizontal screen in which our 

 attention is especially attracted by a line of precipitation almost parallel with the first line of precipita- 

 tion, and only a few millimetres east of it. 



That this secondary precipitation on the horizontal screen is produced by rays that have passed 

 through slit i at the end nearest the terrella, is apparent from the fact that if the terrella is turned so 

 that the first line of precipitation falls either entirely on the west side of the slit, or entirely on its east 

 side, the secondary precipitation completely disappears in both cases. 



The next experiment was to let the rays of the secondary precipitation through slit i at the same 

 time as the first main precipitation came through it; for the distance between the two precipitations was 

 rather less than the width of the slit. 



It appeared that as soon as the secondary precipitation also passed through the slit, new precipi- 

 tation made its appearance both on the horizontal screen and on the terrella, a secondary precipitation 

 suddenly appearing in the polar regions on the night-side, similar to the primary precipitation lying 

 immediately to the west. 



Nos. 10, ii and 12 are from an experiment which shows this. The pressure was 0.0008 mm., the 

 discharge-current 17 milliamperes, and the magnetising current 24 amperes. The photographs were taken 

 from the same positions as the preceding ones. 



We see distinctly that the third polar precipitation consists of two consecutive precipitations. That 

 on the east is the secondary. 



The experiment was repeated several times without being photographed. Again and again it 

 appeared that when the first secondary precipitation upon the screen passed through the slit, a new 

 precipitation was formed nearer the second slit, the innermost part of it falling through that slit at the 

 end nearest the terrella, thereby producing the secondary precipitation upon the night-side of the ter- 

 rella, in the polar regions, farther out on the night-side than the first, which was there already. 



It was distinctly seen that the second secondary precipitation upon the horizontal screen formed 

 a much smaller angle with the first precipitation than did the second principal precipitation. It was the 

 outer part of the first secondary precipitation which, by passing through the first slit, produced a new 

 line of precipitation not more than 50 farther east upon the horizontal screen. Only because the end 

 of the second slit nearest to the terrella was comparatively wide, did a pencil of rays from this preci- 

 pitation pass through there, and occasion the secondary precipitation after the second polar precipitation 

 upon the night-side. 



We have previously touched upon the possibility that the connected polar precipitations upon the 

 terrella (in the auroral zone) were composed of a whole series of close-lying secondary precipitations 



