560 BIRKELAND. THE NORWEGIAN AURORA POLARIS EXPEDITION, IQO2 1903. 



It should be remarked that the position of the various lines of precipitation upon the horizontal 

 screen, depends somewhat upon the magnetising of the terrella. If, for instance, the current upon the 

 terrella is increased from 8 amperes to 30, the first line of precipitation will move back (westwards) a 

 little, with a parallel movement. The second and third lines of precipitation move at the same time; but the 

 angle between the first and second, and between the second and third lines of precipitation continues more 

 or less to be about uo 100. This angle, however, also diminishes somewhat under stronger magnetisation. 



In this experiment, another circumstance was also investigated. A thin screen, about 3 mm. in 

 height, formed of a strip of copper, was placed on its edge upon the terrella, and running from the 

 latter's north pole a little way down a meridian. The screen was then divided into three branches, and 

 was also coated with tungstate of lime. It was so placed that the rays which came through holes i and 

 2 would strike the terrella in the polar regions just where the little three-armed screen was. The in- 

 tention was to determine the direction in which those rays moved which struck the terrella in its polar 

 regions. It appeared from the experiments that the rays in the second polar precipitation, which belonged 

 to the very northernmost part of the precipitation up in the auroral zone, come fairly perpendicularly in 

 towards the terrella, though with some slight movement eastwards, while the rays both in the south- 

 western and south-eastern parts of the precipitation on the terrella to the north of the horizontal screen, 

 had a strong tangential movement, with direction from west to east. 



STUDY OF THE RAYS OF GROUP A. 



109. Experiment in which the Terrella had only a Vertical Screen. We shall begin by 

 describing a series of experiments which were made with the same terrella as before, in which the 

 magnetic poles coincide with the geographical; but the terrella now has only one screen. 



This screen, which maintains a vertical position during the rotation of the terrella, was produced 

 by an abrupt bending of the former vertical screen, so that the latter comes to consist principally of two 

 plane portions, which intersect one another at an angle of about 100 in a vertical line, which is in 

 contact with the terrella in its magnetic equator. We will call the screen the vertical screen. The photo- 

 graphs here reproduced give a sufficiently clear idea of its form. 



Fig. 203 shows 12 pictures from experiments with this arrangement with a vertical screen pro- 

 vided with a horizontal slit. 



Nos. i, 2 and 3 are from experiments in which the terrella was magnetised with 8 amperes, the 

 discharge-current was of 25 milliamperes, and the pressure answered to o.ooi mm. The outer plane 

 part of the vertical screen formed an angle of 45 with the central line between the centre of the ter- 

 rella and that of the cathode. We shall simply, in the following pages, express this by saving that tlu- 

 screen had an hour-angle of 45, referring only to the outer plane part of the screen. 



The photographs were taken in a horizontal plane through the centre of the terrella, from places 

 with hour-angles of respectively 90, 180, and 270. 



When the screen here has an hour-angle of 45, it does not to any great extent shut oft' the rays, 

 and the light-figures on the terrella (Nos. i, 2 and 3) are very much the same as if there had been 

 no screen. We recognise them from fig. 68 in Section I. 



Photographs 4, 5 and 6 are from experiments made under very nearly the same conditions as the 

 preceding, except that the hour-angle of the screen is 135. The photographs are taken from the same 

 positions respectively. 



We here obtain a capital representation of the way in which the screen acts when the slit does 

 not fall near one of the lines of intersection of the rays, those lines which, on the horizontal screen, 

 we called lines of precipitation. 



