658 BIRKELAND. THE NORWEGIAN AURORA POLARIS EXPEDITION, igO2 1903. 



When the magnet was in operation during the discharge, a luminous column of magneto-cathode 

 rays, cruciform in section, was sent out towards the sheet of glass, where it formed a reduced represen- 

 tation of the cross. 



The question that interested me here, however, was how the metal corpuscles expelled from the 

 palladium cross would be deposited. 



It appeared that if the cathode were cold, the discharge-current being kept small, a light cross 

 upon a darker ground was thrown upon the sheet of glass at the bottom of the tube. There was very 

 little deposit upon the sheet of glass where the column of light with cruciform section had heated the 

 glass, while beyond this there was the normal deposit of palladium. 



The whole thing was different when we employed up to 30 milliamperes and more per square 

 centimetre of the entire surface of the cathode, which thereby became highly incandescent. 



After the experiment, which lasted about one minute, there was an intensely metallic cruciform 

 deposit where the column of light had struck the glass, a cross of reduced size, opaque and shining 

 while outside it was a dark, semi-transparent, normal palladium-deposit. 



In addition to this normal deposit upon the plane sheet of glass, there was a strongly-marked 

 ring of evenly dark deposit upon the cylindrical surface of the vacuum-tube, nearest the cathode-cross. 



There were thus distinctly two kinds of metal corpuscles ejected by the cathode, first the normal 

 corpuscles that seem to be expelled from the cathode without being influenced to any great extent, 

 either by electric or magnetic forces; and secondly a kind of corpuscle that accompanies the magneto- 

 cathode rays, and these corpuscles are capable of attaching themselves to the glass wall, provided the 

 velocity with which they reach it is sufficiently great. 



This circumstance may possibly favour RIGHI'S idea that these magneto-cathode rays consist of almost 

 neutral "double stars" of positive and negative ions, expelled in exactly the direction of the magnetic 

 lines of force, thus possibly a combination of a negative electron with a positive metal ion, which, under 

 certain circumstances, can be deposited and form a metallic coating upon the glass wall of the vacuum-tube. 



I then went on to find out whether I could discover any twisting of this cruciform metallic deposit, 

 such as I have proved in the case of a cross of cathode-rays under similar circumstances (see the pre- 

 viously-mentioned paper in Archives des Sciences Phys. et Nat.). These crosses of cathode-rays turn 

 clockwise when a magnetic north pole is employed and the cross looked at from the pole. We can 

 thus find out whether the metal corpuscles in the cruciform deposit were negatively or positively charged, 

 by noting the direction in which the cross was eventually turned. 



It soon appeared that the twisting was at any rate too small to be demonstrated directly by these 

 experiments. The experiment was therefore modified by forming the cathode as a plane, long rectangle 

 of thin palladium, which .was attached in such a manner that it stood edgewise upon the sheet of glass, 

 with its long side parallel to it. 



The intention with this arrangement was to obtain a sharp linear deposit where the magneto- 

 cathode rays came in contact with the glass. 



The experiment was carried out at first with a south pole in front of the sheet of glass, and then, 

 at the same distance, a north pole, while the discharge was going on evenly all the time. 



Neither was it possible, however, in this way to obtain a double turning-angle of measurable size, 

 perhaps because the deposit-lines were not particularly sharp. 



Photographs of the cruciform and linear deposits are here reproduced, from three experiments in fig. 246. 



In the first and second experiments 15 milliamperes was used in the discharge and 10 amperes 

 the magnet. In the third the discharge-current was the same, but the current to the magnet was first, 

 22 amperes in one direction, during one minute, and then in the reverse-direction for one minute, sc 

 that the pole before the cathode changed from S. to N. 



