154 BIRKELAND. THE NORWEGIAN AURORA POLARIS EXPEDITION, 1902 1903. 



These figures will throw considerable light upon the questions we are endeavouring to solve. 

 The arrangements for the experiments are made plain by fig. 67, in which a is the discharge-tube 

 with terrella, b-b a 2o,ooo-volt generator with motor, c a static Kelvin volt-meter up to 20,000 volts, 

 d, d, d are photographic apparatuses, e-e. is an oil-pump with motor, from Siemens-Schuckert, / a mer- 

 curial pump worked by hydraulic pressure, for measuring the gas-pressure in the discharge-tube, and 

 g a Gaede pump with motor from Leybold's Nachfolger, the best mercurial pump that I know of for 

 obtaining a high degree of exhaustion in large tubes. 



The nine photographs in fig. 68 are taken with the terrella always in the same position, but under 

 three different electric and magnetic experimental conditions. The photographs are taken, as fig. 69 



shows, simultaneously from three sides of the terrella. The photographs 

 i, 2, 3, fig. 68, belong to one experiment, 4, 5, 6 to another, and 7, 8, 9, 

 to a third. In all the experiments, it is noon at the magnetic south pole, 

 the cathode representing the sun. 



The intention of the three experiments is to show how the descent 

 of rays upon the terrella alters when the stiffness is continually decreasing. 

 The first experiment shows the result when the stiffness of the rays is very 

 great in proportion to the magnetisation employed upon the terrella. The 

 stiffness of the rays is altered most simply by altering the pressure of the 

 gas in the discharge-tube. With an exceedingly low pressure, however, 

 the disadvantage is that so much gas is evolved from the cathode during 



i- 1 



the experiment, that it is not easy to photograph the phenomena, as they 

 change. 



In the first experiment (i, 2, 3) therefore, I have been obliged, for the sake of the photographing, 

 to keep a comparatively high pressure in the discharge-tube, but on the other hand I have employed 

 a lower magnetising current upon the terrella than in the next two experiments (4, 5, 6 and 7, 8, 9). It 

 has, however, been proved with certainty that the light-figures will be the same if, in the first experi- 

 ment, the same high degree of magnetisation be employed as in the second and third experiments, when 

 the discharge-tube is exhausted sufficiently. 



In the first experiment, the magnetising current was 15 amperes, answering to a magnetic moment 

 M, of the terrella, of 6200 C.G.S. The pressure in the discharge-tube was 0.018 mm., the discharge 

 current was 8.9 milliamperes, and the difference of potential between the electrodes was 4200 volts. 



In the second experiment the magnetising current was 33 amperes, answering to about M = 10,000. 

 The pressure was about 0.006 mm., the current 9.5 milliamperes, and the tension 5 500 volts. 



In the third experiment M 10,000, as in the second. The pressure was 0.03 mm., the strength 

 of the current 8 milliamperes, and the tension 3300 volts. 



As most of the experiments described in this volume were made with the same terrella, marked 

 No. 5 there may be some interest in seeing the curve for its magnetic moment at about 20 C. for 

 various intensities of the magnetising current. Fig. 70 shows this moment-curve. 



The values for high current-intensities are not very exact, owing to the great changes of tempera- 

 ture during the measurements. 



There are various circumstances that appear in the experiments represented in fig. 68, to which 

 we will pay special attention. 



It should first be remarked that if the rays become still more pliant than in experiment 3, the 

 conditions in the fundamental experiment represented in fig. 47 can be exactly obtained. 'In that experi- 

 ment, three regions for the descent of the cathode rays were distinctly seen in a zone round each of 

 the magnetic poles. 



