PART II. POLAR MAGNETIC PHENOMENA AND TERRELLA EXPERIMENTS. CHAP. IV. 595 



We have then on the one hand a magnetic terrella with a radius of 4.1 cm., near whose poles 

 the magnetic intensity amounts to 4500 C. G. S., and round which circle cathode rays whose velocity 

 is T *j c, answering to 2000 volts, when c indicates the velocity of light (cf. LENARD, Ann. d. Physik, 1903, 



317. P- 73 2 )- 



On the other hand we have the earth, with a radius of 6.4 X io 8 cm., and with a magnetic inten- 

 sity in the neighbourhood of the magnetic poles that may be put at 0.68 C. G. S., and round which 

 circle corpuscular rays with a velocity of $c, 



I now believe, that when the terrella is so strongly magnetised that the polar light-rings have the 

 same spherical diameter as the auroral zone on the earth, the cosmic ray system about the earth, which 

 occasions aurora and magnetic storms, is similar to the cathode ray system around the terrella. Thus all 

 details can be elaborated from our terrella-experiments and the results be applicable to the earth with a 

 suitable proportional factor. We shall also make repeated use of this important proposition. 



Now jf the conditions in the one case are, so to speak, a true copy of those in the other, the 

 radii of curvature of the corpuscular rays at all corresponding places must be as much larger than those 

 of the cathode rays as the proportion between the radii of the earth and those of the terrella. Thus 



Qo 4.1 



The proportion between the magnetic intensity at sets of places in the vicinity of the earth and in 

 the vicinity of the terrella will be 



H 0.68 

 7T = 4500 



Now we have, as is well known, 



m . u 

 ' ' a = - 

 e 



where // is the intensity of the magnetic field, Q the radius of curvature of the rays, m the mass of 

 the electric particle, e its charge, and u its velocity. 



For the corpuscular rays round the earth we have therefore 



and for the cathode rays round the terrella 



From this \ve obtain the important relation, 



fJ Q = 2.35 X 10< HoQo = 3 j x 10 



Even from this we may conclude that the rays in question must be unusually stiff magnetically. 

 HQ must be between i and io millions. We know only slightly penetrating positive rays which have 

 approximately so great an inflexibility, as H.Q for u rays from radium may have a value of 4 >( io 5 . 



W. WIE.N observed that on the negative side also of the magnetic spectrum of kanal rays, there 

 was a slightly deflected patch of fluorescence. These may possibly be almost inflexible negative 

 ion-rays. 



The y rays hitherto not magnetically deflected are presumably very much like Rontgen rays in 

 their nature. The opinion has been put forward that they are exceedingly stiff/!/ rays (PASCHEN), or that 

 they consist of neutral corpuscles (BRAGG). Possibly the corpuscles are not absolutely neutral either. 

 Even rays in which H . Q equals ten millions, there is hope of being able to deflect perceptibly by means 



