KR. BIRKELAND M.-N. Kl. 



auroral rays are formed by just such distinct, comparatively small groups 

 of cosmic rays, which successively detach themselves from a larger bundle 

 of rays, after having passed through the magnetic equator ;/, (;/ + i), {n -\- 2), 

 &c. times, answering to the consecutive rays in the curtain. 



It is comparatively easy, from experiments with the terrella, to calculate, 

 within a small percentage, the difference in time that there would thus be 

 between the entrance into the atmosphere of the «*'^ and that of the {n -\- pf^ 

 auroral rays, at the time when the curtain is formed. 



Such terrella-experiments must of course be made with special care 

 in order to insure the great accuracy that is here necessary. 



From previous experiments, which were made without any reference 

 to the present special problem, I can make the deduction that the time- 

 difference in question must be something like V5 of a second from one 

 ray in the curtain to the next. This period, which represents very nearly 

 the time the ray takes to travel from the southern auroral zone to 

 the northern, will also depend upon the actual magnetic stiffness of the 

 cosmic rays. 



This period may also be calculated, and perhaps with a sufficiently 

 close approximation, by the mathematical theory alone, considering the 

 earth as an elementary magnet. 



Oil the other hand I am at this very moment able to point to a sure 

 method by which we can ascertain, by photographic registering, the moments 

 at which the rays in an auroral drapery make their appearance one after 

 another, with an accuracy in the time-differences of a few per cent. 



For eight weeks I have been working in Africa at various observa- 

 tions referring to the Zodiacal Light, and among other experiments have 

 made photometric investigations of variations in the intensity of that light. 

 In these investigations I employed an Elster and Geitel light-cell, and a 

 Cranier-Edelmann's wire electrometer. The oscillations of the electrometer 

 were photographically registered, and gave exceedingly interesting curves 

 for the light-variations in the zodiacal light, resembling the variations in a 

 magnetogram. The results of these investigations will soon be published. 



The apparatus employed for the investigation of the zodiacal light can 

 now be very simply arranged for rapid registering, so that it can register 

 the different light-impulses following one another in quick succession in 

 the formation of an auroral drapery, each time a new beam of rays darts 

 down from the sky to unite itself to the drapery. The observations should 

 be made when a northern-lights drapery is forming near the zenith, and the 

 observer need only approximately follow the margin of the auroral curtain 

 with the tube of the glass that contains the light-cell. 



