PART II. POLAR MAGNETIC PHENOMENA AND TERRELLA EXPERIMENTS. CHAP. V. 



637 



Fig. 234. 



a globe as cathode in a large vacuum-tube, several concentric luminous envelopes separated by dark 

 spaces. These different envelopes are more distinctly seen when the globe used as cathode is mag- 

 netised. In this case the originally spherical envelopes will be flattened so as to form a ring in the 

 magnetic equator. Fig. 235 gives a representation of such an experiment. Such envelopes, as we know, 



contract or expand according as the gas-pressure in the vacuum-tube 

 becomes greater or less. The very singular phenomenon of the contrac- 

 tion of the comet's head with the approach of the comet towards the sun 

 can be reasonably explained by this view. Instead of expanding, as one 

 would naturally expect it to do under the action of solar heat, the comet's 

 head contracts when near the sun, just because the gas pressure about 

 the comet becomes higher there, and the electrically-formed luminous 

 envelopes therefore contract. 



On some occasions comets have been furnished with several tails 

 in a manner that is not quite easy to explain by the assumption that 

 an emanation of tail-material from the comet could directly give rise to 

 all the tails. 



Figs. 2363 &b show respectively the famous Donati's comet (1858) 

 from a drawing by BOND, and the comet of 1744 by M'i? KIRCH at the 



Berlin Observatory. It seemed to me it would be worth while examining whether all the luminous 

 streaks or tails that were seen were perhaps not separate tails, but might possibly be compared with 

 positive strata in the electric discharge from the negative comet-head such as in the discharge repre- 

 sented in fig. 231, 2. 



I have taken two ways for determining this. First the 

 angle was calculated, the angle that a plane through the 

 centre of the earth and a luminous streak in the tail, formed 

 with the plane of the comet's orbit. The result for Donati's 

 comet was 



= 58-99 



I'm- a streak that passed over and / Coronet Borenlis on Oc- 

 tober Qth. The calculation is based upon a description by WIN- 

 NECKE, quoted in Bond's "Account of the Great Comet of 1858" 

 Ip. 61), Annals of the Astronomical Observatory of Harvard 

 College, Vol. Ill, Cambridge, 1862. 

 It was further found that 



= 6 9-53 



for a streak that issued from the head, and kept separate from 

 the tail, passing over 6 Scrficntis and ft Herculis, according to 

 a drawing of the comet on October Qth (1. c.). 

 For the comet of 1744 it was found that 



a = 87.36 



for a streak that, according to a description by LOYS DE CHESEAUX at Lausanne, of the appearance of 

 the comet on the night of the 7th March (quoted in J^-.GERMANN'S "Mechanischen Untersuchungen iiber 

 Cometenformen", pp. 397 & 398), passed through the middle of EQUULEUS and ended in a point of which 

 the longitude was 319 55', and latitude -f- 34 35'. 



Fig. 235. 



Birkeland. The Norwegian Aurora Polaris Expedition, 1902 1903. 



81 



