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



671 



Fig. 254 shows the whole arrangement with the new vacuum-box of 320 litres. Floor and ceiling 

 re here made of 12 mm. steel plates, the pillars between are of bronze, and the sheets of plate-glass 

 t the sides are 30 mm. in thickness. 



The experiment shows the "zodiacal-light ring". It requires little magnetising of the globe (i 1.3 cm. 

 i diameter), but a great discharge-current (up to 100 tnilliamperes). Similar experiments are shown in 

 gures 255, i and 2. In the former the magnetic globe is only 2.5 cm. in diameter; but it was easy, 

 specially with greatly rarefied hydrogen gas in the box, to obtain a plane of rays about the globe that 

 nt all four glass walls in brightly phosphorescent, straight stripes from 5 to 10 millimetres wide. 



It is easy to prove that the plane of rays is partly formed of rays from the upper hemisphere of 

 cathode, that are bent down towards the equator, and rays from the lower hemisphere that are bent 



Fig. 256. 



awards. It will without doubt be possible to produce, with a very small cathode-globe, a ring greater 



proportion to the globe than is the real zodiacal-light ring in proportion to the sun, even if the latter 



ng be assumed to go right outside the earth's orbit. It is only by careful adjustment of the magneti- 



ition of the globe, however, that the ray-masses are made to coincide, so to speak, exactly in one plane. 



In general, the ray-masses from above and from below intersect one another in the plane of the 



juator; and it is easy to form round the circle of intersection a strongly luminous ring, floating in 



>ace round the globe, and resembling a nimbus such as painters in olden times painted round the heads 



' saints. 



The intersecting groups of rays may often be found upon the glass walls in the form of two se- 

 irate parallel phosphorescent bands of light that can be moved to and fro by slight variation in 

 agnetisation. I believe I have seen these groups of rays twice form circles of intersection (node- 



