i82 Dr. E. Goldstein on the Electric 



at extremely small densities. But the kathode-light extends 

 through the whole vessel, without being influenced by the 

 proximity of the anode, so far as rectilinear rays radiating from 

 a can reach. A bundle of rays penetrates into the widest of the 

 three cylinders (Z 3 ), whose diameter is determined by that of 

 the communicating opening between the cylinders. When 

 the exhaustion is continued, the bundle penetrates as far as 

 the wall B, and the ends of the rays excite there bright green 

 phosphorescence in the form of a circular disk, which is the 

 section of the bundle of rays by the wall B. If now c be dis- 

 connected from the induction-coil, and the electrode b in the 

 second be made the anode while a remains the kathode, then 

 there appears, as represented in the figure, a long stratified 

 column of positive light, beginning several centimetres from 

 the mouth of Z 1? completely filling Z 2 , and reaching to the 

 anode b. Z 3 remains, as before, free from positive light; but 

 the bundle of blue negative light and phosphorescent surface 

 on B remain visible as before, affording a decisive proof that 

 the kathode-light penetrates into and through the positive 

 light. 



The green disk disappears as soon as c or b instead of a is 

 made the kathode — when, in short, anv electrode is made ka- 

 thode whose rays have some other direction than those issuing 

 from a. The quantitative differences which the positive and 

 negative light show remain when they mix, as if in the space 

 which they together fill each had separate and independent 

 existence. 



The assumption that the discharge from the negative light 

 propagates itself into the positive layer next the negative pole, 

 then into the second layer, and so on, necessitates the further 

 assumption that the discharge in the last-considered phase, 

 after it has traversed the negative light to its end (in the posi- 

 tive light), leaps back again to form the first layer of the 

 positive light, and then to traverse a second time as positive 

 light the space already once traversed as negative light under 

 the same conditions. But even with this the complication is 

 not yet exhausted of new assumptions to which this represen- 

 tation of the discharge, at the first sight so simple, leads. I 

 have convinced myself that even the secondary negative light 

 which radiates towards the anode from each contracted portion 

 of the tube penetrates into the positive light which follows 

 after the contraction ; we should therefore have in each tube 

 the leaping-back of the electricity, and its passage, once as 

 positive light and once as negative light, as many times as the 

 tube possesses contracted points. 



If, now, we have again, as in fig. 8, a plane at right angles 



