Magnetic Rays, 315 



to the conclusion that the magnetic rays are much more 

 easily produced in air and artificial mixtures of nitrogen and 

 oxygen than in the other mixtures we tried ; and they, in 

 tarn, are more sensitive to the influence of the field than 

 simple gases, including the chemical compound carbon 

 dioxide. 



As regards what has been called the induced column, that 

 is, the glow proceeding from the end of the magnetic rays 

 which acts as if it were a current of charged ions moving in 

 opposite directions from a virtual anode situated at a point 

 in the column, we could obtain it in a well developed form 

 only for air. While we also observed it with other mixtures 

 containing nitrogen, the characteristic phenomena were less 

 certain ; in other mixtures and in simple gases it was absent. 

 This seems to be correct, because if either pure nitrogen or 

 oxygen were in the tube and a little of the other gas intro- 

 duced, the induced column immediately declared itself. 



These results are difficult to account for by Professor 

 Bighi's hypothesis. If the magnetic rays consist of a stream 

 of doublets, each composed of a centre of a positively charged 

 gas-ion, or metal-ion detached from the cathode, why should 

 the doublets form more readily in a medium containing gas- 

 ions of one gas rather than another, and especially when two 

 gases not chemically combined are present. 



Again, Professor Righi supposes that those doublets, which 

 are made more stable by the action of a magnetic field and 

 are driven by it toward regions of less intensity, have their 

 stability decreased as their distance from the magnet in- 

 creases. These doublets are finally destroyed by collisions, 

 and their dissociated parts move on and ionize the gas. The 

 range of these particles is shown by an excess of free positive 

 electricity at a certain place in the tube which he calls a 

 virtual anode. It is difficult to see why i:he existence of this 

 virtual anode should depend on the gas. It might be claimed 

 that for a light ion like hydrogen the virtual anode would 

 be too near one end of the column, and for a heavier ion like 

 nitrogen or oxygen too near the other end to make the 

 reversed bending on the two sides of this anode evident. 

 But this can hardly be the explanation because, while the 

 column should then curve as a whole in one direction, yet 

 the directions should be opposite for the columns of the two 

 gases. This is not the case, for it always curved the same 

 way for all gases and mixtures when the column was present, 

 and the virtual anode could not be located. 



Lastly, the hypothesis requires that a doublet be driven 

 from an intense to a weaker part of the field, and this is in 



