Experiments on Positice Rays. 231 



at some places than at others, thus having a beaded appear- 

 ance. In others again (fig. 18) the line due to the atom 

 is of uniform intensity, while that due to the molecule is 

 very thin at the head and remains so for some distance, 

 and then develops great intensity. The point, however, on 

 which I wish to lay stress is that when there are such differ- 

 ences between the lines due to the atom and molecule of hydrogen, 

 there are similar differences between the lines due to the atom 

 and molecule of oxygen. In other words, that all the atomic 

 lines have common characteristics, as have also all the 

 molecular lines, but that chese characteristics are frequently 

 quite different. A similar effect can be observed in mixtures 

 of hydrogen and nitrogen, and indeed in all the mixtures I 

 have tried. As a general rule, the intensity along the lines 

 corresponding to the atoms is more uniform than that along 

 the lines corresponding to the molecules: the uniformity of 

 the atomic line is especially marked in the lines corresponding 

 to atoms with two charges, such as C ++ N ++ ++ . It is 

 also noteworthy that on the lines corresponding to the 

 monatomic elements helium and mercury w T e get a blending 

 of the characteristic features of both the atomic and the 

 molecular lines. 



The difference between the atomic and the molecular lines 

 suggests that the charged atoms and the charged molecules 

 are produced by different processes ; that in fact there is more 

 than one process of ionization at work, and that the process 

 which produces the charged atom is different from that 

 which produces the charged molecule. AVe have in the 

 discharge-tube rapidly moving cathode particles moving 

 away from the cathode, and positively electrified atoms and 

 molecules moving towards it; each of these systems can, as 

 is well known, produce ionization by collision. With regard 

 to the ionization due to the cathode rays, there are several 

 reasons for thinking that in the dark space itself this is not 

 considerable. For if a cathode ray, starting from the cathode, 

 produced on an average in its course through the dark 

 space another corpuscle, this second corpuscle would, under 

 the influence of the strong electric field w T hich exists in that 

 space, acquire a considerable velocity, though not of course 

 so great a one as that of a particle which had started from 

 the cathode and passed through the whole of the dark space. 

 There would thus be a considerable range of velocity in the 

 beam of cathode rays emerging from the dark space, and no 

 approach to a condition where all the cathode rays in the 

 bundle are moving with the same velocity. Now 7 , although 



