234 Sir J. J. Thomson on 



reproduced in fig. 13 ; it is now convex to the origin and of 

 the type shown in fig. 11. The value of mje for this curve is 

 12 times that for the hydrogen atom, and we conclude that 

 the ions which produce it do not start between the cathode 

 and the photographic plate, but are complex ions (He 3 ) + , 

 which are formed in the dark space in the discharge-tube and 

 are liable to lose their positive charge on their journey from 

 the cathode to the photographic plate. 



The secondary rays present some very interesting features. 

 The following are some of the results obtained by the study 

 of a long series of photographs. We shall for brevity call 

 the ratio of mje for any ray to the value of m/e for the atom 

 of hydrogen the electric atomic weight of the particle forming 

 the ray. 



Secondary rays consisting of the atoms of hydrogen for 

 which ejm = 10 4 are found in all gases when the pressure is 

 not too low. They occur negatively charged as well as 

 positively, and often there is a bright head at the end of the 

 negatively charged rays corresponding to the head of the posi- 

 tively charged ones ; this is shown in the photograph repro- 

 duced in fig. 8. It is due, I think, to secondary rays which 

 have been produced before the magnetic field was reached, 

 and which have all, therefore, been exposed to the full 

 magnetic and electric field and, having suffered the same 

 deflexion, strike the photographic plate at the same spot. 



Secondary rays consisting of the molecule of hydrogen. 

 Though the primary rays corresponding to the molecule of 

 hydrogen occur on nearly every plate, the secondary radia- 

 tion due to the hydrogen molecule is generally conspicuous 

 by its absence. An example of this is shown in fig. 5, which 

 represents the appearance when the gases in the tube are the 

 residual gases left after the air which originally filled the tube 

 has been pumped out. The absence of the secondary corre- 

 sponding to the hydrogen molecule is very marked, while the 

 secondary corresponding to the hydrogen atom is quite 

 distinct. I have found the same thing when the tube con- 

 tained hydrogen, oxygen, carbonic oxide, marsh-gas, cyanogen, 

 or hydrochloric acid gas instead of the residual gas. When 

 helium is in the tube there is at some pressures, though not 

 at all, secondary radiation corresponding to an electric atomic 

 weight of 2. A photograph of a plate taken with helium in 

 the tube is reproduced in fig. 16, and the secondary radiation 

 of this type is very prominent. I am inclined to regard it 

 in this case as clue to an atom of helium with two charges 

 rather than to a molecule of hvdro£en with one charge. I 

 had taken so many photographs before I obtained any 



