236 Sir J. J. Thomson : Farther 



not in this simple way detect the existence of doubly charged 

 molecules, because the parabola corresponding to the doubly 

 charged molecule would coincide with that due to the singly 

 charged atom. We can, however, indirectly get some 

 evidence on this point. Some of these doubly charged 

 systems lose one of their charges before passing through the 

 electric and magnetic fields : they are deflected by these fields 

 into the parabola corresponding to the singly charged atom ; 

 but inasmuch as they had when in the electric field in the dark 

 space a double charge, they will have acquired when in that 

 field twice the kinetic energy of a singly charged atom. They 

 will therefore occur on the parabola at a position which is only 

 half the distance from the vertical line of that occupied by 

 the atoms which have had a single charge during the whole 

 of their career. The effect of this is to make the parabolic 

 arc corresponding to the atom have a kind of beak. An 

 example of this is shown in fig. 14 (PI. IX.). Sometimes 

 this prolongation of the parabola is of nearly uniform in- 

 tensity, as in fig. 14; in other cases it is much brighter at the 

 tip than at any other position, giving the curve a beaded 

 appearance. Now it is remarkable that this prolongation of 

 the parabolic arc towards the vertical axis occurs, as far as 

 my experience goes, only on the lines corresponding to the 

 atoms, and not on those corresponding to the molecules : this 

 I think shows that the ions with double charges are atoms 

 and not molecules. I have not observed on the plates any 

 indications of molecules with two charges. Another que>tion 

 as to which we can get some information from the plates, 

 is whether the doubly charged atoms get their double 

 charge at one collision, or wmether they get one charge at 

 one collision, and the other at a subsequent one. If it 

 took two collisions to give them the two charges, then we 

 should expect that the number of these particles with an 

 amount of energy approximately that due to the fall of the 

 double charge through the whole of the dark space would be 

 small compared with that of those which had a smaller 

 amount of energy. For consider the condition which must 

 exist for the particle to have the maximum kinetic energy: 

 the two collisions which cause it must occur close together 

 at the boundary of the dark space away from the cathode. 

 Compare this with the region open for collisions which ionize 

 the gas so that it falls into the cathode with about half this 

 kinetic energy. Then, if the first collision occurs near the 

 boundary of (he dark space, a second collision near the 

 cathode will be required, but the particle in this case is not 

 obliged to make the first collision close to the boundary— it 



