996 Sir J. J. Thomson on a 



down to the state in which it is showing striations, there is 

 a complicated distribution of positive and negative charges 

 through the tube which must have involved as a preliminary 

 considerable movements of the positive ions along the tube. 

 Let us consider a column of gas containing two electrodes 

 between which a large difference of potential is suddenly 

 established. Any electrons which may be in the gas will be 

 driven by the electric field towards the anode, acquiring, if 

 the field is strong enough, sufficient energy to ionize the 

 gas and produce fresh electrons ; these in turn will acquire 

 energy enough to produce other electrons, and thus the 

 number of electrons will increase with great rapidity as we 

 approach the anode. The electrons will be driven to the 

 anode, while the assemblage of positive ions in the neigh- 

 bourhood of the anode will sweep down the tube towards 

 the cathode, giving ont light whenever they get neutralized 

 by combining w T ith an electron. Thus there will be a patch 

 of luminosity sweeping down the tube towards the cathode 

 before the orderly regime indicated by the stratified discharge 

 is established. The experiments made by Spottiswoode and 

 myself give indications of an effect of this kind. 



Potential Difference between consecutive Striations. 



A considerable number of measurements have been made 

 of the potential difference between consecutive striations; 

 some observers seem to have expected that this difference 

 should be equal to the ionizing potential of the gas. From 

 the considerations given above, however, we should not 

 anticipate that this relation would be always fulfilled. If, 

 as in the case considered on page 988, the kinetic energy 

 of an electron falls to a very low value between the bright 

 parts of two consecutive striations, then since at these parts 

 the energy of the electrons must be greater than that 

 corresponding to the ionizing potential, the potential differ- 

 ence between the centre of a bright part and the centre of 

 the adjacent dark parts must be greater than the ionizing 

 potential. But in the case discussed on page 986, where 

 the variations in the energy of the electrons along the line 

 of discharge is small, and where the electrons have not to 

 pick up their energy between one striation and the next, 

 there is no need for a relation of this kind. And, in fact, 

 we can easily see that there must be wide variations with 

 the pressure in the potential difference between the bright 

 parts of consecutive striations. AVe found (see page 987) 

 that the distance between consecutive bright portions is 



