998 Sir J. J. Thomson on a 



Effect of Impurities on the Striations. 



Suppose the gas in the discharge-tube contains a trace of 

 impurity, and that this impurity has a higher ionizing 

 potential than the original gas. Suppose, too, that the 

 quantity of impurity is so small that the graph representing 

 thQ distribution of the energy of the electron along the tube 

 is not affected. Thus, to find the places where the light due 

 to the impurity will be excited, draw in fig. 1 the horizontal 

 line for which E = E ' where E ' is the ionizing energy for 

 the impurity. The parts of the tube where the light from 

 the impurity is excited will be the parts corresponding 

 to points on the graph of E which are above this line. If 

 E ' is so large that the horizontal line through it does not 

 cut the graph, the light from the impurity will not be 

 excited. If the line cuts the curve, then since E ; is greater 

 than E , the light from the impurity will be confined to a 

 narrow patch in the middle o£ the striations due to the other 

 gas. This will emphasise the brightness of the centres of the 

 striations and make the contrast between them and the dark 

 parts more distinct. We see in this way that the presence of 

 an impurity of the right kind might increase the brilliance 

 and distinctness of the striations. It is well known that stria- 

 tions are often made much more pronounced by impurities; 

 indeed, some physicists bave suggested that a pure gas does 

 not show striations. 



Since the maximum energy of the electrons is greater 

 when the bright parts are broad than when they are thin, 

 it is advisable, if we wish to observe the spectrum of an 

 impurity, to adjust the conditions so as to get broad 

 striations. 



In a very strongly electronegative gas like oxygen, where 

 an electron quickly gets attached to a molecule of the gas, 

 the absorption of the electrons would be very rapid, so 

 that the coefficient a which occurs in equation (11) would 

 be exceptionally large, and the condition k > I which is 

 necessary for striation would require the pressure to be 

 much lower than for a gas like nitrogen or argon, in which 

 the electrons do not combine with the molecules and where 

 the coefficient a has a much smaller value. The pressures 

 at which striations are possible in these very electronegative 

 gases might be so low that in tubes of: moderate dimensions 

 there would be no room for the positive column. If, how- 

 ever, the electronegative gas were an impurity in a gas like 

 nitrogen or argon, the free path of the electron would be 

 much longer than in the electronegative gas at the same 

 pressure, and the pressure at which striations occur might 



