Delta Rays from Gases. 89 



a little hydrogen was likely to be a decrease in the current 

 carried by the 8 rays from the electrodes. The numbers 

 given in the Table are obtained by subtracting from the 

 measured current that carried by those rays in a vacuum ; 

 if the rays carry less current when gas is present, the 

 numbers in the Table will all be smaller than than they 

 would be if (1) were applicable. If, owing to the presence 

 of gas, the current due to the 8 rays from the electrodes is 

 decreased by more than the extra current due to the ion- 

 ization of the gas, the negative figures in the first column 

 would be explained. The minimum in any column at about 

 3 volts could be explained if the effect of the field in counter- 

 acting the effects of reflexion do not begin to be appreciable 

 until V is greater than 3. Again, of course, on this view 

 the observed numbers will all be less than those calculated, 

 until the current carried by the 8 rays from the electrodes 

 becomes inappreciable compared with that due to the gas. 

 But when this stage is reached the effect of collisions in 

 increasing the current due to the 8 rays from the gas is 

 important, so that we immediately pass to the stage where 

 the observed numbers are greater than the calculated. The 

 data concerning the reflexion of the electrons by collision are 

 not so definite that any attempt can be made to explain the 

 results quantitatively ; and certainly it seems impossible to 

 deduce from these results any conclusion concerning the 

 initial velocity of the 8 rays from the gas. 



b\ Now let us turn to the results for air. Here again-, 

 the measurements when V is 30 volts or more are affected 

 by ionization by collision. Again, also, there is a general 

 increase in the numbers in each row with Z/X, but the in- 

 crease is not nearly so marked as before ; the numbers are 

 approximately constant between 1/\=1 and //X = 4, and in 

 this range they show some approximation to those calculated 

 from (1). The numbers in each column increase regularly 

 with V , but the rate of increase is small in the region where 

 in the previous case there was a minimum. 



In interpreting these results we must remember that, of 

 the two gases in air, nitrogen has no affinity for electrons * 

 and probably acts like hydrogen, while oxygen has a strong 

 affinity. Accordingly the nitrogen will tend to decrease the 

 current carried by the 8 rays from the electrodes, and so 

 make the numbers smaller than those predicted by (T), while 

 the oxygen tends to increase both this current and that due 

 to the 8 rays from the gas, and so make the numbers too 



* See the measurements on the velocities of the ions by Franck, 

 Deitfsch. Phys. Gcsell. xii. p. 613 (1010). 



