'90 Delta Rays from Gases. 



large. It is possible, accordingly, that the smallness of the 

 numbers for //\ = 0'35 is due to the preponderant action of 

 the nitrogen, while the constancy of the numbers for some- 

 what greater values of I/X is due to the compensating action 

 of the two gases. At any rate, the difference between the 

 results for hydrogen and air, the fact that the numbers are 

 almost uniformly greater in the case of air for the same 

 value of Y and l/X, is exactly what would be expected to be 

 the result of the presence of a strongly electronegative gas. 



But the main interest of the experiments lies in the 

 question whether they can give any information as to the 

 initial velocity of the 8 rays from gases. It appears necessary 

 to suppose that these 8 rays have a finite initial velocity, and 

 that it is of the same order of magnitude as that of the 

 8 rays from metals. No other explanation can, I think, be 

 given of the fact that in no case is the current through the 

 gas saturated when Y = 20, even though the currents due 

 to the 8 rays from the electrodes and that due to the positive 

 ions is saturated with that potential. If the constancy of 

 the numbers in the Table over a certain range of l/\ means 

 that within this range (1) is applicable, we should have to 

 conclude that the 8 rays from the gas are, on the average, 

 a little faster than those from the electrodes, since the num- 

 bers in the Table are a little smaller than those calculated 

 on the assumption that / (V) ==f(V). 



However, it does not appear that this method is capable of 

 giving any very precise indications concerning the properties 

 of these rays. They do appear to show that it is necessary 

 to suppose that the rays have a finite initial velocity, and so 

 provide evidence for an important proposition which was 

 universally accepted but had never been directly proved. 

 But this discussion has been published mainly because a 

 brief account of the measurements had been given without 

 any attempt at an explanation of them. It seems that they 

 are generally in accordance with the recent addition to 

 •our knowledge of the interaction between electrons and 

 molecules. 



Summary. 



The theory is considered of the conduction through an 

 ionized gas when the pressure of the gas is so low that the 

 free path of an electron is not small compared with the 

 distance between the electrodes, and the number of ions 

 made in the gas is not large compared with the number of 

 S ravs made at the electrodes. The discussion is based 



