4:88 Mr. A. B. Meservey on the Potentials required 



it would occupy if this effect were not present, and would 

 fall farther below the position as the pressure continued to 

 diminish. If the effect were sufficiently great, it would 

 eventually balance the effect of the non-uniformity of the 

 field upon the ionization by collision, the two curves would 

 cross, and below that point the radiation effect would keep 

 the positive curve below the negative. This process would 

 correspond exactly with what takes place in every case in 

 the present experiments, for the small diameters ; for the 

 larger diameters the tendency is the same, though the actual 

 point of crossing does not come within the limit of available 

 potential difference as the field becomes more and more 

 nearly uniform and the curves approach that for parallel 

 plates. 



The fact that the curves cross in different instances at very 

 nearly the same pressure might also be explained on the 

 radiation hypothesis. As the diameter of the inner cylinder 

 decreased, the drop of potential near the positive inner 

 electrode would become more abrupt, so that on the average 

 the negative ions would strike the metal with greater velocity. 

 The amount of the radiation effect tending to lower the 

 positive curve would therefore be greater. On the other 

 hand, the initial difference between the curves, due to the 

 regular process of ionization by collision, would be greater 

 in the less uniform field. The relative increase in the radia- 

 tion effect with the diminution of pressure would probably 

 remain the same, so that the larger amount of radiation 

 might balance the larger initial difference at the same pres- 

 sure at which the smaller amount of radiation would balance 

 the smaller initial difference. 



In addition to the fact that the radiation hypothesis fits in 

 with the characteristics exhibited by these curves, it may 

 be mentioned that the preliminary readings of some 

 experiments' now being carried on indicate the presence of 

 some kind of radiation, at the pressures under consideration, 

 the amount of which varies directly with the force and 

 inversely w 7 ith the pressure. 



The curves in fig. 5 are just what the collision theory 

 would lead us to expect for pressures somewhat above the 

 critical pressure, (hving to the concentration of force near 

 the inner electrode, the sparking potential when that electrode 

 is negative should be lower than when it is positive and 

 lower than the parallel plate value. Decreasing the diameter 

 of the inner cylinder produces a still greater concentration 

 of force near that electrode, and a corresponding weakening 

 of the rest of the field. Such a decrease, therefore, should 



