336 CARNEGIE INSTITUTION OF WASHINGTON. 



On the conduction of electricitj' through an ionized gas, more particularly in its relation to 

 Bronson resistances. W. F. G. Swann and S. J. Mauchly. (Read at the meeting 

 of the American Physical Society, Washington, April 21, 1916.) 



The paper forms an experimental and theoretical investigation of certain 

 phases of the peculiarities exhibited by Bronson resistances. 



In most of the experiments the active substance used was ionium. The 

 current potential curve, in some types of cell, shows an increase of resistance 

 with field for small fields, but as the field is further increased, the resistance 

 starts to diminish. This phenomenon has nothing to do with ionization by 

 collisions, since the fields concerned are too small; it is shown, however, that 

 the Ijchavior of the cells in the above, and in other respects, can be satis- 

 factorily accounted for bj'^ taking into consideration the 5 rays which are 

 emitted by the a particles. The theory of the action of the 5 rays in causing 

 the current to increase more rapidly v»dth the potential difference between 

 the plates than would be accounted for by a linear law is practically the same 

 as that applicable when a field is maintained between two plates, one of 

 which is raised to incandescence so as to cause it to emit negative ions. The 

 nature of the phenomena is such as to indicate that 8 rays are emitted from 

 the inactive plate even when the latter is out of the range of the a particles, 

 and it is necessary to assume that these 5 rays are emitted as the result of 

 bombardment by other 5 rays emitted from the active plate Avith speeds suffi- 

 ciently high to enable them to travel across the air-space between the plates. 

 The amount of such high speed 5-ray radiation seems very considerable, and 

 from different standpoints it is shown that the ratio of the number of 8 rays 

 which finally appear, direct or indirectly, to the number of a particles emitted 

 from the active plate, is much larger than the numbers which have been found 

 in experiments at low pressures. Two lower limits obtained from independent 

 standpoints indicate for this ratio values 2,500, and 4,000, respectively. 



It was found that the behavior of a Bronson cell, as regards its constancy, 

 depends largely on the nature of the material employed for the plates. In the 

 case of aluminum, a large change in resistance was produced when the inactive 

 plate was sandpapered; indeed, it was found impossible to secure satisfactory 

 constancy with aluminum, the resistance fluctuating from day to day, and de- 

 pending, among other things, on the time which had elapsed since the inactive 

 electrode had last been cleaned. Effects of this kind were traced to variations 

 in the 5-ray emission from the upper electrode as a result of time variations in 

 the condition of its surface, these variations being somewhat analogous to those 

 experienced in the case of the photo-electric effect. It was found that the 

 action of the cell was greatly improved when silver plates were employed. 



Since the effect of the 5-ray emission is to cause the current to increase 

 more rapidly with the potential than would be accounted for by a linear law, 

 while the normal departure from linearity is in the opposite direction, it is 

 possible to choose the conditions so as to make the two effects partially com- 

 pensate. In this way linearity may be obtained over fairly large ranges of 

 potential. In practice it has been foimd desirable to make the inactive 

 electrode in two parts. It consists of a large disk supported by a tube, and 

 a smaller disk supported by a rod which slides in the tube. Approximation 

 to the desired linear relation is then secured by suitably adjusting the posi- 

 tions of the parts of the compound electrode with respect to the active plate. 

 The active material is spread over a shallow tray sunk in a thick metal plate; 

 it is covered with a piece of thin mica hermetically sealed to the plate, and 

 the mica is covered with silver foil held down by a silver-plated ring. 



Reasons are given for the very large temperature coefficients shown by sonie 

 types of Bronson cells and various other features of the cells are discussed. 



