SOME CONTEMPORARY ADVANCES IN PHYSICS— XI 



467 



its negative end — is taken as the zero from which the other potentials 

 are measured. In the sketch (Fig. 1) this is marked F. Close to the 

 source there is a gauze (d) maintained at the controllable potential V 

 and thus providing the potential-rise by which the electrons are accel- 

 erated. It is clearlv desirable that the electrons should move at their 



G, 



G, 



-tWVWVWWVW* 1 



-t I I I ' I I I I 



Commutato. 



Fig. 1 



known maximum speed over as long a path as possible in the gas; con- 

 sequently a second gauze (Go) is set up beyond G\, and maintained at 

 nearly the potential Fso that there is a nearly equipotential region be- 

 tween them. (Generally the potential of Gi is raised a fraction of a volt 

 above V so that there may be a slight impulsion of the ions formed be- 

 tween G\ and Gi toward the collector.) Beyond Gi are the collector C 

 and its protecting gauze 6*3, maintained at potentials lower than the 

 filament so that no electrons may reach them. The current of which 

 the sign indicates whether it is due to radiation or ionization, as was ex- 

 plained above, flows through the galvanometer at A. 



With such an apparatus as this it seems to be easy enough to measure 

 ionizing-potentials correctly within one or two volts. As soon as 

 greater accuracy is sought after, the real troubles begin. The electrons 

 do not all leave the source with negligible speed ; their speeds are dis- 

 tributed over a finite range. The potential to which they climb in pass- 

 ing through the meshes of a gauze is not quite equal to the potential 

 of the gauze-wires themselves. The potential-difTerences between the 

 different electrodes are not accurately given by voltmeters, for there 

 are contact-potential-differences superposed upon the values indicated. 

 The filament is not an equipotential surface if it is heated by a current, 

 although this difficulty can be overcome if the experimenter thinks it 

 worth the trouble. Electric charges marooned upon the walls of the 

 tube, electrons ejected by radiation from the gauze Gi and accelerated 

 backwards to Gi with a final speed higher than the electrons from F 



