Conducting Gases on tke Direction of the Electric Field. 215 



thrown on the question by examining the potential-gradient 

 'between the plates when a steady P.D. is applied. 



In the present ease, where the plate on which the radio- 

 active substance was spread was horizontal, it was not con- 

 venient to use a water- or mercury-dropper to determine the 

 potential at any point between the plates. Recourse was had 

 to the method used by Zeleny, in which an insulated wire 

 connected to the electrometer takes up the potential of the 

 gas in which it is placed. The wire W, shown in fig. 1, was 

 bent into a circle in a horizontal plane, and connected to a 

 vertical rod inside an insulated glass tube, so that the potential 

 to which the wire was raised was not in any wslj influenced 

 by the vertical support. The wire and -all the metal surfaces 

 of the vessel were of tin, and in this way contact-differences 

 of potential were to a large extent avoided. The wire takes 

 up the potential rapidlyat a point where positive and negative 

 ions are plentiful, but more slowly at points near the upper 

 plate where the ions are nearly all of one sign, and not nearly 

 so numerous. 



Unless special precautions are taken, there is a danger that 

 near the upper plate the average potential may not be 

 indicated. Suppose, for example, iu the present «case, that the 

 lower plate is positively charged and the upper plate earthed. 

 The ions near the upper plate are nearly all of positive sign, 

 and the wire will continue to receive a positive charge and 

 rise in potential until it has reached the potential correspond- 

 ing to the gas in the neighbourhood o£ the wire w T here the 

 potential is Jdgliest. Since there are no negative ions present, 

 any accidental excess of charge on the wire cannot be got rid 

 of, and the indicated potential may be too high. A mercury- 

 or water-dropper has the advantage that it takes up the 

 potential at a point rapidly whether the ions of one or both 

 signs are present. The wire, however, gives accurate results 

 if care is taken, and is often more convenient to use tlian 

 the mercury- or water-dropper*. 



In determining the potential-gradient the method shown in 

 fig. 2 was employed. 



The insulated battery of small accumulators was connected 

 to the terminals of a Thomson- Varley slide Rof 100,000-ohms 

 resistance. By means of the movable earth connexion E 

 the potential-difference at the terminals of the resistance was 

 divided in such a ratio that the potential of the air was zero 

 close to the potential wire. When this is the case, the electro- 

 meter shows no deflexion on separating the quadrants. This 



* This point has been discussed in a recent paper by C. D. Child, 

 rh\s. Rev. March 1901. 



