McClelland and Dowling — Electrical Conductivity of Poivders. 53 



Layers of lead peroxide were specially tested, as this substance is usually 

 quoted as one that acts as an anti-coherer. "With this substance we 

 obtained an increase of conductivity after the application of the transverse 

 field, provided the layer was very thin, but a decrease of conductivity with 

 thicker layers. This suggests that the effect we are studying is distinct from 

 the ordinary coherer effect, and has its seat not in the body of the powder, 

 but at the surface of separation between the powder and the insulator. 



Another experiment leading to the same conclusion was as follows. A 

 layer was formed as usual on the paraffin surface, and a metal plate mounted 

 parallel to it in air. When the electric field was applied between this plate 

 and the layer, a very slight increase of conductivity took place ; but when the 

 field was applied between the powder and the plate embedded in the paraffin, 

 the usual great increase of conductivity occurred. 



4. The rate at which the increased conductivity disappears was also 

 investigated. In the case of graphite layers the decrease of conductivity is 

 slow, and months would be necessary before the conductivity would return to 

 approximately its original value. In the case of layers of bronze metal 

 powder the rate of decrease is more rapid. Some such layers regained almost 

 their original resistance in half an hour, but two or three hours were more 

 usual. The rate of decay was carefully studied for bronze layers ; but it was 

 somewhat irregular and did not follow any simple law. Afterwards, when 

 some of the experiments were being repeated with the layers in a dry 

 atmosphere, we found that the rate of decay of the conductivity did 

 approximately obey a very simple law. The layer formed in the usual way 

 was placed in a closed vessel with a drying substance (P, 5 ), and after 

 standing for some hours the transverse electric field was momentarily applied, 

 and measurements of the conductivity made at intervals afterwards. In the 

 absence of moisture the increased conductivity remained for a longer period. 

 When ordinary undried air was admitted into the vessel containing the layer, 

 the conductivity began to fall more rapidly. The curves (fig. 3), p. 54, show the 

 results of plotting the logarithms of the conductivity against the time, the 

 curves A and B referring to two different layers. The vessel was opened 

 and the layer exposed to the undried air of the room at the time indicated by 

 the arrows. In the case of layer B we obtain a straight line in the dry 

 atmosphere showing that the conductivity decreases exponentially with the 

 time. For the layer A the rate of fall is more rapid at first, but the curve is 

 approximating to a straight line before the admission of the moisture. The 

 more rapid decay of conductivity after the admission of the moist air, and 

 especially the suddenness with which it begins, are very striking. 



5. A number of experiments were carried out bearing on the intensity 



[8*] 



