58 Proceedings of the Royal Irish Academy. 



over the layer ; and when this happens, the resistance soon becomes infinite. 

 An inspection of the layer shows, however, that it is broken where the sparks 

 occur, and the increase of resistance is not due directly to an electrical action. 

 "When the E.M.F. is much less than that which produces sparking, the passage 

 of the current does not produce an increased rate of fall of conductivity. For 

 an E.M.F. near to that which produces visible sparking there is a more rapid 

 decrease of conductivity when the current is flowing, but the source of the 

 action is somewhat uncertain. 



7. The explanation of the effects described above is by no means clear, 

 and there are many further experiments that should be tried before a full 

 theory is attempted. Some conceptions as to the nature of the effects may, 

 however, be formed from the results we have described. 



The relation between current and E.M.F. for a powder layer which has 

 not been subjected to the transverse field, 



G = aV+bV\ 



is of the type which applies to cases where the ionisation is confined to a 



surface, and ions of one sign are drawn away from the surface by the electric 



field. This type of relation holds when electrons are produced at one plate 



by ultraviolet light, and drawn across to a parallel plate by an electric field, 



and an identical relation has been verified for the current between terminals 



in a hot gas. In the case of the thin layers of powders we may perhaps 



imagine that the grains of powder are surrounded by conducting films which 



do not extend from grain to grain, and that the conductivity is due to the 



electrons drawn across the intervening narrow spaces by the electric field. 



"When the layer of powder has been subjected to the transverse field, a 



very great increase of conductivity takes place, and the current-E.M.F. curve 



is of a new form, viz., 



= Jc(V-v). 



This new form could be explained if we imagine the conducting films to have 

 spread so as to bridge over the gaps between the grains of powder. 



We have stated early in the paper that the effect studied is not due to 

 simple coherer action. For that opinion there is much evidence, including 

 the results with lead peroxide and the large voltages we use as compared 

 with those necessary in the case of the coherer. At the same time it is 

 necessary to remember that there is no satisfactory theory of the action of 

 the coherer, and it may well be that the coherer action and the effect we are 

 studying are different manifestations of the same type of phenomenon. 



We have pleasure in thanking Mr. Coghlan, ll.SC, for carrying out many 

 of the observations. 



