228 Dr. Norman Campbell on 



of the gap or of the sparking distance. It was again shown 

 that electrodes of Pt, Zn, Cu, Ph, Fe, Ni, Co, Or, Ag, Al, 

 Hg, C, Sn, Wj were all liable to hardness, but also that they 

 might all be soft ; no difference in respect of liability to 

 hardness could be discovered. But one new proposition of 

 importance w r as discovered. 



(5) Hardness is a property of the cathode only. No 

 treatment of the anode w T ill ever change a gap from soft to 

 hard or vice versa : if the cathode of a hard gap be sub- 

 stituted for the cathode of a soft gap, that gap becomes hard; 

 but the anode from a hard gap will not make a soft gap hard 

 if still employed as an anode. A gap may thus be hard in 

 one direction but soft in the other. 



The circumstances which make a cathode hard are ex- 

 ceedingly various and difficult to define in detail ; but their 

 general nature was obvious. Tims a cathode could always 

 be made hard by rubbing it with fine oily emery-paper ; it 

 could usually be made soft again by heating to redness in a 

 Bunsen name : but such heating always makes a platinum 

 cathode hard, while it could generally be softened by boiling 

 in nitric acid. A freshly turned surface was usually soft. 

 Hardness is therefore due to some adherent film on the 

 surface which usually consists of oily matter ; but it seemed 

 that the state of the atmosphere exerted some influence, for 

 on some days it was impossible to make any gap really soft. 

 If hardness is due to a surface film, the reason why it affects 

 only the catbode is apparent; for the liberation of electrons 

 from the cathode is an essential part of the mechanism of 

 the discharge, while the anode usually plays the part of a 

 mere terminal. 



Time-lag in hard gaps. 

 Hardness of this nature, then, is essentially different from 

 that which distinguishes point and sphere gaps. By the 

 measurement of the actual sparking potential in different 

 conditions, by the method mentioned on p. 222. an attempt 

 was made to determine whether such hardness is due to the 

 existence of a definite time-lag. Measurements of Q and C 

 w r ere made when the rate of rise of potential was varied. The 

 apparatus was adjusted so that the potential rose at a rate nearly 

 uniform until a value slightly greater than the actual sparking 

 potential was reached, and then remained nearly constant. 

 If the rate of rise is known, the time-lag can be deter- 

 mined as the difference betw T een the time when the potential 

 reaches the steady sparking value and that at which the 

 actual sparking potential, deduced from Q, is reached. 

 No great accuracy in determining the exact rate of rise was 



