230 Time-Lag hi the Spai'k Discharge. 



They may serve also to show why it is so difficult to 

 measure accurately comparatively low alternating potentials 

 (up to 5000 volts) by means of a spark gap. A definite 

 sparking potential in such circumstances can only be 

 obtained if the gap is completely free from hardness ; but if 

 it is so free, wonderfully definite measurements can be made 

 even without any initial ionization of the gap. An accuracy 

 of \ per cent, at 3000 volts can certainly be obtained either 

 with a steady source of potential or with a magneto or 

 induction-coil; though the method is always inferior to that 

 described in a recent paper*. However, since a spark gap 

 is often used for such a purpose, two practical hints for 

 abolishing hardness may be given. First it is a good plan 

 to allow a strong spark to pass between the electrodes for a 

 few seconds immediately before observations ; second it is 

 a very bad plan to clean electrodes by rubbing with fine 

 emery cloth. Unless the surface is actually pitted, discolo- 

 ration due to previous discharges is positively an advantage; 

 to rub the surface with fine emery is to rub the " hard " film 

 in, rather than to rub it o$\ and so to prevent its removal by 

 other means. 



Summary 



It has been shown by Peek that the peak potential of an 

 alternating source which is required to send a spark across 

 a gap between needle-points is greater than the spark 

 potential for a steady source. This dependence of the 

 sparking potential on frequency is attributed to a time-lag 

 in the initiation of the discharge. 



Two theories of this time-lag have been proposed or 

 suggested. One, due to Peek, that the formation of the 

 corona must precede the spark ; the other, that the presence 

 of casual ions is necessary. Each theory appears to account 

 for part of the facts. 



Some additional facts are recorded, which show that there 

 are two kinds of time-lag, one regular and one irregular. The 

 latter is accounted for by the necessity for the presence of 

 casual ions, the former by a slight modification of Peek's theory. 



In normal gaps between spheres the regular time-lag is 

 inappreciable, and in many conditions the irregular as well. 

 But if the surface of the cathode is not quite clean, a time-lag 

 may occur and the sparking potential may be dependent on 

 the frequency. This time-lag becomes inappreciable if the 

 steady sparking potential is sufficiently great. It has not been 

 determined fully what constitutes "cleanliness" in this matter, 

 but some practical methods of obtaining it are suggested. 

 * C. C. Patersou and N. R. Campbell, Phil. Mag. March 1919. 



