I 



M . B. W. Fcddersen on the Electric Spark. 507 



sufficient proof that every partial discharge, under all circum- 

 stances, endeavours to proceed on the same track which the 

 immediately preceding oue had taken. 



Relation between intervals of time to those of space of tioo par- 

 tial discharges. — It appeared to me of interest to ascertain the 

 interval of time which elapses between two partial sparks for 

 different distances of discharge and different resistances. The 

 observations, wliich were made at different times, did not, however, 

 present the concordance which was required, so that the time 

 ultimately obtained cannot lay claim to be an absolute measure. 

 Such an absolute determination would, moreover, be of small 

 value as long as bodies are employed for poles having an irre- 

 gular form, such as those of the micrometer formed by the knob 

 with clamp-screw and matrix. If, however, the smallest of the 

 numerical values found be always taken, we appear to approximate 

 most to the truth, for then the greatest agreement prevails. 

 Inasmuch as the intervals increase towards the end of the series 

 of sparks, I naturally only estimated them at the commencement. 

 The following Table shows in what sense the time-interval observed 

 depends upon the distance of discharge. For the first half of 

 the Table I measm-ed the intervals of the sparks by estimating 

 how many sparks appeared to cover a scale of 2^ millims. or 

 5 millims. in length. The observations were here so made that 

 the discharge-spark in the circuit was brought about by ap- 

 proaching the knob of the spark-micrometer to the charged flask 

 after removing it from the electrifjdng machine. For the second 

 half, and in all other experiments, the distance was directly esti- 

 mated, and the flask left in connexion with the conductor during 

 the discharge. I had, however, previously satisfied myself that 

 neither this combination nor even the addition of a second flask 

 exerted an appreciable influence upon the distance of the partial 

 discharges, so that (of course within certain limits) the extent 

 of the electrical surface was without influence. Generally speak- 

 ing, the latter mode of observation appeared to give more dis- 

 cordant results than the first one ; and although the following 

 numbers show with sufficient clearness that the interval of time 

 diminishes as the distance of discharge increases*, still the ob- 

 servations with the resistance c can only serve to give an idea 

 of the irregularities which often occur, especially with freshly 

 filled resistance-tubes. 



* This is undoubtedly a remarkable fact, although jjreviously jircdicted 

 by Riess. I am a\\;ue that the eye is generally disjjosed to estimate tlic 

 intervals of short hues as greater than those of longer oues, although 

 they be in reality equal ; I believe, however, that I have guarded i^jaiust 

 error of this kind. 



