510 M. B. W. Feddersen on the Electric Spark. 



by employiDg smaller resistances, to measure the time of a total 

 electrical discharge. A new difficulty, however, appeared here : 

 the residue in a Leyden jar, after its discharge through a column 

 of water, is, ccefeiis paribus, by uo means constant ; for the relation 

 of the discharged electricity to the original charge may be three 

 times as- great in one extreme of two cases to what it is in the 

 other : I shall return to this again. As, however, a mean value 

 between these two .extreme ones is by far of most frequent oc- 

 currence, I noted only those lengths of the luminous band which 

 I observed to lie iu the majority , of cases upon the scale. At the 

 same time care was taken that the entire discharge occurred at 

 a time when the circuit-completing spring of the rotary appa- 

 ratus was slipping upon the metal. The jars employed were 

 always connected y?ith the conductor and the one pole of the 

 spark-micrometer,: while the water column was brought into 

 connexion with the other pole. 



In this manner, although I have made only a few experiments 

 on different days atid. with different velocities of rotation, yet 

 these may serve to «h'ow generally, what is a priori evident 

 and otherwise sufficiently supported by indirect experiments, that 

 the duration of a total erec.trical discharge increases both with 

 the resistance and with the electrical surface. Further, it seems 

 probable that while the diiration increases with the distance of 

 discharge, it does so much uiore slowly; for with a fivefold di- 

 stance the duration is not even doubled*. In judging of the 

 observations, it must further be r'epiembered that the phsenomena 

 of discharge almost always ei^ded as suddenly as they began, and 

 that a gradual decrease without a sharply defined limit, probably 

 only occurs with the smallest resistances. A s/««// diminution of 

 the intensity of the light towards the end of the luminous band, 

 however, was always observed. 



* The fact which is shown subsequently, that tlie residue in the Leyilen 

 jar increases more slowly than the entire quantity of charge, is here of im- 

 ))ortance. For if the residue be considered as always being a constant 

 fraction of the charge, and if we ado))t the hypothesis of partial discharges 

 only gradually approaching one another as the distance of discharge in- 

 creases, we should be rather inclined to expect a diminution of the total 

 duration as the distance of discharge increased. 



