

Carves of the Oscillating Spark. 711 



break of the primary current are given off from the cathode. 

 This usually includes the period during which the oscillating 

 spark occurs together with a little time before and after it. 

 The curves are spirals (figs. 1 & 2, PI. XX.). The spirals 

 due to successive sparks do not repeat themselves exactly on 

 the screen, and it was difficult to get good photographs by 

 long exposures with a camera. The figures in the Plate 

 were photographed by simply placing the sensitive plate in 

 direct contact with the back of the fluorescent screen. There 

 is ample light by this method to photograph the spirals due 

 to single sparks, although naturally the definition suffers a 

 little. The spirals are always elliptic (or circular) with their 

 axes vertical and horizontal (the eccentricity may be altered 

 by adjusting the position of Lj). This shows that, as 

 theory demands, the potentials and currents are damped 

 harmonic curves which differ in phase by 90°. The decrement 

 of successive oscillations can be read off directly from the 

 curves : it varies considerably with the conditions (compare 

 the two figures), but it usually approaches more nearly to a 

 uniform than to a logarithmic decrement *. The bright 

 horizontal bands (sometimes mere dots) at each end of the 

 spiral are due to the p.d. of the jars gradually rising before 

 and after the oscillating discharge. By increasing the 

 capacity or the spark-length, the first band may be 

 lengthened at the expense of the second — an effect obviously 

 produced by the resulting delay of the spark relatively to 

 the cathode-ray emission. These bands often exhibit stria- 

 tions, or alternations of bright and dark portions, which show 

 that the charge of the jars does not rise uniformly but in a 

 series of steps — the period of these is no doubt that of the 

 discharge of the jars through the secondary of the induction- 

 coil. 



If Y is connected to the terminal H of the spark-gap we 

 get the spark-current plotted against the p.d. between the 

 terminals of the gap, and the type of figure which results is 

 shown in fig. 3, PL XX. The dot A represents the initial 

 sparking p.d. and the " cross " B the current-potential curve 

 during the oscillations. In order to see the detail here it is 

 necessary to use a tube which gives bigger deviations for the 

 electric field than those of the previous figures. This was 

 done by increasing the length of the parallel plates from *4 

 to 4 cm., and also putting them closer together. The curves 

 then appear as in fig. 4, PL XX. The initial sparking band 

 is now deflected off the screen (it can be brought on again 

 with a magnet), and a distinct displacement is to be seen 

 * First rioted bv J. Zenneclr, Ann. d. Phys. xiii. p. 822 (1904). 



3 A2 



