130 Mr. E. L. Nichols on the Alternating 



with the simultaneous curve of current (A, fig. 2). The 

 corresponding curve F, which was taken while the discharge 

 was passing between the ball and point, is in its essential 

 features of the same character as curve D (fig. 2). The 

 interval of uniform positive potential is of the same length, and 

 it is coincident with the interval of excess of current which 

 shows itself in the positive branch of curve B. It is note- 

 worthy that this interval of uniform potential which marks 

 the duration of the arc, occupies in both cases the same portion 

 of the cycle (between scale-divisions 10 and 19, approxi- 

 mately), although there is otherwise a difference of phase, 

 due to self-induction, amounting to at least four scale-divisions. 

 Curve C, for instance, reaches its maximum in the neighbour- 

 hood of scale-division 8, curve E at scale-division 12. The 

 exclusion of the coils of the indicating galvanometer from the 

 circuit reduced this difference of phase to a small quantity, 

 and it suppressed altogether the remarkable oscillation of 

 electromotive force (see curve D) which in all preceding 

 experiments had introduced the formation of the arc. 



The results exhibited graphically in these six curves afford 

 abundant verification of the theory of the ball and point phe- 

 nomenon, given in a previous paragraph ; and they establish 

 the fact that in the secondary circuit of a transformer, such as 

 that made use of in these experiments, the striking-distance 

 from ball to point exceeds that from point to ball. It follows 

 that whenever the space between the ball and point is less 

 than the former and greater than the latter distance, discharge 

 will occur only during that portion of each alternation for 

 which the ball is positive, and that under such circumstances 

 a galvanometer placed in the circuit will show a constant 

 deflexion. 



Complete corroboration of the foregoing conclusion was 

 obtained by studying the image of the arc in a revolving 

 mirror. With an arc of less than the critical length the 

 discharge was seen to consist of two distinct sets of sparks, all 

 of the same duration but differing in colour. Each alternate 

 discharge was purple, the intermediate ones being of a green- 

 ish cast. The spark-images were everywhere equidistant, 

 and their duration was about four times as great as the inter- 

 vening intervals of darkness. The extension of the sparking- 

 distance beyond the critical point resulted in the complete 

 suppression of the series of purple images, the intermediate 

 ones remaining undisturbed in position, duration, and appear- 

 ance. The intervals of darkness were then estimated to occupy 

 six tenths of each cycle, the discharges four tenths ; a ratio 

 which corresponds with that of the duration of positive poten- 



