If, now, the Ruhmkorff be excited, the following phenomena result: If the point of contact between the conducting wire and the rectangle be moved along the latter, it will be found that, for most places, a spark passes between the balls of the rectangle, which varies in intensity, and at one place entirely disappears. This place, if we suppose the opening in the rectangle to be in the middle of one end and both balls to be of the same size, is in the middle of the other end. If, now, while no spark is passing in the rectangle, an insulated conductor be brought into connection with either ball, the sparks again appear. These, again, may be caused to disappear by moving the point of contact toward the manipulated terminal. The same effect would also be produced if, instead of changing the point of contact, an equal insulated conductor were touched to the other ball.
The length, resistance, and quality of the conducting wire have no influence upon the sparks; neither does the resistance or material of the rectangle affect it noticeably: e. g., one half of the rectangle being made of thick copper wire and the other of very fine German-silver wire did not alter the phenomena. Another conductor being brought in contact with the joint between the conducting wire and the rectangle has no influence.
The size of the rectangle has a great influence upon the size and length of the spark between its terminals; the larger giving, within certain limits, always the longer spark.
The air distance of the Ruhmkorff discharger is of great importance; under five and more than fifteen millimetres proved to be infelicitous.
Hertz's explanation of these phenomena is the following: At the moment when a discharge takes place between the terminals of a Ruhmkorff coil, in the whole circuit, and in all conductors in contact with it, powerful wave disturbances are agitated, which follow each other in such infinitesimal portions of time that the time which is required to travel with enormous velocity even a short wire is appreciable. These waves, arriving through the conducting wire at the rectangle, divide and traverse simultaneously both branches. If both sides are electrically symmetrical, the two wave-branches arrive at the balls of the rectangle in exactly the same phase, but oppositely directed, and interfere; there can then be, of course, no spark. If, however, they are not symmetrical, as when the contact is not in the middle, they do not interfere totally, but a spark passes. As the contact moves around the rectangle, the spark at its terminals will be less or more powerful as the interference is more or less total.
The electrical symmetry depends not alone upon the length of the wire, but upon its self-induction coefficient and its capacity.