98 Dr. E. Weintraub on the Arc in 
of the arc BD, the are does not start between C and A, 
nor is heating of the mercury by a Bunsen burner, so as 
to fill the space with mercury vapour, any help. In presence 
of the arc BD a luminous vapour spreads out from the path 
of the are in all directions, and fills the space between C 
and A. A direct-current ammeter in the circuit AC shows, 
under these conditions, a slight current, which varies according 
to the circumstances from ‘01 to ‘1 ampere. Under these 
conditions, the arc between C and A will eventually start, 
the current increasing, quite suddenly, up to 4-10 amperes, 
according to the resistance in series with the are. Under 
apparently the same conditions the starting of the are usually, 
however, does not take place at all. If all shaking of the 
tube is carefully avoided, this condition can be kept up 
indefinitely. The impression one receives from this experi- 
ment is that to allow the passage of a large current in the 
form of an arc, something must happen on the surface of the 
electrodes. This conclusion is made almost a certainty by 
the following experiment. 
Experiment 2.—Let us suppose that in fig. 1 the cup A is 
connected to the negative pole of the direct-current source, 
and while the are is running between B and D, as pre- 
viously, let us bring in contact for one moment the mercury 
surfaces A and B, B being one of the electrodes of the 
arc BD. Immediately on bringing these two surfaces into 
contact the are starts up between the two electrodes C and A. 
No starting takes place if the electrode A is electrically 
connected to the positive pole of the direct-current source g’. 
From this fundamental experiment one derives the con- 
clusion that in order that an are should start between two 
mercury electrodes, placed inside a highly evacuated space, 
and connected to a source of moderate voltage (magnitude of 
a few hundred volts), the cathode must first be rendered active, 
and we will interpret this in the light of the ionic theory 
by assuming that an tonization process must be started at the 
surface of the cathode in order to allow the passage of an are 
through metallic vapours. The anode behaves differently and 
receives the arc, the ionization process being once started at 
the cathode, without any difficulty. It is, however, irre- 
levant whether the active electrode B with which the others 
are brought into contact is anode or cathode of the are BD. 
The fact that the cathode of the new arc becomes active 
by contact with the anode of the running arc is due to a 
secondary effect. On connecting the two mercury surfaces 
of opposite signs, the current of one machine goes through 
the other, in the direction CDgBAg’C (CD-arc, D being 
