110 . Dr. E. Weintraub on the Arc in 
mined by practical considerations. Thus, the anode must 
have a relatively high melting-point, must not enter into 
combination with mercury, must have in its normal con- 
dition as little gas occluded as possible, &e. The material 
of it does not seem, however, to affect the nature of the 
are in the least. In the case of a mercury anode, to be sure, 
vaporization of the mercury takes place, and this affects 
somewhat the stability of the are in accordance with the 
principle expounded above. Butthis is obviously unimportant 
in this connexion. 
Another property which is characteristic of the cathode is 
that it is difficult, if not altogether impossible, to realize such 
conditions that more than one cathode in multiple should be 
present in the same arc. Attempts to produce more than one 
cathode invariably fail. One of the experiments may be 
described here. 
Experiment 11.—In the tube ABC (fig. 6) there are three 
cups filled with mercury, of which C and B are connected to 
the negative, A to the positive pole of the source of current. 
B and A are close to each other, so that by shaking, the are 
BA can be started. There is enough mercury in the tube to 
cover the surface of the glass between C and B when the 
tube is a little inclined. When the mercury-surfaces in C 
and B are thus connected, and the wandering bright spot 
(from our point of view the centre of ionization) happens 
to move away from B towards C, the mercury-surfaces are 
separated at a point near to B, the bright spot remaining now 
on the surface C. In this way the cathode is transferred from 
3B to C, and B rendered inactive. The arc goes exclusively 
to C, although the arc CB represents an additional resistance. 
If Band C are made anodes, and A the cathode, and the 
same experiment as before performed, the arc sticks to B, 
and by placing resistance in the branch leading to B, so as 
to more or less compensate the resistance of the additional 
length of arc, the current can be distributed between the two 
anodes. 
This impossibility of having more than one cathode is 
intimately connected with the necessity, from our point of 
view, of creating a centre for production of ions on the 
cathode. 
If the cathode is the seat of the ionization process, a 
mercury arc, having the mercury as anode, and a piece of 
solid material, such as graphite or iron, as cathode, must 
produce some peculiar phenomenon on that solid cathode. 
If, in fig. 2, connexions are reversed, B made the anode, 
D and A cathodes, on starting the side-branch arc AB the 
