Metallic Vapours in an Exhausted Space. bial 
main are BD does not start, even after the ionized vapour 
has filled the whole space BD. This has already been 
mentioned. Jt would, however, be too hasty a conclusion to 
say that graphite or iron is incapable of being the cathode in 
a mercury are, since we have seen that the cathode of an are 
must be rendered active, which was not the case with the 
graphite cathode D. The simplest way to make a cathode 
active is that used in all ares, i. e. separating the two 
electrodes after having brought them into contact. This is 
realized in the following experiment. 
In a vertical carefully-exbausted glass tube there is a 
mass of mercury at the bottom and a rod of graphite 
suspended from a platinum wire, and reaching within a 
short distance from the mass of mercury at the bottom of 
the tube. The graphite rod is connected to the negative, the 
mass of mercury to the positive terminal of the source. By 
shaking the tube it is easy to bring the mercury and the 
graphite rod into contact and separate them again, whereupon 
the are starts up with a graphite cathode. In this case the 
following phenomena are observed :—There is a wide hot spot, 
wandering about the surface of the rod, just as there was a 
bright spot on the surface of the mercury cathode. Rapid 
disintegration of the cathode takes place from the very 
beginning, and a deposit of carbon forms all over the walls 
of the tube and the surface of mercury. This disintegration 
goes on as long as the arc lasts, and takes place whatever the 
the material of the cathode (graphite, iron, &c.). 
The cathode is, therefore, the electrode which disintegrates 
in the are. This mechanical disintegration is probably in- 
timately connected with the ionization which takes place on 
the surface, and the most probable assumption is that the 
current leaves the graphite cathode in form of carbon ions, 
and is transported through the arc by means of mercury ions. 
Whether these latter are generated exclusively at the surface 
of the graphite cathode or whether they come in part from 
the mercury anode cannot be decided. If a piece of graphite 
and a mass of mercury are both connected to the negative 
terminal, the are will go almost exclusively to the mercury 
cathode. A small current of about ‘01 ampere flows to the 
graphite, but even this small current is accompanied by a 
slight disintegration. 
To summarize what has been said above, there ts strong 
evidence in favour of the assumption that the cathode is the 
electrode at which the primary generation of ions takes place, 
while the question whether the flow of ions in the are is 
unidirectional—negative ions going from the cathode to the 
