120 . Dr. E. Weintraub on the Are in 
has, however, been ascertained that it is influenced by the 
magnitude of the alternating voltage and by the frequency 
of the alternating current. Jt zncreases in a very marked way 
with the length of tube. The most probable explanation of 
the phenomenon would consist in the assumption that the 
electromotive force must attain a certain value before it can 
bridge over the space. Another assumption would be that 
the “lag”? measures the time necessary for the ions to cross 
the distance between the two electrodes of the alternating 
are. Probably each one of the two assumptions has a part 
of truth in it. The experiments are being continued, and 
will, I hope, throw some light on this, in my opinion, very 
interesting phenomenon. The necessity for having a certain 
value of the electromotive force before the current can 
assume an appreciable value places this phenomenon in 
analogy with the well-known fact of the existence of a lower 
limit of electromotive force necessary to break the resistance 
of a gap between two electrodes placed in a gaseous medium. 
In the case before us the medium separating the two elec- 
trodes is an already, to a certain extent, ionized vapour, and 
this limit of electromotive force, which is low, can be readily 
measured, and the phenomenon therefore studied very 
thoroughly. 
Blondel has already observed a similar phenomenon in 
carbon-metal ares in air, but in this case the cause was in 
the self-induction placed in series with the arc, while here, 
in all the experiments, a practically non-inductive resistance 
was used to limit the alternating current. The influence of 
a self-induction in series will also be studied. 
Considered as a practical rectifier, the arrangement de- 
scribed in Experiment 14 has two great disadvantages. The 
first is the use of a direct current. ‘The second is the sup- - 
pression of one half of each wave, so that only pulsating 
currents are derived, instead of a continuous one. ‘The first 
improvement made is illustrated by fig. 10 (PI. IIL.). 
Experiment 16.—In the tube ABC (fig. 10) the two mer- 
cury cups A and Bare connected to a source of direct current, 
the cup B and the graphite terminal C to a source of alternating 
current. In shunt to the lamp a reactive coil is placed. The 
half-wave of the direction CB goes through the arc; that of 
the direction BC, being unable to pass through the arc, chooses 
the path of the reactance, which, in discharging itself after- 
ward in the proper direction CB through the arc, causes at 
least a part of the half-wave of the wrong direction to become 
rectified. A somewhat similar arrangement, in which a 
transformer is used, is shown in fig. 11. 
