ELECTRIC WAVE TELEGRAPHY—FLEMING. Nis 
charged but charged up again in the opposite direction.” It parts 
with the excess of electrons forming the negative charge on its plate 
in connection with the negative arc terminal, and that plate is left 
with a deficiency of electrons, that is with a positive charge. Then 
the process repeats itself over again. Two conditions seem neces- 
sary for the automatic continuance of this process. First, the are 
must be formed between terminals of such nature and in such sur- 
roundings that rapid variations of current through 1t must cause 
correspondingly rapid and large 
changes in the potential difference 
(P. D.) of the terminals in an in- 
verse sense, that is, as H. T. Si- 
mon has shown, there must be a 
steep falling characteristic curve 
for the arc. (See fig. 10.)’ Sec- 
ondly, the are must have the 
power of restarting itself if en- 
tirely extinguished for a short 
time, but this should not take 
place until the P. D. between the 
terminals exceeds a certain value, 
that is, it must not take place too 
easily or at too low a voltage. If 
the are is formed between solid 
carbon terminals then it appears 
that these conditions are only ful- Amperes. 
filled up to a certain frequency, Fig. 10.—Characteristic curves for con- 
that is when employing a rather tinuous-current are in air and hydrogen 
6 : (Upson). Are length = 1.25 mm. 
large capacity in the condenser 
circuit. We then obtain Mr. Duddell’s musical or singing are, 
which emits a sound because the rapid variation of current through 
@The amplitude of the potential difference of the condenser terminals may 
and does become very much greater than the mere steady potential difference 
of the electrodes between which the arc is formed. Thus, with a P. D. of 220 
or 300 volts across the are the R. M. S. of the condenser plates may reach 
1,000 or 1,500 volts. 
6A careful study of the phenomena of the electric arc between metal and 
metal and carbon terminals in air and hydrogen has recently been made in my 
laboratory, under my direction, by Mr. W. L. Upson. It has been found that 
for an arc between a cold metal and a carbon terminal in hydrogen for the 
same length of arc, the rate of decrease of terminal voltage with increase of 
current is always greater than for an are in air between two carbon terminals. 
In other words the volt-ampere characteristic is steeper. Also it has been 
found that in the case of a carbon are in air the current can be interrupted for 
a much longer time without permanently extinguishing the arc than is the case 
for the metal-carbon arc in air or hydrogen. 
