RADIOTELEGRAPH Y—FLEMING. 179 
ions give to the residual gas a unilateral conductivity, as shown by me 
in a Friday evening lecture given here nineteen years ago. Moreover, 
the ionized gas not only possesses unilateral conductivity, but its con- 
ductivity, like that of the crystals just mentioned, is a function of the 
voltage applied to it. Hence, if we apply an electromotive force 
between the hot filament and the cool metal plate, we find that nega- 
tive electricity can pass from the filament to the plate through the 
ionized gas, and that the relation between the current and voltage is 
not linear, but is represented by 2 characteristic curve bending upward 
which has changes of curvature in it (see fig. 17). The sharp bend 
upward at one place implies a large increase in the current corre- 
sponding to a certain voltage, which means that, corresponding to a 
certain potential gradient and therefore velocity of the electrons, con- 
siderable ionization of the residual gas is beginning to take place. 
The current, however, would not increase indefinitely with the volt- 
age, but would before long become constant or saturated. It will be 
seen, therefore, that 
at points on _ the 
curve where there is 
a bend or change of 
curvature the second 
differential coefficient 
of the curve may 
have a large value. 
Hence, if we consider 
the current and volt- Applied Voltage. 
age corresponding to Fic. 17.—Characteristic curve of rarefied gas ionized by 
this point, it will be hot negative electrode. 
seen that any small increase in the voltage increases the current 
more than an equal small decrease in voltage diminishes it. If 
then we superimpose on a steady voltage corresponding to a point 
of inflexion of the curve an alternating voltage, the average value 
of the current will be increased. This then points out two ways 
in which this oscillation valve or glow lamp can be used as a radio- 
telegraphic detector. First, we may make use of the unilateral 
conductivity of the ionized gas in the bulb and employ the glow 
lamp with cylinder around the incandescent filament as a rectifier 
of trains of oscillations to make them affect a galvanometer or 
telephone. This method was described by me in papers and speci- 
fications in 1904 and 1905. In that case the valve is arranged in con- 
nection with a receiving antenna, as shown in figure 18, and used with 
a galvanometer or telephone. Mr. Marconi subsequently added an 
induction coil and condenser, and employed in 1907 the arrangements 
shown in figure 19. In this case the trains of oscillations set up in 
the antenna could not by themselves affect a galvanometer or a tele- 
000 
Current in Micro-amperes. 
