182 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1907. 
That the action is not altogether due to the removal of polarization 
films is shown by the fact that the fine platinum wire in the Schloe- 
milch form of detector wears away or is dissolved in the nitric acid 
Fic. 13.—Oscillation valve (Fleming). 
when oscillations are passed for some 
time through the cell, and there is some 
evidence that gold and platinum can 
be made to dissolve even in dilute acids 
by the action of electric oscillations. 
In 1904 I was so fortunate as to dis- 
cover another and quite different prin- 
ciple on which a sensitive electric wave 
detector can be based. If a carbon 
filament glow lamp has a metal plate 
carried on a third terminal sealed into 
the bulb, it is well known that a cur- 
rent of negative electricity flows from 
the plate to the positive terminal of 
the lamp, when the filament is ren- 
dered incandescent by a continuous 
current. This is the so-called Edison 
effect. It is also now known that in- 
candescent bodies discharge negative 
corpuscles or electrons from their surface, and incandescent carbon, 
when in a vacuum, exhibits this power in a marked degree. 
electricity escapes freely 
from it, but not positive. 
In 1904 I was endeavoring 
to find some way of recti- 
fying electric oscillations, 
that is, of separating out 
the two sets of alternate 
currents and making them 
separately detectable by 
an ordinary galvanome- 
ter. It occurred to me to 
make use of a carbon fila- 
ment lamp, having a metal 
cylinder insulated in the 
bulb surrounding the fila- 
ment, the cylinder being 
connected to a platinum 
Negative 
a 
Aho Eo} ; = 
et 
Fic. 14.—-Oscillation valve or glow lamp detector, 
used as a receiver in electric wave telegraphy. 
wire sealed through the bulb. (See fig. 18.) This lamp was then used 
as follows: A circuit was connected between the terminal of the metal 
plate and the negative terminal of the filament, the latter being made 
