184 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1907. 
damped and undamped waves involves many factors, and is not fair 
unless we compare together transmitters taking the same mean power. 
Generally speaking, however, we may say that not only this glow- 
lamp detector, but all forms of thermal detector, give greatly 
increased effects when employing undamped oscillations. I find, 
for instance, that if undamped oscillations are created in a closed 
wire circuit which forms part of a circuit containing capacity and 
inductance shunted across a Poulsen arc, I can induce powerful 
secondary oscillations in a similar closed and syntonic secondary 
circuit at a considerable distance, and detect these by the use of my 
oscillation valve and a galvanometer placed. In fact, the use of 
undamped oscillations in a closed primary circuit, and this oscillation 
valve used with a telephone in a closed secondary circuit, brings to 
the front again the possibility of making use of so-called wireless 
telegraphy by electro-magnetic induction over very large distances. 
The old form of electro-magnetic induction telegraphy as practiced 
by Trowbridge, Preece, Lodge, and others made use of low-frequency 
alternating currents (50 to 100) in a closed primary circuit, and 
employed a telephone in a distant closed secondary circuit to detect 
the magnetic field so produced, signals being made by interrupting 
the primary current. I have, however, found a means of greatly 
improving this form of wireless telegraphy. In a closed primary 
circuit I establish continuous undamped oscillations of, say, a quarter 
of a million frequency by the are method. At a distance I place a 
syntonic secondary circuit containing my oscillation valve as a 
detector, a telephone being used with it connected between the 
middle plate and negative filament terminal. Both the primary 
circuit and secondary circuit are connected to earth at some point. 
The signals are made by breaking and making the earth connection 
of the transmitter in accordance with Morse code. When the 
earth connection is made at both ends a sound is heard in the tele- 
-phone, but not when it is broken. This seems to depend upon the 
fact that the oscillations produced by the are method are not abso- 
lutely continuous, but cut up into groups, as already proved by the 
experiment with the rapidly moving neon tube and helix. 
I have found that it is not necessary to employ a high-voltage 
carbon filament, a small lamp with 4-volt filament, taking about one 
ampere, works quite as well as a wireless telegraph receiver as a 12 or 
100 volt lamp. The filament has, however, to be at a certain critical 
temperature to obtain the best result; the vacuum also has to be 
extremely good. There are, no doubt, many possible variations of 
the above-mentioned type of oscillation valve wave detector. Every 
glass vessel containing rarefied gases or mercury vapor having elec- 
trodes of different sizes or shapes or temperatures, has some degree 
of unilateral conductivity, and can be used in the above manner 
