140 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1911. 
The elements of the apparatus thus far include a generator of sus- 
tained high-frequency oscillations, an interrupter to modify the am- 
plitude of these oscillations into groups of a period within the range of 
audition, some form of detector to rectify these oscillations, and a tele- 
phone receiver. Manifestly here are all of the elements that are neces- 
sary for telegraphy, using the telephone receiver to interpret the 
signals. 
Tf in the above mentioned chain of apparatus the interrupter is 
replaced by some form of telephone transmitter, such as the micro- 
phone, this is all that is necessary for the transmission of speech. 
Experiments were made over local circuits with apparatus arranged 
in this order over a range of frequencies from 20,000 to 100,000 per 
second, with the result that speech was transmitted very satisfac- 
torily. Upon removing the detector from the above arrangement all 
perceptible effect in the telephone receiver ceased; in fact no arrange- 
ment of connections of a telephone receiver to such a high frequency 
circuit which did not include some form of detector was found to be 
operative for telephony, unless certain low resistance telephones were 
used in which case the speech was so much weaker as to be of an 
entirely different order of magnitude. 
The presence of a detector in this chain of operations, is not abso- 
lutely necessary in the case of telegraphy, since if the interrupter 
automatically produces a definite number of wave-trains per second, 
each train consisting of at least several complete oscillations’, an effect 
may be produced upon a telephone receiver directly without a detector. 
The physiological effect, however, is quite different, the clear funda- 
mental note corresponding to the frequency of the interrupter being 
no longer audible, but, instead, a peculiar dull hissing sound. If, how- 
ever, a telephone receiver was used, which, instead of having a per- 
manent magnet as a core, had one of soft iron, no effect without the 
detector was produced with the energy used. 
As stated above in the case of telephony, the energy required for 
telegraphy without a detector is of a different order of magnitude. 
Having determined the necessary and sufficient conditions for the 
accomplishment of telegraphy and telephony by means of electric 
waves guided by wires upon local circuits, the next step was to apply 
these means and conditions to an actual commercial telephone cable 
line, the constants of which have been given above. 
The machine was run at a frequency of 100,000 cycles per second 
with the circuit arrangements as shown in figure 1, where one wire of 
the telephone cable was connected to one terminal of the secondary 
of an air-core transformer, the other terminal being connected to 
earth. 
At the receiving end of the line, which was the Signal Corps con- 
struction laboratory, at 1710 Pennsylvania Avenue, Washington, 
