182 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1909. 
by means of, the aerial vibrations created by the voice speaking to 
some form of microphone. We can not employ an intermittent spark 
generator because each spark would give rise to a sound in the tele- 
phone, and these sounds, if occurring at regular intervals, would 
produce a musical note in the telephone. If, however, we make the 
sparks run together into what is practically a high voltage are tak- 
ing a small current, then, in an oscillatory circuit shunted across this 
arc, we have set up persistent high frequency oscillations, as first 
achieved by Mr. Duddell. We can greatly increase the energy of 
the oscillations by immersing the arc in a strong transverse magnetic 
field and also in a hydrocarbon gas, as shown by Poulsen, or we may 
employ a number of arcs in series. KE. Ruhmer has lately also em- 
ployed a high-tension are between aluminum electrodes (see fig. 21, pl. 
2),shunted by a condenser and inductance as a means of generating 
persistent oscillations. As an alternative, it is possible to create them 
_by a mechanical method, viz, by a high frequency alternator, subject, 
however, to certain limitations as to frequency. Both these types of 
generator have their advantages and practical objections. There is 
good evidence that radiotelephony has been accomplished over dis- 
tances of 100 miles or more by each of these methods in the hands of 
experts, but what is now required is the reduction of the apparatus 
to such simple manageable and practical form that it can be applied 
in regular work. The wave-generating apparatus must be capable of 
producing uniform persistent oscillations of high voltage and fre- 
quency, not less than 30,000 or 40,000 per second, or at least above the 
limits of audition, and the amplitude of these oscillations must be 
capable of being varied by some form of speaking microphone placed 
in the oscillation circuit or in the radiating antenna, or in a secondary 
circuit coupled to it. No ordinary simple carbon microphone will 
safely pass sufficient current for this purpose. A type of multiple 
microphone has been used successfully and also a duplex microphone, 
the invention of Ernst Ruhmer. 
It is not, however, possible to speak of radiotelephony at the pres- 
ent time as having reached the same level of practical perfection as 
radiotelegraphy. But the possibilities of it are of such a nature that 
it will continue to attract the serious attention of inventors. This 
is not the place to enter into a full discussion of the causes which 
limit submarine telephony through cables, but there are well-known 
reasons in the nature of submarine cables as at present made which 
impose very definite limits upon it, owing to what is called distortion 
of the wave form. Electric wave telephony is free at least from 
this disadvantage, and if (as has been asserted) arc generators can 
be made self-regulating and capable of being worked for hours auto- 
matically, or even for ten minutes without being touched, then the 
remaining difficulties with the microphone are not insuperable. 
