730 EVENING DISCOURSES. 
in and out from the centre, which produces a movement which much more nearly 
corresponds with the actual current in the vertical wire as used in spark telegraphy. 
It is necessary here to explain the relationship that exists between the wave- 
length, the frequency, and the velocity of propagation of Hertzian waves. ‘The 
waves travel with, as far as we know, the same velocity as light—namely, 
300,000,000 metres, or 186,000 miles, per second. Between these quantities we 
have the relationship that the product of the wave-length by the frequency is 
equal to the velocity of propagation, or, as I have already mentioned, the velocity 
of light. 
The wave-lengths which are of practical use in wireless telegraphy at the 
present time range between 100 and 38,000 metres, though, of course, it is quite 
possible to use for special purposes wave-lengths outside these limits. The 
corresponding frequencies in practical use are therefore between 3,000,000 and 
100,000 complete periods per second. We require, therefore, to produce in the 
vertical conductor alternating or oscillating currents of any frequency within this 
range, and to have a sufficient number of oscillations following one another without 
interruption to allow of good syntony being obtained. 
There are three methods of producing these currents—namely, the alternator, 
the spark, and the arc methods. 
There are great difficulties in the way of constructing an alternator to give 
such high-frequency currents, and I can best illustrate this by taking an example. 
Suppose that it is required to build an alternator to work at the lowest 
frequency, namely, 100,000 periods per second, and let us assume that we can 
drive this alternator by means of a turbine at the high speed of 30,000 revolutions 
per minute. This alternator could not have a diameter much above 6 inches for 
fear of bursting; and, as it makes 500 revolutions per second, it would have to 
generate 200 complete periods for each revolution, so that the space available for 
the windings and poles for one complete period will be less than ;4, inch, a space 
into which it is quite impossible to crush the necessary iron and copper to obtain 
any considerable amount of power. In spite of the small space that we have 
allotted to each period, as there are 100,000 periods per second, the speed of the 
surface of the moving part works out at over 500 miles per hour. A small 
alternator has been built to give over 100,000 frequency, but the amount of power 
it produced was extremely small. Several experimenters have stated lately that 
they have built alternators giving these high frequencies and a considerable 
amount of power, but, so far as I am aware, there is no reliable data available as 
to the design of these machines. 
If it should prove possible to construct alternators for these very high fre- 
quencies, we shall be able to obtain a sufficient number of consecutive oscillations 
of the current in the érial of definite frequency to enable very sharp syntony to 
be obtained. Ne* only will this greatly reduce interference troubles in wireless 
telegraphy, but sh alternators will be of the greatest value for wireless tele- 
phony. 
The earliest method of producing high-frequency oscillations was proposed by 
Lord Kelvin, who pointed out that if a Leyden jar or condenser be allowed to 
discharge through a circuit possessing self-induction or electrical inertia, then 
under certain conditions the discharge of the jar is oscillatory, that is to say, that 
the electricity flows backwards and forwards in the circuit several times before 
the jar or condenser becomes finally discharged. I think that perhaps the best 
‘way to make this matter clear is by demonstrating experimentally with an oscillo- 
graph the nature of the discharge of a condenser, and how it is affected by the 
resistance and self-induction in the circuit. As a mechanical analogy one may 
look upon the charged condenser as a weight attached to a spring which has been 
pulled away from its position or rest. To discharge the condenser we let go the 
weight and it begins to oscillate backwards and forwards, and, after making a 
greater or less number of oscillations, finally comes to rest. The number of oscil- 
lations per second will depend upon the strength of the spring and the mass of the 
weight, which correspond with the capacity and self-induction in our electrical 
circuit, The number of oscillations before the weight finally comes to rest is 
