Scientific American, July 19, 1902, pages 37-38:

    There are at least five different methods by which articulate speech may be transmitted electrically without connecting wires between two given points. The first and oldest of these is by conduction through land and water. In this system four conductors are earthed, two at the transmitting and two at the receiving end. In this way a portion of the current, passing through the transmitting circuit, is shunted by means of the earth between the instruments and acts upon the receiver, since this path offers the least resistance. Collins Wireless Telephone
    As early as 1825 James Bowen Lindsay operated a system of wireless signals by this method, but by substituting a telephone transmitter for a telegraphic key and a telephone receiver for the galvanometer speech may be as easily sent as a signal. This is usually the first method suggested to the inventor seeking to transmit articulate speech without wires, but a very few quantitative tests will show that the limitations appear almost before its commercial value begins.
    The second and most beautiful form of wireless telegraphy is due to the effects of mutual induction or the magnetic lines of force exerted by one coil of wire on another placed in the same field of force by mutual induction. This is the ideal system, since no earth connection either at the receiver or transmitter is necessary to effect transmission, but the action is due entirely to the electric whirls or vortices set up in the ether. In this case the effective distance to which speech may be sent is limited by the number of turns of wire on the coil; their distance apart and the mutual induction will then depend upon the current flowing in the primary. Like the former system, the limits are soon reached.
    The radiophone and speaking telephone are two forms employing a beam of light to transmit telephonic messages. A pencil of light is allowed to fall on a mirror fastened to the diaphragm of a telephone transmitter, and by means of lenses the light is focused on a selenium cell at a distance of two or three hundred feet. In series with the selenium cell is a telephone receiver and a battery. When the sound waves of the voice impinge on the diaphragm of the transmitter, its vibrations cause the light to be displaced and its intensity on the selenium cell varied. Now selenium possesses the property of transmitting an electric current with twice the conductivity value when in the light that it possesses in the dark, so that there is a wide divergence of conductivity assured when the constantly varying beam of light falls upon it, and thus articulate speech is reproduced.
    The fourth system is that employing Hertzian waves, but as the enormously high-frequency oscillations produced by the disruptive discharge of a high potential current is much too rapid to make itself manifest in a telephone receiver, the oscillation circuit which emits the waves must be damped down by the addition of capacity in the form of Leyden jars or condensers and its relation to inductance sustained by supplementing the capacity with coils of wire until the telephone receiver will respond to a vibration of electric oscillations. This system of wireless telephony offers the most interesting experimental field of investigation, but its functions are so complicated that a very limited distance has yet been obtained with it. Diagram of Collins Wireless Telephone
    In making some tests in 1899 I found a method by which the disadvantages of the very rapid oscillations set up by a disruptive discharge in free air, as the spark of a Ruhmkorff coil produces, and without resorting to the loading of the oscillating circuit with artificial capacities and inductances. This was accomplished by permitting the discharge to take place in the earth instead of the air. To render this process clearer, let us employ, not only as a mere analogue, but as a similar proposition, the fact that electric oscillations emit electric waves, just as an electrically charged vibrating atom sends forth waves which are likewise of electromagnetic origin formed by the polarization of the ether. Even alternating currents of comparatively low frequency of a few thousand per second will emit long electrical waves in space, as Guarini has shown in his experiments in wireless transmission between Antwerp and Brussels. The length of the waves depends on the periodicity of the oscillations, the oscillations on the inductance, capacity and resistance of the circuit, and these in turn on the constants of the ether.
    The constants of the ether are its elasticity and its density. The elasticity of the ether is not known absolutely, but is measured by its reciprocal or dielectric constant, which is the ether modified by its relations with gross matter, and is called its specific inductive capacity. Ether, when in close proximity with gross matter, apparently assumes a greater density than in vacuo or free air, however paradoxical it may seem; it is now well known that it is not the conductor or wire joining an electrical circuit which conducts the electricity, but the tube of ether including the wire. The atoms of which the earth is composed are likewise permeated with the ether to a much greater extent than the atoms of gases forming the air. To this condition Tesla has given the name of bound ether. Similarly as mediums of greater densities transmit sound waves to greater distances than mediums of lesser densities, so the bound ether of the earth will propagate electric waves of proper length to greater distances than those of the ether-bound air. As an illustration, in the case of sound waves, if a bell is struck in free air it can be heard at a distance of a mile, it could be heard at a distance of twelve miles if struck under water, for water has a density twelve times that of air; now, when a rapidly alternating current of high potential is discharged into the earth and there allowed to restore the equilibrium, electric waves are emitted and propagated through the earth; the length of the waves is determined by the frequency of alternation and the distance of propagation will depend upon the density of the medium.
    These waves are, of course, normally radiated in every direction, but it has been found possible to reflect them and so make them unidirectional within certain limits. Fig. 1 shows photographically the wireless telephone transmitter the author devised for field work. Fig. 2 is a diagrammatic drawing of the system which has been patented in the United States and Great Britain. In the patent specifications a telegraph key is substituted for a telephone transmitter, as the system is interchangeable and may be used either for wireless telephony or telegraphy with some minor changes and additions.
    Referring to Fig. 2, A is a transmitter and B the receiver. The primary coil is shown at 1 and is in series with the battery, 2, and the key, 3. One terminal of the secondary winding, 4, is connected with a special form of transmitter, 6, and this to a large capacity, 7. The opposite terminal of the induction coil is earthed at 8, and bridged across the terminals of the secondary is the condenser; 10 is a "variator," which will be again referred to. The receiver is quite simple and consists essentially of a transformer coil, 1, a telephone receiver, 2, and a battery, 3; the condenser, 4, of large and equal capacity to that employed in the transmitter, and 5 the earthed terminal.
    The action of the instruments is as follows: When the key, 3, closes the primary circuit the current is automatically varied by a special device, 10, which takes the place of the ordinary interrupter; this produces alternations in the secondary coil, 4, giving rise to high potentials at the terminals, 7 and 8. This potential difference is, however, modified by the transmitter, 6. The surging of the alternating currents through the circuit formed by 7 and 8, emits waves principally at 8, and these traveling with the speed of all other electromagnetic waves reach the earth plate, 4, and, finding an ether path of greater density surrounding the circuit, 4 and 5, it traverses that circuit in preference to passing onward through the earth, since the former offers the least resistance. This sets up alternating currents in the transformer coil, 1, and these are impressed on the telephone receiver, 2. The capacity areas, 4 and 7, should be large and of special construction to secure the best effects. The capacities, 4 and 7, are not elevated, and the larger the capacities the greater the distance over which articulate speech may he carried without wires.
    Both the transmitter and receiver are mounted on tripods providing the operators with testing apparatus almost as portable as a camera. The tests, from the incipiency of the idea of wireless telephony, have been made at Narbeth, Pa., where the conditions were all that could be desired. In 1899, speech was transmitted by this system a distance of 200 feet; in 1900 a mile was covered, when with the equipment shown in the engravings articulate speech was transmitted across the Delaware River at Philadelphia, and in 1902 with the instruments placed on hills separated by a railroad, valleys, wooded lands and numerous streams a distance of three miles was attained. The results have shown the possible commercial value of this system of wireless telephony, which is soon to be perfected for actual use.