Telegraphs operate by sending electrical currents along wires. In Samuel Morse's original lines, two wires were used -- a sending plus a return wire -- to create a complete electrical circuit. However, in 1837 Carl August von Steinheil of Munich, Germany found that, by connecting the end of the sending wire to plates buried in the ground, the return wire could be eliminated. At the time, a common but incorrect belief was that the return current was now traveling through the ground back to the sending point, in order to complete the circuit, which in turn led to speculation about telegraphy through the ground without using any wires. This turned out to be impossible using standard electrical currents, but eventually it would prove possible using radio waves, although radio was not yet known at the time.
 
History, Theory and Practice of the Electric Telegraph, George B. Prescott, 1860, pages 398-400:

STEINHEIL'S  TELEGRAPH.

    "Ampère required more than sixty wires, whereas thirty or so were sufficient for Soemmering. Wheatsone and Cooke reduced their number to five; Gauss, and, probably in imitation of him, Schilling, as likewise Morse, made use of but a single wire running to the distant station and back. One might imagine that this part of the arrangement could not be further simplified; such, however, is by no means the case. We have found that even the half of this length of wire may be dispensed with, and that with certain precautions its place is supplied by the ground itself. We know, in theory, that the conducting powers of the ground and of water are very small compared with that of the metals, especially copper. It seems, however, to have been previously overlooked, that we have it within our reach to make a perfectly good conductor out of water or any other of the so-called semi-conductors. All that is required is, that the surface that its section presents should be as much greater than that of the metal, as its conducting power is less. In that case, the resistance offered by the semi-conductors will equal that of the perfect conductor; and as we can make conductors of the ground of any size we please, simply by adapting to the ends of the wires plates presenting a sufficient surface of contact, it is evident that we can diminish the resistance offered by the ground or by water to any extent we like. We can, indeed, so reduce this resistance as to make it quite insensible when compared with that offered by the metallic circuit, so that not only is half the wire spared, but even the resistance that such a circuit would present is diminished by one half. This fact, the importance of which in the erection of galvanic telegraphs speaks for itself, furnishes an additional feature in which galvanism resembles electricity. The experiments of Winckler, at Leipsic, had already shown that, with frictional electricity, the ground may replace a portion of the discharging wire.
    "The inquiry into the laws of dispersion, according to which the ground, whose mass is unlimited, is acted upon by the passage of the galvanic current, appears to be a subject replete with interest. The galvanic excitation cannot be confined to the portions of earth situated between the two ends of the wire; on the contrary, it cannot but extend itself indefinitely, and it became, therefore, now only dependent on the law that caused the excitation of the ground, and the distance of the exciting terminations of the wire, whether it was necessary or not to have any metallic communication at all for carrying on telegraphic intercourse.
    "I can here only state, in a general way, that I have succeeded in deducing this law experimentally from the phenomena it presents; and that the result of the investigation is, that the excitation diminishes rapidly, as the distance between the terminal wires increases.
    "An apparatus can be constructed in which the inductor, having no metallic connection with the multiplier, by nothing more than the excitation transmitted through the ground, will produce galvanic currents in that multiplier sufficient to cause a visible deflection of the bar. This is a hitherto unobserved fact, and may be classed among the most extraordinary phenomena that science has revealed to us. It holds good, however, for small distances. It must be left to the future to decide whether we shall ever succeed in telegraphing at great distances without any metallic communication at all. My experiments prove that such a thing is possible up to distances of fifty feet. For distant stations we can only conceive it feasible by augmenting the power of the galvanic induction, or by appropriate multipliers constructed for the purpose, or, finally, by increasing the surface of contact presented by the ends of the multiplier. At all events, the phenomenon merits our best attention, and its influence will not perhaps be altogether overlooked in the theoretic views we may form with regard to galvanism itself."