Diagram of the experiment

Today in 1875, during an experiment conducted by Alexander Graham Bell and his assistant Thomas Watson, a receiver reed failed to respond to the intermittent current supplied by an electric battery. Bell told Watson, who was at the other end of the line, to pluck the reed, thinking it had stuck to the pole of the magnet. Mr. Watson complied, and to his astonishment Bell heard a reed at his end of the line vibrate and emit the same timbre of a plucked reed, although there was no interrupted on-off-on-off current from a transmitter to make it vibrate. A few more experiments soon showed that his receiver reed had been set in vibration by the magneto-electric currents induced in the line by the motion of the distant receiver reed in the neighborhood of its magnet. The battery current was not causing the vibration but was needed only to supply the magnetic field in which the reeds vibrated. Moreover, when Bell heard the rich overtones of the plucked reed, it occurred to him that since the circuit was never broken, all the complex vibrations of speech might be converted into alternating currents, which in turn would reproduce the complex timbre, amplitude, and frequencies of speech at a distance.

Edwin S. Grosvenor and Morgan Wesson write in Alexander Graham Bell: The Life and Times of the Man Who Invented the Telephone:

The same day, the two men drew up designs for telephone transmitter that would use a membrane to respond to the vibrations caused by words spoken into it… Bell wrote of his insight to his parents–“I think the transmission of the human voice is much more at hand than I had supposed”–and to Gardiner Hubbard–“I have accidentally made a discovery of the very greatest importance in regard to the Transmitting Instruments… I have succeed today in transmitting signals without any battery whatsoever!” Though today historians see this discovery as a major step on the road to telephones, Bell’s parents were not impressed, and Hubbard dismissed it altogether, gently guiding his inventor back toward multiple telegraphy.