Extending the range of telephone calls was a key engineering challenge, much tougher than for telegraphy because the higher frequency of voice-based signals caused them to fade faster as they traveled along a wire. Early on, a device called a loading coil offered a partial cure. Independently invented in 1899 by Michael Pupin of Columbia University and George Campbell of AT&T, it basically consisted of a coil of wire that was placed along a line every 6,000 feet or so, greatly diminishing attenuation in the range of frequencies suitable for voice transmission. Two years later commercial service began between Philadelphia and Chicago, and by 1911 a long-distance line stretched all the way from New York to Denver. But transcontinental service remained out of reach until Bell engineers began experimenting with the triode vacuum tube, patented in 1907 by the radio pioneer Lee De Forest as "A Device for Amplifying Feeble Electrical Currents."

De Forest's tube used a small, varying voltage on a gridlike element to impose matching variations, even at high frequencies, on a much larger flow of electrons between a heated filament and a plate. The inventor's understanding of his device was imperfect, however. He thought that ionized gas in the tube was somehow involved. In 1913 a Bell physicist named H. D. Arnold showed that, on the contrary, the completeness of the vacuum dictated the performance. Arnold and his colleagues designed superior tubes and related circuitry to amplify long-distance telephone transmissions, and service was opened between New York and San Francisco in 1915. Alexander Graham Bell made the first call, speaking to Thomas Watson, who had helped him develop a working telephone four decades earlier. The transcontinental path had 130,000 telephone poles, 2,500 tons of copper wire, and three vacuum-tube devices to strengthen the signals. A 3-minute conversation that year cost $20.70.

By the mid-1920s long distance lines connected every part of the United States. Their capacity was expanded by a technique called frequency multiplexing, which involves electronically shifting the frequencies of speech (about 200 to 3,400 cycles per second) to other frequency bands so that several calls could be sent along a wire simultaneously. After World War II, the Bell system began to use coaxial cable for this kind of multiplexing. Its design—basically a tube of electrically conducting material surrounding an insulated central wire—enabled it to carry a wide range of frequencies.

Stretching coaxial cable beneath oceans posed difficulties so daunting that the first transatlantic link, capable of carrying 36 calls at a time, wasn't established until 1956. But radio had been filling the oceanic gaps for several decades by then while also connecting ships, planes, and cars to the main telephone system or to each other. After mid-century, a previously unexploited form of radio—the microwave frequencies above a billion cycles per second—took over much of the landbased long-distance traffic. Microwaves travel in a straight line rather than following the curvature of the earth like ordinary radio waves, which means that the beam has to be relayed along a chain of towers positioned 26 miles apart on average. But their high frequency permits small antenna size and high volume. Thousands of two-way voice circuits can be crammed into a single microwave channel.