A SHORT HISTORY OF SHORTWAVE & CW
Long before voices crossed the air, the radio spoke in dots and dashes.
The code
In the 1840s Samuel Morse and Alfred Vail gave the telegraph a language: dots and dashes that could be tapped down a wire. Vail did much of the practical work on the alphabet itself, and the pair demonstrated the finished system publicly on 24 May 1844, sending What hath God wrought from the Capitol in Washington to a railway depot in Baltimore. Within two decades wire ran across continents and under oceans, and the dot-dash alphabet was the only thing moving on it.
Two alphabets, one survivor
The code Morse and Vail built — American Morse, or railroad Morse — was not the code we use now. It had dashes of three different lengths and, worse, characters containing an internal pause: C was keyed as two dots, a gap, then a dot. On a clean land line with a skilled operator that worked well and was slightly faster than the alternative. On a long cable, on a noisy circuit, or later on a radio signal fading in and out, an ambiguous silence is indistinguishable from a letter break, and the errors piled up.
In 1848 Friedrich Clemens Gerke reworked the alphabet for the Hamburg–Cuxhaven line, reducing it to two element lengths and one gap length with no internal pauses. The German-Austrian Telegraph Union adopted his version in 1851, and the International Telegraph Congress in Paris standardised a near-identical code in 1865. That is International Morse, still current and still maintained as an ITU recommendation. It won because it is unambiguous: every character is a solid run of elements, and the only silences that mean anything are the ones between characters and between words. American Morse hung on in North American railroad and press service into the middle of the twentieth century, then went quiet.
Spark, and then continuous wave
The first wireless transmitters were spark-gap sets. A spark does not produce a tone — it produces a damped burst of energy smeared across an enormous swathe of spectrum, heard in the receiver as a harsh buzz. It was loud, it was simple, and it was hopeless at sharing a band. Two spark stations within range of each other were two stations jamming each other.
Continuous wave changed the arithmetic completely. Instead of a burst, the transmitter holds a single steady frequency and the key simply turns it on and off. The Poulsen arc and the Alexanderson alternator got there first, and the vacuum-tube oscillator made it cheap. Because all the power now sits in a narrow slice of spectrum, the receiver can throw away everything outside that slice, and a beat oscillator turns the bare carrier into a clean audio tone the ear can follow through noise that would bury anything else. A CW signal occupies a couple of hundred hertz where a voice signal needs some two and a half thousand. That is the entire reason CW gets through when nothing else does. Spark was prohibited for most services by international agreement in 1927 and had effectively disappeared during the 1930s.
Marconi, and across the Atlantic
Guglielmo Marconi filed his first wireless patent in 1896 and spent the following years pushing range for its own sake. On 12 December 1901 at Signal Hill in St John's, Newfoundland, he reported hearing the letter S — three dots — sent from Poldhu in Cornwall. The claim was contested at the time and is still argued over, since there was no recording, no independent witness, and the daytime path should not have supported the wavelength he was using. What is not in doubt is what followed: documented transatlantic transmissions from Glace Bay in Nova Scotia in 1902, and a commercial transatlantic message service opened in 1907 between Clifden in Ireland and Glace Bay. Long-distance wireless was a business within a decade of being a curiosity.
Why shortwave
Nobody could explain Marconi's results. Signals should travel in straight lines and the Earth curves away. In 1902 Oliver Heaviside and Arthur Kennelly independently proposed a conducting layer high in the atmosphere that bent the signal back down; Edward Appleton demonstrated it experimentally in the mid-1920s and later took a Nobel Prize for it. The ionosphere is real, it is layered, and it rearranges itself between day and night.
The practical discovery came from amateurs. Regulation had pushed them onto the short wavelengths below 200 metres on the assumption that nothing useful happened there. In late 1923 a station in Nice and two in Connecticut worked each other across the Atlantic on roughly 110 metres, with power that the great longwave stations would have considered a rounding error. Commercial and government services moved within a few years: the enormous longwave installations were superseded by compact shortwave beam stations, and the high-frequency bands — roughly 3 to 30 MHz, with the 160-metre band sitting just beneath them — became the backbone of long-distance communication for the rest of the century. Signals skip off the ionosphere and over the curve of the Earth, and a few watts into a wire antenna can cross an ocean. Navies, embassies, broadcasters and spies all built on that one fact.
Morse at sea
The distress signal SOS was agreed at the Berlin radiotelegraph convention of 1906 and came into force in 1908. It stands for nothing. It was chosen because, run together as a single character, its rhythm is unmistakable even through heavy interference.
On the night of 14–15 April 1912 the Titanic's operators, Jack Phillips and Harold Bride, sent both the older CQD and the newer SOS. Ships close enough to help had no one on the headphones. The regulatory response was immediate and permanent: operator licensing and a continuous radio watch, written into United States law that same year and into the first international Safety of Life at Sea convention in 1914. Out of it came the discipline that governed maritime Morse for eighty years — 500 kHz as the calling and distress frequency, and twice an hour, at a quarter past and a quarter to, three minutes of enforced silence across the band while every station in range listened for someone in trouble. The famous messages page collects the traffic that survived from that night and others.
The wartime fist
In the Second World War, signal-corps operators hunched over field sets in tents and trenches, copying Morse by ear under fire. Every operator had a recognisable fist — a personal rhythm in the keying, the particular weight given to a dash, the habitual hesitation before a difficult letter. Skilled listeners could tell who was sending from the cadence alone, and the Allied intercept service turned that into a method: operators at the listening stations logged the keying characteristics of individual enemy operators, which let analysts follow a unit across a change of callsign or a move to a new frequency. The content stayed encrypted; the handwriting gave the sender away. CW got through when nothing else would — weak, jammed, and buried in static.
The long decline
Single-sideband voice took over the point-to-point traffic from the 1950s, radio teleprinters took the rest, and satellites removed the reason to bounce anything off the ionosphere at all. Maritime Morse held on longest and then went in one step: the Global Maritime Distress and Safety System was phased in through the 1990s and came fully into force on 1 February 1999, ending the requirement for a Morse watch at sea. Later that year the last commercial Morse station in the United States closed its own watch, signing off with Morse's 1844 message and the prosign SK — the one that means the conversation is over, not merely paused.
Still on the air
Amateur licensing kept a Morse test for most of a century, then dropped it: the international requirement went at the 2003 World Radiocommunication Conference, and the United States removed its last five-word-per-minute element in February 2007. The obvious prediction was that CW would follow the requirement into history. It did not. More people are learning it now than were learning it when it was compulsory, because the ones doing it chose to.
They have reasons. A CW signal is narrow, so a receiver can filter down to a sliver of bandwidth and pull a readable note out of noise that would swallow a voice. The equipment is simple enough to build on a kitchen table. Low-power and portable operating live on it almost entirely. And there is a craft to it that a microphone cannot offer — a skill that takes months to acquire and never quite stops improving. There is more on that in Why CW Refuses to Die.
Shortwave is a small tribute to that craft: a bakelite-and-brass field set you can key from a browser. If you want to actually learn the code rather than read about it, start at the primer.