"Music plus electricity equals the sound of the 20th century" - Joseph Schillinger, 1918.

They are near-mythical beasts, these pioneering, half-forgotten electric-electronic instruments, with their excellently strange names - the sfaerofon, the sphärophon, the choralcelo, the gnome, the pianorad, the trautonium, the croix sonore, the mellertion, the hellertion, the orgatron, the connsonata, the neo-bechstein - and stories clustering round them, many facts not exactly reliable.

The first and most fabulous monster is Thaddeus Cahill's telharmonium: 200 tons, 60 feet across, taking up a whole floor and the basement below. It looked, surviving pictures tell us, like a church organ mated with a weaving loom. Cahill, a Canadian, built it in Holyoke, Massachussetts; partially funded by the New England Electric Music Company - whoever they might have been - it cost a then-phenomenal $200,000, and was moved in 1906 to Telharmonic Hall in New York. The idea was to transmit 'Telharmony' across America, to hotels, restaurants, theatres and private homes, via local telephone exchanges. The telharmonium itself was a kind of keyboard-operated dynamo organ; the bulk of the machine consisted of vast teethed gears on engine-driven spinning shafts which caused alternating currents in batteries of magnets. There were no loudspeakers in those days - radio was only five years old, and Lee De Forest's audion tube, which amplified signals many thousand-fold, wouldn't exist for at least another decade - so it fed straight into the telephone system. Unfortunately, it needed huge voltages and caused interference over the rest of the telephone network, such as it then was - so that one day an enraged businessman burst in, broke it up and threw the machinery into the Hudson river, or so the story goes.

Actually there were no less than three telharmoniums, spread over some 20 years: the first Cahill had started in 1895 in Washington DC, patented in 1897, finished in 1900; the Holyoke-NYC model was the second; a third, begun in 1908, was finished in 1911 and certainly still in use in 1916. But by the mid-teens, radio broadcasts into the home were the coming thing, and the project went broke for lack of subscribers (though a similar device, the choralcelo, contemporary, smaller, more obscure, is reported to still have been in use in the 50s).

For a short while, however, the telharmonium was big news. A story in McClure's Magazine, 'New Music For An Old World', brought it to the attention of Ferruccio Busoni, a virtuoso classical pianist and critical intellectual, Italian by birth, German by temperament, respected across all Europe. Busoni (whose pupils included Edgard Varèse) cited the telharmonium in a polemic he was then writing (for some reason he calls it the "dynamophone"). His 1907 Sketch Of A New Aesthetic Of Music proposed that music pass beyond its 19th century framings - harmony as the possible combination of a mere 12 notes, a highly selective and conventional instrumentation - to embrace the "infinite" gradations within the octave structures: "The question is important and imperious, how and on what these tones are to be produced. Fortunately, while busied with this essay, I received from America direct and authentic intelligence which solves the problem in a simple manner. I refer to an invention by Dr Thaddeus Cahill. He has constructed a comprehensive apparatus which makes it possible to transform an electric current into a fixed and mathematically exact number of vibrations."

At which point Busoni hurtles intoxicatingly into an airborne rhetoric that flatters Cahill's 200 ton apparatus: "Who has not dreamt that he could float on air? And firmly believed his dream to be reality? Let us take thought, how music may be restored to its primitive, natural essence; let us free it from architectonic, acoustic and aesthetic dogmas; let it be pure invention and sentiment, in harmonies, in forms, in tone-colours (for invention and sentiment are not the prerogative of melody alone); let it follow the line or the rainbow and vie with the clouds in breaking sunbeams; let Music be naught else than Nature mirrored by and reflected from the human breast; for it is sounding air and floats above and beyond the air; within Man himself as universally and absolutely as in Creation entire..."

Air is to the point, or ether is. The technologies Busoni invokes are simultaneously magical and domesticated, that is the instant transmission of sound - of the voice, of consciousness, of will - over vast distances. Today, such technologies - radio, telecommunications - are so familiar it's hard to recapture a sense of them anew, but in the early years of this century they were the essence of sorcery.

In 1907, radio was still largely dedicated to the transmission of Morse Code signals for shipping. The telephone, some 30 years old, was for speech - though concerts had been transmitted (in stereo, no less) down phone lines from the Paris Opera to select homes in 1881 by one Clement Adler, and in Budapest in 1893 by the Telefonhirmonde company; a similar system, called the electrophone, was established in London in the 1880s.

But telephone earpieces make for lousy speaker systems, even today, and not enough punters seemed to want to pay for this early Dial-A-Tune service to fund the necessary improvements. (The telharmonium's huge and problematic voltages were connected to its apparently impressive volume, which was said to be as loud and clear as an orchestra.)

Around 1917, several Paris-based radio engineers - among them cellist and telegraphist Maurice Martenot, engineer Armand (or maybe Joseph) Givelet and a Russian emigre called Nikolay Obukhov - discovered that the mechanical howl of a poorly-tuned radio receiver could be affected by human movement close by. But no one found a satisfactory way to harness this phenomenon to a keyboard:a player's hands interfered with the note produced by the pressed key. Of course, it was wartime, and other matters were perhaps pressing. Nonetheless, within the year a young Moscow engineer, Lev Sergeyevich Termen, had solved the problem: he simply dispensed with the keyboard and waved his hands through the air. The aetherophone, he called it, and it invoked the spirit of radio, the very geist of the wireless age. It was a polished wooden box with a whisker antenna poking up into the air at one end, a ring of metal sprouting at right angles from the other. Sound was produced by the player moving his hands through the electromagnetic field set up by the two protruding bits of wire, a manual vibration of the ether itself. A telephone earpiece, perhaps with a cardboard horn attached to it, emitted an electronic buzz or whine. The sound was far from unpleasant, sobbing and swooping its way through recognisable tunes.

Lenin had the new instrument demonstrated to him in 1921, and loved it. By 1923, Termen was a roving ambassador for Soviet ingenuity and vision, visiting France, Germany, Britain and the US. In fact, he was so successful, and so many imitations of his device appeared, with similarly clinical sci-fi geek names, that he decided to change not only his invention's name but his own. He became Professor Leon Theremin, and it became an advert for him: the thereminvox, soon shortened to the theremin.

Leon had discussed possible improvements to the instrument with Maurice Martenot in Paris in 1923, and Martenot went on to build his own ondes musicale: it debuted around 1928, at first using some kind of tugged string mechanism. A Danish bandleader called Jens Warny built his own aetherophone, dubbed it the sfaerofon, and began touring. In Germany (or possibly in London) Jörg Mager - working independently on 'radio-howl' instruments -had invented his elektrophon in 1921, issuing a pamphlet, A New Epoch In Music Through Radio.

Perhaps influenced by Warny, the elektrophon subsequently became the sphärophon. Early models were worked with a dial (the kurbelspärophon): Mager was interested in microtonal music and initially felt keyboards reinforced diatonic tonality. Subsequently, he developed a keyboard system which allowed for quartertones, the klaviatursphärophon. Finally, he devised a polyphonic version, the partiturophon (this, in other words, was a version that played more than one note at a time, which organs and the telharmonium could, and the rest of the above could not). In pictures, we see Mager is a portly, balding man, and that his kurbelsphärophon and its knobs, is rather more enigmatic: resting partly on a table beside him and partly on a stand in front of him, it looks nothing like the elegant theremin. Mager appears to be operating a small lever on the stand, and reaching beside him to turn a dial. His research gave him a degree of success in the established music world: he was contracted in 1931 by Bayreuth to create the bell-tones in Parsifal (apparently he amplified some Javanese gongs). "At the short-lived peak of his career," wrote Hugh Davies in The Grove Dictionary Of Musical Instruments, "he was given the use of a small castle in Darmstadt, to which he moved in 1929, and where he founded the Studiengesellschaft Für Elektro-akustische Musik." (Mager died in 1939, and none of his machines survived the Allied bombing of Germany. Still, is it mere coincidence that after the war Darmstadt became a centre for experimental and electronic music?)

By 1929, Nikolay Obukhov had built his croix sonore, a theremin in all but name, according to most sources, though visual descriptions are intriguing: "A cross about four feet high surmounting a globe about two feet in diameter." More routinely thereminoid were the electronde, the elektronische Zaubergeige and the ethonium. Theremin's experiments with something like a cello fingerboard - to replace air-gestures - had been taken up by his compatriots back in Moscow, resulting variously in the violena, the sonar, the emiriton and the ekvodin. Back in Paris, one Rene Bertrand, a close friend of Varèse, applied Mager's dial idea to his dynaphone and exhibited it in Barcelona, Prague and Budapest. Martenot picked up a tip from Theremin and renamed the ondes in honour of himself: the ondes martenot was adapted to use a fingerboard, or, as it's often described, misleadingly but attractively, a 'ribbon': it was designed to facilitate microtonal scales (in particular the finely divided scales of Indian raga). Less successful, despite similar principles and aims - in particular this same abiding but largely unsatisfied obsession with non-standard pitch-divisions - were Langer and Halmagyi's emicon; the mellertion, which divided the octave into ten rather than eight steps; and, not to be confused with it, Bruno Helberger and Peter Lertes's hellertion, which had four touchtone fingerboards made, gloriously, of leather.

In the late 1920s, Dr Friedrich Trautwein, a professor of acoustics at Köln, exhibited his trautonium (its monophonic fingerboard, originally based on the hellertion's, evolved into a stretched wire, which was pushed down to contact the surface beneath it). Photos of a 1930 performance of one of composer Paul Hindemith's seven trios for trautonium could as well be pictures from inside some newly automated 20s office: the machines which three men are sitting at resemble typewriters more than anything else (larger models favoured the telephone-switchboard look).

By now, with radios, microphones and electrical phonographic recording commonplace, loudspeaker technology was pretty much evolved. The use of photoelectric cells to generate sound - basic to the working of the film soundtrack - was also being translated into instrumentation: for instance, emigre Ivan Eremeyev's 1932 gnome worked on an electromagnetic/photoelectric tone-wheel principle.

From the late 20s onwards, the ways in which electricity could be applied to the routine amplification or improvement of standard instrumentation proliferated. Jörg Mager's assistants Oskar Vierling and Harald Bode worked on electrically amplified violins and pianos (the neo-bechstein) and electronic organs, as did countless others, on such makes as the Rangertone, the Orgatron (which was bought up by Wurlitzer, the jukebox manufacturer), the Allen, the Baldwin, the Connsonata, the Leslie and the legendary Hammond. The science fiction writer and radio-parts magazine editor Hugo Gernsback built several electric pianos in the mid-20s, among them the staccatophone and the pianorad. (In producing these instruments, Gernsback may have been elaborating on a 1916 patent of Lee De Forest's for musical uses of the audion tube: a photo of the pianorad shows it to resemble a drinks cabinet, full of big light bulbs.) In 1929, Edouard Coupleux, an organ designer, joined forces with Armand Givelet to build the Coupleux-Givelet organ. This was an early 'synthesizer', in that it could be programmed like a player piano with punched paper rolls. Somewhere in the American Deep South in the late 20s, the brother of the World War One air ace Eddie Richenbacher had built a funny little metal steel guitar and fitted pick-ups to it - Country music would never be the same.

Last July, at an exhibition and conference at the American Centre in Paris, called Glo-bal Tekno One, the Futurist Luigi Russolo was reclaimed as the rightful ancestor and visionary precursor of all things rave. This is not entirely unfair, since Futurism was very much about the celebration of machine-age sensation. But despite the fact that Marconi had put Italy on the map in terms of radio technology, Russolo's instruments - the roarers, squeakers and honkers of his intonarumori, also lost during the war - seem to have been home-made contraptions of wood and canvas, though some may have had small electric motors. (An electric or electronic instrument is one where the sound is a direct product of the translation of electromagnetic waves into soundwaves. A church organ that pumps its bellows by electricity is not, in any essential sense, an electric organ; nor is one whose key release mechanism is electrified. In both cases, the means of producing sound is still the vibration of air in a pipe.)

So why is Russolo venerated today as one of the founding fathers of Techno, while the likes of Trautwein and Gernsback, for instance, languish in obscurity? It might have something to do with this: the Italian became emblematic for a technologised present that turns its face towards the future partly because he rejected the musical conventions and routines of his day. By contrast, the inventors/performers of much of the exotic musical flora and fauna so far mentioned here had a rather timid willingness to adapt to established rules, which is probably what prevented them taking root.

Initially, some kind of coming together between these instruments and a new type of music that probed their sonic possibilities seemed possible. Certainly works were written specially for many of them. In the USSR, the theremin had made its concert debut in 1924, performing Andrei Pashchenko's Symphonic Mystery for theremin and full orchestra, while Vladimir Solokov is said for a while to have composed for nothing else.

Theremin's countryman Joseph Schillinger wrote his First Airphonic Suite for the theremin in 1929. Arthur Honegger wrote a ballet - Roses De Metal - for three of Bertrand's dynaphones and piano. Varèse incorporated two theremins in his 1934 ecuatorial (however the published score replaces them with two ondes martenots, which were used in its first performance in 1961). In 1935, Percy Grainger - another Busoni disciple - wrote Free Music #1 for four theremins; the following year he composed Free Music #2 for six; and he went on to build some kind of punch-card programmable free music organ. Aaron Copland used a theremin in his opera The Second Hurricane (1936). Charles Ives added an optional theremin to his Fourth Symphony, and helped finance the rhythmicon, a machine for generating rhythms which classical percussionists couldn't cope with. It was designed by Theremin and composer Henry Cowell, and used by Cowell in his Rhythmicana (but this was not actually performed until 1971. Intriguingly, a new CD issued by Mute, John Came's Rhythmicon, uses the machine's composing principles in a contemporary setting).

The trautonium won itself a degree of support within Germany thanks to Richard Strauss and Hindemith: besides his seven trios, Hindemith wrote a concerto for it. Strauss lost interest and Hindemith was forced by leave Germany by the coming of Nazism, but though the Nazis had active interest only in large-scale electric instruments, suitable for large public occasions, the trautonium somehow survived.

Ravel gave permission for the first movement of his string quartet to be rescored for martenot. Honegger's Les Milles Et Une Nuits (1937) used four. Olivier Messiaen's F&e234;tes Des Belles Eaux used six, while his massive Turangalila Symphony is an ondes soloist's showcase. (A version of the theremin, activated by on-stage dancing, the terpistone, was built in 1932 and used in several ballets.)

RCA manufactured theremins from 1929 ("Not a radio! Not a phonograph! Not like anything you have ever heard or seen!"), but even though Theremin was by then regularly touring and lecturing in the States, they only sold about 300. 200 trautonia were built, but only a handful were sold - perhaps less than 20 - and most of the rest recalled and dismantled. Ondes martenots were manufactured individually, to order only. Most of the others only ever existed as prototypes, long since traded for scrap, or smashed by falling bombs.

However, electric organs - using mechanisms that are direct descendants, miniaturised, of the telharmonium - were becoming commonplace by the 40s, in cinemas, in churches, in the home. It was no longer simply the case that electrification was a bar to popularity. The problem was that few of the above instruments had any such integrated role. Of course new works could be written for them, but just like microtonal composition (also briefly fashionable during these years) these were intruders on the standard classical repertoire, and very much hostage to their own untried quality. And as many as may have been written, how many were ever performed? A registered score does not a public performance signify - or even an extended rehearsal. Of the above compositions, only ecuatorial and Turangalila could be said to have won a place in the international postwar repertoire.

Despite the microtones, the 30s were a backward-looking time for music. The likes of Varèse, Grainger and Cowell were far more reviled than respected, whether as Futurists, as musical Bolsheviks, or as destroyers of magnificent tradition for the sake of trendy machine-age progression. And performance on the new instruments, as Cowell's pupil John Cage complained angrily in 1937, was almost apologetic about its very novelty: "When Theremin provided an instrument with genuinely new possibilities, thereministes did their utmost to make the instrument sound like some old instrument, giving it sickeningly sweet vibrato, and performing on it, with difficulty, masterpieces of the past." (Theremin's own favoured repertoire was not a help here: for that 1921 demonstration of the instrument to Lenin, he'd played Mikhail Glinka's "The Lark", a sentimental oldie and mouldy even then.)

If anything, the 30s proved that the classical orchestra is simply an impossibly conservative institution: it would have to give up too much of its sense of timeless, uniformed, reverenced self to allow for any such changes. Though ragtime and jazz had conclusively demonstrated how limited its expressive range was, they had themselves filled the very gaps they proved existed: and these were musics with a built-in relationship to an audience. A new instrument needed to do more than arrive, cap in hand, and beg to sit in: it had to create a meaningful dramatic space around it, and force its acceptance on its own terms, by means - as it were - of its paying fan club. The unlimited new arenas of the miked studio and the recorded surface allowed the drum kit, the saxophone and the electric guitar just such a space: which is why they became the foremost 'new' instruments of the century, numerically and creatively (and also why Cage and disciples more or less stopped writing for established instruments).

The ondes martenot lived on - Hugh Davies believes that more than a thousand works have been composed for it, as well as ballets, theatre and film music - but only in France. Honegger and Messiaen, as well as Milhaud, Koechlin and Jolivet, all turned their patriotic pens to composing for this magnificent local device, but no one else did.

Trautonium virtuoso and proselytiser Oskar Sala, who later modified Trautwein's instrument into his own mixtur-trautonium, also bartered a tricky postwar position - survival under Nazism impressed few in those years - into a useful pop cultural role: by 1960 he was working with Hitchcock on the soundtrack of The Birds. Beyond this, his mixtur-trautonium remains strictly a German thing. (By way of illustration, Sala and his instrument have recently been 'adopted' by the Frankfurt-based Ambient composer Pete Namlook. Namlook's Fax label has just issued a CD, My Favourite Instrument, which features Sala performing his own compositions for trautonium - intriguing, but he's no Varèse - and Namlook has combined a monophonic trautonium with a new computer program which he says generates "subharmonic chords and tones".)

Pre-empting the use of a trautonium in The Birds, from the late 30s theremins were more in demand in Hollywood than they'd ever been in vanguard classicism: the instrument's weirdly unplaceable expressionism supplied alien soundtrack and jittery sensation to psychodramas and sci-fi shockers alike, including King Kong, The Lost Weekend, Hitchcock's Spellbound, The Day The Earth Stood Still and Forbidden Planet. From there it graduated to pop/counter-culture status, supplying strange atmosphere to songs by The Beach Boys and Led Zeppelin (in pop, "Good Vibrations" and "Whole Lotta Love" remain the iconic theremin moments), Captain Beefheart, Todd Rundgren, The Bonzo Dog Doo Dah Band, the sublimely named Lothar And The Hand People (Lothar being the name of the group's theremin) and, a generation later, Portishead, Jon Spencer Blues Explosion, Pere Ubu, whose David Thomas plays a small, hand-held theremin, Bruce Wooley (!), Paul Weller and so on and so on.

Why, out of this long list of potential non-keyboard devices, is it the theremin that has landed safe in posterity? The point must be this: unlike the theremin in Hollywood (or, in a parallel world, the electric organ, which was widely used at a very unelevated level, or the electric guitar, which in the 40s and 50s expoded into pop), most of the non-keyboard devices mentioned here failed to make an impression in the realm of the non-virtuoso amateur. Those that did were rejected by most 'serious' composers for that very reason: that's to say, post-war avant gardists wanted bigger, better, computer-controlled composing machinery, not some silly electric whining-stick that makes everyone think of giant bugs from Mars.

The fact is that any instrument with no institutional grounding of second - and third - raters, no spectral reserve army of amateurs, will wither and vanish: how can it not? Specialist virtuosos may arrive to tackle the one-off novelty - the theremin's Clara Rockwell, the ondes martenot's Jeanne Loriod, the trautonium's Oskar Sala- but there's no meaningful level of entry at the ground floor, and, what's worse, no fallback possibility of hack careerism if things don't turn out (again, with the exception of the theremin). Why would any child - let alone a large enough group of them to ensure momentum - choose to begin learning to play such devices? Only from-birth mavericks would even have considered it; and by the 40s, many kinds of music offered such people far more enticing futures elsewhere. Orchestral classicism was, if not for dullards, then certainly for the unadventurous. And the unadventurous don't have ambitions to be professional croix sonoristes.

It's here, of course, that the revolutionary machine-dreams of modernism faltered. Classical expertise and technical mastery cannot be cut adrift from layers and layers of ordinary, unassuming music usage; and not just because both can be pitilessly boring to outsiders. The mundane and the ultramontane are part of a continuum, not opposites.

Today the Termen/Theremin story seems more romance than radicalism; it would make a fit subject for opera, even for Hollywood: an inventor of musical devices becomes an ambassador for Leninist future-art and is then named as a traitor-defector under Stalinism. He becomes a successful novelty-name on the American culture circuit but is driven - at the height of segregation - into semi-retirement from polite society when he marries a black dancer, Lavinia Williams. He is kidnapped and forced in secret by the KGB to develop surveillance equipment, on pain of death. During the 50 years he's missing, presumed dead, his most devoted disciple and fellow-exile Clara Rockwell keeps the faith, until the day they are reunited.

However much of this is fact, however much is subtly exaggerated or rounded into melodrama, it's undoubtedly the tale of an honest man whipped ragged by the gale of life, and the forces of politics and history. And then, at the close of his life, he is proved right after all, validated not by the avant garde but by pop culture. Which is a good story.

And yet his acclaim feels a bit too pat, a bit too safe. Heard today, Clara Rockwell's theremin performances make no concessions to the heroic modernism that informed the instrument's origins, nor to Russolo, nor Varèse, nor even Jimmy Page. Her repertoire is absolutely non-radical - Glazunov, Rachmaninoff, Ravel, Tchaikovsky - and her playing style is a throwback to the Romanov era: these pieces drip with portamento and brittle art-deco warbling. Her hands are liverspotted, arthritic claws, but still she flicks and jabs and curls her fingers in the genteel, tiny, reserved, almost aristocratic gestures she invented more than half a century ago. There's a distinct emotional pull to her sound, and how she gets it, that's nothing to do with revolution, but everything to do with utopia (which also applies to the instrument's take-up by some of the pop/rock groups listed above).

Nostalgia for an age yet to come? For past visions of the future perfect? Consider: cyberpunk, the dominant pop-sci-fi mode, is prophetic not of gleamingly antiseptic spaceway-ranging utopias, but rather of the shabbily ordinary present, everywhere extended: the modern city street, stretching from exclusive malls and palatial consumer dromes outwards to endless junkyards, landfills, desolation. The awareness that more and more resources are tied up in discarded and obsolete machinery is transmuted into a kind of glamour. The thrilling future landscapes of modernity functioned better as an inspiration than they ever did as material fact: not least because, put into practice, they seemed suddenly thick with reactionary links and fearful borrowings from the world they were supposed to be replacing. Think again of the grainy look - in all those badly reproduced and hard-to-find photos - of the sphärophon, the pianorad, the trautonium. Think again of the loss of these instruments (and the bad judgement and confused aesthetics that led to that loss). One has to wonder if what makes these failed machines so attractive - as well as the once-was, never-was future of music they ought to have belonged in - is actually that they failed. Tonight, live, from the dustbin of history...