THE ANGELS’
INSTRUMENT

A case file on the strangest accusation ever brought against a musical instrument — and on the four forces that made a continent believe it.

This is an investigation into how that happened — and into a figure that has been repeated so often it now passes for history: the claim that Franklin’s instrument killed some fifty musicians.

Two things can be said at the outset. The first is that the figure does not survive contact with the record. The second is that the story underneath it is far stranger than the body count ever was.

It began, as these things often do, with the boredom of a fashionable salon. In mid-eighteenth-century London there was a craze: performers drew wet fingers around the rims of wine glasses filled to different levels, and out of the crystal came a thin, unearthly ringing. Franklin, in the audience at one such concert, was captivated — and then, being an engineer, immediately began wondering how to do it better.

By 1761 he had his answer. Thirty-seven glass bowls of graduated size were threaded like beads onto an iron spindle, the largest producing the low notes, the smallest the high. A foot treadle spun the assembly like a spinning wheel; the player simply laid moistened fingers against the glass as it slid past. The bowls were colour-coded so the notes could be told apart. Up to ten notes could sound at once — full chords, impossible on a rack of tumblers. The glassblower Charles James built it in London to Franklin’s own instructions.

He called it the armonica, from the Italian for harmony. Franklin described the instrument in his own hand in a long letter of 13 July 1762 to the Italian physicist Giambatista Beccaria — a document that survives and can still be read today. He never patented it and never took a penny from it.

“Of all my inventions,” Franklin would later say, “the glass armonica has given me the greatest personal satisfaction.” From a man who had already tamed lightning, that is a remarkable sentence.

Europe fell in love at once. The armonica sang in the salons of Paris and Vienna. Marie Antoinette learned to play. Mozart, Beethoven, later Donizetti, Saint-Saëns and Richard Strauss all wrote for glass. The young Mozart first met the instrument in 1773, in the Vienna house of a physician whose name will matter a great deal before this investigation is finished.

But the human centre of the story is a woman named Marianne Kirchgessner. Born in Bruchsal in June 1769, she lost her sight to smallpox at four, taught herself keyboard in the dark, and was eventually apprenticed to a conductor who — conveniently — ran a glass harmonica factory in Karlsruhe. She became the reigning virtuosa of the instrument and toured Europe by carriage for years, blind, in an age when that journey was punishing for anyone.

In 1791 Mozart heard her play in Vienna. He was months from death. He wrote her two pieces: the Adagio for solo armonica, K. 356/617a, and the Adagio and Rondo, K. 617, dated 23 May 1791 — the last chamber music he ever composed. She premiered it with Mozart himself in the ensemble at the Kärntnerthortheater that August.

Hold that image, because everything that follows depends on it: a blind woman drawing from spinning crystal a sound that listeners compared to angels, while the greatest composer in Europe sat in the audience with half a year left to live. For an age just beginning to develop a taste for the supernatural, it was almost too perfect a scene.

It started as whispering. Players complained of odd afflictions: trembling hands, nerves stretched to breaking, fainting, cramps, dizziness, a black melancholy arriving without cause. Then the rumour widened. It was not only the players. It was the listeners.

In 1799 a physician, Anthony Willich, publicly condemned the instrument, holding that it induced a great degree of nervous weakness. From there the thing rolled downhill and outgrew medicine entirely. The armonica woke the dead. It drove dogs to madness. It ruined marriages and convulsed household animals. Listeners, it was said, went home and killed themselves.

Then came the episode that turned rumour into policy. At a concert in Germany, a child died. The connection to the singing glass was coincidence and fear, nothing more — but it was enough. Several German towns banned the instrument outright.

And on 9 December 1808, Marianne Kirchgessner died. She was thirty-nine. Too many people nodded: there it is. The instrument had taken its own high priestess.

Here is the first place the record contradicts the legend. Kirchgessner died of a chest infection. The claim that her nerves had been destroyed by the instrument’s piercing vibrations was rumour attached to the fact afterwards — and it sits awkwardly against a documented complaint from a London review of her playing: that her part could barely be heard over the other instruments. The armonica is, physically, a quiet instrument. The sound accused of shattering a woman’s nervous system struggled to carry across a concert hall.

This brings us to the figure that travels from article to article and mouth to mouth: that the glass armonica sent roughly fifty musicians to their graves.

The uncomfortable finding of this investigation is that there is nothing to support it.

Search as far as the sources go — museum archives, historical accounts, the modern literature — and only individual cases are ever named. Kirchgessner in 1808. The child at that German concert. Beyond those two, the record dissolves into fog: many performers were said to have fallen ill; some believed; it was rumoured. There is no list of names, no tally, no bodies. “Around fifty” is not data. It is a legend that acquired a number somewhere along its passage through the centuries, the way a ghost story acquires a date to sound more solid.

So if you came for a serial-killer instrument, strictly speaking there isn’t one. What there is instead is a better question: why did an entire continent believe it?

The most material of the theories is lead.

Early armonicas were made of lead crystal. The paint that colour-coded the bowls also contained lead. Now picture the working life: a musician drawing wet fingers across that glass and that paint, hours a day, year after year. Lead is readily taken up through skin and mucous membrane. The symptoms of chronic lead poisoning — tremor, irritability, fainting, mood disturbance, confusion — line up with the players’ complaints with a precision that is genuinely unsettling.

It is a good theory. It may even be partly true. It has three holes.

First, Franklin played the instrument to the end of his long life and showed no sign of poisoning. Second, lead poisoning was not understood as a disease in the eighteenth century, so not one of those deaths was ever confirmed by autopsy or diagnosis — only by talk. Third, the favourite flourish of this theory’s admirers, that Mozart died of lead absorbed from the armonica, is pure speculation; the cause of his death is still argued over, and his contact with the instrument was occasional at most.

Lead may well have harmed individual musicians. It does not explain the mass terror, and it explains nothing at all about the hypnotic effect that started this inquiry. Lead is silent. The armonica sang.

Here is where the “magic” actually lives.

The voice of the glass armonica sits in a band of roughly 1,000 to 4,000 hertz. Something strange happens to human hearing in that window. To work out where a sound is, the brain compares the signals arriving at the left and right ear. Below about 4,000 hertz it relies not on the difference in loudness but on the minute difference in timing and phase between the two ears. And the armonica’s tone — pure, level, with no percussive attack and no clear beginning, a sound that does not strike but seems to materialise and go on indefinitely — offers the brain almost none of the landmarks it normally uses.

The result: you hear the sound, but you cannot place it. It is not to your right or your left, not near or far. It is everywhere and nowhere. What follows is a mild disorientation, a sense that something is not quite right — a waking trance. Add the ghostly sympathetic ringing of neighbouring bowls, layering overtones no single note accounts for, and you have a sonority that bypasses the rational part of perception and goes straight to the floor where dreams and fears are kept.

One honest caveat. This acoustic account is an elegant popular hypothesis rather than a demonstrated medical fact. Real physiology sits under it — the timing-based localisation of low- and mid-frequency sound is textbook hearing science. But the step from mild disorientation to madness was taken by the era, not by the instrument.

Remember the Viennese physician in whose house the teenaged Mozart first touched the glass? His name was Franz Anton Mesmer.

Mesmer was the father of “animal magnetism,” a supposed invisible force by which he healed patients by inducing peculiar states in them. The glass armonica was his constant companion: he used it to put people into trance. The armonica was, in the most literal sense, the first soundtrack of hypnosis in history. Leopold Mozart, writing home to Salzburg in 1773, reported that Mesmer played the instrument unusually well — and that Wolfgang had tried it too.

The ending is delicious. In 1784 a royal commission was appointed in Paris to investigate animal magnetism, and Benjamin Franklin sat on it. The commission found no evidence of any magnetic fluid; the effects, it concluded, were produced by imagination. The inventor of the instrument helped demolish the doctrine that had made it famous. Mesmer left town and died in obscurity in 1815.

But his method outlived him. Mesmerism crossed to Britain and in 1841 the Scottish physician James Braid renamed it hypnotism. The doctor’s own name survives in the language we use for this very article: to be mesmerised.

So the “hypnotic effect” of the armonica’s music is neither metaphor nor superstition. It is literal historical fact: people were genuinely put into altered states with this instrument. The magic simply was not in the glass. It was in the human brain — and in the man who knew how to work it.

An honest investigation names what it could not close. Six things remain open.

• 01 How many armonica players actually suffered lead poisoning — no clinical record exists, because the diagnosis did not yet exist.
• 02 Whether the timing-based localisation account explains the instrument’s eerie quality, or merely describes it plausibly after the fact.
• 03 Who first attached a number to the legend, and when the count of “about fifty” entered circulation.
• 04 What the German child at that concert actually died of.
• 05 How much of the instrument’s decline was fear, and how much was the plain nuisance of transporting and tuning several dozen fragile bowls.
• 06 Why this particular timbre, and almost no other, still reads to modern ears as the sound of the uncanny — a question for perception research, not for history.

The instrument did not quite die. In the 1980s the Massachusetts glassblower Gerhard Finkenbeiner began building armonicas again, and his workshop has produced instruments for the small number of people who play them. Today perhaps a dozen musicians worldwide perform on the glass armonica regularly — among them William Zeitler, Dennis James and Dean Shostak, keepers of an almost-forgotten sound. It has been creeping back into film scores and concert halls, for listeners who want to hear the music people were once afraid of.

If one evening you hear it — level, pure, drifting in from no direction at all — do not be alarmed. You are not being driven mad. Your brain has simply lost its footing for a moment, and cannot work out that the sound was made by nothing but glass, water and the skin of ten fingers.

Or perhaps Deborah Franklin had it right, and it really is the music of angels. Two and a half centuries on, nobody is in a position to say for certain.

This article distinguishes throughout between what the historical record establishes, what is disputed among researchers, and what is legend repeated as fact. Where a claim could not be substantiated — including the widely circulated figure of some fifty dead musicians — it is identified as unsubstantiated rather than omitted. Corrections and sourced objections from readers are welcome and will be published.