The Dorabella Cipher

Summary

The composer Edward Elgar enclosed a note in his wife's thank-you letter to a friend. It is eighty-seven characters long, written in a private alphabet of hooked semicircles. The recipient never read it. Neither has anyone else. The most interesting fact about the case is that we appear to have Elgar's own key, and it produces nothing.

What is documented

The note. On 14 July 1897, Alice Elgar wrote from Great Malvern to thank the Penny family for a stay at the Wolverhampton Rectory. Edward Elgar, then forty and still an unsuccessful music teacher, put a folded note into the envelope and pencilled "Miss Penny" on the back.

The recipient. Dora Penny, then twenty-two, was the daughter of the Reverend Alfred Penny; her stepmother was a friend of Alice Elgar. Elgar and Dora remained friends until his death in 1934. He named the tenth of the Enigma Variations "Dorabella" for her.

The cipher. Eighty-seven characters across three lines. They are drawn from an alphabet of twenty-four distinct symbols. Each symbol consists of one, two or three approximate semicircles, oriented in one of eight directions. Three arc-counts multiplied by eight orientations gives exactly twenty-four. The orientation of several characters is ambiguous in the surviving image. A small dot appears after the fifth character of the third line. Nobody knows what it is for.

The original is lost. The note sat in a drawer for forty years. It became public when Dora Penny reproduced it in her 1937 memoir Edward Elgar: Memories of a Variation. The original was subsequently lost. Every analysis for the past ninety years has been performed on a printed reproduction in a book.

Elgar and ciphers. Elgar was a serious enthusiast. In 1896 he solved a cipher challenge published in the Pall Mall Magazine that had been presented as unbreakable. The same family of semicircle symbols appears elsewhere in his papers: on the so-called Cryptogram card associated with the Pall Mall solution, in a notebook of the 1920s alongside diagrams resembling clock faces, and in symbols annotated on an 1886 concert programme, over a decade before the Dorabella note.

The key that does not work. In his notebook, Elgar sets out this symbol alphabet and assigns the symbols to the letters of the English alphabet in order, folding I with J and U with V to fit twenty-six letters into twenty-four symbols. Apply that key to the Dorabella note and the output is gibberish. This is the central documented fact of the case. We are not missing the key. We appear to have it, in the author's own hand, and the door does not open.

It is not a simple substitution. In 2023 Viktor Wase applied modern cipher-solving algorithms to the note. He first demonstrated that those algorithms successfully solve ciphers as short as this one, then showed that they fail on this one, concluding that it is unlikely to be a monoalphabetic substitution in English or Latin. The point is methodologically important: the note's brevity is the standard excuse, and it does not hold.

Leading explanations

A cipher with an additional layer (possible, unproven). If Elgar's notebook alphabet is the outer shell, some further step, a transposition, a keyword, a phonetic shorthand, would explain why the direct substitution fails. No such step has been demonstrated.

A melody rather than a text (proposed, unproven). Javier Atance and others have suggested the eight orientations map to the eight notes of a scale and the one, two or three arcs to natural, flat and sharp. Elgar was a composer, the idea is not silly, and it has never been shown to produce anything that behaves like an Elgar melody by any test he would have recognised.

Deliberate nonsense (proposed, and not disprovable). Several analysts have argued that the note is a private joke: a text-shaped object with no plaintext behind it, sent to a friend who would be delighted and baffled in the correct proportions. This would explain the failure of every solver, including Elgar's own key. It has the weakness of every unfalsifiable claim, and the strength of fitting all the evidence.

Claimed solutions (none accepted). Eric Sams offered a reading in 1970 whose plaintext runs to 109 letters against the note's 87 characters; he attributed the surplus to phonetic shorthand. Tim S. Roberts derived a key from a phrase and produced near-nonsense. Wayne Packwood published a claimed complete decryption in Musical Opinion in 2020, having rearranged the characters by dot position and then shifted the presumed value of each glyph independently until a message emerged; the logic governing the shifts is not stated. In 2007 the Elgar Society ran a competition for the 150th anniversary of Elgar's birth and offered a prize. Entries were received. None was accepted, and the Society named the recurring flaw exactly: having matched the symbols to the alphabet, entrants ended up with an arbitrary sequence of letters.

Current status

Unsolved. A short document, an author who liked ciphers, a recipient who could not read it, an apparent key that does not work, a lost original, and a hundred and twenty-eight years of failure. The registry notes without comment that a message nobody can read is indistinguishable from a message that was never there.

Last reviewed: July 2026. Records are provisional. Where the evidence changes, the entry changes. Found an error? Tell us.