Edward Elgar's Dorabella Cipher hides a readable secret message that a solver could still recover
Where the evidence lands: UnresolvedThat the Dorabella Cipher, the 87-symbol note Edward Elgar sent to Dora Penny in 1897, encodes a specific, coherent message in ordinary language that would be legible if the correct key were found, and that one of the various proposed decipherments has, or could, recover it.
Believed by: Elgar enthusiasts, amateur and professional cryptanalysts, and puzzle hobbyists; it sits alongside the Voynich manuscript and the Beale ciphers in the small canon of famous unbroken texts
The full story
What is documented
Start with what is not in doubt, because a surprising amount is settled. In July 1897, Edward Elgar, then a forty-year-old provincial music teacher years away from fame, spent a few days with the Penny family at the rectory in Wolverhampton. Shortly after, a thank-you letter went to the rector's stepdaughter, Dora Penny, a young woman in her early twenties. Folded in with it was a small card in Elgar's hand, written entirely in cipher.
The note is short and precise: 87 characters across three lines, each character a cluster of one, two, or three approximate semicircles, or loops, turned in one of eight directions. That scheme yields 24 distinct symbols. There is no key, no plaintext, and no explanation. Dora kept it, puzzled over it, and, by her own account, never had the faintest idea what it said. In 1899 Elgar would make her famous anyway, naming Variation X of his Enigma Variations“Dorabella” in her honor.
Two more facts matter. First, Elgar was a documented cipher enthusiast: in 1896 he solved a Nihilist cipher that the writer John Holt Schooling had published as unbreakable, painted the answer on a wooden box, and kept index cards on his method. Second, when Dora reproduced the note in her 1937 memoir, Edward Elgar: Memories of a Variation, it passed into public circulation and has resisted every reader since. So the question this file weighs is not whether the note exists or who wrote it. It is the narrower one: does it contain a message anyone can recover?
The case for a hidden message
The believer's case is genuinely strong, and worth stating at full strength. Begin with the author. This was not a random scribble; it came from a man who loved ciphers and was good at them, who cracked a cipher billed as insoluble for the sheer pleasure of it, and whose letters and manuscripts are strewn with anagrams and codes. When someone like that hands a friend a page of careful symbols, the natural inference is that the symbols mean something.
Then there is the length. An 87-character substitution cipher is, by the ordinary rules of cryptanalysis, short but very much breakable. Automated solvers routinely read messages this length. The whole discipline of frequency analysis was built for exactly this problem, and it works on far nastier texts every day. On paper, Dorabella looks less like the Voynich manuscript and more like a Sunday puzzle that simply has not had the right eyes on it.
And the note did not come from nowhere. It sits inside a relationship, a warm, playful friendship between an older composer and a younger admirer, which gives any hidden line a plausible emotional charge. It is easy to imagine an affectionate joke, a confession, or a private endearment folded into the loops, meant for Dora alone.
A cipher expert, a note to a friend, only 87 symbols. Every instinct says there is a sentence in there, and that someone patient enough will read it.
That is the honest steelman: a real note, by a real codemaker, of a length that ought to yield, addressed to someone he cared about. If a message is hiding anywhere in the world's unsolved ciphers, this looks like one of the likelier places to find it.
Where the claim breaks down
The trouble is that the ought has been tested, and the note has not yielded. Over a century of attempts has produced proposed solutions in abundance and an accepted solution never, and the reasons run deeper than bad luck.
Consider the proposed readings themselves. Eric Sams's 1970 solution in The Musical Times, the most serious early attempt, required an intricate and irregular chain of steps and emerged with strained, barely idiomatic text. When the Elgar Society ran its competition, the judges observed that even the most ambitious entries, having matched the symbols to letters, invariably ended in an arbitrary sequence. That is the signature of a method with too many free choices: give a solver enough latitude and the cipher can be steered toward almost any phrase they already have in mind. A decipherment only persuades when the cipher forces the answer, and none of these do.
Now the decisive piece. In 2023, Viktor Wase put the note against modern solvers and reported a pointed result: those tools break comparable 87-character English and Latin substitution ciphers about 98 percent of the time, yet they fail on Dorabella. The very argument the believer leans on, that a cipher this short should crack, turns against the claim. If ciphers like this one fall almost every time and this one refuses, the likeliest reading is that it is not a simple English substitution at all.
The statistics point the same way. Eighty-seven characters drawn from 24 symbols, spread more evenly than the steep frequency curve of natural English, is not what plain text disguised letter-for-letter looks like. It is consistent with something nonstandard, a phonetic or musical mapping, a homophonic scheme, or a private code, and it is consistent with the note simply being too short to read reliably. None of that recovers a message. It explains why the recovery keeps failing.
Why the puzzle endures
A note this small should not command this much devotion, and the reasons it does say as much about us as about Elgar.
First, it is real and unsolved, a rare and potent combination. Most tantalizing mysteries collapse into hoax or error on inspection; this one does not. The author is certain, the note is authentic, and the puzzle is open, which lends it a legitimacy that pulls serious cryptanalysts back again and again.
Second, it looks winnable. Eighty-seven symbols on a single card feels like a weekend's work, not a life's, so every hobbyist can believe the glory is one clever insight away. The long trail of failures rarely discourages the next attempt, because each newcomer assumes the failures simply lacked their particular idea.
Third, the romance does a great deal of work. An older, married composer sending a coded note to a younger woman is a frame that almost writes its own secret, and readers supply the confession the loops withhold. The blank is filled with whatever longing the reader brings.
The note is a mirror. Because no one can read it, everyone can pour into it the message they most expect to find.
Finally, it belongs to a famous family. Named in the same breath as the Voynich manuscript and the Beale ciphers, Dorabella inherits the glamour of the great unbroken texts. Membership in that small club keeps the attention flowing long after the note's modest length would otherwise justify.
Where the evidence lands
Hold the two claims apart, because the gap between them is the whole of the case. That Elgar wrote Dora a note in 87 looped symbols, that she never read it, and that no one has read it since, is documented and settled. That the note conceals a specific, recoverable message, and that some proposed solution has or could reach it, is the rated claim, and it remains unestablished. Every published reading rests on subjective, bendable steps and none has won acceptance, while the strongest statistical work suggests the note may not be ordinary plaintext in the first place. On that claim the verdict is Unproven.
Notice what unproven does and does not mean here. It does not mean hoax; the note is real. It does not mean empty; a private phonetic or musical scheme could still be lurking. It means only that no one has shown what is inside, and that the confident solutions on offer are not compelled by the cipher. It is even possible the note is effectively unsolvable as usually approached: too short, too idiosyncratic, and keyed to a private code that died with its author.
That places Dorabella squarely among its siblings. Like the Voynich manuscript, whose text mimics real language closely enough that neither meaning nor hoax fits cleanly, and like the Beale ciphers, where one solved sheet keeps hope alive for two that may hide nothing, the Dorabella note lives in the honest middle: genuinely unread, plausibly meaningful, and not yet, or perhaps never, legible. The fascination is earned. The solution is not yet in hand.
What's still unexplained
- Whether the note is ordinary plaintext at all. Wase's work leaves open that it uses a nonstandard scheme, phonetic, musical, or homophonic, in which case decades of substitution attempts have been aimed at the wrong kind of target and unsolved shades toward unsolvable-as-assumed.
- What the 24 symbols actually map to. The loop-count-and-orientation structure looks systematic, but whether it encodes 24 letters, a reduced alphabet, sounds, or musical values has never been pinned down.
- Whether Elgar intended a decodable message or a private flourish. He expected Dora to work it out and seemed unbothered that she could not, which is at least consistent with a half-playful gesture meant to be felt rather than rigorously read.
- Whether anything in Elgar's surviving papers, the index cards or coded sketches, uses the same method and could serve as a crib. Nothing found so far has cracked the note, but the archive has not been exhausted.
Point by point
The claim: One of the published readings, such as Eric Sams's 1970 phonetic solution, recovers the true message, so the cipher is effectively solved.
What the record shows: No proposed reading has been accepted. Sams's solution required an irregular chain of ad hoc steps and produced strained, awkward text; the Elgar Society judges noted that the strongest competition entries, having matched Elgar's symbols to the alphabet, still ended in an essentially arbitrary sequence of letters. When a method involves enough free choices, it can be bent toward almost any target phrase, which is exactly why a decipherment is only convincing if it is forced by the cipher rather than coaxed out of it. None of the Dorabella readings clears that bar.
The claim: An 87-character substitution cipher should be easy to break, so a real message must be sitting there waiting.
What the record shows: This is the strongest point for the puzzle, and it cuts in an unexpected direction. Wase's 2023 study confirmed that automated solvers break comparable 87-character English and Latin substitution ciphers roughly 98 percent of the time, yet the same tools fail on Dorabella. That gap is telling: the note's very resistance to methods that should work suggests it is not a straightforward English or Latin substitution at all. The solvability of similar ciphers makes Dorabella's silence evidence against the simple-plaintext assumption, not for it.
The claim: Elgar was a proven codemaker who cracked a cipher declared unbreakable, so the note surely hides a genuine, recoverable message.
What the record shows: That Elgar loved and understood ciphers is documented: he solved Schooling's Nihilist challenge and kept notes on how. But expertise cuts both ways. A skilled enthusiast writing a private note to a friend is exactly the person who might build something idiosyncratic, a phonetic shorthand, a musical mapping, a personal joke, rather than a textbook substitution of ordinary English. His skill raises the odds that the note is clever; it does not guarantee that it is a clean sentence anyone else can read.
The claim: The note's structure is just an ordinary alphabet in disguise, one loop-symbol per letter.
What the record shows: The statistics sit awkwardly with that. The 87 characters draw on 24 distinct symbols with a flatter, more even distribution than the steep letter frequencies of natural English, where a handful of letters dominate. That profile is part of why substitution attacks stall, and it is consistent with a nonstandard scheme (homophonic, musical, or private phonetic) or with a text short enough that its statistics are simply unreliable. The structure neither proves a hidden message nor rules one out; it mainly shows the note does not behave like plain English on the page.
Timeline
- 1896Elgar solves a Nihilist cipher that the writer John Holt Schooling had published in the Pall Mall Gazette as effectively unbreakable. Pleased with himself, Elgar paints his solution on a wooden box and keeps a set of index cards detailing his method. His letters and sketches are peppered with anagrams and codes, establishing a lifelong fascination with hidden writing.
- 1897-07-14After Elgar and his wife Alice stay with the Penny family at Wolverhampton, a thank-you letter is sent to the rector's stepdaughter, Dora Penny, then in her early twenties and some seventeen years Elgar's junior. Enclosed is a card in Elgar's hand covered in 87 looped symbols, with no key and no explanation.
- 1899Elgar premieres his Variations on an Original Theme, the Enigma Variations, the work that makes his name. Variation X is titled Dorabella and dedicated to Dora Penny, binding her permanently to Elgar's love of puzzles and hidden meanings.
- 1920sBy Dora's own later account, she asks Elgar about the note; he expresses surprise that she never worked it out and offers no help. He dies in 1934 without ever explaining it, taking any intended reading with him.
- 1937Dora Penny, by then Mrs. Richard Powell, publishes her memoir Edward Elgar: Memories of a Variation, reproducing the cipher in an appendix and admitting she never had the slightest idea what message it conveyed. This is how the puzzle reaches the public.
- 1970The musicologist and cryptographer Eric Sams publishes a detailed analysis in The Musical Times, proposing a phonetic reading of the note. His method is intricate and hard to follow, and his solution never wins broad acceptance among specialists.
- 2007To mark the 150th anniversary of Elgar's birth, the Elgar Society runs a Dorabella Cipher competition with a prize of 1,500 pounds for a convincing solution. Entries arrive, some ambitious, but every one bottoms out in a fairly arbitrary string of letters. No prize is awarded, in 2007 or 2008.
- 2023Viktor Wase publishes Dorabella unMASCed in the journal Cryptologia, showing that state-of-the-art solvers crack 87-character English and Latin monoalphabetic substitution ciphers around 98 percent of the time yet fail on Dorabella, arguing that it is unlikely to be a simple substitution cipher in either language.
Unresolved. The documented record is not in dispute: in July 1897 the composer Edward Elgar sent Dora Penny a note written entirely in a cipher of 87 looped symbols, she never read it, and she reproduced it in her 1937 memoir. The rated claim is narrower, that the note conceals a specific plaintext message and that one of the many proposed solutions is correct. That remains unproven. Every published reading rests on subjective, ad hoc steps and none has won acceptance; a 2023 study showed that modern solvers crack comparable 87-character substitution ciphers almost every time yet fail here, which suggests the note may not be simple English plaintext at all. It is a genuine unsolved cipher, and possibly an unsolvable one, but no message has been shown to be inside it.
Sources
- 1.Dorabella Cipher, Wikipedia
- 2.Elgar's Cipher Letter to Dorabella (The Musical Times, 1970), Eric Sams (ericsams.org) (1970)
- 3.Dorabella unMASCed: the Dorabella Cipher is not an English or Latin Mono-Alphabetical Substitution Cipher, Viktor Wase, Cryptologia (Taylor & Francis) (2023)
- 4.The Dorabella Cipher, Cipher Mysteries
- 5.The Artist of the Unbreakable Code, Nautilus
- 6.Solving the Dorabella Cipher, Tim S. Roberts, CQUniversity (unsolvedproblems.org)
- 7.Unsolved: Dorabella Cipher, Derek Bruff, Historical Cryptography
Help us investigate
This is a living case file. If you spot an error or know evidence we missed, tell us, and weigh in on where you land.
Where do you land?
Cast your read on this one.