Dorabella cipher (Elgar)

The Dorabella cipher is an 87-symbol cryptogram that Sir Edward Elgar, one of the great British composers of his era, sent to Dora Penny on 14 July 1897. The message is written on the third page of an otherwise mundane letter in which Elgar thanks her for a recent visit.

Dora Penny (1874–1964), goddaughter of Elgar’s brother-in-law and a close friend of the composer, never managed to decrypt it. She kept the letter all her life and published it in her book Edward Elgar: Memories of a Variation (1937). Since then, 125 years of amateur and professional cryptanalysts have broken their teeth on it. The cipher remains unsolved.

Principle

The 87 symbols

The ciphertext is composed of 24 distinct symbols, each made of 1 to 3 arcs pointing in one of 8 cardinal directions (N, NE, E, SE, S, SW, W, NW):

1 arc:   Ↄ (E),  ⊃ (SE),  ⊂ (NE), … (8 variants)
2 arcs:  ⫝ (E,E), ≯, ≮, … (8 variants)
3 arcs:  ⫛, ⫝̸, … (8 variants)

Total: 8 directions × 3 arc counts = 24 symbols in theory. Elgar effectively uses 24 distinct symbols.

Observed distribution

Across the 87 symbols of the message:

• The most frequent appears 9 times.
• The rarest appear once each.
• The average frequency is 3.6 occurrences.

That’s a distribution close to a monoalphabetic substitution on a short English text — suggesting that each symbol encodes a letter (perhaps with a few homophones for vowels).

Why it resists

Several technical hypotheses explain the systematic failure:

1. Text too short (87 symbols)

87 characters is too short for robust frequency analysis. The most common letter appears 9 times — enough for an E “candidate” but not enough to distinguish between E, T, A, O, N (the 5 most common letters in English).

2. Multi-layer hypothesis

Several cryptanalysts (Eric Sams, 1970; Tim Roberts, 2007) suggest Elgar applied two successive operations: substitution + transposition, or substitution on a keyed alphabet + filler insertion. Without knowing the key, analysis becomes combinatorial.

3. Linguistic hypothesis

Elgar was passionate about music but also about crosswords and codes. Several readers have argued that the plaintext is not in standard English but in:

• Archaic English (Victorian vocabulary).
• A private code between Elgar and Dora (made-up words, abbreviations, internal references).
• A blend of musical phrases transposed into letters (B-A-C-H, D-O-R-A).

This hypothesis would render statistical analysis inoperative: no predictable structure.

4. “It isn’t a cipher” hypothesis

The most cynical hypothesis, proposed by some musicologists: it is not a message at all, but an Elgarian graphic joke designed to intrigue Dora — an aesthetic drawing without real linguistic meaning. Dora herself seemed sceptical: she wrote that Elgar told her “she would know”, but she never did.

Notable attempts

A few proposed plaintexts that have made noise:

• Eric Sams (1970) — proposes STARTS: LARKS, IT'S CHAOTIC, BUT A CLOAK OBSCURES MY NEW LETTERS, A, B… — statistically unconvincing.
• Tim Roberts (2007) — proposes a partial musical reading, never validated.
• Tony Gaffney (2013) — decrypts B HUDSNUFF SWEPT IT NIGHTLY, B SUNG O HEY NATURAL… — incoherent.
• Jeff Lefebvre & Spencer Garrett (2018) — use a musical-keyword analysis, partial result, debated.

No reading has reached consensus. The British Museum and the Elgar Society still consider the cipher unsolved.

The human context

Elgar’s affection for Dora Penny is documented but ambiguous. Elgar (40 in 1897) was married to Caroline Alice Roberts; Dora was 22. Their regular correspondence lasted until Elgar’s death in 1934.

Elgar also composed the thirteenth Enigma Variation (op. 36, 1899) in Dora’s honour — the famous “Variation XIII”, marked *** in the score (the three asterisks hiding her identity). The music of this variation directly quotes Mendelssohn’s Calm Sea and Prosperous Voyage — why? Yet another Elgar enigma.

The Dorabella cipher fits within a series of enigmas Elgar scattered through his life and work: the famous hidden theme of the Enigma Variations (never definitively identified), the Cello Concerto whose central theme remains mysterious, and several encrypted letters to various correspondents.

Why it matters

The Dorabella cipher is one of the very few unsolved short ciphers in classical cryptography. Alongside:

• Voynich (15th century, 240 pages, unknown language).
• Zodiac Z13 and Z32 (1970, ~50 symbols, serial killer).
• Beale 1 and 3 (1820, two of three Beale texts).
• Kryptos K4 (1990, 97 symbols, CIA sculpture).

Its resilience comes from the convergence of several factors: short text, possible linguistic ambiguity, private context, possible musical component.

The 24 glyphs

Reconstruction of the Dorabella alphabet as found in most modern decryption attempts. J merges with I, V with U.

A

B

C

D

E

F

G

H

I/J

K

L

M

N

O

P

Q

R

S

T

U/V

W

X

Y

Z