The Dorabella Code

---

What does the Dorabella Code mean?
Send us your thoughts using the form below.

A coded letter full of arcane cipher symbols penned by the composer Edward Elgar in 1897 remains one of cryptography's most enduring mysteries. It has achieved the status of a classic challenge, stubbornly immune to myriad attempts made to crack it by some of the world's most expert cryptographers.

Edward Elgar was a compulsive fan of crosswords, codes and conundrums. His manuscripts and papers are littered with word puzzles, ciphers, puns and anagrams. Early biographers specifically highlighted his interest in cryptography. He collected a series of articles published in the Pall Mall Magazine in 1896, which describe in some detail a large number of historic ciphering systems. The series contained various cipher challenges for readers to attempt, concluding with a complex, multi-stage puzzle described as uncrackable. Undaunted, Elgar solved it. As he unravelled various stages through to the keyword COURAGE, he used the margins of his Cockaigne Overture manuscript to record his progress. Elgar seemed sufficiently impressed at this feat to make a neat copy of the solution steps on to a set of nine carefully crafted cards, now at the Elgar Birthplace museum, leading to the final solution 'He who fears is half dead'.

Dora Penny, recipient of the mysterious coded missive, was the 22-year-old daughter of the Rector of Wolverhampton who had known the Elgar family for about 2 years, and would become immortalised as Dorabella of the Enigma Variations (Prom 31). Dora possessed a keen interest in machines and the emerging technologies of the period, the flyleaf of her 1898 diary was decorated with Morse Code symbols, and it is reasonable to assume that Elgar would have expected her to decipher his code.

Letter patterns

Victorian publications for both youngsters and adults were full of code-solving puzzles and enigmas. These typically used substitution ciphers where arbitrary symbols replace letters to produce something looking like this:

%%* *?=?%+}£?#

To solve an example as short as this we have to look at patterns of symmetry and repetition in the symbols. The first three-letter word starts with identical letters - probably vowels. Potential candidates such as EEL, AAH or OOH seem unlikely so it is worth checking consonants as well. Trying BB suggests 'BBC' - quite plausible in this context. Substituting equivalent symbols in the second word gives this result:

BBC C?=?B+}£?#

The ? appears three times and being positioned next to C and B will almost certainly represent a vowel, probably E:

BBC CE=EB+}£E#

from which likely solutions can be guessed. It could be BBC CELEBRATES or BBC CELEBRATED - there's no way of knowing - and, indeed, other solutions may work. Generally, the shorter the code, the greater the ambiguity, unless there are other external or contextual hints available.

Although the absence of word-breaks makes life very difficult, the cryptologist and musician Eric Sams, and others, have tried using these sorts of pattern-finding techniques on the Dorabella Code but their results are not very convincing. It does, nevertheless, contain some interesting patterns. Do the matching symbols at the end of the first two lines imply a rhyme? Part of a poem, perhaps? Is there any significance in the four-fold match of symbols 6 to 9 in the first line with symbols 4 to 7 in the second? There is an unusual absence of any other repeated sequences - though enough doubled symbols and other symmetries to imply a piece of bona-fide text.

Letter frequencies

To solve longer coded messages a standard approach is to count the frequency of occurrence of each symbol, rank them in order, and then match to a definitive ranking of letters by their expected frequency in English: E T A O I N ... etc., a sequence burnt into the brain of all keen code-breakers! (With Latin the equivalent series begins E T U A S ...) The match will not always be perfect, and letter frequencies will wobble slightly according to whether the text is technical, vernacular or literary. Heavy use of rhyme and repetition in poetry, in particular, can skew the balance. Nevertheless, with a little trial and error, sometimes shuffling letters a place or two in the ranking order, it offers a good place to start.

100 characters is usually considered to be the minimum message length to produce a sufficiently large sample of letters for this approach to work. At 87 characters the Dorabella Code is a bit brief for comfort, but the frequency distribution of the symbols - which is a reasonable match to what might be expected for a straightforward substitution cipher - makes it a tempting exercise. Again, many have tried, but without success.

Cryptic sheep

What about the symbols themselves? Described as looking like a flock of sheep in one cryptological tome, do they provide a clue? Each symbol has one, two or three 'humps' in one of eight orientations - a total of 24 available symbols. Many traditional ciphers with which Elgar would have been familiar use a 24-letter alphabet where I/J and U/V are treated as equivalents, and that appears to be the case here. In the late 1920's Elgar wrote out a number of examples of this same code, along with a cipher key, in an exercise book now at the Birthplace Museum. Here A, B, C are respectively matched to a single, double then triple upright - D, E, F continues with the same sequence rotated 45 degrees - and so on to complete the alphabet. Elgar included two brief coded messages as examples, which when deciphered read MARCO ELGAR (the name of his pet spaniel) and A VERY OLD CYPHER. However when the same key is used on the Dorabella Code a sequence emerges which makes very little sense. On the same pages alternative cipher alphabets are started, but none work any better. It is possible that late in life Elgar was consciously dropping hints, but simply couldn't remember the precise order he had used some thirty years earlier.

There are, though, other instances where Elgar used the same symbols. In particular a sequence of triple cipher characters in the sequence shown below, followed by one upright double and a single, was scribbled on the top card of the COURAGE cipher card set described above, probably at about the same time that he composed the Dorabella Code. It implies the match suggested here, with the second row matching double-humped characters, and the third row single. Using this key to decipher the code produces quite an interesting result. Try it!

A number of plausible words seem to emerge out of the surrounding gibberish, suggesting that it might be on the right track, but not quite there. Should one or more lines or half-lines of letters be reversed perhaps?

Additional disguises

It is possible that Elgar thought the transcription of symbols to letters would be so straightforward that an additional level of disguise was fair game. The symmetry in the symbols, though, seems to have been completely lost on Dora. Many years later, following Elgar's death and after consulting with his daughter, Carice, they sent a copy to the Director of London University's School of Oriental and African Studies, an expert on Middle Eastern and Persian languages, evidently under the impression that it might be a form of Arabic script.

The fact that the statistical distribution of the symbols roughly matches that expected for a straightforward substitution cipher makes it unlikely that Elgar used a more sophisticated multiple ciphering system, such as the use of changing alphabets with a keyword, which has the effect of evening out the frequency differences. It also means that non-textual solutions are highly unlikely - such as numbered directions, a piece of music, or perhaps a knitting pattern! A number of musical solutions have been proposed, but it's very easy to devise a superficial 'musical' mapping of almost any code or data set. The results rarely make real musical sense and the exercise would have to produce something very distinctive and recognisable to be convincing.

Elgar probably used a simple system of shuffling, such as swapping round letter pairs, exchanging first and last letters of words, having part of the text flow backwards, or else used spoilers such as inserting alternate random letters, or missing out certain vowels. Another example in the 1920's exercise book reads DO YOU GO TO LONDON TOMORROW? Designed to confuse, as well as lacking any E's, the most common letter in English, the O's are marked with little downward strokes and remaining letters with upward strokes, hinting that it was to emerge as D YU G T LNDN TMRRW? Elgar, a great fan of gadgets and new technology would have had no problem with txts! When in playful mood Elgar's spelling also became rather eccentric, sometimes featuring excessive alliteration, rhymes, puns and spoonerisms. Yet more hurdles for the hapless code-breaker!

Football, racing and a knot garden

Events leading up to the composition of the Dorabella Code may provide some clues. In the preceding week the Elgars had been staying with Dora and her family at the Rectory in Wolverhampton. Dora's stepmother was from the Baker family who figure prominently in the Enigma Variations. She had been a fellow Geology student with Elgar's wife Alice, and they remained firm friends. It fell to Dora to entertain Edward, and there is no doubt that in spite of an age difference of nearly 20 years, for a while they enjoyed a close platonic bond. They had a mutual appreciation of kite flying, cycling and football, being enthusiastic supporters of Wolverhampton Wanderers. During their stay in July 1897, the glorious summer of Queen Victoria's Diamond Jubilee, Dora took Edward to the races (the rectory garden backed on to the racecourse), a rope factory, and nearby historic Boscobel House.

Boscobel was full of romantic associations with the English Civil War. Its Catholic residents had hidden the future Charles II in the famed nearby Royal Oak tree, and in a priest hole at the house, as he made his escape following the disastrous battle of Worcester, events with special regional and spiritual resonance for Elgar. Running down the side of a knot garden next to the house, Latin letters set into three lines of cobblestones read:

sept id sept 1651 In hac domo Carolus Secundus
tutela quinque fratum de stirpe Penderel potitus
est eoreom denque ope incolumnis evasit

(The 'official' translation runs: "On the 6th September 1651 Charles II in this house was safely preserved by the Penderel Brothers Five from which he escaped safe and sound" - many thanks to the helpful English Heritage staff at Boscobel House for their generous assistance tracking this down.)

Note the matching ending of the first two lines, and a repeated 4-character nque sequence. Is it possible that Dora and Edward spotted the obscure outline of 'cobbled' words in the garden, wondered what it was, Elgar managed to find out back home, and sent it on to Dora - but playfully edited and disguised? What if letters are arranged in some sort of knot-pattern? Might a convincing process work backwards to match this plausible text to the Dorabella Code? Or does the code conceal less innocent secrets?

Well, the enigma of the Code remains. In this anniversary year the
Elgar Society is offering a prize of £1500 for its solution. You may be tempted to tackle it, but be warned. It can become a seriously addictive pursuit!

©Kevin Jones 2007

Kevin Jones researches and writes about historical and contemporary links between music and science, with a special interest in music and codes. He won the THES/OUP Science Writing Prize in 1999. He is also a composer and was Professor of Music at Kingston University until 2007.

A Hidden World of Musical Codes

Variations on an Enigma

---

Have you tried to crack the Dorabella Code?

If you have any comments or questions about Elgar's mysterious letter, please send them to us. We will publish as many as we can.

What you said: Peter Brooks As a postscript, it's worth noting that Elgar used the symbols in his cipher on an earlier occasion (I'm not sure of the exact date, but around April 1886). Mike Combridge, in collaboration with the Elgar Birthplace Museum supervisor, Chris Bennett (I believe), reported the details to the Yahoo! group in May 2006, although it has been known about for many years and even mentioned in Jerrold Northrop Moore's book "Edward Elgar: A Creative Life" (p114). In the margin of a concert note for a performance in honour of Liszt, there are 18 characters arranged in groups of 3 and 6, with one symbol apparently superscripted. It too has so far resisted attempts to decrypt easily (efforts published so far are not convincing). It's worth mentioning that the Yahoo! group also has a Truetype font (Dorabella.ttf) freely available in its FILES section that allows you to more easily experiment with ideas by using the symbols themselves; this also helps a great deal in documentation. Jean Palmer This cipher is a ‘simple substitution’ applied to a message that was written in the form of a puzzle for Dora. After replacing the symbols with the correct alphabet the puzzle message literally reads –BLTACEIARWUNISNFNNELLHSYWYDUOINIEYARQATNNTEDMINUNE HOMSYRRYUOTOEHO’TSHGDOTNEHMOSALDOEADYA The form of this puzzle is twofold – a variation of ‘backslang’ and phoneticised spellings. Elgar was well known to use the latter – the only slight evidence for the use of ‘backslang’ I can find is the following quote from another of Elgar’s letters to Dora – “are you living backwards like the Queen in Alice”. Unravelling the puzzle, as best I can make out, (it contains a few enciphering errors) what Elgar meant to say was – B (Bella) hellcat i.e. war using (enough??) hens shells why yourantiquarian net diminishes, am sorry youtheo O ‘tis God then me so la do E (Elgar) Adieu The first line didn’t make much sense to me until I remembered seeing the following in a letter from Elgar to Dora Penny dated Sept. 24th 1898 ‘... and then some Sunday at Wolverhampton you can give us tea and fire eggs at me as of yore.’ The middle line – ‘antiquarian net’ – would seem to mean circle of elderly friends. The last line – On Elgar’s visits Dora would often sit turning the pages, or sing, for Elgar at the piano – I assume there must have been occasions when asked to do this she had declined (her father being a Rector) and religious duties taking priority. Stephen Colbourn Why did Elgar write only one cryptogram in this manner (or only one has survived) and to Dora? It is rather short for analysis, but there is a probable letter distribution. Unless he expected Dora to sit down and decrypt the message, which she would not have found easy to do, he must have outlined a method to her during his visit to Wolverhampton in the second week of July 1897. The likely method is a simple substitution followed by a simple trnsposition; then the result is transcribed into Elgar's personal compass diagrams. For such a message to be a communication, there must be agreement between sender and recipient. Perhaps poor Dora didn't follow him or she forgot. Or, Elgar did not want her to read the message? Kevin Jones replies: Elgar did indeed use this cipher elsewhere. As well as the examples in the 1920's exercise book, Elgar scribbled an 18 character code using the same cipher symbols in the column of printed programme notes for a concert he attended at Crystal Palace in April 1886 - opposite a musical example from Liszt's "Les Preludes". (Copy at the Elgar Birthplace Museum.) Annotations on other pages are not ciphered - so it's possible that this may have been added at a later date. He heard the same piece at a concert he attended in June 1898 not long before starting work on the Enigma Variations. Although there is no specific evidence either way, it is conceivable, of course, that Elgar didn't intend Dora to read the message whilst still being a genuine code harbouring a concealed secret - like the hidden melody in the Enigma Variations where a latent spirit sleeps below the surface. Dora's father had just returned from Melanesia where he had been a missionary for many years. Fascinated by local language and culture, he possessed a few traditional talismans decorated with arcane glyphs. Perhaps such an item surfaced as a conversation piece during the Elgar's week in Wolverhampton? And if Dora recalled this when writing her memoirs, it might account for the fact the coded message was referred to as an 'inscription' when communicating with the director of SOAS many years later. Peter Brooks For those interested in examining possible paths to a solution I'd recommend the moderated Yahoo! group Elgar-Cipher, started by Rob Marston some six years ago. For my part, I am increasingly confident that the message consists of two parts separated by an evident period on the last line. The entire message appears to have been written in reverse (or rotated upside down before being dated - the effect is the same) and rotated by one character. The resulting "first part" appears to be in Latin, while the "second part" appears to have been written in English and obscured further by being arranged not horizontally but vertically. Others have identified what they see as a phonetic solution (similar to that proposed by Eric Sams), while still others are exploring possible musical and graphical origins. The nice thing about this cipher is that it's not the private playground of professionals - anyone can try their hand at solving it, and you need nothing more than an inquisitive mind and pencil and paper (although a computer does help sometimes :) ). Your comment and name: Your email address: Please note that we cannot guarantee to publish all comments. See Terms of Use below for more information.