Everything anyone knows about the Beale treasure, and everything anyone has to work with on the Beale ciphers, rests on one anonymous pamphlet that sold for fifty cents. There is no independent record of the gold, no iron box in a museum, no original sheet of ciphertext anyone can examine. The pamphlet is the entire evidentiary base. That makes reading it closely the only test available, and it is a test the pamphlet does not comfortably pass: its own printed key does not cleanly produce its own printed solution.
Key Takeaways
- The 1885 pamphlet is anonymous. The only name attached to it is the one on the copyright entry, and the narrator says outright that he has withdrawn his own.
- The Declaration of Independence the pamphlet prints as the key is not the text held by the National Archives. There are 28 differences between them: six shift the numbering itself, so the two texts drift in and out of step, and five more change the first letter of a word, which is the only part of it a book cipher uses.
- Working cipher No. 2 straight off that printed key gets most of the way to the pamphlet's stated plaintext and no further. In the first 200 letters, 159 come out right and 41 come out wrong.
- No word in the pamphlet's Declaration begins with x, y or z, yet the plaintext needs both x and y. The published rule cannot produce the published answer.
- Cipher No. 1 contains ten numbers larger than the Declaration is long, so that document cannot be its key at all.
- A researcher who read the Lynchburg newspapers across the relevant years found Robert Morriss everywhere in them and Thomas J. Beale nowhere.
Where do the Beale ciphers actually come from?
The title page announces "AUTHENTIC STATEMENTS REGARDING THE Treasure Buried IN 1819 and 1821 NEAR BUFORDS, IN BEDFORD COUNTY, VIRGINIA, AND Which Has Never Been Recovered." Beneath that: "PRICE FIFTY CENTS." Overleaf sits the copyright line, "Entered according to act of Congress, in the year 1885, by J. B. Ward, in the Office of the Librarian of Congress, at Washington."
The narrative itself is unsigned, and the narrator explains why in the closing paragraphs: "I have decided upon withdrawing my name from the publication, after assuring all interested that I have given all that I know of the matter, and that I cannot add one word to the statements herein contained."
He also tells the reader, in passing, that nothing printed is an original. Of the three ciphers and the letters he writes: "They were carefully copied, and as carefully compared with the originals, and no error is believed to exist." What became of those originals, of the iron box, or of Robert Morriss's own papers, the pamphlet never says. Nothing physical is offered for anyone to inspect, then or since.
The story also builds in its own reason why the missing key can never be produced. Beale's letter, as printed, says he left it "in the hands of a friend in this place, sealed, addressed to yourself, and endorsed not to be delivered until June, 1832." It never arrived, and the friend is never named.
Which Declaration of Independence unlocks the Beale ciphers?
The pamphlet does not simply say the Declaration of Independence unlocks cipher No. 2. It prints its own Declaration for the reader to number. That printed text is not the one the National Archives holds.
A note on the counting that follows, since it does the work in this section and the next. The figures come from the Wikisource transcription of the pamphlet, which reproduces the 1885 printing in full and is linked below; its reading "inalienable" was checked against a scanned facsimile of the pamphlet held in the National Security Agency's released file before being relied on here. Hyphenated compounds are counted as the two words they are made of, which is the counting under which cipher No. 2's solution appears at all.
Set the two texts side by side and 28 differences show up. Six change the number of words. The other 22 swap one word for another and leave the count alone: "shewn" becomes "shown", "endeavoured" becomes "endeavored", "harrass" becomes "harass", "depository" becomes "depositary".
The pamphlet's Declaration runs to 1,324 words. The National Archives transcript runs to 1,322.
The six count-changing differences do not accumulate tidily, and this is the part worth getting right. The two texts stay exactly in step for their first 154 words. Then the pamphlet's "to institute a new government", where the archived text reads "to institute new Government", puts it one word ahead. At word 520 that is cancelled out: the pamphlet's "in the meantime" is one word where the archived "in the mean time" is two, and the texts fall back into step. They diverge and converge again at 909, at 1070, at 1138 and at 1200, finishing two words apart. Anyone solving from a modern Declaration is therefore not uniformly offset. They are in step across some stretches and out of step across others, with nothing to mark the boundaries.
The substitutions matter more than they look, because a book cipher uses only the first letter of the word it lands on. Of the 22, five change that letter:
Positions below are counted in the pamphlet's Declaration, since that is the text the cipher reads. Where the running offset puts the archived word at a different number, both are given.
| Position |
Pamphlet's Declaration |
Yields |
Archived Declaration |
Yields |
| 95 |
inalienable |
i |
unalienable |
u |
| 210 (archived 209) |
now |
n |
more |
m |
| 549 |
of |
o |
for |
f |
| 653 |
offered |
o |
affected |
a |
| 825 |
powers |
p |
Forms |
f |
Word 95 is the cleanest illustration, because it sits well before the first count-changing difference. At that position the two texts are still word for word in step, so the comparison is like for like and nothing about the counting can explain it away. The number 95 appears in cipher No. 2. From the pamphlet's Declaration it yields i. From the archived Declaration it yields u.
The pamphlet's own numbering does not add up
To help the reader count, the pamphlet threads bracketed position markers through its Declaration, every ten words: (20), (30), (40), and so on. There are 87 of them.
Check them against an actual count and 23 hold. The other 64 do not.
The markers behave until (240). From (250) onward they run one word ahead of the text. Then (480) is printed twice, ten words apart, and after that the printed figures fall behind the true count instead. The gap widens as the document goes on.
The last six markers break the pattern in a different way. Having run in tens the whole way, the pamphlet suddenly numbers six consecutive words one at a time, printing (811) through (816) against "fundamentally, the powers of our governments". Then the numbering simply stops. Marker (816) falls on word 828, which means the printed guide covers the first 63 percent of the document and abandons the reader for the rest, in a stretch that includes one of the five substitutions where the pamphlet's text yields a different letter from the archived Declaration.
None of this is fatal by itself, since the numbers in the ciphers are what matter and a reader can always count for themselves. But it does mean the printed guide to the key is unreliable, in a pamphlet whose entire claim to seriousness is that its papers were copied with care.
No word in the key begins with X or Y
This is the harder problem, and it is checkable in a minute.
The Declaration as the pamphlet prints it contains no word at all beginning with x, y or z. Not one, across all 1,324 words.
The plaintext of cipher No. 2, as the pamphlet gives it, needs x and y repeatedly. It has "excavation", "six feet below the surface of the ground", "exchange", and "exact". It has "county", "jointly", "eighty", "securely", "roughly" and "locality".
So the stated solution cannot be produced from the stated key by the rule the pamphlet describes. Somebody took liberties the pamphlet never mentions.
You can watch this happen. Number the pamphlet's Declaration straight through and read off the first letters that cipher No. 2 points at. The opening emerges unmistakably, and imperfectly. Here are the first forty characters the key produces, above the forty the pamphlet says should be there:
iharedeposctedinthecountfofgedfordaboatf
ihavedepositedinthecountyofbedfordaboutf
And characters 41 to 80:
oirmilesfrombtfordsiiaaehcarationorraalt
ourmilesfrombufordsinanexcavationorvault
Thirteen of those eighty characters are wrong, five in the first block and eight in the second. Across the first 200 letters, 159 land on the right letter and 41 do not.
Now find the x and the y in those eighty characters and the argument closes on itself. Character 25 of the first block has to be the y of "county"; cipher No. 2 gives the number 811 there, and by a true count, not the pamphlet's printed marker of the same number which sits twelve words later, word 811 is "For", so out comes f. Character 25 of the second block has to be the x of "excavation"; the cipher gives 1005, word 1005 is "have", so out comes h. Those are not near misses. They are the two places where the key has nothing to offer, turning up exactly where the letters it cannot supply are needed.
That is far too close to be coincidence, and far too rough to be a working method. Cipher No. 2 is genuinely keyed to something very like this Declaration, and it is not keyed to this Declaration by any rule the pamphlet states.
What liberties were taken cannot be settled from the pamphlet, but the possibilities are narrow. Either the solver worked from a copy of the Declaration numbered differently from the one he printed, or he supplied by hand the letters his key could not give him, or both. The pamphlet itself points away from anything systematic. Its narrator never describes a procedure at all: the solution, he writes, came when "accident revealed to him the explanation of the paper marked '2'", and his advice to everyone who follows is that "accident alone, without the promised key, will ever develop the mystery." That is the difficulty facing anyone attacking the other two the same way. There is no clean rule to copy, because the one worked example was never presented as one.
Why the Declaration cannot be the key to cipher No. 1
Cipher No. 1, the one that is supposed to give the location, rules the Declaration out on arithmetic alone.
It runs to 520 numbers, and ten of them are larger than the Declaration is long: 1431, 1496, 1629, 1701, 1706, 1780, 1817, 2018, 2160 and 2906. A document of 1,324 words has no word 2,906. Whatever keytext No. 1 was built on, if it was built on one, it was at least twice the length of the Declaration.
Cipher No. 3 sits differently. Its 618 numbers top out at 975, comfortably inside the Declaration's range, so on size alone the Declaration remains possible for it.
| Cipher |
What it is said to hold |
Numbers |
Highest number |
Declaration possible on size? |
| No. 1 |
The exact location of the vault |
520 |
2906 |
No |
| No. 2 |
The contents of the vault |
762 |
1005 |
Yes, and it is the one with a stated solution |
| No. 3 |
Thirty names, next of kin, residences |
618 |
975 |
Yes |
That asymmetry is worth holding onto. Ruling a candidate keytext out is cheap and can be done with arithmetic. Ruling one in requires the text itself, which is the position every book cipher without its book leaves you in, and it is why this puzzle has proved more durable than most of the unsolved codes and ciphers it sits alongside. The sculpture Kryptos is unsolved because its cipher is hard. Beale Nos. 1 and 3 are unsolved because the thing that would make them easy has never been found.
Does cipher No. 3 hold what it is supposed to hold?
Beale's letter instructs Morriss to divide the treasure "into thirty-one equal parts", keeping one for himself. That leaves thirty associates. The accompanying note says paper No. 3 carries "the names of all my associates" and, "opposite to the names of each one", the "names and residences of the relatives" they leave their shares to.
Cipher No. 3 holds 618 numbers. Spread across thirty men, that is about 20.6 characters each, and each man needs his own name, a relative's name, and a place of residence.
Twenty letters is roughly one full name and nothing else. "Thomas Jefferson Beale" is exactly twenty letters on its own, before any relative or any county is named. Fitting three items per man into that allowance would need abbreviation severe enough that the result could not be read back with confidence, which is the opposite of what a document naming heirs is for. The pamphlet does not address this anywhere, and it is the plainest arithmetic problem in the whole affair.
Is there any record of Thomas J. Beale?
The most useful check on the story is not cryptographic at all. It is archival, and some of it sits in a file the National Security Agency released under the Freedom of Information Act in December 2001.
In that file is a research report by Carl W. Nelson Jr., submitted to a private group called the Beale Cyphers Study Committee on 15 April 1972. Nelson read the Lynchburg press issue by issue, "beginning with the 26 December 1816 issue and ending with the 5 April 1822 issue", which covers the years the pamphlet places Beale in Virginia.
His finding on Morriss is emphatic. The innkeeper's identity, he writes, "is fully established by many references in the Lynchburg (Va) Press to Robert Morriss' business, social, and community activities." Morriss was a real, traceable, thoroughly documented Lynchburg man.
His finding on Beale is the opposite: "no records which confirm a Thomas Jefferson Beale or a Thomas J. Beale relationship to the Beale Party alleged activity have been found."
And one detail cuts closer still. The pamphlet has Morriss recalling, "It was in the month of January, 1820, while keeping the Washington Hotel, that I first saw and became acquainted with Beale." Against that entry in his own chronology Nelson notes, in brackets, that "no other reference has been found to a Washington Hotel or to Robert Morriss' association with it."
There is also a word worth pausing on. Beale's letter, dated 4 January 1822, describes the party following a buffalo herd, "securing many and stampeding the rest." The Online Etymology Dictionary dates the noun "stampede" to 1836, and the transitive verb, meaning to cause a panicked flight, to 1838. First-attestation dates are ceilings on surviving written evidence rather than proof of when a word was first spoken, so this settles nothing on its own. It is one more thing that reads more comfortably as 1885 than as 1822.
What the code-breakers concluded about the Beale ciphers
The ciphers have had serious professional attention, and the honest summary is that it has not resolved anything.
The cryptologist William F. Friedman was still corresponding about them in 1949. In a letter of 15 August that year, addressed to a correspondent at Madison Heights in Virginia, he explained that he had been "much engrossed with other matters" and had "found it necessary to put off dealing with The Beale Papers", adding: "However, I do not intend by any means to drop the study which I initiated on them sometime ago." What the letter establishes is modest and worth having: as of August 1949 the papers were unsolved, had been formally studied, had been set aside, and were still an open question for the man writing.
The most substantial technical study in the released file is by Carl Hammer, then director of computer sciences at UNIVAC, presented as an invited paper to the Third Annual Simulation Symposium in Tampa on 14 January 1970. Hammer ran the three ciphers against computer-generated random sequences and reached a conclusion that cuts against the simplest hoax theory: "none of the Beale Cyphers was constructed with the help of early random number tables, or by tossing coins or rolling dice."
His counts for two of the three match the pamphlet exactly: 520 entries in No. 1 running from 1 to 2906, and 618 in No. 3 running from 1 to 975. For No. 2 he records 763 entries running from 1 to 994, where the pamphlet as printed gives 762 running from 1 to 1005. The gap is small, and it is a useful reminder that even the ciphertext has drifted between printings, so a figure quoted for these ciphers is worth tracing to the version it came from.
Hammer's result is worth stating carefully, because it is often overstated. It says the numbers are not noise. Someone sat down and produced them by working along a text. It does not say that text described a treasure, or that the treasure existed, or that Thomas J. Beale did. Other statistical work on the pamphlet has reached the opposite conclusion, and the field has never converged.
So is the Beale treasure real?
Nothing in the surviving record establishes that it is, and nothing establishes that it is not.
What can be said is narrower and firmer. A pamphlet was published in 1885. It contained three sets of numbers and a solution to one of them that is close enough to its printed key to be no accident, and rough enough to show the key as printed is not the whole method. The innkeeper in the story was real. The man who supposedly buried the gold has left no trace in the record where he ought to appear. And no key, no box, no original paper and no treasure has ever been produced.
That is an unsatisfying answer, and it is the one the evidence supports.
Frequently asked questions
Who wrote The Beale Papers?
Nobody knows. The narrative is unsigned, and its narrator states plainly that he decided "upon withdrawing my name from the publication". The only name printed anywhere in the pamphlet is J. B. Ward's, and it appears on the copyright entry rather than as an author credit. Whether the narrator and the copyright registrant are the same person is not established by the pamphlet itself, and no correspondence, manuscript or contemporary account has surfaced to settle it.
Can I solve the Beale ciphers with a copy of the Declaration of Independence?
Not with a modern accurate one, and not cleanly with the pamphlet's version either. The Declaration printed in the pamphlet differs from the National Archives transcript in 28 places. Six of those change the word count, so the two texts drift in and out of step rather than sitting at a fixed offset, and five more change the first letter of a word, which is the only part of it the cipher reads. On top of that, no word in the pamphlet's Declaration starts with x, y or z, while the plaintext of cipher No. 2 needs x and y many times over. Whatever produced the published solution, it was not the published rule applied consistently.
What did the National Security Agency conclude about the Beale ciphers?
It did not publish a conclusion. What was released in December 2001 under the Freedom of Information Act is a collection of material the agency had gathered: outside research reports, magazine clippings, an association's membership form, and a technical paper given at a simulation conference. Treating that file as an official finding would misread it. It is a folder of other people's work, kept because the puzzle kept arriving at the agency's door.
How much treasure does cipher No. 2 describe?
By the pamphlet's own reading of it, two deposits. The first, dated November 1819, held one thousand and fourteen pounds of gold and three thousand eight hundred and twelve pounds of silver. The second, dated December 1821, held nineteen hundred and seven pounds of gold and twelve hundred and eighty-eight pounds of silver, plus jewels bought in St. Louis and valued at thirteen thousand dollars. The narrator puts the lot at "more than three-quarters of a million" in the money of his day.
Why has nobody cracked ciphers No. 1 and No. 3?
Because there is no rule to inherit. A book cipher is only as approachable as its keytext, and the one worked example on offer does not behave consistently enough to generalise from. Cipher No. 1 rules the Declaration out outright, since ten of its numbers exceed the length of that document, the largest being 2906. Cipher No. 3 stays inside the range, but that only means the Declaration has not been eliminated on size, which is a long way from a lead. Tools that classify an unknown message, such as a cipher identifier, can narrow what family a text belongs to; none of them can hand you a document nobody has found.
Sources
- The Beale Papers (1885), Wikisource's transcription of the Lynchburg pamphlet in full: title page, copyright entry, both Beale letters, the pamphlet's printed Declaration of Independence, all three ciphers and the stated solution to No. 2. Every quotation from the pamphlet above, and every count, comes from this text. Because it is a transcription rather than a scan, its reading "inalienable" was checked against a facsimile of the pamphlet inside the released file listed below before being relied on. Retrieved 10 August 2026.
- Declaration of Independence: A Transcription, U.S. National Archives, used for every comparison against the pamphlet's version and for the 1,322-word count. Retrieved 10 August 2026.
- Beale Papers, the National Security Agency's file released on 26 December 2001 under Freedom of Information Act case 41052. It contains Carl W. Nelson Jr.'s report to the Beale Cyphers Study Committee of 15 April 1972, quoted above on the Lynchburg press and the Washington Hotel, and Carl Hammer's paper "Signature Simulation and Certain Cryptographic Codes", Third Annual Simulation Symposium, Tampa, 14 January 1970. Retrieved 10 August 2026.
- Study of the Beale Papers, 15 August 1949, William F. Friedman's correspondence, from the Friedman collection released by the National Security Agency. Retrieved 10 August 2026.
- stampede, Online Etymology Dictionary, for the dating of the noun and of the transitive verb. Retrieved 10 August 2026.
Want to try the underlying trick for yourself? Start with the A1Z26 cipher decoder, which is the same idea at its smallest scale: numbers standing in for letters, with a key short enough to hold in your head.
