The Somerton Man: Australia's Tamam Shud Cipher That's Never Been Solved

On the morning of December 1, 1948, a well-dressed man was found slumped against a seawall on Somerton Beach in Adelaide, South Australia. He had no identification. The labels had been cut from his clothes. In a hidden fob pocket of his trousers was a small rolled piece of paper printed with two words: "Tamam Shud." Those words, Persian for "it is finished," were torn from the last page of a copy of the Rubaiyat of Omar Khayyam.

The case became one of Australia's most enduring mysteries. A code was found. A phone number led to a nurse who said she did not know the man. The body was buried without a name. Nearly 78 years later, the code is still unread, and even the man's identity, seemingly settled by DNA in 2022, is now disputed.

This post lays out what is actually known as of August 2026, what the code looks like, why every attempt to decode it has failed, and where the DNA identification stands. You can run your own frequency analysis on the code letters with our letter frequency analyzer.

A passing cyclist found the man at around 6:30 a.m. on December 1, 1948. He was lying against the seawall at Somerton Beach, south of Adelaide, his head resting on the wall. Witnesses had seen a man in the same spot the previous evening. A conductor remembered a man who may have been him boarding the 10:50 a.m. train from Adelaide to Henley Beach on November 30, then alighting at Adelaide Railway Station and boarding a bus toward Somerton. None of this was ever confirmed.

The man was around 40 to 45 years old, in good physical condition, about 5 feet 11 inches tall. He wore a sharp suit, a tie, polished shoes. He carried no wallet, no ID, no train ticket. The labels had been cleanly cut from every garment, which is unusual and deliberate. A railway ticket from Adelaide to Henley Beach, a bus ticket, a comb, a packet of chewing gum, and a small packet of cigarettes were found, but nothing that named him.

The post-mortem found no visible wounds. The coroner, Thomas Cleland, could not establish a cause of death. The man's spleen was enlarged, his liver damaged, and his stomach was congested, but no common poison was identified with the tests available in 1948. Cleland recorded an open finding. The body was embalmed and buried in West Terrace Cemetery in Adelaide under a headstone reading only "Here lies the unknown man who was found at Somerton Beach." The case file stayed open.

The Wikipedia article on the Tamam Shud case tracks the public record of the investigation, and the South Australian Police maintained the file as an active missing-persons matter for decades.

The scrap of paper found in the fob pocket was a photograph of a printed line, "Tamam Shud," torn from the last page of a book. "Tamam" is a Persian and Urdu word meaning "finished" or "ended," and "Shud" is a grammatical marker. Together the phrase means "it is finished" or "it is ended." It is the final line of the Edward Fitzgerald translation of the Rubaiyat of Omar Khayyam.

Police appealed to the public for the book the scrap had been torn from. In early 1949, a man came forward to say he had found a copy of the Rubaiyat on the back seat of his unlocked car, parked near Somerton Beach, around the time of the death. The car had been unlocked, the book tossed onto the seat. The last page was torn out. It matched the scrap.

Inside the front cover of the book, faintly pencil-impressed, were two things. The first was a phone number, which traced to a woman in the Adelaide suburb of Moseley Street, a nurse later identified as Jessica Thomson, also known as Prosperine. She told police she did not know the dead man and did not know why her number was in the book. She gave her name as Thomson, which was not her birth name. She died in 2007 without ever publicly elaborating. Her daughter, Kate Thomson, later told reporters her mother had admitted she knew who the Somerton Man was but would not say.

The second thing impressed in the cover was a sequence of capital letters, crossed out and partly rewritten. This is the "Somerton Man code." It is the cryptographic heart of the case, and it has never been solved.

A detail that matters for any book-cipher theory: the Rubaiyat recovered from the car was long assumed to be a Whitcomb and Tombs edition, a New Zealand printer. Later research showed the physical dimensions did not match any known Whitcomb and Tombs printing. The book was a Collins dual-translation edition, containing both the first and second Fitzgerald translations. This was confirmed by measuring the page dimensions against surviving copies. Any interpretation that treats the code as a book cipher, indexing into the Rubaiyat's text, has to use the correct edition, and the Collins dual-translation text differs from the more common single-translation printings.

The code as transcribed from the book cover is a short string of capital letters, arranged in lines, with the second line crossed out and rewritten. Different transcriptions circulate, and the exact letters depend on who read the impressions and when. The commonly cited reading is roughly:

`` W B A B O A O I O B B P A B M W A O I O I A B B ``

The second line appears struck through and replaced, which suggests the writer corrected themselves, or that two attempts were made. The letters do not spell English words in any obvious reading order. There are no digits, no spaces marked, no punctuation.

Attempts to decode it have come from several directions and all have failed to produce a verified solution. The South Australian Police consulted military and university cryptographers in the 1940s and 1950s. Nothing came of it. In later decades, amateur and academic researchers tried treating it as a Vigenere cipher, a transposition cipher, an initialism, and a book cipher indexing into the Rubaiyat. None produced a plaintext that matched known facts about the case or that other researchers could independently reproduce.

The structural problem is severe. The code is short, roughly 20 to 25 letters depending on the transcription. A short ciphertext is underdetermined: many different keys and many different plaintexts can be made to fit, so any proposed solution is hard to verify unless it predicts something checkable. This is the same problem that dogs the short Zodiac ciphers, where a 13-character string can be forced to spell almost any name. Our cipher identifier can classify longer ciphertexts by type, but a string this short resists classification because it lacks enough statistical signal.

One persistent theory, advanced on the Tamamshud blog by Gordon Cramer, is that the code relates to Cold War tradecraft. Cramer argues the letters are consistent with methods taught in the 1943 Special Operations Executive manual Syllabus of Instruction: Codes and Ciphers, and that micro-writing and ink-erasure evidence on the book and other items points to intelligence activity. He connects the case to SOE operational practices and to the geopolitical context of 1948. Mainstream investigators have not adopted these findings, and the micro-writing claims in particular are disputed, but the tradecraft angle remains one of the more developed alternative theories.

Another possibility, taken seriously by sober researchers, is that the letters are not a cipher at all. They could be working notes, a mnemonic, a draft of something unrelated, or the idle scribbling of whoever owned the book. The fact that a line was crossed out and rewritten is consistent with someone thinking on paper, which is not how a finished ciphertext usually looks. If it is not a cipher, then the failure to decode it is not a failure of cryptanalysis. It is a category error.

For 74 years the man had no name. That changed in 2022, when Professor Derek Abbott of the University of Adelaide announced that DNA analysis had identified the Somerton Man as Carl "Charles" Webb, a Melbourne man born in 1905.

Abbott's team had been working the case for years. The breakthrough came after the body was exhumed in 2021 under South Australian Police's Operation Persevere. The team extracted DNA from hairs embedded in a plaster bust that had been cast of the man's face and torso in 1948, before burial. From those hairs they built a genetic profile and used commercial genealogy databases, primarily Ancestry.com, to trace family lines. The genealogical work pointed to Carl Webb, an electrical worker and salesman from Melbourne with documented links to a woman in Adelaide.

Webb had disappeared from the historical record in the late 1940s. He had a wife, Dorothy Robertson, known as Doff, who had herself gone missing around the same period. The proposed connection to Adelaide runs through Webb's brother-in-law and through the Thomson family, though the exact links are tangled and rely on genealogical inference rather than direct documentation.

The identification was widely reported as a solution. It is not yet official. As of July 2026, South Australian Police had not formally confirmed the identification through the coroner. ABC News reported on July 6, 2026 that a police report to the coroner was expected "very soon," and that the coroner was expected to present findings by the end of 2026. Until the coroner rules, Carl Webb is a strong lead, not a closed case.

The Somerton Man case is unsolved on two fronts, and they get conflated. The first is the identity of the man. The second is the code. Solving one does not solve the other.

On identity, the Carl Webb identification is the strongest lead in the history of the case, but it is contested. A 2025-2026 academic paper posted on Academia.edu, titled "Methodological Concerns in the Carl Webb Identification," raised critical objections to the DNA evidence and the research protocols behind it. The paper questions the chain of custody for the hair samples, the statistical strength of the genealogical match, and the handling of alternative candidates the database surfaced. The identification has not been independently replicated by a second laboratory, and the coroner has not ruled on it. Researchers who support the Webb identification argue the genealogical convergence is strong enough to stand. Researchers who doubt it argue the methodology has not met the bar for a formal identification, especially for a case this old and this contaminated by decades of handling.

On the code, the situation is worse. No proposed decryption has been verified. The ciphertext is too short to constrain solutions, the correct edition of the Rubaiyat was misidentified for decades, and there is no agreement on whether the letters are even a cipher. If the code is a one-time pad fragment, it is unbreakable by information-theoretic arguments, because a one-time pad's key is the same length as the message and is used once, leaving no statistical structure to exploit. If the letters are random notation, there is nothing to break. If they are a book cipher, the correct book and edition must be known, and even then a short book cipher is underdetermined.

You can test the difficulty yourself. Take the transcribed letters and run them through the Vigenere cracker with different candidate keys, or feed them to the letter frequency analyzer to see how little statistical signal a 20-letter string carries. The exercise shows why every claimed solution has failed peer review. There is simply not enough data.

The honest position, as of August 2026, is that the Somerton Man's identity is probably close to settled but awaits coroner confirmation, and the code is either unsolvable or not a cipher at all. Both could be true. The man may be Carl Webb, and the letters in the Rubaiyat may be nothing more than someone's private notes, misread for 78 years as a message they were never meant to be.