The Zodiac · an archive

Named suspects · none charged

Suspects

Fifty-six years of accusation has produced one man who was investigated seriously and a long tail of names attached by books, documentaries and press releases.


No one has ever been charged with a Zodiac murder. That is worth stating before any name, because a suspect list reads like a shortlist and this one is not.

What follows separates the evidence from the argument. Where a claim rests on a single witness or a single researcher, that person is named.

Arthur Leigh Allen

Born 18 December 1933, died 26 August 1992. The only publicly named suspect who was investigated at length, searched three times, and never arrested for the Zodiac crimes. He is the suspect in Robert Graysmith's book and the 2007 film, and most public belief about the case traces to that portrayal.

What put him under suspicion

  • Don Cheney's account. In July 1971 Cheney told police that Allen had once described hunting people, mounting a flashlight on a gun, shooting children coming off a school bus, and signing letters “Zodiac” — supposedly before the crimes.
  • The 1991 search of his Vallejo home produced a Zodiac-brand watch with a crosshair logo, a typewriter, pipe bombs, weapons and ammunition.
  • Michael Mageau's identification. In 1991 the Blue Rock Springs survivor picked Allen from a 1968 photograph: “That's him.”
  • The letter timeline. Authenticated letters stop in March 1971, resume in January 1974, and end around Allen's 1974 arrest and subsequent commitment.

What excludes him

Physical evidence

Fingerprints: no match to the latents from Paul Stine's cab.

Handwriting: excluded by state examiner Sherwood Morrill, and again by Lloyd Cunningham, who said in 2009 that none of Allen's writing came close.

DNA: the partial profile developed in 2002 from a letter stamp excluded him.

Voice: the Vallejo dispatcher who took the killer's call said Allen was not the caller.

Where it is genuinely contested

Two of those exclusions are weaker than they sound, and honesty requires saying so.

The DNA sample was reportedly taken from the outside of a stamp rather than from under it, which means the profile may belong to a postal worker rather than to the killer. It excludes Allen from that sample; it does not prove the sample is Zodiac's. Contested

Cheney's account has also shifted. The date moved from January 1968 to January 1969 over the years, and the Zodiac-watch detail appears only after Graysmith's 1986 book. Michael Butterfield treats the account as contaminated by retelling; Tom Voigt and Graysmith treat it as the cornerstone of the case. There is a logical problem either way: an early-1968 conversation in which Allen used the name “Zodiac” is more damning, and also impossible, because the name did not exist publicly until August 1969. Contested

Mageau's 1991 identification came twenty-two years after he was shot, from a photograph, and does not match the description he gave in 1969. The detective who worked the case doubted he ever got a clear look at the shooter.

Where that leaves him

Allen is the strongest circumstantial suspect and is excluded by every piece of physical evidence that has been tested. Both statements are true at once, and anyone who tells you only one of them is arguing rather than reporting.


Gary Francis Poste

Named in October 2021 by a private group calling itself the Case Breakers, who cited forehead-scar comparison and a claimed cipher decode. The FBI responded that the case remains open and the claim did not change its status. Cryptanalysts who examined the claimed decode found it did not follow from the ciphertext.

The claim received enormous coverage and has no evidentiary support. Not supported

Others

Advanced by particular researchers or authors
NameAdvanced byStatus
Lawrence Kane Researchers working the Lake Tahoe disappearance of Donna Lass No physical evidence.
Ross Sullivan Riverside-focused researchers, via the Cheri Jo Bates case The Bates link itself was retired in 2021.
Richard Gaikowski Blaine Blaine and later documentary treatments Rests largely on a voice comparison.
Earl Van Best Jr His son, Gary L. Stewart, in a 2014 book Handwriting claims widely rejected by examiners.
Paul Doerr A 2022 book by Jarett Kobek Circumstantial; no forensic link.
Ted Kaczynski Internet theorising, amplified by documentaries Contradicted by his documented whereabouts.

What the ciphers add

This is the one place where the cryptanalysis on this site speaks directly to the suspect question, and its answer is uncomfortable for everyone.

Tested

Every name above was tested against Z13's four equality constraints in multiple forms — full name, first and last, with and without a middle initial, common nicknames. Not one fits. Each fails either on length or on the pattern the ciphertext forces.

That is much weaker evidence than it first appears, and the weakness runs in both directions. Z13 is thirteen characters long, which is far below the length at which a cipher has a unique solution — so failing to fit does not clear anyone, and fitting would not convict anyone either. There are 49 real first-and-last names that fit, and if the Zodiac is allowed a single encoding error the set expands to 3,668,004.

The more interesting result is structural. Z13's repeat pattern disfavours male names at p = 0.0001. If he really did encode his own name, he chose a letter pattern that only 8.5 per cent of male-named candidates can produce. The likeliest reading is that Z13 does not contain a name at all — which would mean every attempt to extract a suspect from it, including the famous ones, has been mining an empty seam.