The Zodiac Killer murders
Northern California, 1968–1969
This Brief is an AI-generated synthesis of the public record. It may contain errors, omissions, or out-of-date information, and is not legal advice or original reporting. Verify against the primary sources before relying on it.
THE BRIEF: The Zodiac Killer Murders
SECTION 1 — VERDICT
The established record is that between December 1968 and October 1969 an unidentified killer who called himself “Zodiac” murdered at least five people in four Northern California attacks and wounded two others, then sustained years of taunting letters, ciphers, and threats to newspapers and police. He claimed 37 murders; investigators worked from a total of seven victims, five of them homicides. No one has ever been charged. The first cipher was solved within a week; the 340 cipher remained unread for 51 years before three private codebreakers solved it in 2020, and the FBI confirmed the solution. The case remains open.
The record shows no federal investigative case of its own: the FBI supported local agencies but never opened an investigation because the murders did not fall under federal jurisdiction. At the Stine murder scene, a dispatcher error told officers the suspect was Black, and officers did not stop a man matching the witness description. The SFPD marked the case inactive in 2004 before reopening it, and the record shows a partial DNA profile was collected from the outside of a stamp rather than from behind it or the envelope seal, while earlier DNA work on a glove and a car door was unproductive. Numerous offers, mostly from media outlets, were made to pay for a complete SFPD DNA analysis; further SFPD DNA analysis was described as “a tightly held secret.” A wet palm print from the Napa payphone was never matched. These facts do not establish that better coordination or earlier forensic work would have produced an identification, but they cluster in one direction: the official record contains avoidable failures at exactly the points where the case was most likely to be broken. This reading cannot be proven from available public evidence. It also cannot be dismissed.
What the evidence cannot establish is whether any specific failure was decisive, whether the agencies’ fragmentation was the primary reason the killer was not caught, or whether any named proposed suspect was the Zodiac. It also does not establish that the Cheri Jo Bates killing or the later disappearance of Donna Lass were Zodiac crimes, and it identifies no living person as the perpetrator.
SECTION 2 — CASE SUMMARY
The Zodiac case began on December 20, 1968, when David Arthur Faraday, 17, and Betty Lou Jensen, 16, were shot and killed on Lake Herman Road in Benicia, Solano County. On July 4 or 5, 1969, Michael Renault Mageau, 19, and Darlene Elizabeth Ferrin, 22, were shot at Blue Rock Springs Park in Vallejo; Mageau survived and Ferrin died. On September 27, 1969, Bryan Calvin Hartnell, 20, and Cecelia Ann Shepard, 22, were attacked at Lake Berryessa in Napa County; Shepard died two days later and Hartnell survived. On October 11, 1969, taxi driver Paul Lee Stine, 29, was shot in the head at Washington and Cherry in San Francisco’s Presidio Heights, and his wallet and car keys were taken.
The killer mailed letters and ciphers to newspapers including the San Francisco Chronicle, taunting police and claiming far more victims than were ever linked to him. The first cipher, sent in the July 31, 1969 letter, was solved by Donald and Bettye Harden within days. The 340 cipher, sent in November 1969 with pieces of Stine’s bloody shirt, was not solved until 2020. A final confirmed Zodiac letter arrived at the Chronicle in January 1974 and signed off “Zodiac — 37 SFPD — 0.” As far as investigators can determine, Stine was the last confirmed Zodiac killing.
No one has ever been charged. The only suspect police ever publicly named was Arthur Leigh Allen, who was never arrested for the Zodiac crimes; he died in 1992 at age 58. Forensic comparisons excluded Allen as the source of tested DNA and fingerprints, and a handwriting examiner said none of his writing came close, though the 2002 DNA sample’s collection location was later questioned, and the lead SFPD detective, Dave Toschi, said in 2010 that all the evidence against Allen ultimately “turned out to be negative.” Many other proposed suspects have been advanced by authors, researchers, and advocacy groups; no authority has adopted any of them as a charge or conviction.
The investigation’s institutional shape is a central part of the record. The San Francisco Police Department had primary jurisdiction over the Stine murder, while Vallejo, Napa County, Solano County, and the California Department of Justice also carried parts of the case. The FBI never opened its own investigation because the killings did not fall under federal jurisdiction; it provided handwriting, cryptanalysis, fingerprint, and latent-print support. The SFPD marked the case inactive in 2004, then reopened it, and all the relevant files remain open.
SECTION 3 — FULL RECORD
Evidentiary posture
The available record is a mixture of contemporaneous law-enforcement statements, official case-status disclosures, a city-commissioned internal review, and a large volume of suspect-identification claims from authors, researchers, and advocacy groups. The physical evidence is old, fragmentary, and in places of uncertain provenance. Key items — letters, envelopes, a bloodied glove, a car door, a payphone — were handled by multiple agencies over decades, and the record itself notes that DNA recovered in the case may not necessarily be the Zodiac’s because postal workers, newspaper employees, or police officers could have contaminated it. The absence of a single federal investigative file means there is no unified master record described in the available material.
Observed facts vs inferred claims
Observed facts include the four attacks and their dates, the letters and ciphers, the dispatcher error at the Stine scene, the FBI’s support-only role, the SFPD’s 2004 inactivation and later reopening, and the negative forensic results obtained for named suspects.
Inferences include the claim that multi-jurisdictional fragmentation and evidentiary failures materially compromised the investigation, and the competing claim that the killer simply left no traceable signature within the forensic capacity of the era. Both are readings of the same documented facts, not established conclusions.
Figure inventory
| Person | Connection to the case | Outcome or status in the record |
|---|---|---|
| The Zodiac | Unidentified killer | No arrest, charge, or confirmed identification |
| David Arthur Faraday | Victim, Lake Herman Road | Deceased, Dec 20, 1968, gunshot |
| Betty Lou Jensen | Victim, Lake Herman Road | Deceased, Dec 20, 1968, gunshot |
| Michael Renault Mageau | Survivor, Blue Rock Springs | Survived the 1969 attack; he identified Arthur Leigh Allen in a 1991 photo lineup, an identification that no agency adopted |
| Darlene Elizabeth Ferrin | Victim, Blue Rock Springs | Deceased, July 4/5, 1969, gunshot |
| Bryan Calvin Hartnell | Survivor, Lake Berryessa | Survived the 1969 attack |
| Cecelia Ann Shepard | Victim, Lake Berryessa | Deceased two days after the Sept 27, 1969 attack |
| Paul Lee Stine | Victim, Presidio Heights | Deceased, Oct 11, 1969, gunshot to the head |
| Cheri Jo Bates | Riverside 1966 homicide, disputed Zodiac link | Deceased, Oct 30, 1966, multiple stab wounds |
| Donna Lass | Missing person, possible but not definitively linked Zodiac victim | Deceased; skull found 1986, identified by DNA in 2023 |
| Dave Toschi | SFPD lead Zodiac detective | Deceased at age 86 |
| Arthur Leigh Allen | Vallejo suspect; only publicly named Zodiac suspect | Deceased 1992, age 58; never arrested for the Zodiac crimes |
| Gary Francis Poste | Suspect named by the Case Breakers | Deceased 2018 |
| Lawrence Kane | Proposed suspect | Deceased 2010 |
| Marvin Margolis / Marvin Merrill | Proposed suspect named by researcher Alex Baber/Barber | Deceased in California in 1993 |
| Joseph James DeAngelo Jr. | Golden State Killer case; cited as catalyst for Zodiac DNA technique | Charged with eight of the 12 Golden State Killer homicides; the record does not state a conviction, acquittal, or current procedural status |
| Don Cheney | Allen’s former friend; first to suggest Allen might be the Zodiac | Subjected to DNA comparison and excluded as a contributor |
| Donald Harden | Solved first Zodiac cipher with wife Bettye | No death reported in record |
| Bettye Harden | Solved first Zodiac cipher with husband Donald | No death reported in record |
| David Oranchak, Sam Blake, Jarl Van Eycke | Solved the 340 cipher in 2020 | No deaths reported in record |
| Robert Graysmith | Author of “Zodiac”; advanced the Allen theory | No death reported in record |
| Tom Voigt | Runs Zodiackiller.com | No death reported in record |
| Terry Poyser | Vallejo PD detective who discussed DNA testing; named in the 2020 badge-bending report | Retired; the 2020 Giordano report found no evidence badge bending marked a line-of-duty killing and insufficient evidence of a department custom or “Badge of Honor”; no disciplinary outcome is stated in the record |
| Kent Tribble | 2020 Giordano report found by a preponderance of the evidence that he bent badge tips of Matt Komoda and David McLaughlin in 2016 and his own in 2003 | The same report found no evidence badge bending marked a line-of-duty killing and insufficient evidence of a department custom or “Badge of Honor”; no disciplinary outcome is stated in the record |
| Zach Jacobsen | 2020 Giordano report found by a preponderance of the evidence that he bent the badge tip of Ryan McMahon in 2018 | The same report found no evidence badge bending marked a line-of-duty killing and insufficient evidence of a department custom or “Badge of Honor”; no disciplinary outcome is stated in the record |
| Fayçal Ziraoui | French engineer who claimed to have cracked the remaining unsolved Zodiac ciphers and named a person who is not identified here | No authority adopted the claim; the named person is not repeated here |
| Gary Stewart | Author of a 2014 book that claimed his biological father was the Zodiac | The FX documentary placed the author’s father in Europe during the Zodiac activities; the director said he thought the father was not the Zodiac |
| Dennis Kaufman | Claimed in 2007 that his stepfather was the Zodiac and turned several items over to the FBI, including a hood | No authority adopted the claim |
| Alex Baber/Barber | Researcher who claimed to have broken the Z13 cipher using an AI program and identified the key “Elizabeth” as pointing to Marvin Merrill/Marvin Margolis; also claimed links to the Black Dahlia killing | No authority adopted the claim; Merrill’s son denied the theory |
| Kathleen Johns | Identified Lawrence Kane in a photo lineup as the man who abducted her in 1970 | No authority adopted Kane as the Zodiac |
| John Lynch | Vallejo Police detective who interviewed Arthur Leigh Allen on Oct. 6, 1969 | — |
| Richard Grinell | Researcher who said in 2022 that “there are probably 50 or 100 suspects named every year” | — |
| Ken Mains | Cold case detective who said he believed the Zodiac Killer was potentially one of two men, including Lawrence Kane | No authority adopted the view |
| Dale Julin | TV news anchor who filed court affidavits saying he interviewed Poste in 2017 and Poste admitted to being the Zodiac | No authority adopted the claim |
| Richard Nichelman | Vallejo Police lieutenant who discussed DNA retesting | No death reported in record |
| Ken Narlow | Napa County detective on the case from the outset | No death reported in record |
| Bill Armstrong | SFPD detective assigned to the case | No death reported in record |
| Lloyd Cunningham | Retired police handwriting expert | No death reported in record |
| Nancy Slover | Vallejo police dispatcher contacted by the Zodiac | No death reported in record |
| Don Fouke | SFPD officer near the Stine scene | No death reported in record |
| Ryan J. Railsback | Riverside Police public information officer | No death reported in record |
| Sheriff Bill Pooley | Tuolumne County sheriff skeptical of the Case Breakers | No death reported in record |
| Robert Giordano | Author of the 2020 Vallejo badge-bending report | No death reported in record |
Source weighting
The most reliable material in this record is official institutional action and finding: case-status statements from the SFPD, Vallejo PD, the FBI, the California Department of Justice, and Riverside Police; forensic outcomes such as exclusions and non-matches; and the city-commissioned Giordano report. Next are contemporaneous investigative facts reported consistently across the record, such as the dispatcher error and the negative forensic results. Suspect-identification claims from authors, researchers, volunteer groups, documentaries, and private individuals carry less weight, and the record consistently shows no official adoption of any of them. The case’s offered DNA comparisons are official investigative acts, but their weight is limited by the record’s own description of potential contamination and by the later disclosure that the 2002 sample came from outside the stamp.
Anomalies
HIGH — The Stine-scene dispatcher error. Three teenage witnesses described a husky white man in a dark or black jacket wiping down the cab. A dispatcher error led police to believe the suspect was Black, and officers who spotted a man matching the witness description did not stop him. The Zodiac later claimed in letters that he was the witness who spoke to two police officers at the scene.
HIGH — Initial misclassification of the Stine murder. Police initially assumed the Stine murder was a robbery and not necessarily related to the Zodiac.
HIGH — No federal investigation. The FBI never opened its own investigation because the murders did not fall under federal jurisdiction. Its role was support only. This left the primary investigative burden split among city and county agencies with no single lead agency described in the record.
HIGH — DNA evidence collection problem. A partial DNA profile developed in the early 2000s ruled out Arthur Leigh Allen in 2002, but it was later reported that the sample was collected from outside the stamp rather than from behind it or from the envelope seal, meaning Allen could still be a suspect as of March 2018. The record therefore contains both an exclusion and a later qualification of that exclusion, without a final resolution stated.
MODERATE — The 2004 inactivation. The SFPD marked the case inactive in 2004, then reopened it before March 2007. The available record does not explain why the case was inactivated or what changed.
MODERATE — Offers to pay for DNA analysis and the “tightly held secret.” Numerous offers, mostly from media outlets, to pay for a complete SFPD DNA analysis were made, and further SFPD DNA analysis was described as “a tightly held secret.” The record does not report that the offers were accepted.
MODERATE — Lead detective’s misconduct. In 1978, lead SFPD detective Dave Toschi made three fake fan-mail letters praising himself and sent them to columnist Armistead Maupin, disrupting the Zodiac investigation. He admitted using fake names and was reassigned from homicide to the pawn shop detail.
MODERATE — The unmatched payphone palm print. A wet palm print was lifted from the Napa payphone used to report the Lake Berryessa crime. Detectives were never able to match it to any suspect.
LOW — Possible contamination. The record notes that DNA recovered in the case may not necessarily be the Zodiac’s because postal workers, newspaper employees, or police officers could have contaminated it.
Motive and mechanism
The motive question belongs to the killer, and the record does not establish a motive for the Zodiac murders. The mechanism question in this case has two parts. The first is whether jurisdictional fragmentation and evidentiary failures left critical leads unpursued; that is supported by the dispatcher error, the absence of a federal investigation, the hiatus around 2004, the DNA collection problem, and the failure to resolve the payphone print. The second is whether the killer left no traceable signature within the forensic capacity of the era; that is supported by the negative results on fingerprints, handwriting, and DNA, and by the fact that no authority has ever charged anyone. The record does not permit the two mechanisms to be cleanly separated: a killer who leaves no traceable signature and an investigation that mishandles what trace exists can both be true.
Competing theories
| Theory | What it claims | Confidence and status |
|---|---|---|
| Arthur Leigh Allen as Zodiac | Promoted by Robert Graysmith and the 2007 film “Zodiac”; Mike Mageau identified Allen in a 1991 photo lineup | Rejected by agencies; never arrested for the Zodiac crimes; DNA excluded him as a contributor in 2002, subject to later sample-location caveat; fingerprints did not match Stine prints; handwriting expert said none of Allen’s writing “even came close”; Dave Toschi said all evidence ultimately “turned out to be negative” |
| Gary Francis Poste as Zodiac | Advanced by the Case Breakers, including forehead scars, a heelprint, a paint-splattered watch, darkroom photos, anagrams, and a claimed DNA “strong match” | Rejected or not adopted by officials; SFPD said identification must come down to hard evidence like handwriting and fingerprints, which the analysis lacks; FBI and SFPD declined to confirm any suspect; TV news anchor Dale Julin filed court affidavits saying he interviewed Poste in 2017 and Poste admitted to being the Zodiac; no authority adopted the claim |
| Lawrence Kane as Zodiac | Identified by Kathleen Johns in a photo lineup; Darlene Ferrin’s sister identified a photo of Kane; Officer Don Fouke said Kane resembled the man observed near the Stine scene | Fingerprints did not match those from Stine’s cab; no authority adopted him as the Zodiac |
| Marvin Merrill / Marvin Margolis as Zodiac | Researcher Alex Baber/Barber claimed an AI break of the Z13 cipher pointed to Merrill/Margolis | No authority adopted the claim; Merrill’s son denied the theory |
| Fayçal Ziraoui cipher claim | French engineer claimed to have cracked remaining unsolved Zodiac ciphers and named a person who is not identified here | Many Zodiac sleuths disputed the solutions; FBI and SFPD declined to comment; the named person is not repeated in this record |
| Book and documentary parentage claim | A 2014 book and an FX documentary claimed the author’s biological father was the Zodiac | The documentary placed the author’s father in Europe during the Zodiac activities; the director said he thought the father was not the Zodiac |
| Dennis Kaufman claim | Claimed his stepfather was the Zodiac and turned items over to the FBI, including a hood | No authority adopted the claim |
| Cold case detective’s serial-killer link | A former cold case detective linked another serial killer to the Zodiac murders and other unsolved cases | No authority adopted the link; the named person is not treated as an official suspect in the record |
| Voice identification | Former dispatcher Nancy Slover identified a recording as the same voice as the Zodiac’s | No authority adopted the identification |
THE STRONG CIRCUMSTANTIAL READING: A COMPROMISED INVESTIGATION
The reading is that the Zodiac investigation was not simply defeated by an absent forensic signature; it was materially weakened by a fragmented jurisdictional structure and by identifiable failures in the official response. The record supports this reading with specific indicators.
The first indicator is structural: no single agency owned the case. The FBI never opened its own investigation because the murders did not fall under federal jurisdiction, leaving support functions only. City and county agencies — San Francisco, Vallejo, Napa County, Solano County, and the California Department of Justice — each held parts of the case, with no unified investigative file described in the record.
The second indicator is the Stine-scene error. At the one moment when the killer was seen and a suspect description was available, a dispatcher error led police to look for a Black suspect even though three teenage witnesses had described a husky white man. Officers who spotted a man matching the witness description did not stop him. The Zodiac then claimed in letters that he had been the witness who spoke to two officers at the scene. Police initially assumed the Stine murder was a robbery and not necessarily related to the Zodiac.
The third indicator is the forensic record’s disarray. The 2002 partial DNA profile excluded Allen, but it was later reported that the sample came from outside the stamp rather than from behind it or the envelope seal, reopening the very question the exclusion had appeared to close. Earlier DNA work on the Stine crime-scene bloody glove and the Lake Berryessa car door was unproductive. A wet palm print from the Napa payphone was never matched to any suspect. The record itself allows that recovered DNA may not be the killer’s because of possible contamination.
The fourth indicator is institutional behavior after the active period. The SFPD marked the case inactive in 2004 before reopening it. Offers to pay for full DNA analysis were made but the record does not show they were accepted, and further SFPD DNA analysis was described as “a tightly held secret.” Vallejo PD separately sent two Zodiac letters and envelopes to a private lab for DNA testing.
Tom Voigt, who runs Zodiackiller.com, described the Zodiac case as having been investigated part time by a police department in a city that filed for municipal bankruptcy, and told the San Francisco Chronicle: “There is no reason in the world why they shouldn’t make another run at all the DNA in this case, especially considering what has happened in the Golden State Killer case. Just send it all to a lab.”
In 1978, the lead SFPD detective, Dave Toschi, made three fake fan-mail letters praising himself and sent them to columnist Armistead Maupin, disrupting the Zodiac investigation. He admitted using fake names and was reassigned from homicide to the pawn shop detail.
A separate city-commissioned review adds institutional context for Vallejo. The 2020 Giordano report found by a preponderance of the evidence that three named Vallejo officers had bent badge tips: Kent Tribble on the badges of Matt Komoda and David McLaughlin in 2016 and his own in 2003; Terry Poyser on the badges of Jason Bahou in 2015 and Zach Jacobsen in 2017; and Zach Jacobsen on the badge of Ryan McMahon in 2018. The same report found no evidence that badge bending marked a killing in the line of duty and insufficient evidence that it was a department custom or “Badge of Honor”; the record does not state any disciplinary outcome. That report does not establish that the Zodiac investigation was mishandled, but it is part of the documented institutional environment in which Vallejo’s handling of the case must be assessed.
What is missing is direct proof of consequence. No specific failure has been shown to have caused the killer not to be identified. The counter-explanation — that the killer left no usable signature — is supported by the negative results themselves. The strength of the compromised-investigation reading is that so many failures occurred at precisely the points where the case was most breakable. This reading cannot be proven from available public evidence. It also cannot be dismissed.
What the evidence best supports
The evidence best supports that the Zodiac case remains unsolved because the killer left little usable evidence and the official response failed to exploit what it had. Neither explanation is sufficient alone. The record documents a series of operational and forensic failures — the dispatcher error, the absence of a federal investigation, the initial robbery misclassification, the 2004 inactivation, the uncertain DNA collection, the unmatched palm print, the stalled offers of paid analysis, and the lead detective’s own documented disruption — that make it impossible to attribute non-identification solely to the killer’s lack of a traceable signature. At the same time, the negative fingerprint, handwriting, and DNA results show that no proposed suspect has been connected to the crimes by the available forensic evidence.
SECTION 4 — WHAT REMAINS UNKNOWN
It remains unknown who the Zodiac was. No one has ever been charged, no authority has confirmed an identification, and the available physical evidence has not matched any named suspect. It is unknown whether the dispatcher error or any other specific failure was decisive in allowing the killer to escape, and unknown whether a unified federal investigation or earlier comprehensive DNA work would have produced an identification. It is unknown whether the 2002 DNA sample actually represents the killer, because the record says it was taken from outside the stamp and allows for contamination. It is unknown whether the Cheri Jo Bates homicide or the Donna Lass disappearance were connected to the Zodiac, and unknown what motive, if any, drove the killings. The record also does not state the present procedural status of every person named in connection with the broader genetic-genealogy developments.
SECTION 5 — METHODOLOGICAL NOTE
The central difficulty of this case is that the killer left an unusually large archive of his own words while leaving no physical evidence that modern forensic methods have been able to attach conclusively to a named person. That mismatch produces a record in which the killer is constantly speaking and yet remains unidentified, and in which every proposed suspect ultimately fails the available forensic tests. The unresolved question is therefore not only who he was, but whether the failure to identify him reflects the limits of the evidence or the limits of the official response.