The Cleveland Torso Murders: twelve official victims, three names, no killer.
Between 1934 and 1938, the remains of at least a dozen people turned up in and around Cleveland's Kingsbury Run. Only three victims were ever identified. Public Safety Director Eliot Ness took on the case, and the killer was never caught.
Case file at a glance
This is the oldest case in our set that science is about to revisit. The suspect list is short and weak, and the real prize is identifying the nine victims who never got names, which could finally show who they were and where they came from.
- Most victims were poor and unnamed.Many were drifters living in the Flats shantytowns. Only Edward Andrassy, Florence Polillo, and, tentatively, Rose Wallace were identified.
- Eliot Ness could not stop it.Ness oversaw the hunt, ordered the shantytowns burned in 1938, and was taunted by the killer, who left remains in view of his office. The killings stopped without an arrest.
- The suspect list is thin.Frank Dolezal confessed, recanted, and was found hanged in custody. Dr. Francis Sweeney, who failed two polygraphs, was never charged.
- Science has returned.In 2024, medical examiners and the DNA Doe Project began exhuming unidentified victims to identify them through genetic genealogy.
Editorial assessment of public sources, not a finding of fact. Nobody named here has been charged unless this page says so.
What happened
Dates come from the Wikipedia summary of period records. Counts differ between sources.
The "Lady of the Lake"
A woman's remains are found on the Lake Erie shore. Some researchers count her as the first victim; the official count begins later.
Edward Andrassy
The remains of Edward Andrassy are identified by fingerprints. He had a 1931 Cleveland Police record. Another man, never identified, is found nearby.
Florence Polillo
Florence Polillo, victim No. 3, is identified by fingerprints.
More victims
Further remains are found in and around Kingsbury Run and the Flats. The killer usually decapitated victims and often dismembered them; most male victims were castrated. The press calls him the "Mad Butcher of Kingsbury Run."
Ness and the shantytowns
Ness orders the Flats shantytowns searched and burned. The raid finds nothing incriminating. The killer leaves remains in view of Ness's City Hall office and later sends him postcards.
Frank Dolezal
Frank Dolezal is arrested, confesses, and retracts. He is found hanged in his cell. He is later exonerated posthumously.
Willie Johnson executed
Willie Johnson, who had been considered a suspect, is executed for a different murder.
A possible later victim
Robert Robertson's remains are found; some researchers count him as a later victim.
Sweeney dies
Dr. Francis Sweeney, a World War I veteran and alcoholic whom Ness had interviewed, dies.
Exhumations begin
The Cuyahoga County Medical Examiner's Office and the DNA Doe Project begin exhuming unidentified victims to identify them using genetic genealogy.
Victims
The official count is twelve. Only three victims were identified; the others remain unnamed.
1931 Cleveland Police photo. Photo: gargantuen (Flickr), CC BY-SA 2.0, via Wikimedia Commons. Shown unmodified.
DiedEdward W. Andrassy
The first official victim identified, by fingerprints from a 1931 police record. His remains were found in 1935.
Cleveland Police photo, c. 1934; the only known photo of her. Photo: gargantuen (Flickr), CC BY-SA 2.0, via Wikimedia Commons. Shown unmodified.
DiedFlorence Polillo
Victim No. 3, identified by fingerprints. The only known photo of her is a police photo from about 1934.
A 1939 sketch based on a lost newspaper photo; sketch credited to Michael Nevin. Photo: gargantuen (Flickr), CC BY-SA 2.0, via Wikimedia Commons. Shown unmodified.
DiedRose Wallace (tentative)
Tentatively identified as the eighth victim. Only a 1939 sketch, made from a lost newspaper photo, is known.
Nine unidentified victims
Most official victims have never been named. Medical examiners and the DNA Doe Project are working to identify them through genetic genealogy.
The three photos shown are credited on each card and in the credits table. We show no photographs of remains or of crime scenes. Descendants or others who want a photo added, changed, or removed can use the contact form at the bottom of this page.
Evidence
Little evidence survives, and the key records are old police files.
| Item | What is known | Disputed or unknown |
|---|---|---|
| The remains | Found in Kingsbury Run, the Flats, and along the lake. Victims were usually decapitated and often dismembered; most male victims were castrated. | Cause of death and the killer's skill (some suggest medical training) are debated. |
| Identifications | Andrassy and Polillo by fingerprints; Rose Wallace tentatively. | Nine others remain unidentified. |
| Taunts | Remains left in view of Ness's office; postcards sent to him. | Authenticity of each is debated. |
| Confession | Frank Dolezal confessed in 1939. | He retracted it and died in custody; later exonerated. |
| Polygraphs | Dr. Francis Sweeney failed two. | Polygraph evidence is unreliable and he was never charged. |
Public documents
The surviving record is scattered across museums, libraries, and newspaper archives.
| Document | Holder | What it shows | Access |
|---|---|---|---|
| The Kingsbury Run Murders (Cleveland Police Museum, by Dr. James Badal) | Cleveland Police Museum | Police-based case details, including the fingerprint identification of victim No. 3. The page blocked our automated check. | Online, free |
| Crime Library: Kingsbury Run (excerpt of a September 1935 police report) | Reproduced by Crime Library | An excerpt describing the first bodies found. We did not open it. | Quoted in news |
| University of Akron review of Badal's "In the Wake of the Butcher" | University of Akron | Notes that the book drew on detailed police files not used before. | Online, free |
| Special collections: Cleveland Public Library, Cleveland Press archives (Cleveland State University), Western Reserve Historical Society | Library and university archives | Files, scrapbooks, and photographs. Whether items are digitized is not confirmed. | Request needed |
| Cuyahoga County Medical Examiner records (2024 exhumations) | Cuyahoga County | The identification effort. No public results found. | Not located |
Access labels: Online, free = we could open it. Fan-hosted copy = a private researcher posted scans; originals are held by the named agency, and we cannot vouch for completeness. Quoted in news = reporters read it; we did not see the document itself. Request needed = held by an agency or court. Withheld = the holder has not released it. We cannot reach non-public records, and we do not try to.
Online discussion (unverified)
Where people talk about this case online, and how to read it.
| Platform | Live search | What you tend to find |
|---|---|---|
| Google: Reddit | site:reddit.com search | Public Reddit threads that Google has indexed, including old discussions. |
| Google: X and Twitter | X and Twitter posts | Public posts and profiles that appear in search results. |
| Google: Facebook | Public Facebook pages and posts | Only public pages and groups that Google can see. Private groups never appear. |
| Google: Instagram and TikTok | Instagram and TikTok | Public posts, memorials, and videos. |
| Google: forums | Forums and message boards | Long-running true-crime forums. |
| Google News | Latest news | Newest reporting, which is usually more reliable than posts. |
| Search Reddit | Long discussion threads and theories, often by amateur researchers. No sign-in needed to read. | |
| X (Twitter) | Search X | Short posts, news links, and anniversary notes. May require sign-in. |
| Search Facebook | Public groups and family or community pages. Requires sign-in. | |
| Search #clevelandtorsomurders | Memorial and true-crime posts. Requires sign-in. | |
| YouTube | Search YouTube | Documentaries, news clips, and interviews. Quality varies widely. |
Be careful with online theories. Naming a private person as a suspect can harm an innocent family, and many "identifications" on forums have been wrong. Our own tools could not read these platforms: Reddit blocks automated access, and the others require an account. The Google searches above work from your browser. To get new public mentions by email, make a free Google Alert for the case name in quotation marks.
Coverage timeline and verification
What local and national reports say, and when we last checked for a solution.
| Date | Outlet | What it reported | Link |
|---|---|---|---|
| 2024 | |||
| CBS News | |||
| DNA testing under way to identify victims; two bodies exhumed. | |||
| Read | |||
| Undated | |||
| News 5 Cleveland | |||
| Dr. James Badal says he is "99.9 percent sure" he knows the killer's identity; no name or evidence given. | |||
| Read | |||
| June 2026 | |||
| Cleveland International Film Festival | |||
| A film about the case was set for release. | |||
| Read |
Status is re-checked on a regular schedule. If you know of a development we missed, use the clue form near the top of this page or the contact form, and include a link.
Suspects
The suspect list is short. Most early leads fell apart.
Frank Dolezal
Why named: Arrested in 1939 and confessed.
Against: He retracted the confession and was found hanged in his cell. He was exonerated after his death.
Dr. Francis E. Sweeney
Why named: A World War I veteran and alcoholic whom Eliot Ness interviewed. He failed two polygraphs.
Against: The evidence was circumstantial; he was never charged. He died in 1964.
Willie Johnson
Why named: Considered as a suspect.
Against: Never conclusively linked. Executed in 1944 for a different murder.
Where a card shows a silhouette, we found no freely licensed photo. Anyone who wants a photo added, changed, or removed can use the contact form at the bottom of this page.
Latest lead: the 2024 exhumations
There is no new suspect. The newest development is a plan to identify the victims, which may lead to the first new facts in decades.
What is happening
- In 2024, the Cuyahoga County Medical Examiner's Office and the DNA Doe Project began exhuming unidentified victims.
- The goal is to use genetic genealogy to give names to victims who have been unnamed for about 90 years.
- We found no published identifications from this project as of Oct 2026.
Why it matters
- Knowing who the victims were and where they lived could show how the killer found them.
- A pattern of origins, such as one neighborhood, could narrow the suspect pool.
- Descendants may be able to give family names back to the dead.
What would settle it
- Name the victims through genealogy and records.
- Compare their last known whereabouts and associates.
- Test any surviving evidence against known suspects' descendants where possible.
Names first, suspects later
We put the chance that the exhumations identify at least one more victim in the next five years at roughly 55%. We put the chance that the killer is ever named at roughly 5%. The most likely suspect named in public, Dr. Francis Sweeney, has a roughly 10% chance in our view, because the evidence is a polygraph and circumstance.
Editorial judgment, not statistics.
Elimination matrix
"—" means no public information, not a clean result.
| Person | Confession | Physical link | Medical access | Outcome |
|---|---|---|---|---|
| Frank Dolezal | ? retracted | — none published | — | ✗ exonerated posthumously |
| Francis Sweeney | ✗ none | — none published | ✓ physician | ? never charged; failed polygraphs |
| Willie Johnson | — | — | — | ✗ executed for a different crime |
✗ excluded or contradicted · ✓ supports · ? disputed · — nothing public. No cell is an official finding.
Our hypothesis
Editorial reading of public records. A working theory, not a finding, and it names no individual.
A killer who preyed on people no one would report missing, and a gap that identification could close
- He chose people who would not be missed. Most victims were drifters or poor, living in the Flats. That is why so few were identified.
- He had the means and a private place. Dismemberment and disposal took time, tools, and somewhere to work unobserved.
- He stopped, or was stopped, around 1938. The killings ended soon after the Flats shantytowns were cleared, possibly because his hunting ground disappeared.
- The next break is a name, not a clue. Identifying even a few victims may reveal where he found them.
Judgment, not statistics.
Missing links
Questions raised by gaps in the public record.
1. Nine unnamed victims
The central gap. The exhumation project is the first serious attempt in decades.
2. A complete case file
A public index of Cleveland police files, coroner reports, and Ness's own papers.
3. The postcards
Where the postcards to Ness are today, and whether any handwriting or DNA work was done.
4. Victim habits
Where each identified victim lived, drank, and worked, mapped against the dump sites.
5. The 1938 raids
What was seized from the Flats shantytowns and what became of it.
6. Sweeney's records
Medical, military, and hospital records for Dr. Sweeney, and whether anyone has compared them with the case.
7. The Black Dahlia comparison
Authorities examined a possible link to the 1947 Los Angeles case. No connection has been established.
8. Cases after 1938
Whether later unsolved dismemberment cases, such as Robert Robertson in 1950, belong to the series.
Evidence gaps
"Not public" means we could not find a report.
| Item | Tested | Gap |
|---|---|---|
| Remains of unidentified victims | Being exhumed and sampled (from 2024). | Results: not public. |
| Police and coroner files | Created in the 1930s. | Public access: not confirmed. |
| Ness's papers and the postcards | Referenced in histories. | Location and analysis: not public. |
| Dolezal confession record | Used in the 1939 case. | Full record: not public. |
Forensic tech
| Tool | What it could do | Status here |
|---|---|---|
| Genetic genealogy (DNA Doe Project) | Identify the nine unnamed victims. | Begun 2024; no results published |
| Fingerprint database search | Match victims to old police records. | Identified Andrassy and Polillo |
| Digitized records search | Match victims to missing-person and census data. | Available |
| Facial reconstruction | Create likenesses from skulls. | Death masks exist; we show no images of remains |
| Geographic profiling | Map victims' homes and dump sites. | Not public |
Solved cases
DNA Doe Project successes
The organization has used genealogy to name long-unidentified remains across the country. The lesson: names can come back after decades if remains were kept.
Golden State Killer
Genealogy led to a 2018 arrest of a killer from the 1970s–80s. The lesson: it needs killer DNA, which may not exist here.
What differs here
The victims, not the offender, are the target of the new genealogy work.
Case details come from public news coverage. Confirm in primary records.
False leads
| Claim | Source and date | Where it stands |
|---|---|---|
| Frank Dolezal was the Butcher | 1939 confession | Retracted; found hanged; exonerated posthumously. |
| Dr. Francis Sweeney was the Butcher | Polygraphs; later authors | Circumstantial; never charged. |
| The Cleveland killer also killed Elizabeth Short (Black Dahlia) | Various authors | Authorities examined it; no connection established. |
| There were 40 or more victims | Some researchers | The official count is twelve; the infobox gives 13 to 20+. |
| Dr. James Badal says he is "99.9 percent sure" he knows the killer's identity | News 5 Cleveland (undated) | No name, reasoning, or evidence is given in the report; unverified. |
| The killer was a doctor | Popular theory | Unproven; inferred from dismemberment skill. |
Where
Most remains were found in Kingsbury Run, a ravine near downtown Cleveland, and in the Flats and along the lakeshore.
Open in maps: Kingsbury Run · The Flats. We show no photographs of the sites or remains.
Source ratings
| Source | Type | Rating | Use with care because |
|---|---|---|---|
| Cuyahoga County Medical Examiner / DNA Doe Project statements | Agency / nonprofit | A− | Few results published so far. |
| 1930s Cleveland newspapers | Primary press | B+ | Sensationalized. |
| Wikipedia: Cleveland Torso Murderer | Tertiary | B | Good summary; counts differ. |
| Books and documentaries | Advocacy | D | Often argue one suspect. |
Records
Cleveland Division of Police
Historic files may be in city or county archives.
Cuyahoga County Medical Examiner's Office
Leading the exhumation project.
DNA Doe Project
Nonprofit working on identification: dnadoeproject.org.
Cleveland Public Library and Western Reserve Historical Society
Likely holders of period newspapers and papers.
Sample public-records request
Under the Ohio Public Records Act, I request copies of non-exempt records relating to the 1934–1938 Kingsbury Run (Torso) homicides, including coroner reports and identification records. Please send electronic copies. If any part is withheld, please identify the reason.
Records this old may be held by archives rather than by the police department.
Remembering the victims
Most were poor and unnamed. Efforts to give them names, and descendants the chance to claim them, are the most meaningful work still being done.
News & databases
In the news
Loading headlines…
Sources
Every fact on this page comes from one of these. Where they disagree, we say so above.
- Cleveland Torso Murderer, Wikipedia (victims, Ness, suspects, 2024 exhumations, Black Dahlia).
Photo credits
| Image | Author / source | License (per Commons) |
|---|---|---|
| Wikimedia Commons | gargantuen (Flickr) | CC BY-SA 2.0 (credit required; shown unmodified) |
| Wikimedia Commons | gargantuen (Flickr) | CC BY-SA 2.0 (credit required; shown unmodified) |
| Wikimedia Commons | gargantuen (Flickr); sketch credited to Michael Nevin | CC BY-SA 2.0 (credit required; shown unmodified) |
Names, dates, and counts come from one Wikipedia article and are as reliable as its sources; we did not check primary records. We found no published results of the 2024 exhumations. The Sweeney polygraph claim and the Black Dahlia comparison are as reported there.
Change log
- Oct 8, 2026: File published.