Key Takeaways
Treat a conviction as a finding, not proof the case is complete.
A case went cold. In 1962, engaged couple Tim McKillop and Joyce Sterrenberg were shot dead in the Arizona desert. Investigators failed to secure the scene, cast tire tracks, or test a clump of hair found nearby. Years later, teenager Linda Primrose described witnessing the murders and led deputies to the site, but the investigation stopped without finding the people she named. Leads went unconnected. Ernest Valenzuela separately confessed to the killings, and a psychiatrist described him as dangerous. He was released, and records show no effort to connect him to Primrose’s account. Twelve years later, Bill Macumber was arrested, but the discarded leads and missing records left the case looking more settled than it was.
The case illustrates how an investigation can narrow not because uncertainty has been resolved, but because institutions stop asking questions. That distinction matters in any high-stakes inquiry: a file can become administratively closed while important evidence remains unexplored. Modern investigative practice emphasizes documenting decisions and preserving material precisely because later reviewers cannot reconstruct what was never recorded. Still, an abandoned lead does not prove the lead was true. Primrose’s changing account and later recantation complicate the story. The lesson is not to accept every alternative account, but to investigate it before treating the official explanation as complete.
Demand a traceable chain before trusting forensic evidence.
Access was too casual. At the murder scene, officers collected four .45-caliber shell casings and a live round, but the surviving record did not establish a reliable path from the desert to storage. Years later, former evidence technician Dave Brewer said employees could reach latent-print cards and crime reports, and that casings sat in an often-unlocked desk drawer. Bill’s brother recalled being allowed to handle evidence and take casings to a Motorola laboratory for examination. The physical evidence vanished. A court-authorized evidence purge in 1982 destroyed the exhibits, leaving later investigators dependent on photographs and old testimony. The palm print that became central at trial had been considered too unclear to classify in 1962, then was identified as Macumber’s in 1974.
Forensic evidence gains authority from its scientific appearance, but reliability depends on mundane safeguards: labeling, access logs, secure storage, and independent review. A broken chain does not establish that evidence was altered. It does make claims about its identity and history harder to verify, especially once the original objects are gone. This is a general lesson in measurement science: precision in the final reading cannot repair uncertainty about the sample. The more decisive a single print or casing becomes, the more important it is to preserve the material and document every transfer.
Record interrogations before disputed words become the prosecution's strongest evidence.
Sixteen hours went undocumented. On the day of Macumber’s arrest, deputies questioned him from morning until midnight. They recorded neither the interview nor contemporaneous notes, although a stenographer had taken down Carol’s accusation. Officers later testified that Macumber admitted telling his wife he had killed the couple. Macumber denied making that admission; his friend Carl Pace also said he heard no confession. The ambiguous phrase Macumber acknowledged using, about saying something to keep Carol from leaving, was presented as incriminating. No recording could settle it. The jury had to choose among recollections shaped by the investigation and the defense, rather than hear the conversation itself.
Recording protects both sides. It can expose coercion or inaccurate paraphrase, but it can also preserve a genuine admission and prevent later disputes about what was said. This is a procedural safeguard, not a presumption that police testimony is false. Human memory is reconstructive, particularly under stress and after repeated retelling, so contemporaneous records reduce dependence on confidence alone. The asymmetry in this case is striking: Carol’s account was carefully transcribed, while the suspect’s lengthy questioning was not. When a statement may carry the weight of a confession, transparency should be standard rather than optional.
Let juries hear credible alternative suspects, even when legal rules complicate disclosure.
Privilege blocked a competing account. Valenzuela confessed the Scottsdale murders to lawyers Thomas O’Toole and Ron Petica while they represented him on another murder case. Both later said they believed him. Because the client had died, the lawyers initially faced attorney-client privilege, a rule designed to let clients speak candidly to counsel. Valenzuela’s mother eventually waived the privilege, but the judge still excluded the testimony, finding it insufficiently trustworthy. The jury never weighed it. Nor did the first trial jury hear Primrose’s account in full. The U.S. Supreme Court’s Chambers v. Mississippi decision had recognized that excluding reliable third-party confession evidence can violate due process, but Macumber’s courts distinguished his case.
Confidentiality is not a technical nuisance. Without it, clients may conceal facts that lawyers need to defend them, and the justice system loses candor at its source. Yet when a confession points to another person and the accused faces life imprisonment, silence carries a grave cost. The hard question is how to evaluate reliability without letting a judge substitute intuition for a jury’s assessment. Repetition, details unavailable publicly, corroboration, and consistency can help, but none is conclusive alone. Macumber’s case shows why legal systems need principled exceptions, not blanket rules or case-by-case instinct.
Turn suspicion into testable evidence before asking a court to reopen a case.
The review took years. The Arizona Justice Project began examining Macumber’s case after Judge O’Toole urged Larry Hammond to reconsider it. Volunteers assembled records, tracked down witnesses, and searched for physical evidence. Their early theory that the palm print could not have come from the Impala’s door did not survive repeated expert review. They could not prove the print had been planted, and missing evidence limited what could be checked. A new foothold emerged. A 2009 National Academy of Sciences report described weaknesses in forensic comparison methods. Three experts later challenged the certainty of the original ballistics identification, which had relied on ejector marks despite other firearm features not matching. Their affidavits helped support a post-conviction petition.
The project’s most valuable discipline was distinguishing a compelling story from a legally useful claim. Suspicious circumstances can justify investigation, but courts generally need new evidence tied to a specific legal threshold. The long delay also exposes a structural problem: volunteer labor can bring extraordinary commitment, yet rotating students and unpaid experts make continuity difficult. Scientific skepticism should be symmetrical, too. The state’s original certainty deserved scrutiny, but an expert’s inability to explain a fingerprint anomaly did not itself prove fabrication. Good review tests both the prosecution’s claim and the defense’s proposed alternative.
Treat rehabilitation as evidence of capacity, not proof of innocence.
Prison became a record of contribution. Macumber helped establish the Roadrunners, a prison Jaycee chapter, and led projects including snack bars, sports facilities, classes, and charity fundraising. He later taught, mediated disputes, worked in prison programs, and wrote extensively. His disciplinary record remained exceptionally clean, and prison staff and fellow inmates trusted him with responsibilities unusual for a person serving life without parole. That record mattered, but differently. It could not establish who killed Tim and Joyce. It did give the clemency board evidence about Macumber’s conduct, abilities, and likely life outside prison. His old-code sentence offered no ordinary parole route, making clemency one of the few available paths to release.
A person’s conduct after conviction answers a different question from whether the conviction was correct. Conflating those questions can either overstate rehabilitation as proof of innocence or dismiss it as irrelevant. Macumber’s record showed that a prisoner can create value and responsibility inside an institution that often reduces people to risk categories. It also gave decision-makers concrete evidence about his capacity to live safely in the community. The case invites a broader correctional question: if prison work, teaching, and mediation demonstrate capability, should such evidence matter only when a prisoner has a path to parole?
Use clemency as a safeguard, but recognize how exposed it is to politics.
A different forum listened. In 2009, Arizona’s Board of Executive Clemency heard the Justice Project’s account, including Valenzuela’s confessions, Primrose’s statements, disputed forensic evidence, and Macumber’s prison record. The board unanimously recommended release for time served, describing substantial doubt about his guilt. Governor Jan Brewer rejected the recommendation after announcing she would run for reelection. At a second hearing in 2012, the board split two to two, so no recommendation went forward. Clemency was not a retrial. It offered room to consider fairness and human consequences after ordinary appeals had failed, but the process depended on decision-makers who were not required to agree with the evidence or the previous board.
Clemency can address cases that procedural rules leave stranded, especially when evidence has disappeared or a conviction cannot be reopened under ordinary standards. Its flexibility is also its vulnerability: the same discretion that permits correction can make outcomes inconsistent or sensitive to political risk. A public board’s reasoning and a governor’s response therefore matter, not only the final decision. The Macumber hearings show both promise and limitation. A forum outside the courtroom can recognize serious doubt, yet recognition alone does not equal release, and release still requires political approval.
A son's return reopened the case, but could not repair every family fracture.
Ron searched for his father. Separated from Macumber for nearly three decades, he had grown up believing his father was a murderer. After the Justice Project explained the case, Ron wrote to him in 2003, asking to see the evidence. Their letters became calls, then a prison visit with Ron’s daughter. Ron came to believe his father was innocent and testified for him at the 2009 clemency hearing. The family remained divided. At the 2012 hearing, Macumber’s two older sons supported his continued imprisonment and condemned him, while Ron stood by him. The contrast made the case’s central problem personal: shared history did not produce a shared account of what had happened.
Ron’s return was consequential even before it helped the legal campaign: it restored a relationship and gave Macumber a family connection he had thought lost. But family testimony is not a neutral instrument. Memory is shaped by loyalty, injury, separation, and the stories relatives hear while growing up. The older sons’ opposition cannot simply be dismissed, just as Ron’s support cannot prove innocence. Their conflict underscores a distinction between emotional truth and forensic truth. Both deserve attention, but neither can substitute for the other.
A plea deal can free someone without settling whether the conviction was wrong.
The court allowed one last route. In 2012, Judge Bruce Cohen granted an evidentiary hearing on third-party guilt and ballistics. Before it began, the state agreed to a no-contest plea to two second-degree murder counts in exchange for setting aside the first-degree convictions and sentencing Macumber to time served. The state said lost evidence and the risk of post-conviction relief made a retrial difficult. Freedom did not equal exoneration. The judge acknowledged evidence on both sides and significant unanswered questions. Macumber was released after thirty-eight years, but the plea resulted in a legal finding of guilt, and he continued to maintain his innocence. The case ended with freedom, not a definitive account of the 1962 murders.
The ending resists the clean categories that public narratives prefer. The plea protected Macumber from continued imprisonment and spared the state a retrial it might not have been able to conduct fairly, given the lost evidence. It also left a permanent ambiguity: legally, he stood convicted of lesser offenses; factually, the central question remained unresolved. Such settlements can be humane and practical without being fully satisfying. They should not be advertised as exonerations, but neither should the plea alone erase the record of procedural failures and substantial doubt.
Analysis
Manifest Injustice is not a conventional whodunit with a final revelation. Its force comes from making uncertainty itself the subject. The evidence against Macumber was not imaginary: a palm print was identified as his, an FBI examiner linked shell casings to his pistol, and officers testified that he acknowledged telling his wife he had killed the couple. But the competing case was also substantial: a violent man confessed repeatedly, an eyewitness account appeared to corroborate him, the physical evidence was poorly preserved, and the interrogation that allegedly produced an admission was never recorded. Neither narrative can be responsibly reduced to a single decisive clue.
The book exposes an epistemic imbalance in criminal adjudication. At trial, the state’s evidence arrived through official witnesses and forensic experts, while the defense struggled to find comparable expertise and could not present the confession. Once the verdict hardened, the burden shifted dramatically: the convicted person had to uncover new evidence from a record that had been mishandled and partly destroyed. That is not proof that courts acted maliciously. It is a system design problem in which finality is administratively valuable, while correcting an error is costly and rare.
The Justice Project’s long effort is both admirable and imperfect. Its patience uncovered witnesses, challenged overstated forensic certainty, and helped reconnect a father and son. Its volunteer structure also produced years of delay, shifting personnel, and an abandoned investigative path when experts could not substantiate the fingerprint theory. The story therefore supports persistence while warning against persistence becoming a substitute for measurable progress.
The outcome is morally important but legally incomplete. Macumber left prison, his first-degree convictions were set aside, and a judge acknowledged serious doubt. Yet a no-contest plea left him legally guilty of lesser murders, and the book cannot establish with certainty who killed Tim and Joyce. The title’s idea of manifest injustice is best understood here as a condition requiring correction even when perfect historical certainty is unavailable, not as a synonym for proven innocence.
Review Summary
Manifest Injustice receives overwhelmingly positive reviews, with readers praising Siegel's detailed research and compelling narrative. Many found the book eye-opening, highlighting flaws in the justice system. Readers were moved by Bill Macumber's story, his resilience, and the efforts of those working to exonerate him. Some noted the book's technical legal details and repetition but found it necessary for understanding the case. Overall, reviewers were emotionally invested, expressing frustration with the justice system and admiration for those fighting wrongful convictions.
People Also Read
Glossary
Manifest injustice
obvious unfairnessThe Arizona Justice Project used this phrase for cases involving an apparent serious unfairness in a conviction or sentence, including cases where evidence of innocence or procedural failure warranted renewed scrutiny. The claim was broader than a simple assertion of innocence, but it still required credible grounds for intervention.
Arizona Justice Project
post-conviction review groupA nonprofit initiative founded in Arizona in 1998 to investigate claims of innocence and manifest injustice. Unlike DNA-focused innocence organizations, it considered cases involving other forms of evidence and procedural unfairness. Its work relied heavily on volunteer lawyers, law professors, law students, investigators, and forensic experts.
attorney-client privilege
protected legal communicationsA legal protection that generally keeps confidential communications between a lawyer and client from being disclosed in court. In Macumber’s case, Valenzuela’s lawyers said he confessed the murders to them. Courts initially treated those communications as privileged even after Valenzuela died, raising a conflict between confidentiality and a defendant’s ability to present evidence of another person’s guilt.
third-party guilt
evidence implicating someone elseA defense theory that evidence points to a person other than the accused as responsible for the crime. In this case, the theory centered on Ernest Valenzuela’s repeated confessions and Linda Primrose’s account of witnessing the murders. The issue was whether the evidence was reliable and legally admissible enough for a jury or reviewing court to consider.
Chambers v. Mississippi
third-party confession precedentA 1973 U.S. Supreme Court decision involving a defendant convicted of killing a police officer after another man had confessed. The Court held that excluding testimony about the confession under the circumstances violated due process. The case became central to arguments that defendants must sometimes be allowed to present reliable evidence implicating another person.
Rule 32
Arizona post-conviction procedureThe Arizona criminal procedure rule governing post-conviction relief for people convicted of crimes. Macumber’s petition relied on claims of actual innocence and newly discovered evidence. Under the standard described in the book, he had to show clear and convincing evidence sufficient to establish that no reasonable fact-finder would have found him guilty beyond a reasonable doubt.
old-code lifer
lifer without parole eligibilityA label for Arizona prisoners sentenced under the older criminal code, which did not provide the same automatic parole eligibility as the later code. Macumber’s life sentence offered no ordinary parole route, leaving executive clemency as one of his few possible paths to release.
Herrera v. Collins
clemency as a safety valveA 1993 U.S. Supreme Court decision discussing claims of innocence based on newly discovered evidence after ordinary judicial remedies have been exhausted. The opinion described executive clemency as a traditional means of preventing miscarriages of justice. Arizona clemency board chair Duane Belcher relied on this reasoning when considering Macumber’s case.
Strengthening Forensic Science in the United States: A Path Forward
forensic science reviewA 2009 National Academy of Sciences report reviewing the condition of forensic science in the United States. It identified limitations in several forensic comparison disciplines, including firearms identification. The Justice Project used the report to challenge the certainty of testimony that linked Macumber’s pistol to crime-scene shell casings.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.