Key Takeaways
Treat a spouse's insurance windfall as motive, not proof of murder
A tidy accusation. In December 1966, Henry Stockton was found shot five times in his home. Nearby were his life insurance policy, naming his wife Sandra as beneficiary, and a receipt for a .22-caliber revolver purchased in her name. A suspicious fire had started in a bedroom. The clues seemed to point straight at Sandra, yet the fire might also have destroyed evidence, and the gun receipt did not establish who possessed the weapon or fired it.
Forensics stayed inconclusive. A fingerprint examiner first told investigators that all readable prints belonged to Henry. At trial, he disclosed two unidentified prints. Further examination found that one was Henry's; the other was a palm print that could not be identified. No fingerprint or recovered weapon connected either defendant to the killing. A suspicious scene can support a theory without proving it.
What's striking is how easily a coherent story forms around a few vivid objects: a policy, a gun receipt, and a fire. Cognitive science calls this tendency to favor a neat explanation over competing possibilities narrative bias. Investigators can resist it by asking what each clue proves independently, and what alternative account could produce the same clue. The insurance gave Sandra a possible motive, but motive is not an action, and ownership is not possession at the crime scene. The book's forensic reversals make that distinction concrete.
Follow the money, then test the timeline
Cash suggested a connection. After Sandra received $15,000 from Henry's workplace policy in February 1967, she bought $3,000 in traveler's checks and went to Las Vegas with Alan Palliko under assumed names. The checks were spent at casinos and hotels. Alan gambled, lost money, and soon deposited $2,500 in his bank account, though he also had other possible sources of funds. Later, Sandra received $17,677 from a second policy after a legal settlement.
Dates complicated the story. Michael Brockington said Alan collected an envelope of cash near Sandra's workplace before the insurance settlement arrived. Bank records showed Alan's first entry into a safe-deposit box was September 1, after Sandra received the money on August 29. Weather records also fit Michael's memory of an extremely hot day better than the earlier dates he named. The trail became more plausible, but cash never supplied a receipt proving its source.
Financial records can turn vague suspicion into a testable sequence. Dates, deposits, withdrawals, and access logs let investigators challenge a witness's recollection against events that were recorded at the time. Yet timing is not provenance: money is fungible, and a withdrawal near an insurance payment does not by itself identify the recipient. The strongest analysis keeps both ideas in view. Transaction patterns can corroborate a relationship or expose a false timeline, while still falling short of showing exactly who paid whom, or why.
Build a circumstantial case from links that withstand alternatives
Patterns can add force. Circumstantial evidence supports a conclusion through connected facts rather than a witness directly describing the crime. In the Stockton case, there was no eyewitness, murder weapon, matching fingerprint, or bullet comparison tying Alan or Sandra to Henry's killing. Prosecutors instead connected the insurance, the Las Vegas aliases, large cash withdrawals, Alan's changing explanations, and his later reported admission to Brockington.
Alternatives still matter. The defense proposed other explanations: Alan might have extorted Sandra, taken money as her lover, or learned of a murder she had arranged with someone else. The law required proof beyond a reasonable doubt, not merely a story that fit the evidence. A conspiracy is an agreement to commit a crime, accompanied by an act to advance it. The prosecution had to persuade jurors that this explanation fit better than the alternatives.
A network of clues can be more persuasive than a single dramatic fact, especially when independent sources converge. But apparent convergence can mislead if several clues depend on the same assumption. For example, a cash withdrawal, a presumed payoff, and a later bank deposit are not three independent proofs if the payoff is inferred from the withdrawal alone. Good circumstantial reasoning tests whether each link adds new information. It also takes alternatives seriously, not as a demand for certainty, but as a safeguard against mistaking coherence for proof.
Do not use grief or composure as a lie detector
Behavior invited suspicion. Sandra had an alibi: she was visiting her friend Vicky Stowe in Twentynine Palms when Henry was killed. When her father told her Henry was dead, she cried, then went horseback riding with the children. Later, during a police interview, she appeared composed and collapsed when asked for fingerprints. Investigators could read these reactions as guilt, but they could not establish what she knew or had planned.
The same problem returned. After Judy's murder, Alan's friends described him as devastated. Other witnesses said he appeared calm. The prosecution emphasized those calmer reactions; the defense emphasized the emotional displays. The trial offered no simple emotional measure of innocence or guilt. A person's demeanor may raise a question, but it cannot settle it without corroborating evidence.
Human beings often treat emotion as a transparent window into truth: grief should look a certain way, and an innocent person should react predictably. That assumption is unreliable. People differ in temperament, culture, shock response, and willingness to show emotion, while observers interpret behavior through their existing beliefs. Research on lie detection likewise finds that confidence in reading deception from demeanor tends to exceed actual accuracy. The book's competing readings of Sandra and Alan are a useful warning: behavior can be evidence, but it is especially vulnerable to confirmation bias.
Corroborate the witness who carries your whole case
Brockington mattered enormously. Alan's friend and employee eventually told police that Alan had privately admitted killing Henry. Brockington had delayed coming forward because he feared Alan. At first he also withheld information that made Alan look bad, which supported the claim that he had initially tried to protect him. But the confession was unrecorded, and Brockington's account had a damaging error about when Alan collected cash near Sandra's workplace.
Other evidence checked him. Bank safe-deposit records and weather reports helped show that his timing was mistaken, not necessarily invented. A separate dispute concerned Judy's life insurance: Brockington said he was present when Alan agreed to the policy, while insurance agent Jack Dodd claimed he had pressured Alan into it. Dodd later acknowledged Brockington had been there, but the accounts remained in conflict. The jury had to weigh credibility alongside corroboration.
Witness credibility is not a simple choice between trustworthy and dishonest. A person may be accurate about one event, wrong about a date, and self-protective about another. Memory research describes recollection as reconstructive, which is why contemporaneous records can be so valuable. Still, corroboration should not become a ritual that merely confirms the favored theory. Investigators should ask whether documents independently support the witness, and whether contrary evidence has been given equal scrutiny. This case shows both the power and the fragility of testimony.
Update the theory when new evidence changes the map
A witness changed the question. Prosecutors initially argued that Alan had arranged for someone else to kill Judy, partly because bar employees placed him at work during the murder. Near the end of the case, apartment tenant Walter Wasson said the man he had seen hurrying away from the building looked and walked like Alan. Wasson had not seen the man's face, so the identification was suggestive, not certain.
The timeline allowed a possibility. Investigators had first thought Alan had too little time to travel between the bar and the apartment. A closer check showed the ambulance left nearly half an hour after the shooting, enough time for the trip. The prosecutor added the possibility that Alan had killed Judy himself, while retaining the theory that he had arranged it. The jury convicted him, but its members reportedly divided over whether he was the shooter.
Changing a theory in response to new evidence is a strength when the change is explicit and proportionate. In Bayesian reasoning, evidence updates the relative plausibility of explanations; it does not magically convert uncertainty into fact. Wasson's account made direct involvement more plausible, but the lack of a face identification limited what it could establish. There is also a risk: flexible theories can become impossible to disprove if every outcome is treated as support. The responsible move is to name the uncertainty and show precisely what the new evidence does, and does not, resolve.
Require proof of agreement before assigning a partner's crime
The law reached beyond the trigger. The prosecution argued that Sandra and Alan agreed to murder Henry for insurance money. Under the conspiracy rule described at trial, an agreement to commit a crime plus an act advancing it can make each conspirator responsible for crimes committed to carry out that plan. This is vicarious liability: one person's criminal act can legally count against another conspirator.
Association alone was not enough. Sandra was not at the house when Henry was killed, and her Twentynine Palms alibi was not disputed as a physical location. The case against her depended on the alleged agreement, supported by the insurance, her relationship with Alan, the money trail, and their behavior afterward. The jury found her guilty of first-degree murder. That verdict reflects its judgment that the evidence proved participation in the plan, not that mere friendship or an insurance benefit automatically established guilt.
Conspiracy law addresses a real feature of organized wrongdoing: plans are often divided among people, so the person who performs the final act may not be the only responsible participant. At the same time, this doctrine makes the boundary between association and agreement crucial. Shared motives, contact, or benefit can be evidence, but they are not interchangeable with proof of a common plan. The book demonstrates why jurors must assess the whole pattern while resisting the shortcut that proximity to a crime equals participation in it.
Let verdicts settle legal guilt, not every historical question
The jury reached a verdict. In 1969, Alan was convicted of murdering Henry and Judy, and of attempting to murder Katherine. Sandra was convicted of Henry's murder. The jury sentenced Alan to death and Sandra to life imprisonment. After the United States Supreme Court invalidated existing death-penalty schemes in 1972, Alan's sentence became life imprisonment. Sandra was paroled in 1976, after seven years in prison.
The shooter remained unknown. Bugliosi believed Alan had engineered Judy's murder, but said he could not establish who fired the gun. Some observers suspected Brockington, Sandra, or another person with a grudge. The jury's verdict settled legal responsibility under the evidence and law presented at trial; it did not identify the person who physically killed Judy. That distinction is the book's lasting unease: a case can be legally closed while a central historical fact remains unresolved.
Courts must convert incomplete evidence into decisions that carry real consequences. Historical inquiry has a different task: it can preserve uncertainty when the record does not answer a question. Treating a verdict as proof of every detail risks claiming more than the jury decided, while treating unresolved details as proof of wrongful conviction makes the opposite mistake. The most honest account keeps institutional findings, the prosecutor's interpretation, and unanswered factual questions distinct. Here, the outcome is firm, but the identity of Judy's killer is not.
Analysis
This is both a murder narrative and an account of how institutions reason under uncertainty. Its engine is not simply the question of who killed Henry and Judy, but the changing relation between clue, inference, and proof. The fire, insurance documents, aliases, cash movements, conflicting witnesses, and courtroom disputes repeatedly offer explanations that seem complete until a new fact unsettles them. The movement from a police investigation into a prosecutor's retrospective account makes that instability part of the story.
The book's strongest contribution is its demonstration that circumstantial evidence is not inherently weak, but that its force depends on how carefully each inference is tested. The prosecution builds a case from converging behavior and financial records, yet its own mistakes matter: a fingerprint examiner's error, a witness's mistaken chronology, and uncertainty about the second murderer's identity. Those details make the narrative more than a celebration of prosecutorial skill. They show the difficulty of seeking a just verdict when the evidence is incomplete and the accusation is grave.
The account also distinguishes legal resolution from historical certainty. The jury convicted Alan of Judy's murder, while its members reportedly disagreed about whether he had fired the shots himself. Bugliosi believed Alan arranged the killing, but the perpetrator's identity remained unknown. That ending is not a failure of storytelling. It is a serious reminder that a verdict answers a legal question under a defined standard, not every question a reader can ask.
One claim deserves particular caution: the book argues for the deterrent value of capital punishment partly by comparing changes in California execution and homicide rates. A rise in homicide after executions declined does not, on its own, establish that one caused the other. Many social conditions can change over the same period. The book is more persuasive when it exposes the demands of proof than when it treats broad historical comparisons as conclusive. Its lasting value lies in holding narrative conviction and evidentiary restraint together.
Review Summary
Till Death Us Do Part receives generally positive reviews, averaging 4.05/5. Readers praise Bugliosi's meticulous attention to detail and compelling courtroom narrative. The book's two-part structure—covering the murders and subsequent trial—is widely appreciated, though many find the trial portion overly technical and slow. Fans of Bugliosi's Helter Skelter frequently compare the two, noting this case is less sensational but still gripping. Critics cite repetitiveness and self-congratulatory writing as weaknesses, while admirers highlight the fascinating circumstantial prosecution and valuable legal insights.
People Also Read
Glossary
Circumstantial evidence
Evidence supporting an inferenceEvidence that establishes a secondary fact from which a fact in dispute may be inferred. Seeing a suspect possess stolen property, for example, may support an inference that the suspect stole it, but does not directly show the theft. In this case, aliases, insurance payments, and financial behavior were used to infer a murder conspiracy.
Conspiracy
Agreement to commit a crimeA criminal agreement between two or more people to commit an unlawful act, accompanied by an act intended to advance the plan. The prosecution argued that Alan and Sandra agreed to murder Henry for insurance money. Because the alleged agreement was not directly documented, jurors had to decide whether the surrounding conduct proved it beyond a reasonable doubt.
Vicarious liability
Responsibility for a partner's crimeThe principle that a conspirator may be held legally responsible for another conspirator's criminal act when that act advances the shared plan. The prosecution relied on this rule to argue that Sandra could be guilty of Henry's murder even though she was not physically present when he was killed.
Negative fingerprint testimony
Context for missing printsBugliosi's term for expert testimony explaining how rarely investigators recover usable fingerprints belonging to a suspect. He cites LAPD figures indicating that readable prints were obtained in about 30 percent of examined crime scenes, and that prints matching a defendant were found in about 10 percent of those cases. He used the figures to counter the claim that absent prints proved a suspect was never present.
Bifurcated trial
Separate guilt and penalty phasesA capital trial conducted in two stages. First, jurors decide whether the defendant is guilty. If they return a qualifying murder conviction, the same jury then hears evidence about punishment and chooses between death and life imprisonment. The Palliko-Stockton trial followed this structure, separating proof of guilt from the penalty decision.
Beyond a reasonable doubt
The criminal proof standardThe level of certainty the prosecution must establish for a criminal conviction. It does not require eliminating every imaginable possibility, but it requires more than suspicion or a conclusion that guilt is merely likely. The jury in this case had to decide whether the evidence against each defendant met that demanding standard.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.