Mary Anne Fagan - They Didn't Fit The Timeline

They thought they had him. For more than sixteen hours, his answers kept changing. He was never charged.

In April 1978, two months after Mary Anne Fagan was murdered in her own home, the Homicide Squad sat a man down in Russell Street and worked through his movements, his money, and what he had said about her. By the end, he had been caught in contradiction after contradiction. The Coroner would later name him in open court as far from honest.

He walked free.

For forty-eight years, the gap between what police believed and what they could prove has remained open.

Here, we stop assuming and start testing.

Four conditions any explanation of this murder has to meet — the door, the blood, the cigarette, and the departure — are held against both men the investigation pursued: the man at the corner, and the man at the gate nobody ever named.

One thread comes apart. The other has been sitting in an institutional shadow since 1978.

Content warning: this episode discusses murder and includes explicit sexual language drawn from sworn inquest testimony. Listener discretion is advised.

Transcript

Content warning: this series discusses the murder of a woman in her home.  This episode revisits the police interrogation record and includes references to sexually explicit language given in sworn testimony at the inquest.  The language is offensive and is included only because it forms part of the evidentiary record.  Listener discretion is strongly advised.  If you or someone you know needs support, contact Lifeline on 13 11 14, or 1800 RESPECT, on 1800 737 732, or your local crisis service.

In April 1978, two months after the murder of Mary Anne Fagan, a homicide detective sat across a table from a council labourer in Russell Street.

The interview ran into the night.

By the close, the Homicide Squad believed it had enough to charge him in connection with Mary Anne's death.

He had lied in his first statement. He had lied in his second. He had been caught out, on his own evidence, in lie after lie about where he had gone that morning, where the money in his pocket had come from, what he had said about Mary Anne to the man on the fence beside him.

The Coroner, fourteen months later, would name him in open court as far from honest.

And he was not charged.

Seventy pages of transcript. Hours of questioning. Further interviews after that. A man brought back to the table again and again, until even he conceded, on oath, that he had not told the police the truth.

So when the case against him is tested — not against the assumption of his guilt, but against what the evidence of the case actually requires — does it survive?

And if it doesn't — who walked away?

Part 5: Mary Anne Fagan

This case has constraints.

Across this series, we have built them, piece by piece, from the primary record. The conditions any explanation of Mary Anne's murder has to satisfy if it is going to hold up.

The killer entered the house without forcing entry. The only broken pane of glass was the one Mary Anne's children smashed that afternoon to get inside. The door was opened from within.

The killer was in the company of a woman who was halfway through dyeing her hair. Dye paste was still in her hair when her children found her. The toothbrush used to apply it was still on the bathroom vanity. Hair-dyeing, on her husband's evidence, was an extremely personal activity — one she would not interrupt for an ordinary caller. Whatever caused her to leave the bathroom, walk down the hall, and open the front door overrode every normal instinct about privacy and appearance.

The telephone was found disconnected when police arrived. The plug had been pulled from the wall socket in the hallway. Whether that happened during the attack or in the chaos that followed is not perfectly clear.

The killer used materials that were already in the room. The bindings on Mary Anne's wrists and ankles were strips torn from a towel from her own household. The gag was a piece of striped cloth from the same house. The plastic sheet beneath her body was almost certainly her own drop sheet, set out for the hair-dyeing.

The killer brought a weapon and took it away. Despite the most extensive ground search the Homicide Squad would conduct that decade — including digging up the road repair in case the weapon had been entombed in the bitumen — the knife was never found.

No blood distinguishable from Mary Anne's was identified in the samples tested at the Forensic Science Laboratory. Every sample returned Group B — Mary Anne's blood group. Within the limits of 1978 serology, no second source was identified.

The killer smoked a cigarette in the room. On the mantelpiece, near the bed, a butt was found burnt to the filter, the ash column still intact. It was not a Peter Stuyvesant — Mary Anne's brand. The Detective Senior Sergeant on the scene that evening described it at inquest as more likely a Winfield. A cigarette burnt to the filter takes between five and seven minutes. Someone stood in that room for the duration of a full cigarette.

The killer took Mary Anne's red handbag — containing approximately one hundred and eighty dollars in cash drawn from her bank account the day before, an eternity ring, two cigarette lighters, religious medals, bank passbook, chequebook and keys. The bag was never recovered. But rings on her hands, and a wristwatch on her left wrist, were left untouched.

The killer left through the front door. A man was seen walking out of the front gate at approximately ten minutes past midday. He paused. He looked back at the house. He looked both ways along Dandenong Road. He walked west, at an unhurried pace, toward Glenferrie Road.

Thirty-five minutes later, a second independent witness, approximately a kilometre and a half further west, saw a man running across Dandenong Road from the central plantation, signalling for a lift. The man would say, when asked where he was going, "anywhere." He would get out short of Williams Road and disappear.

The man wore a service uniform. The first witness — a former Navy man — identified it as the new-style RAAF summer dress. The second witness, asked at inquest about a Royal Australian Air Force belt, said it looked very familiar.

Neither witness has been contradicted on the central elements of the sighting. Neither has, in forty-eight years, been credibly explained away. Neither man has ever been identified.

These are the conditions any explanation of Mary Anne's murder has to meet.

Not every condition does the same work.

Here, we take the two threads the investigation pursued — the council labourer and the man at the gate — and test each against those conditions. We then acknowledge that a third possibility, in the strict sense, has never been formally closed; and test that possibility against the same conditions.

We start with the man on whom the investigation spent seventy pages of interrogation.

His name is James.

******

By the close of the interrogation on the 20th of April 1978, the Homicide Squad believed it had a case. Across that day and into the night, on his own evidence and the evidence of others, James had been caught out in concealment after concealment.

That he had spoken to Mary Anne that morning, when he had first said he had not. That he had walked into her open driveway, when he had first said he had stood at the kerb. That he had left the worksite for forty-five minutes during the period the investigation came to treat as the murder window. That he had sat on a fence facing her house and made explicit comments set out earlier in the series — comments his de facto wife, on oath, had never heard him make. That he had returned to the worksite with cash he could not consistently account for, cycling through four different stories about its origin across hours of questioning. That his bookmaker — the named other party — had given evidence on oath that James had not collected money from him that morning, did not collect money from his home before five o'clock on any day, and in fact owed him forty dollars.

That, on cross-examination at inquest, when it was put to him that he had not told the police the truth, he had said: yes.

The note found in the working file — statements true, subsequent statements, had not told the police the truth — had been written about him. And he himself had now confirmed it.

The Coroner, fourteen months later, would say so in open court. The newspaper headline the next morning would carry it into the public record.

And on the test the law requires, none of this was enough. He was not charged.

For forty-eight years, the gap between what police believed and what they could prove has stood open.

There is a reason for that gap.

*****

Take what the seventy pages contain — and ask, of each piece of it, whether it satisfies the conditions of the case.

Four constraints discriminate. Some of the conditions described earlier — the improvised bindings, the weapon brought in and out, the taken handbag — could in principle be satisfied by anyone with access to the house and a knife. They do not separate one candidate from another. Four do. The door. The blood. The cigarette. And the departure.

These are the four where each thread either holds, or comes apart.

*****

The first constraint. The killer entered the house without forcing entry.

James had spoken to Mary Anne about silt in her driveway. He had walked with her into her open garage area. He had walked through her front gate at about twenty past eleven with the man on the fence beside him. He never knocked on the door. By his own admission, repeatedly, across three separate statements: he did not knock.

If James was the man who entered that house, he had to find a way to get Mary Anne to open the front door — a woman who, while polite, did not know him, and who was halfway through a private hair-dyeing process her husband described as one she would not interrupt for anybody.

He had no professional pretext that would override that interruption. He was a man in council work clothes, a blue terry-towelling hat, industrial boots, whom she had met once that morning about the council's responsibility for clearing silt.

There is no mechanism on the record that would have caused Mary Anne to abandon the bathroom, walk down the hall, and open the front door to him.

The pathologist who reconstructed the crime gave the most feasible version this way. Mary Anne was in the bathroom dyeing her hair when the killer first encountered her. She was made to walk from the bathroom to the front bedroom, undress, and lie down on the bed.

That reconstruction depends on something the same pathologist conceded. Mary Anne would not have admitted an ordinary visitor in the middle of dyeing her hair. The process required precise timing. Her husband described it as private. The pathologist agreed it was unlikely she would have admitted someone she knew socially in that state.

Either way, the door had to open. Either it was already unlocked — which her husband said was not her practice — or Mary Anne admitted the person despite the dye. If she admitted him, the reason had to override the timing.

A council labourer with a private quote she had not asked for did not override that timing.

However, a man in service uniform, at the door of a service household, while the serving member was at the base, did.

*****

The second constraint. No blood distinguishable from Mary Anne's was found at the scene.

Fourteen stab wounds produced a great deal of blood. The chest cavity was full. The bedding was sodden. The wall beside the bed was spattered. Every sample tested in the room was Group B — Mary Anne's group. Within the limits of 1978 serology, no second source was identified.

If James was the killer, he stood in that room — close enough to deliver fourteen wounds, in multiple directions, into the back of a woman lying face down — and left no detectable trace of his own blood and no transfer of hers onto his person.

He then returned to a worksite fifteen metres from the front door of the house in which she was lying dead. He had nowhere to wash. No clean clothes to change into. No private space. He worked alongside the same colleague who had been on the fence with him that morning. He had lunch with the rest of the council depot crew. He shovelled hot mix until the wind-up whistle.

That evening he went to two hotels. He played pool. He paid back five dollars to a council workmate. He drank in the same blue council clothes he had worn all day, until police arrived to take a brief statement from him. He had not changed.

The colleague on the fence — a man with twenty-two years on the council, who had worked alongside James all morning, lunched with him, drunk with him at the Armadale Hotel that evening, and been questioned by police in his presence — noticed nothing different about the way James looked, smelled, or behaved when he came back to the corner.

If James was the killer, he had crossed back into the workday — and into the public bar of an inner-suburb hotel — wearing clothes that should have been forensically catastrophic. There is no evidence in the record that they were.

*****

The third constraint. The killer smoked a cigarette in the room.

The butt on the mantelpiece, burnt to the filter, was identified by the senior officer on scene that evening as more likely a Winfield.

James smoked Benson and Hedges. By his own admission, given in the same series of interviews in which he conceded most of what he had previously denied, the cigarettes he bought at the milk bar that morning were Benson and Hedges.

The brand on the mantelpiece does not match the brand he was carrying.

There is also what the cigarette tells us about the person who smoked it.

A cigarette burnt to the filter is not a panicked smoke. It was placed, lit, and left. Someone stood in that room long enough to smoke it down to nothing — and walked out without touching it again.

That is not what a man working a public corner — fifteen metres from the front gate, with two colleagues on the road and a foreman expected back from the depot at any minute — could hazard. Every minute he was inside that house, his absence from the corner was observable to anyone who looked.

A man who could not be recognised by the people on the road could stand in that room and smoke. A man who could be recognised could not.

*****

The fourth constraint. The killer left through the front door, in daylight, wearing what they had worn coming in.

A retired man crossed Dandenong Road at about ten minutes past midday. He saw a man come through the front gate of 575 Dandenong Road, pause, look back at the house, look both ways along Dandenong Road, and walk west, at an unhurried pace, toward Glenferrie Road.

The man he saw was wearing a service uniform. New-style RAAF summer dress. Light blue shirt, dark trousers, peaked cap. Not work clothes. Not council blues. Not a blue terry-towelling hat.

At approximately ten minutes past midday, James was at the council depot. He had returned for lunch in the truck driven by his workmate. By his own account, and the accounts of the other men at the depot, he was sitting in the meal room, having just eaten. Cards were being played. Banter was happening. He was there.

He was not at the gate of 575 Dandenong Road. He was not in RAAF uniform. He was not walking west.

There is no version of the timeline in which James can be both the man on the fence outside Mary Anne's house at twenty past eleven, and the man at the council depot at midday, and the man walking out of her front gate in service uniform at ten minutes past twelve.

The man at the gate is somebody else.

*****

There are other, smaller threads.

The bitumen flecks on the singlet. The Forensic Science Laboratory officer, asked at inquest, said the dark substance that had come off the deceased's singlet "appeared and behaved like" bitumen. The council crew were laying hot bitumen all morning. There was a tar pot on the public footpath outside the property. Dark particulate could find its way onto a household garment by a dozen ordinary routes. It did not require a killer to enter the bedroom.

The boot print at the rear gate. The print is industrial-soled. James wore industrial boots. He had also, on his own evidence, walked down that driveway with Mary Anne mid-morning to look at the silt, and walked down it a second time with the foreman. The print could have been left by him on either occasion. It could also have been left by another man wearing industrial-soled work boots — and the council crew on that corner was three men, not one.

No positive forensic match is recorded in the inquest brief. And even if a match had been made, James had already conceded — on the record — being in that driveway twice that morning. A print of his boot in that location was not, on the case as it stood, evidence of anything he had not already admitted.

*****

Constraints, though, are not the only test the case faces.

Mary Anne, when she was found, was wearing a wristwatch on her left wrist, three rings on her left ring finger, and one ring on her right ring finger. She had been wearing them earlier that morning when she stood in her driveway and spoke to James about silt. He had been close enough to her hands to see them.

Nothing on her body was taken.

The only item taken from the house was her handbag.

If the killer was James, his motive on the case the investigation built around him was either sex or money. A killer motivated by money does not leave four pieces of small, portable, sellable jewellery on a defenceless woman's body. A killer with the time to disconnect a phone and smoke a cigarette to the filter had the time to take a wedding ring.

He did not.

The handbag-but-not-the-rings pattern is wrong for a robbery, and wrong for a robbery with sexual elements added. A killer who came for what was in the bag had no reason to leave what was on her hands. A killer who came for what was on the hands had no reason to take the bag.

A simpler reading fits both observations.

The handbag was not taken for what was inside it. The handbag was taken because it could not stay.

In the course of the attack — in the bindings, the stabbing, the moving of the body — contact between the killer and the handbag is possible. We do not know what, if anything, would have transferred. Skin cells. A hair. A trace. Something the bag carried — on its surface or its strap or its lining — that the rest of the room did not. Foreign material from the man who killed her.

Taking the handbag may not have been the act of a robber. It may have been the act of a man removing the one item in the room that could later connect him to it. The cash, the eternity ring, the bank passbook, the keys — incidental. Possibly even useful, as misdirection toward a robbery narrative the police would treat seriously enough to follow.

That reading helps explain, in one move, both the absence of any identifiable foreign trace from the killer in the room and the absence of any clear motive in the case against James. He did not come for the bag. He left with the bag because by then the bag carried him.

*****

The knife was the same problem on a smaller scale.

It was long enough and sharp enough to deliver fourteen wounds, several of which penetrated the back wall of the lung and the wall of the left ventricle. It was not improvised. The kitchen knives at 575 Dandenong Road were checked. Collins Fagan stated that, to his knowledge, none were missing. The killer had brought a knife to the house and taken it away.

If James was the killer, he had to have carried that knife to the corner of Bailey Avenue and Dandenong Road that morning, in council blues, without his colleagues seeing it. He had to have kept it concealed through the morning's work. He had to have taken it into the house and out of the house. He had to have carried it back across a public corner, into a council truck, into a depot meal room, through the afternoon's bitumen work — and disposed of it, at some point that day or that evening, in a way that left no record.

His de facto wife, asked at inquest whether he ever carried a knife, gave evidence that she had never seen him with one. The single time she had given him a knife to take to work — a fruit knife, to cut a tomato at lunch — he had returned it the same day.

The handbag is the same problem at scale. It cannot be folded into a pocket. It cannot be slid under a council shirt. It cannot be carried across a public corner past three colleagues without being noticed.

The handbag has never been recovered. It is not at any of the places James demonstrably went that day.

A man in council blues, working a public road in summer, had no obvious way to remove a handbag from the scene unseen. A man whose hands could carry something without it looking out of place did.

*****

What is now known about alcohol and memory was not known in 1978.

James was a heavy drinker, by every account in the case file. His de facto told the Coroner so. He told the detective sergeant himself, partway through the interrogation: "My memory's not too good. Too much booze." He had been drinking the night before. He drank for six hours on the evening of the murder. He played pool, on his own evidence, until police arrived to interview him in the bar.

In 1978, this was understood broadly as a problem of unreliability. A man with drink in him forgot things, slurred things, contradicted himself. His evidence was suspect.

In 2026, the science is sharper. Alcohol disrupts the consolidation of new memories. It interferes with retrieval under stress. It compounds the effect of prolonged interrogation, where fatigue and pressure are themselves degrading recall. A heavy drinker, interviewed for multiple hours the day after a heavy session, may cycle through possibilities — not because he is lying about the truth, but because he cannot reliably locate it.

This does not explain the lie about the silt conversation, which is a small, vivid, recent event lied about for an obvious reason. It does explain some of the cycling on the money and the bookmaker — events of an ordinary working morning, blurred by alcohol, retrieved under pressure, asked about for hours.

It also creates a different problem.

A man whose recall of the previous Friday morning was as compromised as the depositions describe is not, in any easy theoretical move, the same man who is then said to have planned and executed a killing with forensic discipline. The memory that lost the bookmaker is the memory that would have had to hold, in clean sequence, the order in which the door was opened, the binding cut, the cigarette burned to the filter, the handbag taken, the route of departure chosen.

A man whose recall of his own working morning could not be relied on does not, easily, become the man who managed the operational facts of a murder in the same period.

*****

Two witnesses sat across the case from each other.

The first was Con, the SP bookmaker. He gave evidence at inquest that he had never paid out a bet to anyone before five o'clock on any day, and that he had not paid James on the morning of the seventeenth of February.

The second was Ken, the man on the fence. He gave evidence that at about a quarter past eleven, James had returned from the direction of the hotel with a roll of notes in his hand, had shown them to him, and had asked him to check the day's paper for the price the horse had paid. They had looked at the paper together. The price was six to four. Ken estimated three tens visible in the bundle, possibly more.

The investigation, in the months that followed, treated the bookmaker's denial as the more reliable account.

That choice has a structural problem.

The bookmaker was operating an illegal trade. SP — starting price — bookmaking was a criminal offence in Victoria in 1978. The Gaming Squad had visited him. A bookmaker who admitted, on oath, to paying out cash before five o'clock at his home, on a working morning, to a worker who had walked there from his shift — was admitting to operational facts that strengthened the case against him as an SP operator. A bookmaker who denied it lost nothing he had not already lost.

"He probably doesn't want to admit he's an SP," was what James himself said about him, on the record, during the interrogation. It is the one observation in the seventy pages that is worth taking seriously on its own logic.

The witness on the other side had no such structural reason to lie. Ken had known James for a few months at most. There was no friendship to protect. In the first weeks of the investigation, Ken was a suspect alongside James. If he was protecting himself, the move was to distance from James, not to corroborate him. To say: I don't know what he did during the gap. I don't know where the money came from. He didn't show me anything.

He gave the opposite account. He placed James away from the corner during the gap. He placed cash in James's hand on his return. He walked through the paper with him to confirm the price. The detail he gave — three tens visible, six-to-four well-backed-in, the bookmaker named — was specific in a way that an invented corroboration would not need to be.

The investigation's preferred witness was the one with the motive to lie. The witness with no motive to lie said something else.

*****

What survives of the case against James, once it has been tested, is this.

The lies were real.

He had spoken to Mary Anne about the silt and concealed it. He had walked into her driveway and concealed that. He had left the worksite alone for forty-five minutes and concealed that as well. He had sat on a fence opposite her house and said vile things about her to a man he barely knew. He had returned with cash he could not consistently account for. He had said his bookmaker had paid him out, when the bookmaker said, on oath, that he had not.

These were not the lies of a man being open and forthcoming. They were the lies of a man concealing several things that were independently worth hiding. Illegal starting-price gambling. Unauthorised absence from a council worksite. A private quote, mentioned to nobody, that would have given him a pretext to return to the Fagan property on his own time. A graphic, sexually explicit conversation about a woman he had spoken to that morning. A reputation for sexual bravado dressed up around expressions a woman at home had never heard him use.

Each had its own internal logic. Each, on its own, explained the concealment of that thing — without needing to involve a murder.

A man protecting a murder, by contrast, has one core lie. The lie that he was not there. The lie that has to hold under sustained pressure, because the alternative is unthinkable.

James's lies were not that. They were a tangle of small shameful secrets, each separately defensible, none of them load-bearing for the question of murder. They collapsed in the order pressure was applied to them, and they collapsed quickly — which is also not the pattern of a single great lie.

Step back from the lies for a moment, and look at the man behind them.

The bet he collected on the morning of the seventeenth was illegal. SP bookmaking was a Gaming Squad offence. Collecting a payout on council time was an additional offence against his employer. If James wanted to keep one thing quiet that morning, the bet was it.

He told Ken. He showed Ken the cash and asked him to check the starting price in the paper. He named the bookmaker. He named the hotel. By the time the detective sergeant sat across from him in April, half of working Armadale knew James had been to Con’s on Friday morning.

This is not a man who could hold a small secret for half a day.

And we are asked to believe he held the largest possible secret from the seventeenth of February 1978 until his death, in the 1990s. That he sat through, by his count, sixteen or seventeen hours of questioning, watched the bookie story collapse, watched the money story cycle four times, watched the sexual conversation drag out of him line by line — and held, behind all of it, in perfect silence, the one secret that mattered.

A man whose nature was to talk. Who talked his way into every problem he was in that day. Who could not stop talking even when talking was costing him.

That man does not, on the 17th of February 1978, walk out of a forensically silent crime scene and never speak of it again.

The lies came apart. The murder did not attach to them.

And every test of the case against him collapses in the same direction. None of them places him in the room.

*****

A council worker who had spoken to Mary Anne that morning had been treated, for two months and through the inquest the following year, as the likely killer. Across the same period, in the institutional view that took over the investigation in April 1978, the man at the gate had been almost certainly the husband on a different day.

The investigation had followed the thread the constraints could not carry.

Now turn that around.

Take the man at the gate, and put him through the same constraints.

*****

The first constraint. The killer entered the house without forcing entry.

A man in RAAF uniform, at the front door of a Group Captain's house, while the Group Captain was away overnight at base, is not the same as a man in work clothes at the same door.

Mary Anne was a service wife. She had been a service wife for seventeen years. Her husband was the Commanding Officer of the Number 1 Stores Depot at Tottenham. He had stayed overnight at the base. He had, the night before, attended a function at the Sergeants' Mess.

If a man came to the door of that house in service uniform, while she was halfway through dyeing her hair, the meaning of that man at that door was not social. It was institutional. A service uniform at the door of a service household, while the serving member is away at base, signals an emergency. An accident. A notification. Something that requires the door to be opened immediately, regardless of what the woman inside happens to be doing at the time.

You don't finish your hair. You don't check through the screen. You walk down the hall and you open the door.

The mechanism that the constraints require — something that overrode every normal instinct about privacy and appearance — is in this case named in the institutional context of who Mary Anne was.

The uniform was not just clothing. It was a key.

*****

What happens next, on this reading, is not a confrontation.

It is a conversation. The man on the doorstep tells Mary Anne, in some form, that something has happened to her husband at the base. She is listening. She is not yet aware of any threat. She walks back from the door, into the hallway, into the house. The nearest room is the front bedroom — perhaps twelve feet from the door. By the time the framing of the encounter collapses, she is already inside it.

You do not need to coerce someone very forcefully to move twelve feet if the premise of the encounter has not yet collapsed. The room entry can happen before she fully understands the uniform is a deception. That is compliance through confusion, not compliance through force. It fits the relative order of the scene — no signs of serious struggle until the bed itself — better than the official reconstruction.

Patrick, the youngest of Mary Anne's five children, was seventeen months old. He was at home that morning. He was found unharmed. Patrick's presence may have mattered. He was not a witness in the ordinary sense, but his presence in the house may have affected Mary Anne's choices and the offender's control of the situation. The record cannot take us further than that.

*****

The phone in the hallway was found disconnected when police arrived. The history of the phone that afternoon is not perfectly clean — Anthony, the eldest son, spoke to his father from the house after the children found Mary Anne, on a line that was working — and the disconnection may have happened at the killer's hand, or in the chaos that followed. The evidence does not let us be certain.

What can be said is this: if the uniform meant what it appeared to mean, then the failure of the connection may not have been external at all. It may have sat inside the system itself. If the man at the door has just told a service wife that something has happened to her husband, the next thing that wife is going to do, after he has come inside, is ring the base. Not for confirmation — for detail. To find out what has happened. To know what to do next.

One call to the Tottenham switchboard, and the deception collapses. Collins is alive. Collins is in his office. Collins is well. No notification has been issued. The man in the house is fraudulent.

A killer wearing that uniform had reason to pull the phone before the call could be made. A council labourer at the door did not. If the disconnection was the killer's act, it is consistent with the deception. If it was not, it is consistent with nothing in particular.

The mechanism — uniform at the door, the call that the uniform would naturally suggest — does not depend on the phone for its coherence. It requires a person who knew enough about Collins Fagan's service, his rank, his workplace, and his routine to construct that deception in the first place.

*****

The second constraint. No blood distinguishable from Mary Anne's was found at the scene.

This is harder to read in the uniform thread than in the council-worker thread, because the man at the gate has not been identified. We do not have his clothes. We do not have his person. We cannot examine him for transfer.

What we can say is this. The man who walked out of the gate at ten past twelve was described by the witness as wearing a uniform that appeared ruffled, and not neat. He was not described as bloodied. He was not described as soaked. He was described as a man who had stopped at a gate to compose himself before walking away.

A killer who stood behind a woman lying face down and stabbed her fourteen times in the back could, with care, avoid the worst of the spatter. The wounds were inflicted from one side, into a body that had been bound and was not free to move. A killer who took five to seven minutes to smoke a cigarette in the room afterwards had time and composure to assess himself before opening the front door.

We do not know what was on his shirt. We never will. The shirt has never been recovered.

We do know that he then disappeared. He did not return to a worksite. He did not return to a hotel. He did not stand at a bar in his blue uniform that night. He hitched a lift, said "anywhere", refused to converse, got out short of Williams Road, and was never seen again.

A man who could disappear had the option to dispose of clothing that could not be disposed of by a man who had to go back to work.

That option is structural. It is built into the difference between someone whose presence on the corner that morning was witnessed by his workmates, and someone whose presence on the corner was witnessed by nobody other than the woman he had come to deceive — and who was now dead.

*****

The third constraint. The killer smoked a cigarette in the room.

The mantelpiece butt was identified as likely a Winfield. We do not know what the man at the gate smoked. He has never been identified.

What we can say is that the act of smoking a cigarette to the filter while standing in a room containing the body of a woman he has just killed is the act of someone whose presence in that room is not, at that moment, rushed. A man who has come in through the front door, done what he came to do, and is now standing in the bedroom long enough to finish a cigarette is a man composing himself for the next stage of his departure.

A worker fifteen metres from the corner does not have five to seven minutes to stand still in someone else's bedroom. A worker has to be back at the corner before he is missed. A worker has to explain himself when his workmates return from the depot or the laneway.

A man who is leaving on foot, and who does not have to be back anywhere by any particular time, can stand in the room for as long as it takes him to think about how he is going to walk out of the gate.

The disciplined departure that the witness described — the pause at the gate, the look back, the look both ways, the unhurried walk west — is consistent with the disciplined cigarette inside the room.

They are both behaviours of a man with time.

*****

The fourth constraint. The killer left through the front door, in daylight, wearing what they had worn coming in.

Two witnesses. Two independent sightings. Thirty-five minutes apart. Both describing a man in military-style uniform. Walking on foot. Heading west.

The retired man at the gate had walked that stretch of Dandenong Road for three years and had never seen an adult male at the address. The hitchhiker pickup, thirty-five minutes and roughly a kilometre and a half further west, was made by a stranger who would later, at inquest, say that the belt he was shown — a Royal Australian Air Force belt — looked very familiar.

For two unrelated witnesses to converge on this kind of alignment by coincidence would require an exceptional set of accidents. The simpler explanation is that they saw the same man.

That man fits the constraint of the front-door departure precisely. He left the way he came in. He walked west on the road he had walked east. He did not change. He did not double back. He did not run.

He composed himself at the gate and he walked.

*****

One assumption of the official theory has to be tested before the analysis closes.

The murder window the investigation worked from — between roughly half past ten and half past eleven on the morning of the seventeenth — was built around two things. The pathologist's estimate of time of death. And the report of female screams heard by neighbouring residents on the day.

The pathologist examined Mary Anne at the scene that evening. He gave the time of death as "well in excess of two or three hours" before his examination. Working backwards, that places the killing before about three in the afternoon. It does not place it before noon. A time of death around noon is fully consistent with the pathology.

The screams report is the harder problem.

Two witnesses gave evidence of female screams near the Fagan property. A builder working at the back of the Fagan house heard what he called a muffled scream, sometime between one and half past one in the afternoon. A resident in the block of flats across Dandenong Road heard a louder female scream about half an hour later. Neither witness wore a watch. The flat resident gave her time as two o'clock in her first statement, corrected it to half past one before the inquest, and described her own state at the time as nervous and on medication.

What neither witness could have heard, on the evidence of the scene, is Mary Anne. If the sound they heard was Mary Anne, then the timing attached to that sound cannot be right. By the official window, she had been dead for more than an hour before either scream was heard. By any reading, she was bound at the wrists and ankles with strips of her own towel, and had a piece of striped cloth forced into her mouth and tied across her face. The medical evidence does not support a long, half-conscious dying period that would have allowed her to scream intermittently into the afternoon.

And the closest people to the source — the workmen back on the corner from a few minutes before one — heard nothing at all.

Whatever the two witnesses heard, those screams cannot safely anchor the murder window. They may matter. But they cannot do the work the investigation later made them do.

What remains as a foundation for the official morning window is the pathologist, who allows for a noon time of death as easily as for ten thirty. The murder may have happened in the morning. It may also have happened closer to noon — at, or close to, the moment the man at the gate was seen leaving.

*****

The coronial file acknowledges both threads in handwritten notes.

In the working papers of the case file, in the same hand that wrote statements true, subsequent statements, had not told the police the truth about James, a separate page records, in the same hand:

Evidence of two witnesses as to someone in uniform seen in location.

The investigation knew. The investigation had both threads on the table. The investigation had, in the first months of 1978, treated the uniform sighting as the prime line of enquiry — the head of the Homicide Squad, Chief Inspector Noel Jubb, telling Melbourne newspapers on the 3rd of March 1978 that police were "ninety-five per cent certain it was an RAAF uniform", that the man "may have worn a RAAF uniform to lull Mrs Fagan into opening the door", and that the sighting was "clearly the best lead we have so far".

That was the head of homicide in the initial stages of the investigation.

He was replaced in April 1978.

A senior investigator connected to the case, asked about the case twenty-three years later, gave a different view of the sighting. He said it was almost certainly the husband on a different day. That all the evidence was that Mrs Fagan was a very moral woman, and that there was no reason it would be anybody else at the house.

That view came to dominate the institutional position on the case. It came to dominate it for the next four decades. And from April 1978 onwards, the investigation's attention shifted from the man at the gate to the man at the corner.

The retired man, taken at inquest to look at the husband across the courtroom, had said: "I would say a similarity. But I would say no."

The witness himself had ruled it out.

The investigation, in its own working file, had recorded both threads.

And the door that opened between them — the institutional door, in April 1978 — closed against the one that fitted the constraints.

*****

It is fair to acknowledge one further possibility.

The two threads tested in this episode are the threads the investigation actually pursued — the man at the corner, and the man at the gate. The Coroner's finding, in June 1979, was open. Murder, by a person unknown. That phrase, in its plain meaning, does not pick out either thread. It leaves space for a third.

Test that against the constraints.

To get Mary Anne to open the door, a third actor would need a mechanism. The only documented mechanism on the public record is the uniform. The forensic discipline required is the same. The departure pattern — observed by two independent witnesses on Dandenong Road — has to be explained. If a third actor departed by some other route, at some other time, in different clothing, then either the man the two witnesses saw was unconnected to the murder, or two people were on the property in the same window without crossing paths, and only one of them was seen leaving.

Neither is impossible. Both stack coincidences.

A third possibility remains on the table because the Coroner's finding leaves it on the table. It is thinner than the uniform thread on the public record. A third actor has to be built almost entirely from absence. The uniform lead is built from sightings, timing, clothing, direction of travel, and institutional context. It exists in the space between the two named threads.

The man at the gate could also have been a real person who had nothing to do with what happened inside the house. He could have been on the property for an unrelated reason, walking past, approaching and being turned away. This is structurally possible. It requires only that we accept a coincidence — a service uniform at the door of a service household, on a working morning, on the day of the murder, observed independently by two witnesses, the first of whom watched him pause at the gate, look back at the house, look both ways along Dandenong Road, and walk west.

The behaviour described is the behaviour of someone leaving a place they have just been — deliberately, with reason to check whether they have been seen. An uninvolved man — a man whose presence on that corner was innocent — had every reason to come forward when the case was reported in the days that followed. Forty-eight years on, no such man has.

The coincidence reading cannot be ruled out. It can only be ranked.

On the available evidence, it ranks behind the readings that put the uniform inside the house.

*****

When the investigation made its choice in April 1978, it picked the thread the constraints could not carry. It picked the man whose lies were loud, whose contradictions were visible, whose presence on the corner could be measured to the minute, and whose deception was systematic enough to feel, in interrogation, like a confession waiting to come apart.

The lies came apart.

The murder did not attach to them.

The man at the gate, in the same period, walked into the institutional shadow of almost certainly the husband on a different day, and stayed there for forty-eight years. The witness who saw him did not flinch when shown the husband at the inquest. "I would say a similarity. But I would say no." He had not been wrong.

The investigation had followed the thread the constraints could not carry.

*****

The man at the gate has been almost completely silent for forty-eight years.

He has not come forward. He has not, on the public record, been named. The photofit produced from the witness's description was published in newspapers across Victoria, and in other states, and is still on file. He did not match anyone at any of the Victorian RAAF parades the witness was taken to. He did not match anyone the Royal Australian Air Force police could produce. He did not match anyone identified through the checks made at Air Force supply stores.

That does not eliminate him from RAAF connection. It eliminates him from a category — the category of a serving RAAF member who could be paraded.

One category remained capable of explaining why the RAAF checks failed: a former or discharged airman who still had access to a uniform. A man who had been discharged from the Air Force — for any reason, in any era — would still possess his uniform. He would still know how to wear it. He would still know what effect it would have at the door of a Group Captain's house. He would not appear in any parade. He would not be identified through any check at any supply store. He would, by every measure the investigation took, have been invisible.

This is not to name anyone. It is to say that one category capable of explaining the failed checks is a category the investigation, in 1978, did not look in.

It is also to say that the silence of forty-eight years is itself a kind of evidence. Seventy pages of lies told by James were the evidence the investigation acted on. The complete absence of any equivalent talk from the man at the gate has been treated, institutionally, as if it meant he wasn't there.

The opposite is also possible.

*****

There is one further question this analysis has not asked.

The investigation, in its own summary, recorded that enquiries into Mary Anne's background had found nothing. No association with another man. No grievance from anyone. A normal and happily married woman.

That is a finding the investigation made. It is also a question, asked in only one direction.

There is another direction to ask it from. The next episode walks it.

*****

In the next and final episode, we step back from the threads, and ask what could still close this case.

Twenty-seven items of physical evidence were collected. Some of them may matter differently now than they did in 1978 — including fibres found in Mary Anne's closed hand.

The handbag has never been recovered. The weapon has never been recovered. The man at the gate has never been named.

But the conditions of the case are clear. The constraints have not weakened with time. They have sharpened.

Someone, somewhere, knows the answer. Someone may have known it for forty-eight years. Someone may have suspected, for forty-eight years, that the answer they carry — the conversation they overheard, the absence they noticed, the uniform they once saw hanging in a wardrobe — matters.

Mary Anne Fagan was a forty-one-year-old mother of five, on an ordinary Friday morning in February, dyeing her hair for a function the next night. She opened her front door to someone she believed she had to.

She never washed the dye out.

Her children came home from school and found her.

For forty-eight years, those children — now adults — have been waiting to know who walked into their house that day.

The question Episode 6 will sit with is whether forty-eight years is enough — and what it would take, finally, for the answer to come back.

If you have information that could assist police in relation to the murder of Mary Anne Fagan, please contact Victoria Police Crime Stoppers on 1800 333 000, or submit a confidential report online at police.vic.gov.au/crime-stoppers.

If you served with the Royal Australian Air Force in Victoria in the 1970s, and you recall a colleague — a former colleague, a discharged member, anyone connected to the service — who you have wondered about for years, that information may matter.

If you were in the Armadale, Malvern, or Toorak area on Friday the 17th of February 1978, and you saw something on the corner of Bailey Avenue and Dandenong Road, or on Dandenong Road heading west between midday and one o'clock, please come forward.

Some names and non-essential identifying details have been changed for privacy. All core events and timings are drawn from publicly available records, inquest materials, and contemporaneous reporting.

Unsolved. Unforgotten. Unfinished.