The Zodiac Killer wrote to newspapers, and because a threatening letter in the American mail crosses a state line, the FBI opened a file on him. That file is public: seven parts, seven hundred and forty-nine pages, and it never writes down a name.
Between December nineteen sixty-eight and October nineteen sixty-nine, five people were murdered in the San Francisco Bay Area and two survived being attacked. The man who did it was never identified. No one has ever been charged. That is the part everybody knows, and it is where almost every account of this case starts and ends: a list of suspects, and an argument about which of them it was.
This film is about something else. It is about the paper. Because this man did something the others did not. He wrote to newspapers.
He mailed letters and cards and, twice, he mailed a cipher - a message written in symbols instead of letters, which has to be broken before it can be read. And the moment a threatening letter crosses a state line in the American mail, it stops being only a local murder case and becomes a federal matter. So the FBI opened a file. Not a murder file - the murders belonged to the county sheriffs and the city police departments in California.
What the Bureau opened was a laboratory file: a record of objects sent in, examined, and reported on. That file is public. Seven parts, seven hundred and forty-nine pages. The Bureau file number is nine dash four nine nine one one.
The latent fingerprint case - latent meaning an invisible mark left by skin, brought up with powder - carries its own number, A dash one zero zero four two. The subject line typed at the top reads: ZODIAC, semicolon, EXTORTION. Nobody in that file ever writes down a name for him. What they write down is exhibit numbers.
It starts with two teenagers on a road outside Benicia, California. On the twentieth of December, nineteen sixty-eight,
David Faraday, seventeen, and Betty Lou Jensen, sixteen, parked on Lake Herman Road. It was a place local couples went. A man came to the car and shot them both with a twenty-two calibre pistol. The file records one detail about the weapon that matters more than the calibre.
He had taped a pencil flashlight to the barrel. A pencil flashlight is exactly what it sounds like - a torch the size and shape of a pen. Taped alongside a gun barrel, it does one job: it lets you see what you are aiming at in the dark, without holding a torch in your other hand. That is not a detail about a weapon.
It is a detail about preparation. Both of them died. There was no letter, no cipher, no name. For seven months this was simply an unsolved double murder in Solano County, and there was nothing federal about it at all.
On the night of the fourth of July, nineteen sixty-nine, a car was parked at Blue Rock Springs Park in Vallejo. Darlene Ferrin, twenty-two, was in it with Michael Mageau.
A second car pulled in, sat there, and drove away. It came back. A man walked up to the passenger side with a flashlight and fired a nine millimetre pistol into the car. Darlene Ferrin died.
Michael Mageau was hit and lived. The park is less than two miles from Lake Herman Road. About forty minutes later, a man telephoned the Vallejo Police Department, reported the shooting, gave its location, and then claimed the murders on Lake Herman Road as well. Then he hung up.
Up to this point everything in this case is ordinary police work: two scenes, two calibres, one survivor, one anonymous call. There is nothing to send to Washington. That changes at the end of the month. On the thirty-first of July, nineteen sixty-nine, three newspapers received mail.
The Vallejo Times-Herald, the San Francisco Chronicle and the San Francisco Examiner each got an envelope. The Bureau's laboratory worksheet describes one of them exactly
as it arrived: postmarked San Francisco, California, thirty-one July nineteen sixty-nine, hand printed address, S.F. Chronicle, San Francisco, California - and then the words that were written twice, once on the front and once on the flap: Please Rush to Editor. Inside each envelope was a two-page hand printed letter. The Laboratory records its first line too.
Dear Editor. I am the killer of the two teenagers last Christmas. The letter then set out details of both shootings - the position of the car, the number of shots, the brand of ammunition - things that had not been printed anywhere. And with each letter came one third of a cipher.
Four hundred and eight symbols, divided in three, one part to each paper, with an instruction: print it on the front page, or he would go on a weekend of killing. The papers printed it. And the police now had a piece of paper that a laboratory could work on. This is the moment the case becomes federal, and it is also the moment the case becomes strange, because of what happens next.
The cipher went to the FBI. It also went, by way of the front page, to about a million readers. And it was not the Bureau that broke it. On the ninth of August, nineteen sixty-nine,
the Vallejo Times-Herald ran an article. The Sacramento field office cut it out and posted it to Washington, and it sits in part one of the file today. It reports that Donald G. Harden, a high school teacher in Salinas, California, claims to have broken the cipher.
He had worked on it with his wife, Bettye. It had taken them about twenty hours. Their method was a guess about vanity: that a man who writes to newspapers about killing people would probably begin by saying so. They tried the letters of "I like killing people" against the opening symbols.
It fitted. The decoded message runs for several paragraphs. It says that killing people is more fun than killing wild game in the forest, because man is the most dangerous animal of all. It says the people he kills will become his slaves in paradise.
It is misspelled throughout. And at the end, where a signature would go, there are eighteen leftover symbols that do not resolve into anything. The message does not contain a name. The Bureau added it to the file as an exhibit and carried on numbering.
Seven weeks later, on the twenty-seventh of September, nineteen sixty-nine, Bryan Hartnell, twenty, and Cecelia Ann Shepard, twenty-two, were sitting by the water at Lake Berryessa in Napa County. A man walked up to them wearing a black hood that covered his head, with a symbol on the front of it - a circle with a cross through it. He had a gun. He told them he was an escaped convict and wanted their car and their money.
He gave Shepard lengths of cord and told her to tie Hartnell up, and then he tied her. Then he took out a knife and stabbed them both. Hartnell ten times. Shepard twenty-four.
He walked back to their car and wrote on the door in felt pen: the same circle and cross, and under it the dates of the two previous attacks and this one. Then he drove to a phone box in Napa and reported the stabbing himself. Hartnell survived. Cecelia Shepard died two days later.
The Bureau's caption for the fingerprint work on this one is not a name. It reads: unknown victims, Lake Berryessa area, Napa County, California, September twenty-seventh, nineteen sixty-nine, assault. That is filed as latent case A dash one two three nine seven, alongside the first. Two weeks after that, on the eleventh of October, nineteen sixty-nine, a man got into a taxi in downtown San Francisco.
The driver was Paul Stine, twenty-nine, a doctoral student driving cabs at night. They drove to Presidio Heights, a residential district of large houses. At the corner of Washington and Cherry Street the passenger shot Stine once in the head. Then he did something the previous scenes had not involved.
He tore away a section of Stine's shirt. Three teenagers watching from a window across the street called the police while it was happening, and a patrol car passed the man walking away.
Within minutes police radios were describing him. Two days later the San Francisco Chronicle received an envelope. Inside was a letter, and with the letter, a piece of bloodstained cloth cut from Paul Stine's shirt. A murder in a taxi had just become a laboratory exhibit.
From here the file stops being about California and becomes about paper. Every item is given a number. Q for a questioned document - one whose origin is what is being tested. Qc where the Bureau held a photocopy rather than the original, because the original stayed with the local police force that received it.
Q one. Q eight. Qc one, Qc two. Qc nine and Qc ten - the two-page letter beginning Dear Editor, I am the killer of the two teenagers.
Qc twenty-one and Qc twenty-two - an envelope postmarked the eighth of November nineteen sixty-nine, addressed to the Chronicle, Please Rush to Editor, and the six-page letter inside it beginning This is the Zodiac speaking.
Qc forty-six, Qc sixty-two, Qc sixty-three - and on, for years. That phrase, This is the Zodiac speaking, is where the name comes from. He gave it to himself in a letter in August nineteen sixty-nine, and the newspapers used it because he had told them to. The Laboratory tested what could be tested.
Handwriting, against samples sent in by field offices from suspects all over the country. Latent fingerprints, lifted from the letters and the envelopes. Paper stock. Postmarks.
Twenty-two pages of this file are handwriting comparisons: a field office writes in with a suspect's known writing, the Laboratory compares it against the letters, and answers. The answers are all the same kind of answer. Here is one of them, from an examination of two questioned items: it was not determined whether Qc one and Qc two were prepared by the writer of the letters in the case, because of a lack of comparable material. Not "he did not write it".
Not "he did". It was not determined. That sentence, in different wordings, is most of what a hundred and fifty pages of laboratory reports in this file actually say. The fingerprints have the same shape of answer.
There were prints - from the letters, from the envelopes, from the taxi. At one point the file contains a field office asking Washington the obvious question: how many latent prints believed to be the Zodiac's are available, and is there any possibility of classifying them well enough to permit a search through the FBI's fingerprint files. That question is worth stopping on, because of what it assumes. In nineteen sixty-nine the Bureau's fingerprint collection held tens of millions of cards, and it was searched by hand, using a classification system that turns the pattern of a print into a code.
To search it, you need enough of a print to code it. A partial smudge lifted from an envelope cannot be coded, so it cannot be searched - it can only be compared, one suspect at a time, after somebody has already named that suspect. Which means the prints in this case could never find him. They could only ever confirm or eliminate a man that somebody else had already pointed at.
The Bureau was extremely good at telling the difference between what it had established and what it had not, and it wrote the difference down every time. In his letters, the man kept score. He would write in with a running total of how many people he had killed, and the total climbed into the dozens - far past anything any police force could connect to him. Police attributed five murders and two survivors.
His letters claimed many times that. There is no way to test a claim like that, and the file does not try. What it does hold is the threat that came with the boasting. He wrote that he would attack a school bus - shoot out the front tyre and pick off the children as they came out - and he drew a diagram of a bomb wired to a clock and a photoelectric cell, to be buried by a road.
The bus attack never happened. The bomb was never found. But the Bureau circulated the diagram, because a diagram from a man who had already killed people is not a joke. The file also holds a question that has never closed.
In October nineteen sixty-six, more than four hundred miles south, a young woman named Cheri Jo Bates was murdered outside the library at Riverside City College, and letters were later sent about that killing too. Some investigators believed it was the same man. The file notes the geography plainly: all of the other attacks fell inside the Sacramento and San Francisco divisions of the Bureau. Riverside did not.
It is still argued about. The file does not settle it, and neither will this. The killing stopped. The mail did not.
For the next two years he kept writing, and the file keeps numbering. Q ninety-two, a five-page letter posted the twenty-sixth of July nineteen seventy. Q ninety-three, a Halloween card posted the twenty-seventh of October nineteen seventy - the Laboratory's description records the envelope in his hand printing, addressed S.F. Chronicle, Fifth and Mission, San Francisco, and the card inside beginning with the words By Fire.
There is a postcard claiming a murder in the Lake Tahoe area. There are letters to a San Francisco lawyer, Melvin Belli. And there is one episode in October nineteen sixty-nine that shows what the case had become. A man telephoned a live television talk show in San Francisco, on air, claiming to be the Zodiac, and asked to speak to Belli.
The Bureau's own note on this is two sentences long and completely deadpan: a man claiming to be the killer spoke on a television
news programme to the attorney Melvin Belli, and the Bureau telephonically notified San Francisco to conduct no investigation in the absence of a request from local authorities. Somebody claiming to be the most wanted man in California was on live television, and the correct federal answer was: not our case, do nothing. That is not indifference. It is jurisdiction.
The Bureau's authority here ran to the mail and the laboratory, and not one inch further - and the file is scrupulous about the line, over and over, for twenty years. The second cipher was mailed to the San Francisco Chronicle on the eighth of November, nineteen sixty-nine. Three hundred and forty symbols, on a card, in the same hand printing. Nobody could break it.
Not the Bureau, not the Navy, not the National Security Agency, not fifty years of amateurs. It became the most famous unsolved cipher in the world, and the reason was that it was not built the way people assumed. The first cipher had been a straight substitution - one symbol for one letter, several symbols per letter to defeat frequency counting. The second used the same trick and then did something else on top of it: the text was written into a grid and read out diagonally, in knight's-move steps, so that even a correct substitution produced nonsense unless the reading order was recovered at the same time.
On the fifth of December, twenty twenty, three men submitted a solution: David Oranchak, a software developer in Virginia; Jarl Van Eycke, a Belgian warehouse worker who wrote the code-breaking program; and Sam Blake, an Australian mathematician who worked out how many different reading orders were possible - about six hundred and
fifty thousand - and had the computer try them. The FBI's San Francisco field office confirmed the solution six days later. What it says is this. He hopes they are having lots of fun trying to catch him.
He was not the man who had gone on a local television programme claiming to be him. He is not afraid of the gas chamber, because it will send him to paradise all the sooner, where he now has enough slaves to work for him. Fifty-one years of work, by cryptographers and code-breaking software and three men who did it in their spare time, and the message on the other side is a taunt and a boast. It does not contain a name either.
And the file does not end in nineteen sixty-nine. Teletypes in part six of it are dated January nineteen ninety-one - twenty-two years after the last confirmed attack - still asking the Identification Division for latent fingerprint work under case number A dash one zero zero four two. Here is everything that physically survives of the Zodiac. Letters and cards and envelopes, hand printed, held in evidence by three California police departments and photographed into a federal file.
A piece of a murdered man's shirt. Handwriting that has been compared against hundreds of suspects and matched none of them. Latent fingerprints that have identified nobody. Two ciphers, four hundred and eight symbols and three hundred and forty symbols, both now solved, containing between them a description of why he did it and no indication at all of who he was.
That is the case. Not a mystery about a man - a mystery about a stack of paper, which has been examined by the best laboratory in the world, repeatedly, for over fifty years, and which has given up everything it has to give except the one thing everybody wants. The people who broke the first cipher were a schoolteacher and his wife, in about twenty hours, because they guessed he would start by bragging. The people who broke the second were three strangers on the internet, fifty-one years later, because one of them worked out that the message had been folded before it was encrypted.
Both times, the man had written down everything about himself except his name. He chose what to put on the paper. That is the only part of this he ever controlled, and he controlled it completely.





