On the morning of August 26, 2016, in the city of Baiyin, in the northwestern province of Gansu, a man of fifty-two stood behind the counter of a little shop inside a technical school, selling cigarettes and bottled water and instant noodles.
Plainclothes officers came in and put him against the shelves.
He did not run. He did not shout.
Two days later he was in custody. The detention centre fixed a monitor to his cell wall and watched him around the clock. In the interrogations that followed, he walked through the things that had begun twenty-eight years earlier, one by one.
A shopkeeper in a school. It is the kind of detail that makes a Chinese true crime story memorable. But the part of this case that changed Chinese policing is not the man. It is a blood sample taken from somebody else, a relative he barely knew, for a completely unrelated reason.
One. The first one, and the fourteen years
On May 26, 1988, a young woman was killed in her own home in Baiyin.
That was the first. Over the next fourteen years the same city produced the same kind of case eleven times. Every victim was female. The youngest was eight years old. The killings spread from Gansu into Baotou, in Inner Mongolia.
In Baiyin in those years there were almost no surveillance cameras on the streets, and witnesses were almost never found. The only witness at each scene was the scene itself. The killer left something behind every time: fingerprints, biological traces. Sometimes he did not even bother to wipe them. But a trace on a shelf is not a person, and the traces piled up without ever resolving into a face.
Forensics in that era was still done by eye and by hand. DNA testing could establish blood type and little more. Lifting a print was not simple either. A print was taken up from a surface, fixed to a card, numbered and filed, and the only way to check it against a particular man was for a detective to sit down with a magnifying glass and follow the ridges, one print at a time.
Because everyone understood how valuable such material was, it was handled with great care. Print cards and blood samples were numbered one by one, boxed, and moved into evidence stores. Nobody could say on which day any of it would become useful. They knew only that it could not be thrown away.
The police built a profile. Male. Rigid in method. Bold. Working alone. They narrowed the field to men born between 1966 and 1974, and later widened it to 1958 through 1975. They went through the men of the city in waves, hundreds of thousands of them in total. Fingerprints were compared by detectives with magnifying glasses. Over twenty-eight years they compared more than a hundred thousand prints.
Two. A hunt that could not close
In 2001 the series was listed as a case under the direct supervision of the Ministry of Public Security. In 2004 the Baiyin police released the details to the public and offered a reward of two hundred thousand yuan.
The reward produced leads. It did not produce a man.
For several years life in Baiyin was altered. At the worst of it, women did not dare travel to work or to school alone and had to be accompanied by a man from the family. An unsolved case had reshaped the ordinary routine of a whole city, and then, for more than a decade, it went quiet.
The detectives who worked this case did not have an easy decade. After the early killings, nearly every police officer, armed police unit and neighbourhood patrol in the city was put on the streets in three shifts, around the clock. One officer who worked the case later said that about two hundred and fifty people in the local force had been pulled into it at some point, and that all of them carried it with them, on duty and off. Meanwhile the public was asking, loudly, whether the police were being paid to do nothing.
Under that pressure the Baiyin public security bureau held a rally at which a hundred officers stood in a hall and shouted a pledge. Then they went back to comparing fingerprints.
And then, after February 2002, the killing stopped.
A man who had killed across fourteen years simply went quiet. Nobody could tell whether that was good news or worse news. Perhaps he was dead. Perhaps he had moved away. Perhaps he was still there in the city, living quietly among the people looking for him, and had only stopped.
The case settled into the cold file. But the evidence did not disappear. Prints, blood samples and other traces were catalogued, sealed and stored, waiting for the technology to catch up with them.
The wait lasted more than a decade. The files were moved from one building to another. One generation of investigators replaced the next, and in the Baiyin force the first detectives who had walked those scenes were reaching retirement age. The case was never closed. It was simply no longer discussed.
Three. One blood sample, one surname
The year is 2016. In March, the Gansu provincial department of public security decided to reopen the investigation.
The break came from somewhere else entirely, and it came by accident.
A man surnamed Gao had been placed under residential surveillance in a bribery matter. As part of the procedure, officers took a sample of his blood. That sample went into the database and was put through a test that was by then well established but had not yet been used at scale in Chinese homicide work: Y chromosome testing.
The Y chromosome passes down the male line only. A man gives it to his sons, and they give it to theirs, with no contribution from the mother. That means every man descended from the same paternal ancestor carries nearly the same set of markers at particular points along that chromosome. So a Y-STR profile recovered from male biological traces at a crime scene can point, first, at a surname and a family rather than at an individual.
The result came back showing that the genetic data of the man surnamed Gao corresponded to the profile left behind at the Baiyin scenes.
The police began a family-line search. Working along that family, outward and upward, they arrived at a name: Gao Chengyong, a distant nephew of the man whose blood they had taken.
The rest was ordinary. They collected his fingerprints and his DNA and compared them with the prints and traces left at scenes twenty-eight years earlier. They matched.
On August 26, 2016, the police found him in a small shop inside a technical school in Baiyin. He was criminally detained the next day and formally arrested on September 30. He confessed to eleven attacks between May 1988 and February 2002, in Baiyin and in Baotou, and to the deaths of eleven women.
The case that had a number instead of a name, Gansu-Inner Mongolia 8.05 series, was closed twenty-eight years after it began.
Four. Why the family, and not the face
People asked afterwards why this approach worked when everything else had failed.
It worked because it went around the deadlock. That is also why Gansu 8.05 is now treated as a turning point in crime in China, and not only in this one city.
A fingerprint identifies a person, but only if that person is in a database or is brought in for comparison. The Baiyin case was stuck exactly there. The police did not know what the killer looked like, what he was called, or where he was. They could only compare prints one at a time, and after more than a hundred thousand comparisons they still had no name.
What the male traces at the scenes also carried was something that works even when you have no idea who the man is: his paternal bloodline.
It is worth being precise about the difference between Y-STR testing and ordinary autosomal DNA. Autosomal DNA is personal. Half comes from each parent, and it identifies this individual. Y-STR is familial. It travels only down the male line, and it identifies this family. So the standard route has two steps. First, use the Y-STR profile to pull the search into a single family. Then use autosomal DNA and fingerprints inside that family to find the one man.
That blood sample taken in 2016 opened the door. What actually nailed the case shut was still the scene from 1988, and the fact that the evidence from it had been kept properly for all those years until the day it could finally be read.
One qualification belongs here. This route is not unique to China. In 2018, police in California closed the Golden State Killer case by the same logic: they uploaded crime-scene DNA data to a public genealogy site, matched a distant relative, and worked backwards down a family tree. The technology has no nationality. Wherever there is stored evidence and a family tree, it starts to work.
There is also a social condition that made this easier to do in China, and it is worth naming. Household registration and the old pattern of clan settlement are still largely intact, so police can trace a surname backwards and sideways across a county. In many countries that map no longer exists. It is one reason Chinese investigators have been able to return to cold cases that were thought closed for good.
The case became a reference point. In the literature of the Chinese criminal case, and in the wider study of cold cases, Gansu 8.05 marks the moment a family-line search moved out of theory and into ordinary practice.
The other half of the answer is duller, and it matters just as much in any account of Chinese true crime. The material from the Baiyin scenes had been kept for nearly three decades, and that is not something to take for granted anywhere. A family-line search only works if there is something left to search against. In this case the something was a set of print cards and sealed vials that nobody, across twenty-eight years and several changes of command, had thrown away.
Five. The trial
On April 24, 2017, the Baiyin municipal people’s procuratorate charged Gao Chengyong with four offences: intentional homicide, rape, robbery and desecration of a corpse.
Because the case touched the privacy of the victims, the Baiyin intermediate people’s court heard it behind closed doors on July 18 and 19, 2017.
A closed hearing is not the same as no hearing. What is closed is the gallery. The court sits, the defence is heard, the evidence is tested. The public is simply kept out, along with the press, so that the names, the identities and the last hours of eleven women are not walked through again in front of a crowd.
The families were allowed in. The court said that the close relatives of every one of the victims joined civil claims to the criminal case and appointed lawyers to represent them. On the other side, the court assigned a legal aid lawyer to the defendant.
The hearing ended at a quarter past eleven on the morning of July 18. The reports record one detail from it: Gao Chengyong showed no gesture of apology to the victims’ families.
On the morning of March 30, 2018, the court delivered its first-instance judgment: the offences were combined and punished as one, and the sentence was death with deprivation of political rights for life, plus compensation for the material losses of the plaintiffs in the attached civil claims. Gao Chengyong accepted the judgment in court and did not appeal. The prosecution did not appeal either.
Under Chinese procedure a death sentence still has to be reviewed by the Supreme People’s Court before it can be carried out. That review was completed, and on the morning of January 3, 2019, Gao Chengyong was executed on the authority of the Supreme People’s Court, with procurators present to observe.
He was born on November 10, 1964, in Yuzhong county, Gansu. He had done well at school, well enough to apply to a flying school, and to sit the entrance examination for an aviation academy a year after repeating his final year. He missed the mark and was not admitted. He went out to work, at one point selling restricted knives. He married in 1986 in Qingcheng, in Yuzhong. Later he and his wife moved into Baiyin city, they said, so the children could study there, and the couple ran a small shop. In May 1988, in the city where that shop stood, he began.
Six. The shape of the case
Three things are hard to get around.
The first is time. Fourteen years separated the first killing from the last, and another fourteen separated the last from the arrest. Twenty-eight years is not an abstract span here. It is a specific number of days, and it is a stack of files that moved from a magnifying glass to a database that could compare prints by itself.
The second is that the hunt never had a single moment of success. Hundreds of thousands of men were questioned. More than a hundred thousand fingerprints were compared. A reward notice stayed up for years. It was ground out of time and manpower, slowly, until technology finally arrived at the one place that mattered.
The third is what the killer left behind. He was bold enough not to wipe his prints. But what gave him up was not the print he carelessly left. It was something he never knew existed: a string of bases in his father line that does not change.
From the first killing to the moment the police led him out of that little shop, twenty-eight years passed. In those years the police compared more than a hundred thousand fingerprints and went through hundreds of thousands of people. What finally stopped the count was a sample of blood from another man, a man he barely knew, taken for a reason that had nothing to do with murder at all.
