In March 1873 a jury at Durham heard evidence about roughly twenty deaths and convicted Mary Ann Cotton of one.
The prosecution had been permitted to lead the other deaths as evidence of system, a device that let the Crown show a pattern without having to prove each death separately. So the jury was told about the husbands, the children and the lodgers. It was asked to return a verdict on her stepson Charles Edward Cotton, and only on him. It took ninety minutes. She was hanged at Durham gaol on 24 March 1873 by William Calcraft, a convicted murderer once over, with twenty more deaths hanging in the record behind her and no verdict attached to any of them.
That gap is the subject of this piece. Twelve people in this archive carry a number in public that no court ever returned, and the reason is different in almost every case.
The commonest reason is the plainest. The evidence ran out. Maria Catherina Swanenburg was found guilty at The Hague in 1885 of poisoning committed three times, four victims named in the judgment because the third count covered a married couple. The figure that follows her name is forty, and it comes from an unsourced commentary paragraph printed beside that judgment rather than from the court. She was sentenced to life, the maximum available in a country that had abolished the death penalty, and died in prison at Gorinchem in 1915. Helene Jegado went to the guillotine at Rennes in February 1852 on three convictions with at least twenty-three killings attributed and none of the others proved. Margaret Waters was hanged at Horsemonger Lane Gaol in October 1870 for the murder of one child, John Walter Cowen, against at least nineteen attributed. Dorothea Puente was sentenced to life without parole in December 1993 on three convictions with at least nine deaths attributed. Max Gufler was convicted of four murders at Vienna in 1961 and died in prison in 1966, with eighteen or more attributed.
Five cases, five countries, a hundred and forty years apart, and the same shape every time: the court took what it could prove and left the rest in the file.
The second reason is that the number came from the killer.
Donald Leroy Evans claimed around sixty. Almost none of it was substantiated. He was convicted of two, sentenced to death in 1993, and stabbed to death by another death row inmate at the Mississippi State Penitentiary in January 1999, which means the claims were never tested and now never will be. Jake Bird claimed up to forty-four. Investigators in other states corroborated eleven of them, which is the useful detail: somebody checked, and the number came down by three quarters. He was convicted in Tacoma in November 1947 and hanged at Walla Walla in July 1949. Stephen Lawrence Anderson claimed eight and was charged with none of them. He was executed at San Quentin in January 2002 on a single conviction.
A confession is not evidence of the thing confessed to unless something else supports it. Bird is the case that proves it, because his was the only one anyone appears to have audited.
The third reason is that a bargain closed the case before the counting was done. Bobbie Sue Terrell was convicted of one murder and related offences in February 1988 and sentenced to sixty-five years under a plea arrangement that avoided a capital trial. Twelve deaths are attributed to her. She died in prison in August 1995, aged forty. A plea ends the proceedings; it does not adjudicate the deaths that were not part of it, and those deaths keep being counted in every account written afterwards.
The fourth reason is geography. Charles Sobhraj has twenty or more deaths attributed to him across Thailand, India, Malaysia and Nepal, and two murder convictions, both in Nepal, ten years apart. He served twelve years in India for offences that were not murder. His first murder conviction came in 2004, when he was sixty, for a killing in 1975. No single jurisdiction ever had the whole file, and no jurisdiction has to account for a death that happened in another one.
And then there is the case with no number at all, which is where this ends.
Masten Wanjala was arrested in Nairobi on 14 July 2021. Kenyan police attributed ten to thirteen killings of children to him. He was produced in court within days, but plea taking was deferred twice at the investigators' request, and thirteen counts of murder were prepared without ever being put to him. On 13 October 2021 he escaped from Jogoo Road police station, hours before he was finally due to enter a plea; three officers on duty were later arraigned for allowing and assisting the escape and said there had been a power cut. Two days later he was beaten to death by a crowd at Mukhweya in Bungoma County. He was twenty. His parents have disowned him.
Thirteen counts existed on paper and were never read aloud to the man they named. He has no convictions, and the archive records none, because none is the true figure. The thirteen is real in the sense that police prepared it and unreal in the sense that no court ever heard it, and both of those things have to be said at once or the record is wrong.
What the twelve say together is that the number attached to a name is almost never a judicial finding. It is a prosecutor's pattern evidence, a detective's tally, a confession nobody could test, a charge sheet that a plea made redundant, or a file that stopped at a border. Courts produce the smallest number in the room, every time, because a court is the only body in the process obliged to prove a death before counting it.
That is not an argument that the smaller number is the truer one. Swanenburg almost certainly killed more than four people. It is an argument that the two figures are answers to different questions, and that an archive which prints only one of them is answering a question nobody asked. What a court found is a fact about the case. What was attributed to someone is a fact about the investigation, and sometimes a fact about the man himself, and occasionally, as with Jake Bird, a fact that falls apart the moment somebody checks it.
Mary Ann Cotton's jury heard everything and convicted on one death. It is the oldest case here and the clearest. The information and the verdict were in the same room, and they still came out at different numbers.