In the winter of 1977, 35 officers from nine police departments moved into a headquarters in Southfield, Michigan, and began working through tips about four dead children. By the time the unit closed in December 1978 it had received more than 11,000 of them. No one has ever been charged. Twenty-seven years later, in Wichita, Kansas, a lieutenant who had spent most of his career on one old file led a smaller group against a man who had not killed since 1991. He had his suspect in custody less than a year after the man chose to make contact again.
The task force is the American answer to the serial case, and it has been copied from Atlanta to New South Wales. It pools detectives, laboratories and records from agencies that otherwise compete for credit, budgets and press attention. It also inherits their habits. The nine investigations that follow, from 1977 to 2022, show that a task force rarely fails for want of officers or tips. It fails on what it believes about the offender, on the information it never sees, and on the people it leaves outside the room. When it succeeds, the cause is usually narrower than the headlines suggest: one record, one analyst, one detective who kept the file open.
Mark Stebbins was 12 when he died in February 1976. Jill Robinson, also 12, died in December 1976. Kristine Mihelich, 10, was missing for 19 days before her death came to light in Franklin Village on January 21, 1977. Timothy King, 11, of Birmingham, was the fourth, in March 1977. The children came from four neighboring suburbs north of Detroit and had no link to one another in life.
The joint effort began, according to WDIV’s reconstruction of the case files, as an idea at a meeting called by a Southfield police lieutenant to discuss the Robinson killing and the Mihelich case. It was formally set up and funded only after the fourth child was taken. Its leaders thought in large terms. A 1977 summary report by a State Police lieutenant and a Southfield Township youth officer lists eight special projects. Under one, flyers carrying a profile and a composite sketch went to postal workers, utility crews, meter readers, bank tellers and real estate agents, anyone who saw many houses and many faces in a day. Under another, investigators gathered every reported sexual offense in southern Oakland County. Prosecutor L. Brooks Patterson called it the strongest effort he had ever seen in the county.
The volume became the problem. Tips arrived faster than detectives could follow them, and the public face of the case rested on one witness’s description of a car, a blue AMC Gremlin, and on a sketch. Officers stopped Gremlins on sight and questioned owners across the county. Decades later, the chief assistant prosecutor in Oakland County, Paul Walton, told WXYZ that the Gremlin was probably the wrong car and that the sketch was a composite of several descriptions, not a portrait of one man. The task force spent much of its short life chasing both.
It was disbanded in December 1978, after the killings stopped. What followed showed how badly the case suffered from divided ownership. Timothy King’s death came to light in Livonia, in Wayne County, which gave a second prosecutor jurisdiction. In 2009, according to sources quoted by WXYZ, investigators learned that human hair held by the State Police had been mislabeled as animal hair. When the samples finally reached the FBI laboratory, mitochondrial testing linked hairs from the Stebbins and King investigations to each other and to a car that had belonged to a man named in reports only as a person of interest, A.S. The hairs were not his, and he was never charged. When Oakland County Prosecutor Jessica Cooper announced the link in July 2012, Wayne County Prosecutor Kym Worthy said the news conference had compromised the investigation. Barry King, Timothy’s father, said it was heartbreaking to watch the two counties run in different directions.
In 2020 Walton told WXYZ that his office had completed a full review of the forensic evidence and asked every original agency for help, something he did not believe had been done before. It was 43 years after the first task force opened its doors. He also said it was not proven that one person had killed all four children. The case remains open.
The Atlanta investigation began with a gap between the police and the families. Edward Hope Smith, 14, and Alfred Evans, 13, died in the summer of 1979. Other deaths followed, and for months, the Atlanta Journal-Constitution’s chronology records, investigators saw no thread connecting them. On April 15, 1980, Camille Bell, whose son Yusuf had died the previous November, and other parents formed the Committee to Stop Children’s Murders to press the city to treat the cases as linked.
On July 17, 1980, after nine deaths, the Atlanta police assigned five investigators to a special task force. It grew to more than 50 officers. The FBI, by its own account, entered cautiously at first, following out-of-state leads and offering its laboratory and its behavioral unit while it weighed whether federal law applied. In November 1980, after pressure from politicians and the press, the Bureau opened a major case with more than two dozen agents and staff assigned full time.
The decisive move was tactical. In late April 1981, the FBI says, the task force put surveillance teams in place across the metropolitan area. On May 22, 1981, one of those teams questioned a 23-year-old freelance cameraman named Wayne Williams and let him go for lack of probable cause. Nathaniel Cater’s death came to light two days later. Williams was questioned for 12 hours in June, placed under round-the-clock surveillance and arrested on June 21, 1981, charged with Cater’s murder. On February 27, 1982, a Fulton County jury convicted him of murdering Cater and Jimmy Ray Payne, two adult men. The case at trial rested on fiber and hair comparisons and on ten other deaths from the list that Judge Clarence Cooper allowed the state to put before the jury.
What the task force did next shaped the case for four decades. On March 1, 1982, it announced that it would close 21 more cases, with one to follow, all attributed to Williams. None was ever charged. A task force exists to make arrests and support prosecutions, and here it used its authority to settle cases that no court had heard. Williams has always denied them. To many families the closures left their children’s deaths unanswered. In 2021 the Atlanta police sent evidence to a private laboratory for DNA testing, and a retired deputy chief who had worked the cases told WABE he believed Williams was guilty. In December 2022 relatives asked the city to release the results. "No lab results have ever been given to the families," said a relative speaking for several of them. The police said the investigation remained open.
The King County police set up a task force on August 16, 1982, the day after the deaths of three more young women came to light near Kent. Five deaths in a month had made the pattern plain. What the unit lacked was a way to sort its own information.
Within weeks it had a favorite suspect. A taxi driver, never charged with anything, came forward with a theory about a colleague and soon became the focus of the case. KUOW, reviewing the investigation in 2024, describes "countless hours" of surveillance on him. He was one of about a dozen men being watched at the time. Dave Reichert, then a 31-year-old detective, later said the episode taught him that there were a great many men like the eventual killer at large, and that a task force could not afford to fix on one of them.
Gary Ridgway, a truck painter from Kent, passed through the files more than once. He was arrested on a patrol aimed at men soliciting in 1982. In 1984, after a tip, he spoke voluntarily with a task force detective, Randy Mullinax, and passed a polygraph. "We moved on, as we did on so many other tips, and he was sort of put to the side," Mullinax told KUOW. In 1987 investigators searched his house and vehicle and took a saliva sample. They could not connect him to the victims beyond one eyewitness account, HistoryLink’s history of the case records, and prosecutors did not charge him.
Then the case went cold and the task force shrank, as task forces do when arrests stop coming. Most members moved on to other cases. The saliva sample stayed in storage. Reichert became sheriff in 1997 and reopened the investigation. In 2001 state crime laboratory scientists, working on old evidence with newer methods, linked the 14-year-old sample to biological evidence from several victims. Former task force members came back together, and Ridgway was arrested in Renton on November 30, 2001. On November 5, 2003, he pleaded guilty to 48 counts of aggravated first-degree murder under an agreement that spared him the death penalty in return for a full accounting. He pleaded guilty to a 49th count in 2011.
Ridgway later mocked the sheriff’s claim to credit with a line of his own: technology, he wrote, had caught him. He was not entirely wrong, and not entirely right. The laboratory made the match in 2001, but only because detectives in 1987 had collected a sample from a man they could not charge, and because someone kept it for 14 years. The forensic work that identified the last of the unnamed victims, completed in January 2024 with genetic genealogy and private laboratories, belonged to the same tradition of keeping the file alive long after the task force had dissolved.
In New South Wales the formation of a task force marked the moment the police accepted what they had. On October 8, 1993, after forensic examination indicated that the deaths of two Victorians, James Gibson and Deborah Everist, and of two British travelers, Caroline Clarke and Joanne Walters, were probably the work of the same person or people, the state formed Task Force Air under Superintendent Clive Small. Within a month the deaths of three Germans, Simone Schmidl, Anja Habschied and Gabor Neugebauer, had come to light in the same district of the Southern Highlands. On November 5 the reward was raised from 100,000 to 500,000 dollars, matching the highest previous reward for a major crime in the state.
Small told Australian Associated Press in 2014 that rumors about the Milat family surfaced early. The breakthrough, he said, came from a British man who had survived an encounter in 1990 with a driver matching Ivan Milat’s description, and who reported it to investigators from Britain in 1993. Small called it "a tremendous step forward." Milat, a laborer from Sydney’s southwest, became the prime suspect. Investigators watched him for about two months before arresting him in Sydney on May 22, 1994. He was charged with the seven murders at the end of that month and convicted on July 27, 1996.
Small also described the threat a task force can pose to itself. A detective was removed from the unit in 1994 for leaking information to the press, and Small feared the leak could hand Milat an argument that he had not received a fair trial. Small said he considered suing a current affairs program over its coverage and chose to "grin and bear it" instead, rather than give the defense grounds for appeal. The convictions survived appeal and a failed application to the High Court in 2004.
Task Force Air’s work did not end with the verdict. In 1993 it had reviewed 58 cases of missing young people from every state and territory for any link to the seven deaths. The list stayed confidential until August 2025, when the state’s police minister released it. Milat refused to cooperate with police until his death in custody on October 27, 2019, and he was never charged with any other death. Inclusion on a list reviewed by police is not a finding that he was involved.
The Spokane task force was a joint venture of the city police and the county sheriff, formed late in 1997 with help from the FBI. Over its life it received more than 6,000 tips. It also built the kind of database that later critics would say it failed to use: a record of every stop of a white Corvette, a list of every Corvette registered in Washington and Idaho, and the membership rolls of Corvette clubs.
Robert Lee Yates Jr. was in that data early. An officer stopped him in a white 1977 Corvette in September 1997 and did not cite him. A state trooper stopped him for speeding in the same car that November. Officers recorded his name in July 1998 and saw him again that November. In March 1999 a caller reported a license plate that belonged to him. In September 1999 detectives interviewed him about the 1997 stop, and he told them he had sold the car. He was not a suspect.
In January 2000 detectives traced the Corvette through its new owner and sent material from it to the state crime laboratory. On February 28, 2000, the task force held a public meeting and said it had DNA that would link the killer to ten victims. The laboratory reported a match in April, and Yates was arrested in Spokane County on April 18, 2000, charged with the murder of Jennifer Joseph. In October 2000 he pleaded guilty to 13 murders, among them ten women killed in the Spokane area between 1996 and 1998 and a college couple shot near Walla Walla in 1975, and to the attempted murder of a woman who survived. A Pierce County jury convicted him in 2002 of two more murders.
In May 2001 the former Los Angeles detective Mark Fuhrman published a book arguing that the task force could have caught Yates about two years earlier by following up on the Corvette contacts. Sheriff Mark Sterk called the claim a fabrication, and the task force leader, Sergeant Cal Walker, said the facts did not support the speculation. The documented contacts are not in dispute. Whether any one of them should have produced an arrest, in a database of thousands of cars and thousands of tips, is still a matter of opinion. The Associated Press reported in December 2001 that the task force had wrapped up its work.
On July 29, 2002, investigators in Baton Rouge announced that DNA linked the deaths of Gina Wilson Green, Charlotte Murray Pace and Pam Kinamore to one man. Local, state and federal agencies formed the Multi-Agency Homicide Task Force. Its tip line had taken some 1,200 calls by February 2003, and it asked men named in tips for cheek swabs. By then, The Christian Science Monitor reported, nearly 1,000 men had been tested.
The task force worked from a picture. Sightings of a white pickup truck had been released to the public, and Newsweek reported in 2003 that authorities had relied for months on an FBI profile that assumed the killer was white. The longer profile the task force published in December 2002, as printed by WAFB, did not state a race, and the reporting has never settled what the original said. The effect is clearer. When DNA linked the death of Trineisha Dené Colomb in Lafayette to the same man in December 2002, investigators had at first doubted the connection because she was Black and the earlier victims were white. In March 2003, shortly after a Florida company offered an ancestry estimate from trace DNA, the search was opened to men of all races. A forensic statistician consulted on the case, Charles Brenner, wrote in 2004 that he had reached essentially the same conclusion from the standard crime laboratory profile.
The name that solved the case had been offered to the task force before it existed. On July 24, 2002, officers from the small police department in Zachary met the Baton Rouge police chief and other investigators and said that Derrick Todd Lee, a laborer from St. Francisville with a record of peeping and stalking, was their prime suspect in two cases of their own. Zachary detectives were not brought into the task force. Joey Watson, Zachary’s police chief at the time, told Newsweek that leads were lost because of the profile.
The break came from outside. Zachary police and the state Attorney General’s Office, still working their own cases, obtained Lee’s DNA in May 2003 under a court subpoena. The state crime laboratory matched it to the task force’s profile later that month. Lee was named as the prime suspect on May 26 and arrested in Atlanta the next day. The Louisiana Supreme Court later held that the subpoena had not met the standard for a search, but admitted the evidence anyway: a sketch released on May 23 had produced tips naming Lee independently, and the task force’s standard procedure would have led it to swab him. The court’s reasoning is a backhanded credit to the task force’s method and a quiet indictment of its priorities. Lee was convicted in 2004 of the murders of Geralyn Barr DeSoto and Charlotte Murray Pace and died on death row in 2016.
The Wichita killings of the 1970s and 1980s produced no arrest, and the case sat largely dormant after 1991. It came back to life because the offender wanted attention. On March 19, 2004, a letter reached The Wichita Eagle containing a photocopy of Vicki Wegerle’s driver’s license and photographs from her 1986 killing. By March 25, Lieutenant Ken Landwehr, who had worked the case for more than 20 years and led a new task force, had confirmed the letter was genuine. Within days police had 365 tips through a new telephone line and email address.
The task force did something few have done well: it treated the offender’s own correspondence as the investigation. Rader sent 11 communications before his arrest, police said. In January 2005, in one of them, he asked whether he could communicate by floppy disk without being traced. Police answered, as he asked, with a classified advertisement on January 30: "Rex, it will be OK." On February 16, 2005, a disk reached KSAS-TV. Data on it, the Associated Press reported, led to his church and to the name "Dennis." An internet search showed that the president of the congregation was Dennis Rader, a Park City code enforcement officer.
The task force then needed DNA without alerting him. It obtained a sample taken from his adult daughter, and DNA from three of the cases was a familial match. On February 25, 2005, Rader was arrested in Park City in an operation involving more than 200 officers from five agencies. He gave about 30 hours of videotaped interviews. "His arrogance helped catch him," said Larry Welch, director of the Kansas Bureau of Investigation. "He should have kept his mouth shut." Rader pleaded guilty on June 27, 2005, to 10 counts of first-degree murder.
Thirty years of conventional work had not named him. Investigators had collected swabs from about 1,600 men over the years, none of which matched, the Sedgwick County district attorney, Marc Bennett, said in 2025. What the 2004 task force had that its predecessors lacked was continuity in one person and a decision to engage. "Kenny was a better cop than Rader was a criminal," Bennett said of Landwehr. Landwehr had kept the case alive long enough for the man who started it to make a mistake.
On October 1, 2006, the death of Joan Diver, 45, a nurse and mother of four, came to light in Newstead, outside Buffalo. Investigators asked early on whether it was tied to a series of sexual assaults on women in the Buffalo area going back to 1981, and to the murders of Linda Yalem in 1990 and Majane Mazur in 1992. A joint task force was formed, drawing on 12 agencies, WGRZ reported. Its core was the Erie County Sheriff’s Office, the Buffalo and Amherst police departments and the New York State Police, with the University at Buffalo police as liaison.
DNA recovered in the Diver investigation matched the profile of the man responsible for the earlier assaults. That told investigators they were looking for one man. It did not tell them who. The name came from paper. Betsy Schneider, a State Police intelligence analyst assigned to the FBI’s Buffalo field intelligence group, went back through decades of records and flagged Altemio Sanchez, a factory worker from Cheektowaga, as a man who had been looked at and dropped. She linked a car in a 1981 license plate report to his uncle. Investigators went back to a house where Buffalo police had questioned people 25 years earlier, and the uncle admitted that his nephew had been driving the car. By mid-November, three separate investigative units had each pointed to Sanchez, the sheriff said. On January 13, 2007, detectives collected the glasses and tableware he had used at a restaurant. The DNA matched.
Sanchez was arrested in Buffalo on January 15, 2007. On May 16 he pleaded guilty to second-degree murder in the deaths of Linda Yalem, Majane Mazur and Joan Diver. None of the sexual assaults was ever tried. New York’s five-year statute of limitations for rape had expired on all of them, and the 2006 law that removed the limit for the most serious sexual offenses did not reach back.
The task force’s success exposed an older failure. A man had spent more than 20 years in prison for two of the earlier assaults. Task force members visited him, and he again said he was innocent. Locating the old evidence took time: a state review later concluded that officers had not followed the medical center’s procedure for requesting it. When slides from the two cases were located in March 2007, DNA testing excluded him and matched Sanchez. On April 2, 2007, a judge vacated both convictions in a hearing that lasted seven minutes. The National Registry of Exonerations lists mistaken witness identification as the cause. The state settled his claim in 2010 for 4.25 million dollars. The task force was honored by the State Police and the New York State Bar Association. Schneider said later that she could hardly believe the man had been identified at last. Her work had been done at a desk, with files other people had closed.
For more than a decade the Long Island case stood as the textbook example of a stalled investigation. Part of the reason lay inside the Suffolk County Police Department. Its chief of department, James Burke, was arrested on December 9, 2015, and later sentenced to 46 months in federal prison for a civil rights violation and conspiracy to obstruct justice, after beating a handcuffed man in a precinct interrogation room in 2012 and pressing detectives to hide it. Two days after his arrest, CBS New York reported, the FBI agreed to assist Suffolk police in the investigation, five years after the first deaths came to light.
The task force that solved the case was formed in early 2022, with the Suffolk County police, the State Police, the Suffolk County Sheriff’s Office, the FBI and the district attorney’s office. Accounts differ on the month and on who deserves credit for creating it. They do not differ on the speed of what followed. On March 14, 2022, according to the bail application, a State Police investigator searching vehicle registration records learned that Rex Heuermann, a Manhattan architect who lived in Massapequa Park, had owned a first-generation Chevrolet Avalanche at the time of the killings, the model a witness in the Amber Lynn Costello case had described in 2010. He also matched the witness’s description of the man.
From that record the investigators built a case almost entirely from data that had sat in other people’s files. A grand jury authorized more than 300 subpoenas, search warrants and other legal process. The historical location records for Heuermann’s own phone no longer existed, so investigators used billing records. They identified no occasion, the bail application says, on which his phone was somewhere different from the prepaid phones used to contact the victims. Prosecutors said an email account linked to him had been used for more than 200 searches about the case between March 2022 and June 2023, among them, on March 28, 2022, a search about the new task force itself. The man the task force was hunting was reading about it.
A DNA comparison reported in June 2023 completed the picture. Heuermann was arrested in Manhattan on July 13, 2023. The case grew through superseding indictments, and on April 8, 2026, he pleaded guilty to three counts of murder in the first degree, for Melissa Barthelemy, Megan Waterman and Amber Lynn Costello, and four counts of murder in the second degree, for Maureen Brainard-Barnes, Jessica Taylor, Sandra Costilla and Valerie Mack. He admitted killing Karen Vergata, for which he had not been charged. On June 17, 2026, he was sentenced to life without parole.
The largest task force in this group, the one in Southfield, produced the least. The one in Atlanta produced a conviction for two killings and then closed 22 more on its own authority. The early Green River unit spent its first months watching the wrong man and shrank to a skeleton while the right man’s saliva sat in storage. Each of them had officers, money and tips in quantity. What they lacked was a way to stop believing something.
The profile in Baton Rouge and the Gremlin in Oakland County did the same damage in different decades. Each began as a reasonable inference from early information. Each became a filter, and the filter screened out the man investigators needed. In Baton Rouge the right name had been delivered in person by a neighboring department, which was then left out of the unit. In Oakland County, the hair that might have mattered had been labeled as animal hair and the two counties with jurisdiction could not agree on how to speak in public.
The successes share a smaller, plainer trait. Someone went back. Schneider in Buffalo reread files that had been closed for 25 years. Landwehr in Wichita carried one case through most of a career. The Ridgway sample survived 14 years because no one threw it out. The decisive Long Island record was a vehicle registration, searched within weeks of the new unit’s formation by an investigator from an agency that had not owned the case before. In Spokane the database existed and the name was in it; the dispute is over how quickly anyone should have seen it there.
None of this makes the task force obsolete. It makes the structure less important than the habits inside it. A unit that shares information with the departments around it, keeps its evidence, tests its own assumptions and stays alive after the press leaves has a chance. The four families in Oakland County are still waiting to learn what a task force that did all of those things would have achieved.
• February 1976 to March 1977: four children killed in southern Oakland County, Michigan; a joint task force with headquarters in Southfield forms in early 1977.
• December 1978: the Oakland County task force is disbanded.
• April 15, 1980: Atlanta parents form the Committee to Stop Children’s Murders.
• July 17, 1980: Atlanta police assign five investigators to a special task force, which grows to more than 50 officers.
• November 1980: the FBI opens a major case in Atlanta.
• June 21, 1981: Wayne Williams is arrested in Atlanta and charged with the murder of Nathaniel Cater.
• February 27, 1982: Williams is convicted of the murders of Jimmy Ray Payne and Nathaniel Cater.
• March 1, 1982: the Atlanta task force announces it will close 22 more cases attributed to Williams, none charged.
• August 16, 1982: King County police form a task force on the Green River deaths.
• 1984: Gary Ridgway speaks with a task force detective and passes a polygraph.
• 1987: investigators search Ridgway’s house and vehicle and take a saliva sample.
• October 8, 1993: Task Force Air is formed in New South Wales under Superintendent Clive Small.
• May 22, 1994: Ivan Milat is arrested in Sydney after about two months of surveillance.
• July 27, 1996: Milat is convicted of seven murders.
• Late 1997: Spokane police and the Spokane County sheriff form a joint task force with FBI help.
• April 18, 2000: Robert Lee Yates Jr. is arrested in Spokane County.
• October 2000: Yates pleads guilty to 13 murders and one attempted murder.
• 2001: state crime laboratory testing links Ridgway’s 1987 sample to several Green River victims; he is arrested in Renton on November 30.
• December 2001: the Spokane task force wraps up its work.
• July 2002: DNA links three Baton Rouge deaths, and the Multi-Agency Homicide Task Force forms; on July 24 Zachary officers tell Baton Rouge investigators that Derrick Todd Lee is their prime suspect in two cases.
• March 2003: the Baton Rouge search is opened to men of all races.
• May 27, 2003: Lee is arrested in Atlanta, after Zachary police and the Attorney General’s Office obtain his DNA under a subpoena.
• November 5, 2003: Ridgway pleads guilty to 48 counts of aggravated first-degree murder.
• 2004: Lee is convicted of the murders of Geralyn Barr DeSoto and Charlotte Murray Pace.
• March 2004: a letter to The Wichita Eagle revives the Wichita case; Lieutenant Ken Landwehr leads a new task force.
• February 25, 2005: Dennis Rader is arrested in Park City, Kansas.
• June 27, 2005: Rader pleads guilty to 10 counts of first-degree murder.
• October 1, 2006: the death of Joan Diver comes to light in Newstead, New York; a joint task force forms.
• January 15, 2007: Altemio Sanchez is arrested in Buffalo.
• April 2, 2007: the two earlier rape convictions of a wrongly convicted man are vacated.
• May 16, 2007: Sanchez pleads guilty to three counts of second-degree murder.
• July 2012: Oakland County prosecutors announce a mitochondrial hair link between two of the child killings.
• December 2015: Suffolk County Police Chief James Burke is arrested; the FBI agrees to assist the Long Island investigation.
• Early 2022: a new multiagency task force forms in Suffolk County; on March 14 a State Police investigator links Rex Heuermann to a vehicle described in 2010.
• July 13, 2023: Heuermann is arrested in Manhattan.
• April 8, 2026: Heuermann pleads guilty to seven murders and admits an eighth killing.
• June 17, 2026: Heuermann is sentenced to life without parole.