Years tick by and the cold case of one dead man in Largo is given an occasional nudge. Then changing technology gives detectives new tools and new leads.
By CHRIS TISCH
Published October 18, 2004
[Photo courtesy of the Largo Police Department]
Ten years after William Cosco’s murder, police still had nine fingerprints from the apartment that had not been matched with a person, including some from the mug on the coffee table.
Part I of III The truth is waiting
A Largo man is found dead in his apartment. Police have fingerprints, and a neighbor saw someone leave the building. What would it take to catch the killer? (Oct. 17, 2004)
[Photo courtesy of the Texas Department of Public Safety]
Jill Kinkade, a fingerprint examiner for the Texas Department of Public Safety, first checked fingerprints from the William Cosco crime scene in 1996 and found no matches. She checked the prints again in 2004 after Texas’ computer system was upgraded. That’s when she hit pay dirt.
Fingerprints on a coffee mug at Wiliam Cosco’s apartment placed Jeffrey Scott Payne at the scene but did not prove he was involved in the murder.
[Photo courtesy of the Largo Police Department]
After sifting through aliases, police determined fingerprints in the apartment belonged to Jeffery Scott Payne, whose last arrest had been in California in 1985. Then, he seemed to vanish.
Coming Tuesday: The search for a killer.
There are two kinds of murders: dunkers and whodunits.
Dunkers are the ones that go down easy. They come with a smoking gun, a crowd of witnesses or a suspect begging to talk. Whodunits are the ones that infect homicide detectives' dreams.
Detective John Carroll had a whodunit on his hands.
The brutal murder of William Cosco, who was stabbed to death in his Largo apartment on June 30, 1986, would not go down easy. Investigators had culled fingerprints and a bloody palm print from Cosco's apartment, but they had no suspect who matched the prints.
Leads had dried up. The case had gone cold.
Carroll was promoted out of detectives four years later. From time to time, new investigators took fresh looks at the murder, but their efforts led nowhere.
In Cosco's hometown, Utica, N.Y., his family hoped it would someday see justice.
* * *
When William Cosco was murdered, crime-solving technology was limited. Over the years, it improved. The use of DNA grew in sophistication. So, too, did fingerprint technology.
For decades, police officers printed the fingers of people who were arrested. Those prints were placed on cards that were stored in filing cabinets at police stations or county jails.
If a detective pulled a fingerprint from a crime scene and developed a suspect who had been arrested before, investigators could compare the print from the crime scene with the suspect's to see whether they matched.
But when detectives had no idea who committed the crime - as in the Cosco murder - they had no practical way to search for a match.
In the late 1980s, law enforcement agencies around the world began developing computer databases for their fingerprints. That technology led to the Automated Fingerprint Identification System, or AFIS, a computer database that could store millions of prints.
If a detective had a print from a crime scene but no idea to whom it belonged, fingerprint analysts could enter it into the computer, which would search the database for a match.
That leap in technology was stunning.
Hundreds of law enforcement agencies - from city police to sheriff's departments to state patrols - bought AFIS systems and developed databases. The systems helped solve untold numbers of cases for each department.
But there was one problem: The databases weren't connected.
If a man committed a murder in Miami and left bloody fingerprints everywhere, then was arrested and fingerprinted for bank robbery in Tampa the next day, neither city's police department would know that the robber and killer were the same man.
That's the way it was in 1996 - 10 years after William Cosco's murder - when two Largo detectives gave the Cosco case its first good shake since Carroll had left the division.
* * *
Detectives Steve McMullen and Bill Shaw decided to apply the new crime-solving technology to the case. They still had the cigarettes that had been found in Cosco's ashtray, but when they had them checked for DNA, they got no leads.
Next, they focused on nine fingerprints found in the apartment that hadn't been matched to anyone. The detectives figured one of them could belong to the killer.
Shaw got a list of nearly all the police agencies in the country - and some outside of it - that had AFIS systems. Then he made copies of all the fingerprints pulled from Cosco's apartment.
He wrote letters to more than 100 police agencies asking them to run the fingerprints through their systems. If the killer had been arrested by one of the agencies, he might get a match.
Shaw couldn't do much with the bloody palm print. Though fingerprints were being entered into databases around the world, no one was compiling palm prints.
Over the next few weeks and months, Shaw received letters back from all the agencies. Not one of them had gotten a hit.
Once again, the investigation went cold.
One of the letters Shaw received, however, was from Jill Kinkade, a fingerprint examiner for the Texas Department of Public Safety. Kinkade politely wrote that she had flushed the prints through the Texas database but had gotten no hits.
But that letter would not be the last time Largo detectives would hear from Texas.
* * *
In February 2003, Detective Joe Coyle's captain asked him to take a fresh look at the Cosco case.
Coyle, a cop since 1990, had been a patrol officer for 10 years before graduating to detectives. The work fascinated him. When he started in homicide, he pored over old cases, studying how other detectives had solved them, or failed to.
Now he had a challenge of his own. He lugged the Cosco murder books home. He leafed through the two thick, black binders of information while watching television in his living room.
Shaw and McMullen had probed the physical evidence, but Coyle - a boyishly jovial guy whose personality disarms nervous witnesses and murder suspects - wanted to try prying people's memories.
He reinterviewed Timothy Jankowski, one of the friends who had found Cosco's body, and Robert Palmiero, Cosco's cousin who lived in Clearwater at the time of the murder. He interviewed a couple of other people who knew Cosco, but he didn't unearth any new leads.
Coyle also knew that fingerprint technology had continued to advance in the seven years since Shaw and McMullen's efforts.
AFIS systems across the county were connecting. The FBI also had gathered almost all the fingerprint records in the country into one database of millions of prints; it went online in 1999.
Coyle took the nine unknown prints from Cosco's apartment to the Pinellas County Sheriff's Office fingerprint lab. He asked examiners to run the prints through the FBI system.
The examiners found no matches. Again, the case went cold.
Nine months later, Coyle was driving to work on a Monday morning in March when his cell phone rang. It was his partner, Keith Barton, calling from the police station.
"Guess what?" Barton said. "They got a hit on Cosco."
* * *
The call had come from Texas.
Jill Kinkade, the fingerprint examiner who had eight years earlier sent a letter to Largo saying no matches had been found, had called to say that had changed.
Kinkade, who has been in the fingerprinting business for 26 years, had received the prints from the Cosco case in 1996. She had run them through the Texas database then but got no matches.
Four years later, Texas upgraded its fingerprint database, adding thousands of prints and expanding the database to more than 4.4-million people.
After the upgrade, fingerprint examiners began running old cases through the new system. It took them several years to get through the backlog. It wasn't until March 2004 that Kinkade ran, for the second time, the Cosco prints. She got no matches.
Recognizing that this was a murder case, Kinkade thought it was worth it to also run the prints through the FBI's integrated system - the system the Pinellas sheriff's examiners had used months earlier.
To understand what happened next requires some knowledge of fingerprinting. Though experts in the field say fingerprinting is indeed a science, searching for a match in a database is an art.
How an examiner enters a print into the system can have a big effect on how the computer searches for a potential match. There is no right or wrong way, but each examiner is likely to enter the print differently. How Kinkade entered the print in the computer was certainly different from how the Pinellas examiner did it.
Fingerprints are never left in perfect condition, either. Police often get partial prints that are blurred or smudged.
The prints from Cosco's apartment, for instance, "were really bad prints," Kinkade said.
When an examiner runs a print through a database, the computer doesn't search for just one match. It unearths a number of prints that have similar characteristics, then ranks them in the order of likelihood that any one is a match.
"It's sort of a needle in a haystack," said Bill Schade, fingerprint records manager for the Pinellas County Sheriff's Office.
Then, just like in the old days, the fingerprint examiner must match the print with their eyes.
"No machine can make an absolute match," Schade said.
Kinkade asked the computer to spit out the 20 most likely matches. She went through the first five it gave her; none was right.
But No. 6 looked to be a keeper. Kinkade saw that all the distinctive points on the Cosco prints and the print from the database appeared to match up.
The prints from No. 6 belonged to a man who had been arrested for shoplifting in West Palm Beach in 1978. With one finger looking like a possible match, Kinkade called the FBI fingerprint lab and asked analysts to fax her copies of prints of all 10 of the shoplifter's fingers.
When she got the fax the next day, she took a magnifying glass to the prints. What she saw gave her a surge of excitement.
She immediately called the Largo Police Department.
* * *
The shoplifter from West Palm Beach had left a number of prints on a mug on a coffee table in Cosco's apartment.
Kinkade told Coyle that the prints belonged to a man who went by the name Jeffrey Scott Thomas.
Coyle started to run background checks on Thomas and found that he used a dozen aliases and almost as many Social Security numbers and birth dates.
He appeared to be a career criminal with a slew of arrests in Arizona, Wisconsin, Minnesota, Washington and California. He also had been arrested several times in Florida in the 1970s. His 14 arrests ranged from trespassing to a couple of bank robberies out west.
The man's most recent arrest had been in California on Jan. 15, 1985, more than a year before Cosco's murder. After that, the man appeared to have dropped off the map. He had not been arrested since.
Coyle determined that of all the aliases Thomas used, the one that showed up the earliest and most often - and therefore was most likely his real name - was Jeffrey Scott Payne. He was born in Minnesota and would have been 48 years old.
Coyle started calling all the police agencies that had arrested Payne. He asked them to send copies of his fingerprints and mug shots.
When the mug shots arrived, Coyle looked into the face of a man who at times was tanned with bushy-blond hair that hung down to his collar. That matched the description of the man a neighbor had seen leave Cosco's apartment the night of the murder.
All the fingerprints also matched the prints taken from the coffee mug in Cosco's apartment.
But just because Payne had left some fingerprints in Cosco's apartment didn't make him a killer. Payne could have just visited and had a cup of coffee, then left peacefully.
It would take a lot of work for Coyle to know whether Payne was the killer. The best method would be to find Payne and ask him to provide a palm print, which could be compared with the bloody print taken from Cosco's closet door jamb.
But with Payne having vanished after 1985, Coyle had no idea where to start looking for him.
Then came a lucky break.
Coyle discovered that after Payne had been arrested for robbing two banks out west, the authorities took his photograph and his fingerprints. But the FBI did something else as well.
They took prints of his palms.
Coyle asked the FBI to send him copies of Payne's palm prints. When they arrived, he took them to the Pinellas Sheriff's Office print lab and had them compared with the print left in blood on Cosco's closet door.
Fingerprints on a coffee mug at Wiliam Cosco’s apartment placed Jeffrey Scott Payne at the scene but did not prove he was involved in the murder.
