A 24-year-old shot dead on the back steps of his Kingsessing home. Rapper Meek Mill's cousin gunned down outside a South Philly takeout joint. A 68-year-old woman stabbed to death near her front door in Northeast Philadelphia, allegedly by her friend, who then killed himself. A decomposed body with dozens of stab wounds, found inside a burning East Germantown home.
Philadelphia is recording homicides at the fastest pace since 2012, and those killings and others last week added to a murder total that by Friday afternoon stood at 203, an increase of about 9 percent over the same point last year, police statistics show.
"It's one of the things that keeps you up at night," said Police Commissioner Richard Ross, who noted that shooting victims also have increased at a similar pace.
Other cities across the nation also are experiencing rising homicide rates. Thirty-seven of the 61 police departments responding to a midyear survey by the Major Cities Chiefs Association - including Atlanta, Dallas, Las Vegas, and Washington - reported rates higher than at the same point last year. Chicago has had a dramatic and well-publicized increase, surpassing 500 homicides earlier this month - already its highest annual total in a decade.
Police leaders and criminologists have found no clear explanation for the uptick in Philadelphia, let alone nationwide.
"I don't think there are any good answers," said Alfred Blumstein, a criminal justice expert and public policy professor emeritus at Heinz College at Carnegie Mellon University in Pittsburgh.
Still, interviews in recent days with experts and police officials yielded a variety of theories - from the unpredictable and cyclical nature of urban crime, to the possibility that unsolved cases could fuel additional violence, to the potential impact of "the Ferguson Effect" in which officers ratchet back their aggressiveness to avoid generating controversy.
A bloody Tuesday
Last weekend in Philadelphia, no one was murdered - a rarity, particularly during the summer.
But on Tuesday, nine people were shot - three fatally - in a series of unrelated incidents throughout the city.
That unpredictability, both in terms of timing and geography, can present a challenge for police.
Deployment decisions are routinely made to prevent retaliatory violence or drug-related crimes in specific areas. But Ross said it can be harder to focus resources when the violence is widely distributed.
And although the numbers of homicides and of shooting victims have increased by roughly the same percentage this year, the number of shooting incidents is actually down about 17 percent, Ross said. That means shooters are either getting more accurate or hitting more people at once.
Total violent crime - which includes rape, robbery, and assault - is also down about 6 percent overall, Ross said, making homicide the only violent crime category that has increased over last year.
Darrel W. Stephens, executive director of the Major Cities Chiefs Association, added that every city experiencing higher homicide rates has different factors that dictate when and where crimes might happen, such as gang conflicts or drug markets. As a result, finding a common thread among cities is complicated.
"It's not the same thing in all the cities that have seen an increase," he said.
Faulty Ramsey forecast
In 2014, after Philadelphia posted a second straight year of historically low murder totals, then-Commissioner Charles H. Ramsey said, "I don't believe we're as low as we can go," and predicted the city one year might have fewer than 200 homicides.
But the pendulum has been swinging the other way.
If the current pace holds, the city will finish with its highest murder total since 2012, when 246 people were slain, and homicides will have increased for the third consecutive year.
Criminologists point out, however, that those year-end totals still would be lower than in previous decades - and that as recently as the 1990s, Philadelphia regularly ended the year with more than 400 homicides.
"The argument begs the question: What is the expected level of violence in the city of Philadelphia?" said Robert J. Kane, professor of criminology and justice studies at Drexel University. "Some people would say zero, but that's not going to happen. So the question is: What's reasonable?"
As Philadelphia's homicide rate has been rising, the rate at which police have solved those killings - the so-called clearance rate - has been declining.
Between 2008 and 2014, the annual clearance rate exceeded 69 percent five times and never dipped below 57 percent, according to police statistics.
But through July of this year, the rate was 54 percent, police said.
Ross acknowledged that a falling clearance rate could contribute to increased levels of violence: If neighborhood residents know that a murderer is still on the street, for example, others might feel safe to commit similar crimes - and witnesses might be afraid to come forward.
"If the shooter is not taken off the street, they're either likely going to shoot someone else, or they're going be a victim," said Ross.
Some Philadelphia homicide investigators privately grumble that interrogation policies instituted two years ago have made it tougher to solve crimes. Among other things, interrogations now are videotaped, and detectives are required to remind witnesses that they can leave interviews whenever they want.
But Ross, who once commanded the homicide unit, does not think those policies present an insurmountable obstacle to solving cases.
"I believe that psychologically there are many people in homicide that think [the policy changes] have had [a negative] impact," Ross said. "But we do know that we're not going backwards in this department."
Among rank-and-file officers, a popular theory for the homicide uptick is that police are hesitant to patrol aggressively and then end up on a video that might generate negative attention. The idea has become known as as the "Ferguson Effect," named after the Missouri city where protests erupted in 2014 after a white policeman fatally shot an unarmed black man.
FBI Director James Comey said in May that he believed this theory - also known as de-policing - was driving the spike in murders in big cities.
Christopher Herrmann, an assistant professor at John Jay College of Criminal Justice at the City University of New York, agrees, but notes that no data yet support it.
"It's certainly real: You talk to everybody and they tell you they're doing things differently," Herrmann said. "But from a research perspective. . .can you measure what they're doing differently?"
Ross is skeptical that de-policing is widespread in Philadelphia, pointing out that gun-related arrests are up more than 20 percent, and that other violent crime rates are down.
And although homicide numbers currently may be up, Ross said, a lack of definitive explanations won't deter police from seeking to bring those numbers down again.
"This is a constant fight for us," he said.