The Origins of Dogs

It’s the soft part of a summer morning, when the sun is barely up and the day’s heat has not yet taken hold. A young rabbit ventures from a thicket of raspberries, nose quivering. Its movement does not go unnoticed.

Some 20 feet away, one of two dogs sprawled belly-up on the grass opens her amber eyes. The other dog turns over slowly and looks in the direction of the raspberries. They roll to their feet, eyes now trained on the rabbit. The animal darts for the back gate.

The dogs pursue. The female, smaller and more agile, swerves left and right like a cheetah, adjusting course with the rabbit. The bigger male dog thuds behind. They close in on the prey as a team. Without slowing, the rabbit squeezes under the gate to freedom.

The smaller dog, unable to brake in time, slams headfirst into the gate. Behind her, all 70 pounds of the male dog are still hurtling forward. He collides with her haunches, sending her back into the gate. The 6-foot-tall fence shudders with each impact. Both dogs shake their heads and trot back to me, hoping for a consolation prize.

I doubt either of them would last long in the wild. They are thousands of generations removed from the ancestral wolves at the base of the dog family tree. What nature started, we humans have tinkered with for millennia.

The dog, Canis lupus familiaris, was the first animal, and the only large carnivore, to be domesticated. Yet its origins remain a mystery. Now, an unprecedented global collaboration of scientists is at last unraveling the tangled story hidden in dogs’ DNA.

For years, getting researchers to pinpoint the origin of man’s best friend was like walking into a room full of dog owners and asking, “Who’s a good boy?” Everyone would promote, with utmost confidence, their own pet theory.

Recent genetic studies have placed ground zero for dog domestication in Europe, Central Asia, the Middle East, South Asia or Southeast Asia. And they date dog origins anywhere between 10,000 and 38,000 years ago.

Some of these contradictory results arise from dogs’ complex history with humans and other canids. Once domesticated, dogs proved valuable in myriad ways: guards, hunters, shepherds, pack animals, sled pullers.

“Dogs are motivatable to do the jobs humans want them to do, so people took them with them as they moved,” says Adam Boyko, a geneticist at Cornell University. “Cats, by comparison, are not motivatable, so people didn’t scatter them across the world. A single cat might jump on a ship and end up somewhere else, but you don’t have, say, trained war cats going into battle.”

As they traveled, dogs mingled with other populations of dogs but also with wolves, both ancestral and modern, creating a genetic potpourri. A 2015 study in Genome Research, for example, estimated that 25 percent of modern Eurasian wolf DNA actually comes from interbreeding with domesticated dogs.

Another confounding factor in earlier studies: Researchers sampled DNA from modern purebred dogs, which are the result of generations of artificial selection and hybridization by breeders, skewing the genetic timeline of when wolves and dogs parted ways.

But now, researchers are collecting broader modern-dog samples. Boyko, for example, co-authored a 2015 study that included samples from free-breeding dogs around the globe. These “village dogs” are more representative, genetically, of the species. Of the estimated 1 billion dogs in the world today, about 75 percent are free breeding.

At the same time, researchers have made advances in extracting and sequencing ancient DNA, allowing them to see the past rather than make a calculated guess based on modern material.

“Ten or 20 years ago, we looked at modern dogs and modern wolves, and that’s it,” says German geneticist Olaf Thalmann, currently at Poland’s Poznan University of Medical Sciences. “We have realized this is wrong. Now we’re going back to the cradle of domestication to look for answers there [because] the wolves we see today are not what gave rise to dogs.”

In June, Science published a paper that heralds the new direction of research. According to the study, dogs were domesticated not once but twice, on opposite ends of the Eurasian continent at least 15,000 years ago. Previous studies assumed that domestication was a difficult and thus rare event, occurring only once. But the new dual-origin theory found that an ancient European population was replaced by an eastern Asian population as the latter expanded across the continent. Every dog alive today is descended from ancient Asian roots.

In addition to collecting DNA from hundreds of modern wolves as well as mutts and purebred dogs, the dual-origin researchers extracted DNA from dozens of ancient dogs, including a particularly high-value sample from a 4,800-year-old animal unearthed in Newgrange, Ireland.

“The ancient [Newgrange] dog had ancestry not found in modern dogs or in modern wolves,” says Mietje Germonpré, who was not part of the dual-origin team. The Belgian paleontologist has studied the remains of other older canids in Eurasia and believes some of them were early dogs — a controversial theory, but one this new research suggests may be correct.

“It’s the first hint toward what’s out there,” says Thalmann, who also wasn’t involved in the research. “It’s a wake-up call. The theory about multiple origins and timing was out there for some time, but this is the first evidence for it genetically.”

The Science paper is just the prelude to a flood of new research that will emerge over the next decade, largely thanks to Oxford University evolutionary biologist Greger Larson. One of the study’s authors, Larson has advocated collaboration rather than competition as co-director for a dog origins project begun in 2013. Nearly every major researcher in the field is now working with him in some capacity.

“[Larson] got all these people in a room, put the hardcore data on the table and said, ‘OK, what are the data telling us, and what are egos telling us?’ ” says Thalmann, who is now working with Larson.

One of the early recruits was Germonpré, who set off a firestorm in 2008 when she described a 36,000-year-old Pleistocene dog from Belgium’s Goyet Cave. The remains, which include a partial skull, were found in the 1860s in a limestone cave along with lynx, mammoth and other animals. The Goyet individual was labeled a wolf in field notes until Germonpré compared details in the size and shape of its skull and teeth with modern wolves and dogs and other ancient canids. She believes Larson invited her to collaborate precisely because of her controversial theory, and her perspective.

“I have a background in Pleistocene mammals, and the others generally specialize in Neolithic or after the Ice Age,” Germonpré says. As a researcher focused on an earlier period, she was more comfortable than some of her peers in accepting that domestication prior to the advent of agriculture — roughly 12,000 years ago — was even possible.

The archaeological records of domestication and agriculture go hand in hand for all species but one: the dog. The newest studies provide the most robust confirmation yet that the domesticated dog evolved when humans were still hunter-gatherers.

Germonpré’s 2009 assessment of the Goyet canid was largely based on the measurement of observable physical traits, such as the comparatively wide skull and snout, more doglike than wolfish. It’s a relatively old-school approach, but one still crucial to Larson’s project. 

On an early evening walk, my dogs dawdle over an interesting scent, noses to the ground, making snuffling noises. I hear another sound: the heavy pant of a large dog on the run.

It’s our neighbor’s oversized white German shepherd, off-leash again. It’s exactly the reason I carry a high-pressure spray bottle.

Even as I squirt the dog with water to keep it at bay until its owner arrives, I appreciate that it’s a magnificent animal. It’s pure white, and probably a hundred pounds. Dissuaded by the stream of water in its face, it backs off, standing a few feet from my male dog, who is also white.

The contrast between them is striking. Mine has a short, sleek coat. The shepherd has luxurious fur.

Visually connect the shepherd’s triangular ears to the end of its pointed snout, and you get a larger triangle. My dog has a broad, square head and a wide mouth, and is built like a linebacker. Had someone unfamiliar with dogs witnessed the meeting, they might have concluded, as Charles Darwin once did, that different dogs descended from different species.

Today we know that dogs descended from ancestral wolves. But we know very little about what that animal, and the first dogs, looked like. Wolves have been around for half a million years, with at least 32 living subspecies. Researchers have unearthed ancient canid bones from Belgium to Kamchatka, but some scientists are skeptical over whether many of the finds are ancestral wolves, domesticated dogs or something in between.

Skull morphology is key to the domestication debate because it’s where the physical difference between modern wolves and modern dogs is most pronounced. Today, even upper-latitude, “wolfish” varieties like huskies have shorter, wider snouts, and shorter braincases, than wolves — traits typical of domestication syndrome.

First observed by Darwin but best chronicled in a 20th-century Soviet attempt to breed tame foxes, domestication syndrome covers a range of unintended physical traits that emerge as a wild species is selectively bred for more docile behavior. The changes include neoteny, where juvenile appearance is preserved into adulthood.

That permanent puppy face can be seen from Chihuahuas to St. Bernards. This suite of traits did not evolve in a generation, however. The first dogs likely looked like now-extinct ancestral wolves.

“Evolution is a process,” says Larson. “Where do you draw the line between what is a dog and what is not?”

This summer, researchers working with Larson began winding down the collection phase of the dog origins project. They’ll spend at least the next two years studying the roughly 1,500 samples of genetic material from ancient canid remains.

But extracting and analyzing ancient DNA is no easy task. After death, a host of microbes invade the body to begin the process of decomposition. Bacteria are rampant in the samples, leaving their own genetic material and muddying the picture.

Boyko is collaborating with Larson to analyze ancient canid samples, some of which are 99.5 percent bacterial contamination, according to Larson. Geneticists can bypass the bacterial mess by designing custom search tools that allow them to hunt for DNA segments unique to canids, but it’s a time-consuming process.

The genetic data, combined with updated morphological study of some of the oldest ancient canid remains, could strengthen the dual origin theory proposed in June — or it could build a case for even more origin events.

But even if we can one day say, with confidence, when and where and how many times an ancestral wolf became a dog, we’ll still have only half of the story. The remaining half could be even more important: How did an apex predator, a direct competitor for resources during our hunter-gatherer days, become our best friend?

My female dog was picked up as a stray when she was 8 weeks old.

She spent the next four months in a concrete kennel. Her only interaction was with shelter workers who came by twice a day to clean her space and provide food and water. The first six months of a dog’s life include crucial periods of socialization, both with humans and other dogs. Those windows had closed by the time I met her.

Less than a week after bringing her home, I took her to a dog park. The environment of grass, wood chips and other animals must have been a sensory wonderland for her.

One of the dogs took an interest in a pile of something near a back fence. All of them stopped. Sniffed. Some began to eat, others to roll. A retriever got possessive about the find.

The humans, in a knot of conversation a tennis ball’s throw away, began to move toward their pets.

I called my dog’s name. Immediately she came in a clumsy puppy trot, ears flopping this way and that, her eyes on me: her human. Even without the benefit of socialization at all the right times, something in her very dogginess made her respond.

“All animals can be social with people to some extent, but dogs are uniquely good at being with us,” says University of Alberta archaeologist Robert Losey, who studies the human-dog relationship in prehistory. Much of his work has focused on the Lake Baikal area of Siberia, where dogs have been buried with or near humans in cemeteries up to 8,000 years old. The graves are the remnants of a hunter-gatherer culture.

“The dogs buried there were not killed or eaten,” Losey says. “In some cases, human bodies were moved to make room for the dog’s. That tells us there was a social bond, perhaps a belief in a dog’s soul, or something that persisted in the afterlife.”

The Baikal burials are not the archaeological record’s only example of Fido’s final resting place. “There are more dog burials in prehistory than that of any other animal,” Losey says. While evidence such as cut patterns on bones suggests that dogs were sometimes eaten or killed as part of a ritual, there are also examples, like the Baikal dogs, of the animals being accorded human-like status, in death as well as, perhaps, in life.

There may be far older examples of dogs buried with care. One of the ancient canids Germonpré and her team have studied, from Predmostí in the Czech Republic, was laid to rest with a bone in its mouth. The animal is 30,000 years old.

That close relationship between humans and dogs may be based on empathy. A 2016 study in Royal Society: Biology Letters confirmed what today’s dog owners have told anyone who’ll listen: Dogs can interpret a human’s emotional state based on the individual’s facial expression, a rare interspecies feat. Those cognitive skills likely evolved from the already highly organized behavior we believe was present in ancestral wolves.

“The wolf was pre-adapted for life with humans because it had a sophisticated social system of family groups,” says James Serpell, director of PennVet’s Center for the Interaction of Animals and Society in Philadelphia.

While wolves’ complex social system may have set them up for domestication, we can say very little about how they actually made the leap from wild predator to fireside friend. That’s because the dog is the only animal to undergo the process when humans were still nomadic hunter-gatherers. The templates we’ve developed for understanding how other animals were domesticated, after the advent of agriculture, don’t apply so neatly to dogs.

Most researchers define three general paths to domestication: prey, directed or commensal. In the prey model, humans hunt the animal — typically a large herbivore — but instead of killing it, they keep it around for future use in a kind of resource management. The free-breeding animals become a founding population. It’s the path that led to cows, pigs and most familiar farmyard residents.

Horses and other animals typically used for transport are products of directed domestication: Humans take an animal out of the wild with a specific use in mind, and breed subsequent generations for that purpose.

In commensal domestication, however, humans unintentionally create an environment that attracts the animal. Over generations, the human appreciates some benefit that the animal provides and encourages its presence while the animal becomes separated from others of its kind in the wild.

Cats are an example of commensal domestication. Once humans took up agriculture and began storing grain, the ready food source attracted rodents, which attracted cats.

Dogs may have undergone a similar process. “We see how it could have happened in wolves,” says Larson. “There was a population that just started hanging out with us, subsisting off the environment we were creating. Only after generations did humans start intentionally creating populations, and only long after that do we get crazy things like Labradoodles.”

Other researchers, however, including PennVet’s Serpell, doubt human hunter-gatherers would have tolerated large predators near their camps — or that the resource-frugal humans would have left behind enough potential food to sustain a wolf-sized animal. Instead, they argue, it’s possible that prehistoric humans, like many more recent hunter-gatherer groups, had a custom of adopting baby animals. A hand-reared ancestral wolf, Serpell argues, would develop an intense, familial bond with humans.

“That animal, as an adult, would be sufficiently socialized to be safe” in the eyes of the hunter-gatherers, says Serpell.

Meanwhile, other researchers argue that debating the how or why of domestication is beside the point.

“I think of domestication as a multispecies process,” says Losey. “It’s not something we did to wolves. It’s something we’ve done together.”

If two apex predators managed to find mutual benefit in a close relationship once, as early research held, or twice, as the newest analysis indicates, why not three or four or 10 times?

“In some ways, we have embedded the idea that domestication is so complicated and rare and unusual that it could have only been done once,” says Losey. “I don’t see that. I think there were probably hundreds, if not thousands, of domestications that lasted only a few generations. For whatever reason, they didn’t persist.”

As our oldest companion outside our own species, the story of dogs is most closely linked with our own. (Sorry, cat fanciers.) The more collaborative, more precise direction research is heading promises at last to tell us the opening chapter of that long relationship.

“We need to reconsider how we think of dog domestication,” says Losey. “It’s an ongoing process that continues today ... and it’s a process that has no end. It’s an evolutionary partnership.