The side view of a Palaeolithic dog fossil recovered from a cave in Belgium.

Where did dogs come from? That simple question is the subject of a scientific debate right now. In May, a team of scientists published astudy pointing to East Asia as the place where dogs evolved from wolves. Now, another group of researchers has announced that dogs evolved several thousand miles to the west, in Europe.

This controversy is intriguing even if you’re not a dog lover. It illuminates the challenges scientists face as they excavate the history of any species from its DNA.

Scientists have long agreed that the closest living relatives of dogs are wolves, their link confirmed by both anatomy and DNA. Somewhere, at some point, some wolves became domesticated. They evolved not only a different body shape, but also a different behavior. Instead of traveling in a pack to hunt down prey, dogs began lingering around humans. Eventually, those humans bred them into their many forms, from shar-peis to Newfoundlands.

A few fossils supply some tantalizing clues to that transformation. Dating back as far as 36,000 years, they look like wolfish dogs or doggish wolves. The oldest of these fossils have mostly turned up in Europe.

In the 1990s, scientists started using new techniques to explore the origin of dogs. They sequenced bits of DNA from living dog breeds and wolves from various parts of the world to see how they were related. And the DNA told a different story than the bones. In fact, it told different stories.

In a 2002 study, for example, Peter Savolainen, now at the Royal Institute of Technology in Sweden, and his colleagues concluded that dogs evolved in East Asia. Eight years later, however, Robert Wayne, a geneticist at the University of California, Los Angeles, and his colleagues analyzed some new dog breeds and concluded that the Middle East was where dogs got their start. (All such studies suggest that a few breeds may have been independently domesticated, although they differ on which ones and where.)

Dr. Savolainen and his colleagues continued to sequence DNA from more dogs, and they published more evidence for an East Asian origin of dogs — narrowing it down to South China.

While early studies of canine origins were limited to fragments of DNA, scientists are now starting to sequence entire genomes of dogs and wolves. In May, for example, Dr. Salovainen and Chinese colleagues reported that Chinese native dogs had the most wolflike genomes. By tallying up the mutations in the different dog and wolf genomes, they estimated that the ancestors of Chinese village dogs and wolves split about 32,000 years ago.

If this were true, then the first dogs would have become domesticated not by farmers, but by Chinese hunter-gatherers more than 20,000 years before the dawn of agriculture.

Dr. Wayne and his colleagues think that is wrong.

A dog may have wolflike DNA because it is a dog-wolf hybrid. In a paper that is not yet published, they analyze wolf and dog genomes to look for signs of ancient interbreeding. They cite evidence that, indeed, some of the DNA in dogs in East Asia comes from wolf interbreeding.

“That’s going to pump up the resemblance,” Dr. Wayne said.

Now Dr. Wayne and his colleagues are introducing a new line of evidence to the dog debate: ancient DNA. Over the past two decades, scientists have developed increasingly powerful tools to rescue fragments of DNA from fossils, producing “an explosion in the samples,” said Beth Shapiro of the University of California, Santa Cruz, a collaborator with Dr. Wayne.

On Thursday in the journal Science, Dr. Wayne, Dr. Shapiro and their colleagues report on the first large-scale comparison of DNA from both living and fossil dogs and wolves. They managed to extract DNA from 18 fossils found in Europe, Russia and the New World. They compared their genes to those from 49 wolves, 77 dogs and 4 coyotes.

The scientists examined a special kind of DNA found in a structure in the cell called the mitochondrion. Mitochondrial DNA comes only from mothers. Because each cell may have thousands of mitochondria, it is easier to gather enough genetic fragments to reconstruct its DNA.

The scientists did not find that living dogs were closely related to wolves from the Middle East or China. Instead, their closest relatives were ancient dogs and wolves from Europe.

“It’s a simple story, and the story is they were domesticated in Europe,” Dr. Shapiro said.

Dr. Shapiro and Dr. Wayne and their colleagues estimate that dogs split off from European wolves sometime between 18,000 and 30,000 years ago. At the time, Northern Europe was covered in glaciers and the southern portion was a grassland steppe where humans hunted for mammoths, horses and other big game.

“Humans couldn’t take everything, and that was a great treasure trove,” Dr. Wayne said. Some wolves began to follow the European hunters to scavenge on the carcasses they left behind. As they migrated along with people, they became isolated from other wolves.

Dog evolution experts praised the scientists for gathering so much new data. “I think it’s terrific,” said Adam Boyko, a Cornell biologist. Dr. Savolainen agreed. “I think it’s a fantastic sample,” he said.

But Dr. Savolainen said the analysis was flawed. “It’s not a correct scientific study, because it’s geographically biased,” he said.

The study lacks ancient DNA from fossils from East Asia or the Middle East, and so it’s not possible to tell whether the roots of dog evolution are anchored in those regions. “You just need to have samples from everywhere,” Dr. Savolainen said.

He also rejects Dr. Wayne’s argument that interbreeding in East Asia creates an illusion that dogs originated there. Dr. Savolainen points out that the study suggesting interbreeding was based on a wolf from northern China. “What they need to have is samples from south China,” he said.

There’s just one catch. South China is now so densely settled by people that no wolves live there. A similar problem applies to the fossil record.

“It may be impossible to go this way,” Dr. Savolainen said.

Dr. Wayne is not quite so pessimistic. He and his colleagues are hoping to widen their scope and find more DNA from fossils of dogs outside of Europe, while also looking at the genes of living dogs that might hold important clues. Yet he thinks it unlikely that the new evidence will change the basic conclusion of his latest study.

“But there have been so many surprises in the history of this research on dog domestication that I’m holding my breath till we get more information,” Dr. Wayne said.