Our coronavirus coverage is free for the first 24 hours. Find the latest information at Please consider subscribing or donating.

  1. Archive

Train that human. Good dog

Few relationships are so laden with mutual benefit as that between human being and dog. Much of the credit for this unusual state of affairs, it now turns out, may lie on the canine side of the equation.

Three studies in today's issue of Science shed light on the questions of when, where and how dogs were domesticated from wolves. One suggests that a few female wolves, perhaps from the same population somewhere in east Asia, are the ancestors of almost all dogs alive today.

Despite some researchers' belief that dogs were domesticated independently in the Old World and in the New, domestication may have happened only once, probably about 15,000 years ago. Dogs seem to have quickly become highly prized and were brought along by the settlers who reached North America via Beringia, the land bridge that spanned the Bering straits until the last ice age. This is the conclusion of a second study, based on DNA retrieved from ancient dog bones from Mexico, Bolivia and Peru, which found all the pre-Columbian dogs belong to Eurasian dog lineages.

A third study examines the psychology of dogs, showing that although chimpanzees may have brain power of far greater wattage, there is one task at which dogs excel, that of picking up cues from human behavior. This interpretive skill was perhaps the ability for which they were selected.

The origin of dogs, as judged by their mitochondrial DNA sequences, was first addressed five years ago by Dr. Robert K. Wayne and colleagues at UCLA. Wayne showed that dogs were indeed derived from wolves, as long suspected, but he set their date of origin as a separate population at 135,000 years ago.

Archaeologists found the date implausible because the earliest known dog bones date to only 14,000 years ago, more than 100,000 years later. Dr. Peter Savolainen, a former colleague of Wayne now at the Royal Institute of Technology in Stockholm, has now proposed a date that is more palatable to archaeologists. On the basis of DNA from several wolf populations and from the hairs collected off 654 dogs around the world, he calculates a much earlier date for domestication _ either of 40,000 years ago, if all dogs come from a single wolf, or of about 15,000 years ago, the date he prefers, if three animals drawn from the same population were the wolf Eves of the dog lineage.

Savolainen believes dogs originated from wolves somewhere in East Asia, because there is greater genetic diversity, often a sign of greater antiquity, in Asian dogs than in European dogs.

Separately, Wayne and another colleague, Dr. Jennifer Leonard, have analyzed the DNA of New World dogs, expecting to find they were domesticated by American Indians from local wolf populations. To exclude dogs brought from Europe, Leonard gathered pre-Columbian dog bones from archaeological sites and extracted their DNA. The samples matched Eurasian dogs, not American wolves, showing that dogs, of at least five lineages, must have been brought from the Old World to the New by pre-Columbian settlers.

These pre-Columbian dog lineages have disappeared. Even New World breeds of dog like the Eskimo dog, the Mexican hairless and the Chesapeake Bay retriever derive from dogs brought in from Europe. It is not clear why the pre-Columbian dogs were lost, but possibly American Indians preferred the European dogs for some reason and prevented their own dogs from breeding with them.

The dates yielded by dog DNA suggest that wolves were domesticated by hunter-gatherers, before the invention of agriculture and permanent human settlements. But domestication is an arduous process, in which animals must be selected for a particular trait through many generations. It is hard to see how hunter-gatherers could have foreseen the payoff from domesticating wolves, or would have known what traits to select for.

Two experiments bear on this puzzling issue. One was started by the late Dmitry K. Belyaev, a biologist at the Institute of Cytology and Genetics in Novosibirsk, Russia. He spent 26 years domesticating the silver fox, using tameability as the sole criterion of selection. Dr. Lyudmila Trut, who continued his work, reported recently that after selecting from 45,000 foxes over 40 years the institute now has 100 fully tame foxes. Tameability has brought with it other changes, like floppy ears and white-tipped tails.

Another experiment, reported in today's Science by Dr. Brian Hare of Harvard and colleagues, shows dogs have a special ability to pick up human cues. Chimpanzees will notice where a person is looking but do not take the hint that the box under gaze is the one that holds the hidden food. Dogs get the picture immediately, Hare reports. Wolves, though very smart, are much less adept than dogs at following human cues, suggesting dogs may have been selected for this ability.

So were dogs' ancestors selected for tameability or trainability? Ray Coppinger, a dog behavior expert at Hampshire College, believes that neither is the case. Wolves domesticated themselves, he argues in a recent book, Dogs, written with his wife, Lorna Coppinger. Wolves, which are scavengers as well as hunters, would have hung around the camp site for scraps, and those that learned to be less afraid of people survived and flourished, in his view.

"It was natural selection: The dogs did it, not people," Coppinger said. "The trouble with the theory that people domesticated dogs is that it requires thousands of dogs, just as Belyaev used thousands of foxes." From the half-tamed, camp-following wolves, he believes, people may then have adopted some cubs into the household and found they could be trained.

Hunter-gatherer peoples often bring back baby wild animals and keep them as pets until they become unmanageable. Dr. James Serpell, an expert on dog behavior at the University of Pennsylvania, believes this is more likely as an explanation of dog domestication than that people adopted scavengers. The particular population of East Asian wolves identified by Savolainen's genetic studies, Serpell suggests, might have had some special feature that made them easier to train.

Once dogs had been domesticated, they would have been of great value to hunter-gatherers, though it is hard to know what specific quality the domesticators sought. "They could have been useful as guard dogs, for hunting, as an emergency food supply, as bed warmers," Leonard said.

When two species live together for a long time, each usually influences the genetically conferred qualities of the other. People may have selected preferred abilities in the dog, but dogs too may have fostered their favorite qualities in people _ not of course deliberately but simply by giving people who used dogs a better chance of surviving than people who did not.

"This is a symbiotic relationship with substantial time depth," said Dr. Richard Klein, a Stanford University archaeologist. "You could imagine dogs would be useful for giving warning signals, or tracking other animals, so you can see how both sides would benefit."