© Getty ImagesThe 33,000-year-old remains of an animal in Siberia suggest it was partly domesticated. Its bones suggest it resembled the modern Samoyed dog, shown here.

Some dogs were domesticated by at least 33,000 years ago, but these canines did not generate descendants that survived past the Ice Age, suggests a new PLoS ONE study.

The theory, based on analysis of a 33,000-year-old animal that may have been a partly domesticated dog, explains why the remains of possible prehistoric dogs date to such early periods, and yet all modern dogs appear to be descended from ancestors that lived at the end of the Ice Age 17,000-14,000 years ago.

The ancient animal identified as being a partly domesticated dog was found in Razboinichya Cave in the Altai Mountains of southern Siberia.

"The Razboinichya dog find demonstrates that the right wolf/human conditions suitable for getting domestication started were present at least 33,000 years ago," co-author Susan Crockford told Discovery News. "However, such conditions would have had to be present continuously -- stable -- for many wolf generations, perhaps 20 over about 40 years for the domestication process to generate a true dog."

"It appears that such stable conditions were not present until after the Ice Age, sometime after 19,000 years ago," added Crockford, a researcher at Pacific Identifications Inc. and author of the book Rhythms of Life. "Even after the Ice Age, domestication of wolves could have got started at several different times and places, and still failed because the conditions were not continuous enough for the changes to become permanent."

The Siberian animal was unearthed some years ago, but was only recently dated to 33,000 years ago by three independent radiocarbon dating facilities. Crockford and her colleagues conclude that it was a partly domesticated dog because of its mixture of dog and wolf features.

Based on its skull and other remains, the scientists believe it was about the size and shape of a large male Samoyed dog. Its teeth were still wolf-sized, however, and "it probably behaved more like a wolf than a dog."

Its remains were excavated from a cave area containing wild animal bones. Usually fully domesticated dogs, even very early ones, received more careful burials, often being placed in graves with, or next to, their owners.

Since no other dog-like animals were found at the site, the researchers think this animal was an "incipient" dog in the early stages of domestication. The scientists hold that domestication can happen naturally, without direct human intervention, when wolves are attracted to settlements and gradually adjust to a human-dependent lifestyle.

The Ice Age, however, changed the abundance and migration patterns of the animals that the people in the Altai Mountains of Siberia hunted for food.

"As a result, the people probably had to move more often than they did before," she explained.

Without the conditions that fuel domestication, the dog or dog-like animals gradually died off, the researchers suspect. Dogs reemerged after the Ice Age, reproducing and becoming the ancestors to today's modern dogs. It is unclear when the first pre-Ice Age dogs emerged, but a dog-like skull dating to 36,500 years ago was found at Goyet Cave in Belgium. It's possible then that the first dogs appeared in parts of Europe and Asia much earlier than commonly thought.

Other experts contacted by Discovery were interested in the new study, but would like to see more specimens to strengthen the theory. For example, Keith Dobney, chair of human palaeoecology at the University of Aberdeen's Department of Archaeology, said, "This is a very interesting and potentially important find -- potentially pushing the domestication of the dog much further back."

Without more specimens, however, he said it cannot be ruled out that the Siberian dog, and possibly some of the other pre-Ice Age animals, were different representatives of now-extinct wolves.

Richard Meadow, director of the Zooarchaeology Laboratory at Harvard University's Peabody Museum, echoed Dobney's reservations about the study's conclusions.

Crockford admits that the paper presents "a new way of thinking about domestication, but it fits the evidence better than the idea that people deliberately created dogs for some specific purpose."