TREKKING THROUGH REMOTE NORTHeastern Tibet last September on a field study of the region's distinctive wild horses, French anthropologist Michel Peissel and his colleagues were raked by icy winds and hailstorms as they followed precarious trails through the mountains. Unusually deep snowdrifts blocked many passes, forcing the expedition to detour through the isolated terrain in the Riwoche region. In early October they reached the rim of a secluded, steep-sided valley.

There they saw, grazing between the trees, a herd of 20 to 25 horses like none Peissel, a leading expert in the species, had ever seen before. They bore a striking resemblance to the horses depicted in prehistoric cave drawings, which are thought to be extinct. Peissel and his fellow expeditionists quickly realized they had perhaps made an astonishing discovery: at the very least, a breed of horse that had never been identified before. Some suspect that this may be a prehistoric missing link in equine evolution.

On closer inspection the scientists observed that the horses differed greatly from breeds found nearby, such as the Nangchen ponies, which Peissel had discovered in 1994. "The beige coat, black and bristly mane and the stripes on its back legs and back are similar to [features of] the most ancient breeds we know," says Peissel. "The angular shape of the body, and the head in particular, is like that of the horses found in the Stone Age cave paintings."

Enthralled by their prize, the team spent two days taking photographs and making notes about the animals, which they named Riwoche horses, before continuing on to the Tibetan capital, Lhasa. Peissel decided not to bring any of the horses back to Europe, preferring instead to prepare a future expedition for that task. The valley's residents are the Bon-po people, pre-Buddhist pagans who round up the horses when they are needed for work. The team managed to extract blood samples for DNA testing to determine what relationship the Riwoche has to other members of the Equus genus--and just how far back in the horse's evolutionary chain the newly discovered breed belongs. Scientists believe that the modern horse is 5 million years old.

Could the Riwoche be yet another ancient species? "The Riwoche horse is primitive and has similar characteristics to the Przewalski's horse, but is certainly unrelated," Peissel says. "I think you have to treat the two as entirely different."

One factor that supports Peissel's assertion is the location of the herd. In the 27-km-long valley where the horses roam, the team also spotted rare white-lipped deer. "The fact that this forest is in such a secluded area and harbors such rare species of animals makes me think that it may be a remnant of what was once a large ecosystem," Peissel says. He theorizes that the roughly 5,000-m altitude of the passes leading into the valley created a "lost world" that has sealed the Riwoche off from other breeds for thousands of years and has therefore allowed the horse to remain unevolved from its relatively primitive form. "The creation and continuation of breeds arises from isolation," Peissel says, "and this may well be the most secluded area in the world."

Neither this isolation nor the unusual physical characteristics, however, definitively prove that the breed originated in ancient times. Stephen Harrison, a geneticist at Britain's Royal Agricultural College who will perform the DNA testing on the Riwoche, cautions that any modern breed of horse if left in the wild for a few generations might revert back to the primitive coloring--the dark dorsal stripe and zebra pattern on the legs. He also points out that a number of known breeds resemble wild horses. Geneticists have found some domesticated Mongolian riding ponies that carry genetic markers for Przewalski's horses, as if they had been crossbred. "It would be premature to say these [Riwoche] horses are a new species," Harrison says. "Without tests you cannot tell whether it's a population of wild horses that have evolved in isolation or a feral population that was once domesticated and has gone wild. The Chinese, of course, were great horsemen even before we had horses in Europe."

Whatever the DNA tests show, Peissel is sure of one thing: finding the Riwoche so completely by chance offers hope for similar new discoveries, even in a modern world in which there is little left unexplored. In 1994, on his previous expedition into Tibet, in addition to discovering the Nangchen ponies, Peissel chanced upon what he says is the source of the Mekong River. "True exploration involves setting out to find one thing and discovering another," he says. "After all, if you set out to discover something you are already convinced is out there waiting, it's not really a discovery when you find it."

--Reported by Bruce Crumley/Paris and Helen Gibson/London