By JOHN NOBLE WILFORD

Scientists have found what they say is the first direct evidence of cannibalism among prehistoric Indians in the American Southwest, belying the image of these people as steadfastly peaceful farmers.

The finding may well reignite a long-smoldering controversy over whether ancestral Indians ever made it a practice to eat human flesh, a conclusion deeply resented by their descendants.

The latest evidence sprang from the site of an ancient Anasazi settlement near Mesa Verde in southwestern Colorado, where archaeologists came upon butchered human bones and stone cutting tools stained with human blood. A ceramic cooking pot held residues of human tissues. But the most telling of the evidence was found in human feces: biochemical tests revealed clear traces of digested human muscle protein in the dried coprolite, or fossilized prehistoric feces.

''Analysis of the coprolite and associated remains at last provides definitive evidence for sporadic cannibalism in the Southwest,'' Dr. Brian R. Billman, an archaeologist at the University of North Carolina, said yesterday in an announcement of the discovery.
 

The biochemist who conducted the tests, Dr. Richard A. Marlar of the University of Colorado School of Medicine in Denver, said the only way this particular human protein, myoglobin, would be present in the feces was if the person had eaten human flesh. The tests and other research were repeated many times over the last two years.

''We felt we had to be absolutely sure on such a sensitive and controversial issue,'' Dr. Marlar said in an interview. ''I have no doubt about this data.''

The research team -- Dr. Marlar, Dr. Billman and colleagues -- described and interpreted the findings in a report being published in today's issue of the journal Nature. In an accompanying commentary, Dr. Jared M. Diamond, an anthropologist at the University of California at Los Angeles, called the research ''compelling evidence'' of the practice of cannibalism by some ancestral Indians.

The chief descendants of the Anasazi are the Hopi and Zuni tribes, which have never practiced cannibalism and object to archaeologists' inferences of cannibalism by their prehistoric ancestors.

But Dr. Donald Grayson, an anthropologist at the University of Washington who specializes in prehistoric people in the American West, said in an interview that he was not surprised by the biochemical finding. ''I was already convinced by the evidence of butchering marks and bones from archaeological sites'' elsewhere, he said.

Archaeologists think that cannibalism among people in the American Southwest occurred between A.D. 900 and 1150 but was then fairly rare, probably occurring when the community was faced with starvation. After that, for about 50 years, the evidence suggests, there was an outbreak of what is known among anthropologists as ''customary cannibalism.'' Dr. Diamond described this consumption of human flesh, from either deceased relatives or slain enemies, as a nonemergency practice and not just what a group might resort to in the face of starvation as the Donner party of American pioneers did in the winter of 1846-47.

The new findings are hardly the first indication of cannibalism among the Anasazi. Butchered human bones and other mounting evidence were described by Dr. Christy Turner 2d, an anthropologist at Arizona State University, in his book, ''Man Corn: Cannibalism and Violence in the Prehistoric American Southwest.'' Dr. Turner is on an expedition to Siberia and could not be reached for comment on the new findings.

But the evidence is still disputed by some anthropologists and especially by modern American Indians. Archaeologists who previously reported suspected cannibalism said they were sometimes accused of contributing to racism and genocide; some early European colonists used stories of Indian cannibalism as justification for their conquests of the land.

The debate, now decades old, intensified in 1979 when Dr. William Arens, an anthropologist at the State University of New York at Stony Brook, argued in an influential book, ''The Man-Eating Myth,'' that there had been no reliable firsthand accounts of cannibalism. So he doubted the very existence of the practice, except perhaps in circumstances of near-starvation.

Efforts to reach Dr. Arens for comment yesterday were unsuccessful. But in a 1998 article in The New Yorker, he was quoted as having apparently changed his mind. Asked about Dr. Turner's research, Dr. Arens said cannibalism was a ''possible interpretation, even a good interpretation.''

But all previous evidence of occasional Indian cannibalism, though increasingly persuasive, was still circumstantial. The discovery of the human feces, in an ancient dwelling in Colorado, opened the way to gathering direct evidence through biochemical analysis.

The prehistoric feces were found as a result of a preservation effort by the Ute tribe. When archaeological sites in the foothills of Sleeping Ute Mountain were about to be destroyed by newly irrigated agricultural fields, the Utes called in Soil Systems Inc., an archaeological excavation company in Phoenix. Dr. Billman, who worked for the company then, led the investigation of a community of 70 to 125 people that existed there between A.D. 1130 and 1150.

It was a time of severe drought, one of the worst in the American West in the last 2,000 years, and this provoked conflicts between farmers who had settled there and intruders from surrounding areas.

''We believe the entire community was extinguished in a single episode of violence and terroristic cannibalism during a period of social chaos brought on by the drought,'' Dr. Billman said.

The raiders responsible for the killing may have considered the local farmers themselves to be intruders on withered farming land who had to be driven away. After killing, cooking and eating seven people in that one dwelling, one of the raiders squatted over the cold ashes of the hearth and defecated, a parting sign of contempt.

When Dr. Billman gave that account at an archaeological meeting, Dr. Marlar, an amateur archaeologist as well as a biochemist, offered to try to develop a technique for testing the feces for any traces of human tissues. Biochemical assay techniques had been used before to identify animal meat residues in cooking pots at archaeological sites. And the presence of the human myoglobin, the headline on Nature's commentary said, was ''Incontrovertible Evidence of Cannibalism.''