For decades, Federico Solorzano has gathered old bones from the shores of Mexico's largest lake -- bones he found and bones he was brought, bones of beasts and bones of men.
The longtime teacher of anthropology and paleontology was sifting through his collection one day when he noticed some that didn't seem to fit: a mineral-darkened piece of brow ridge bone and a bit of jaw that didn't match any modern skulls.
But Solorzano found a perfect fit when he placed the brow against a model of the Old World's Tautavel Man -- member of a species, Homo erectus, that many believe was an ancestor of modern homo sapiens.
The catch: Homo erectus is believed to have died out 100,000 to 200,000 years ago -- tens of thousands of years before men are believed to have reached the Americas.
And archeologists have never found a trace of Homo erectus in the Americas.
"Most people sort of just shook their heads and have been baffled by it," said Robson Bonnichsen, director of the Center for the Study of the First Americans at Texas A&M; University. "That doesn't mean it's not real. It just means there's not any comparative evidence."
Solorzano's find was described at a September conference here that drew academics from Europe and the Americas to discuss new research on early man in the Americas.
That primitive brow ridge from Lake Chapala "is in a category by itself," Bonnichsen said.
It is so strange -- and so out of context -- that it has been largely ignored even as other discoveries are raising basic questions about the story of human beings in the Americas: when they arrived and where they came from.
Until recently, most U.S. archeologists believed the first Americans arrived about 13,500 years ago when a temporary land corridor opened across the Bering Strait. The migrant Clovis people, named for a site near Clovis, N.M., apparently hunted mammoths and other large animals, leaving scatterings of finely worked spear tips and other tools across North America and, some argue, South America.
A sometimes vehement minority still holds to that "Clovis first" position. The evidence of what could have come before remains sparse, scattered and controversial. Archeologists have proposed possible alternative routes to the Americas -- across the Pacific from Asia or Australia, across the Atlantic from Europe or Africa -- although most say a trip from northeast Asia is most likely, perhaps by people advancing along a frozen coast in small boats.
South American researchers say they have found numerous sites that are 10,000 to 15,000 years old, and argue that Clovis people could not have migrated to Tierra del Fuego, at the southern tip of South America, so soon after the ice-free corridor opened from Asia to Alaska.
Argentine archeologist Laura Miotti agrees that the settlers likely came from the north. But she and others say there are no Clovis-like finds in the part of Asia from which the migrants supposedly came, and they question why North American sites don't appear to be older than those in South America.
The evidence for earlier human habitation in the Americas, however scanty, is tantalizing. It includes:
* A possible hand scraper splotched with blood more than 34,000 years ago at Monte Verde in Chile.
* Possible stone tools at a site in Brazil that is 40,000 to 50,000 years old.
* A not-yet-published report of human remains dated as much as 28,000 years old near Puebla in central Mexico.
Most crucially, a majority of archeologists are convinced that a second site at Monte Verde dates to at least 14,000 years ago -- some 500 years before the land bridge from Asia opened more than 9,000 miles to the north.
Yet the early dates are still often questioned. A claim of 250,000-year-old human tools near Mexico's Valsequillo reservoir was widely laughed at in the 1970s, although researchers are once again working at that area.
Clovis-first advocates suggest that the early dates may reflect variations or errors in the still-developing technologies of dating old samples.
They say that natural breakage could account for some of what look like early tools and that the dating of others was likely confused, as when streams, floods or human beings mix new material into old.
As for human remains, only two teeth in Brazil seem to have been directly dated to pre-Clovis times.
"If you are trying to break through a barrier that is well established, you need well documented, incontrovertible proof," said archeologist Stuart Fiedel, author of a textbook on early Americans and a proponent of the Clovis-first model.
Both sides say new research on DNA and climate history supports their claims, or at least fails to undermine them.
Solorzano's finds raise so many unanswerable questions that they have remained a curiosity.
Solorzano, 83, is a respected researcher who has taught generations of university students in the city of Guadalajara. His home office has a cabinet full of bones -- some of them human -- topped by 14 realistic models of hominid skulls.
He says the brow bone raises "many questions, one of them being its great and amazing resemblance to primitive hominid forms whose presence in the Americas has not been generally accepted."
The few additional scientists who have analyzed the bones closely agree that they look human -- not animal -- and are very, very old.
"They were definitely human," said Joel Irish, a specialist in bioarcheology at the University of Alaska, Fairbanks.
He suggested that they could be from "a very primitive-looking modern human," but said they would be "very early."
Efforts to date the pieces using modern techniques so far have failed because of a lack of surviving tissue.
Most frustrating for archeologists, who are accustomed to fussing over the tiniest details, is that nobody knows quite where the bone came from or even when it was found. It was apparently picked up when drought exposed a large ring of the Chapala lakebed from 1947 to 1956.
Archeologist Stanley Davis, then at Texas A&M;, spent several seasons accompanying Solorzano on surveys of the region and said he located places that he would like to investigate further.
"It takes a lot of money. That's the reason I'm not down there working right now," he said.
Davis said that in the same area, other human bones that are about 6,000 to 7,000 years old lack the mineralized darkness of age found in the brow and jaw pieces.
Davis said the Chapala area was interesting because the lake was very old and was a likely spot for coast-hopping migrants to come inland.
Yet relatively few people have investigated the area. Until recently, Mexican archeologists tended to focus on the spectacular indigenous cultures of the Olmecs, Mayas, Aztecs and others that arose in the last 3,000 years or so.
Davis said the Chapala-area finds included 12 scattered skulls of a long-extinct horse species. All have been smashed between the eyes.
"Either we have a herd of very stupid horses ... or we have some other action responsible for their death. That action is probably human," he said. He estimated that the horses were probably 10,000 to 20,000 years old.
A cache of swamp-deer teeth included several that were grooved, apparently for use in a necklace, he said. A radio carbon test showed that one was roughly 20,000 years old. "That tells us we may have something."