By Martin Redfern
BBC radio science unit
It was a sensational discovery - human footprints said to be 40,000 years old, preserved by volcanic ash in an abandoned quarry in Mexico.
The announcement, in July last year, created a flurry of excitement, but was then promptly dismissed by a second team of researchers who re-dated the rocks at 1.3 million years old, impossibly ancient to bear human traces.
The original claim has not gone away, however.
The first widespread evidence for the human occupation of North America came from the town of Clovis in New Mexico.
The beautiful fluted stone-spearpoints made by the Clovis people are found on many sites and date back 11,500 years or so. They are believed to have been left by people who crossed a land bridge that once existed between Siberia and Alaska.
But there is an increasing body of evidence for earlier occupation of the Americas, dating back to a time when the overland route through the ice would have been impossible.
The best evidence probably comes from Monte Verde in Chile and dates back at least 12,500 years. To have reached so far south by then, people must have entered the continent earlier still.
There have been many claims of earlier dates, but few have been substantiated. So the announcement of 40,000-year-old footprints from Mexico was greeted with scepticism and caution.
It came from a team led by Silvia Gonzalez, a Mexican working at Liverpool John Moores University (JMU), UK.
In 2003, she was visiting a site at Cerro Toluquilla in the Valsequillo Basin. It is near Puebla, about 100km southeast of Mexico City.
It is a dry environment with many small volcanoes and, in the distance, the smoking peak of Popocatepetl. She was hoping to find the geological context of deposits that had yielded animal bones showing possible butchery marks and dating back 20 or 30,000 years.
The researchers were looking for a vertical section through the rocks in the side of a small quarry, but it was overgrown and strewn with debris.
As they were about to give up, they noticed that the floor of the quarry was made of a single layer of hardened volcanic ash called the xalnene tuff. It looks a bit like a badly asphalted car park.
Silvia Gonzalez had studied much younger human tracks in Lancashire and thought she could see similar markings in the volcanic ash.
"Suddenly, I began to see some marks on the top surface of the ash... and I recognised them as human footprints," she told the BBC's Unearthing Mysteries programme.
"I felt quite shocked, because I knew already that this ash was very old."
At first, her colleagues laughed at her, but soon they were brushing away the dust with excitement.
A large area of the quarry has now been cleared and the researchers are making a detailed digital laser survey of the marks.
Some are clearly animal tracks left by perhaps deer and buffalo and running for several metres.
Others do seem to resemble human footprints though there are few in a continuous track. They are of several sizes suggesting both adults and children.
The ash is too coarse to have left clear toe prints but some certainly appear to be from left or right feet with a raised arch and material bunched up behind the toes as the person pushed forwards up the gentle slope.
The famous Popocatepetl volcano is visible in the distance
Gonzalez' colleague at JMU, Professor Dave Huddart, demonstrated the likeness to his own feet: "If I put my foot beside it, size 8½, it looks a typical size; it's got the characteristic figure-of-eight shape and the big toe is there, so it's a left foot."
Professor Mike Waters of Texas A&M University is sceptical. He thinks the marks are products of the quarrying process and subsequent erosion. There are certainly some marks at the site that are very obviously due to that and seem much fresher - pick marks and tracks from mechanical diggers.
More controversial still are the dates. Colleagues of Silvia Gonzalez at Oxford University used a technique called optically stimulated luminescence (OSL) that records the last time rocks were exposed to sunlight or heat.
That gave a variety of dates from the overlying sediments, but when applied to small fragments of what looked like brick or burnt clay within the volcanic ash, it produced a date of about 40,000 years.
That initially shocked Dr Gonzalez as it implied by far the earliest evidence of humans in the Americas. But it fitted in with dates of up to 38,000 years based on carbon 14 in shells in the sediments above.
POSSIBLE MIGRATION ROUTES
(1) Kennewick Man remains are about 8,800 years old
(2) Santa Island Rosa bones date back 10,960 years
(3) Clovis site with stone spearpoints - 11,500 years ago
(4) Toluquilla 'footprints' are claimed to be 40,000 years old
(5) Peñon Woman III in California - 10,800 years ago
(6) Monte Verde site in Chile - 12,500 years ago
In December, however, Paul Renne of the Berkley Geochronology Center in California published dates for the volcanic ash itself based on the powerful argon-argon technique.
That gave an age of 1.3 million years, far too old to be compatible with human footprints.
"I don't think that they are [footprints]," he told the BBC. "There are no trails of footprints that are consistent, the shapes don't really look like footprints, and, most importantly, there's a huge diversity of shapes, sizes and arrangements of these things."
Professor Renne measured the age of several different grains in the ash and got the same age for each. Dr Gonzalez, though, says that the volcano responsible is complex.
It interrupted explosively underneath a lake and lots of older material and lake sediment may have been caught up in the ash, distorting the date.
Furthermore, there do not seem to be the signs of erosion and weathering that would be expected if there had been a gap of more than a million years between the ash and the overlying sediments.
Professor Renne also looked at the magnetisation of the rocks, partly to see if they might have been jumbled up and redeposited from an earlier material, which, he says, they were not.
But he did find that the magnetic polarity was the opposite of the Earth's present magnetic field. The Earth's magnetic poles do occasionally flip.
BBC presenter Aubrey Manning with Silvia Gonzalez
"The last time the Earth's magnetic field had consistently reversed polarity was about 790,000 years ago, so the fact that we found reversed polarity magnetisation in this rock tells us that it's older than 790,000 years," he said.
Silvia Gonzalez' view? "We know that there are short-term 'excursions' of the magnetic field, and one of those happened 40,000 years ago, very interestingly."
Professor Renne: "How did I know they were going to say that? There is a finite possibility that that is correct, but the probability is extremely low."
It seems this debate really is going to run and run.
To answer the criticisms, Dr Gonzalez and her colleagues hope now to get permission to excavate for further footprints that would not be associated with any quarrying marks and to get more secure and consistent dates for the rocks.
"That would convince even the most intense critics," she said. "We need to talk to each other to make a continental model of human migration across the Americas. It won't be done in a few years. It will take a lifetime, but we are not afraid to do that."
If she succeeds, this little quarry could become one of the most important archaeological sites in the Americas.
A final comment from Professor David Meltzer, from Southern Methodist University, Dallas. He has researched and written extensively on the subject of the "first Americans". He told the BBC: "I'm not averse to the idea of 40 000-year-old people in the New World - but I'm sceptical because we've been fooled before.
"We want to see it confirmed with all the evidence laid out so that we're not buying in to something that isn't there."
A BBC Radio 4 team has been to the site to assess the evidence for themselves. Unearthing Mysteries with Aubrey Manning was broadcast on the network on Tuesday, 17 January. It has now been archived on the Unearthing Mysteries website