By Paul Rincon
Science reporter, BBC News, Houston
A meteorite which ploughed into the Peruvian countryside last year should have shattered and dispersed long before reaching the ground.
That is the conclusion of scientists who have been examining samples of the space rock and the 15m-wide crater it dug out in Carancas last September.
The discovery of a water-filled hole, following reports of a fireball in the sky, made headlines around the world.
Now experts say the event challenges conventional theories about meteorites.
This has nothing to do with the mass panic that famously followed the impact; rather it has to do with the science of space impacts.
Usually, only meteorites made of metal survive the passage through Earth's atmosphere sufficiently intact to scoop out a crater.
But the object which came down in the Puno region of Peru was a relatively fragile stony meteorite. During the fiery descent through Earth's atmosphere, these are thought to fragment into smaller pieces which then scatter over a wide area.
Yet pieces of the estimated 1m-wide meteorite are thought to have stayed together during entry, hitting the ground as one.
Details of the work were unveiled at the Lunar and Planetary Science Conference in Houston, Texas.
Peter Schultz told the conference that the meteorite was travelling at about 24,000km/h (15,000mph) at the moment of impact - much faster than would be expected.
"This just isn't what we expected," said Professor Schultz, from Brown University in Providence, US. "It was to the point that many thought this was fake. It was completely inconsistent with our understanding of how stony meteorites act."
Typically, fragments shoot off in many directions as the meteorite hurtles towards the ground - the so-called "pancake" model of atmospheric descent.
Professor Schultz believes fragments from the Carancas meteorite, which crashed to Earth on 15 September last year, may have stayed within the speeding fireball until they struck the ground.
This might have been due to the meteorite's high speed.
At the velocity it was travelling, fragments could not escape the "shock-wave" barrier which accompanies the meteorite's passage through the atmosphere.
Instead, the fragments may have reconstituted themselves into another shape, which made them more aerodynamic. Consequently, they encountered less friction during their plunge to Earth, holding together until they reached the ground.
"Although [the meteorite] is quickly broken up, it is behaving like a solid mass," Professor Schultz told the conference.
Dr Thomas Kenkmann, from Humboldt University in Berlin, Germany, offered an alternative view of the Carancas impact. His modelling of the event suggests it was probably caused by a meteorite travelling at low speed and a slanting angle.
Under this scenario, the space rock would have broken into just a few pieces rather than many - the largest of which would have gouged the crater.
Some scientists, however, remain doubtful of either interpretation. After his conference talk, Dr Kenkmann was pressed by one scientist on whether a magnetic survey of the crater had been carried out to look for signs of an iron, rather than a stony, projectile.
Fact and fiction
The crater is located on a dry riverbed near Lake Titicaca, on the border with Bolivia. The 3m-deep depression filled with water from below ground in the first 15-30 minutes after the impact.
At the time, scores of local people who visited the crater complained of headaches, vomiting and nausea. Some commentators had speculated that a chemical reaction might have released toxins such as sulphur and arsenic. But mass hysteria is thought to be the most likely explanation.
Professor Schultz said he hoped his work would "distinguish fact from fiction".
He commented: "Reports about arsenic, bubbling [of water in the crater] and sickness were probably overblown. People were frightened, but some were also afraid they were under attack from a nearby country."
Eyewitnesses reported a cloud of dust travelling outwards from the impact site after the meteorite fell.
Reports of numerous livestock deaths are believed to have been exaggerated, though the researchers confirm that a bull's horn was ripped off in the impact.
Professor Schultz added that the Carancas event raised the possibility there were many other small craters caused by stony meteorites which go unrecognised.
Large buried iron meteorites are easier to detect, while pieces of stony meteorite become intimately mixed with terrestrial soil.
"Maybe there are more of these things and we just don't recognise them because they're rock. When these things get pounded into cement, you won't see them," Professor Schultz explained.