The stones that move across Death Valley were filmed doing it in 2013

The trails are real

That is worth establishing before anything else, because this is a subject where a lot of the photography online has been enhanced past the point of honesty. It has not been here. Rocks on Racetrack Playa do sit at the ends of long furrows scored into the lakebed, and the furrows do run for hundreds of metres. The longest movement recorded by instruments was 224 metres in a single season.

Look at the trail in the photograph. It is cut into the mud, not scraped across baked ground. That detail turns out to be the whole answer.

Why it stayed unsolved so long

Researchers had been at this since the 1940s. The obstacle was not that the movement was inexplicable — it was that nobody had ever seen it. Every visit found rocks in new positions with fresh trails behind them, and no witness.

That produced exactly the theory-generating conditions you would expect. Hurricane-force winds were proposed. So were slick algal films, and thick rafts of ice buoying the rocks up and carrying them along. All of them were attempts to explain a lot of movement with a lot of force, because the trails looked like they needed a lot of force.

What actually happens

In winter 2011 a team installed GPS units inside fifteen rocks and a weather station on the playa, and then did the unglamorous part: they waited two years.

On 4 December 2013, and again on 20 December, and once more on 9 January 2014, the instruments caught it.

A few centimetres of water had pooled on the playa and frozen overnight into a sheet of ice between three and six millimetres thick — the researchers call it “windowpane” ice, and the name is exact. It is thin enough to see through and thin enough to break with your hand.

In the late morning the sun begins to melt it, and it fractures into floating panels tens of metres across. Then a light wind — four to five metres per second, which is not much more than a breeze — pushes those panels across the water. The panels meet the rocks and lean on them, and because a panel that size has a great deal of area and the mud beneath is soft and saturated, the rocks go.

Two to five metres per minute. Walking pace. On 20 December 2013, more than sixty rocks moved at once.

Nobody had guessed it because everybody had been looking for a large force. The actual answer is a very small force applied over a very large area, at the precise moment the ground has turned to grease.

The part that is still open

It would be easy to end there, and most coverage does. The National Park Service does not, and neither will we.

Richard Norris, who led the work, notes that the team never observed the largest rocks moving, and raises the question of whether the same mechanism accounts for them. That is a real gap. Five movement events in one unusually wet winter is a solid, direct observation of a mechanism. It is not proof that it is the only one that has ever operated here.

There is also a small discrepancy worth naming: the Park Service’s account describes the December ice-breakup on the 21st, while the published paper records movement on the 20th. We have followed the paper, because it is the primary record. We mention it because a file that quietly picks one date and hides the other is doing the same thing we criticise elsewhere.

Why the claim is Misleading rather than false

Every component of “nobody has ever seen them move and scientists can’t explain it” was accurate for about seventy years. It was not invented, and the people repeating it are not being credulous — they are repeating something that used to be true and was never updated in their feed.

That is the most common failure mode in this entire subject area, and it is worth naming plainly: most internet mysteries are not lies. They are correct statements that expired.