Mystery Atlas
Debunked Myths

How Were Death Valley's Sailing Stones Finally Caught Moving?

Last updated 20 August 2026 · 7 min read

Direct Answer

For decades, boulders weighing up to several hundred kilograms left long trails across the dry lakebed of Death Valley's Racetrack Playa with no one ever witnessing them move, prompting theories involving wind alone, large sheets of lake ice, and even pranksters. Geologists Robert Sharp and Dwight Carey tracked and named 30 stones from 1968 to 1974, ruling out large rafting ice sheets for at least one tracked stone with a corral experiment, but the exact mechanism remained unconfirmed for another four decades. In December 2013, Richard and James Norris equipped rocks with GPS loggers and set up time-lapse cameras, directly recording the actual process: thin, wind-driven panels of ice, just a few millimetres thick, form after rare rain or snow and break up under light morning winds, pushing the rocks a few metres at a time.

Background

Racetrack Playa is a dry lakebed in the Panamint Mountains of Death Valley National Park, California, whose flat, cracked mud surface is scattered with dolomite and syenite boulders, some weighing several hundred kilograms, each trailing a long, curving furrow behind it, sometimes hundreds of metres long. The trails, some running roughly parallel to each other, others curving or doubling back, showed the rocks had moved across the playa, yet no scientist, park visitor, or researcher had ever directly witnessed one in motion, despite documented interest in the phenomenon dating to at least the 1940s. Early researchers, including geologist George M. Stanley, who published a 1955 study on the tracks, generally attributed the movement to wind acting on the playa's surface after rare rainfall left it wet and slippery, though the precise conditions required, and why some stones moved while others nearby did not, remained unclear.

In 1968, Caltech geologists Robert Sharp and Dwight Carey began a more systematic investigation, selecting, naming, and staking the positions of 30 stones, including one roughly 320-kilogram boulder they nicknamed "Karen," and returning to survey the playa roughly twice a year for seven years. Over that period they recorded substantial cumulative movement, in one case more than 260 metres for a single stone, but caught no movement directly in progress. To test whether the stones were being rafted by large, continuous sheets of lake ice, a leading theory at the time, Sharp and Carey built a corral of metal stakes around one tracked stone; when the stone later moved roughly 8.5 metres and emerged from the corral without disturbing the stakes, the finding suggested that if ice was involved, it could not be forming as one large sheet enclosing multiple rocks, since a large ice raft carrying the stone would have had to drag against or dislodge the corral itself.

The mechanism was not directly recorded until December 2013, when a team led by Richard D. Norris and James M. Norris installed a weather station, time-lapse cameras, and fifteen rocks fitted with custom GPS loggers on the playa, recording position and temperature data at regular intervals through the winter months when conditions were most likely to produce movement. During a visit that December, the team witnessed and recorded rocks moving in real time, publishing their findings in the journal PLOS ONE in August 2014.

Main Theories

The confirmed ice-panel mechanism

The Norris team's direct observations showed that after a rare rain or snowfall left a shallow pond of a few centimetres of water across parts of the playa, a subsequent cold, clear night could freeze the pond's surface into a thin sheet of ice, often no more than a few millimetres thick. As morning sun warmed the ice the following day, it began to melt and fracture into large, thin floating panels driven by light winds, typically around 4 to 5 metres per second. Where a floating ice panel encountered a partly submerged rock, the panel's edge could push against the rock with enough force to shove it slowly across the underlying soft, water-saturated mud, leaving a track as the rock moved, sometimes for a few minutes and occasionally for as long as roughly 16 minutes, at recorded speeds of 2 to 5 metres per minute. Because this combination of conditions, standing water, a hard overnight freeze, and a subsequent breezy, sunny morning, occurs only rarely at Racetrack Playa, most stones move just a handful of times across many years, which is why direct observation had eluded researchers for so long.

Superseded and competing explanations

Before the 2013-14 study, the leading alternative to a wind-and-ice mechanism was that large, continuous sheets of ice, thick enough to fully raft and float multiple rocks at once, moved across the playa surface during rare flooding and freezing events, carrying stones with them much as pack ice can carry embedded rocks and debris. Sharp and Carey's 1968-74 corral experiment provided the first strong field evidence against this specific version of the theory, at least for the stone in question, though it could not rule out an ice mechanism altogether, since the thin, localised ice panels the Norris team eventually documented were not the same phenomenon as a large rafting sheet. Popular, non-scientific speculation over the decades also floated magnetic fields, deliberate human pranks, and, at the more sensational end of amateur coverage, paranormal explanations; none of these attracted serious scientific investigation, given the consistent physical evidence, correlated tracks, weather-dependent timing, and, eventually, direct GPS and video confirmation, pointing to an ordinary, if unusually specific, physical process.

Common Misconceptions

A common misconception holds that the sailing stones remained a complete scientific mystery until 2013, with no serious research done in the interim. In fact, researchers had been studying the phenomenon since at least the 1940s, and Sharp and Carey's 1968-74 project had already narrowed the possible mechanisms considerably; the 2013-14 study's real achievement was catching the confirmed mechanism directly in the act, not originating the wind-and-ice explanation from nothing.

A second misconception treats every stone on the playa as equally likely to move. Field studies found that some stones, generally smaller and more angular, moved far more often and further than larger, smoother ones, and that a stone's specific position relative to the playa's subtle drainage patterns strongly influenced whether it was ever reached by a shallow pond deep enough to freeze and drive the mechanism at all.

Current Consensus

The mechanism behind Death Valley's sailing stones is now considered a solved scientific question, one of a small number of long-standing "impossible" physical phenomena that direct field instrumentation eventually confirmed rather than merely theorised about. The 2014 PLOS ONE study's GPS telemetry, time-lapse video, and weather data converged on a single, internally consistent explanation, thin, wind-driven ice panels acting on a saturated playa surface, that has not been seriously challenged by subsequent research. What remains is refinement rather than genuine dispute: exactly how often the necessary weather combination recurs, and why certain stones move dramatically more than their neighbours, are questions of degree within an already-confirmed mechanism, not open questions about whether the mechanism itself is correct.

Why This Mystery Endures

The sailing stones belong to a small category of once-baffling physical mysteries that science eventually solved not through a single flash of insight but through decades of patient, incremental fieldwork, each study narrowing the possibilities the next one would test. Ball lightning followed a similar arc, a phenomenon reported for centuries and taken seriously by scientists long before it could be reliably instrumented, confirmed as a genuine electrical phenomenon only once field studies and, eventually, spectroscopic recordings caught it directly. Both cases show that the gap between "documented but unexplained" and "solved" can be a matter of instrumentation and patience rather than the phenomenon itself being any less real in the meantime. Dyatlov Pass, resolved by a different kind of instrumented modern science, computational slab-avalanche modelling rather than direct observation, shows the same underlying pattern: a mid-20th-century field mystery finally converging on a mundane explanation only once researchers had the specific tools the original investigators lacked.

The case also endures as a small, reassuring counterexample within a site otherwise dominated by claims that remain genuinely open. Unlike the Bermuda Triangle or the Bloop, where competing explanations still coexist without full resolution, the sailing stones offer a clean example of the scientific method working exactly as intended: a testable hypothesis, a controlled field experiment that ruled out a competing theory, and, eventually, direct instrumented confirmation that left little room for ongoing dispute. That combination, a genuinely strange, decades-documented phenomenon with a fully confirmed mundane cause, is precisely what makes the case a useful point of comparison whenever this site discusses claims that have not yet reached the same resolution.

Frequently Asked Questions

How fast do the sailing stones actually move?
According to the 2013-14 GPS study, the fastest recorded rocks moved at 2 to 5 metres per minute, sustained for as little as a few minutes and up to about 16 minutes at a time, and only during specific weather windows: clear mornings following a sub-freezing night, when thin ice panels formed by a shallow overnight pond break up under light winds of around 4 to 5 metres per second. Individual stones can accumulate hundreds of metres of total track length over many separate movement events spread across years.
Did anyone ever seriously believe the stones moved by supernatural or alien means?
Not among scientists, though the phenomenon attracted plenty of popular speculation once photographs of the trails circulated from the 1940s onward, ranging from magnetic fields to deliberate hoaxes to, at the more sensational end of popular coverage, paranormal or extraterrestrial explanations. Every serious scientific study, from George Stanley's 1955 paper through Robert Sharp's 1968-74 field project to the 2014 GPS study, treated the movement as a physical, explainable process and focused on identifying the specific mechanism rather than entertaining supernatural causes.
Why did it take so long to solve if geologists had been studying the stones since the 1940s?
The stones move rarely, briefly, and unpredictably, requiring a specific and infrequent combination of rain or snow to form a shallow pond, followed by a cold, clear night to freeze it, followed by a sunny morning with light wind to break the resulting ice into moving panels. Robert Sharp and Dwight Carey's seven-year field study (1968-74) visited only twice a year and never caught a movement in progress, despite ruling out some competing theories with controlled experiments. It took the Norris team's continuous, unattended monitoring, GPS loggers recording rock positions every 60 minutes and weather instruments running through the winter months when the conditions occur, to finally record an actual movement event as it happened.

References

Connected to

How this topic links to the people, places, and ideas around it — drawn from our knowledge graph.

Related Mysteries

  • Yuba County FiveFebruary–June 1978

    California contains Yuba County Five.

  • Dyatlov Pass Incident is frequently compared to Somerton Man — Both are mid-20th-century unresolved-death cases where forensic and physical modelling narrowed, without fully closing, decades-old public speculation.

  • Levelland UFO Incident2-3 November 1957

    Ball Lightning is frequently compared to Levelland UFO Incident — The Air Force's and Menzel's proposed explanation for the case; disputed because no storm was recorded during the sighting window and because ball lightning has no accepted mechanism for stalling car engines or headlights.

  • Ball Lightning is frequently compared to Pioneer Anomaly — Both are precisely documented physical puzzles in this site's physics-theories cluster that briefly or repeatedly attracted exotic-physics proposals; the Pioneer anomaly yielded fully to archival data recovery, while ball lightning's central obstacle is that almost no natural event has ever been captured by instruments at all.

  • Dyatlov Pass Incident is frequently compared to Hinterkaifeck Murders — Both cold cases were reopened decades later by an institutional review (Dyatlov Pass by Russian prosecutors in 2019-20, Hinterkaifeck by German police cadets in 2007-09), each producing a leading explanation without full public consensus.

  • Ball Lightning is frequently compared to Earthquake Lights — Both are rare, contested luminous phenomena with a serious scientific mechanism proposed yet individual cases that remain difficult to verify against mundane alternatives.

People

Events

Organisations & Programmes

Historical Context

  • Dyatlov Pass Incident occurred during Cold War — The Soviet state's reflexive regional secrecy and the case's rapid 1959 closure fed decades of cover-up speculation.

Science & Technology

  • Ball Lightning is frequently compared to Zero-Point Energy — Both are genuine physics subjects in this site's physics-theories cluster where public confusion outruns the settled science, though in opposite directions: zero-point energy is a precisely understood phenomenon that attracts pseudoscientific extraction claims, while ball lightning's basic reality is not disputed but its underlying mechanism remains authentically unsettled.

Objects & Artifacts

  • Sailing Stones of Racetrack Playa is frequently compared to Klerksdorp Spheres — Both are natural objects whose motion or form was popularly read as impossible before geologists documented the mundane mechanism behind each.

Related Questions