On a perfectly flat, sun-baked lake bed in Death Valley, heavy rocks, some weighing hundreds of pounds, sit at the end of long trails scraped deep into the cracked earth. No footprints, no tire tracks, no machinery. For a century, the sailing stones of Racetrack Playa were one of geology's most stubborn mysteries: stones that apparently took a stroll whenever nobody was looking, sometimes traveling farther than a football field.

This episode follows the hundred-year stakeout that finally caught the culprit: from the first 1915 prospector's report, through the seven-year study that named rocks like Nancy (860 feet of travel at two and a half inches wide) and Karen (700 pounds, refused to budge), to the steel-rebar "jail cell" a one-pound rock casually escaped. The answer, captured at last by GPS trackers, is a slow-motion ballet of buoyant ice panels that only performs in a ten-second window of perfect conditions, and climate change and vandals may be closing the show.

• The crime scene: 700-pound dolomite blocks, 330-foot trails, and tracks that diverge like dogs bolting after squirrels
• Why the obvious wind theory failed, and how George Stanley's 1955 ice hypothesis got it half right
• The corral experiment: a geological jail cell built from rebar, and the rock that slipped right past it
• The GPS breakthrough: floating ice shoves, a perfect freeze-thaw recipe, and rocks moving in plain sight
• The new threats: a 4-to-1 decline in rock movement, an SUV joyride, stolen stones, and the volunteers who repaired the playa by hand