Roopkund Lake
A glacial lake in Uttarakhand ringed with several hundred human skeletons, reported by a forest ranger in 1942. Genome sequencing in 2019 showed they are not one population and did not die in one event.
Field Guide Entry
The ranger's report
A forest ranger, H K Madhwal, reported the bones in 1942, when the summer melt exposed them at about 16,500 feet. Wartime suspicion came first and pointed at Japanese soldiers moving through the mountains. The skeletons turned out to be far too old for that.
The explanation that settled in afterward was a single disaster. A large party caught in the open by a sudden hailstorm, killed by hailstones the size of cricket balls, the bodies kept by cold and thin air. A Garhwali folk song describes something close to it, a goddess striking down travelers who had offended her, and the Nanda Devi Raj Jat pilgrimage still passes nearby once every twelve years.
Three groups, a thousand years apart
In August 2019, a study in Nature Communications, led by Éadaoin Harney with collaborators at Harvard and the Max Planck Institute, published genome-wide data from 38 of the skeletons, with radiocarbon dates and isotope analysis alongside. The single-disaster account did not survive it.
The 38 separated into three groups. Twenty-three had ancestry in the range of present-day South Asians and dated to around 800 CE, and even those were not all deposited at the same time. Fourteen had ancestry typical of the eastern Mediterranean and dated to around 1800 CE. One had Southeast Asian-related ancestry.
The second group is the part nobody has accounted for. Fourteen people of eastern Mediterranean descent died at a high Himalayan lake roughly a thousand years after the first group, and no record of any such party has been found in any archive.
The lake now
Roopkund sits on a trekking route, and the bones surface for a few weeks a year as the ice pulls back. Decades of visitors have moved them, taken them, and rearranged them for photographs, which is part of why assembling a clean sample for the 2019 study was as hard as it was.
The 1800 CE group still has no origin, no name, and no documented reason to have been on that mountain.
Visiting Information
Reaching the lake is a multi-day trek at altitude, usually staged from Lohajung in Chamoli district, and it is not a day walk from a road head. The window is narrow: the approach is under deep snow for much of the year and the bones are only exposed for a few weeks after the melt.
Access is regulated by the Uttarakhand forest department and conditions have changed more than once in recent years, including periods of restriction, so the current position is worth confirming with the department or a registered operator rather than with a trekking blog. Registered guides and permits are the normal route in.
The remains are archaeological material in an open landscape with no barrier around them. They have been picked over for decades, which is a large part of why the site is as difficult to study as it is.
Field Gallery
No approved field photos yet. You could be first.
Community Notes
No community notes yet.
Explore More Mysteries
Arkaim
Found by accident in 1987 during a dam survey, this circular Bronze Age fortress in the Ural steppe was built with gates and doorways researchers say align to eighteen solar and lunar events, a claim, if accurate, of more precise astronomical design than Stonehenge.
Bada Valley Megaliths
Scattered across a remote Sulawesi highland valley, more than 400 carved stone figures, some over four meters tall, were made by a culture that left no name, no writing, and no agreed-upon date behind.
Actun Tunichil Muknal
Fourteen sets of human remains lie in the upper chamber where they were left more than a thousand years ago, some of them children. Pots sit among them with holes punched through the base so nobody could use them again. Cameras have been banned since 2012, when a visitor dropped one onto a skull.
Ale's Stones
Fifty-nine boulders stand in the outline of a ship on a ridge above the Baltic, the largest weighing about five metric tons, hauled up a slope to a spot with no shelter and no water. Radiocarbon evidence places the monument broadly in the later Iron Age, around 500 to 1000 CE. An amateur astronomer has spent decades arguing it is thousands of years older and is a solar calendar, and no working archaeologist accepts it.