Feldspar grains date South Dakota's Mammoth Site to at least 160,000 years

Bone apatite is the mineral part of bone, and in groundwater it does not hold its carbon. Emma Krolczyk, a geologist with the U.S. Geological Survey, said the old carbon in apatite can dissolve and recrystallize with new carbon from the surrounding groundwater ^[10]. "This essentially resets the radiocarbon signal," she said ^[11]. The sinkhole is a good setting for that: four to five metres deep, fed by water at 32 to 35 degrees Celsius, with sides too slick for a mammoth to climb ^[4]. The bones held no collagen, the protein radiocarbon labs normally date, so apatite was what the early work had, and the study notes apatite is notorious for dubious ages in water settings like this one ^[5].

How wrong the old chronology was depends on which pair of numbers you set side by side. Divide the oldest luminescence age, 232,000 years, by the youngest apatite date, 23,000, and the error is a factor of about ten ^[13]. Divide the youngest luminescence age, 160,000, by the oldest apatite date, 42,000, and it is under four ^[14]. The six apatite dates also disagree among themselves by 19,000 years, nearly two to one ^[15].

The cleanest check came from saturated grains. Quartz, the usual mineral for luminescence, had already absorbed as much radiation as it could hold and yielded only a minimum age of around 75,000 years ^[8]. That floor sits 33,000 years above the oldest radiocarbon date ^[16]. Feldspar saturates later, and it is what produced the 160,000 to 232,000 range ^[2].

Luminescence dates sediment, not bone. It measures when a mineral grain was last in sunlight, using the radiation the grain absorbed while buried ^[7], so these ages describe the sand and carry over to the animals only if bone and sand were buried together. One sample from the middle of the sinkhole came back roughly 40,000 years younger than the samples around it ^[17]. Paul Hanson of the University of Nebraska-Lincoln said a number of things could explain that, and pinning down one is hard ^[18]. "There is a good reason that we don't rely on just one or two ages ...," he said ^[19].

Jim Mead, an emeritus researcher at the Mammoth Site and the University of Arizona, had doubted the apatite dates for years, having worked as a graduate student in the university's radiocarbon lab and found them unusual ^[20]. "I have always wanted to get a different chronology method to date the site ...," he said, "and a decade later here we are" ^[21]. The sampling was done inside the museum built over the excavation, where visitors could walk up and talk to the researchers ^[22]. Shannon Mahan, a research geologist with the USGS, said: "It was the purest form of science education because it did not matter what that person's religion was, what their political affiliation was, or even their educational level" ^[23].

The older ages put the mammoths in a deeper and probably colder part of the Pleistocene, and make this one of very few sites dated that far back on the Northern Great Plains ^[24]. The exchange Krolczyk describes needs only wet bone and time, so the same failure is available at any spring or sinkhole deposit where collagen is gone. The study does not say whether other chronologies in the region currently rest on apatite dates ^[25].