The Hadza of Tanzania and the Tsimane of the Bolivian Amazon share 1,231 gut microbial species. Their ancestors parted tens of thousands of years ago. They do not share a diet, a river, or a recent contact. Justin Sonnenburg and colleagues at Stanford report the overlap in Nature, published October 7, 2026. Of the species recovered from the Tsimane, 87.4 percent also turned up in the Hadza. Most of that shared set is scarce in industrialized guts. The paper counts 848 of the 1,231 species, 60.2 percent, as rare or absent from people in industrial societies. The finding shows more than 600 species that may once have been common and are now missing from most societies. The 848 are the shared species that fail to appear at any useful rate in industrial samples. The 636 are the subset the team could push further, each one yielding at least four reconstructed genomes in each population. The Tsimane samples were collected by the Tsimane Health and Life History Project. The Hadza comparison rests on the Stanford group’s earlier deep sequencing, published in Cell in 2023: 351 stool samples from 167 Hadza people, taken in 2013 and 2014. In that earlier work the average Hadza gut carried about 730 species and the average Californian gut about 277. A species-accumulation curve in the new paper estimates a population ceiling of 2,443 species for the Hadza (95 percent confidence interval 2,385 to 2,503) and 1,590 for the Tsimane (1,561 to 1,620). The sampled Tsimane individuals together held about 1,400 species. The team then asked whether the overlap was inheritance or a coincidence of life without antibiotics and refrigeration. They looked at strain-level mutations inside the 636 species. Of those species, 545, or 85.6 percent, showed no recent strain sharing between Tanzania and Bolivia. Reproductive isolation between the two sets of strains was stronger than the isolation seen in comparable microbes from industrial populations. A dating method that accounts for bacterial recombination put the median last genetic exchange between Tsimane and Hadza strains at 17,090 years, with a standard error of 863 years. Exchange between industrial populations was far more recent: a median of 1,109 years between Europe and North America, and 4,637 years between Asia and North America. The difference was large (P < 10⁻⁴). Seventeen thousand years is not the date humans left Africa. It sits near the later crossing into the Americas. The paper says the split times approximately align with migration out of Africa and into the Americas. That is a fit to a window, not a clock on a single departure. Some species in the shared list, Ruminococcus bromii among them, are still common worldwide. The claim that matters is about the other 848. The loss could be pretty bad, a changed microbiome may have contributed to disorders common in industrial countries, from allergies to dementia. The paper does not show that. Its last sentence on the point is that many of these species are vanishing from industrialized populations and the consequences for human health remain uncertain. A long co-residence is evidence that human biology had time to come to depend on functions those microbes supply. It is not evidence that any named disease rose because they left. The functions themselves are mostly unmeasured in this study. The samples show who is missing. They do not show what the missing organisms were doing, or what replaced them.