The Underground Network

On the eastern flank of the tiny constitutional monarchy of Lesotho, about 225 kilometers from Durban, South Africa, sits the village of Ha Mokoto. Its residents are eking out a living in a manner not dissimilar to how their ancestors eked out a living 200 years ago, when Basotho people, led by King Moshoeshoe I, fled to the mountains to escape colonial strife in what is today South Africa. They tend to livestock and grow corn and sorghum, a starchy grain similar to quinoa or bulgur that’s cooked in large pots and stirred with a long stick into a thick paste. Circular homes called rondavels are fashioned from a base of stones and mortar made out of dung and fine dirt harvested from termite mounds and topped with a conical roof of grass supported by beams from poplar trees. The only electricity in Ha Mokoto, which means “place of wild dogs,” comes from a few solar panels that an entrepreneur brought to the village so people could charge phones. The unpaved road into town has been described as “tormenting.”

Nevertheless, the tentacles of science are long and tenacious, and in late 2022 the evolutionary biologist Toby Kiers, a professor at Vrije Universiteit in Amsterdam, followed the direction of a machine learning algorithm developed by colleagues at ETH Zürich and GlobalFungi, an open-source database, that drew on a geo-tagged database and hundreds of data layers of ecological variables to predict areas with high levels of biodiversity of mycorrhizal fungi, her area of study, and traveled to Ha Mokoto. Farming in Lesotho occurs almost entirely without the chemical fertilizers that render these fungi—which have been transferring nutrients and sometimes water to plants for 450,000,000 years—superfluous. So the region may have evolved unknown fungal species with special properties.