HOW LONG have we worn shoes? The oldest surviving footwear dates back less than 9,000 years, but subtle changes in human toe bones suggest that our ancestors may have been shod at least 25,000 years ago.

Although well-preserved and well-dated prehistoric footwear is largely confined to desert North America, where woven sandals have been preserved in dry caves, there is also some circumstantial evidence suggesting its existence in Eurasia. The arrangement of beads, apparently sewn on to clothing, around the feet of an adult skeleton and two children found at Sunghir, Russia, and dated to 23,000-24,000 years ago, “imply that they were buried with foot protection”, Erik Trinkaus notes in the Journal of Archaeological Science.

Other sites of the same period have yielded woven materials, but none are certainly identifiable as shoes. However, the site of Pavlov, Ukraine, yielded clay models of what seem to be boots. Most of the footprints found in Palaeolithic cave sites, where people went to create painted and engraved images on walls from 30,000 years ago onwards, were made by bare feet. But Professor Trinklaus says that within a few millennia “archaeological data suggest that foot protection and insulation were readily available by the second half of the Upper Palaeolithic”.

A different avenue of approach is to look at the changes in bones occasioned by wearing supportive footwear, he suggests: in bare feet, peak forces go through the heel on landing and through the front of the foot and toes on taking a stride. Forces exerted on the lateral toes are about a third to half of those on the big toe (hallux) and heel. A significant increase in the use of footwear would result in modest reduction in the robustness of the big toe, and a greater decrease in the other toes: a relative shift between the first and other four toes “should indicate an increase in the use of protective footwear”, he suggests.

This thesis was tested by examining toe bones from European Palaeolithic burials from Neanderthals and early anatomically modern humans, and comparing them with prehistoric and protohistoric Native American bones, including those from warm areas, who went barefoot, and Arctic peoples, who would have worn thermally effective footwear for most of the year. Ethnographic examples show that the soles of these were usually of stiff sealskin, with the uppers of softer leather. A third comparative sample came from recent Euroamericans, who habitually wore industrially made, rigid-soled shoes.

Body masses and robustness of the toe bones were calculated, and a consistent reduction in apparent strength from Neanderthals on was observed, with middle Upper Palaeolithic man, from 20,000 years ago, proving very close to the recent human average. Recent human samples followed the predicted pattern, with the barefoot Native American having the most robust toes, followed by the Inuit with his seal-soled boots, and then Euroamericans in their manufactured shoes.

The contrast with Neanderthals, the similarity of the middle Upper Palaeolithic sample to the feet of recent humans, and the reduction in load on the lateral toes relative to the hallux, together with the lack of changes elsewhere in the leg seem to be “due to contrasts in the extent to which they were shod”, Professor Trinkaus says. “It is hard to explain these differences other than through the increased use of a device that reduced the role of the lesser toes in locomotion and thereby decreased habitual loads on them.”

Journal of Archaeological Science Vol 32: 1515-1526.