Evidence for ancient metalworking is sparse, and now historians who recreated Bronze Age smelting techniques know why — the clues naturally disappear. The finding explains why, despite the discovery of 10 British mines dating from 2050 to 1500 BCE, very few remains of actual metalworking sites have been excavated around the world. For the study, students at the University of London's Birkbeck college conducted a number of smelting experiments, including the construction of crude furnaces, at Butser Ancient Farm Project in Hampshire.
     "One of the reasons for making these furnaces and for monitoring how they weather on abandonment is to show how little archaeological record they might leave," said Simon Timberlake, excavations director of the Early Mines Research Group and one of the metalworking project's leaders. "Such ephemeral, once-only furnaces, particularly if they were dug out after use, might be difficult to detect archaeologically, and even harder to interpret as such if found."
     Timberlake's team constructed three types of possible early furnaces. The first was a clay-lined hearth with air intake provided by homemade leather bellows and pipes, called tuyeres, made out of clay, straw, crushed flint grit, and sheep dung collected from a farm. The second unit was similar, only it was not lined with clay, and was dug into the side of an enclosure bank. The third and final furnace was the crudest of all. Researchers simply put a rough clay dish at the bottom of a narrow hole with a tuyere inserted into one side. As with the other furnaces, air intake was controlled by a homemade bellows, which this time was stuck into a tuyere.
     "We smelted both high-grade malachite (hydrous copper carbonate) and tin ore (ground cassiterite-tin oxide) successfully in this furnace, preheating the hole with charcoal to dry it out and warm it, then adding the crushed cassiterite or malachite onto the top, then piling fresh charcoal over the top," Timberlake told Discovery News. The speed of combustion and the sealing of furnace gases were expedited by the addition of a "black hat" pile of unburnt charcoal or a piece of cut turf.
     Timberlake and his colleagues found that malachite smelted at 800° Celsius, but copper and tin required temperatures higher than 1,000° C for the metal to pool. No slag was produced at those low temperatures. The clay and bellows would have eroded over time. Any tiny bits of leftover metal would have mixed into the soil. In short, the metalworkers' evidence naturally disappears.
     John Prag, professor of archaeological studies at the University of Manchester and keeper of archaeology at the Manchester Museum, agrees with the team's findings. "Little, if any, of these processes were described by ancient authors, and therefore experimental archaeology of this kind when carried out in collaboration with research scientists is really our only way of discovering how early smelting was carried out," he said.

Source: Discovery Channel News (26 April 2005)