Virtually all the ore used by the ancients was handpicked with only the most worthwhile material being taken. Originally it was probably smelted by the Sumerians in shallow pits using charcoal as the fuel. How they first derived the necessary draught to raise the temperature sufficiently to melt the ore is still a matter of speculation, but it may have been done by banking over the furnace with clay and leaving an opening directed towards the prevailing wind. Bellows were certainly known by about 2500 B.C. and some form of bellows must have been employed still earlier in order to account for the more ancient bronzes. Not until 2000 B.C. or later did these improvements reach Egypt, where hieroglyphs of that period show air being blown into the furnace through a straight tube. The bellows type reached Egypt a little later.
Long afterwards smelting furnaces acquired the shape that they essentially maintained right down to Victorian Swansea days, i.e. a small stone or brick chamber with some means of pouring or ladling out the molten metal, a hearth below, and a brick chimney, usually low and stumpy, which provided an updraft and allowed the waste gases to escape. The state of the atmosphere around such furnaces can be imagined, but both life and labor were cheap. The Romans improved on the early primitive methods, but without altering the fundamental principle; and, as already mentioned, they successfully smelted sulfide ores, which would have defied the more ancient metallurgists.
In Britain there has been found in Anglesey a number of circular cakes of copper which were cast in Roman times. These cakes were 11 to 13 in. in diameter, 2 to 2 1/2 in. deep, and weighed from 30 to 50 lb. apiece. They were formed 'by pouring the metal into a shallow tray. A comparatively smooth band, which is always found on the upper surface immediately within the perimeter, is due to the more rapid cooling of the metal where it touched the cold tray. The central part shows a marked rising, due to the evolution of sulfurous acid, and proves that at that time sulfide ores were being smelted. The metal must have been ladled from the furnace, not tapped. One came from Amlwch; it weighs 42 lbs. and is stamped "IVLS." Another is stamped "SOCIO ROMAE." (2)