Smelting: first metal

Bog iron — rust-orange to dark brown nodules and crusts in wet ground, slow streams and peat — is the easiest iron on earth to find and needs no mining. Look for an oily, rainbow-sheened film on still water nearby that breaks into plates when poked — that is iron bacteria, not oil. Copper often shows as green (malachite) or blue (azurite) staining on rock.

Prove it: heavy for its size; on the unglazed back of a tile, bog iron leaves a yellow-brown streak, haematite a red-brown one and magnetite a black one. A roasted sample that is pulled by a magnet almost certainly contains iron. See De Re Metallica.

Roast the ore in an open fire, to a dull red heat, about 600–800 °C, for an hour or more, turning it. This drives off water and sulphur, cracks the lumps, and converts bog iron to a more reducible oxide. Roasted ore is darker and crumbles easily.

Crush it to pea size — about 5–10 mm — and sieve out the dust, which blows out of the furnace or clogs it. Fine, uniform ore reacts; lumps do not. Weigh the roasted, crushed ore — you will need the number to judge the yield.

A clay shaft about 1 m tall with an internal diameter of 25–30 cm, walls 10 cm or more of clay mixed with sand and chopped straw or dung. The tuyère — the air pipe — enters about 20 cm above the floor, angled downward at around 20–25°. Leave an arch at the base that can be opened to tap slag and remove the bloom. See the bloomery.

Dry it slowly for days, then warm it with a small wood fire for several hours before the smelt. Water in the walls turns to steam and cracks the furnace open during the burn.

Fill the shaft with charcoal and blow for about an hour to bring the whole column to heat before any ore goes in.

Then charge in alternate layers: a measure of charcoal, a measure of ore, roughly equal weights per charge, adding a new charge every 10–15 minutes as the column sinks. Keep the shaft full. A typical run for a furnace this size is 4–8 hours, and records of each charge — time and weight — are what let you improve the next one.

Insufficient or intermittent air is the single commonest cause of failure: the furnace never reaches temperature and you get a lump of roasted rock. Air comes from bellows worked in shifts, without stopping, for the whole run. Plan a rota of at least four people for a single furnace.

Signs it is right: a bright yellow-white glow in the tuyère, a steady purring roar, and slag dripping and forming at the base. If the tuyère glow goes orange or red, the air is short — increase it before anything else. Tap slag through the arch when it builds up, so it does not drown the bloom.

A bloomery does not melt iron. It reduces it in the solid state, leaving a spongy mass — the bloom — of iron and slag at the base. Expecting liquid iron and discarding the bloom as a failure is the commonest mistake of all.

At the end of the run, break open the arch, pull the bloom out with tongs, and immediately start hammering it on a stump or stone — lightly at first, because it is fragile — to squeeze out the slag and weld the iron together. Reheat and hammer repeatedly over hours. Most of the mass is lost; what remains is wrought iron. Expect a bloom of roughly 10–25% of the ore weight.

Copper melts at about 1,085 °C and can be poured, so every step gives visible feedback. Smelt copper ore in a smaller furnace or a crucible with charcoal and forced air, then cast it in a mould of stone, clay or sand.

Bronze — copper with roughly a tenth tin — melts lower, casts beautifully and is harder than either metal alone. If you have tin, bronze is a faster route to good tools than iron. Always dry moulds completely and preheat them: molten metal on a damp mould throws out steam and metal together.

Add crushed limestone or shell — see lime — with the ore when the slag is thick and sticky; it makes the slag run freely and leave the iron. Start small, a handful per charge, and increase only if the slag stays viscous.

Record every smelt: ore source and weight, charcoal weight, charge interval, air rota, duration, slag behaviour, and bloom weight. This is how the second smelt beats the first. And have three times the charcoal you think you need — running out halfway through a smelt wastes all of it.