Agricola, De re metallica (1556): copper smelting, roasting, cupellation, assay and liquation
Covers one book, read as archive.org OCR text:
- G. Agricola, De re metallica (Basel, 1556), translated by H. C. and L. H. Hoover (London, 1912), archive.org
georgiusagricola00agririch. Page numbers are Hoover's printed folios, read from the OCR running heads; OCR is rough and the figure legends are the only place some letters survive. Book VII pp. 223-250 (assay), Book VIII pp. 270-279 and 348-351 (stamps, burning, roasting), Book IX pp. 353-427 (smelting), Book X pp. 464-487 (cupellation, refining), Book XI p. 491 (liquation, Hoover's outline only), Appendix C pp. 616-618 (measures). - Hoover's footnotes and "historical notes" are quoted as HOOVER; they are 1912, not 1556.
- Not read: the rest of Book XI (the liquation, drying and copper-refining furnaces), Books I-VI, XII.
Facts are the book's; INFERRED marks the note's own reading.
0. Units and what the book does not give
- Agricola's lengths, Hoover's table (Appendix C, p. 617): digit .726 in, palm 2.90 in, pes ("foot") 11.61 in, cubit = 1.5 foot. Hoover keeps the Latin and says (p. 616) Agricola means the German foot, "fairly closely approximate to the English". INFERRED metric: digit 1.84 cm, palm 7.4 cm, foot 29.5 cm, cubit 44 cm. HOOVER GIVES NO METRIC VALUES; every cm figure below is mine.
- 4 digits = 1 palm; 4 palms = 1 foot. So "four feet and three palms" is 4.75 ft.
- Weights: libra 4946 troy grains (320 g); centumpondium = 100 librae = 494,640 grains (Appendix C p. 617; "about 70.6 lbs avoirdupois", fn 31 p. 472), 32 kg. The old German centner is heavier (about 114 lb avoirdupois, 52 kg). Hoover says (p. 616) the text sometimes means the German scale. Which is meant is not settled; Roman is assumed below.
- Not in the book: temperatures (no degrees, only colour and "fierce/mild fire"), fuel quantities per smelt except the basket counts below, tuyere air volumes, slag chemistry, anything on foot-bellows used on a smelting furnace, and any furnace built only for copper. Agricola describes one general blast furnace and four ways of running it (section 2).
- There is no furnace of one low shaft built for copper alone: the nearest are the first/second/third furnace (section 2), the tin furnace (section 2.8) and the Carni lead hearth (section 2.7).
1. Roasting: heaps, enclosures and furnaces (Book VIII, pp. 273-279, 348-351)
1.1 Why burn or roast (p. 273)
- "Ore is burned for two reasons ; either that from being hard, it may become soft and more easily broken ... or that the fatty things, that is to say, sulphur, bitumen, orpiment, or realgar may be consumed." Sulphur "is most harmful of all to iron, and less to tin than to bismuth, lead, silver, or copper"; "in a very vigorous furnace fire, sulphur resolves metal into ashes and makes slag of it."
- Hot ore is wetted to soften it: "after fire has dried up the moisture in the ore, it breaks up more easily while it is still hot, of which fact burnt limestone affords the best example" (p. 275).
1.2 The common heap (p. 273-274)
- A quadrangular area is dug "of fair size, open at the front". Firewood is laid close, other wood across it, "this pile of wood a crate"; repeated to a height of "one or two cubits" (44-88 cm, INFERRED). Ore broken small with a hammer goes on top: "first the largest of these pieces, next those of medium size, and lastly the smallest", built into "a gently sloping cone".
- To hold the cone: fine sand of the same ore soaked and smeared on and beaten with shovels, or charcoal dust "just as do charcoal-burners". At Goslar the cone is smeared with amentum sutorium rubrum (leached roasted pyrites; Hoover cannot identify it, fn 5 p. 274).
- Roast count: "In some districts the ore is roasted once, in others twice, in others three times, as its hardness may require" (p. 274). No duration is given for a heap.
- The third roast of Goslar pyrites exudes "a certain greenish, dry, rough, thin substance ... no more easily burned by the fire than is asbestos" (p. 274; Hoover fn 6 reads it as alum efflorescence).
1.3 The walled stall (p. 275)
- This is the three-walled bay. "By digging out the earth they make the areas much larger, and square ; walls should be built along the sides and back to hold the heat of the fire more effectively, and the front should be left open." Agricola gives it for tin ore.
- Charge order: wood "about twelve feet long ... in the area in four layers, alternately straight and transverse"; larger ore pieces on the wood, smaller on top "and also placed around the sides"; fine ore sand spread on the pile and pounded with shovels "to prevent the pile from falling before it has been roasted"; then fired.
- Lead ore in a like area but sloping, wood over it, "a tree trunk ... laid right across the front of the ore to prevent it from falling out. The ore, being roasted in this way, becomes partly melted and resembles slag" (p. 275). This is the failure mode of too hot a roast.
- No stall inside dimensions are given beyond "twelve feet" wood. INFERRED: a bay about 12 ft (3.5 m) deep.
- Hoover's footnote marks the tin furnace "similar to an oven" with a front mouth, poker and two-tooth rake (p. 348, fig.); the burned tin-stone "frequently ... runs together" and is re-washed (p. 349).
1.4 Heap of copper schist, faggots (p. 279)
- At Eisleben the bituminous copper schist is not piled on logs but on "bundles of faggots". Method now: stones are left to lie "for some time in such a way as to allow the air and rain to soften them"; a faggot bed is made beside the heap, stones carried onto it, faggots laid in the emptied place, and so on "until all the stones have been piled mound-shape on a bed of faggots." Fired on the lee side "lest the fire blown up by the force of the wind should consume the faggots before the stones are roasted and made soft"; the fire passes stone to stone and "the heap very often burns continuously for thirty days or more."
- The schist rich in copper exudes the same asbestos-like stuff (p. 279). HOOVER fn on cadmia bituminosa (p. 273) quotes Agricola's own account: rain falling on the burning heap makes it "burn more brightly and soften more quickly".
1.5 Roasting that saves the sulphur (pp. 276-278)
- Sulphurous or bituminous ore on "an iron plate full of holes, and roasted with charcoal placed on top ; three walls support this plate"; pots of water below catch the vapour (p. 276). A vaulted two-chamber furnace with bottomless pots of 100 pounds of ore each is also given (p. 278). Not needed for copper oxide; noted because it is the only closed roast.
1.6 Roasting the cakes (matte) (pp. 349-351)
- "Cakes from pyrites, or cadmia, or cupriferous stones, are roasted in quadrangular pits, of which the front and top are open, and these pits are generally twelve feet long, eight feet wide, and three feet deep" (p. 349; 3.5 x 2.4 x 0.9 m, INFERRED). Cakes of melted pyrites "usually roasted twice over, and those of cadmia once". Cadmia cakes are first rolled in mud moistened with vinegar so the fire does not consume the copper with the sulphur.
- "The cakes of pyrites are first roasted in a slow fire and afterward in a fierce one, and in both cases, during the whole following night, water is let in" to wash out alum, vitriol or saltpetre (pp. 349-350). Cakes stand "on wood piled up in the form of a crate, and this pile is fired."
- Cakes of copper smelted from schist: broken, set in a furnace on faggot bundles, "generally roasted seven times and occasionally nine times"; "if they are bituminous, then the bitumen burns and can be smelted" (pp. 350-351).
- That furnace: "a structure like the structure of the furnaces in which ore is smelted, except that they are open in front ; they are six feet high and four feet wide." "Three of them are required for one of those in which the cakes are melted." Cakes pass first, second, third furnace "and afterward back to the first ... until they have been roasted seven or nine times" (p. 351).
- Wood is the roasting fuel everywhere in Book VIII; charcoal is named only for the plate roast. For smelting HOOVER p. 355: "Charcoal was the universal fuel for smelting down to the 18th century."
1.7 Stamps, sieves and washing as they bear on the charge (pp. 270-295)
- Hard ore is burned, then crushed on a paved area with iron "threshing sledge" hammers or under iron-shod stamps; wet ore is washed in a strake (pp. 272-273).
- The dry-stamp box has a wire-net floor "through which broken rock of the size of a hazel nut cannot pass" (p. 287). Iron ore for the iron furnace is "fragments of ore not larger than nuts" (p. 423).
- Washed concentrates are fed to the furnace mixed with litharge, hearth-lead, slag or cakes so they do not "fall back ... or fly out" (p. 401); a furnace is dampened with water for the same reason (p. 379).
- Ore mills may be turned by men treading, horses, asses or goats (p. 295). That is ore grinding only; no beast turns a bellows in the book.
2. The smelting furnace and forehearth (Book IX, pp. 353-412)
2.1 The works and the furnace body (pp. 355-362)
- Usually six furnaces, "rarely more, and often less", built against a thick "second wall" (2 ft 2 palms thick, 15 ft high) of brick or stone, 6 ft apart "in order that the smelters and their assistants may more easily sustain the fierceness of the heat" (p. 356).
- Three furnace walls (back and two sides) are natural stone, soft and crack-free: "bricks soon become defective and crumble away" when the accretions are chipped off; hard cracked stone "is burst asunder by the fire" (p. 355). The front wall is brick.
- Wall heights: front not over 5 ft "so that the ore may be conveniently put into the furnace"; sides 6 ft; back 7 ft; all 3 palms thick (p. 356).
- INSIDE SIZE: "five palms wide, six palms and a digit long ... the upper part of the furnace widens out somewhat" (p. 356). INFERRED: 37 x 46 cm plan, 1.8 m side height (a low, narrow, open-fronted box, not a tall stack).
- Mouth in the front wall: "three palms wide and one and a half feet high, when the hearth is completed" (p. 355; 22 x 44 cm).
- Tuyere: a hole "slanting upward, three palms long ... through the back furnace wall, at the height of a cubit, before the hearth has been prepared"; through it and a foot-long hole in the second wall (arched behind) goes a pipe "of iron or bronze" holding the nozzles (p. 355). The pipe's mouth must not protrude from the lute "lest sows form round about it" (p. 376). The pipe: rolled copper or iron plate, "a foot and two palms and the same number of digits" long, bore "three digits wide, and two and a half digits high in front", with a half-digit crack in the front widening behind; nozzles sit 5 digits from its front end (p. 369). INFERRED: front bore about 5.5 x 4.6 cm, one pipe, two bellows.
- Blast strength is tuned by the pipe slope: low hearth and gentle slope for easily fused ore; high hearth and steep slope for ore that fuses slowly (pp. 379-380).
- Beneath hearth and forehearth, about a cubit down, a stone- or brick-lined water tank "three feet long, three palms wide and a cubit deep" with a stone cover and a walled vent 6 digits wide sloping to a pipe through the back wall; its job is to stop the ground drawing moisture into the hearth, "the moisture would vitiate the blast, and part of the metal would be absorbed and part would be mixed with the slags" (pp. 356-358). A tank without a vent "is indeed to be condemned".
- Roof, hood and dust chamber: a hooded front area, ore and charcoal store beside it (p. 361). Vaulted dust chambers over the furnaces (twice-yearly sweeping of soot and "pompholyx") catch metal in the fume; "profitable for all metalliferous ores" (pp. 394-396).
2.2 Hearth and forehearth materials and making (pp. 374-379, 386)
- Hearth powder: charcoal crushed under wooden stamps and sieved, mixed with dried, sieved earth: "two parts of pulverised charcoal and one part of powdered earth" (p. 376), moistened until it balls "like snowballs". For furnaces with an always-open tap-hole: "consists mostly of equal proportions of charcoal dust and of earth, or of equal parts of the same and of ashes" (p. 386). Stored in a covered pit.
- New furnace: inside lined with lute to fill cracks and "preserve the rock from injury by fire". Old furnace: when the rocks have cooled the assistant chips off the accretions with a spatula, breaks them with a hoe or five-tooth rake, fills cracks with rock or brick fragments, re-lutes with a long wooden spatula from a ladder (p. 376).
- Hearth: powder thrown in and rammed with a truncated-cone rammer (3 palms long, 5 digits across the bottom, 5 ft handle) so "the hearth is made solid"; the forehearth pit outside is rammed the same way, then "from a point a digit under the mouth of the copper pipe the hearth slopes down into the crucible of the forehearth" (pp. 376-377).
- Forehearth crucible: round, "a foot in diameter and two palms deep if it has to contain a centumpondium of lead, or if only seventy librae, then three palms in diameter and two palms deep" (p. 378; 30 x 15 cm, 22 kg-class, INFERRED), hollowed with a curved blade and pounded with a bronze rammer. The forehearth sits "very near the furnace in order that it may keep very hot and the alloy thus be made purer" (p. 386).
- Tap-hole: a rod laid in the mouth to the forehearth, "higher up if the ore to be smelted readily fuses, and lower down if it fuses with difficulty"; the lower part of the mouth is not rammed hard "for it could not breathe" (p. 386).
- Method with tap-hole closed (first method, for rich silver/gold ore): mouth closed with lute dipped in charcoal dust, a charcoal block and a rod pulled out later; the master holds the charcoal block in with his hand (p. 378).
2.3 Lighting and warming (pp. 378-379)
- Evening before: the assistant fills the furnace "four or five larger baskets full of charcoal ... right up", lays charcoal and glowing coals in the forehearth, and shuts the tap-hole with lute or potsherds so flame does not enter. Some warm the forehearth only in the morning with firebrands "prepared by the watchman of the works".
- "At the fourth hour the master begins his work": a glowing coal through the nozzle pipe and the bellows "thus within the space of half an hour the forehearth, as well as the hearth, becomes warmed, and of course more quickly if on the preceding day ores have been smelted in the same furnace".
- Cold or damp hearth: "If the hearth and forehearth are not warmed before the ore to be smelted is thrown in, the furnace is injured and the metals lost ; or if the powder ... is damp in summer or frozen in winter, they will be cracked, and, giving out a sound like thunder, they will blow out the metals and other substances with great peril to the workmen" (p. 379).
2.4 Charge order, tapping, products (first method, pp. 379-382)
- Slag first, to prime the forehearth: "the master throws in slags, and these, when melted, flow out through the tap-hole into the forehearth." Tap-hole stopped with a lute-and-charcoal plug on a rammer; slags drawn with a hooked bar. For gold or silver ore a centumpondium of lead (half if poor) goes in the forehearth and is melted with firebrands.
- Then, per fill: "as many cakes melted from pyrites" as the ore needs; then "two wicker baskets full of ore with litharge and hearth-lead, and stones which fuse easily by fire of the second order, all mixed together; then one wicker basket full of charcoal, and lastly the slags." Smelted slowly. Ore is not piled against the back wall "lest sows should form around the nozzles". Water is sprinkled on the charcoal so fines stick to it rather than being blown out (p. 379).
- "After a quarter of an hour, when the lead ... is melted, the master opens the tap-hole with a tapping-bar" (iron, 3.5 ft, hollow back on a 3 ft handle). Slag runs first (stones, earth, "solidified juices"), then the pyrite cakes, then the metal settles; in the forehearth slag floats, the cakes sit in the middle, the metal is heaviest (p. 381). Slag, cakes are drawn with the hooked bar or an iron fork; the alloy is ladled into lute-lined iron or copper pans (p. 382). Slag heaps are kept by layer: top little metal, middle more, bottom much; re-smelted with the right flux, "if it emits much odour, there is some metal in it" (p. 382).
- Repeat until the day's ore is done: "if the ore was rich the work is finished in eight hours ; if poor, it takes a longer time. But if the ore was so rich as to be smelted in less than eight hours, another operation is in the meanwhile combined ... and both are performed in the space of ten hours" (p. 382).
- Close of campaign: a basket of litharge or hearth-lead goes in to wash metal from the accretions; then the furnace mouth is broken out "with a crowbar", rabble and five-tooth rake pull the accretions and charcoal, and once cool the walls are chipped with the rectangular spatula (p. 382). The furnace is therefore emptied and re-made after each day's smelt in this first method (INFERRED from the sequence; the text does not say "daily" outright).
- If the ore melts too slowly or does not unite: add fluxes; too readily: reduce them (p. 384). If tapping shows scattered flaming, "the metal is consumed by the violence of the fire" (p. 384).
2.5 The always-open furnaces: first, second, third (pp. 385-390)
- Three methods with the tap-hole always open. "The tap-hole of the first of this kind is deeper in the furnace and narrower than that of the third, and besides it is invisible and concealed." Forehearth "one and a half feet higher than the floor of the building, in order that below it to the left a dipping-pot can be made"; slag is skimmed off the forehearth, then the alloy and the matte run into the dipping-pot, where the matte is made into cakes; cakes are broken and thrown back (p. 386). Ore that melts with difficulty goes to the back at the hottest part; easily fused ore, one or two baskets at the front "so that the fire, being driven back by it, may also smelt the ore and the sows" (p. 386-388). Called "very ancient among the Tyrolese" (p. 388).
- Second method (Styria): two crucibles, one half inside the furnace and half outside holding lead that absorbs the metal; a lower crucible takes alloy and matte; the upper is tapped "once or twice" (p. 388).
- Third method (copper, lead and the poorest silver ore): "the furnace is higher and wider than the others, and its bellows are larger ... a larger charge of the ore can be thrown into it." "They smelt in the same furnace continuously for three days and three nights, providing there be no defect either in the hearth or in the forehearth. In this kind of a furnace almost every kind of accretion will be found." A second forehearth crucible takes the run when the first is full; "When a smelter has spent twelve hours' labour on this work, another always takes his place" (p. 388). No inside dimension is given for the third furnace.
- Fluxes named for the third furnace (p. 390): cakes of melted pyrites ("cupriferous pyrites which easily melt in the fire"), lead ore, "stones easily fused by fire of the second order and sand made from them, limestone, tophus, white schist, and iron stone". Hoover fn 15 (p. 380) reads the first two orders as partly fluorspar and the third as quartz and silica. Iron-rich ores (fixing lead): hammer-scales in the hearth powder and iron slag as the main flux (p. 408).
- A frozen tap-hole is broken with a third kind of hooked bar (fig. legend, p. 389).
- Charcoal is carried to the furnaces in wicker baskets from a heap, measured in a wicker box on a barrow (p. 389, fig. legend).
2.6 Copper from ore, matte and black copper (pp. 401-408)
- Oxide ores: "Pure copper ore, whether it is its own colour or is tinged with chrysocolla or azure, and copper glance, or grey or black rudis copper, is smelted in a furnace of which the tap-hole is closed for a very short time, or else is always open." Hoover fn 41 (p. 401): its own colour, probably cuprite; chrysocolla and azure, partly malachite and azurite; "lead coloured", chalcocite. If silver is large, run into the forehearth over lead; "If there is a small amount of silver in the ore, no lead is put into the forehearth ... if there be no silver, copper is made direct" (pp. 401-403).
- Ore carrying hard minerals (pyrites, "cadmia metallica", iron-stone): add crude pyrites "which easily fuse" and slag, make cakes, roast as needed, re-smelt: "the copper is made" (pp. 403-404).
- Inferior rudis ore, Tyrolese, first furnace, 12 h: "To as much rudis copper ore as will fill eighteen vessels, each of which holds almost as much as seven Roman moduli, the first smelter ... adds three cartloads of lead slags, one cartload of schist, one fifth of a centumpondium of stones which easily fuse in the fire, besides a small quantity of concentrates collected from copper slag and accretions" and "makes six centumpondia of primary cakes and one-half of a centumpondium of alloy" (p. 405). HOOVER fn 44 (p. 405) works the ore at about 5,400 lb (about 2.4 t). Cakes carry half a libra of silver per centumpondium. Weekly (6 days): 36 centumpondia of cakes and 3 of alloy (p. 406).
- Second smelter: 18 centumpondia of cakes + 12 of hearth-lead and litharge + 3 of lead-smelting stones + 5 of rich hard cakes + 2 of exhausted liquation cakes + slag, 12 h, gives 18 of secondary cakes and 12 of copper-lead-silver alloy poured into copper or iron moulds (p. 406); the third smelter makes quaternary "hard cakes", and "from the 'final' cakes, thrice roasted and re-smelted, is made black copper" (pp. 406-407). Hoover fn 46 (p. 407) tabulates the circuit and gives silver contents.
- Poorly silvered copper: smelt in the third furnace, make cakes, "after being seven times roasted are re-smelted" to melt out copper; the copper cakes then go to "a furnace of another kind" for a third melting (liquation, Book XI) (p. 407).
- Copper schist: roast, smelt with stones fusing at "second order"; the slags float on the cakes; cakes "usually roasted seven times and re-melted"; the copper "occupies the bottom of the crucible" (p. 408). Poor schist is burned, stamped, washed, sieved, and the concentrates melted into cakes (p. 408).
- The first-melt lead pyrites at Goslar throw a white liquor that "consumes" silver and is drawn off before the lead; the walls exude "conterfei" (Hoover fn 48, p. 408: metallic zinc) (p. 408).
- Smelt time, matte yield and loss per charge are only given in the examples above. No copper yield percentage is stated.
- HOOVER historical note on copper (p. 402): Bronze Age smelting "a simple shallow pit in the ground into which the ore was placed, underlaid with charcoal"; "Rude round copper cakes eight to ten inches in diameter"; Egyptian furnaces with forced draught by blow-pipe, bellows about 1500 B.C.; Theophilus (12th century) smelts "incessantly forged by day and night" in layers of coal and stone (p. 405). Pliny: "all copper fuses best when the weather is intensely cold" (p. 403); pine is best fuel for copper and iron (p. 403).
2.7 Lead smelting hearths (pp. 390-394) - closest to a low hearth
- Carni: "Between the two low walls of a hearth, which is inside a furnace made of ... rock that resists injury by the fire", green wood with a dry layer on top, burned ore on that; "when the wood is kindled the lead drips down" onto a sloping hearth of charcoal and earth into a half-inside crucible; a rectangular hole at the back lets in draught "and ... the smelter can crawl through it into the furnace" (p. 391). No bellows.
- Saxons (Gittelde): baking-oven shape, wood put in at the back, lead trickles into a forehearth, then is tapped with a bar (p. 391). Westphalian heaps: 10 wagon-loads of charcoal on a hillside, ore on top, wind-fanned; Poles: a brick hearth 4 ft high (pp. 391-394). Hoover fn 23 (pp. 390-392) says the Roman lead furnaces in Cumberland drew draught through a tapering tunnel driven into a hill.
2.8 The tin furnace (small low shaft) (p. 411)
- "narrower than the other furnaces, that there may be only the small fire which is necessary for this ore ... higher, that the height may compensate for the narrowness". Hearth: sandstones "two and three-quarters feet long, the same number of feet wide, and two feet thick", on a slight slope. Furnace around them "eight or nine feet high" of sandstone, lute-lined, interior at the top "two feet long and one foot wide, but below it is not so long and wide". Front closed at top, steps on the open back side. Hood walls lead the fume to a dust chamber (p. 411-412). Charcoal ladder to feed. The only furnace in the book with a stated sub-two-foot interior other than the main one in 2.1.
- Lusitanian small tin furnaces with a round hand-drawn bellows (leather folds on iron hoops, iron disc at the nozzle end, wooden disc behind with a blow-hole and a haft): "the smelter is not able during a whole day to smelt much more than half a centumpondium of tin" (pp. 419-420). This is the only hand bellows on a smelting furnace in the book.
2.9 Bellows, drive and crew for the smelting furnace (pp. 362-373)
- Two bellows per furnace ("twelve bellows, if there are to be six furnaces", p. 366), both nozzles in one copper pipe (p. 368).
- Body: upper board a palm thick, "five feet and three palms long", 2.5 ft wide at the back and a cubit at the head; ox hide (horse inferior) tacked with T-nails; glued pine with linden edge strips; head a cubit long; nozzle of thin iron plate, 3 digits across in front, a palm by two palms behind, "three feet long" (pp. 362-365). Top-board aperture with a sliding lid "if he fears the hides might be burst when the bellows are too vigorously and quickly inflated"; lower board has a goat-skin flap valve in a one-foot air-hole (pp. 362-364).
- Drive: a water-wheel on an axle outside the works; inside, a rundle drum (two double hubs, 30 rundles) meshes with a toothed wheel (60 teeth) on a cam axle with 24 cams (two per lever); the cams depress the levers (8 ft 3 palms long) which compress the bellows; a counterweight box of earth and stones on an upper lever (8 ft) raises the board "and the air is drawn in" (pp. 369-373). "this is only one of many such axles having cams and a water-wheel" (p. 374). INFERRED: one wheel drives all twelve bellows of a six-furnace row, staggered by the cam holes.
- Wheel type is not stated (the figure legend says "water-wheel" only). Horses or men for a smelting bellows: NOT IN THE BOOK. The only man-worked bellows are the assay double bellows on a hand lever (p. 227) and the Lusitanian hand bellows above.
- HOOVER fn 6 (p. 362): bellows are Egyptian, a foot-bellows figured in a tomb of Thothmes III (about 1500 B.C.); Ephorus ascribes bellows to Anacharsis (about 600 B.C.). Foot-bellows in the Egyptian drawing: not described by Agricola.
- Crew per furnace: the master smelter and an assistant (deputy); a watchman of the works lights; mine captains (two) weigh the cakes; in 3-day smelts, 12-hour shifts of smelters (pp. 378, 383, 388). Copper operation (Tyrol): three smelters in series (pp. 405-407).
2.10 What goes wrong (smelting)
- Sows (lumps) form round the tuyere if ore is piled at the back, if the pipe mouth protrudes, or if the nozzle is too narrow for the bellows, "the melted material would be chilled" (pp. 376-380); cleared by dividing the charge with a hooked bar and pushing it down the pipe. Hoover fn 10 (p. 376): "like most modern metallurgists, Agricola offers no method for treating them."
- Hearth rammed too hard or the mouth rammed hard: it "could not breathe" (p. 386). Hearth too sloping with slow-melting ore: the fluxes run out first and leave metal in the accretions (p. 386).
- Fines choke the tap-hole (p. 380); frozen tap-hole (p. 389); damp hearth blows out (p. 379); damp ground under the hearth (p. 356); hides burst (p. 363).
- Fumes: lead and cadmia fumes are caught in dust chambers; Strabo's high stacks to carry off the "dense and pestilent" vapour (HOOVER fn 23, p. 392).
- Fire risk: roof boards covered with lute; leather buckets and a brass squirt pump stand ready should the furnace roof catch fire (fig. legend, p. 377).
3. The iron furnace and bloom (Book IX, pp. 420-427)
- Two furnaces. (a) The hearth bloomery: "Very good iron ore is smelted in a furnace almost like the cupellation furnace. The hearth is three and a half feet high, and five feet long and wide; in the centre of it is a crucible a foot deep and one and a half feet wide, but it may be deeper or shallower, wider or narrower" (p. 420; 1.0 m high, 1.5 m square, 44 cm crucible, INFERRED).
- Charge: "throws charcoal into the crucible, and sprinkles over it an iron shovel-ful of crushed iron ore mixed with unslaked lime. Then he repeatedly throws on charcoal and sprinkles it with ore ... until he has slowly built up a heap" (p. 420). It melts "when the charcoal has been kindled and the fire violently stimulated by the blast of the bellows, which are skilfully fixed in a pipe." No flux other than lime is named. No quantity of ore per charge is given beyond the bloom weight.
- Crew: one master (wearing a face cap with eye and breath holes) and an assistant. The master works a bar at the hearth side "which he raises as often as is necessary, when the bellows blow too violent a blast, or when he adds more ore and charcoal"; the same bar draws slag and opens or closes the sluice gates to the water-wheel (p. 421). Bellows are water-driven, as in 2.9; a bar is the only control.
- Time: "He is able to complete this work sometimes in eight hours, sometimes in ten, and again sometimes in twelve" (p. 421).
- Product: "a mass weighing two or three centumpondia may be made, providing the iron ore was rich" (p. 421; 140-210 lb avoirdupois by Hoover's 70.6 lb each, 64-96 kg). Drawn: "the master opens the slag-vent with the tapping-bar, and when all has run out he allows the iron mass to cool", stirs it with the bar, then it is taken down to the floor and "beat ... with large wooden mallets having slender handles five feet long" to chip off slag and flatten it (p. 421). Then on the anvil under a cam-raised water-driven iron hammer, cut into 4, 5 or 6 pieces, re-heated in the smith's forge and drawn to square bars, ploughshares or tyres; "Four, six, or eight of these bars weigh one-fifth of a centumpondium" (p. 423); a youth ladles water on the glowing iron (p. 423).
- Failure/other product: "The masses, if they remain and settle in the crucible ... become hard iron which can only be hammered with difficulty, and from these they make the iron-shod heads for the stamps" (p. 423). HOOVER fn 55 (p. 420) reads this as the first hint of cast iron.
- (b) Cupriferous or hard-to-melt iron ore needs a fiercer fire: dry-stamp, roast, wash; smelted "in a furnace similar to the blast furnace, but much wider and higher"; stairs on the side; filled "partly with fragments of ore not larger than nuts, and partly with charcoal"; "once or twice smelted" to give iron for the smith (p. 423). No dimensions are given. Hoover fn 55 calls it the German Stuckofen, "a high bloomery furnace".
- Steel by cementation is described in the smith's crucible (about 1.5 ft wide, 1 ft deep, best charcoal; four iron masses of 30 lb heated 5 to 6 hours with stirring, then quenched; pp. 425-426). Not detailed here.
- Not in the book for iron: bellows size, tuyere pipe size, wall material, hearth lining, yield percentage, fuel quantity. Hoover's iron history (pp. 421-423) is 1912.
4. The cupellation furnace (Treibherd) (Book X, pp. 464-487)
Agricola names it "the second furnace"; Hoover renders "cupellation furnace" and gives the German as Treibeherd (fn 26, p. 464).
4.1 Building and place (pp. 464-467)
- Housed in a building near the smelting works: middle wall 21 ft x 15 ft high, rear wall 19 ft behind, 36 ft long x 14 ft high; the furnace sits between the second and third transverse walls beside the stamps and the bellows with their wheel (pp. 464-467). Chimney: a vertical wall on the middle wall and a sloping chimney wall on beams, 12 ft from ground to roof-beams (p. 466).
- Bellows: the same wheel-driven cam bellows as the smelting works; a drum and cams turn the other way for the stamps (p. 467).
4.2 The furnace body (pp. 467-468)
- "Lead is separated from gold or silver in a cupellation furnace, of which the structure consists of rectangular stones, of two interior walls of which the one intersects the other transversely, of a round sole, and of a dome. Its crucible is made from powder of earth and ash" (p. 467).
- Circular wall: "four feet and three palms high, and one foot thick" (1.4 m high by 30 cm thick, INFERRED); "usually as many as fourteen stones", each "a foot and a palm wide" outside, narrower inside, sunk "a foot and a palm" in the earth, joined at the top by iron staples set in molten lead. A ledge cut from 2 ft 3 palms up carries the stone sole. INFERRED from 14 stones of 1.25 ft: outside circumference 17.5 ft, outside diameter about 5.6 ft (1.7 m). The book states no diameter.
- Six air holes "at a height of a foot above the ground", each 2 palms high by a palm and 3 digits wide, one by the litharge channel; they let ground moisture out: "if it were not for these openings the crucible would absorb the moisture and be damaged ... a lump would be raised, like that which a mole throws up from the earth, and the ash would float on the top, and the crucible would absorb the silver-lead alloy" (p. 467).
- Inside: two crossed brick walls (one brick thick, with four air holes) forming four compartments, each filled with a wheelbarrow of slag and a big basket of charcoal dust (drainage and insulation); the walls rise a cubit and the stone sole (palm and 3 digits thick) lies on them and the ledge; the front is sloped to the litharge channel; copper plates on the sole there "so that the silver-lead or other alloy may be more rapidly heated" (pp. 467-468).
- Dome: a half sphere of iron bands (3, each a palm wide, a digit thick, 1 ft and 2 ft apart) and 18 curved bars riveted to them, iron plates wired inside and lute-lined; "the dome is only one foot and three palms high" (p. 468). Four apertures: a rear one 2 ft wide at the bottom (workman's access and the litharge side); the second over the litharge channel, 2.5 ft wide; two small ones, 3 palms wide by about a foot, for the tuyere pipes (rolled iron plate, 2 palms 3 digits long, bore 3.5 digits); a lid on top with three curved bars, finger-size holes in the lining plates to grip lute (pp. 468-469). Three rings on the middle band take chains from a crane so the dome can be lifted off for each charge (pp. 469, 476-480).
- Lute on sole, copper plates and rock: lute mixed with straw "three digits" deep rammed to one; the stone top is lute with straw a palm thick; "All this, as soon as it becomes loosened, must be repaired" (p. 469).
- Variants: at Freiberg the upper part is vaulted like an oven, 4 ft high, with a front litharge aperture (1.5 ft x 1 ft) and two side access holes (p. 482); in Poland and Hungary the vault is the same but the bed is solid, with a passage along one side covered by iron bars, wood laid on the bars to throw flame into the crucible, charge 80 to 100 centumpondia (p. 483); the Moravians and Carni, who make little silver ("a bes or five-sixths of a libra"), use an open crucible without a cover "exposed to the wind", cakes under dry wood and green wood, bellows-fanned (p. 483).
4.3 The hearth of ashes (the test) (pp. 469-473)
- Made by one artificer in two working days: "distributes the operation into two shifts of two days" (p. 469). Day 1:
- Sprinkle ash in the lute and sweep; throw "sifted ashes" (ashes already leached for lye, because fatty ashes must first be burnt again) and dampen to ball-ability; smooth with the hands; slope the crucible to the middle; tamp with a rammer; form the litharge channel with two small wooden rammers (p. 469-470). The artificer then "with shoes on" tramples it in every direction, tamps again, marks a circle with his bare right foot and cuts it with a curved iron plate or sieve-rim strip (p. 471). Channel tamped; aperture blocked with a stone and board while ashes are poured in "again and again" and rammed; dry ash sieved over; three baskets of damp ash round the margin; the dome lowered; he climbs in and builds the ash rim "lest the molten alloy should flow out"; a basket of charcoal and a shovel of glowing coals are thrown in (p. 471).
- Drying: this "work and labour is finished in the space of two hours". Then burning charcoal on an iron plate and wooden block under the channel; the crucible "is heated in an hour"; the charcoal is stirred with the hooked bar at hourly intervals "or some blackness would remain in the crucible and that part would be damaged, because it would not be sufficiently dried"; "which takes three hours; then the crucible is left quiet for the remaining two hours" (p. 471).
- At "the hour of eleven" he sweeps out the ashes, wipes the whole crucible with a linen rag in ash-water (two buckets of 5 Roman sextarii each) "lest the crucible, when the metals are being parted, should break open", rubs it with a doeskin, fills cracks. At the left of the channel he sets two pieces of hearth-lead (better a brick, which "gets much hotter" and so forms litharge faster) as a baffle so the bellows wind does not blow the litharge away. The middle of the crucible is "two palms and as many digits deeper" (p. 472). Some smear with incense and egg white, or egg and bullock's blood, or sift lime (p. 472).
- Charge: lead alloy weighed; "he sometimes puts a hundred centumpondia into the crucible, but frequently only sixty, or fifty, or much less" (p. 472; 3.2 t maximum; Hoover fn 31: 3.5 short tons if Roman, 5.7 if German). Three small bundles of straw strewn "lest the lead by its weight should break the surface". Cakes of lead alloy are stacked in the channel, along the dome sides and inside through the top aperture; the remainder on the second day. Dome lid on, joints luted, half a basket of charcoal thrown through the tuyere aperture. "These hours all reckoned up make a sum of eight hours" (pp. 472-473).
4.4 The blow (day 2, pp. 473-476)
- Morning: two shovels of live charcoal, small pieces of fir or pitch-pine laid on, water-gates opened so the bellows wheel turns. "In the space of one hour the lead alloy is melted" (p. 473).
- Fuel for the blast: four sticks of wood, 12 ft long, through the rear dome hole and as many through the channel, weighted at the ends and supported on trestles with an iron plate against litharge splash; the remaining cakes laid on the sticks with a fork. The wood burns off at the front end and the master "drives into them pointed iron bars" to push them forward (p. 473-474). Fuel total: forty long sticks suffice; the variant with sticks through the tuyere aperture needs sixty (p. 474). After the lead "has been heated for two hours, it is stirred with a hooked bar, that the heat may be increased." The fuel is billets and charcoal, not charcoal alone. No quantity of charcoal is given for day 2.
- Silver ore (native, glance, grey, ruby) flattened, cut, heated in an iron crucible and poured into the molten lead "in order that impurities may be separated" (p. 473).
- Hard-to-part alloy: "he throws copper and charcoal dust into the molten silver-lead alloy"; impure alloy: argol and Venetian glass, or sal-ammoniac, or glass and Venetian soap, or 2 argol to 1 iron rust, some with a little saltpetre, "a bes or a libra and a third" per centumpondium (p. 474). The scum ("the lead makes this scum when it becomes hot") is drawn off through the channel with a rabble and charcoal; frequent stirring makes less (pp. 474-475).
- "Within the space of a quarter of an hour the crucible absorbs the lead ; at the time when it penetrates into the crucible it leaps and bubbles." The master takes an iron-ladle sample and assays the silver content (p. 475).
- Litharge: when hot, "he extracts with a bar the litharge which comes from the lead and the copper"; "it might more rightly be called spuma of lead than of silver" (p. 475). Roman practice: the litharge flowed to a lower crucible and was lifted with a stick and rolled to tubes; "now ... they make it into bars" (p. 475; Hoover fig. p. 481). If the alloy threatens to flow out with the litharge, a pointed cylinder of lute on a hooked bar is set against it (p. 475).
- Finish: "when the colour begins to show in the silver, bright spots appear, some of them being almost white, and a moment afterward it becomes absolutely white." The assistant shuts the water-gates and the bellows stop; the master pours several buckets of water on it (some beer), the cake is raised with a pointed iron bar (3 ft, wooden handle 4 ft) and laid on a stone; hearth-lead and litharge are chipped off and it is scrubbed with brass wire (p. 475). Product: the blicksilber cake. Not stated: cake size or weight, though the charge and the assay give it (see yields).
- Total blow time on the second day: NOT STATED. INFERRED only: the text lists 1 h to melt, 2 h to the first stir, then the absorbing, litharge and colour phases; the book does not add them up.
4.5 Yields and losses
- Silver content is assayed from the ladle sample and "when the lead is separated from the silver, more silver is frequently found than when it was assayed" (3 unciae 3 drachmae per centumpondium becoming 3 unciae and a half; p. 475; Hoover fn 36: about 82 oz and 85 oz per short ton).
- The lead is not all converted to litharge: "Often the hearth-lead remaining in the crucible is a palm deep"; it is dug out with the ashes, sieved and the pieces added to the hearth-lead. Hearth-lead (molybdaena, saturated furnace bottom) and litharge go back to the blast furnace as charge and flux (pp. 475, 410; Hoover fn 37, p. 475).
- HOOVER fn 26 (p. 464): four periods of cupellation; Winkler's average from a 100 centner charge: 2 centners abzug (64% lead), 5.5 centners abstrich (73%; OCR "5i"), 21.5 centners herdtplei (60%), 18 centners impure litharge (85%), 66 centners litharge (89%), 113 centners total; "lead loss at from 8% to 15%", blicksilber about 90% silver. Freiberg practice of 1837, not 1556.
- De-silvered lead from re-melting hearth-lead: about a drachma of silver per centumpondium, against 3 unciae before (p. 410).
- Cupellation product mass for a 100-centumpondium charge at 3 unciae per centumpondium: INFERRED 300 unciae = 25 librae (8 kg) of silver.
4.6 Poison and crew
- Figure legend (p. 474): "The foreman when hungry eats butter, that the poison which the crucible exhales may not harm him, for this is a special remedy against that poison."
- Crew: the master (artificer) preparing and tending the crucible; an assistant (stirs the charcoal, laps lute, closes the sluice); a foreman; a labourer to carry out the cake; a second workman to lute the dome joint (fig. legends, pp. 470, 475-476). Two or more men; not stated as a number.
4.7 Repair between campaigns (p. 476, 469)
- The ashes that pass the sieve are re-used for new cupels. The yellow pompholyx on the furnace walls and dome rings near the apertures is cleared "when much of it has accumulated" (p. 476). The lute, dome lining and sole are repaired "as soon as it becomes loosened" (p. 469). The crucible of ash is rebuilt for each charge (4.3).
4.8 Refining the silver (test, p. 483-487)
- A second, smaller furnace: arched brick chamber 3 ft high over a hearth 5 ft x 4 ft, with a round pit "a cubit wide and two palms deep" holding a clay "test" packed with equal parts bone ash and crucible ashes, tamped with a toothed pestle, cut, sifted and rolled smooth with a hard-wood ball (pp. 483-485). Tests hold 15 to 50 librae of silver (Hoover fn 40: 155-515 troy oz). Cakes are heated, broken, set standing in the test, blown with a 4 ft 2 palm double bellows in a bronze pipe; "The silver melts when it has been heated in the test for about an hour"; fir billets over it; refined "in two or three hours"; a bar-dip assay on a brass block; iron door lowered to protect the refiner's eyes; granulated copper or lead dropped in if impure (p. 487). "Of each bes of silver one sicilicus is consumed" (p. 487): INFERRED 1/32, about 3 percent; "or occasionally when very impure, three drachmae or half an uncia".
5. The assay furnace, muffle and cupel (Book VII, pp. 223-241), briefly
- Three assay furnaces, all small and charcoal-fed: brick (a cubit high, 1 ft wide inside, 1 ft 2 digits long, grate plate 5 digits above the hearth, mouth a palm high by 5 digits, p. 224); iron (four iron bars 1.5 ft high, plate-covered, lute-smeared, p. 225); clay (portable, base two earthen tiles 1 ft 3 palms by 1 ft 1 palm, inside 3 palms 1 digit by 3 palms 3 digits, 1 ft high, mouth a palm high, clay door and lid, p. 226). Temporary: three bricks, an iron plate, three bricks more (p. 226). Set high for scorifiers through the mouth, low for crucibles through the top (p. 226).
- Iron hoop (for crucible copper assays): 2 palms high, diameter "one foot and one palm", notched for the blast, daubed with lute, on a chimney hearth in front of a double bellows with a hand lever (5.5 ft) and a lead weight on the upper board (p. 226-227). Copper-ore assay: roast 6 to 8 hours, crush, wash, re-roast; 3 centumpondia (lesser weights) of concentrates with 3 each of copper scale, saltpetre, Venetian glass in a triangular crucible in the hoop, gentle then strong blast, break the crucible for the copper button (p. 245). Copper-silver alloy: half-hour of bellows after a quarter hour of heating, "copper which is devoid of lead requires this time to become hot and to melt" (pp. 249-250).
- Muffle: "made of clay, in the shape of an inverted gutter tile ; it covers the scorifiers, lest coal dust fall into them"; a palm and a half broad, a palm high, nearly as long as the furnace, as thick as a stout earthen jar, side windows two to four, back windows two, a space of 3 digits round so charcoal lies beside it (pp. 227-228).
- Cupels: of leached wood ashes, ideally beech; best of deer-horn ash, or bone ash; sprinkled with beer or water, ground, pressed in brass bottomless moulds with a pestle having a round knob; kept dry and long; the small mould for gold/silver from lead, the large for silver from copper and lead (pp. 228-231). Works making silver from copper mould cupels from 2 parts of cupellation- hearth ashes and 1 part bone ash (p. 230).
- Cupelling: cupels are pre-heated "in the empty space" under the muffle or they crack and the lead sputters; "when the heat of the fire has nearly consumed the lead, then is the time when the gold and silver gleam in their varied colours, and when all the lead has been consumed the gold or silver settles in the cupel" (pp. 240-241). Remove it while warm. Too cold: the lead "remains fixed and set, and is covered by a kind of skin". Too much sputtering: cover with glowing charcoal (p. 240).
6. Liquation (Book XI), two lines (HOOVER's outline, p. 491)
- "if a copper-lead alloy, containing a large excess of lead, be heated in a reducing atmosphere, above the melting point of lead but below that of copper, the lead will liquate out and carry with it a large proportion of the silver." Not finished in one pass; the argentiferous lead then goes to the cupellation furnace of section 4 and the residual copper is dried and refined. Hoover quotes losses for 100 lb of argentiferous copper: 32-35 lb lead and 5-6 lb copper (Karsten), Lautenthal 1857-60: 25 percent of the silver, 9.1 percent of the copper (p. 491). Later centuries.
7. Furnaces and heats at a glance
| Furnace | Inside size | Blast | Charge | Heat length | Products | Page |
|---|---|---|---|---|---|---|
| Smelting furnace, closed tap-hole (rich Ag/Au ore) | 5 palms wide x 6 palms 1 digit long (about 37 x 46 cm); walls 6 ft sides, 7 ft back, 5 ft front | 2 water-wheel bellows, one copper pipe, tuyere a cubit up, slanting up | slag, pyrite cakes, 2 baskets ore + litharge + easy-fuse stones, 1 basket charcoal, slag; 100 centumpondia lead in forehearth | warm 0.5 h; tap each charge; day's work 8 h (10 h) | slag, matte cakes, silver-lead alloy | 355-382 |
| First furnace, tap-hole always open (hidden) | as above; forehearth 1.5 ft above floor, dipping-pot below | as above | ore + fluxes + charcoal; lead flux; easy-fuse stones | continuous | alloy, cakes, slag | 385-388 |
| Third furnace (copper, lead, poor ore) | higher and wider; not given | larger bellows | ore with cakes, lead ore, limestone, easy-fuse stones, iron-stone | "three days and three nights", 12 h shifts | matte cakes, copper (after 2nd to 7th roast), alloy | 388-390, 401-408 |
| Copper example, first smelter | not given | water bellows | 18 vessels (about 5,400 lb) ore + 3 cartloads lead slag + 1 cartload schist + 0.2 centumpondium easy-fuse stone + concentrates | 12 h | 6 centumpondia primary cakes + 0.5 alloy | 405 |
| Carni lead hearth (low, no bellows) | two low walls under a vault; rear hole | wind/draught only | green wood under dry wood under burned ore | not given | lead into crucible | 391 |
| Tin furnace | top 2 x 1 ft, 8-9 ft high, sandstone hearth 2.75 x 2.75 x 2 ft | small bellows | tin-stone and charcoal | not given | tin to forehearth | 411 |
| Hearth bloomery (iron) | hearth 3.5 ft high, 5 x 5 ft, crucible 1 ft deep x 1.5 ft wide | water bellows in a pipe | charcoal layers and shovelfuls of ore with unslaked lime | 8, 10 or 12 h | bloom 2 to 3 centumpondia, slag, hard iron | 420-423 |
| Stuckofen (iron, hard ore) | wider and higher than the blast furnace; not given | water bellows | nut-size ore + charcoal from a stair | not given | iron for the smith | 423 |
| Roasting heap | "area of fair size", open front; wood crate 1-2 cubits high | none | wood layers, ore large to small, sand or dust cover | 1 to 3 roasts; schist heap 30 days or more | roasted ore | 273-279 |
| Roasting stall (walled bay) | sides and back walled, front open; wood 12 ft long | none | 4 layers of wood, ore on top | not given | roasted tin or lead ore | 275 |
| Cake roasting pits | 12 x 8 x 3 ft, open front and top | none | crate of wood under cakes; slow then fierce; water by night | 2 roasts (pyrites), 1 (cadmia) | roasted matte | 349-350 |
| Cake roasting furnaces (3 per smelter) | 6 ft high, 4 ft wide, open front | none | faggot bundles under copper cakes | 7 or 9 passes | roasted matte | 351 |
| Cupellation furnace | stone wall 4 ft 3 palms high, 1 ft thick, 14 stones; dome half-sphere 1 ft 3 palms high; crucible diameter not given | 2 water bellows through two tuyere pipes | up to 100 centumpondia lead alloy on ash hearth; 40 sticks 12 ft long | day 1 prepare 8 h; day 2 melt 1 h, blow to white silver, time not given | silver cake, litharge, hearth-lead, scum | 467-476 |
| Refining test | hearth 5 x 4 ft; pit a cubit wide, 2 palms deep | double bellows, bronze pipe | 15 to 50 librae silver | melt 1 h; refine 2 to 3 h | refined silver | 483-487 |
| Assay furnace and muffle | brick: 1 ft wide x 1 ft 2 digits, a cubit high | none, draught holes; iron hoop with double bellows | scorifier, cupel, triangular crucible | not given | buttons | 224-228 |
8. Not in the book
- No temperatures; no tuyere air volume; no slag composition; no yield percentage for copper from ore or iron from ore.
- No foot-bellows or horse-driven bellows for smelting; no hand-worked bellows for the copper furnace (only the Lusitanian tin one and the assay double bellows).
- No dimension for the third (copper) furnace, the high bloomery, the roasting stall interior, or the cupellation crucible diameter; no total blow time for cupellation; no cake weight in the silver cake.
- No furnace built for the oxide-ore-only first copper smelt. The nearest period text is Hoover's note on the Bronze Age pit smelt (p. 402) and Theophilus (p. 405).
- No separate "campaign" length is stated for a smelting furnace; the third furnace runs 3 days and 3 nights "providing there be no defect either in the hearth or in the forehearth".
- Hoover's metric: none; only English inches and troy grains.