Collections: Iron, How Did They Make It? Part II, Trees for Blooms

This week we continue our four-part (I, II, III, IVa, IVb, addendum) look at pre-modern iron and steel production. Last week we prospected our iron ore and extracted it from the ground and did some initial mechanical processing (washing, sorting, crushing). This week, we’re going to make our way from just rocks to an actual mass of metal rather than just some metal-bearing ore. As we’ll see, we are going to do this by applying heat and (more importantly) chemistry:

Note that this week is going to be spent just getting our iron ore into being an iron bloom, the first two steps.

Warning: Many, many trees were harmed in the making of this iron.

As always, if you like what you are reading here, please share it; if you really like it, you can support me on Patreon. And if you want updates whenever a new post appears, you can click below for email updates or follow me on twitter (@BretDevereaux) for updates as to new posts as well as my occasional ancient history, foreign policy or military history musings.

But let’s start with the single largest input for our entire process, measured in either mass or volume – quite literally the largest input resource by an order of magnitude. That’s right, it’s…


The reader may be pardoned for having gotten to this point expecting to begin with exciting furnaces, bellowing roaring flames and melting all and sundry. The thing is, all of that energy has to come from somewhere and that somewhere is, by and large, wood. Now it is absolutely true that there are other common fuels which were probably frequently experimented with and sometimes used, but don’t seem to have been used widely. Manure, used as cooking and heating fuel in many areas of the world where trees were scarce, doesn’t – to my understanding – reach sufficient temperatures for use in iron-working. Peat seems to have similar problems, although my understanding is it can be reduced to charcoal like wood; I haven’t seen any clear evidence this was often done, although one assumes it must have been tried.

Instead, the fuel I gather most people assume was used (to the point that it is what many video-game crafting systems set for) was coal. The problem with coal is that it has to go through a process of coking in order to create a pure mass of carbon (called ‘coke’) which is suitable for use. Without that conversion, the coal itself both does not burn hot enough, but also is apt to contain lots of sulfur, which will ruin the metal being made with it, as the iron will absorb the sulfur and produce an inferior alloy (sulfur makes the metal brittle, causing it to break rather than bend, and makes it harder to weld too). Indeed, the reason we know that the Romans in Britain experimented with using local coal this way is that analysis of iron produced at Wilderspool, Cheshire during the Roman period revealed the presence of sulfur in the metal which was likely from the coal on the site.

We have records of early experiments with methods of coking coal in Europe beginning in the late 1500s, but the first truly successful effort was that of Abraham Darby in 1709. Prior to that, it seems that the use of coal in iron-production in Europe was minimal (though coal might be used as a fuel for other things like cooking and home heating). In China, development was more rapid and there is evidence that iron-working was being done with coke as early as the eleventh century. But apart from that, by and large the fuel to create all of the heat we’re going to need is going to come from trees.

And, as we’ll see, really quite a lot of trees. Indeed, a staggering number of trees, if iron production is to be done on a major scale. The good news is we needn’t be too picky about what trees we use; ancient writers go on at length about the very specific best woods for ships, spears, shields, or pikes (fir, cornel, poplar or willow, and ash respectively, for the curious), but are far less picky about fuel-woods. Pinewood seems to have been a consistent preference, both Pliny (NH 33.30) and Theophrastus (HP 5.9.1-3) note it as the easiest to use and Buckwald (op cit.) notes its use in medieval Scandinavia as well. But we are also told that chestnut and fir also work well, and we see a fair bit of birch in the archaeological record. So we have our trees, more or less.

Forests and Fellers

The bad news is that while ancient sources are often very interested in trees (entire books about them, in fact), they are generally interested in trees used to make things like ships, buildings, furniture and weapons; essentially, elite products. They are not interested in trees used as fuel. Indeed, Latin marks this distinction, where wood for building was materia whereas wood for burning (but also, it seems, bulk wood being transported overseas) was lignum; our sources care greatly about the former and only minimally about the latter. And so as soon as we get very far into the question of the harvesting and preparation of fuel woods, our evidence just about drops away entirely, save for a few poor mentions of this or that tree being good for charcoaling (a crucial process we’ll get to in a moment).

Via Wikipedia, a detail of an Assyrian relief from the palace of Sargon II (r. 722-705) at Dur-Sharrukin, now in the Louvre, Paris, showing the transport of Lebanese cedars; in this case, clearly building timber.

Consequently, our ability to see the fellows felling the forests (say that five times fast) is limited. Medieval ‘foresters’ are often more visible, but much like we noted last time that when Georgius Agricola says ‘miner’ he means ‘mine owner,’ my understanding is that foresters in the Middle Ages were something closer to administrators of the forest (responsible for letting out contracts, catching poachers, etc.; essentially a sheriff but in the woods) rather than simple tree-fellers.

So who did the actual tree-cutting? I must confess, I have found relatively little evidence for the social standing of ancient tree-fellers. In quite a lot of cases, they must not have been meaningfully distinct from the local peasantry or other sources of unskilled rural labor. Clearly a lot of woodcutting was done by the rural population that bordered the forests to clear spaces for fields, gather fuel and firewood and so on, and consequently it seems like the basic skills of tree-felling may have been relatively common. The Latin word for a wood-cutter was a lignator (or sometimes a caesor, which meant ‘cutter’ but could mean of wood (lignorum caesores) or of stone), but that word most often appears in military contexts to mean soldiers tasked with cutting wood for fuel, not full-time lumberjacks. Evidence for the medieval period is somewhat better and also generally suggests that the local peasantry was employed in the wood-cutting itself (for this, note J. Birrell, “Peasant Craftsmen in the Medieval Forest” Agricultural History Review 17.2 (1969): 91-107). As we will see below, often wood cut for charcoaling was cut by the colliers themselves, who we will discuss below. It seems hard to imagine that there wasn’t some division of labor in larger operations (like on Elba or at Populonia), but how that might have been structured is not clear from the limited evidence.

Not all timber works were so easily acquired, of course. While ancient wood-cutters are hard to see in the evidence, ancient sawyers and carpenters are more visible; records from building programs in Athens and Delphi suggest that skilled sawyers (seemingly always assisted by at least one unskilled worker) were paid at least as well as citizen oarsmen in the Athenian navy and in some cases rather better. The presence of English surnames like Carpenter, Cooper, Fletcher, Bowyer, Turner, Sawyer and Wheeler speak to the fact that these were specialized crafts in medieval England; the absence of wood-cutting surnames further suggests that the bulk labor of felling was mostly done by the local rural workforce. Consequently, the social status of the average timber-cutter seems to have been about the same as that of a local peasant, serf or small-farmer, because by and large these seem to have been the same people; while the work done once the tree was down and barked might be done by specialists (but is far less important for trees that are going to be charcoaled). There were also clearly specialist timber merchants, even in the ancient world, and the degree of their visibility, especially in timber-rich regions suggests that they could do quite well for themselves (although, like most merchants, we effectively never see them penetrate into the ruling class), but again, these merchants were likely working with building timbers because, as we’ll see, charcoal wood doesn’t tend to travel very far.

Via Wikipedia, a detail of an Assyrian relief from the palace of Sargon II (r. 722-705) at Dur-Sharrukin, now in the Louvre, Paris, showing the shipment of timber, in this case likely down the Tigris to the Assyrian heartland. High quality building timbers or ship timbers do seem to have been valuable enough to be worth shipping long distances (Athens famously imported its ship-timber from Macedon), but fuel timbers tended to come from closer to hand. But I had this picture and I wanted to use it.

The largest stock of forest-land was typically owned by the state, but private landholders owning their own forests also played a role, albeit generally a small one. In Macedon, the king owned the forests and controlled the supply of lumber, granting or revoking the authority for communities within his territory to take advantage of woodland resources; the practice seems to have been the same, Meiggs (op cit.) notes, in the Near East. In Roman Italy, a large amount of the forest-land was held by the state and contracted out for timber-cutting; Meiggs supposes that figures called saltuarii may have been responsible for making sure that these contracts were carried out properly (much like the later medieval forester, discussed above). We know also from Roman legal enactments later in the empire that it was common for large estates with woodlands and pastures to also have saltuarii, suggesting they might contract out their woodlands in much the same way. Likewise, large forests in the European Middle Ages tended to fall under royal ownership, but Birrell (op cit.) also notes significant timber exploitation from private wood owners in thirteenth century England.

In terms of the practice of timber-cutting, it doesn’t seem to have changed very much through the pre-modern period, although the availability of iron and later steel axes represented a significant improvement in efficiency. Wood-felling axes in the ancient world were the standard single-headed affair (with a heavier axe-head than axes for war; war axes tend to have very thin, light heads). Ovid (Met. 8.775) writes of using ropes on trees ‘loosened by the countless blows of the axe’ being used to pull the tree down, presumably to guide the direction of its fall; we see the same method being used on Pharaoh Seti I’s (r. 1290-1279) relief at Karnak to fell Lebanese cedars:

Pliny and Theophrastus, writing about timber, both place quite a lot of importance on the season of felling, but this was mostly for wood used in things like construction and ship-building; quite evidently (see below), wood would have been cut for the furnaces year round. The practice of pollarding and coppicing (which we have actually discussed before) has more relevance here; pollarded or coppiced trees are pruned in the upper branches to produce a dense set of relatively thin branches for easy harvesting. We know that the Romans did this (Propertius mentions it) and it seems to have been quite common in the Middle Ages for trees intended to supply fuel wood (probably mostly fuel for cooking and home-heating, but possibly also for charcoaling).


With just one exception, however, we cannot jump directly to using that wood to process our ore. Wood, even when dried, contains quite a bit of water and volatile compounds; the former slows the rate of combustion and absorbs the energy, while the latter combusts incompletely, throwing off soot and smoke which contains carbon which would burn, if it had still been in the fire. All of that limits the burning temperature of wood; common woods often burn at most around 800-900 °C, which isn’t enough for the tasks we are going to put it to.

Charcoaling solves this problem. By heating the wood in conditions where there isn’t enough air for it to actually ignite and burn, the water is all boiled off and the remaining solid material reduced to lumps of pure carbon, which will burn much hotter (in excess of 1,150 °C, which is the target for a bloomery). Moreover, as more or less pure carbon lumps, the charcoal doesn’t have bunches of impurities which might foul our iron (like the sulfur common in mineral coal).

Via Wikipedia, A pile of wood for charcoaling. Note the central shaft for the fuel.

That said, this is a tricky process. The wood needs to be heated around 300-350 °C, well above its ignition temperature, but mostly kept from actually burning by lack of oxygen (if you let oxygen in, the wood is going to burn away all of its carbon to CO2, which will, among other things, cause you to miss your emissions target and also remove all of the carbon you need to actually have charcoal), which in practice means the pile needs some oxygen to maintain enough combustion to keep the heat correct, but not so much that it bursts into flame, nor so little that it is totally extinguished. The method for doing this changed little from the ancient world to the medieval period; the systems described by Pliny (NH 16.8.23) and Theophrastus (HP 5.9.4) is the same method we see used in the early modern period.

First, the wood is cut and sawn into logs of fairly moderate size. Branches are removed; the logs need to be straight and smooth because they need to be packed very densely. They are then assembled into a conical pile, with a hollow center shaft; the pile is sometimes dug down into the ground, sometimes assembled at ground-level (as a fun quirk of the ancient evidence, the Latin-language sources generally think of above-ground charcoaling, whereas the Greek-language sources tend to assume a shallow pit is used). The wood pile is then covered in a clay structure referred to a charcoal kiln; this is not a permanent structure, but is instead reconstructed for each charcoal burning. Finally, the hollow center is filled with brushwood or wood-chips to provide the fuel for the actual combustion; this fuel is lit and the shaft almost entirely sealed by an air-tight layer of earth.

Via Wikipedia, a charcoal fire in process. Note that the wood have been covered over with clay to create an impermeable layer that won’t permit excess oxygen into the process.

The fuel ignites and begins consuming the oxygen from the interior of the kiln, both heating the wood but also stealing the oxygen the wood needs to combust itself. The charcoal burner (often called collier, before that term meant ‘coal miner’ it meant ‘charcoal burner’) manages the charcoal pile through the process by watching the smoke it emits and using its color to gauge the level of combustion (dark, sooty smoke would indicate that the process wasn’t yet done, while white smoke meant that the combustion was now happening ‘clean’ indicating that the carbonization was finished). The burner can then influence the process by either puncturing or sealing holes in the kiln to increase or decrease airflow, working to achieve a balance where there is just enough oxygen to keep the fuel burning, but not enough that the wood catches fire in earnest. A decent sized kiln typically took about six to eight days to complete the carbonization process. Once it cooled, the kiln could be broken open and the pile of effectively pure carbon extracted.

Via Wikipedia, a finished batch of charcoal. Do note how different this charcoal is in composition from the charcoal briquettes one purchases these days at the store: it is much more friable, which makes long-distance transport difficult.

Raw charcoal generally has to be made fairly close to the point of use, because the mass of carbon is so friable that it is difficult to transport it very far. Modern charcoal (like the cooking charcoal one may get for a grill) is pressed into briquettes using binders, originally using wet clay and later tar or pitch, to make compact, non-friable bricks. This kind of packing seems to have originated with coal-mining; I can find no evidence of its use in the ancient or medieval period with charcoal. As a result, smelting operations, which require truly prodigious amounts of charcoal, had to take place near supplies of wood; Sim and Ridge (op cit.) note that transport beyond 5-6km would degrade the charcoal so badly as to make it worthless; distances below 4km seem to have been more typical. Moving the pre-burned wood was also undesirable because so much material was lost in the charcoaling process, making moving green wood grossly inefficient. Consequently, for instance, we know that when Roman iron-working operations on Elba exhausted the wood supplies there, the iron ore was moved by ship to Populonia, on the coast of Italy to be smelted closer to the wood supply.

Via Wikipedia, modern charcoal briquettes. Ancient or medieval charcoal will not have looked like this, but instead have been a flaky mess, like the picture one paragraph up.

It is worth getting a sense of the overall efficiency of this process. Modern charcoaling is more efficient and can often get yields (that is, the mass of the charcoal when compared to the mass of the wood) as high as 40%, but ancient and medieval charcoaling was far less efficient. Sim and Ridge (op cit.) note ratios of initial-mass to the final charcoal ranging from 4:1 to 12:1 (or 25% to 8.3% efficiency), with 7:1 being a typical average (14%).

We can actually get a sense of the labor intensity of this job. Sim and Ridge (op cit.) note that a skilled wood-cutter can cut about a cord of wood in a day, in optimal conditions; a cord is a volume measure, but most woods mass around 4,000lbs (1,814kg) per cord. Constructing the kiln and moving the wood is also likely to take time and while more than one charcoal kiln can be running at once, the operator has to stay with them (and thus cannot be cutting any wood, though a larger operation with multiple assistants might). A single-man operation thus might need 8-10 days to charcoal a cord of wood, which would in turn produce something like 560lbs (253.96kg) of charcoal. A larger operation which has both dedicated wood-cutters and colliers running multiple kilns might be able to cut the man-days-per-cord down to something like 3 or 4, potentially doubling or tripling output (but requiring a number more workers). In short, by and large our sources suggest this was a fairly labor intensive job in order to produce sufficient amounts of charcoal for iron production of any scale.

As should be obvious from the complexity of the work, charcoal burning was a specialized profession; gaining a sense of how dense to set the pile (very dense, generally), how much fuel to add, how large to make it, and how to gauge the process of carbonization merely by the sight of the smoke are the sort of things one learned by experience, presumably first as an assistant or apprentice. In some cases, this job was done by peasants who did it part-time (the colliers certainly seem to come from the local peasantry); Birrell (op cit.) notes that in the fourteenth century records for the Cumberland Forest of Inglewood that the majority of licenses for charcoaling were for seasons rather than all-year, suggesting most of the colliers had other occupations. Nevertheless, the presence of ‘Colier’ (or collier), Askebrunner, Ashburner, or the Latin carbonarius or cinearius as surnames or professions appearing in thirteenth and fourteenth century English sources attests to the fact that this was a specialized skill-set, one which defined the practitioner, whether pursued part-time or full-time. In many cases, as Birrell notes, the fellows taking out licenses to burn charcoal (almost certainly for assistants or servants) were themselves smiths who also owned forges in the area.

And yet the colliers themselves seem not to have been well thought of by society despite these skills. Their occupation was a solitary one, since they had to attend their charcoal piles for the week or so it took for them to carbonize, which meant spending weeks at a time out in the forest. Mostly beneath the notice of our ancient sources (Pliny and Theophratus both describe the charcoaling process without mentioning the men doing that process); it seems likely that colliers fell under the same elite derision that attached to many overly smelly jobs (like tanners). German preserves an odd bit of this dislike with Köhlerglaube (“collier’s faith”) meaning a blind faith or loyalty in a rather negative sense, ostensibly because the colliers, alone in the forest (rather than in church) had no opportunity to learn the details of their own faith. Fourteenth century charcoal burners working directly for the English King (Birrell, op cit., 98n.2) were paid 3d (that is, 3 pence) a day, a decent but not extravagant wage at the time (for a broad comparison of contemporary wages, see this website, which amasses medieval price data), suggesting that the collier’s skill demanded some wage premium, but not necessarily a tremendous amount.


With our ore mined and our charcoal made, we are almost ready to get to our smelting, but first we have one final step to prepare our iron ore for the smelting process: roasting. Roasting solves two problems we have. The first problem is water: our ore, even if it appears dry, almost certainly traps small amounts of water inside the ore. We want to remove that before we subject this ore to extreme temperatures, both because water is just going to absorb the energy (heat) from our fuel, wasting it, and also because small pockets of water inside rock heated in excess of 1,200 °C can pose problems. Driving out the water at this stage also has the added benefit of cracking most ores into smaller bits that will reduce even easier once we get to smelting.

The second problem is chemical, because we are never going to melt this iron. Our furnaces can’t get that hot (and even if they could, melting this iron would cause it to absorb a lot of carbon from our fuel, which we do not want). So we’re going to be reducing our iron – that is, getting it to change chemically with the exposure of heat. That means the chemical composition of our iron matters a lot and we have to solve our chemical problems before we can smelt our ore.

The good news is that some ores of iron reduce fairly easily and directly, most notably hematite and the hydroxide-iron ores like limonite and goethite. They reduce fairly easily (but the latter two tend to come with lots of water that needs removing). But then we have magnetite, which while also an iron-oxide, doesn’t reduce nearly so easily as hematite, and siderite (and other carbonates) which has carbon in it, which we do not want. Moreover, our country rock might have some trace amounts of things like sulfur (or iron-sulfides) in them, which we very much do not want. Sulfur will absolutely ruin our final iron product, so we do not want it floating around when we get to the smelting process.

But let’s say we expose some magnetite (Fe3O4) to heat in an environment where there is some oxygen around – we can get a further oxidizing reaction (I dearly hope I have remembered my chemistry) whereby 4Fe3O4 get along with 1 O2 (your garden variety oxygen in the air) to make 6Fe2O3. And good news, that’s something we recognize – that’s hematite (Fe2O3) which we already know will reduce in our furnace in a bit. Likewise through a slightly more complicated reaction, we can get that pesky carbon out of our carbonate ores like siderite (FeCO3) which release its carbon and oxygen as carbon dioxide and end up forming Fe2O3, which again is our good friend hematite. Likewise, my understanding is that small amounts of sulfur will oxidize to sulfur-dioxide which is pleasantly a gas and so will be on its merry way out of ore as well.

(Edit: The more chemistry minded have pointed out that the hydroxide iron ores – which you have to roast to remove water anyway – will also reduce down to hematite when roasted through a reaction that looks like this: 2 Fe(OH)3 -> Fe2O3 + 3H2O releasing water which – just like the water trapped in the ore itself – boils away in the roasting process.)

The actual process of roasting the ore is in many ways less interesting than the chemistry that goes into why we roast the ores. This is the only step generally done with raw wood (typically dried to remove water), rather than charcoal. The temperatures we need to roast ore are fairly low, typically between 300 to 550 °C or so – enough to boil off the water and trigger our chemical reactions, but that’s it. We can accomplish that by setting a pile of dry wood within a brick kiln, putting the iron ore on top of them, and lighting the whole thing on fire. For very small quantities of iron ore, this can be done very simply (pictures at the link); doing larger batches of ore would have required more careful management of fuel and ore.

With that, we can finally move on to smelting in a…


All of that effort now brings us to a frustrating problem and the reason why nearly every ‘making the sword’ scene in a movie is wrong: iron melts at 1538 °C. That is very hot. To achieve those temperatures over the entire mass of ore (necessary to melt the iron out of it) we would need to get to around 2000 °C, which is unsurprisingly the temperature that modern blast furnaces work at. At those temperatures, a whole lot of chemical reactions happen and the resulting iron comes out molten, enabling it to be separated from the waste material (slag). The problems here are two-fold: first, that molten iron would pick up a ton of carbon, turning it into brittle and functionally useless pig iron and second, prior to the early modern period, most of the world’s furnaces couldn’t achieve those temperatures anyway (we’ll talk about exceptions in China and India in an addendum).

Game of Thrones, S4E1, showing the melting of Ned Stark’s sword. There is a lot wrong with this scene, but let’s start with the obvious fact that the smallish charcoal forge they are using couldn’t achieve heats anywhere near the level required to melt and cast iron (you have to exceed the melting point of something by quite a lot to cast it) and also that this is clearly not iron. Molten iron is bright, sparking white hot (the hot glow of something is a product of its temperature and iron is well past white-hot by the time it melts); this is almost certainly aluminum.
For more on the nonsense of these scenes, Lloyd made a great video a few years back.

So we can’t melt the iron out of the ore. What if we melt the ore out of the iron? This is half of the fundamental principle of a bloomery (the other half being chemical reactions that take place in a carbon-rich environment with lots of heat and iron-oxides).

Via Wikipedia, a modern reconstruction of a basic small bloomery furnace. The hole at the base is for tapping slag and potentially withdrawing the bloom when the process is complete.

A bloomery was a type of furnace that produces a bloom, a sponge-shaped mass of (mostly) pure iron. The basic structure of a bloomery furnace is fairly simple: a conical shaft (or sometimes a lower bowl-shape) made out of some fairly heat-tolerant material (bricks, stone, clay) with one or several holes near the base (called tuyères) for air to enter at the bottom and either a bit beneath the shaft for slag materials to drain into or a larger hole at the base to allow the slag to be tapped. Some types of furnaces (particularly the bowl-types and furnaces which used a bit rather than having slag-tapping to deal with slag accumulation) were effectively single use, being broken up in order to extract the bloom. I am going to focus here on the more sophisticated low-shaft furnaces with slag-tapping because, as Tylecote (op cit.) notes, these were the sort that the Romans used and seem, so far as I can tell, to have provided the basis for later medieval European bloomeries.

That description sounds very complicated, so it is time for one of my trademark badly made diagrams:

After Healy (op. cit.).

With air entering through the tuyères (the diagram shows just one, but furnaces often had several, which may also have had air pushed through them by bellows), it feeds the combustion of the charcoal. To understand what that does for us, we need to get back into the chemistry.

The charcoal – a (nearly) pure mass of carbon – takes some heat and some O2 from the air and produces carbon dioxide (CO2) and a bunch more heat, which both heats up the iron ore we’ve stacked on top of it and the rest of the charcoal, which will react with the first oxygen it meets to repeat the process. But this process is rapidly going to be oxygen starved (this is important) and so some of these reactions are going to produce carbon monoxide (CO) which then, because it is very hot, goes racing up through our iron ore. And carbon monoxide is a lonely fellow – give him some heat and he starts looking for one more oxygen to form carbon dioxide…and would you just look at the hip, happening place he has stumbled into, because he is now surrounded by hematite (Fe2O3) with all that oxygen to spare. So each lonely-lad carbon monoxide grabs one more oxygen dance partner: Fe2O3 + 3CO -> 2Fe + 3CO2. The CO2, being a gas, exits the furnace at the top, flying off into the sky to utterly ruin our climate in revenge, leaving just the pure, metallic iron behind. This process actually begins to happen at only 800 °C or so.

But wait! That hematite we’re reducing didn’t come to us pure, it came to us embedded in other kinds of rock (called gangue). And this is why we need all of that heat. Most of these impurities from the ore are going to be silica (SiO2) and alumina (Al2O3). We need those gone too. In modern furnaces, a flux (typically limestone) is introduced into the mixture because it will readily bind to these compounds to get them out of the iron. In the Roman period, at least, this wasn’t understood for the smelting stage of the process; I’m not sure when exactly ironworkers realized that they could use limestone to deal with this. But iron can also do the job; exposed to enough heat, 2FeO (an iron-oxide we’ll have in abundance because of the processes above) + SiO2 forms Fe2SiO4 (alumina reacts much the same, but I cannot find the formula for the life of me). Of course, that means we lose a decent amount of our iron from having it serve in place of a flux, which further lowers our efficiency. Those new slag compounds melt around 1150 °C, well below the melting temperature of iron.

Via Wikipedia, an early modern (16th cent) bloomery from Georgius Agricola’s De Re Metallica.

And so at long last we are getting somewhere. Our iron-oxides have mostly converted, by reaction with carbon monoxide (and a lot of heat) into pure metallic iron, while most everything that isn’t pure metallic iron has grabbed some of our iron (unfortunate) or some limestone (if we have it), lowering its melting point so that it becomes liquid, dripping down to the base of the bloomery furnace, where we can tap it out.

Once the charcoal burns all the way down and all of the ore has been reduced, what we’ll be left with is a mass of metallic iron – with some slag impurities still left, but tolerably few – in a sponge-shape. In contrast to a blast furnace, which will give us iron with a lot of carbon, our bloom’s carbon content will be extremely low, because the carbon that we had was burned into carbon-monoxide (and then oxidized into carbon dioxide) and so wasn’t sticking around to be absorbed into our growing bloom of iron. Which is fantastic, because we have no efficient way with medieval or ancient technology to decarburize iron (that is, pull the carbon out), so getting iron with super-high carbon counts would be really very bad (we’ll talk about iron and carbon more in the last post of this series when we get into steel).

Via Wikipedia, a hot iron bloom fresh out of the bloomery, encrusted with slag which can be expelled from the iron mass by hammering (we’ll get to that next week).

Efficiency and Ecology

Ok, that was all a lot of fascinating chemistry, but you may well ask what does that mean from a practical standpoint? How much wood and how much ore do we need to produce a given amount of iron?

Obviously, those figures vary wildly based on the quality and type or ore, the quality of our furnace (how well it keeps heat, for instance) and any number of other considerations. Healy, using some of Tylecote’s figures, computes an estimate that Roman smelters, working continuously (two batches in 24 hours) might, in a day produce 16kg of iron from 100kg of ore and 80kg of charcoal. Healy also notes an experiment by E.J. Wynn with a more primitive bowl-furnace took 16lbs of charcoal to produce 1lbs of finished iron. Sim and Ridge, working from a different set of experiments, suggest that about 1kg of finished iron might require 12.3kg of ore and 14.6kg of charcoal (obviously a furnace batch would be larger than this). The variance here is considerable, but remember that the iron-content of the ore isn’t constant, so this is really a range of possibilities (there’s also a difference in experimental procedures here, particularly if bar- and billet-smithing are counted, or if we’re measuring the weight of the bloom).

And, for that charcoal, as you will recall, the charcoaling process averages a mass of finished carbon around 14% of the starting wood. If we take Sim and Ridge’s figures, the 14.6kg of charcoal required for each 1kg of iron would require around 105kg of wood (plus some fuel for the charcoaling process) to produce. And to be clear, we are not done: every stage of iron-working past this also involves losing mass, in some cases because we’re ejecting slag that’s found its way into our iron (so we’re not losing anything of value, but our mass is decreasing because we’ve counted things-not-iron in our 1kg iron bloom) and in some cases because our iron is oxidizing and we’re ejecting the resultant iron-oxide (read: rust) in the forging process, and in some cases because we’re going to need to polish, file and sharpen, which involves stripping off small amounts of iron.

To put that in some perspective, a Roman legion (roughly 5,000 men) in the Late Republic might have carried into battle around 44,000kg (c. 48.5 tons) of iron – not counting pots, fittings, picks, shovels and other tools we know they used. That iron equipment in turn might represent the mining of around 541,200kg (c. 600 tons) of ore, smelted with 642,400kg (c. 710 tons) of charcoal, made from 4,620,000kg (c. 5,100 tons) of wood. Cutting the wood and making the charcoal alone, from our figures above, might represent something like (I am assuming our charcoal-burners are working in teams) 80,000 man-days of labor. For one legion.

The ecological impact of pre-modern iron production was also significant. We know, for instance, that Elba was almost totally deforested during the Roman period to fuel the bloomeries smelting the ore and Pliny notes in his Natural History that smelting (not always of iron) had substantially reduced forest-stocks in parts of Gaul and Campagnia. Roman iron production in the eastern High Weald of England may have deforested something like 500km2 over the course of three centuries and there is reason to believe that Roman-period iron production in this area stopped because of scarcity of fuel, rather than ore. Iron-working was hardly the only factor in the steady deforestation of Europe, but it was a major factor.

Nevertheless, our trees have not all died in vain. We started this post with some crushed up iron-bearing rocks and some trees and we have managed to produce a bloom – our sponge-shaped mass of iron. Next week, we’ll apply even more heat (did I mention this process requires a lot of fuel?) and some mechanical force in the form of hammers when we finally get to forging.

165 thoughts on “Collections: Iron, How Did They Make It? Part II, Trees for Blooms

  1. > German preserves an odd bit of this dislike with Köhlerglaube (“collier’s faith”)

    The same expression also exist in French, “la foi du charbonnier”.

    > Mo.14dern charcoaling is more efficient

    Got a typo here

    1. Does this date the term to before the France/Germany split? Or are the two so closely linked even after the Franks disbanded?

  2. And, of course, if it took 80,000 person-days to prepare the charcoal for the iron used by one legion, then it took still more person-days to feed, shelter, and clothe the woodcutters and colliers.
    Before reading this, when I read about how expensive a knight’s armor was, I imagined the time it took for a skilled smith to make it. I wasn’t thinking about all the woodcutters standing behind that smith.

  3. This was an issue even into the early industrial period. I’ve read about the records of a mid-sized smith in a mid-sized town sometime around 1800 showing where he bought his fuel. Something like 250 different suppliers were listed, ranging from professional woodcutters with wagons full, all the way down to random kids who brought in armloads of sticks for extra money. (All this is quoted from memory, so don’t take it as gospel, but it’s in that ballpark).

    Switching to coal around 1800 was a huge environmental win per unit fuel burned, however weird that is to modern eyes. (It caused Victorian coal filth because they burned a lot more fuel, of course, but imagine how bad it’d be if they tried to do that with plant matter). It also enabled the Industrial Revolution in no small part, because it meant we could actually produce metal in real bulk.

    1. I’m sure you know this as well, but a big part of the reason they even switched to coal to begin with was that they couldn’t to that with plant matter even if they really wanted to, and they really did, since wood, as mentioned, is actually a much easier fuel source to work with. Unfortunately the tree murdering continued for so long that by the early modern period the deforestation of places like the British Isles were more or less complete, and it was not a matter of choice to use coal as fuel or not, there simply were not enough trees left.

      1. Yup. Simply getting enough fuel to feed the fires was getting very difficult, and doubly so once you add in needs like household fuel, civilian constriction, and naval construction. (The latter wasn’t the biggest, but it did have the most unique requirements, because very old trees of specific species were uniquely valuable, and you need very well-seasoned wood to prevent rotting. The handful of old wooden ships still in use have some trouble finding the necessary timber for repairs and refurbishments even today – has some descriptions of the difficulties of getting wood for restorations historically.)

        The Fangorn scene from LotR seems fairly accurate, all told. A bit oversimplified, but reasonable enough. (Which Bret discussed here:

          1. Well Isengard is near and used the trees of Fangorn Forest for lumber. In the movie Isengard seems surrounded by Fangorn Forest so either name is accurate.

          2. True,But it looks as if you referred to the metal-working scene, which is beneath Isengard not out in the Forest.

        1. You also needed conifers to be light enough for the masts. They were importing those from Scandanavia, and then a new source: America! And the trees were even better, so you didn’t have to build a tall enough mast.

          The lumberjacks of America HATED that because it was a lot harder than ordinary lumber, and they didn’t get paid more. And even before Lexington and Concord, the Americans were doing things to prevent the export. Being cut off entirely was hard. They had even lost the skills to build a mast out of a shorter tree from Scandavia.

      2. An Englishman spoke contemptuously of American trees because they were just trees. If you had towering old trees like that in England, you knew these were the lands of a great land owning family.

    2. Actually, coal was big MUCH earlier. Read The Domestic Revolution: How the Introduction of Coal into Victorian Homes Changed Everything by Ruth Goodman and realize they all were actually cooking and heating their homes with coal in London by the end of the reign of Elizabeth I. (Which had many consequences beyond that. For instance, the iron pots needed to cook helped trigger the Industrial Revolution, too.)

  4. Something I’ve wanted to ask since the last time I saw that video of Lloyd’s – what *would* someone do in Tywin Lannister’s situation where he’d captured an enemy’s sword and *really* wanted two family members to reach have a sword they could point to and say “this was made from the sword of our defeated enemy”

    1. Strip off any grip materials or other flammables, get a smith to heat it orange-hot in their forge, split the sword into two, and forge swords from each half. It’s basically the same as forging a sword from bar stock, at least mechanically.

      I can’t speak to the proper metallurgical considerations for Valyrian steel, but it’s plausible that’ll make it a bit tougher to produce really good swords. (It’ll still be better to re-forge than to cast, though.)

    2. We’ll talk about forging next week, but a smith could break the blade apart, hammer it into two bars and then resmith the final product, though in practice you will probably not have enough material in any one sword to make two, since you will lose some material in the process.

      1. Given how much fuel went into producing workable iron, there must have been attempts to salvage as much loss as possible at every step. Obviously leftover stock would be used in the next project (and hammer-welded where necessary) but were there any known techniques for collecting and reusing smaller bits of waste from filing and cutting? Do any sources go into this?

      2. I am wondering about this- Ice is, after all, supposed to be an unusually large two-handed sword. I imagine it as comparable to the sword which is supposed to have belonged to the Frisian folk hero Grutte Pier, and is now in the Frisian Museum in Leeuwarden. This is 213 cm long and weighs 6.6 kg.

        (While we don’t know if Grutte Pier actually used this particular sword, it is the right age)

        Even losing over half of the material in the process of re-forging, you still have two normal-sized swords’ worth of material in there. And Widow’s Wail was made for a 12-year-old (in the books) so might be smaller than normal…

        1. Ice is a big sword, but it was made for a man of ordinary size.

          As for Widow’s Wail, it wasn’t really made for a boy. Making a weapon for someone too young to fight would be an outrageous waste of this extremely rare metal. The smith would expect Joffrey to grow up and need a man-sized sword, and pass the sword on to his descendents for centuries to come.

          1. Well like with a lot of things Martin does, Ice is an exaggeration and it’s freaking huge. Bigger than historical greatswords and even if we handwave about Valyrian steel being lighter than real steel and otherwise special it’s pretty damn unwieldy.

            Checking the book wiki it’s “Ice is a greatsword as wide across as a man’s hand.The blade is taller than an adolescent Robb Stark and near as tall as Ser Ilyn Payne.”

            That seems like plenty of material to make two swords, especially since with Valyrian steel the smiths would go to EXTREME lengths to produce as little wastage as possible which wasn’t anywhere near as much of a priority with actual iron due to Valyrian steel being vastly more expensive than iron.

          2. Coming back to this much later, I notice something in your reply.

            On the one hand, “Ice was made for a man of average size.”

            On the other hand, the very correct observation that a Valyrian steel sword is made to last for centuries and to be passed down from generation to generation.

            Is it not possible that Ice quite simply was made for some prodigiously strong giant of a man, originally? There is no guarantee that the line of Stark, the kings of the North, were the first owners of that sword, and even if they were, surely some of the Starks were larger and brawnier than others, across the generations.

            Given that it is usually impractical, in modern Westeros, to reshape or rework Valyrian steel weapons, the fact that Ice is a little too big to be a very practical sword for most of the Starks is just… something they would have to deal with. You can’t very well put aside a priceless and irreplaceable family heirloom just because it’s rather unwieldy.

            Some of the Starks might not choose to use it in battle, granted- but they’d keep the blade, in the form that it occupies, even if it was suboptimal. And given the associations of power and status that go with possession of Valyrian steel swords in Westeros, they’d probably try to use the sword in battle unless it was obviously quite suboptimal.

            So it’s quite believable that Ice was bigger than it really ought to be, in order to be the best sword for Ned Stark. If Ned wanted a weapon sized to him personally, he’d hire a normal swordsmith to make him a normal steel sword. If he wanted a Valyrian steel blade, his only choices were ancient family heirlooms, and he’d hardly be looking to trade in his for a smaller one that used to belong to someone else.

  5. Great article, as usual. I had no idea how much raw fuel went into the smeting process, and it’s fascinating to learn about. As usual though, this provokes a question. There are a bunch of places which had human habitation but not much in the way of forestry. Places like Egypt, the huge steppes, what’s now Iraq were hardly noticed for their abundance of wood.

    Yet all of them, at some point, made the switch over to primarily using iron tools and weapons. If the iron forging process is so fuel intensive, and fuel effectively means wood, where did they *get* all their iron from? Did they import all of it? Did they import the staggering amounts of wood necessary to smelt things themselves? How did these places function, and if they did trade for it all, how did they produce enough wealth that someone would be justified coming out there with wagons or ships full of bulk material to flog to them?

    1. I’ve also read that Egyptian architecture was mostly wood; they only used stone for tombs and temples. Maybe they had more wood back then, at least near the river?

      1. I’m not an expert on Egypt by any means whatsoever, but at least if what I pulled out of Old Kingdom was in any way accurate, they were importing wood from what’s now Lebanon shortly after setting up the Ways of Horus, which was towards the end of the First Dynasty, and way before the iron age. (Although they were using copper stuff then and I suspect that copper smelting would also be rather fuel intensive). I can’t imagine they’d be bringing in lots of wood if they had it available locally, but like I said, I’m not really all that knowledgeable about the period, so I could be blowing smoke.

        1. The wood imported by Egypt from lebanon was Cedar, specifically for making ships (and later chariots) and was quite important to the rulers, see the relief of cedars being felled in the blog post. It should, however be noted, that while not as densly vegitated as the forests of continental Europe most of North Africa and the Nile valley was much greener in those times than they are now, and there could be enough timber for some minor local supply. However, the bigger issue is the copper itself. The densly populated Nile area is poor in copper deposits, and the only real deposits in the vicinity are in the middle of the desert to the east towards the Red Sea, in Sinai and in southern palestine. While copper ore most certainly was extracted there, the mines seem to have primarily have been for other things, such as turquise and other precious stones. Copper was extensively traded throughout the mediterranean with Cyprus (and possibly mainland Greece as well) being a prime trader in the metal.

      2. the typical city/town building in ancient egypt was mudbrick.. bricks made from a blend of clay rich mud and straw, either sundried or baked in a kiln. these would be covered with a layer of Clay-Mud for additional structural strength and weather protection. Wood was used mainly for structural elements like roof supports. this method of construction though made it hard to build large buildings.

        Stone was used for the large ‘monumental’ construction.. the palaces, the temples, and so on.

        building a structure entirely from wood would have been ridiculously expensive and something that only the royal family could ever hope to afford.

      1. Makes sense. And it’s probably cheaper to translate finished iron products than it is to transport iron and then inevitably waste some crafting them into tools, weapons, and whatever else it’s being used for.

    2. Wood was mostly too bulky to transport overland (although it does float well – so rafting wood from the Auvergne down the Seine to fuel Paris was a thing). Egypt grows acacia and other scrub trees, which are fine for charcoal. Same for Mesopotamia. There was also a flourishing iron industry along the upper Niger, again not exactly forest country. The timber Egypt imported from Lebanon was large cedar baulks for building.

    3. This is making me really curious what the process looked like for Bronze, that this much work to produce iron in bulk was *easier*.

      1. Bronze is a lot easier, which is why it developed first. The problem is rarer materials — *much* rarer in the case of tin.

        “Tin is the 49th most abundant element in Earth’s crust, representing 2 ppm compared with 75 ppm for zinc, 50 ppm for copper, and 14 ppm for lead.[43] ”

        Iron is roughly 5% of the crust by weight. Converting to ppm is hard because of different atomic masses (assuming ppm is by atom type count), but if we ignore that, iron would be 50,000 ppm. Slightly more abundant.

        (Silver is 0.07 ppm and gold 0.001 according to another page. Which is about the same ratio as their modern prices.)

        1. Iron is about 2% by atom.

          Silicon and aluminum are about half the atomic weight of iron, oxygen is about 1/3, and these make up the vast majority of the crust, so iron’s percent by molecule is about half its percent in weight.

          Or you can calculate it directly, just doing it I got very near 2%.

          1. Each of these things can be found in ‘bronze’ (or more correctly, copper alloy) but result in different qualities in terms of strength, hardness, malleability, etc. and so are not really replacements for the mostly superior copper-tin alloy. Bronze used for military equipment in the ancient world are almost all primary alloys of copper and tin, with the occasional very limited admixture of lead, arsenic or zinc.

      2. I’ve heard that the bronze for classical Greek statues and armor came from so far away, the Greeks didn’t even know where it was. Turns out it was mostly Cornwall, but they really had no concept of the British Isles.

  6. So what’s different between charcoaling wood and coking coal? Looking it up, the basic idea seems to be the same; what’s the additional difficulty that allowed coking to take so much longer to invent in Europe?

    1. I’m curious about this, too. Random searches online don’t really explain why it wasn’t done earlier, though I’ve found several claims that before an understanding of chemistry the reason for both charcoaling and coking was to remove bad smells that’d affect whatever was being produced (foods and otherwise).

      Maybe trees were just easier to access for the payoff? Until more advanced kilns, maybe coke wasn’t sufficiently better than charcoal to merit mining it? Or maybe pit- or mound-coking didn’t remove enough volatiles?

    2. I mean for one thing, coking requires higher temperatures, double I believe, of what it takes to produce charcoal. It was simply harder and more expensive to do. It is also important to remember that coal and coke are not inherently better fuels than wood based fuels, in fact somewhat the opposite without modern industrial techniques. The industrial revolution switched to coal eventually because you can mine coal in vastly larger quantities than you can fell wood, and there’s a larger supply around, not necessarily because coal was actually a better fuel source. In ancient times when you still had woodlands to fell there was really no particular economic incentive to switch to coal.

  7. Could a collier leave the charcoal to cook unsupervised overnight? I imagine that putting it out each evening and restarting each morning would make the whole process even slower and more labor-intensive.

    1. My gut says that starting and stopping would probably be a lot worse for the yield than anything that could drift slightly off-spec overnight, but I’d love to hear an answer from someone with proper domain knowledge.

    2. Well, based on the information provided by Arthur Ransome (noted below), the answer is No. The two burners in the small-scale operation took turns sleeping.

    3. No, once the klin runs, it runs. If you put it out and want to start it again, you have to open the whole thing up (remeber those were one time constructions).
      They probably worked shifts, with two or three people working one klin. I can imagin that in times when everybody worked around the day-night-cycle this added to the percived “oddness” of the colliers.

    4. As mentioned by others previously, no. In the 1930s, when the U.S. Government purchased Hopewell Furnace, which had last fired in 1883, the National Park Service hired one of the colliers, 82-year-old Lafayette Houck, to demonstrate how to make charcoal over the course of three weeks (with his son as an assistant). It was written up as “American Charcoal Making in the Era of the Cold-blast Furnace” (1941).

      According to Houck, they’d usually start the burn late in the evening so they could get one last good night’s sleep, since it wouldn’t need watching that early, but after that they’d be regularly tended around the clock. On the flip side, a collier and one or two assistants would often run 8 or 9 pits at once from a central hut. Each was “around a city block” apart, and contained 25-50 cords of wood. A 30-cord pit would take 10 days to 2 weeks to burn off, so using the 560 kg of charcoal per cord of wood, the 2-3 workers could produce roughly 135,000-150,000 kilograms of charcoal from 8 pits of 30 cords each in two weeks (plus the time to build the pits) given sufficient wood to work with.

      At Hopewell, the work was seasonal. Because of high winter and spring winds, the pits only fired from May to October, with the colliers acting as woodcutters for the rest of the year. The area to be cut was divided into strips 20 axe-handles wide and allocated to each collier, which reminds me of the strips of farmland from village farming. Each tract would be basically clear-cut, and then left fallow for 30 years to regrow.

  8. Other typos noted below, but first a question for Bret or others:

    As I read about the deforestation, I immediately thought of Treebeard’s explanation of why Fangorn was only the “East End” of the original forested land in Middle-earth. I had always thought in term of the shipbuiding of Numenor to explain that, but now I’m realizing how much more important the iron-making process probably was. Has anyone computed numbers for the deforested area to estimate equivalents? Now I’m curious!

    Possible corrections shown below:
    Note below diagram of fuel use: is going to spent –> is going to be spent
    frequently experiments with and –> frequently experimented with and
    a staggering amount of trees –> a staggering number of trees
    most office appears in military –> most often appears in military
    process. Mo.14dern charcoaling –> process. Modern charcoaling
    have to solve and pressing chemical problems –> (did you intend “our”?)
    (slag). The the problems here –> (slag). The problems here
    bright, sparking white –> (is “sparking” the intended word here?)
    carbon that we had was used was burned –> carbon that we had was burned

    1. > Has anyone computed numbers for the deforested area to estimate equivalents?

      Crude estimate, via holding my fingers over the map scale and then parts of Enedhwaith and Miniriath: on the order of 300×600 = 180,000 square miles. And that’s probably a minor part of the Numenorean empire: if the Faithful got to keep Belfalas to Eryn Vorn, the King’s Men majority must have had many colonies well south of Umbar.

      (A reminder that Numenor was sub-tropical, probably south of Umbar: Fonstad’s map here

      Textually the inhabited part of Valinor is equatorial — “Girdle of Arda” — and Numenor could look west to see Tol Eressea; if anything Fonstad’s Numenor looks too far north, though there’s also Tolkien’s description of the climate to reconcile… I’m not sure it’s all consistent, honestly.)

  9. Also meant to mention that apparently small charcoalers were still operating the in the English Lake District even into the early 20th century. Arthur Ransome has his Walker children spend some time with a couple old charcaol-burners and observing their process as part of his book Swallows and Amazons.

    1. yep. even in the book, Saruman’s armaments project for his army would have consumed a huge amount of lumber to support all the smelting and forging being done. given his Orcs were being given uniform gear (short swords, armor, shields, spearheads, etc.. from the description they were equipped with gear not too dissimilar to a Roman Legionnaire, and he had thousands of them to arm, at least a legion’s worth if not more)
      and the film’s version also would require a ton of wood to make all those industrial devices (the cranes and scaffolds, and such) that we see.

      it is also likely that he was using some for chemical engineering.. we know he made explosives and something akin to napalm (seen during the Ent’s attack in the book.) both would require access to Nitrates, the easiest of which to make would be potassium nitrate.. which can be made by mixing Urine and manure with large amounts of wood ash and letting the mix sit (moistened frequently) and interact for months, as evaporation brings a layer of saltpeter crystals to the surface.
      and he’d need wood or charcoal to feed fires used for refining various chemicals as well.

  10. “80,000 man-days of labor. For one legion.”

    That is only about 16 man-days per man. How long did all the ironmongery last before having to be replaced? On the face of it, swords and armour might be expected to last many years. If they lasted 16 years, that would work out at one man-day per legionnaire-year.

    1. One complication is, these are pretty low-surplus societies when it comes to agriculture. The miners have to eat, and the food has to be carted/shipped to them, preferably sourced locally… which in turn creates a pressure against having very many people working at any given iron production site, especially since a BIG site will rapidly deforest the entire surrounding countryside and become ineffectual.

      So for any given pre-industrial iron production center to operate sustainably, it has to operate at some minimum size- realistically determined by ability to supply the miners with food and the operation with fuel. Which means you need quite sizeable populations by ancient standards to support any given operation, and your 80,000 man-days of labor look more like “80 men working for 1,000 days” than “1,000 men working for 80 days.”

      Plus, of course, the legions are only a small fraction of overall demand for iron. Dr. Devereaux just pointed out that the legions will use a bunch of iron tools not covered in the “80,000 man-day” reckoning, and while civilian populations use far less iron per capita than soldiers, there are many, many, many times more of them, many of whom are putting a lot more wear and tear on their tools. How often do the woodcutters felling the trees need new axe blades just to support this operation? Probably not enough to meaningfully eat into production, but it begins to illustrate the scope of the problem.

      1. All true. Also, people doing continuous heavy labour need to eat really well. Food translates directly into production of ore. A miner probably burned through two or three times the ration of an average peasant.

      2. I’m not disagreeing, but those figures do suggest to me that manning the legion and keeping it fed would be bigger problems than supplying it with weapons. At least in the long term.

        1. The issue is, you also need to keep the men making iron weapons for the legion fed. And the folks making clothes, and so on.

          The fundamental problem is that everything you do, including the fighting itself, has to be supported by farmers who produce fairly small surplus. I discuss this problem and how it shakes out in a lot more detail in my dissertation, which I am (slowly) in the process of hopefully bookifying. But the upshot is that the metal-workers and the legion itself are competing for essentially the same scarce and very limited pool of resources.

          1. Indeed, but the metal workers would seem to be requiring a lot less from those resources of food and manpower than the legionnaires do, simply because so many less of them are required, in the long term. (Less than 1%, in my rather arbitrary example of one metalworker-day per legionnaire-year.)

            I would think the problem would be the short term. What do you do when your city or tribe needs to mobilise a lot of men very quickly? If armour lasts a long time, you do not need to make much in one year to replace wastage. But if war breaks out and you suddenly need five times as many soldiers as you did last year, then the fact that armour lasts a long time after it has been made, does not help you suddenly make a lot of it when you need it.

          2. Only coming back to this in retrospect, but I think the answer to the question “so what do you do to manufacture decent armor and weapons for a whole lot of soldiers when you’re in a hurry” is “you don’t.”

            Because with a handful of exceptions where the most centralized, developed societies in the world had a truly massive ironworks (such as Elba for the Romans), it looks like there wasn’t a way to do it, even in theory. The norm for these societies would be, then, that there IS no way to mass-produce iron to provide armor or metal-intensive weapons for thousands of men. Some degree of accelerated production is likely possible, but not a sudden acceleration to the scope that lets you crash-equip whole armies. At a minimum producing the iron required would take months, as would working it. Maybe more, depending on the scale involved.

            So you’d be restricted to the kinds of weapons that could be manufactured with little metal, or out of the limited amounts of ‘scrap’ metal that would presumably be available for sale if the price was right (e.g. broken tools that are not practical to fix, rusted nails that can, in principle, be de-oxidized in a bloomery as if they were just more high grade iron ore)

            With that, if you were trying to recruit thousands of soldiers in a hurry, soldiers who didn’t already have their own weaponry, your options would be limited. Shields are straight-up made out of wood with a minimum of metal and leather parts, so you could probably give most of your soldiers a shield. Spears use metal tips, but not a lot of metal- basically a knife blade. You could arm a lot of people with spears in a relative hurry, probably explaining their popularity.

            Iron body armor would be a non-starter. Maybe helmets- as noted, helmets are the most popular form of body armor, require relatively little material, and in theory you could even make one out of wood or something. Body textile protection would be nice if you could get it, though textile production had its own limitations and you probably couldn’t mass produce thousands of heavy quilted padded jackets in a hurry either (maybe by cannibalizing other cloth products, if there were any to be had).

            So there’s your typical hastily raised levy. No prior military training, armed probably with spears, bludgeons, or repurposed farm tools. Shields with any luck at all, textile protection, maybe a metal helmet if they’re lucky.

            This is why most pre-modern societies have a “bring your own weapons” clause to their style of warfare, and very especially a “bring your own armor and mount” rule. And why membership in a martial or warrior caste was usually predicated on owning enough land and having enough tenant farmers that you can extract revenue capable of paying and/or feeding someone to make the equipment you’d need.

  11. This for whatever reason has me thinking about that fantasy trope of Dwarves building huge fortresses underneath mountains where they mine and do a lot of metalworking. Which presents a bit of a problem; they’re gonna need a LOT of timber and fuel for all that mining and metalworking, especially if they’re like the Dwarves in Erebor from the Hobbit that made statues from a metal which looks like gold but melts like mercury or gallium (as an aside, dear lord what happened during the filming of the Hobbit for that to slip through?).

    So in a setting with Tolkien-esque Dwarves, you’d likely be able to tell which mountain has a Dwarf fortress underneath it by looking at the trees on and around the mountain; if a mountainside has been stripped completely of anything even remotely tree-like, that probably means that there’s a Dwarf city somewhere in the general area.

    1. On the other hand, it could be that Dwarves build underground dwellings precisely because they’ve figured out how to coak coal for the processing and working of iron! There’s a lot we don’t know about the Dwarves, so I count it as plausible.

      1. Perhaps, but the timber would be a necessary intermediary unless Dwarves just popped into existence with all of their technical know-how. Which, given that they often exist in settings with magic oozing out of just about everywhere, isn’t an inherently absurd idea.

        1. Gloranthan dwarves were pretty wacky, making canned food from scratch somehow. Tolkien, less so:

          > Then Manwë and Yavanna parted for that time, and Yavanna returned to Aulë; and he was in his smithy, pouring molten metal into a mould. ‘Eru is bountiful,’ she said. ‘Now let thy children beware! For there shall walk a power in the forests whose wrath they will arouse at their peril.’

          > ‘Nonetheless they will have need of wood,’ said Aulë, and he went on with his smith-work.

        2. As meticulous about his worldbuilding as JRRT was, I’d rank that as a supremely absurd idea. Sure, the Seven Fathers of the Dwarves likely learned smithwork at Aulë’s own forge, but they’d have started with bringing him timber to charcoal down for the bloomeries and forges.

        3. Certainly! It’s been a minute since I last read the Silmarillion, so I can’t remember the details of where the Dwarves first “awoke” and how long it took them to settle their under-mountain cities in Khazad-Dum and Ered Luin. Could be that part of the trade between the Dwarves of Khazad-Dum and the Elves of Eregion involved shipments of wood that the Dwarves had no means of harvesting themselves, or it could be that they clear-cut the mountains and then developed coal-coking techniques as a replacement.

          1. There are hardly any details to remember; Tolkien never said much about dwarven history. Khazad-dum was definitely settled in the pre-Noldor First Age, though. I think Ered Luin was settled after the Petty-dwarf exiles were in Beleriand; there’s one text where the Sindar hunted the Petty-dwarves like apes until they met the normal dwarves and realized “oops these are people”. Mim hates Noldor more than Sindar though, so I’m inclined to regard that as one of Tolkien’s dangling ideas, and one of the poorer ones.

            Ered Luin dwarves swelled Moria after the First Age, so it was around, but got bigger and wealthier. I think dwarves can chop their own trees, and there were a lot more of those before the Numenoreans clear-cut everything and Sauron burned anything left.

          1. “It is told that in their beginning the Dwarves were made by Aulë in the darkness of Middle-earth; for so greatly did Aulë desire the coming of the Children, to have learners to whom he could teach his lore and his crafts, that he was unwilling to await the fulfilment of the designs of Ilúvatar. And Aulë made the Dwarves even as they still are, because the forms of the Children who were to come were unclear to his mind, and because the power of Melkor was yet over the Earth; and he wished therefore that they should be strong and unyielding.”

            — Silmarillion

    2. I’ve been thinking about this too, and doing all that smelting and metalworking underground is going to be awful for the air quality. They’re going to need a lot of ventilation, and they’re going to need to protect their ventilation from enemies. I’m imagining a small fort on the mountain, above the workshops, with a big chimney blowing smoke.

    3. This is why the woodcutter is such an important job in the game Dwarf Fortress. Unless you’ve got a lot of coal, you’ll need to cut down a lot of wood.
      Which will annoy all the elves in the area which is why crossbowdwarf and deathtrap maker are also important jobs

      1. My rather limited impression of Dwarf Fortress is that the creator of the game has thought of just about every possible complication that could potentially happen with the sorts of settlements that fantasy Dwarves (Dorfs?) build, so the woodcutter being an important job in Dwarf Fortress doesn’t surprise me one bit.

      2. Until you get access to magma, that is.

        But I suspect that if the real world had even a hundredth as much access to reliable sources of molten rock as the average fantasy world does, then magma-smelting would have a whole section in this article

        1. Access to it isn’t necessarily the problem, it’s that getting near molten rock will scorch your lungs and kill you. Frodo standing over a lake of lava in Return of the King is pure fantasy.

    4. I heard that they DID know that gold was incandescent when molten, but they decided to change it because it didn’t LOOK enough like gold. Silly, I think, but sometimes filmmakers take visual shortcuts that don’t translate well into reality. Like the supremely common ‘casting a sword’ bit- which I think most production crews know isn’t correct, but which LOOKS awesome and industrial and primal.

      In a lot of fantasy, underground civilizations use wood-like fungus for their wood needs- perhaps something similar could work here- especially as fungus grows particularly quickly.

      Though it seems to me that the Erebor dwarves (in the film at least) probably managed to find and tap into natural gas reserves for their forges. Probably the best fuel source for underground, because it’s so much more clean-burning than coal, wood or charcoal.

    5. Until we come up with a good explanation for what the Dwarves eat I’m not going to worry too much about what they use for fuel.

      1. I’d imagine that the Dwarven food supply would be effectively the same as castles in real life, with every square inch of arable land outside the Dwarf’s fortress being dedicated to farming. Potatoes if they have access to them would be the perfect crop for a Dwarf hold, given that hills often sit in front of mountains.

        More fantastical ideas I’ve seen floating around include edible and more importantly farmable mushrooms; perhaps Dwarves make like Leafcutter Ants and use all the foliage from the prodigious amount of trees they’d be harvesting as compost for these edible mushrooms.

        1. If I ever run dwarves I might give them a leguminous potato-like tuber with a heat-sensitive toxin, or maybe a toxin to anything that isn’t a dwarf. Land efficient, possibly labor efficient, grows well on mountain terraces (feels like that should be dwarfy), self-fertilizing, low competition.

          I might also give them goat herds, though Tolkien said his dwarves didn’t keep animals.

          I’ve had the fungus idea; leaf-cutter or termite styles make more sense than the common handwave of “mushrooms grow underground, right?” No, you need an energy input. Which might be outright magical radiation (hello Underdark) but I want clarity (and that point you might as well have plants.)

          Or maybe dwarves get the supertubers while orcs get the mushroom farms: the latter are raiding, but often stripping the landscape down to the dirt like locusts. Feels pretty Morgothy while giving them a food base beyond simple hunting or raiding of small human settlements.

          1. In Moria, at least, Tolkien doesn’t describe gardens but he is quite specific about the dwarves using a series of mirrors and lenses to bring sunlight underground. Perhaps enough for agriculture, if only in the parts of the city closest to the surface?

        2. That’s an option but it still requires a significant departure from the standard Dwarf archetype, in that it means probably a majority of the population are spending most of their working hours outside and above-ground.

          1. Erebor in The Hobbit was stated to get by entirely on trading to obtain food for the dwarves living in the mountain. Since Khazad-dûm had very good relations with the Elves around it (in Hollin and Lothlorien), it may be that it relied heavily on trade as well for supplies, possibly also with Human nations as well . though it is so much larger than Erebor in both size and population, that i suspect they probably would have had to grow at least some of their own food. given the sheer size of the shafts they sank to light up the upper halls the Fellowship passed through, i would not be all that shocked if they couldn’t have had entire galleries with light shafts opening onto carefully tended gardens. or that they’d have fish-farming using blind cave fish varieties.

  12. Fun fact 1: the dearth of trees in England (well, it was not always so, but as time went on, the trees began to run out) coupled with relative abundance of coal is often listed as one of the reasons why the Industrial Revolution started in there.

    Fun fact 2: in modern scientific parlance, what a collier does burning all that wood is the process of pyrolysis of wood. As in, you heat up the wood so it begins to decompose into base components (like, say, elementary carbon). The actual burning is an oxidation of wood.

    Fun fact 3: when I saw that scene with casting Valyrian steel, I did not think the scene is unrealistic. Instead, I was like, “wow, so it turns out Valyrian steel actually isn’t a steel after all”. I figured it’s actually a resilient alloy, like, for lack of a better example, duraluminum, and which is only called steel because people in Westeros don’t really know better.

    1. Perhaps they figured that the translated term would mean nothing to the audience who didn’t know any better.

  13. Your legion figure is 9 kg of iron per soldier; that feels low for armor + weapons, mostly the armor.

    China had blast furnaces, are you saying just that medieval Europe lacked a way to turn cast/pig iron into wrought iron?

    1. For the legion figure, see Devereaux, “The Material and Social Costs of Roman Warfare in the Third and Second Centuries B.C.E.” (2018), 125-276, 492-500. I’m pretty confident in the figures. Note that Roman mail armor has less coverage than a high medieval mail harness.

  14. Charcoal is a zero-carbon-footprint fuel, because the carbon in it has been pulled from the air over the previous few years as the tree grew. (The same principle is used to justify all biofuel ideas, such as bioethanol, biodiesel, etc., even if the motivations are better explained in terms of messed-up US agricultural policy.) This, unfortunately, ruins some of the jokes:
    “if you let oxygen in, the wood is going to burn away all of its carbon to CO2, which will, among other things, cause you to miss your emissions target”
    “The CO2, being a gas, exits the furnace at the top, flying off into the sky to utterly ruin our climate in revenge”

    “man-days-per-cord”: As far as I can tell, the calculation assumes that a kiln consists of 1 cord of wood, but in the attached images, the piles seem to me to be several times larger than that. If the kilns are, in fact, much more than 1 cord, then the process becomes far less labor-intensive. (E.g. a single worker cutting 10 cords in 10 days, assembling, running, and disassembling the kiln in another 10 days, on average the process taking 2 days per cord.)

    Fe2O3 + 3CO -> 2Fe 3CO2. (missing + on the right side)

    “alumina reacts much the same”: not a chemist, but it should give an AlO4(5-) anion, which would make a slightly unwieldly Fe5Al2O8 product.

    At the temperature in a bloomery, limestone converts itself into quicklime (CaCO3 -> CaO + CO2). To some extent, the excess CO2 isn’t a problem; it can encounter some fresh, oxygen-starved charcoal and share (CO2 + C -> 2 CO). This latter reaction actually absorbs some heat (it is endothermic); however, it spares us from having to blow more air through the works, which would enter cold and leave hot.

    “In contrast to a blast furnace, which will give us iron with a lot of carbon, our bloom’s carbon content will be extremely low, because the carbon that we had was burned into carbon-monoxide”
    As far as I know, the contrast is not there, but in the temperature and allotropes iron. Below 910 °C, iron stays in the ferrite phase, which can dissolve only 0.02% carbon (by weight). Above 910 °C, it rearranges into austenite, which can dissolve a whopping 2% carbon by weight. However, staying below 910 °C implies that a lot of slag will stay in the bloom.

    1. “Charcoal is a zero-carbon-footprint fuel” so long as you are providing somewhere for the atmospheric carbon to get locked back up in, by, say, replanting. If the practice had been sustainable (more coppicing and replanting and less felling and clear-cutting), not only would it have been carbon neutral, we might have spent a lot longer using wood instead of moving on to coal…

      Hell, even coal is carbon-neutral when used at the right rate… It’s just that taking all that carbon that accumulated over sixty million years (the early Carboniferous before fungus and bacteria learned how to break down lignin) and putting it back in the atmosphere in under a thousand years isn’t the right timescale.

    2. In addition to George: Ho sustainable/how much emissions charcoal gives out depends on how much wood exists at a time with charcoal production vs. how much exists without production. If charcoal production leads to deforestation, it results in a release of carbon dioxide into the atmosphere (from the trees that don’t exist anymore.)

  15. Chemist here, one minor correction. Heating iron hydroxides leads to a disproportionation reaction, driving off water, not hydrogen, and leaving you with iron oxide; ie. in a simplified equation (iron hydroxides and oxides are almost always complicated mixtures) something like 2Fe(OH)3 + heat —> Fe2O3 + 3H2O.

    Very interesting post, as always!

  16. I can see why there’s such a disconnect, technologically and organizationally– city-states rather than tribal villages– between metallurgic cultures and non-metallurgic. One doesn’t (in the 21st c.) usually consider the mere possession of usable metal to be something so labor-intensive!

      1. Well, they had some metal working of the more precious metals, even copper smelting. But apparently mostly for ceremony/adornment rather than as tool metal, so the point stands. Or goes in reverse? You need a state to produce a lot of labor-intensive metal[1], but you don’t need the metal to have a state.

        [1] Exceptions for native metal, and perhaps bog iron, which I’ve read was a lot easier to extract.

    1. A lone youtuber, Primitive Technology (, has been able to generate small amounts of iron bloom using homemade charcoal and iron-rich bacteria. Granted, that’s with the benefit of modern knowledge, but it would have been technically possible at the scale of a small village to have small amounts of usable metal without a near-industrialization-scale structuring of society. (An idea for an alternate history or fantasy setting, I suppose…)

      I think in our own history, the organizational disconnect is more a factor of scale and demand; we had such a demand for it we couldn’t even manage to not denude whole countries of trees to make as much of it as possible as fast as possible…

      1. Similarly, one of my favourites, How To Make Everything, is in the middle of a years-long project to build up all the tools for his projects, all the way from hitting rocks to (eventually) a steam engine.

        He’s just recently finished iron smelting – And while it wasn’t exactly an expert smelting job, he does show the process (and the difficulties of it) pretty well.

      2. Jack Vance wrote a story set on a water world. The settlers live on giant lily pads. Their experiments lead to metallurgy using blood (copper from lobster-equivalents, iron from donations) and large lenses. See Blue World. Vance was fairly science-literate.

        1. Sure, he was fairly science literate. More importantly, he could tell a good story.

          In any case, an 80 kg human would contain about 4.8 grams of iron. ( so one could, in theory, use humans as a source for iron. To forge a fairly typical sword (perhaps 3 kg, plus wastage, maybe another three) would require something like 1200 people.

          Forget bronze weapons

  17. And then there’s metallic impurities! Though actual copper had the most impressive issues. Nickel is from Kupfernickel — the devil’s copper — because its ore does not give copper when smelted. Cobalt got its name from kobold ore — goblin ore — because smelting them gave off arsenic fumes and did not produce copper — or nickel, once they wanted that.

    1. TIL “nickel” comes from an old word for a malicious demon. So calling the Devil “Old Nick” wasn’t the 19th-century coinage I thought it was.

  18. For making charcoal was green wood used or was the wood dried first?

    I can sort of see how people figured out charcoal production. If you have a lot of green wood and you need dry wood then you need to cure the wood and it doesn’t seem like too much of a jump to go from curing wood to making charcoal but for the smelting that’s going on here some of it gives me headaches when I’m trying to think of how the hell people invented all of that.

    There are so many steps and each step in isolation seems useless (except for making charcoal, which can be used for other stuff) so how did all of that get put together? Now obviously this WAS hard for people to figure out since they didn’t figure it out all through the Bronze Age. Do we have any idea HOW it was figured out? I’m guessing that they made bronze and tin smelting more efficient and this carried over to iron? Still pretty damn impressive.

    As you note the efficiency is of course a lot lower for pre-modern techniques. I wonder how much of this efficiency gap is really simple things that a modern expert in metallurgy could improve easily without needing any additional technology vs. stuff that required more modern equipment.

    I don’t know the first thing about metallurgy but I know brewing and have read pre-modern brewing instructions and they’re a bizarre combination of incredible ingenuity and baffling inefficiency. For example, it was common among Medieval brewers to put hotter water into the mash first to “open up the grain”, and then they’d drain that off and put in cooler water and then sometimes repeat. This is completely backwards and is a good way to denature the enzymes in the mash and murder your efficiency. And then right next to that are really clever ways to gauge water temperature without a thermometer but it’s bewildering that it apparently took a long time for people to figure out that starting with cooler water and then heating it up (or even just holding the water at a more moderate temperature) would get you a lot more bang for your buck. It also seems that brewers didn’t realize that by using pale malt instead of more heavily roasted varieties they could improve efficiency by around 30% until someone started applying hydrometers to brewing. You’d think they’d have noticed 30% differences in efficiencies between different kinds of malt but apparently not…

    A modern brewer who knew a bit about malting could achieve much greater efficiencies if transported back in time without even making any additional equipment, I wonder if the same is true for smelting etc.

    1. Guy who studied chemical engineering here (who than switched fields for graduate studies, so no job experience but some knowledge.)

      Most obvious improvement might be better heat recycling/heat management.

      For the iron itself, modern steel production tends to keep the material hot all the way through, instead of cooling and than reheating (as it looks like future blog posts may describe).

      For charcoal production, doing a quick google says the energy needed to pyrolyze wood is a lot smaller than the energy released from burning wood (around several hundred times as much), or that some pyrolyzing is exothermic (wood is complex, depends on measurement conditions, “pyrolysis” may not have charcoal as an end product, etc.) It suggests a lot of charcoal wood could be processed with a small amount of fuel wood. Probably a separate mostly insulated charcoal chamber, either completely separate and heated from the outside by fuel to stay warm, or mixing a small amount of hot fuel exhaust into the charcoal chamber (this would also keep (oxygen lower), to replace heat losses to environment and pyrolysis, plus some preheating of charcoal wood (using fuel exhaust that had already heated the charcoal) if this were continuous, would be the way to go.

      The 40% wood to charcoal number in this blog post looks about right if most weight loss comes from non-carbon stuff in the wood getting baked out, with a little fuel used and other random losses, so modern charcoal production almost certainly does something like this, and non-modern has god room for improvement. If bloomeries can be built from bricks/clay/rock/etc., than lower temperature charcoal stuff should be physically buildable as well.

      For the iron production part: Since roasting and charcoaling take place at a much lower temperature, stacking some ore , maybe wood, on top of the main iron reducing chamber would effectively recycle some heat. (Both roasting and ore –> iron consume heat, but if iron must be at a high temperature to be reduced, the exhaust from that process will be hot enough to do some roasting, maybe charcoaling as well, depending on how much impurities are actually removed by roasting.) I’m assuming these processes shouldn’t be mixed (otherwise iron makers would presumably have just thrown some raw ore into a bloomery furnace instead of using 2 steps), but the hot air from the bloomery seems usable for charcoal production (it is low oxygen and hot), it may not work for roasting*, so directly using exhaust may work for charcoaling, but separate chambers needed for roasting.

      Preheating input air is also useful: use exhaust or just heat from the furnace to heat the air as it flows in should save some heat. The slag could in theory be used as well, though making sure you aren’t freezing the slag to heat input air may make this not make much sense.

      Better control of all of this may allow less charcoal to be burned to carbon dioxide (to produce heat), and more to carbon monoxide (to produce less heat and reduce the iron), although non-modern people may not have had the equipment/knowledge/ability to even in theory control the process that accurately.

      1. Should also add: Everything I’ve written assumes good heat insulation. I’m not sure how well this could be done. (Obviously thick walls insulate well for everything, but this assumes a furnace will be used over and over, o some quick construction method exists. Also might not work so well for batch processes, the thick walls being heated would themselves take a lot of energy, how much this hurts depends on how much iron is being made at a time. Possibly some brick types/clay types/rock types could be made or found that insulate better.)

        I could also see two stage burning being used: use some wood to heat air to its burning temperature, than either heat a second stream or air, or use the exhaust with lots of oxygen, to burn charcoal. This would use more overall energy than just burning charcoal (wood releases a lot of stuff that charcoal does not, this is a part of why the temperature is lower), but if charcoal production is still pretty inefficient it may save some wood overall.

      2. Naively, building a new ‘kiln’ for every batch of charcoal seems pretty wasteful of both wood and labor. I’m guessing it makes sense when transporting wood is hard and you’re in the middle of the forest with lots of wood. But having a permanent ceramic kiln you bake wood in would seem more effiicient… if you could bring the wood to sufficiently large or numerous kilns.

        Hmm, if we imagine a more sustainable ‘high tech’ or ‘elven’ society, there might be less wood to process at once anyway, since someone’s trying to avoid net deforestation.

        1. i think the main problem with transporting wood to a dedicated efficient charcoal-making facility would be feeding the horses or oxen. remember, a load of charcoal is 14% of the wood required to make it inefficiently in temporary forest structures, so if you were shipping wood (instead of charcoal) to an efficient permanent kiln either at the edge of the forest or near your foundry (worse), you’d require about 7 times the animal-food to move raw wood as compared to charcoal over the same distance, plus the food for all the extra animal handlers. and after you use the nearest trees, you’d have to transport the raw wood further…all that for a ~3x yield increase from 14% to 40%? nah, that activation barrier seems too high for people to consider the R&D and capital investment in a dedicated charcoal facility.

          if one could cut sufficient costs by using river or ocean transport for fresh wood, that might make it feasible.

          since modern societies have a highly specialized and productive food system, and no longer rely on animals for transport, we can afford this. but just THINK about all the extra squeezing of peasant farmers that would have been required in Roman times…

      3. Thanks for such a detailed response. Looks like there’s nothing specific about the response that makes modern people smack their forehead and say “ye gods, what were these people thinking!” the way the some pre-modern brewing instructions does to modern brewers (some of which is undeserved, a lot of pre-modern techniques that are strange seem to have been work-arounds for using the poorly malted grain they had due to a whole host of other problems).

        Obviously insulation matters for a lot but with that simple quantity would help a lot. Insulation doesn’t have to be fancy.

    2. “but for the smelting that’s going on here some of it gives me headaches when I’m trying to think of how the hell people invented all of that.”

      As for discovering iron smelting. Some copper deposits exist of the pure metal. Once you realize copper is useful, you’re going to mine for those pure nuggets. The rock surrounding those nuggets is often copper ore. The discovery of metal smelting could have started out with trying to find better ways to extract nuggets of pure copper from the rock they’re embedded in. Or, someone who worked at a copper mine used a chunk of ore as fireplace liner, and noticed that bits of metal that weren’t there before appeared after building a fire. Once the conceptual leap is made that you can *make* metal out of some kinds of rock by applying heat, it’s a natural development to look for *other* kinds of rock that will yield metal, learning to recognize ores as a class or rocks, and then discovering that there are some very common rocks that look like ores but which stubbornly refuse to give any metal when you heat them the same way you’ve been heating your copper ores.

      The other half of the story is that early smiths had a perpetual problem of needing tons and tons of wood to process copper ore. They were always eager to discover ways to need less wood, and ways to somehow get more metal out of a given amount of ore. Learning how to make hotter fires is a natural consequence of those two quests. And with every new innovation in copper smelting, someone inevitably tried it out again on those enticingly common but stubborn ores. Once they made a fire that was just barely hot enough in a few places in the furnace, they would have noticed a tiny bit of iron starting to appear. From there, it was a (long) process of trial and error to discover ways (introduce more air, use charcoal) to make the smelter hot enough to produce a usable amount of iron.

      1. The best explanation I’ve seen for the origin of metallurgy is it developed out of glazing pottery, copper ores turn interesting colors and can be smelted at temperatures achievable by pottery kilns

  19. I am absolutely loving the parallel between the way the charcoal-burners monitored their process, and the way the College of Cardinals tells us they’ve selected a new Pope.

  20. Assuming that all the figures are derived experimentally, then I think Healy would be more accurate for the final charcoal : ore: iron figures – a bigger furnace would use less fuel because of the surface area : volume ratio.

    Of course it would also be more difficult to control, for much the same reason.

  21. I suspect that what wood was used for charcoal was something people thought about more when the charcoal was going into gunpowder, where the precise combustion characteristics became more important- willow charcoal IIRC made slower-burning powder for artillery, while the faster-burning powder used in handguns used charcoal made from hazel twigs.

    Does anyone know how the transport worked in this case? Did they only use wood cut close to the gunpowder mill, was wood brought there to be made into charcoal, or did they transport charcoal?

    1. I’m not especially knowledgeable about the gunpowder used in firearms, but I get the impression that it was not generally of particularly fine quality, aside from the primer (which has to flash, and powder that flashes has to be reasonably fine).

      Certainly accounts of historical production or of the limitations of logistics tend to concentrate on producing the saltpetre in enough quantity, rather than any limitations on the charcoal or sulphur – if they needed particularly high-grades of charcoal, or were limited to particular types of tree, then I’d expect to have run across that concern.

      Modern black powders are more often used in fireworks than firearms, and a variety of different woods are used to get different effects (denser charcoal will burn slower but will have more energy per volume). I used to be acquainted with some people with expertise in that area, and they would sometimes do their own miniature charcoal burns (with woodchips) but mostly bought in the charcoals. I know they even used animal charcoals (pyrolising flesh – non-rotted but unfit for human consumption) for certain purposes. As an aside, these animal charcoals are different from the substance generally known as “animal charcoal”, which is “bone char”, i.e. charcoal made from bones and has a variety of calcium and phosphorus compounds in it making it unsuitable for burning – it’s used for purifying water, for whitening sugar and as a black pigment in artist’s painting.

  22. Didn’t see it in the comments so wanted to drop a note from my few experiences charcoaling agricultural byproducts (corn sheaths & stalks etc) for household cooking/heating: after black smoke follows yellow smoke (which I’ve heard is sulfurous), and sometimes blue smoke. Above a certain temperature the smokes are joined by (clear) biogas, which you can ignite to increase efficiency. To ensure a clean-burning charcoal (at the cost of some lost energy) the oil-barrel “kilns” I used weren’t covered until smokeless, with just waves of heat coming out.

    The process I followed is here:, though it doesn’t mention the smoke colors.

  23. It isn’t clear whether Prof. Devereux is talking about both ancient and medieval merchants, or only the former, but isn’t it somewhat exaggerated to say that merchants never made it into the ruling class? It might take a few generations, as it did for the de la Poles, but couldn’t a wealthy merchant who bought a manor start his family up the ladder? And weren’t merchants effectively the ruling class in many towns, though their concerns and their power might be almost exclusively local?

    1. I think Prof Devereux is speaking in generalities, and he is referring to both ancient and medieval societies, where people did not as clearly understand that moving things actually produces value. If you look at the article he linked where he introduced this idea, he mentions that societies which heavily depended on trade and where merchants held positions of power (he suggested medieval Italian City States…and I would add probably the same for the Hanseatic league) had a better view of them then more agrarian societies. de la Pole is a very interesting example though given that was a more agrarian society and I do not know much about the rise of Sir William–perhaps it was a hurdle he had to overcome though? I don’t believe the later de la Pole’s continued to work in business once they had become landholders.

  24. On the eternal Tolkien digression: I don’t think dwarves could magic up fuel but they might have had better smelting techniques than discussed here. Blast furnaces or weirder stuff. There’s also the Noldor, who also had lots of metal, though even more divine instruction. Unlike farmers, elven miners are actually mentioned, at least with Gondolin (Maeglin’s talent and mistake), though where Gondolin got its wood (or food) is unclear… forested mountains, maybe? The valley wasn’t that big, I think Tuor spiraled to the center in half a day. OTOH they did have 300-400 years to clad their 10,000 warriors, and not from scratch — Noldor probably brought arms from Valinor, plus pre-Gondolin residence in Nevrast.

    Numenoreans had lots of people and metal but they canonically did lead to mass deforestation.

    1. I’ve been thinking that people who enjoy these military and mining posts might like Kit Sun Cheah’s Dungeon Samurai trilogy — Kamikaze, Kami no Kishi, and Seisen.

  25. A question about sulfur: would it ruin bronze? I’m currently reading the Book of Daniel, chapter 3 (the famous “fiery furnace”)—which already includes a very large furnace— but in verse 46 (Syriac and Greek versions are longer than the Herbrew–Aramaic), it mentions that the tenders of the furnace were shovelling in “naphtha and pitch and straw and twigs”, and the Syriac Peshitta adds sulfur to this list (it could also be lye—the word is not exact). Is this just someone who doesn’t understand metallurgy adding in something else that burns?

    1. On further reflection, it could be referring to a flux compound like potash (“fuller’s soap” is within the word’s semantic range)—it’s just rarely translated that way.

  26. Reblogged this on Head Noises and commented:

    Folks don’t even think twice about blacksmiths and coal– coal is an incredibly more dense energy source than trees or the stuff you can make for them.

    This series is enjoyable for bringing up things folks don’t think about.

    1. Hindsight is 20/20. Coal has some huge strikes against it. 1, You have to mine it out of the ground. 2, it doesn’t regrow. 3, Most coal burns very dirtily in its natural form. The mining part was the biggest block to using it as a serious energy source, until people finished cutting down all the forests near where they wanted to build industries and had no other choice. Then, and only then, did they start working with coal in a serious way, and realize that properly treated coal burns hotter than wood or charcoal and thus works better for making steel.

      1. In fantasy novels (which my audience would mostly be focusing on 😀 ) it has even MORE strikes against it, because people very rarely bother to have anybody mining except for Dwarves and maybe evil guys.

        So you’d have blacksmiths that make their own iron (to show your story guy is special), other than wanting to buy some of the really high quality Dwarf-steel, on a grassy plain with no coal.

        Which– again for the folks who’d be reading my blog– is going to knock a larger than average number of folks out of their suspension of disbelief exactly because they’ve got at least a hint of an idea that it takes more.

  27. “The charcoal burner (often called collier, before that term meant ‘coal miner’ it meant ‘charcoal burner’)”

    And, unlike Treefeller, Collier _is_ a fairly common surname in modern English-speaking cultures. Presumably also Fr. “Charbonnier”?

    1. Whoops, mixed in some apocryphal sources there. But at any rate, Moria had access to sunlight, which presumably was used for growing things as well as for general illumination.

  28. One extra detail — I’m pretty sure the ropes pulling the trees down weren’t just to change direction. It takes a lot less skill and effort to cut a tree part of the way through and then pull on it hard enough to break the last bit. Especially if you’re sawing or chopping the thing by hand. At least, that’s been my experience with some very amateur manual tree removal.

  29. As if future historians would read this (awesome) article, they may get the idea that charcoal “briquettes” are of common use and that “raw” charcoal is rare nowadays. I’d like to stress out this may highly depends on the local culture. As for France, it’s the opposite: raw charcoal is a common resource in supermarkets, whether briquettes are quite rare. Not sure what to conclude, but I found it interesting there’s still such culture gaps, in our globalization times.

    1. Both are readily available in German super markets. My parents prefer the raw kind (as they are vary of the additives in something they prepare food over), most of my friends prefer the brikettes as they last longer.

      1. Similar story in Britain where both are available, though I think the raw charcoal has gone through some sort of sieving process to remove the smaller pieces and most of the dust. Certainly when I’ve had raw charcoal for cooking over, it contains mostly pieces of roughly the size of the briquettes (though a far greater variety of shape).

        1. Raw charcoal is also available in other parts of the United States. I guess that explains why Carolinans need to put vinegar on their BBQ; it’s to cover up the inferior fuel they’re reduced to using.

          1. Carolinians make barbecue sauce correctly. Other people use some kind of sugary ketchup.

          2. One of the Texas BBQs doesn’t use sauce at all. You rely on the flavor imparted by the smoke and a few spices. Saucing the meat at all defeats the purpose of the smoke. You might as well just cook it on your stovetop.

    2. So, the story I heard (probably some documentary) is that Kingsford briquettes are due to a Mr Henry Ford trying to find a use for sawdust and scrap wood leftover from making wooden parts of the car. Which would imply that they started as an American thing, centered on Detroit. I don’t know enough logistics to say when it made economic sense to ship briquettes overseas, vs cost of making local charcoal. (I also don’t know when and hows of briquette factories becoming widespread).

      From grilling to smelting, I would want to know more about using propane/natural gas/coal gas. That can get iron at least hot enough for forging, and if there’s an efficient method to make the gas, that could be a lesson from Aule.

  30. Fun Fact: China began to use coke in iron-working as early as the 11th century because……they ran out of woods.

    Almost all the capitals of Chinese Empire from Early Imperial (c. 2nd century BCE) to Early Modern period (c. 10-12th century CE) were in the Northern China (Chang’an, Luoyang, Kaifeng). These capitals usually had a population count larger than 1 million, therefore during the winter they would need a lot of charcoals as domestic fuel. These capitals also had many monumental palaces, which require a lot of woods as well. Moreover, the center of iron-working of medieval and early modern China, Hedong (河東), was also in Northern China (modern Shanxi province) and not very far away from the capital; the iron industry need a lot of charcoal as well.

    The above three “great wood consumers” easily deforested a large portion of Northern China around 10-11th century. Many literati and officials reported that a lot of mountains in northern China, esp. in Shanxi and Shandong, literally became “bald” as they ran out of trees; some officials even reported that peasants began to felling mulberry tree for charcoal which impacted local silk industry. These officials and literati began to suggest that it would be better to use “stone charcoals” (石炭) – coal – as fuel in order to deal with the deforestation.

    These suggestions were put into practice: Song Huiyao (宋會要, the book of institutional changes of Song Dynasty) recorded that since Emperor Shenzong of Song (reign 1067-1085), there were over 30 state-controlled coal and coke markets (called 石炭場, literally “stone-charcoal yard”) around Kaifeng, the capital city. All the selling revenues belong to the Song state, and sometimes officials would try to control the selling price when there was a fuel crisis.

  31. Hi Bret, I have two questions:

    1/ I am looking at your “trademark badly made diagram”, but I have seen this with all other furnace diagrams: what is holding the ore and bloom in the air? What prevents it from falling down into the shaft?

    2/ You talk a lot about tapping off, or tapping out the slag. I am not a native speaker and I don’t understand what is meant by that. According to several dictionaries, the phrase is used in baseball and no other meaning is mentioned. Could you elaborate, please?

    Thanks for the fascinating reading!

    1. Hi, Daniel! You don’t have to be a non-native speaker to question the meaning of “tapping off” or “tapping out” the slag! “tapping” in this context is what is called a “term of art,” and I only know that because I went to Wikipedia to read more about “slag.”

    2. The interior of the bloomery is filled with both charcoal and ore. In modern experimentation it is typical practice to fill the stack with an initial load of charcoal and then charge alternating layers of charcoal and ore as the stack falls. The stack falls as the charcoal is burnt and the ore is processed into slag and bloom.

  32. A late thought on this, prompted by listening to Ep. 4 “The Ring Goes South”, in the 1981 BBC radio adaptation of Lord of the Rings – given the increased scarcity and the possibility that it required more technical working, is it feasible that Frodo’s mithril-coat was actually more valuable than eveything in a prosperous agricultural region like the Shire?

    1. I am not expert enough to have a reliable answer, but it doesn’t sound as if you’re considering more than the intrinsic value of the mithril and jewels the mithril-coat was made of. In addition, there is the extrinsic value of its age and its provenance (and other factors?). Gandalf has assigned a market value overall of what it would be worth to a “billionaire” collector, perhaps.

    2. Without a doubt, especially when you consider that mithril is no longer being made/mined in Middle-Earth in that time.

      1. That doesn’t make it “without a doubt” at all. The question isn’t whether things are being made, but how valuable they are to someone. Or how valuable they are to a reasonable someone.

        If you were king of Arnor, would you trade the province of the Shire, and all the food and soldiers it could produce, for a coat of mithril? (Ignoring that it would have to be re-worked with yet more mithril to fit a human king.) Maybe, rich people do odd things. But would acquiring such a coat be more militarily valuable than holding the Shire? That seems very unlikely. Be more economically productive? Even more unlikely.

  33. I think you have the iron chemistry slightly wrong? The iron reduction cascade normally runs:

    3Fe2O3 + CO -> 2Fe3O4 + CO2 (hematite -> magnetite)
    Fe3O4 + CO -> 3 FeO + CO2 (magnetite -> iron oxide)
    FeO + CO -> Fe + CO2 (iron oxide -> iron)

    So roasting magnetite to produce hematite doesn’t seem to be a useful thing to do?
    (It just has to be reduced back during the blooming process anyway).

    This isn’t to say that roasting isn’t useful! Driving off water and oxidizing sulphur et al are totally useful, just not the magnetite oxidation.

  34. “With just one exception, however, we cannot jump directly to using that wood to process our ore”

    I gotta ask: which one is the single exception?

    1. Roasting — our dear host covers this.

      “The actual process of roasting the ore is in many ways less interesting than the chemistry that goes into why we roast the ores. This is the only step generally done with raw wood (typically dried to remove water), rather than charcoal.”

Leave a Reply