25 July 2005
Submitted to Late Imperial China
It has often been remarked that bureaucracy favours the historian, for it produces great quantities of paper which later become historical sources. The Ming state ironworks in Zunhua, Hebei, provides a good example: its administration, especially the management of a large force of corvée, conscript, and convict labour, gave the officials of the Ministry of Works (Gong bu) enough trouble that a good deal of written communication was required, and some of this has survived to tell us about both the technology and the organisation of the works. We therefore know much more about this one ironworks than about any – or all – of the thousands of private ironworks of the Ming period.
We have in particular a description of the blast furnaces in which the iron was smelted and some quantitative information about inputs of ore, fuel, and labour and outputs of cast iron, wrought iron, and steel. Several authors have studied the details of the administration of the Zunhua works, especially the organisation of the labour force. Here I shall concentrate on the more technical aspects of its operation, but also attempt to place it in its historical context.
Zunhua District (modern Zunhua County, Hebei) lies about 150 km east of Beijing, very close to the Ming Great Wall, and was thus a militarily important peripheral region. Numerous garrisons were stationed here and in neighbouring districts, and these provided a significant part of the labour used by the ironworks.
In the early years of the Ming, state involvement in the iron industry was a contentious issue. Starting in 1364, before the final consolidation of Ming power, state ironworks were established piecemeal in various places throughout the Empire; in 1374 there were thirteen (none anywhere near Zunhua). In 1382 the Emperor ordered an official flogged because he proposed the establishment of one more, and in 1385 all the state ironworks were closed, ‘because they weary the people’.
In the following years many of these state ironworks were reopened, but then in 1395 all were again closed. The order was repeated in 1397, presumably because it had not been obeyed; in 1398 it was ordered that the state ironworks should be opened again for one year, then closed.
The political background of all this opening and closing of state ironworks is not easy to see in the sources, but presumably the fundamental factor was conflict between local and national interests. The central government needed a dependable production of iron to supply the armies with weapons, but the ironworks demanded corvée labour to an extent which conflicted with the interests of the local gentry. No doubt there were also many places where the state ironworks competed directly with local ironmasters.
After 1398 the only major ironworks established and administered by the central government was the one in Zunhua. The reason for the special status of Zunhua seems to have been a need for high-quality wrought iron and steel, smelted and fined with charcoal, conveniently near the capital. By the Song period population growth in north China had led to increased pressure on the land, increased demand for iron, and forest destruction, so that it was necessary to substitute mineral coal for charcoal in iron production. (In the south, charcoal iron production continued.) Iron produced using mineral fuel would have had a high sulphur content: this does not matter much in foundrywork, and there is indeed some evidence that in the Ming, in north China, agricultural implements were generally cast rather than wrought. But fining this iron to produce wrought iron would have resulted in an inferior, ‘red-short’, product.
Peripheral regions like Zunhua were undoubtedly the last places in north China where significant forests remained and charcoal iron could be produced economically. Further north, in the province of Liaodong, 25 small ironworks were established in the Ming garrisons (wei) in 1411. Half of the 120 soldiers assigned to the ironworks ‘fined iron’ (chao tie) while the other half engaged in agriculture to support them. Presumably the soldiers also produced the cast iron to be fined, using small blast furnaces, but it was fining which would have required the most labour. The annual production quotas of the individual ironworks varied from 4 to 19 tonnes, with an average of 10 tonnes. The total of about 250 tonnes per year appears to have been the production of 1500 soldiers (half of a total of 120 ę 25); the labour required was thus about 6 man-years per tonne. In the Dabieshan region, production of wrought iron from ironsand and charcoal by traditional methods required only 2–4 man-months per tonne: more labour-efficient by a factor of more than twenty.
In Zunhua and neighbouring districts a number of garrisons were established in the early Ming. That these garrisons, like those in Liaodong, engaged in wrought-iron production is made clear by a memorial of 1403 which mentions convict labourers ‘fining iron’ in Zunhua. This production seems to have become more important as time went on, and in 1426 a major decision was made:
[An official memorialised:] For the production of weapons there is a shortage of wrought iron, and we request permission to purchase it at various places in the Jiangnan region. However, because of the distance involved, we are afraid that it will not be delivered on time, and we propose to send people to the ironworks at Zunhua to fetch 20,000 jin [12 tonnes] to meet immediate needs.
The Emperor replied: If iron is available in Zunhua, why purchase it in Jiangnan? Iron is heavy and clumsy, and transporting it would be a burden to the people. . . . It will be sufficient to obtain it in Zunhua.
As we shall see directly below, this was the beginning of a process which resulted in the creation of an entirely new organisation for the production of iron for the State, including a large walled ironworks in the mountains in the southeast corner of Zunhua District.
The mention of Jiangnan (the Yangtze Delta: southern Jiangsu, southern Anhui, and Zhejiang) as a source of iron is a surprise, for this region is usually said to be very poor in iron, importing most of its iron from Fujian and other more southerly provinces. However, other sources indicate that this was one of the major Chinese iron-production regions in the Yuan period, and that it had some importance in the Ming as well.
Zhejiang in the 19th and 20th centuries had a traditional small-scale iron industry based on charcoal and ironsand, and this technology could well go back to the early Ming or before. At Zunhua, as we shall see, the same fuel and ore were used. We are accustomed to thinking of iron from Guangdong as being the best in China, but it may very well be that there were advantages – real or imagined – in the use of ironsand iron, produced in small furnaces, for weapons production. It is quite possible that the garrison ironworks established in the border regions near Beijing, including Zunhua, used technology imported from Jiangnan.
Three months later a military commander in Jizhou subprefecture (which included Zunhua) noted a ‘recent’ order to reopen the ironworks at Zunhua and resume iron-fining, ‘using military and corvée labour as before’, and reported that the order was giving difficulties. The ‘thousand’ soldiers from four nearby garrisons who had worked at the ironworks had long since been sent home, and the ‘thousand’ peasants from six districts in Yongping Prefecture (east of Zunhua) had been assigned other corvée duties. Further, this was a busy agricultural season, and the peasants should not be disturbed in their work at this time. He therefore requested permission to use soldiers from a long list of military units, some quite far away, and only after the harvest was completed go back to the system of using peasants and soldiers from border districts. This is our only information about the earlier arrangements at Zunhua: if we trust the ‘thousands’ of soldiers and corvée labourers to indicate the order of magnitude of the labour force, it was already a major industrial complex in its earlier period of operation.
Labour problems continued, and in 1432 a military officer at Zunhua reported that the available workers were not familiar with fining; he was given permission to bring in metallurgical experts (an lian zhi ren) from elsewhere. In 1435 the newly enthroned Emperor Ying-zong ordered the Zunhua ironworks closed, but the next year it was reopened. Another source has this reopening in 1438; probably a major reorganisation and relocation occurred in this three-year period, and from this time the ironworks operated continuously for almost 150 years. In 1507 a long memorial concerned with various administrative reforms at the Zunhua ironworks, by an otherwise unknown official, Han Dazhang, included a good deal of archival information on the works as it had operated in 1438. Its early history, and its ecological consequences, are summarised as follows:
Investigation indicates that the Zunhua ironworks was established at Shapogu in the Yongle period [1403–24]. It was later moved to Songzhagu, and in the Zhengtong period [1436–49] it was moved to its present location at Baiyezhuang [‘White Smelter Village’]. At that time the forests were flourishing, and it was not difficult to supply firewood and charcoal. Now, over a hundred years after its establishment, the trees have all been felled and the prices of wood and charcoal are high. If nothing is done to restrict [forest use], within ten years the price will have risen by several-fold.
(Later, in the Qing period, forestry was forbidden in this region because of the proximity of the Imperial tombs, and in 1917 the forests were thriving; but by 1928 deforestation was again a problem here.)
One aspect of the reorganisation of 1438 seems to have been the building of a wall around the new works at Baiyezhuang. The early sources do not mention this wall, but the site is today the village of Tiechang (‘Ironworks’), which according to a modern local gazetteer has a stone wall 10 metres high, 5 metres thick, and 2000 metres in circumference. This massively walled village deep in the mountains is an extreme oddity, and it had military significance long after the ironworks was closed. It was a factor in 1638 in the Qing conquest of the region, and in 1938 and 1942 in the Japanese invasion.
Still another aspect was a reduction of the labour force by half, to 683 corvée labourers and 465 soldiers. These were in addition to convict labourers and skilled craftsmen. There is more information about these labourers, their organisation, their tasks, and their productivity, but before dealing with this material we must look at their technology.
Iron production at Zunhua was described in some technical detail in a book by one of its directors. Fu Jun was a native of Nan’an in Fujian. He received his jin shi degree in 1499, and in 1513 was placed in charge of the Zunhua ironworks. While in office he wrote a book entitled Tie ye zhi, ‘Treatise on the ironworks’, seemingly a kind of local gazetteer. The book is no longer extant, but it is described briefly in the Si ku bibliography of 1798, among books which were considered politically acceptable but not important enough to be included in the Si ku quan shu:
Tie ye zhi, in two juan, by Fu Jun in the Ming period. . . . It includes 23 topics, from ‘Forestry’ [shan chan] to ‘Miscellaneous tasks’ [za zhi], and begins with two illustrations, showing the administrative offices and the ironworks. It records the yearly input and output quantities in great detail, but is useless for evidential research [kao zheng]. . . . 
It seems that we have quotations from Fu Jun’s lost book in two books of the 17th century, by Zhu Guozhen in 1622 and by Sun Chengze between 1654 and 1675. Both describe the blast furnaces at the Zunhua ironworks in the same words, and it appears to be certain that this description was originally in Tie ye zhi.
The Zunhua iron furnace has a depth of 1 zhang, 2 chi [3.7 m]; a breadth of 2 chi, 5 cun 寸 [78 cm] in front and 1 chi, 7 cun [84 cm] in back; to the left and right 1 chi, 6 cun [50 cm]. In front there is a clearing of several zhang [a zhang is 3.1 m], the place for tapping the iron.
Black sand is the basis [ben] [the ore] and ‘stones’ [shi zi] are an auxiliary [zuo] [the flux]. Hour by hour they move downward [xuan xia]. A charcoal fire is used, and two bellows are installed to fan it. Iron can be obtained [i.e. tapped] four times in one day.
The stones are a marvel. They are produced [quarried?] at Shuimenkou. Their colour is between red and white, rather like that of a peach-blossom. The largest are like a hu [grain measure, ca. 50 liters], the smallest like a fist. They are pounded and broken and thrown into the fire; then they transform and become liquid. If the ‘stone centre’ [shi xin, apparently the furnace burden] becomes dry, and the sand [ore] cannot descend, it is remedied with this [the flux]. Then the sand begins to melt and turn to iron.
If not, then the centre [xin, lit., ‘heart’] is ill and does not melt [xiao, can also mean ‘disperse’]. It is like a man whose heart is fiery [xin huo da sheng]. He takes a prescription [liang ji] to remedy it, his spleen and stomach are harmonised, and food and drink advance [are digested]. Such are the wonders of the Transformations [zao hua, i.e. the workings of Nature].
. . .
The ‘refining’ [lian] of pig iron [sheng tie] is completed in three double-hours [shi]. Wrought iron [shu tie] is produced from pig iron by five or six ‘refinings’, and steel [gang tie] is produced from wrought iron by nine ‘refinings’.
The furnace begins feebly and then flourishes; after flourishing it declines. At the most it reaches 90 days, and then it is ruined.
This furnace is 3.7 m high, with an elliptical shaft only ca. 50 ę 80 cm in section, and is thus extremely narrow in relation to its height. It is quite different from other traditional Chinese blast furnaces used for smelting ironsand. That ironsand was the ore seems certain from the description, and this is confirmed by incidental mentions in Han Dazhang’s memorial and by a mention of ironsand production at Zunhua by Song Yingxing in Tian gong kai wu.
Fu Jun, or whoever wrote the description, was greatly impressed by the action of the ‘stones’ as a flux in the blast furnace. Ironsand is normally produced by sluicing river sand. The product of the sluicing is an ore with 75–95 per cent iron oxides, the rest being largely quartz. A flux is not absolutely necessary in the smelting of such a rich ore, but it can greatly improve efficiency. Yang Kuan identifies the ‘stones’ as fluorspar (CaF2, also called fluorite), which is commonly used in modern steelmaking. It has a low melting point, 1386°C, and can depress the melting point of a silicate slag below 1200°C. This identification is certainly possible from the description, but needs confirmation from slag analyses.
. . .
Archaeology seems to be at odds with Fu Jun’s description. In 1991 the remains of eighteen blast furnaces were discovered at a site east of the village of Tiechang. The only publication of this discovery is a very short journalistic article, which states that the furnaces are built into hillsides, and are 1–2 metres high. But Fu Jun states that the furnace is built entirely of stone, so these seem to be of quite a different type.
We have some input and output figures in two sources, the ‘Collected statutes of the Ming dynasty’ of 1587 and the memorial of 1507 by Han Dazhang, already mentioned. Both are primarily concerned with administrative matters, and give a circumstantial (and highly confusing) account of changes through the long period in which the works functioned. However, both give concrete figures for the year 1438; these differ enough to indicate that they are independent sources, but are enough alike to be mutually confirming. The ‘Collected statutes’ give the following data for 1438:
Specialist artisans in Zunhua
Charcoal production, 70 households
producing 143,070 jin (86 tonnes)
Ironsand production, 63 households
producing 447.3 shi (= hectolitres)
Iron production, 60 households
Labourers from nearby places
Corvée, 683 persons, working 6 months per year
Conscripts, 462 persons, working 6 months per year
Craftsmen from more distant places, on rotating duty, 6 months every 4 years, 630 persons
Convict labourers, number not given, who fined iron
Han Dazhang indicates that the labourers, working presumably under the supervision of the specialist households, performed such tasks as cutting wood, digging ironsand, working the furnace bellows, transporting materials, transporting iron to the capital, and maintaining buildings and walls. Some of the soldiers also did guard duty. He gives a good deal of information on the rations of grain and cloth which the state supplied to the specialist households, the labourers, and the convicts respectively.
Annual deliveries of iron to the state from Zunhua varied in the 15th century between 300,000 and 400,000 jin, i.e. 180–240 tonnes. Using the labour figures above and making some casual assumptions about their interpretation suggests that labour expenditure may have been 4–5 man-years per tonne, i.e. of the same order of magnitude as the Liaodong military production noted earlier, perhaps slightly better.
(Here it is necessary to note that the production figures for charcoal and ironsand in 1438 are not sufficient for the production of 180–240 tonnes of iron. Production of 180 tonnes of iron would require at least 500 tonnes of charcoal and 750 hectolitres of ironsand. Probably the figures for 1438 are atypical, for the new ironworks’ first year of production; but another possibility is that they represent only part of the charcoal and ironsand inputs. Perhaps the production of the labourers was registered separately from that of the specialist households.)
Production increased for a time after this, and the labour force was reduced. Eberstein collects from a variety of sources figures for delivery of iron products to the Capital from Zunhua. These, converted from jin to tonnes, are:
15th century 180–240 t
from 1483 180 t
1507 348 t cast iron, wrought iron, and steel
1509 292 t cast iron
125 t wrought iron
37 t steel
1529 152 t cast iron
125 t wrought iron
steel production discontinued
1581 125 t wrought iron
And deliveries of iron are not the whole story, for there was a considerable overproduction, and in 1507 the Zunhua ironworks had a stockpile of no less than 2100 tonnes.
Meanwhile the input of labour was falling. The labour force in 1438, discussed above, had just been halved. In 1504 the corvée and conscript-labour quotas were further reduced by 60 per cent. Detailed calculation of labour productivity is even more risky here than for the earlier case, but in 1509 a tonne of iron cost perhaps somewhere around 2 man-years per tonne. The improvement from 4–5 man-years is no doubt largely attributable to ‘learning by doing’, that old stand-by explanation of economists, along with the organisational improvements with which our sources are primarily concerned. But it would seem that technical improvements were also taking place at this time.
. . .
That there may also have been technical advances is suggested by a passage in the ‘Collected statutes of the Ming dynasty’ of 1587. Perhaps this too is quoted from Fu Jun’s Tie ye zhi. The passage occurs in a chronological survey of the history of the Zunhua ironworks:
In the 4th year of Zhengde , ten da jian furnaces were in operation, producing 486,000 jin [292 tonnes] of cast iron. Twenty bai zuo furnaces produced 208,000 jin [125 t] of wrought iron and 12,000 jin [7.2 t] of steel. [In all 424 t of ferrous products.]
In the 6th year  there were five da jian furnaces and eight bai zuo furnaces in operation, producing cast iron, wrought iron, and steel as before.
From the 8th year of Jiajing , each year three da jian furnaces produced 188,800 jin [113 t] of cast-iron plates and 64,000 jin [38 t] of ‘broken iron’ [sui tie, granulated cast iron?]. In bai zuo furnaces were produced 208,000 jin [168 t] of wrought iron in bundles [shu gua ti], which was delivered to the Capital. Steel [production] was discontinued. One bundle of wrought iron is four ‘pieces’ [kuai], weighing 20 jin [12 kg], for a total of 10,400 bundles.
The context makes it clear that a da jian furnace is a blast furnace, producing cast iron, and a bai zuo furnace is some sort of finery in which both wrought iron and steel are produced from cast iron. Neither term makes obvious technical sense in itself, and neither seems to be found elsewhere.
The passage indicates that some major technical improvements took place around this time, leading to greatly increased output per furnace. This would have saved labour in the construction and operation of the furnaces, and quite possibly have saved fuel as well. There seems to be no way of knowing whether technical improvements were made at the same time in charcoal and ironsand production.
Regardless of whatever productivity improvements were attained, iron from Zunhua remained very costly, as a memorial of 1581 reveals:
The Zunhua ironworks delivers an annual quota of 208,000 jin of [wrought] iron [125 t]. The [market] price of this is not more than 2,700 taels of silver, but the expenditures involved in providing for officials, officers, soldiers, and corvée labourers is over 10,000 taels. We recommend that the works be closed and replaced by a tax in silver, with which iron can be bought for use.
The indication of 10,000 taels as the annual cost of production at Zunhua is probably no more than an estimate, for it is unlikely that accounts were kept in such a way that this information was easily extracted. The market price of 2,700 taels for 125 tonnes of wrought iron, on the other hand, was undoubtedly the result of an investigation. It amounts to somewhat less than a gramme of silver for a kilogramme of wrought iron.
The proposal to close the works was approved, and after this we hear no more about iron production in Zunhua until 1623, when the now-failing Ming dynasty attempted to re-open it, seemingly without great success.
Let us now attempt a connected history of the Zunhua ironworks, drawing on the evidence presented above and filling in the gaps in our knowledge with reasonably well-informed speculation.
In the early decades of the Ming dynasty, political controversy over state-run ironworks was finally settled with a firm decision to leave iron production in private hands. The single exception at Zunhua was a response to a technical need: weapons for the armies required high-quality wrought iron, smelted and fined using charcoal as the fuel. It is likely that most iron production in north China now used mineral fuel, and the iron produced was best suited to foundry rather than smithy work. Buying charcoal iron on the open market meant importing it from south of the Yangtze, and the undeveloped market of the time meant that the government itself had to arrange transport to the capital. In 1426 the Emperor declared that this would be ‘a burden to the people’, i.e. politically dangerous, and therefore arrangements were made for state production of high-quality charcoal iron nearer the capital. Presumably it was known that this iron would be more costly than iron from the south, but the use of corvée labourers for transport would take them far outside their home districts, and this had been known to be dangerous ever since the fall of Qin, sixteen centuries earlier.
There had earlier been military iron production in the region, but we know very little about this. Apparently iron was produced and delivered to the capital by small works spread out in the garrisons of Zunhua and neighbouring districts. The garrison works in Zunhua seems to have been closed at some time after 1403 and reopened in 1426; various administrative problems led in time to a complete reorganisation and the establishment in 1435–8 of a single large works in a walled compound at the village today named Ironworks, Tiechang.
Several sources make it clear that at this new large-scale works ironsand was the ore and charcoal was the fuel. On the blast furnaces used for smelting, the evidence allows more than one interpretation, but the most obvious is that a major change in design was made around 1509. The type of blast furnace which was dug into a hillside was used widely in China from the Song through the Ming; in the 12th century as far north as Heilongjiang. The archaeological evidence suggests that the same type was used at Zunhua. The furnace described by Fu Jun, sometime after 1513, is very different, however, so it is likely that he describes the da jian furnace, a new type introduced around 1509. The efficiency of these furnaces improved with experience: roughly the same quota was produced by ten furnaces in 1509, five in 1511, and three in 1529. On the technology used for fining cast iron to wrought iron we seem, unfortunately, to have no direct evidence at all.
In spite of efficiency improvements, wrought iron produced at Zunhua remained much more costly than what could be purchased on the market. This had probably been known from the outset: the disadvantage of importing it from the south was political rather than economic. But in the course of the dynasty the economy stabilised and long-distance trade became more developed. In the late 16th century it is likely that merchants were bringing high-quality wrought iron from Jiangnan and Guangdong to the Capital region and selling it at prices that made the enterprise at Zunhua absurd. When this was finally realised, the works were closed and the Imperial arsenals began obtaining their raw materials by purchase on the open market.
The troubles at the end of the Ming disrupted trade with the south and led to the brief attempt to re-open production at Zunhua in 1623. From the beginning of the Qing onward there appears to be no evidence at all for iron production anywhere near Beijing; trade was by this time so well-developed that iron from better-favoured regions was in sure supply and cheaper than what could be produced locally.
Ben chao fen sheng ren wu kao (Biographies of Ming-dynasty persons, arranged by province). Comp. by Guo Tingxun (ca. 1570–1625); printed 1622. Facs. repr. with the title, Ben (Ming) chao . . . , 30 vols., Taibei: Chengwen Chubanshe, 1971.
Chang Lijun 1991. ‘Zunhua faxian shiba chu gu liantie lu yizhi’ (Eighteen iron-smelting furnace sites discovered in Zunhua County, Henan). Zhongguo wenwu bao, 28 July 1991, 1.
Chun ming meng yu lu (A work on Beijing and its environs as it was in the late Ming). By Sun Chengze (1593–1675); Written after 1654. Si ku quan shu zhen ben chu ji ed.
Da Ming hui dian (Collected statutes of the Ming dynasty). Comp. by Shen Shixing (1535–1614); Rev. ed. suppl. to 1585, completed 1587. Facs. repr. of the Palace ed., Taibei 1963.
Da Ming yi tong zhi (Comprehensive gazetteer of the Ming empire). Comp. under the direction of Li Xian (1408–66). Facs. repr. of 1461 ed., 10 vols., Taibei: Wenhai Chubanshe, 1965.
Da Qing yi tong zhi (Comprehensive gazetteer of the Qing Empire). Rev. ed. authorised by the Jiaqing Emperor, also entitled Jia qing chong xiu yi tong zhi; completed 1843. First printed Shanghai 1934, with added index; repr. 11 vols., Taibei: Taiwan Shangwu Yinshuguan, 1966.
Diaocha baogao, di san bian: Nong kuang (Investigation reports [for Hebei Province]: No. 3: Agriculture and mining). Ed. by Hebei Provincial Government, Construction Office. N.p., 1928.
Eberstein, Bernd 1974. Bergbau und Bergarbeiter zur Ming-Zeit (1368–1644). Hamburg: Gesellschaft für Natur- und Völkerkunde Ostasiens.
Golas, Peter J. 1999. Joseph Needham, Science and civilisation in China. Vol. 5: Chemistry and chemical technology. Part 13: Mining. Cambridge: Cambridge University Press.
Graham, A. C. (tr.) 1981. Chuang-tzu: The Seven Inner Chapters and other writings from the book Chuang-tzu. London: George Allen & Unwin.
Guo Qingfan (ed.) 1961. Zhuang zi ji shi (Annotated edition of Zhuang zi). Orig. 1894; typeset ed. 1961; facs. repr. Beijing: Zhonghua Shuju, 1985 (Xinbian Zhu zi jicheng).
Huang Ming jing shi wen bian. (Collected memorials and other political writings of the Ming period). Comp. by Chen Zilong (1606–47), Xu Fuyuan (1600–1665), and Song Zhengbi (fl. 1643). Pinglu Tang ed., 1638; facs. repr. Taibei: Guofeng Chubanshe, 1964.
Huang Ming shi fa lu. (Political encyclopaedia of the Ming dynasty). Comp. by Chen Renxi (1579–1634?); Completed 1630, printed after 1632. Facs. repr. Taibei: Taiwan Xuesheng Shuju, 1965.
Li Bozhong 1987. ‘Ming Qing Jiangnan shehui shengchan zhong de tie yu qita jian jinshu’ (Iron and other base metals in social production in the Jiangnan region in the Ming and Qing periods). Zhongguo shi yanjiu, 1987. 2, 153–64.
Li Bozhong 2000. Jiangnan de zaoqi gongyehua 1550–1850 (Early industrialisation in Jiangnan, 1550–1850). Beijing: Shehui Kexue Wenxian Chubanshe.
Li Guoxiang & Yang Chang (chief eds.) 1993. Ming shi lu leizuan:Jingji shiliao juan. (Classified selections from the Veritable records of the Ming: Items relevant for economic history). Wuhan: Wuhan Chubanshe. Numerous typographical errors.
Li Guoxiang & Yang Chang (chief eds.) 1995. Ming shi lu leizuan: Hebei Tianjin juan (Classified selections from the Veritable records of the Ming: Items referring to Hebei and Tianjin). Wuhan: Wuhan Chubanshe. Numerous typographical errors.
Liu Yuncai 1978. ‘Zhongguo gudai gaolu de qiyuan he yanbian’ (The origin and development of the blast furnace in ancient China). Wenwu, 1978.2: 18–27.
Ming hui yao (Important documents of the Ming dynasty). Comp. by Long Wenbin (1821–93). Orig. 1887; typeset edition, 2 vols., Beijing: Zhonghua Shuju, 1956.
Ming shi lu (Veritable records of the Ming dynasty). Taibei: Zhongyang Yanjiuyuan Lishi Yuyan Yanjiusuo, 1963ff, 133 vols.
Ming shi (Standard history of the Ming Dynasty). Critical ed., 28 vols., Beijing: Zhonghua Shuju, 1974.
Ming shu (A history of the Ming dynasty on the pattern of the official histories). By Fu Weilin (d. 1667). Cong shu ji cheng ed.
Nan’an xian zhi (Gazetteer of Nan’an Prefecture, Fujian). Rev. ed. 1671 by Liu You. Facs. repr., 2 vols., Taibei: Nan’an Tongxianghui, 1973.
Qing dai de kuang ye (Sources on the mining industry in the Qing period). 2 vols., Beijing: Zhonghua Shuju, 1983.
Qiu jian xian sheng da quan wen ji (Collected works of Wang Yun [1227–1304]). Sibu congkan ed.
von Richthofen, Ferdinand [1872?]. Baron Richthofen’s letters, 1870–1872 [to the Shanghai Chamber of Commerce]. Shanghai: North China Herald, n.d.
Rosenqvist, Terkel 1974. Principles of extractive metallurgy. New York: McGraw–Hill, 1974. (McGraw–Hill series in materials science and engineering.)
Sakuma Shigeo 1972. ‘Mindai no Junka tetsushō ni tsuite – Gan’ei tetsuya no rōeki taisei’ (The ironworks at Zunhua in the Ming period – The corvée system of a state-administered ironworks). Pp. 203–14 in Yamamoto hakase kanreki kinen Tōyōshi ronsō (‘Jubilee volume presented to Dr. Tatsuro Yamamoto on the occasion of his sixtieth birthday’). Tōkyō: Yamakawa Shuppansha.
Schurmann, Herbert Franz (tr.) 1967. Economic structure of the Yüan dynasty: Translation of chapters 93 and 94 of the Yüan shih. Cambridge, Mass.: Harvard University Press. (Harvard–Yenching Institute Studies, 16.)
Si ku quan shu zong mu ti yao (Annotated bibliography of the ‘Complete library of the four categories of literature’). Typeset ed. with index, 4 vols., Shanghai: Shangwu Yinshuguan, 1933.
Sun, E-tu Zen & Shiou-chuan Sun (trs.) 1966. T’ien-kung k’ai-wu: Chinese technology in the seventeenth century. University Park & London: Pennsylvania State University Press.
Tegengren, F. R. 1921–24. The iron ores and iron industry of China: Including a summary of the iron situation of the circum-Pacific region. 2 vols., Peking: Geological Survey of China, Ministry of Agriculture and Commerce (Memoirs of the Geological Survey of China, series A, no. 2.) Part I 1921–23, part II 1923–24. English text and abridged Chinese translation by Xie Jiarong. Chinese title: Zhongguo tiekuang zhi, by Ding Gelan (Dizhi zhuanbao, A.2). Xerographic repr. available from University Microfilms, Ann Arbor, Mich., order no. OP 12715.
Tian fu guang ji (A work on Beijing and its environs). By Sun Chengze (1563–1638). First publ. 1962; typeset ed. Beijing: Beijing Guji Chubanshe, 1982.
Tian gong kai wu (The products of Heaven and Man). By Song Yingxing, d. ca. 1660; preface dated 1637. Facs. repr. of orig. 1637 ed., together with three other titles, Shanghai: Shanghai Guji Chubanshe, 1988 (Zhongguo gudai banhua congkan, 3).
Voiret, J.-P. 1986. Review of Wagner 1985. Neue Zürcher Zeitung, 15 Nov. 1986, 39.
Wagner, Donald B. 1984. ‘Some traditional Chinese iron production techniques practiced in the 20th century.’ Journal of the Historical Metallurgy Society, 18.2: 95–104.
Wagner, Donald B. 1985. Dabieshan: Traditional Chinese iron-production techniques practised in southern Henan in the twentieth century. London & Malmö: Curzon Press, , 1985. (Scandinavian Institute of Asian Studies monograph series, no. 52.)
Wagner, Donald B. 1993. Iron and steel in ancient China. Leiden: Brill. (Handbuch der Orientalistik, vierte Abteilung: China, no. 9.)
Wagner, Donald B. 1997. The traditional Chinese iron industry and its modern fate. Richmond, Surrey: Curzon Press.
Wagner, Donald B. 2001a. The state and the iron industry in Han China. Copenhagen: Nordic Institute of Asian Studies. (NIAS report series, 44.)
Wagner, Donald B. 2001b. ‘Blast furnaces in Song-Yuan China’, East Asian science, technology, and medicine, 18: 41-74.
Wagner, Donald B. (2001c). ‘The administration of the iron industry in eleventh-century China.’ Journal of the economic and social history of the Orient, 2001, 44 (no. 2), 175–97.
Wagner, Donald B. 2003.‘Chinese blast furnaces from the 10th to the 14th century’, Journal of the Historical Metallurgy Society, 2003, 37.1: 25-37. Abridged version of Wagner 2001b.
Yang Dajin (ed.) 1938. Xiandai Zhongguo shiye zhi (Gazetteer of modern Chinese industry and commerce). Changsha, 1938; facs. repr. in 3 vols., Taibei: Huashi Chubanshe, 1978 (Zhongguo jingji shiliao congshu, E0105).
Yang Kuan 1960. Zhongguo tufa yetie lian’gang jishu fazhan jianshi (A brief history of the development of Chinese regional siderurgical techniques). Shanghai: Renmin Chubanshe.
Yong chuang xiao pin (Short notes on various subjects, in large part related to Ming history). By Zhu Guozhen (ca. 1557–1632); written between 1609 and 1621, printed 1622. Typeset ed., 2 vols., Shanghai: Zhonghua Shuju, 1959 (Ming Qing biji xiaoshuo).
Yuan shi (Standard history of the Yuan Dynasty). Critical ed., 15 vols., Beijing: Zhonghua Shuju, 1976.
Zhu Baojiong & Xie Peilin (comps.) 1980. Ming Qing jinshi timing bei lu suoyin (Index of names of jinshi graduates in the Ming and Qing periods). Shanghai: Shanghai Guji Chubanshe.
Zun hua xian zhi (Gazetteer of Zunhua County, Hebei). By an ad hoc committee. Shijiazhuang: Hebei Chubanshe, 1990.
 This article is part of a larger project, the volume on Ferrous Metallurgy for Joseph Needham’s Science and civilisation in China. Support from The Nordic Institute of Asian Studies, The Danish Research Council for the Humanities, The Carlsberg Foundation, The University of Copenhagen, The Julie von Müllen Foundation, The Needham Research Institute, The Leverhulme Trust, and The Technical University of Berlin is gratefully acknowledged. I have received advice on earlier versions from so many persons that I cannot list them here, but I remain responsible for all remaining errors, misunderstandings, and infelicities of expression.
 It sems important to emphasize here our ignorance of the Ming iron industry, for Robert Hartwell has claimed, without real evidence, that Ming iron production was far inferior to that of the Song. I have examined Hartwell’s claims for the Song critically in Wagner (2001c).
 See e.g. Sakuma Shigeo 1972 and Eberstein 1974: 33–7, together with several other studies which they cite.
 See e.g. Wagner 1993: 261; 2001a: 100.
 Da Ming yi tong zhi, 1: 8a; Eberstein 1974: 33.
 This is stated in a memorial by Wang Yun in the 1260’s. See Qiu jian xian sheng da quan wen ji, 90: 19a–b; cf. Chun ming meng yu lu, 46: 68b; Eberstein 1974: 33.
 The following is largely based on Eberstein 1974: 23–32.
 Ming shi lu, Tai-zu, 176: 5b; Li Guoxiang & Yang Chang 1993: 120.
 We seem to have no sources on this issue with respect to iron production, but one source notes a proposal to start lead mining in Zunhua which came to nothing because of opposition by the local gentry (shi shen). Chun ming meng yu lu, 46: 70b.
 Wagner 2001b; 2003.
 E.g. Chun ming meng yu lu, 46: 66b.
 Ming shi lu, Tai-zong, 115: 5a; Li Guoxiang & Yang Chang 1993: 121; Huang Ming shi fa lu, 61: 8a–16a; Eberstein 1974: 38.
 ‘Fining’ (chao, sometimes translated ‘puddling’ or ‘refining’) is the process by which cast iron is converted to wrought iron or mild steel. See e.g. Wagner 1985: 22–5, 60–69; 1997: 21, 24–5, 39–43.
 The figure of 120 soldiers, of whom half engaged in agriculture, is given in the Ming shi lu (cited above). Unfortunately it is not made explicit whether this was the total number or the number at each of the 25 ironworks. If 120 is in fact the total number, then the labour used would be only 3 man-months per tonne, fitting nicely with the Dabieshan figure; but the higher figure is supported by the later production figures for the state ironworks at Zunhua (see below).
 Wagner 1985: 2. And an engineer, reviewing Wagner 1985 in Neue Zürcher Zeitung, called the Dabieshan productivity ein Elend, miserable, compared with modern methods. Voiret 1986.
 Da Ming yi tong zhi, 1: 19a–b.
 Ming shi lu, Tai-zong, 20a: 4b; cf. Eberstein 1974: 33, 55 n. 37.
 Ming shi lu, Xuan-zong, 14: 5a; Li Guoxiang & Yang Chang 1995: 790. An abridged quotation of this memorial in Fu Weilin’s unofficial history, Ming shu (82: 1669–70) omits the word shu, ‘wrought’, so that the memorial appears to concern iron in general rather than specifically wrought iron (see also Chun ming meng yu lu, 46: 70b). This confused Eberstein (1974: 35), who was unaware of the version in Ming shi lu.
 Li Bozhong 1987; 2000: 299ff. Eberstein (1974: 35) interprets the term broadly, as the entire region south of the Yangzi, but in fact ‘Jiangnan’ was a well-defined toponym with the more limited meaning given here.
 Golas 1999: 157–8; Yuan shi, 94: 2384; Schurmann 1967: 161–2; Eberstein 1974: 26, 42–3.
 See Golas 1999: 164; also von Richthofen 1872: 47, 52; Tegengren 1921–3; Yang Dajin 1938, vol. 2: 313–15; Wagner 1985: 28–9.
 See e.g. Wagner 1997: 58–75, esp. p. 65.
 For example, ironsand iron usually has a significant titanium content, and this can be an advantage in steel.
 Perhaps it is significant that one of the later directors of the ironworks, Fu Jun, came from Fujian, another place where this technology was used. See below.
 Ming shi lu, Xuan-zong, 17: 4a–b; Li Guoxiang & Yang Chang 1993: 122; 1995: 790.
 Ming shi lu, Xuan-zong, 95: 10b; Li Guoxiang & Yang Chang 1995: 791.
 Ming shi lu, Ying-zong, 24: 9b; Li Guoxiang & Yang Chang 1995: 791.
 Da Ming hui dian, 194: 17a, 19a.
 Han Dazhang in Huang Ming jing shi wen bian , bu yi, 2: 7a–b; variant text, Chun ming meng yu lu, 46: 67a–b; Tian fu guang ji, 21: 287–8. All that is known of Han Dazhang is that he received his jinshi degree in 1493 (Zhu Baojiong & Xie Peilin 1980: 1497, 2485).
 Shapogu and Songzhagu are places in the north of Zunhua District, at or near mountain passes.
One reason for the two relocations of the ironworks, as well as the building of its wall after the second relocation, may have been their proximity to the border and vulnerability to attack by Mongol raiders.
The two names are also found in the sources with yu for gu, sha for sha, and peng for zha. Da Ming yi tong zhi, 1: 16a, 23a; Da Qing yi tong zhi, 46: 1a, 1b, 1b–2a; Ming hui yao, 194: 984. Baiyezhuang is presumably named for the nearby mountain Baiyeshan, where a legendary smith named ‘Master Baiye’ produced swords. Da Qing yi tong zhi, 45: 4b.
 Diaocha baogao, pp. 36–7.
 Zunhua xian zhi, pp. 61, 561.
 Ming shi, 273: 7006; Zunhua xian zhi, p. 61.
 Ming shi lu, Ying-zong 英宗, 46: 4b; Li Guoxiang & Yang Chang 1995, p.792; Han Dazhang in Huang Ming jing shi wen bian , bu yi, 2: 1a–2b; Da Ming hui dian, 194: 19b–20a.
 The little we know of Fu Jun comes from the following sources: Si ku quan shu zong mu ti yao, 17: 1759 (only source for the date 1513); Nan’an xian zhi, 14: 1b–2a, 3b–4b (1973 repr. pp. 754–5, 760–62) (brief biographies of Fu Jun, his father, and his son); Ben chao fen sheng ren wu kao, 30: 13a–b (1971 repr. pp. 6431–2); Zhu Baojiong & Xie Peilin 1980: 738, 2488.
 Si ku quan shu zong mu ti yao, 17: 1759.
 The editors go on to cite a single point on which the book can be used in evidential research, correcting an entry in the Ming shi concerning the organisation of the Ministry of Works (Hu bu), which had administrative responsibility for the ironworks.
 Yong chuang xiao pin, 4: 94–5; Chun ming meng yu lu, 46: 68a–69a.
 Liu Yuncai 1978: 25.
 The version in Chun ming meng yu lu has jian gan.
 I have been unable to identify the two types of stone.
 A place about 15–20 km southwest of the Zunhua district seat, probably at or near the modern village of Shimen. Da Qing yi tong zhi, 45: 4b, 7a, 7b, 8a, 46: 1b; Zunhua xian zhi, fold-out map.
 This paragraph is not in Chun ming meng yu lu.
 A reference to Zhuang zi (juan 3a, Guo Qingfan 1961: 262; tr. Graham, 1981: 88–9), in which Heaven and Earth are compared to a furnace, and their Fashioner and Transformer (zao hua) to a great smelting master.
 This fits with the earlier statement that iron is tapped from the blast furnace four times per day.
 Wagner 1984; 1985: 12–21, 48–59; 1997: 16–21.
 Huang Ming jing shi wen bian, bu yi, 2: e.g. 2b, 3b, 4b (1964 repr., vol. 30: 756, 758, 760).
 Tian gong kai wu, xia: 15b; Sun & Sun 1966: 248. Liu Yuncai (1978: 24–5) has a different interpretation.
 Golas 1999: 164–5, 246–55.
 Yang Kuan 1960: 109; also Liu Yuncai 1978: 25.
 Rosenqvist 1974: 333.
 Presumably like the Song-period blast furnaces described in Wagner 2001b: 42–3, 47–50; 2003: 25–8.
 Chang Lijun 1991. The journalistic author seems to be entirely unaware that iron production in Zunhua had a unique history.
 On the other hand, if Fu Jun’s furnace was dug into the hillside, this would explain why he refers to its ‘depth’ rather than its height.
As an anonymous referee for this article has noted, a visit on site might very well solve some questions of this sort. However a visit to this village, far out in the mountains east of Beijing, would be a very expensive undertaking; and it is not entirely improbable that one would find that the local farmers have completely ploughed away all remains of furnaces.
 Da Ming hui dian, 194: 18b–22a; Han Dazhang in Huang Ming jing shi wen bian, bu yi, 2: 1a–9b.
 Huang Ming jing shi wen bian, bu yi, 2: 1a–9b. He also gives some information on production of ironsand and charcoal per labourer which unfortunately is very difficult to interpret.
 Eberstein 1974: 33–4.
 The best fuel efficiency ever recorded for charcoal iron smelting (in Sweden in the early 20th century) is about 3/4 tonne charcoal per tonne pig iron (see e.g. Tegengren 1921–24, vol. 2: 334, fn.); in Dabieshan the corresponding figure was about 3 tonnes per tonne pig iron, and this is a much more realistic estimate for the Zunhua ironworks. Sluiced ironsand contains 50–60 per cent iron by weight, and one hectolitre would weigh at most 400 kg (the specific gravity of magnetite is 5.2).
 Eberstein 1974: 33–4.
 1400 t cast iron, 450 t wrought iron, 136 t steel, and 127 t of a type of iron called song tie, apparently an intermediate product, perhaps ‘spongy iron’. Han Dazhang in Huang Ming jing shi wen bian, bu yi, 2: 6b; Eberstein 1974: 34. Note the typographical error, tong tie 銅鐵 for gang tie, in Han Dazhang’s text.
 Han Dazhang in Huang Ming jing shi wen bian, bu yi, 2: 1b, 3a, 3b.
 But note that this assumption would imply that Fu Jun wrote Tie ye zhi after 1529, sixteen years after his first appointment at Zunhua.
 Da Ming hui dian, 194: 21a–b.
 One possibility is that a da jian furnace was a furnace shaped like a large jian-bowl, i.e. one like that described in Wagner 2001b: 60, fig. 7; 2003: 31–2, figs. 8–9. But these do not fit Fu Jun’s description, written after 1513, so this hypothesis is unlikely to be correct.
 Eberstein (1974: 55, n. 42) has a different interpretation.
 Ming shi lu, Shen-zong, 110: 3a; Li Guoxiang & Yang Chang 1993: 134; Eberstein 1974: 37.
 Eberstein 1974: 37.
 Wagner 2001b: 42; 2003: 25–6.
 See e.g. Qing dai de kuang ye, 493–522.