Hostname: page-component-cd9895bd7-gbm5v Total loading time: 0 Render date: 2024-12-25T18:26:56.824Z Has data issue: false hasContentIssue false
  • English
  • Français

Markets, Technology, and the Structure of Enterprise in the Development of the Eleventh-Century Chinese Iron and Steel Industry*

Published online by Cambridge University Press:  03 February 2011

Robert Hartwell
Robert Hartwell
Affiliation:
University of Chicago
Get access Rights & Permissions [Opens in a new window]

Extract

From about 750 to 1100, China experienced a series of economic changes roughly comparable to the subsequent patterns of European growth from the Crusades to the eve of the French Revolution. The spread in the use of money, development of new credit and fiscal institutions, increase in interregional and international trade, and colonization of hitherto marginal land which took place in the Occident during the half millennium preceding the Reformation was paralleled by an earlier era of progress in East Asia during the two-hundred-fifty years from the rebellion of An Lu-shan (755) to the treaty of Shan-yüan 1004). And the achievements of late sixteenthand early seventeenth-century England, which John Nef terms an “early industrial revolution,” were in many respects even exceeded by the impressive expansion of mining and manufacturing in eleventh-century China.

Type
Articles
Information
The Journal of Economic History , Volume 26 , Issue 1 , March 1966 , pp. 29 - 58
Copyright
Copyright © The Economic History Association 1966

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Nef, John U., Cultural Foundations of Industrial Civilization (Cambridge [Engl.]: The University Press, 1958), pp. 50–62Google Scholar. The Reformation as the point of demarcation between a period of commercial expansion and one of industrial growth is analyzed by Nef in “Industrial Europe at the Time of the Reformation, ca. 1515-ca. 1540,” The Conquest of the Material World (Chicago: University of Chicago Press, 1964), pp. 67–117.Google Scholar

2 The cash (chʼien) was the money of account, and the string its multiple of one thousand. In 738 A.D., 277,000 strings of one-cash bronze coins were minted (Wang Ying-lin, Yü Hai [Chʼeng-tu Wang Shih ed.], 180/23a. Cited hereafter as YH). In 806 A.D., the receipts from the silver tax amounted to 12,000 liang (Hsiu, Ou-yang, Chʼi, Sung, and others, Hsin Tʼang-shu [Shanghai, 1884], 54/7aGoogle Scholar. Cited hereafter as HTS). This figure represents no more than 10 per cent of the total output of silver (Shigeru, Katō, Tō-sō jidai ni okeru kin-gin no kenkyū [Tōkyō: Toyo Bunkō, 1926], p. 534)Google Scholar. The market price of silver in 938, the earliest extant, was 680 cash a liang (ibid., p. 473). Consequently, 81,600 strings of cashʼ worth of silver were produced each year in the middle of the eighth century, bringing the total annual increment in the supply of money to 358,000,000 cash.

3 The price of rice in North China markets was 77 cash a U. S. standard bushel in 746 A.D. (HTS, 51/3a.). In this article, all conversions of weights and measures are based on those of Chʼeng-lo, Wu, Chung-kuo tu-liang-heng shih (Shanghai: Commercial Press, 1937).Google Scholar

4 More than six million strings of cashʼ worth of coins were minted yearly in the 1080's (Sung Hui-yao Chi-kao [Shanghai, 1936]Google Scholar, “Shih Huo,” 11/2a-3a, 6a, 8a-9a, 55/19a. Hereafter cited as SHY:SH. Fu-pao, Ting, Ku-chʼien ta-tzʼu-tien [Shanghai, 1938]Google Scholar, hsia-pien/133a-b, 148a, 245b-246a, 271a-271b, 350b, 446a; YH, 180/27a-27b, 28a-29b, 31a, 32b, 36b-37b; Tʼo, Tʼo et al. , Sung Shih [Shanghai: Tʼung-wen shu-chü, 1884]Google Scholar, 180/3a-3b, 5a-5b, 6b-7a, 8b. Hereafter cited as SS). By the eleventh century, a major use of gold was as a medium of exchange (Katō Shigeru, pp. 273–83). In 1078, the output of gold was 76,000 ounces, or 55,790 liang (SHY:SH, 33/7a-7b, 27a-29b), nearly twice the average annual yield of New World mines between 1503 and 1660 (Earl Hamilton, J., American Treasure and the Price Revolution in Spain, 1501–1650 [Cambridge: Harvard University Press, 1934], p. 42)CrossRefGoogle Scholar, and greater than the 70,000 ounces produced in Europe each year at the beginning of the sixteenth century (Soetbeer, Adolf, Edelmetall-Production und Werthverhältniss zwischen Gold und Silber seit der Entdeckung America's bis zur Gegenwart [Gotha, 1879], p. 107)Google Scholar. The price of gold from 1015 to the end of the eleventh century remained fairly stable at 10,000 cash a liang (Katō Shigeru, p. 473). Consequently, the value of gold output in 1078 was 557,900 strings of cash. The price of silver in 1076 was 1,500 cash per liang (Kaisaburō, Hino, “Gin, Ken no jukyu-jo hori mita Godai Hokusō no sai-hei, sai-shi,” Tōyō Gakuhō, XXXV [Sept. 1952], 8–9)Google Scholar. Consequently, the value of silver output in 1078 was 2,586,338 strings of cash, and the annual increment in the supply of money was about 9,174,238 strings of cash.

5 The price of rice in North China markets was 425 cash a U. S. standard bushel in 1075 (Tao, Li, Hsü Tzu-chih-tʼung-chien Chʼang-pien [Chekiang, 1881]Google Scholar, 265/5b-6a. Hereafter cited as HCP).

6 The population of China in 742 was slightly over 52.5 million. In calculating this total, I have used the data presented in HTS, pp. 37–43, corrected by comparisons with the statistics in Hsü, Liu, Chiu Tʼang-shu (Shanghai: Tʼung-wen shu-chü, 1884), 38–41Google Scholar. In 1078, there were about 16.5 million households in the Sung empire (HCP, 295/13b; Tsʼun, Wang, Yuan-feng chiu-yü chih ***[Wu-ying-tien, 1899 ed.], passim)Google Scholar. According to contemporary estimates, each household averaged five individuals (e.g., SHY:SH, 24/10b).

7 The calculations in this paragraph are based on rough data to be ultimately used in a more refined study on money, prices, and wages in China, ca. 620–1100 A.D. The argument represents a crude application of the techniques developed by Milton Friedman for using monetary data to reconstruct national income growth (“Monetary Data and National Income Estimates,” Economic Development and Cultural Change, IX [Apr. 1961], 267–86)Google Scholar. It should be emphasized that the data only suggest the possibility of growth in real per capita income. There is no way to make even a wild guess about changes in velocity; and a good proportion of the increased stock of money was certainly used in transactions transferred from the subsistence to the exchange sector of the economy. For a discussion of similar problems encountered in attempts to estimate national income growth in contemporary underdeveloped countries, see Bauer, Peter T. and Yamey, Basil S., The Economics of Underdeveloped Countries (Chicago: University of Chicago Press, 1957), pp. 16–24.Google Scholar

8 Hartwell, Robert, “Iron and Early Industrialism in Eleventh Century China” (unpublished Ph.D. dissertation, University of Chicago, 1963), pp. 37–54.Google Scholar

9 This figure is based on the total output of iron, copper, lead, and tin reflected in the mountain-and-marsh tax (shan-tse shui) list in SHY:SH, 33, and the “annual monopoly receipt tax” (sui-kʼo) list in Tuan-lin, Ma, Wen-hsien tʼung-kʼao (Chekiang, 1882–1896), 18Google Scholar: hereafter cited as TK. Although the mountain-and-marsh tax list is not dated, the report comes from the Kuo-chʼao hui-yao and uses circuit names only established in 1059, indicating that the statistics were collected between 1059 and 1077 (for dating of the various hui-yao see Chung, Tʼang, Sung Hui-yao yen-chiu [Shanghai: Commercial Press, 1932])Google Scholar. General surveys were made of the mining industry in 997; in 1021 (TK, 18/27a; HCP, 97/20a-21b); in 1049–53 (HCP, 97/20a-21b; TK, 18/24b-26a, 28a-28b; SS, 185/11a-11b); in 1064–67 (TK, 18/27b-28a, 28a-28b; SS, 185/13a); in 1074 (SHY:SH, 33/3b-4b), and in 1078 (SHY:SH, 33/12b-14a).

Since both the 1074 and 1078 surveys were limited to the sui-kʼo paying units, the only logical date for the mountain-and-marsh lists is 1064–67. The total receipts from the taxes on iron in 1064–67 equal the national total sui-kʼo receipts listed in TK, 18/28a-28b, plus the total mountain-and-marsh iron tax returns (SHY:SH, 33/27a-29b). At the 10 per cent rate (Hartwell, Robert, “A Revolution in the Chinese Iron and Coal Industries during the Northern Sung, 960–1126 A.D.,” Journal of Asian Studies, XXI [Feb. 1962], p. 155 n.)Google Scholar, this reflects a total production of pig iron per year in 1064–67 of 90,400 tons. The annual output of copper in 1067 was 93,000 tons (TK, 18/28a-28b; SHY:SH, 33/27a-29b). This figure assumes a tax rate of 20 per cent. Although Katō Shigeru, pp. 527–28, says that the rate on the base metals was 10 per cent during the Northern Sung, SHY:SH, 34/20b implies that it was 20 per cent on copper before 1085. Otherwise the calculations are the same as above. The output of lead was 65,000 tons (SHY:SH, 33/27a-29b; TK, 18/28b) and of tin, 48,500 tons (ibid.).

10 See Table 1.

11 The English ratios are computed on the basis of the average price of iron and wheat in 1600 (James Rogers, E. Thorold, A History of Agriculture and Prices in England [London, 1887])Google Scholar; in 1693–97 (ibid., 101–49, 610–13); in 1782–90 (Tooke, Thomas, Thoughts and Details on Prices from 1793–1822 [London, 1824]Google Scholar, appendix to Part IV); and in 1818–22 (ibid.). In 997, pig-iron prices of 19.77, 25.55, and 34.98 cash a pound—an average of 26.77 cash a pound—are recorded for three Szechwan prefectures (Ting Fu-pao, hsia-pien/226a-226b). The average price of rice in Szechwan, 998–1003, was 4.24 cash a pound (Chʼi, Han, Chung nsien Han Wei-wang An-Yang Chi [Chou-chin Tʼang ed.], 50/15bGoogle Scholar; Chen, Fan, Tung-chai chi-shih [Shanghai: Ts'ung-shu chi-chʼeng, 1936]Google Scholar, chüan 4, p. 24). In 1080, pig iron sold at 19.77 cash a pound in Hsing-chou (Tʼao, Lü, Ching-te chi [Wu-ying-tien, 1899 ed.], 4/12a)Google Scholar, while the average price of rice in Szechwan between 1077 and 1086 was 11.20 cash a pound (Lü Tʼao, 1/2a; Chih, Liu, Chung-su chi [Wu-ying-tien, 1899 ed.], 5/12a)Google Scholar. In 1074. iron cost 15.21 cash a pound in Shensi (HCP, 256/16a-16b), while the average price of rice in North China in 1074 was 11.22 cash a pound (HCP, 251/26b).

12 This fairly technical term, used throughout this paper, is defined in the 1954 edition of Webster's unabridged dictionary as “the metallurgy of iron and steel.”

13 See map, p. 30.

14 Tegengren, F. R., The Iron Ores and Iron Industry of China (Peking: The Geological Survey of China, 1921–1924), Part II, pp. 400–3.Google Scholar

15 Ibid., passim. Besides Tegengren, this statement is based on a careful study of everything published by the Geological Survey of China, and all other available reports, periodicals, and books written over the past one hundred years which discuss iron reserves in the territory of China formerly governed by the Sung emperors. For a list of Sung iron-mining sites, see Hartwell, “Iron and Industrialism” (cited in n. 8), pp. 178–95.

16 In 1078, the Ku-chen smelter at Wu-an hsien reported an annual yield of 12,812 tons of pig iron (see n. 35). By the Chʼien-lung period (1736–1795) of the Chʼing dynasty (1644–1911), the only reminder of the once flourishing industry at Wu-an was a stone tablet on the ore hill of Hung-shan stating that iron could be made from the rock (Tegengren, p. 334). From the Opium War to the establishment of the Communist regime in 1949, it would appear that not one ounce of iron was produced in this district. In addition to Tegengren, pp. 315, 333, this statement is based on a thorough survey of all publications on Chinese mining published during the past century, including the documents in Academia Sinica, Institute of Modern History, Kʼuang-wu tang (Taipei, 1960)Google Scholar, 8 vols. The important Li-kuo deposits which are later discussed (pp. 44–45) were probably not worked after the twelfth century. Between 1126 and 1883, Chinese sources are silent about these deposits. In the latter year, Chʼing provincial officials—citing Su Tung-pʼoʼs Sung memorial discussing the wealthy thirty-six enterprises at Li-kuo—leased the mines to merchants, but work was soon abandoned owing to financial difficulties caused by the Franco-Chinese War (1883–85). Kʼuang-wu tang, III, 1959. No work was going on when Tegengren visited the field in 1914 (p. 236).

17 For example, ores from important deposits at Ta-yeh, which were converted into 148,424 tons of pig iron in 1922, were only smelted into 347 tons in 1078. Shansi out-put was 69,000 tons in 1919 and only 7,384 tons in 1078. Tegengren, p. 399; Hartwell, “Iron and Industrialism,” Appendix II; Wen-hao, Weng, “Chung-kuo kʼuang-chʼan chih-lüeh,” Memoirs of the Geological Survey of China, Series B, No. 1 (July, 1919).Google Scholar

18 The terminology and analysis used in the following discussion of market areas and the location of siderurgical enterprises are based on Lösch, August, The Economics of Location, translated by Woglom, William H. (New Haven: Yale University Press, 1954).Google Scholar

19 See Table 2.

20 See ns. 35, 36, and 37.

21 I am using the concept “metropolitan market” in the sense developed by Gras, N. S. B., The Evolution of the English Corn Market (Cambridge: Harvard University Press, 1915), p. 95CrossRefGoogle Scholar. “The metropolitan market may be described as a large district having one center in which is focused a considerable trade. Trade between outlying ports of course may take place, but it is that between the metropolitan town and the rest of the area that dominates all. This is chiefly the exchange of the raw products of the country for the manufactured or imported goods of the town. The prices of all goods sent to the metropolitan center are ‘made’ there, or, in other words, prices diminish as the distance from the center is increased.”

22 SHY, “Ping,” 3/3a ff. Hereafter cited as SHY:P. I am indebted to E. A. Kracke, Jr. for the reference to this document.

23 The prefectural population increased by 284,840 people between 980 and 1080, and most of the increment can probably be ascribed to the city's growth (Shih, Yüeh, Tʼai-pʼing huan-yü-chi [Chin-ling, 1882]Google Scholar, 1/4a; Wang Tsʼun 1/2a).

24 See n. 21.

25 SHY, “Fang-yü,” 3/51a. Hereafter cited as SHY:FY.

26 SS, 197/1a; YH, 151/42a-42b.

27 SHY:FY, 3/50b-51a; TK, 161/26a-27a; YH, 151/42a-42b.

28 YH, 151/42a-43b; SS/197/1a.

29 TK, 161/26a-27a; YH, 151/42a-42b.

30 SS, 197/1a; YH, 151/42a-42b.

31 TK, 161/26a-27a; YH, 151/42a-42b.

32 This figure represents the total for the Wen-ssu Yüan (SHY; “Chih-kuan,” 29/1a. Hereafter cited as SHY:CK), the Tung Hsi pa-tso ssu (SHY:CK, 30/7a), the I-luan ssu (SHY:CK, 22/5a), and the Chu-hsieh wu (SHY:SH, 55/19a). Many other workers must have been employed at the Hou-yüan tsao-tso-so which manufactured saws, nails, bells, iron musical instruments, and other iron products (SHY:CK, 36/72b-73a).

33 SHY:CK, 36/72b-73a; 29/1a; 30/7a; 22/6b; SHY:SH, 55/19b.

34 Meng Yüan-lao, Tung-ching meng-hua lu (Hsüeh-chin tʼao-yüan ed.), 3/6a, 3/8b-9a, 6/3b.

35 Hsing-chou, Sha-ho hsien, Chʼi-tsʼun yeh produced 14,126 tons; and Tzʼu-chou, Wu-an hsien, Ku-chen yeh-wu reported 12,812 tons (SHY:SH, 33/12b-14a). The mountain-and-marsh tax receipts reflect an output of 6,937 tons, which was probably produced, in the main, at Li-chʼeng in Hsiang-chou, Lin-lü (Lin) hsien where the government offices had been closed in 1055 (Hung-pʼan, Hsü, Fang-yü kʼao-teng-kao [Chi-ning Fan-shih Hua-chin-chien-kʼo ed., 1932], 29/20b-21aGoogle Scholar; SHY:SH, 33/27a-29b). If the output of the Pan-yang yeh-wu at Hsiang-chou, Lin-lü (Lin) hsien was only the average of 1,300 tons, then the minimum output for the district would be 35,175 tons. SHY:SH, 33/3b-4a. For the location of this smelter, see Li Chien-chʼüan et al., Chʼung-hsiu Lin-hsien chih (1932 ed.), 1/10a, and Hsü Hung-pʼan, 29/21a. For average yield of enterprises supervised by a smelter office, see Jnl. Asian Stud., XXI (1962), 154–55, n. 12.Google Scholar

36 From 2,002 to 14,280 tons at Li-kuo and 2,600 tons at Ta-tʼung chien. The mountain-and-marsh tax statistics were collected in the period from 1064 to 1067 (TK, 18/28a-28b; SHY:SH, 33/27a-29b). At this time, the output of the works at Ta-tʼung chien was taxed as sui-kʼo, and consequently was not included in the mountain-and-marsh tax returns. The chien was abolished before 1074 to 1077, and consequently the output was also not included in the sui-kʼo returns for 1074 to 1077 or 1078 (SHY:SH, 33/12b-14a). We must interpolate in order to estimate probable production. The output of a works administered by a yeh, or smelter office, averaged 1,300 tons annually, and each yeh probably collected taxes from several smelters. Since Ta-tʼung chien administered two yeh, 2,600 tons would seem to be a fair estimate of average annual output, 1074–1078. The Ta-tʼung office administered an eastern and a western smelter (SHY:SH, 33/3b-4a).

The sui-kʼo receipts for Li-kuo chien, reflecting an annual output of 2,002 tons (SHY:SH, 33/12b-14a), are probably incomplete. This would not be unlikely. In at least eight instances out of a total of thirty-five entries, the sui-kʼo figures are completely missing from the listing of quota and receipts reflecting output from 1075 to 1078 (SHY:SH, 33/12b-14a). Evidence concerning the industry at Tʼung-shan during the seven years following the sui-kʼo receipt list of 1078 suggests that the combined output of Li-kuo and Ta-tʼung was much higher than our estimated 4,602 tons. This might have been an increase owing to the adoption of coal after the discovery at Pai-tʼu chen (see p. 51). But it seems hardly likely that such a revolution as a two- or threefold increment could take place within such a short period of time. The annual rate of growth of the smelters included in the quota and 1078 lists was no more than 2.7 per cent between 1077 and 1078, and probably somewhat less, since the quota figure is an average of output in 1075, 1076, and 1077 (Kaisaburō, Hino, “Hokusō jidai ni okeru dō-tetsu no sanshutsu-gaku ni tsuite,” Tōyō Gakuhō, XXII [1934], 109–10Google Scholar; SHY:SH, 33/12b-14a). If 2.7 per cent continued to be the annual rate of growth for all the smelters at Li-kuo Yi, at the estimated output of 4,602 tons in 1078, then the yield by 1085 would still be only about 5,655 tons. But this figure is certainly too low. In 1083, an iron-coin mint, Pao-feng chien, was established at Hsü-chou (Tʼung-shan) to cast 200,000 strings of two-cash iron coins a year (HCP, 356/15b). This number of coins required 2,340 tons of pig iron (HCP, 344/5a), and this quota was apparently estimated on the basis of the sui-kʼo returns for 1078, i.e., 10 per cent of what was supposed to be total output. In 1084, another mint, Pao-feng hsia chien, was established with a similar quota (HCP, 344/5a; 345/9a), bringing the anticipated consumption of iron for casting iron coins alone to 4,680 tons—or 78 tons more than our estimated yield in 1078. But coinage was not the only market supplied by the smelters at Li-kuo Yi during these years. In addition to Kaifeng, the products of Lai-wu and Li-kuo foundries were transported to general arsenals (Tu-tso-yüan) at Tʼung-shan (Hsü-chou), Lin-tzu (Chʼing-chou) and Tung-pʼing (Yün-chou), as well as to branch arsenals (Hsiao-tso-yüan) in other prefectures (HCP, 339/11b; TK, 18/32a). Agricultural implements were manufactured at the foundries to supply the needs of farmers in Shantung (ibid.), and siderurgical products were transported from Li-kuo to Kaifeng, Hopei, and even Korea (Shih, Su, Tung-pʼo chʼi chi [Pao-hua huicheng li-ming chʼeng-hua, 1909]Google Scholar, “Tsou-i,” 2/9b; 13/9a-9b; “Hsü-cni,” 11/21a). Moreover, the structure of enterprise at individual foundries also suggests a much higher output. Each of the thirty-six smelters employed over one hundred men full time to mine the ore, gather the fuel, and run the furnaces (ibid., 2/10a; 11/21b). One furnace man could produce 15 metric tons of iron a year, one miner could dig 30 tons of ore, and one collieryman could dig 60 tons of coal during the early years of the twentieth century with equipment similar to that in use during the Sung (Torgasheff, Boris P., “Mining Labor in China,” Chinese Economic Journal, VI [April 1930], 511, 515, 522)Google Scholar. Assume that the efficiency of labor was not much different in mining coal and in producing charcoal. A blast furnace obtained one ton of pig iron from 2.5 tons of ore and 3 tons of fuel (Tegengren, pp. 327, 338–39). Consequently, one furnaceman, 1.25 ore miners, and .75 charcoal workers, or three men would be necessary to produce 15 metric tons of pig iron a year. One hundred workmen could produce five hundred tons a year. An estimate of smelter output based on the probable equipment of the average enterprise yields a comparable figure. Before 1078, charcoal was the primary fuel (Su Shih, “Tsou-i,” 2/10a; “Shih,” 10/11b-12a), and the crucible process was probably limited to anthracite furnaces at the larger centers during the Sung (see p. 51). Moreover, the coal discovered at Pai-tʼu chen in 1078, and subsequently used in Li-kuo furnaces, was bituminous (see n. 92) and unsuitable for use in the crucible process. Small blast furnaces had a capacity to produce .5 tons of pig iron per diem, while the larger shaft furnaces yielded 1.3 to 1.4 tons daily (see p. 53). During the work year of three hundred days (HCP, 276/13a), one small furnace could produce 150 tons, and a large furnace up to 420 tons. The generally advanced character of enterprise at this district suggests that the shaft furnaces were probably used (see pp. 45 ff.). If the average of 500 tons, 150 tons, and 420 tons (357 tons) is taken as the yield of the typical smelter, then the total output at Li-kuo chien was possibly 12,852 metric tons or 14,280 short tons in 1078, a reasonable figure in light of the evidence presented in this note.

37 The mountain-and-marsh tax receipts probably reflect the output of Chin-ling chen (Tzu-chʼuan) and smaller enterprises near Lai-wu that were taxed by subprefectural rather than industrial prefectural authorities (SHY:SH, 33/12b-14a; 27a-29b). At the beginning of the Sung, iron ore and loadstone were mined at Hsi-shan, which was about twenty-three miles north of the seat of Szechwan hsien (Yüeh Shih, 19/3b), almost the exact location of the deposits reported at Chin-ling chen by modern writers such as Richthofen and Tegengren (Tegengren, p. 141). Anthracite from Hung-shan (ibid.) was probably used to smelt the ore here, as was certainly the case in the eighteenth century (Chang Tʼing-tsʼai and Chang Wu-to, Tzu-chʼuan hsien chih [1776], 1/42a-43b). For a description of Sung iron mining at Lai-wu, see Yüeh Shih, 21/12a-16a.

38 SHY:SH, 33/12b-14a.

39 Ibid.

40 See Table 2.

41 The resulting division of the iron industry into static and progressive sectors appears to have much in common with J. H. Boeke's “sociological dualism.” For a summary of the various theories of dualism in economic development, see Higgins, Benjamin, Economic Development (New York: W. W. Norton, 1959), pp. 274–93, 314–44Google Scholar. For a general outline and critique of Boeke's theory, see Itagaki, Yoichi, “Some Notes on the Controversy Concerning Boeke's ‘Dualistic Theory’: Implications for the Theory of Economic Development in Underdeveloped Countries,” Hitotsubashi Journal of Economics, I (Oct. 1960), pp. 13–28Google Scholar. According to Boeke, classical economic theory was impotent in dealing with premodern “Oriental” economic activity.

42 Schmoller, Gustav, “Die geschichtliche Entwickelung der Unternehmung,” Jahrbuch für Gesetzgebung, Verwaltung, und Volkswirtschaft im deutschen Reich, XIV (1890), 735–47.Google Scholar

43 Nef, John U., “Mining and Metallurgy in Medieval Civilization,” The Cambridge Economic History of Europe, ed. Postan, M. and Rich, E. E. (Cambridge [Engl.]: The University Press, 1952), II, pp. 473–74Google Scholar. Cf. Weber, Max, General Economic History, tr. Knight, Frank H. (New York: Collier Books, 1961), pp. 143–44.Google Scholar

44 For example, a similar industry was located at Ku-shih hsien in Kuang-chou, where fourteen mining and smelter households (kʼeng-yeh-hu) were employed in 1110 A.D. (SHY:CK, 43/123b-124a).

45 Chʼeng, Pao, Pao Hsiao-su kung tsou-yi (Sheng-hsin ko, 1863 ed.)Google Scholar, 7/17a-17b; Tegengren, p. 139.

46 Pao Chʼeng, 7/17a-17b.

47 This was the case for this region in 1929–33 (Buck, John Lossing, Land Utilization in China, Statistics (Chicago: University of Chicago Press, 1937), pp. 307–8.Google Scholar

48 The mean December, January, and February temperatures at Chefoo are 34°, 23°, and 30°, with lows near zero (ibid., p. 7).

49 Native furnaces in early twentieth-century China ranged in per diem capacity from 0.065 tons (Tegengren, p. 355), 0.13 (ibid.), 0.17 (ibid., p. 326), 0.24 (ibid.), 0.34 (ibid., pp. 320, 326), 0.40 (ibid., p. 339), 0.47 (ibid., p. 334), 0.48 (ibid., p. 326), 0.50 (ibid., p. 339; Wertime, p. 48), 0.68 (Tegengren, p. 337), 1.3 (ibid., pp. 333–34; Wertime, p. 55), and 1.4 (ibid.). The big difference is between the crucible-stall and small blast furnaces with daily heats of less than 0.68 tons and the larger shaft furnaces with capacities of nearly one and one-half tons per diem. The latter were never used in seasonal enterprises similar to those at Teng-chou. It would take one small blast furnace, operating full time at a per diem capacity of 0.5 tons, 44 days to yield the 22.2 tons reported for Teng-chou in 1078. This was also approximately the maximum time that a furnace could be operated without being closed down for repairs (Tegengren, p. 354). We get a similar result if we calculate on the basis of the efficiency of labor. The average per diem output of a worker (mining ore, gathering charcoal, and running the furnaces) was 0.017 tons a day (see n. 36). Eighteen households were engaged in the Teng-chou iron industry (Pao Chʼeng, 7/17a). The average number of adult males per household in Teng-chou in 1102 was 2.13 (Hope Wright, Geographical Names in Sung China [Paris: Ecole Pratique des Hautes Etudes, 1956], p. 157). The thirty-eight male workers would require over thirty-four days of steady effort to produce 22.2 tons. An increase in the number of furnaces would not materially alter the time required to yield this amount of iron with the labor force at hand.

50 Chih-wang, Wang, Han-pin Chi (Hu-pei hsien-cheng i-shu ed., 1923)Google Scholar, 8/7b.

51 Ibid., 8/7b-8a. A similar organization is suggested for gold mining in Fukien at the end of the tenth century by SHY:SH, 34/13b.

52 Since the Han (Ku, Pan, Chʼien Han-shu [Shanghai, 1884]Google Scholar, 28B/17b; possibly as early as the reign of Han Kao Tsu (206–195 B.C.) (Tegengren,, p. 304.)

53 Nef, in Cambridge Economic History (cited in n. 43), pp. 473–74.

54 Ibid., p. 474; Nef, John U., The Rise of the British Coal Industry (London: George Routledge, 1932), II, 143–45Google Scholar. Cf. Weber, pp. 143–44.

55 Su Shih, “Tsou-i,” 2/9b-10a, “Hsü-chi,” 11/21b.

56 Su Shih, “Tsou-i,” 2/9a-9b; “Hsu-chi,” 11/21a.

57 Pao Chʼeng, 7/17a-17b.

58 Su Shih, “Tsou-i,” 2/9a-9b; “Hsü-chi,” 11/21a.

59 SHY:SH, 34/14a-14b, 15b.

60 Pao Chʼeng, 7/17a-17b; Wang Chih-wang, 8/7b-8a.

61 Nef, in Cambridge Economic History, pp. 474–75.

62 Nef, Rise of British Coal Industry, II, 4.

63 Although the document in the Wen-chi is not dated, a very brief version in HCP bears with year 1086 (HCP, 390/16b-18a).

64 Lü Tʼao, 4/11b-12b.

65 Ibid., 1/12a, 13a-21a.

66 Eberhard, Wolfram, “Wang Ko, An Early Industrialist,” Oriens, X (1957), 248–52.CrossRefGoogle Scholar

67 HCP, 191/10a-10b.

68 Hartwell, “Iron and Industrialism” (cited in n. 8) pp. 219–22.

69 E.g., Su Shih, “Tsou-i,” 1/7b-26b, “Hsü-chi,” 11/19b-25b.

70 Nef, Rise of British Coal Industry, II, 49.

71 E.g., Lane, Frederic C., “Family Partnerships and Joint Ventures in the Venetian Republic,” Journal of Economic History, IV (1944), 178–96.CrossRefGoogle Scholar

72 Sombart, Werner, Ver modern Kapitalismus (2d. ed.; Munich: Duncker and Humblot, 1916), II, Pt. I, 89–90.Google Scholar

73 Tegengren, p. 236.

74 Ibid., p. 238.

75 At an annual yield of 4,000 tons, which is the minimum estimate for Li-kuo output in the 1070's, only fifty years would have been necessary to produce the estimated 200,000 tons (see p. 39).

76 Tegengren, p. 238.

77 Ibid., p. 237.

78 Su Shih, Tung-pʼo chih-lin (Hsüeh-chin tʼao-yüan ed.), 4/3b-4a.

79 Tegengren, p. 237.

80 These figures are based on an estimated yield of one ton of iron for every two and one-half tons of ore (see n. 36).

81 Yü-hsi, Liu, Liu pin-kʼo wen-chi (Chieh-i-lu Chu Shih Sheng-yü Tsʼung-shu ed., 1905)Google Scholar, 8/9a-10a.

82 Ibid., 8/9b.

83 Kenneth Madison, Professor of Biological Sciences, University of Illinois, a trained scientist as well as an authority on the history of Chinese science, was kind enough to explain the way in which this process would work.

84 That this was not an isolated example is suggested by the use of almost identical terminology in documents separated by three hundred fifty years.

85 This account is a part of the Hsing-chuang, or biography, of Yang Wang-hsiu (Yüeh, Lou, Kung-k'uei chi [Wu-ying-tien, 1899 ed.]Google Scholar, 91/1a-16b).

86 Lou Yüeh, 91/5b.

87 It would be interesting if evidence could be uncovered for the use of gunpowder in mining. It had been employed in warfare since the end of the tenth century. Cf. Goodrich, L. C. and Chia-sheng, Feng, “The Early Development of Firearms in China,” Isis, XXXVI, Pt. 2, No. 104 (1946), 114–23CrossRefGoogle Scholar, and Addendum, ibid., Pts. 3–4, Nos., 105–106, pp. 250–51; Ling, Wang, “On the Invention and Use of Gunpowder in China,” Isis, XXXVII (July 1947), 162–67Google Scholar. Blasting was apparently first used in Europe during the early decades of the seventeenth century (Galloway, Robert L., Annals of Coal Mining and the Coal Trade [London, 1898], pp. 226–27.Google Scholar

88 Su Shih, “Tsou-i,” 2/10a; “Hsü-chi,” 11/21b. The combination of mining and smelting into one enterprise was relatively common. Cf. Eberhard, 249–50.

89 Su Shih, “Tsou-i,” 2/10b; “Hsü-chi,” 11/22b.

90 Idem, “Tsou-i,” 2/10a; “Hsü-chi,” 11/21b.

91 The Sung seat of Hsü-chou and Li-kuo chien.

92 According to modern surveys, there is one seam of bituminous, five meters thick, with a reserve of 31,000,000 tons, at Pai-tʼu chai in the western part of Hsiao-hsien (Liu Chi-chʼen and Ju-chün, Chao, “Report on the Geology and Mineral Resources of Northern Anhui,” Bulletin of the Geological Survey of China, I (July 1919), 10.Google Scholar

93 Su Shih, “Shih,” 10/11b-12a.

94 Needham, Joseph, The Development of Iron and Steel Technology in China (London: The Newcomen Society, 1958), pp. 14–15, 19Google Scholar. Needham suggests that the early success of this technique in China was owing to the availability (and discovery) of good refractory clay.

95 The following description is based on Tegengren's account of the process (pp. 323–31).

96 Ibid., p. 323. Needham, p. 14. Needham points out that amounts of phosphorus up to 6 per cent in the iron reduce its melting point from 1,130° to about 950°. Tegengren's analysis, however, gives no evidence of iron phosphate present in the “hei-tʼu,” and he suggests that it merely acted as a flux.

97 Tegengren, p. 327.

98 Ibid., p. 330.

99 The stall furnaces were constructed almost entirely of broken crucibles, slag, and clay. The main investment was in the wind box which supplied the blast. Ibid., pp. 323–31, passim.

100 Needham, p. 5.

101 Wertime, p. 53.

102 Needham, p. 18.

103 pan Ku, Oüan 91. Kʼai-ying, Chin and Tseng-chʼüan, Hung, “Chung-kuo kʼo-sheng mei-chih fen-hsi,” Bulletin of the Geological Survey of China, XXI (July 1933), 27–161.Google Scholar

104 Needham, p. 47.

105 Cf. Tegengren, pp. 327, 338–39.

106 Wertime, p. 48.

107 These figures are based on the output of shaft furnaces in Hunan, which were roughly comparable to those excavated in Honan. Cf. Tegengren, pp. 333, 338–39; and Wertime, p. 55.

108 Wertime, p. 100.

109 E.g., Kao, Tung et al. , Chʼüan Tʼang Wen (Wu-ying-tien ed., 1818)Google Scholar, 119/2a, 863/16b-17a, 967/10a.

110 Kua, Shen, Meng-chʼi pi-tʼan (Ssu-pu tsʼung-kʼan ed.), 3/7b; Wertime, pp. 194–95Google Scholar; Needham, pp. 26–31. According to Needham, this is “co-fusion steel,” and the process is ancestral to that of Sieman-Martins. Wertime argues that “Needham's leap from Chinese interfusion to the modern Sieman-Martins process seems technically and geographically unjustified.” In any case, the Chinese achieved the same goal—the conversion of high-phosphorus iron (i.e., non-Bessemer iron) into steel.

111 Shen Kua, 3/7b-8a; Needham, p. 31; Wertime, p. 195.

112 Needham, p. 40; Wertime, pp. 194–95.

113 Shen Kua's description is the first mention of the process, and his discussion of it implies that it was not the common method in other parts of China even at the beginning of the last quarter of the eleventh century.

114 The Sung prefecture of Tzʼu-chou administered the important iron-producing subprefecture of Wu-an.

115 Shen Kua, 3/7b-8a. For dating of visit in 1075, see Tao-ching, Hu, Meng-chʼi pi-tʼan chiao-cheng (Shanghai: Shanghai Publishing Co., 1956), I, 135.Google Scholar

116 The use of anthracite in blast furnaces requires the use of a hot blast, first successfully developed in the United States in the late 1830's. Swank, James M., The Manufacture of Iron in All Ages (Philadelphia, 1892), pp. 358–59.Google Scholar

117 Hartwell, , Jnl. Asian Stud., XXI (1962), 159–60.Google Scholar

118 Read, Thomas T., “The Earliest Industrial Use of Coal,” Transactions of the Newcomen Society, XX (1939–40), 119.CrossRefGoogle Scholar

119 Needham, p. 14. He promises, however, to throw further light on the subject in Vol. VII of Science and Civilization in China.

120 Wertime, p. 55.

121 See text, p. 52.

122 Ibid. Substantial anthracite beds are located near Tzu-chʼuan and Lai-wu (Chin Kʼai-ying, pp. 27–161). The coal near Tʼung-shan is good coking bituminous, which could not be used for actually smelting the ore until the technological innovations of the eleventh century (ibid.; Chia-ying, Hsieh, “The Chia-wang Coal Field of Tʼungshan District, Kiangsu Province,” Bulletin of the Geological Survey of China, XVIII [Feb. 1932], 10Google Scholar; Wang, W. H. and Chi, Y. S., “The Liehshan and Liuchiakou Coal Field of Suhsien, Northern Anhui,” Bulletin of the Geological Survey of China, XVIII [Feb. 1932], 17).Google Scholar

123 “Sō-dai ni okeru sekitan to tetsu,” Tōhōgaku, XIII (Mar. 1957), 13.Google Scholar

124 Kʼang Pien, Chʼü-tʼan-lu (Ching-tai pi-shu ed.), hsia, 12b-14a.

125 Miyazaki, 11–28, passim.

126 Nef, Rise of British Coal Industry, I, 248–51.

127 Tʼien-kung kʼai-wu (Shanghai: Commercial Press, 1937), pp. 282–84.Google Scholar

128 Ibid., p. 189.

129 HCP, 164/12b; SS, 180/7a.

130 HCP, 79/4a-4b, 111/3b, 111/10a; Sung Ying-hsing, pp. 266, 282; Li Fang, Tʼai-pʼing yü-lan (Ssu-pu tsʼung-kʼan ed.), 871/3b; Wang An-shih, Lin-chʼuan hsiensheng wen-chi (Ssu-pu tsʼung-kʼan ed.), 3/5a.

131 I have found no example of its use in this sense in any Sung work. The term commonly used for burning charcoal is chih; e.g. Li Fang, 871/3b.

132 Hsiu, Ou-yang, Ou-yang wen-chung-kung chʼüan-chi (Hsiao-en-tʼang ed. 1748)Google Scholar, 118/11b.

133 Kʼang Pien, hsia, 12b-14a. The relevant Chinese characters for the terms employed in this paragraph and in n. 131 are the following:

shao shih-tʼan

tʼan pien lien

lien-tʼan chih.

134 Coal could have been used by the smith for working up the finished product, or it could have been used for a preliminary roasting of the iron stone. But the absence of readily available wood fuel or anthracite in the immediate vicinity of the smelting works suggests that bituminous coal was used in actually smelting the ore. Similar problems faced T. S. Ashton when he was attempting to date the discovery of the coking process in England by Abraham Derby in 1709. Ashton, T. S., Iron and Steel in the Industrial Revolution (Manchester: University Press, 1954), p. 30.Google Scholar

135 SHY:FY, 13/21a-21b; HCP, 106/6b.

136 SHY:SH, 37/10a.

137 SHY:SH, 161/92a-92b.

138 SHY:SH, 17/21a-21b.

139 SHY:FY, 13/21a-21b; HCP, 106/6b.

140 Hung Mai, Jung-chai sui-pi (Ssu-pu ts'ung-kʼan ed.), 11/4a-5a.

141 SHY:FY, 13/21a-21b; HCP, 106/6b.