^{page 290 note 1} And not mistakes in adding up the totals, cf. Bielenstein, H., ‘The Census of China’, in BMFEA., vol. 20 (1947), p. 128.Google Scholar

^{page 290 note 2} There are, of course, two styles of Chinese numerals, usually distinguished as the ‘ordinary style’ and the ‘large style’. The latter cannot easily be altered from one number to another, and are generally used to avoid fraud. In the Ming local gazetteers which I have examined, I have found only one or two which used ‘large style’ numerals for their statistical entries.

^{page 292 note 1} See Franke, Wolfgang, ‘Zur Kompilation und Überlieferung der Ming Shih-lu,’ in Sinologische Arbeiten, 5 Jahrg., Nr. 1/2 (Deutschland-Institut, Peking, 1943), pp. 1–46; espec. pp. 11 and 37–8.Google Scholar

^{page 292 note 2} See Ming Shih-lu, vol. 19Google Scholar; T'ai-tsu, , ch. 135, 5aGoogle Scholar. This edict is noticed in some detail below.

^{page 292 note 3} In most of the Fu, Chou, and Hsien local gazetteers of the Ming period that I have been able to see, population figures are given for these ‘tenth years’ only. The population figures from the Shih-lu, reproduced in Table I below (where the ‘tenth years’ have been marked by an asterisk), provide little evidence that revisions were actually carried out during those years. But see below, p. 311, note 3.

^{page 293 note 1} This is surprising, both in view of the administrative revival during that reign, and because Wan-li 6/1578 is one of the three years for which the Ming Shih Ti-li-chih and the Ta-Ming Hui-tien give detailed (i.e. provincial) totals.

^{page 298 note 1} Franke, Wolfgang, Preliminary Notes on the Important Chinese Literary Sources for the History of the Ming Dynasty. Studia Serica Monographs, Series A, No. 2. Chengtu, 1948. Espec. pp. 76–7Google Scholar. On the Ming History see also: Chin-hua, Li, Ming Shih Tsuan-hsiu K'ao (A History of the Compilation of the Ming Dynasty History), Yenching Monograph Series, No. 3. Peking, 1933Google Scholar; and Wên-lan, Fan, Chêng Shih K'ao-lüeh (A Study of the Standard Histories), Peking, 1931, pp. 283–294Google Scholar. Professor Fan writes: ‘the compilation, completion and checking of this book occupied a period of more than 60 years. That is why its form is so perfect, and its mistakes so very few. It is an excellent history.’ (p. 283.)

^{page 298 note 2} See the Ming bibliography Ch'ien-ch'ing-t'ang Shu-mu, ch. 9. 10 a–b (references are to the reprint in the Shih-yüan Ts'ung-shu). Also Franke, , Preliminary Notes, pp. 42–3Google Scholar. My chapter and page references to the Hui-tien are to the Wan-li print in 12 ts'ê.

^{page 299 note 1} A native of Shan-yin Hsien, in Chekiang, and a chin-shih of Chia-ching 26/1547. He rose in the official career to be Deputy Provincial Judge (An-ch'a-ssŭ Fu-shih) of Yunnan and President of the Imperial Stud Office (T'ai-p'u-ssŭ Ch'ing).

^{page 299 note 2} The Huang-yü K'ao is mentioned, but without comment, in Ch'ien-ch'ing-t'ang Shu-mu, ch. 6, 1b. See also Franke, , Preliminary Notes, p. 73Google Scholar. Much of the information it gives on population seems to have been taken from the Kuang-yü-t'u, an atlas of China and adjacent countries by Lo Hung-hsien. This work ante-dated the Huang-yü K'ao by about a decade (original version 1541; first printed edition c. 1555). I have used the copy in the British Museum, which is a Ch'ing reprint of 1799.

^{page 299 note 3} Detailed figures for the separate prefectures of Ching-shih are given only for 1491 and 1578. A single figure for the whole of the Northern Capital Territory is all we have for 1393. No figures are given for Kweichow for 1393. Its territory was not organized as a province until Yung-lo 11/1413.

^{page 299 note 4} On this question of the dating of the Huang-yü K'ao figures, see below, pp. 311–12. The totals given by the Huang-yü K'ao for Shantung are identical with those that appear in the History and Statutes for the year 1491, but this is an isolated case.

^{page 299 note 5} Provincial figures from the Ming History, ch. 40–6, are also reproduced by Rieger, M. in her article ‘Zur Finanz- und Agrargeschichte der Ming Dynastie’Google Scholar (Sinica, vol. xii, 1937, pp. 236–7Google Scholar), but unfortunately with many mistakes. For Kweichow the number of households given is much larger than the number of persons.

^{page 303 note 1} Provincial Gazetteer of Shantung compiled in the Chia-ching Reign, in 40 chüan, by Lu Yi (c. 1441–c. 1490), a native of K'un-shan Hsien, in Nanking. Chin-shih in T'ien-shun 8/1464. The author's preface to the work is dated Chia-ching, kuei-ssŭ (i.e. 12/1533). Population figures are given in ch. 8, 20b–21b. Available in Peking National Library Microfilms, Series A, rolls 371–2. See Ch'ien-ch'ing-t'ang Shu-mu, ch. 6, 29aGoogle Scholar; Shih-chia, Chu, Union List, vol. 2, Shantung, 1a.Google Scholar

^{page 306 note 1} On this general subject see Yi-t'ang, Hsü, Chung-kuo Nan-Pei chih Jen-k'ou Sheng-chiang (The Rise and Fall of Population in North and South China), in Chung-kuo Wên-hua Yen-chiu Hui-k'an (Bulletin of Chinese Studies), VII, 09, 1947, pp. 27–58Google Scholar. Unfortunately the author has little to say about the Ming period.

^{page 307 note 1} Bielenstein, Hans, The Census of China during the period A.D. 2–742 in BMFEA., vol. 20 (1947), pp. 125–163Google Scholar. The passages quoted are from pp. 142–3.

^{page 308 note 1} See Pulleyblank, E. G.'s forthcoming Background to the Rebellion of An Lu-Shan, Appendix II.Google Scholar

^{page 308 note 2} Professor Herbert Franke writes: ‘Das auffallendste Moment der Bevölkerungsverteilung zur Yüan-Zeit 1st aber die geringe Volkszahl des Nordens im Vergleich zum Süden, dem ehemaligen Sung-Gebiet … Die Südgebiete weisen 1290 89·3% der registrierten Gesamtbevölkerung auf, also fast 9/10 des Reichs! Auch für 1291 ergibt sich immer noch das Verhältnis von 85·3%, also rund 17/20.’ See his Geld und Wirtschaft in China unter der Mongolen-Herrschaft (Leipzig, 1949), p. 130.Google Scholar

^{page 308 note 3} The word li in the sense of a group of households occurs in the earliest Chinese writings, though different texts and commentators assign varying numbers of households to it. The commentary to the Book of Songs explains that five chia make a lin, and five lin a li: a total of 25 households. The Shang-shu Ta-chuan says that eight chia make a lin, three lin a p'êng, and three p'êng a li. According to this system 72 households made up a li. The Fêng-su-t'ung explains the li. as consisting of 50 households; the commentary to the Kung-yang Chuan speaks of 80 households; the Kuan-tzŭ of 100.

^{page 308 note 4} Ming Shih-lu, vol. 19, ch. 135, 5aGoogle Scholar. Similar but not identical texts describe the same edict in the Ta-Ming Hui-tien (TMHT), ts'ê 2, ch. 12, Hu Pu, 7, 64aGoogle Scholar; and in the Ming Shih, ch. 77, Shih-huo-chih (SHC), 1, 1a.Google Scholar

^{page 308 note 5} T'ien-hsia chün hsien, i.e. all the administrative areas of the empire. TMHT writes explicitly: fu, chou, hsien.

^{page 309 note 1} The text reads: t'ui ting liang to chih shih hu wei chang. The same version is found in SHC. TMHT on the other hand omits the character ‘liang’ after ‘ting’. Using this source, Nan-ming, Liu, Contribution à, l'étude de la population chinoise (Geneva, 1935), p. 11Google Scholar, translates: ‘110 families formaient un groupe appelé li, dont les dix families les plus pourvues d'adultes étaient les chefs.’

^{page 309 note 2} TMHT omits all mention of chia-shou. SHC writes that ‘the li-chang and chia-skou will be responsible for managing the affairs of the li and chia’. These local officials had many duties beyond those concerned with taxation and corvée. See the article by Yukio, Yamane, On the duty of the village headman in the Ming period, in Tōhōgaku, vol. 1, iii (01, 1952), pp. 79–87Google Scholar (in Japanese, with an English summary).

^{page 309 note 3} Hence the name fang-hsiang-li sometimes given to the system. This sentence seems out of context here, though it also occurs in this place in TMHT. In SHC the sentence comes later, after the description of the rotation hi office of the li-chang, which makes better sense of the passage.

^{page 309 note 4} Fan shih nien yi chou. The TMHT is equally elliptical. SHC makes clear the posts referred to. It adds that the ten-year system was called p'ai nien.

^{page 309 note 5} Here again there is a discrepancy between the texts of the Shih-lu and TMHT. The former has The latter substitutes the character shu for liang. SHC agrees with the Shih-lu.

^{page 309 note 6} SHC says that the widows and orphans should be written down at the end of the chia lists. It adds that Buddhist and Taoist priests were given certificates (tu-tieh) attesting their status as religious. If they owned laud they were included on the register just like other people. If they did not own land they were inscribed at the back of the lists like the ch'i-ling.

^{page 309 note 7} SHC adds that this copy was bound in yellow paper. Hence the name Yellow Registers.

^{page 310 note 1} A decision as to what? The answer is not clear from the context. It was probably either with regard to (a) the transfer of households, or (b) the readjustment of the tax assessment.

^{page 311 note 1} This official geography of the empire was produced, under the general editorship of Li Hsien (1408–1466) and others in T'ien-shun 5, iv/1461 May–June. In 90 chüan. There were several editions of this work: in Hung-chih 18/1505; in Chia-ching 38/1559; and two in Wan-li. See Ch'ien-ch'ing-t'ang Shu-mu, ch. 6, 1bGoogle Scholar; and Franke, , Preliminary Notes, p. 72Google Scholar. The total number of li for each province is nowhere stated, but the text does give the number of li separately for each chou and hsien. My totals for the provinces have been arrived at by adding these figures together. The two other outstanding general geographies of the Ming era, the Huan-yü T'ung-chih and the Kuang-yü-chi, unfortunately give no information about li.

^{page 311 note 2} In 11 chüan. Preface dated Chia-ching ting-ssŭ (36/1557). Yao chou was in Hsi-an Fu, Shensi. There were two hsien subordinate to it: T'ung-kuan and Fu-p'ing. See Shih-chia, Chu, Union List, vol. 2, Shensi, 4b.Google Scholar

^{page 311 note 3} Yao Chou Chih, ch. 1, 8a–b, 9a, 9bGoogle Scholar. It should be noted that all three of these changes are referred to ‘tenth years’.

^{page 312 note 1} This also suggests that the successive editions of the Yi-t'ung-chih were not mere reprints, but that each brought up to date the material contained in its predecessor.

^{page 312 note 2} Analysis of the chou and hsien figures for Shensi (see below, Section V) shows that this administrative simplification was carried out most thoroughly in the areas of densest population.

^{page 313 note 1} See above, p. 311.

^{page 315 note 1} In 40 chüan. The preface records that the book was completed in Chia-ching, jên-yin year, xii/1543 January–February. See Shih-chia, Chu, Union List, vol. 2, Shensi, 1aGoogle Scholar. I have used the copy reproduced as No. 734 on rolls 371–2 of the Peking National Library Microfilms. The population statistics reprinted in Tables VII and VIII below are from ch. 33, 6a–28b.

^{page 315 note 2} A native of San-yüan Hsien, Hsi-an Fu, Shensi. Chin-shih in Chêng-tê 9/1514. Served as First Secretary (Lang-chung) in charge of the Dept. for Examining Merit (K'ao-kung) of the Board of Civil Office. Appointed President of the Imperial Banqueting Court (Kuang-lu-ssŭ Ch'ing) in Chia-ching 12/1533.

^{page 321 note 1} See note 1, p. 303.

^{page 322 note 1} Prefectural Gazetteer of Ch'ing-chou. In 20 chüan. The preface is dated Wan-li 43/1615. See Shih-chia, Chu, Union List, vol. 2, Shantung, 8bGoogle Scholar. The copy I have used is No. 743, Peking National Library Microfilm Series, on rolls 374–5. Material on population is in oh. 5, 1b–3a.

^{page 322 note 2} A native of Yi-tu Hsien, Ch'ing-chou Fu, Shantung. A Chin-shih of Wan-li 8/1580. Became President of the Board of Public Works (Kung Pu Shang-shu) in T'ien-ch'i 3/1623.

^{page 323 note 1} Compiled by Hsüeh Ying-ch'i, a native of Wu-chiu Hsien, Kiangsu. Chin-shih in Chia-ching 14/1535. Was at one time the Director of Education (T'i-hsüek Fu-shih) for the province of Chekiang. The Gazetteer is in 72 chüan, and is dated Chia-ching 40/1561. See Shih-chia, Chu, Union List, vol. 1, Chekiang, 1aGoogle Scholar. I have used the copy reproduced as No. 823 in the Peking National Library Microfilm Series, on rolls 394–5.

^{page 323 note 2} It cannot be assumed, simply because both types of data are printed in the same work, that both are to be referred to the same date. The two sets of figures occur in different parts of the Gazetteer and have almost certainly been taken from different sources. Comparison of the li figures with those in the Huang-yü K'ao and Ta-Ming Yi-t'ung-chih suggests that their date lies between those of the two latter.

^{page 324 note 1} The late Professor Otto Franke went so far as to write: ‘Es ist ein hoffnungsloses Unternehmen, mit den chinesischen Zahlenangaben, auch wenn sie sich noch so genau gebärden, zu einer Schätzung zu gelangen.’ (Geschichte, vol. iii, p. 236.Google Scholar) Another note (op. cit., p. 342) reads: ‘Es zeigt sich auch hier wieder, dass die chinesischen Angaben über Bevölkerungszahleu wertlos sind.’ Dr. Liu Nan-ming, while not taking so extreme a view, is worried by ‘les variations brusques et discordantes’ which he meets with in the Chinese figures; and in particular by the seeming discrepancies which so often appear between the figures given for households and persons. He writes: ‘Les rapports anormaux constatés entre ces deux sortes de statistiques [hu k'ou] établies pour certaines années montrent l'invraisemblance des résultats obtenus. Nous avons vu, par suite des analyses faites précédemment sur les documents, qu'il y a, soit des contradictions ou des discordances inexplicables, soit des omissions ou des chiffres entachés d'erreurs évidentes. De tout cela, il ressort que les renseignements démographiques contenus dans les statuts officiels et les ouvrages classiques, ne méritent généralement pas de confiance.’ (Contribution, pp. 36–7.)Google Scholar

^{page 325 note 1} Accuracy in Chinese statistics does not seem to be a monopoly of the Ming. The last chapter of ProfessorBielenstein, H.'s forthcoming book on the Restoration of the Later HanGoogle Scholar (which he has been good enough to afford me the privilege of reading in typescript) contains convincing evidence of the general reliability of Han population statistics and of their evidential value in clearing up historical problems.

^{page 325 note 2} A helpful undertaking would be the large-scale collection and classification of copyist's and printer's mistakes in the transmission of Chinese numerical and other material. There is an interesting passage in Chapman, B. W.'s Textual Criticism (Johnsonian and Other Essays and Reviews, Oxford, 1953, p. 88)Google Scholar in which, speaking of the problems of English editorship, he writes: ‘The editors of the new Cambridge Shakespeare tell us that they are guided by a large collection of demonstrable misprints, which show them the kind of error to which Shakespeare's manuscripts were exposed. Much has been done in the investigation of handwriting. But it remains true that we still lack such an apparatus as classical scholars have long possessed. We do not yet know with any certainty what letters and combinations of letters are, at different periods and in different scripts, liable to confusion. We do not know how far printers were likely to omit groups of words by the error known as homoeoteleuton [like ending], to which copyists, modern as well as ancient, are quite extraordinarily prone.’ How much of this applies, with but minor modification, to the problems of emendation of Chinese texts!

^{page 325 note 3} For some striking examples see Ku Yen-wu, Jih Chih Lu, Chou hsien fu-shui (Ts'ung-shu, Kuo-hsüeh Chi-pên edit., vol. 1, ch. 8, 62–64.Google Scholar

^{page 325 note 4} Since writing the above, I have found the following quotation in an article written more than a century ago by Édouard Biot: ‘En rapprochant done les chiffres obtenus dans le même temps pour les families et les mesures de terres cultivees, on peut déterminer la quantité moyenne de terrain cultivé qu'occupait chaque famille, et si cette quantité est considérable, on peut présumer de la quo le recensement néglige une forte partie d'individus.’ This is from the ‘Memoire sur les recensements des terres consignés dans l'histoire chinoise et sur l'usage qu'on en peut faire pour évaluer la population totals de la Chine’, in Journal Asiatique, III Series, vol. 5 (04, 1838), pp. 305–331.Google Scholar

^{page 326 note 1} Though not perhaps quite as simply as in Ching-han, Li's Social Survey of Ting HsienGoogle Scholar, where the author of the section on population writes in his historical summary: 34 110 3, 740 ‘under the Ming the households were inscribed in 34 registers, each register comprised 110 households, so there were altogether 3,740 households.’ See Ting Hsien Shê-nui Kai-k'uang Tiao-ch'a (Peking, 1933), ch. 4, p. 121Google Scholar. The provincial H/L averages for Ching-shih (Hopei, the province in which Ting Hsien is situated) are 131 and 144 (see Table VI). This would give results of 4,454 and 4,896 households respectively.