^{page 526 note 2} There have been some penetrating remarks by Waley in his various works dealing with the T‘ang, and recently a brilliant preliminary sketch of the Buddhist church in T‘ang times was given by Reisehauer in Ennin's travels in T‘ang China, 1955. The works of Kenneth Ch‘en have also shed light on political relationships and he has recently published a study of the economic problem, ‘The economic background of the Hui-ch‘ang suppression of Buddhism‘, HJAS, XIX, 1–2, 1956, 67–105.Google Scholar Most important on the purely economic problems have been Yang Lien-shêng's ‘Buddhist monasteries and four money-raising institutions in Chinese history’, HJAS, XIII, 1 and 2, 1950, 174–91,Google Scholar and his Money and credit in China, 1952.

^{page 528 note 1} There is a very extensive literature on this subject. I refer readers to Ishida Mikinosuke Tempō-jūsai no teiseki ni miyuru Tonkō-chihō no sei‘iki-kei chūmin ni tsuite in Katō Hakusi kanreki kinen Tōyoshi Shūsetsu, 1941, and Pulleyblank, , ‘A Sogdian colony in Inner Mongolia’, T'oung Pao, XLI, 4–5, 1952,Google Scholar 317–56.

^{page 528 note 2} I have taken the household registers printed in Niida's, TōSō hōritsu bunsho no kenkyū, 1937, 676–721,Google Scholar as a basis for this and subsequent discussions.

^{page 528 note 3} Register of 769, fragment printed in Sha-chou wên-lu pu household of So Ssu-li first plot of land. ‘One plot 119 mou. (40 mou yung-ye, 19 mou hsün-t'ien, 36 mou k'ou-fên-t'ien, 14 mou purchased land mai-t'ien.)’—the sum of these areas is only 109 mou. It is clear that the family (who were local gentry, So Ssu-li himself being a retired military officer with high rank) had built this holding up by amalgamating smaller plots. So Ssu-li'a rank entitled him to 6,153 mou of land, and he was thus well within his limit. His total holdings were 240 mou, and he employed four slaves.

^{page 528 note 4} See Chin-shih ts'ui-pien, 113, pp. 47b–51b, Ch'ung-hsiu Ta-hsiang-ssu chi dated 841.

^{page 528 note 5} Even of these, one was a boundary of a field which was not under cultivation. At Tunhuang practically every plot had at least one boundary with a canal.

^{page 529 note 1} See Masao, Nishikawa ‘Tonkō hakken no Tōdai koseki-zankan ni arawareta Shiden ni tsuite’ Shigahu Zasshi, LXIV, 10, 1955, 38–60.Google Scholar

^{page 529 note 2} His rather ingenious proof rests primarily on an analysis of the plot boundaries, where the term constantly occurs. He makes the point, which, superficially, seems valid, that if means ‘own land’ as all scholars are so far agreed, where a tzu-t'ien appears as the eastern boundary of one plot in a family's possession, another should appear as the western boundary in the other plot. However, this assumes that lands were all divided into rectangular plots, and that each had only four neighbours. In fact, lands are rarely rectangular, especially in irrigated areas, and the ssu-chih were not the only neighbours, but four selected so as to give the position of the property. The parallel term tzu-chih which occurs in innumerable inscriptions in the meaning ‘own property’, is clear proof that the conventional meaning is correct.

^{page 529 note 3} See Niida, , TōSō hōritsu bunsho no kenkyū, 1937,Google Scholar 780 ff.

^{page 530 note 1} See the Register of 769 quoted in n. 3, p. 528 above, and Pelliot 2592, Register of 747, Household of Chêng En-yang etc.

^{page 530 note 2} It is possible that the climate of the area was less arid in T‘ang times than at the present day. See Coching, Chu, ‘Climatic changes during historic time in China’, JNOBRAS, LXII, 1931, 32–40,Google Scholar and ‘Climatic pulsations during historic times in China’, Geographical Review, XVI, 1926.Google Scholar However, the frequent occurrences of ‘desert’, ‘wasteland’, etc., in the Registers together with the sum of all other evidence would lead us to believe that the conditions were at least in general much as they are to-day.

^{page 530 note 3} See P. 3947 (p. 129). The author's discussion of the individual properties of monks is an important improvement on the traditional views.

^{page 530 note 4} P. 3354, Register for 747, Family Tu Huai-fêng

^{page 530 note 5} P. 3354, loc. cit.

^{page 530 note 6} Fragment of a Register for Liu-chung hsien Turfan, undated. Niida, op. cit., 691.

^{page 530 note 7} P. 3669, Register dated 701, Family Chang Hsüan-chün

^{page 531 note 1} See, for example, the painstaking but uncritical study of price data by Han-shêng, Ch'üan, ‘T'ang-tai wu-chia ti pien-tung’ Academia Sinica, Bulletin of the Institute of History and Philology, XI, 1947, 101–48.Google Scholar

^{page 531 note 2} There is a constant difficulty when dealing with the terms mi and su in T'ang texts. They may either mean hulled grain as opposed to unhulled grain, or rice as opposed to millet. I take them in the latter sense here.

^{page 531 note 3} A standard length of silk was 4 chang (i.e. 40 feet). This standard size was enforced because rolls of silk were commonly employed as currency.

^{page 532 note 1} See Shigeru, Katō, TōSō-jidai ni oheru kingin no kenkyū 1924 63–4.Google Scholar

^{page 532 note 2} See Katsutoshi, Ono, ‘Tōdai ni okeru ichi kinrei no kōshaku ni tsuite’ Shirin, XXII, 1, 1937, 87–110.Google Scholar

^{page 532 note 3} See Shigeru, Katō, Shina heizai-shi gaisetsu , 1944,Google Scholar 39 ff. Also see Osamu, Shinoda, ‘Go-koku no kigen’ Shizen to bunka, II, 1951, 37–70.Google Scholar On the grain crops mentioned in classical texts see also Ch'êng Yao-t'ien Chiu-hu h'ao for the traditional view.

^{page 532 note 4} So Shinoda (loc. cit.) and Amano Motonosuke (Shizen to bunka, III). But Katō Shigeru considered that wheat was grown from very ancient times (Shina heizai-shi gaisetsu, 40).

^{page 532 note 5} See T‘ien-kung h'ai-wu (Yabuuchi text, pp. 408–10). See also Tenkō kaibutsu no kenkyū 1954, introduction, 49–52 (chapter on agriculture by Amano Motonosuke).

^{page 532 note 6} See Motonosuke, AmanoChin Fu no Nōsho to suitōsaku gijutsu no tankai Tōhō gahuhō , XIX (Kyōt series), 1950, 23–64.Google Scholar The first mention is to be found in the fragments of the Ssu-min yüeh-ling the Later Han author Ts‘ui Shih These are collected in the I-lan-l‘ang ts‘ung-shu.

^{page 533 note 1} See the excellent study of this document by Yamamoto Tatsurō ‘Tonkō hakken keiehō-yō bunsho zankan’ Tōyō gakuhō, XXXVII, 2, 1–60;Google Scholar XXXVII, 3, 83–98.

^{page 533 note 2} This is confirmed by T‘ang liu tien 3 (Konoe edition, p. 17a).

^{page 533 note 3} Ennin, Diary 2 (840, iv, 23).

^{page 533 note 4} Ennin, Diary 2 (839, i, 15).

^{page 534 note 1} See the illustration in Wang Chên's Nung shu 19 and the very similar one in T‘ien kung k'ai-wu 1 (Yabuuchi edition, p. 262). In its essentials the shui-tui represented the application of water-power to a simple treadle-operated device known as a t‘a-tui which is illustrated in Nung shu 16.

^{page 534 note 2} See the extremely important definition of nien and wei which is included in the commentary to Ryō-no-shūge the mid-ninth century Japanese digest of commentaries to the Statutes. This passage undoubtedly derives from a Chinese original earlier than the definition given in the present T‘ang-lü shu-i shih-wên and cited by the author. The passage reads as follows (see Takigawa and Miura edition, pp. 86–7):

Nien-wei has the meaning of a shui-tui The ones used for making hulled grain are called nien, and those used for making flour are called wei.

The Shaku says : ‘They are shui-tui. Those for making hulled grain are called nien… Their tui are made from stone, and their pestles ch'u from wood. Those for making flour are called wei … Their pestles and mortars are both made from stone’.

The tui in the Shaku's definition of the nien cannot be the pestle, for this is mentioned immediately following. Possibly tui is the name of the whole apparatus transferred to a part—in this case the wrong part, that is the mortar.

The Shaku, according to the introduction to Takigawa and Miura's edition, is probably the work entitled in the Honchō hōke monjo mokuroku But it is also a generice name for works giving definitions of legal terms. We know, for instance, of a Shin-ryō shaku a Ryōshaku kōki , etc.

^{page 534 note 3} For such a mill driven by animal power see T‘ien-kung k‘ai-wu (Yabuuchi edition, p. 263). Another type, entitled shih-nien , appears in Nung shu 16. In these mills there was a large horizontal circular stone acting as a fixed bed. The upper stone was a cylindrical stone rotating on a horizontal axle pivoted to an upright in the centre of the horizontal stone. Power was applied at the outer end of the axle in the case of the animal-driven mill. With water-driven mills of this type the central vertical axis was driven by a water-wheel, and this in turn moved the upper stone. See for this type of mill Nung shu 19 under the heading shui-nien An upper millstone rotating in a plane parallel to the lower one around a vertical axis was not, however, unknown. Such a millstone is perhaps envisaged by the late T‘ang fictional work Yao K‘un quoted in T‘ai-p‘ing kuang-chi 454. In this a brutal monk-bailiff flings the hero into a well and blocks the mouth with a millstone. He escapes, thanks to magical intervention, through the axle-hole. Clearly a flat, and not cylindrical stone is intended, and it was the rotating stone which was easily movable, the lower stone being a fixture.

^{page 535 note 1} See P. 2507, line 88. The same document (line 42) mentions the prohibition, in the neighbourhood of Loyang, of a type of mill, otherwise unknown, called a fou-wei This must have been mounted on an anchored raft bearing a water-wheel driven by the current in passing.

^{page 535 note 2} On this and other technical points concerning the milling of grain, I refer readers to an excellent account in Motonosuke's, Amano ‘Chūgoku no “usu” no rekishi’ Shizen to bunka, III, 1952, 21–58,Google Scholar where such querns are illustrated.

^{page 535 note 3} According to P. 2507, lines 52–3, it seems that there was water in the streams at Tun-huang only between the third and ninth months. Irrigation in northern China as a whole usually began in the fourth month. However, it is possible that an alternative reading of the passage in question should be preferred. I deal with this in a forthcoming study of this MS.

^{page 536 note 1} See Yoshiyuki, SudōChūgohu tochi-seido-shi kenkyū 1954, 41–3,Google Scholar etc.

^{page 536 note 2} This is not only pre-Buddhist, but may even be pre-Chinese. It is tempting to link it with the mountain cults so common throughout south-east Asia.

^{page 536 note 3} See Ch‘üan T‘ang wên 788, pp. 3b–5b, Li Pin Ch‘ing tzu ch‘u fêng-ct‘ien shou-shu Shanch'üan-ssu shih tsou The monastery was said to be the home of a dragon, and a propitious place to pray for rain.

^{page 536 note 4} See Chin-shih ts‘ui-pien 108, pp. 31b–37a, A-yü-wang-ssu ch‘ang-chu-t‘ien pei

^{page 536 note 5} See Chin-shih ts‘ui-pien 113, p. 51b, lines 1–4. We are told that apart from seven units of farmed land () this property was all either uncultivated or covered in brush ().

^{page 537 note 1} Ennin, Diary 2 (840, iv, 6). The monastery was the Li-ch‘üan ssu on Ch‘ang-pai shan.

^{page 537 note 2} See Chiu T‘ang shu 38, Hsin T‘ang shu 37.

^{page 537 note 3} See Yüan-ho chün-hsien u‘t chih 2.

^{page 537 note 4} T‘ai-p‘ing huan-yü chi 32 19,577 households.

Yüan-fêng chiu-yü chi 3 24,774 households.

Sung Shih 87 28,370 households.

^{page 537 note 5} An interesting contrast with the monasteries in the west, which also depended to a large extent on non-agrarian exploitation, but did so in spite of the diminished labour force involved.

^{page 537 note 6} See Shigeru's, Katō analysis of the Yüan-fêng chiu-yü chi figures in Shina Keizaishi hōshō 1953 (reprinted from Shigahu, XII, 3, 1933), 338–70.Google Scholar The figures for tenancy in Fêng-hsiang-fu also show a similar decline over the same period.

^{page 538 note 1} These are the Chih-tê-ssu p. 50a, line I, and the Kuang-shêng-ssu p. 49b, line 5.

^{page 538 note 2} See p. 50b, line 7, to p. 51a, line 6. Another Yang, Yang Ch‘ing , also appears in the inscription. Two other landowners who are mentioned several times are Fu I and Chang Te-ch‘êng

^{page 538 note 3} Yang Yen See Hsin T‘ang shu 145, Chiu T‘ang shu 118.

^{page 538 note 4} See p. 49a, line 7. Po Erh-chün who also occurs, must be a relative.

^{page 538 note 5} The inscription calls him Ying-ch‘uan chün Ch‘en-kung The editors of Chin-shih ts‘ui-pien have not identified him, but CTS 17 B makes it clear that he was in fact Ch‘ên Chün-i General of the Shên-ts‘ê army of the Left.

^{page 538 note 6} This is a very complicated problem in T‘ang times, owing to the very fragmentary sources. The position was theoretically that the produce of the ‘mountains and marshes’ belonged to the state, a dictum which was reiterated frequently throughout the dynasty. The responsible authority was the Prefect, and later the Salt and Iron Commission (after 780). During the ninth century Provincial Governors took a large part, but in 836 control again reverted (in theory) to the Prefects. In circa 851 control passed back to the Salt and Iron Commission. But in practice this governmental supervision seems to have been evaded on a wide scale.

^{page 539 note 1} See Michisaburō, MiyazakiTödai no cha-shō to hizen in Miyazaki sensei Hōsei-shi ronshū 1902,Google Scholar 148 ff., Koretada, Kanei ‘Tō no chahō’ Bunka, V, 8, 35–53.Google Scholar

^{page 539 note 2} See Act of Grace dated 821 in T‘ang ta chao-ling chi 10.

^{page 539 note 3} This was one of the policies put into force by Wang Ya and Chêng Chu in 835. See CTS 49, T‘ang hui yao 84, Ts‘ê-fu yüan-kuei 493, Ch‘üan T‘ang wên 541, etc. It lasted only two months, and it is doubtful how far it was ever implemented.

^{page 539 note 4} See Chiu T‘ang shu 49 and other sources quoted in note 3.

^{page 539 note 5} See Chiu T‘ang shu 49, Memorial of Chang P‘ang dated 793, and Memorial of P‘ei Hsiu dated 852.

^{page 539 note 6} Hsin T‘ang shu 40 under Chin-chou

^{page 539 note 7} Ennin, Diary 2 (840, iv, 6).

^{page 539 note 8} Chin-shih ts‘ui-pien 113, p. 48b.

^{page 539 note 9} loc. cit. The text (line 9) reads ‘A shop in Shan-ho ward in the Eastern Market, together 6 chien and a half’ It is possible that tien and shê represent separate buildings. Many prefectural towns had more than one market, so that it is quite probable that Lung-chou is in question.

^{page 540 note 1} See the ch‘uan-ch‘i story Liu-i chuan attributed to Li Ch‘ao-wei T‘ai-p‘ing kuang chi 491, i. See also the lists of tribute goods in Hsin T‘ang shu and T‘ai-p‘ing huan-yü chi.

^{page 540 note 2} See the Ta-ning ho by Fan Ch‘êng-ta in Shih-hu ch‘u-shih shih-chi 12.

^{page 540 note 3} One of the reasons for the dependence of the Chinese on horses supplied by the Uighurs in the period after An Lu-shan's rebellion (see Hsin T‘ang-shu 52) was the loss of their grazing grounds in Kansu and Shensi which had been ravaged by the Tibetans.

^{page 540 note 4} Ennin, Diary 2 (840, iv, 24). It is clear from the continuation of the passage that these flocks were grazed in very barren country where ‘They have had plagues of locusts the past three or four years, and the five grains have not matured’, and where ‘there is no rice or gruel, and they eat small beans () for food’. The Diary also tells us that the mountains near Wu-t‘ai shan were then well wooded.

^{page 540 note 5} This scheme is dated 783 in Chiu T‘ang shu 49 (Section on transport) but 782 later in the same chapter (Section on the tea tax). The earlier date is confirmed by Ts‘ê-fu yüan-kuei 502 and T‘ang hui-yao 88. Irregular taxes on bamboo, etc., were also levied in Ling-nan during the early ninth century. See the memorial from the Censorate dated 833 in Chiu T‘ang-shu 49, T‘ang-hui-yao 84, etc., and these were said to harm the people.

^{page 541 note 1} See P. 2507, lines 29–32.

^{page 541 note 2} See P. 2507, lines 125–6, and compare T'ang liu-tien 7 (Konoe edition, p. 28b).

^{page 541 note 3} See Yu-yang tsa-tsu (hsü-chi) 10.

^{page 541 note 4} Chin Tang shu 184.

^{page 541 note 5} See P. 2507, lines 112–19, line 142, and compare T'ang liu-tien 7 (Konoe edition, p. 28b).

^{page 541 note 6} See p. 17; first item. The artisans' charges are quite separate from the carpenters' salaries (item 13) for constructing the building, and the timbers were purchased already squared.

^{page 542 note 1} See Li Wên-Kung chi 9 , and cf. CTW 634.

^{page 542 note 2} HTS 52 gives the price as 3,200 cash. Han Yü quotes an even lower price of 3,000 cash from a memorial of Chang P'ing-shu . See Han Chang-li chi 40 .

^{page 542 note 3} According to CTS 49, CTS 123, THY 87, ‘In 779, the total annual revenue was some 12,000,000 strings of cash, and salt provided more than a half of this’. Yü Hai 181, p. 20a gives an even higher estimate of the income from salt (9,000,000 strings) which it quotes from the lost Li too yao chüeh of Tu Yu .

^{page 543 note 1} See THY 59.

^{page 543 note 2} See Pulleyblank, The background of the rebellion of An Lu-shan, 30–2, for an account of this.

^{page 543 note 3} See Okazaki Fumio , ‘Ubun YS no kakko-seisaku ni tsuite’ Shinagaku, n, 5, 42 ff.

^{page 543 note 4} See Suzuki Shun ‘Ubun Yu no kakko ni tsuite’ , in the volume of Tōyōshi Ronsō dedicated to Wada Sei, 1951, 319–44.

^{page 543 note 5} The only account which was sympathetic to Yü-wên Jung was that given by Tu Yu in T'ung tien. Yü-wên Jung was considered the first of the finance experts in the aristocratic interest who had attacked the position of the career-bureaucrats during Hsüan-tsung's reign. The majority of the historiographers were, of course, members of the latter group.

^{page 543 note 6} See Pulleyblank, , ‘The political background of An Lu-shan's rising’, Tōyō galcukō, xxxv, 2, 1952, 98–108.Google Scholar

^{page 543 note 7} A good account of the great relative importance of local and other labour dues is to be found in Ichisada, Miyazaki, ‘Tōdai fueki-seido shinkō’ Tōyōshi kenhyū, xiv, 4, 1956, 1–24.Google Scholar Exemption payments for persons selected for special duties () were very high indeed. Miyazaki cites some of 2,000 cash per annum, and other payments in this range are specified in the Shui-pu shih fragment, P. 2507.

^{page 544 note 1} Balázs, , ‘Beiträge zur Wirtschaftsgeschichte der T'ang-zeit’, MSOS, xxxiv, 1931, 16, n. 27.Google Scholar

^{page 544 note 2} The chu-hu and k'o-hu of HTS 52 and of the accounts of the liang-shui tax reform given in CTS 48, THY 83, TFYK 488, etc., have not yet, it seems, quite such a specific and specialized meaning as they had during the Wu-tai and Sung, when they designated landowners and tenants.

^{page 544 note 3} Shih T'ung edition, p. 41c.

^{page 544 note 4} We know from the Tzu-chih t'ung-chien k'ao-i that one of the two series, the Chien-chung shih-lu , was edited by Shên Chi-chi , a partisan of Yang Yen. (See K'ao-i 17, p. 9a.) The Tê tsung shih-lu later duplicated this work, but was less favourable to Yang Yen. The Chiu T'ang shu seems to have based itself principally upon this work. It might be suggested that the Chien-chung shih-lu contained the larger figures and that the figure given in THY 84, TFYK 488, etc., derives from the Tê tsung shih-lu, whose editor is attempting to denigrate Yang Yen. The figure quoted by Tu Yu cannot, however, reflect the Tê tsung shih-lu, as the T'ung tien was already completed in 801 while the Tê tsung shih-lu was not finished and presented to the throne until 810. In any case Tu Yu's figures are even smaller than those of THY, TFYK, etc.

The alternative explanation of a textual corruption would explain that the 1 of 1,800,000 was corrupted into 3 possibly under the influence of the similar number immediately preceding it. In the second case perhaps , was garbled into dropping out.

^{page 545 note 1} This still does not explain how the author came to take them as taxpayers. Balázs, loc. cit., correctly renders them ‘ortsansässige t'u-hu’.

^{page 545 note 2} See TT 6. Balázs made a curious slip in dealing with these figures, in that he took all nontaxpayers to be members of untaxed households. This is, of course, quite untrue. All households contained some untaxable members, women, children, old persons, etc. See MSOS, xxxiv, 1931, 15.Google Scholar

^{page 545 note 3} Under lax conditions of registration it was comparatively easy to omit individuals, or even to enter them in the tax lists in a wrong personal category. But it was much more difficult, without wholesale connivance, to omit a whole family. Moreover it was often a disadvantage for a household to be completely unregistered, since it had no longer any legal standing, nor any title to its lands.

^{page 546 note 1} See TLT 30 (Konoe edition, p. 22b). Tao fo occurs as one item, with no further details, in a long list of the responsibilities of the Kung-ts'ao ts'an-chun-shih. This item was omitted by the compilers of HTS 49B which would suggest that little importance was attached to the duty, but des Rotours, Traité des fontionnaires et de I'armeée, 691, n. 4, suggests that all mention of Buddhism and Taoism was systematically omitted by the compilers of HTS. The mention of Tao fo also occurs in CTS 44.

^{page 547 note 1} On this region during the period in question there is a superficial preliminary survey by Michio, Tanigawa entitled ‘Tōdai no hanshin ni tsuite’ in Shirin, xxxv, 3, 1952, 70–89.Google Scholar

^{page 547 note 2} The reading ling of Li Wei-kung Hui-ck'ang i-p'in chi is clearly superior to chin in CTS 174.

^{page 547 note 3} Suan-shan was on the south bank of the Yangtse 9 li west of Tan-t'u See YHCHTC 25. The crossing is frequently mentioned in Sung texts, as a principal route to Hangchou from northern China.

^{page 548 note 1} See Ennin, Diary 4 (845, iii, 3 ff.).