¹ Definitions of slavery vary considerably, but because forced labour as a form of punishment for prisoners is rarely included, I prefer where possible to use the term ‘bondage’ instead of ‘slave’, ‘serf’, ‘bondsman’ or even ‘servant’, which are heavy with cultural connotations and imprecision in all languages. I am, however, guided by Drescher's definition of slavery: ‘The most crucial and frequently utilized aspect of the condition is a communally recognized right by some individuals to possess, buy, sell, discipline, transport, liberate, or otherwise dispose of the bodies and behavior of other individuals’: see Drescher, Seymour, Abolition: A history of slavery and anti-slavery (Cambridge: Cambridge University Press, 2009), pp. 4–5CrossRefGoogle Scholar. Unless explicitly stated, I do not include in this discussion women ‘given away’ as wives or concubines or young people ‘given’ as apprentices to craftsmen.

² First Historical Archives, Beijing (hereafter FHA), Manwen dang'an 3495–041 159–3130 and 3495–042 159–3134, Qianlong 60.2.25 (15 March 1795). For further discussion of ‘Ghalcha’, see below.

³ Karun was a Manchu term for guard posts. Located at strategic points in the border zone, they served as checkpoints and marked a notional frontier. For reference to the prohibition and the earlier incident in the forty-fourth year of Qianlong (1779/80), see 吳 豐 培, Wu Fengpei (comp.), Songyun Xinjiang zougao 松 筠 新 疆 奏 稿, Mimeographed reproduction (Beijing: Zhongyang minzu tushuguan, 1980), 6aGoogle Scholar.

⁴ FHA Manwen dang'an 3495–042 159–3134, Qianlong 60.2.25 (15 March 1795); see also Da Qing Gaozong huangdi shilu大 凊 高 宗 (乾 隆)皇 帝 實 錄 (hereafter GZSL),1473:19b–20a, 60.run 闰 2.27 (16 April 1795).

⁵ Hsiao-ting, Lin, ‘The Tribute System in China's Historical Imagination’, Journal of the Royal Asiatic Society, 19, 2009, p. 494Google Scholar.

⁶ Murav'ev, Nikolai, Journey to Khiva Through the Turkoman Country, 1819–20 (1871; reprinted London: Oguz Press, 1977), p. 58Google Scholar; Schuyler, Eugene, Turkistan: Notes of a Journey in Russian Turkistan, Khokand, Bukhara, and Kuldja (London: Sampson Low, Marston, Searle and Rivington, 1876), Vol. 2, p. 354Google Scholar. Schuyler gives a vivid account of his visit to a slave market in Bukhara in the early 1870s where he endeavoured to purchase a young boy in order to convince the Russian authorities that the trade had not been halted, despite their assertions to the contrary. See Schuyler, Turkistan, pp. 100–09.

⁷ There were no slave armies in Central Asia akin to the Mamluks of the Abbasid Caliphs or the Janissaries of the Ottoman empire, but the soldiers of conquered peoples were regularly absorbed into the victor's fighting force. For a brief discussion of military slavery in the Islamic world, see Clarence-Smith, William, Islam and the Abolition of Slavery (London: C. Hurst, 2003), pp. 86–88Google Scholar.

⁸ See, for example, Vámbéry, Arminius, History of Bokhara (London: Henry S. King, 1873), p. 221Google Scholar. For the dispatch of Chinese slaves together with an elephant as presents from the khan of Khoqand to the tsar in St Petersburg in the early nineteenth century, see Wathan, W. H., ‘Memoir on the U'sbek State of Kokan, properly called Khokend (the ancient Ferghana) in Central Asia’, Journal of the Asiatic Society of Bengal, 3 (32), August 1834, p. 375Google Scholar.

⁹ Morrison, Alexander, Russian Rule in Samarkand 1868–1910 (Oxford: Oxford University Press, 2008), p. 43CrossRefGoogle Scholar.

¹⁰ Vámbéry, History, pp. 220–21.

¹¹ See, for example, Murav'ev's account of the Russian slave boy, called ‘David’ in the English translation: Murav'ev, Journey to Khiva, pp. 57–59.

¹² British Agent in Kabul to the Secretary to the Government of India, British Library, India Office, London (henceforth BLIO), L/P&S/7/78, 5 February 1895.

¹³ Azumah, John, Islam and Slavery (Northwood: London Bible College, 1999), p. 3Google Scholar. According to Hopkins, however, religion rather than race was the main criterion in Central Asia: see Hopkins, B. D, ‘Race, Sex and Slavery: “Forced labour” in Central Asia and Afghanistan in the early 19^th century’, Modern Asian Studies, 42 (4), 2008, p. 641CrossRefGoogle Scholar.

¹⁴ For the religious mythology surrounding the Islamic idea that the enslavement of dark-skinned people is natural, see Davis, David B., Slavery and Human Progress (New York: Oxford University Press, 1984), pp. 32–51Google Scholar. Nineteenth-century travel accounts are peppered with references to Russian and Chinese slaves held in bondage throughout Central Asia. See, for example, note 11 above and Burnes, Alexander, Travels into Bokhara (London: John Murray, 1834), Vol. 2, p. 274Google Scholar.

¹⁵ For testimony of the trade in Tajik children, but also men from Kanjut and Chitral, for sale in the Bukhara market, see Forsyth, T. Douglas (ed.), Report of a Mission to Yarkund in 1873 (Calcutta: Foreign Department Press, 1875), p. 56Google Scholar. For evidence that Shi'ites were regarded as unbelievers (kaffir) rather than heretics, see Schuyler, Turkistan, p. 102.

¹⁶ Zuhūr al-Dīn was a descendent of Emīn Khoja of Turfan; see below.

¹⁷ A descendant of the Turkic Muslim Khoja brothers, who claimed authority to rule southern Xinjiang and were displaced by Qing forces, Jahāngīr made several attempts to oust the Qing from the region in the early nineteenth century.

¹⁸ Valikhanov notes erroneously that Zuhūr al-Dīn was removed to Khoqand in 1830. Valikhanov, Ch. Ch., ‘O sostoianii Altyshara ili shesti vostochnïkh gorodov Kitayskoy provintsii Nan-lu (Maloy Bukharii) v 1858–9 godakh’, in Sobranie sochineniĭ (Alma-Ata: AN Kaz SSR, 1962), Vol. 2, p. 326Google Scholar. In fact, according to Chinese documentary sources, he was captured in 1826. See Da Qing Xuanzong huangdi shilu大 凊 宣 宗(道 光)皇 帝 實 錄 (hereafter XZSL), 128: 2b–3b, 7.10.16 (4 December 1827); and Zuhūr al-Dīn's deposition made on his return, National Palace Museum Archive (Taipei), no. 058525. I am grateful to David Brophy for the latter reference.

¹⁹ Clarence-Smith, Islam, p. 2.

²⁰ Shaw, Robert B., ‘A Sketch of the Turki Language’, Journal of the Asiatic Society of Bengal, 47 (1), 1878, p.137Google Scholar; and Jarring, Gunnar, Materials to the Knowledge of Eastern Turki (Lund: C. W. K. Gleerup, 1946–1951), Vol. 1, p. 121Google Scholar, note 1.

²¹ George Macartney, the British representative in Kashgar, referred to these people by the name ‘Toghmak’. BLIO L/P&S/7/77, Macartney to the Government of India (4 August 1894). I am grateful to Ilidkó Bellér-Hann for suggesting that this word appears to be ‘tughmaq’, derived from ‘to bear a child’ (or ‘tuğ’ to be born) and therefore possibly denoting someone who was born as a slave (‘qul’).

²² Valikhanov notes that in the 1850s the outskirts of Yarkand were well known for the numbers of liberated Mountain Tajiks from Chitral and Wakhan who dwelled there; see Valikhanov, ‘O sostoianii Altyshara’, p. 295.

²³ The freeing of individual slaves was considered an act of piety and is recognized as having been widespread in Islamic societies. See Segal, Ronald, Islam's Black Slaves: The history of Islam's other black diaspora (London: Atlantic Books, 2001), p. 9Google Scholar.

²⁴ The role of akhund afforded considerable social standing within the local community and, as with other professions, tended to pass from father to son.

²⁵ Khōja-Jahān and his older brother, Burhān al-Dīn, assumed power in Altishahr after the fall of the Junghar Khanate, but after they failed to throw off the Qing mantle, they fled from the advancing Qing forces in 1759.

²⁶ Songyun Xinjiang zougao, 6a–b; FHA Zhupi zouzhe, Minzu lei, 635.1, Jiaqing 20.10.3 (3.11.1815). As a Badakhshani, Ziyā’ al-Dīn's grandfather would almost undoubtedly have been a Mountain Tajik.

²⁷ The number of children cited here is so large and the cost of keeping them fed and housed would have been so great that, unless Grenard was mistaken, this must still have been a thriving trade. Alternatively, the seller may have functioned as a middleman, in which case it may be conjectured that these were local children whose parents had agreed to their ‘sale’ (temporary or otherwise) when and if a suitable purchaser was found. Grenard, Fernand (J.-L. Detreuil de Rhins), Mission scientifique dans la Haute Asie 1890–5 (Paris: Ernest Leroux, 1898), Vol. 2, p. 167Google Scholar.

²⁸ In 1827 Qing officials were prohibited from taking local children into their households. Millward, James, Beyond the Pass: Economy, ethnicity, and empire in Qing Central Asia, 1759–1864 (Stanford: Stanford University Press, 1999), p. 305Google Scholar, note 28.

²⁹ Millward cites the case of the Manchu officer, De-hui, who in 1830 took a Turkic ‘servant’ back to the interior of China with him from Khotan. Although Millward refers to the boy as a servant, it is quite clear that he was not free to choose his own master and, possibly because it was forbidden for officers to purchase or take in local children after 1827, he had been raised as a Chinese. See Millward, Beyond the Pass, pp. 305–06, note 28.

³⁰ Grenard, Mission scientifique, p. 167. I know of no extant contract for the purchase of men, women or children in Xinjiang, but the fact that they existed, at least by the end of the nineteenth century, is clearly shown both by Grenard and Macartney (see below).

³¹ In theory, it would also have meant that children born to them while they were still in servitude were free. Writing in the 1920s, Roerich notes that children were still being sold to ‘good families’, but he gives no further details: see Roerich, George N., Trails to Innermost Asia: Five years of exploration with the Roerich Central Asian expedition (New Haven: Yale University Press, 1931), p. 64Google Scholar.

³² I am grateful to Ayxem Eli of the University of Tasmania for information about her current research on this practice, which persisted into the twentieth century. It is important to note that, despite the relatively smaller numbers involved, the trade in children extended to Mongols as well. See FHS Manwen dang'an 2677–042 108–2901, Qianlong 41.2.27 (15 April 1776).

³³ BLIO L/P&S/7/77, Macartney to the Government of India (9 August 1894). Generally speaking, serfs had rights to some land and could not be sold. See Bush, M. L. (ed.), Serfdom and Slavery: Studies in legal bondage (London: Longman, 1996), p. 1Google Scholar.

³⁴ See Elliott, Mark, The Manchu Way: The Eight Banners and ethnic identity in late imperial China (Stanford: Stanford University Press, 2001), p. 51Google Scholar.

³⁵ Banners were the administrative and military divisions into which Manchus were divided.

³⁶ The Chinese terms nupu 奴 僕, jia ren 家 人 or huxia ren 户 下 人, etc. were similarly considered inadequate, hence the tendency to use a transliteration in Chinese: baoyi 包 衣 baoyi aha 包 衣 啊 哈 (jia de nupu 家 的 奴 僕). See Elliott, The Manchu Way, p. 82.

³⁷ Xiaomeng 刘小萌, Liu, Qingdai Beijing qiren shehui 清 代 北 京 旗 人 社 会 (Beijing: Zhongguo shehui kexue chubanshe, 2008), pp. 450, 507–08, 642Google Scholar.

³⁸ See Li Zilong, who separates the various ways in which people were enslaved in the early Qing into four categories: those who had been taken captive, those who worked the land for Manchus and had lost their tenancies when land was distributed to the bannermen, those who were punished with servitude under the Qing criminal code, and those who were legally purchased. Zilong 李 子 龙, Li, ‘Qing chu taoren wenti chu tan’, 清 初 逃 人 问 题 初 探 Jiangsu shehui kexue, 2, 1998, pp. 124–25Google Scholar.

³⁹ See, for example, Cable, Mildred and French, Francesca, The Gobi Desert (London: Hodder and Stoughton, 1942), pp. 132–45Google Scholar.

⁴⁰ Beihai 苏 北 海, Su and Jianhua 黄 建 华, Huang, Hami Tulufan Weiwuer wang lishi 哈 密 ， 吐 鲁 番 维 吾 尔 王 历 史 (Urumqi: Xinjiang daxue chubanshe, 1993), pp. 117–20Google Scholar, 223–24. For Qing dependence on the elite families of Hami and Turfan, see also Newby, Laura, ‘The Begs of Xinjiang: Between two worlds’, Bulletin of the School of Oriental and African Studies, 61 (2), 1998, pp. 286–87CrossRefGoogle Scholar, 289.

⁴¹ These were the people whom Toqtā Mahmud had led to safety within the pass of Jiayuguan in 1725. See Newby, ‘The Begs of Xinjiang’, p. 286; GZSL 544: 12b–13a, 22.8.4 (16 September 1757); Perdue, Peter, China Marches West: The Qing conquest of Central Eurasia (Cambridge, Massachusetts: Harvard University Press, 2005), pp. 331–33Google Scholar.

⁴² GZSL 619: 7a, 25.8.3 (1 October 1760).

⁴³ GZSL 653: 2b–3a, 27.1.17 (10 February 1762).

⁴⁴ Another 162 women and girls apparently went ‘voluntarily’, which may of course have meant that their impoverished families ‘donated’ them. Bawden, Charles, The Modern History of Mongolia (London: Weidenfeld and Nicolson, 1968), pp. 138–39Google Scholar.

⁴⁵ GZSL 749: 10a, 30.11.23 (3 January 1766).

⁴⁶ I have found no evidence that the term was ever applied to women.

⁴⁷ Clauson, Gerard, An Etymological Dictionary of Pre-thirteenth-century Turkish (Oxford: Clarendon Press, 1972), p. 173Google Scholar.

⁴⁸ Xu Jianying suggests that the reason slavery existed in southern Xinjiang was because it was closely connected with the bek system and that those who tilled the inchu land were serfs (nongnu 农 奴) who were ‘dependent on their feudal masters’. Jianying 许 坚 英, Xu, ‘Qingmo Xinjiang Yingnu wenti ji qi jiejue’ 清 末 新 疆 英 奴 问 题 及 其 解 决, Xiyu yanjiu, 3, 2003, pp. 54–55Google Scholar. However, most modern, and some Qing, Chinese sources refer to the inchu as agriculturalists (zhong di ren 種 地 人) or tenant farmers (tian hu佃 户).

⁴⁹ Pusheng 苗 普 生, Miao (ed.), Boke zhidu 伯 克 制 度 (Urumqi: Xinjiang renmin chubanshe, 1995), p. 50Google Scholar. For an overview of the bek system and various functions of the bek under the Qing, see Tōru, Saguchi, 18–19 seiki Higashi Torukisutan shakaishi kenkyū (Tokyo: Yoshikawa kōbunkan, 1963), pp. 103–92Google Scholar.

⁵⁰ 1 batman = 5 dan 石3 dou斗 = approximately 53 mu 畝 (100 mu is approximately 0.16 acre).

⁵¹ Compare, for example, the figures given in 蘇 爾 德, Suerdeet al. (comps), Huijiang zhi 回 疆 志, (1772; reprinted Taibei: Chengwen, 1968), pp. 178–99Google Scholar, and 傅 恒, Fuhenget al. (comps), Qinding huangyu Xiyu tuzhi 欽 定 皇 輿 西 域 圖 志 (1782; reprinted Beijing: Zhongyang minzu xueyuan tushuguan, 1986), 30: 35aGoogle Scholar. See Newby, ‘The Begs of Xinjiang’, p. 290. According to the Huijiang zhi, a rough estimate of the number of inchu in Altishahr allocated by the Qing authorities would suggest approximately 3,200.

⁵² To date I have been unable to find any information on how the original process of allocating inchu to beks was carried out. See Duman for a useful introduction to the ‘institution’. See Duman, L. I., ‘Feodal'nyi iantsii v. Vostochnom Turkestane v XVIII veke’ in Zapiski Instituta Vostokovedenia Akademii Nauk, SSSR 3, 1935, pp. 87–89Google Scholar.

⁵³ For the difficulties that this regulation caused and its relaxation, see Newby, ‘The Begs of Xinjiang’, pp. 289–90. In the early years after the conquest, in particular, there was a rapid turnover among the higher ranking inchu.

⁵⁴ 托 津, Tuojinet al. (comps), Qinding Huijiang zeli 欽 定 回 疆 則 例, in 呂 一 燃, Lü Yiranet al. (eds), Menggu lüli. Huijiang zeli, 蒙 古 律 例.回 疆 則 例 (1842; reprinted Lanzhou: Quanguo tushuguan wenxian suowei fuzhi zhongxin, 1988), 7: 20a–bGoogle Scholar.

⁵⁵ For the special status of servants and slaves in the Qing code, see Bodde, Derk, Law in Imperial China: Exemplified by 190 Ch'ing dynasty cases (translated from the Hsing-an hui-lan), with historical, social and juridical commentaries (Cambridge, Massachusetts: Harvard University Press, 1967), pp. 172CrossRefGoogle Scholar, 242.

⁵⁶ Family members could accompany convicts voluntarily but others were sentenced to servitude in exile because of their association with the criminal. For a discussion of collective responsibility under the Qing, see Waley-Cohen, Joanna, Exile in Mid-Qing China: Banishment to Xinjiang, 1758–1820 (New Haven: Yale University Press, 1991), pp. 221–22CrossRefGoogle Scholar.

⁵⁷ GZSL 1093: 7a, 44.10.19 (26 November 1779).

⁵⁸ See the examples given by 魏 庆 远, Wei Qingyuanet al., Qingdai nubi zhidu 清 代 奴 婢 制 度 (Beijing: Renmin chubanshe, 1984), p. 71Google Scholar, note 1.

⁵⁹ Several commentators have drawn a distinction between ‘public’ or ‘official’ slavery and ‘private’ slavery under the Qing. See, for example, Meijer, Marinus, ‘Slavery at the End of the Ch'ing Dynasty’, in Cohen, Jeromeet al. (eds), Essays on China's Legal Tradition (Princeton: Princeton University Press, 1980), p. 328Google Scholar; Waley-Cohen, Exile in Mid-Qing China, p. 30; and Schottenhammer, Angela, ‘Slaves and Forms of Slavery in Late Imperial China (seventeenth to early twentieth centuries)’, Slavery and abolition, 24 (2), 2003, p. 146CrossRefGoogle Scholar.

⁶⁰ GZSL 1353: 40b, 55.4.29 (11 July 1790).

⁶¹ Wei Qingyuan, Qingdai nubi zhidu, pp. 70–71.

⁶² Xing'an huilan 刑 案 匯 覽 (reprinted Taipei: Chengwen chubanshe, 1968), 6: 12a–b (13 April 1813).

⁶³ Qi Qingshun demonstrates how severe their living conditions were in comparison to those in the military; see 齐 清 顺, Qi Qingshun, ‘Yiqiliuqinian Changji ji qianfan baodong bu ying kending’ 一 七 六 七 年 昌 及 遣 犯 暴 动 不 应 肯 定, Xinjiang daxue xuebao, 4, 1986, pp. 63–65Google Scholar. Note that Qi makes the common mistake of dating the rising to 1767 instead of 1768.

⁶⁴ 紀 昀, Ji Yun, ‘Wulumuqi zaji’ 烏 魯 木 齊 雜 記 in Yingqi, Wang (comp.), Xiaofanghuzhai yudi congcha (Shanghai: 1891–97), 2: 121bGoogle Scholar.

⁶⁵ Waley-Cohen, Exile in Mid-Qing China, p. 31.

⁶⁶ See Qi Qingshun 齐 清 顺, ‘Qingdai Xinjiang qianfan yanjiu’ 清 代 新 疆 遣 犯 研 究, Qingdai shi yanjiu, 2, 1988, pp. 84–85； GZSL 1382: 23b–27b, 56.7.13 (12 August 1791).

⁶⁷ GZSL 1382: 23b–27b, 56.11.13 (8 December 1791).

⁶⁸ GZSL 1384: 14a–17b, 56.8.8 (5 September 1791); Waley-Cohen, Exile in Mid-Qing China, pp. 164–66.

⁶⁹ Wei Qingyuan, Qingdai nubi zhidu, p. 63.

⁷⁰ Qi Qingshun, ‘Qingdai Xinjiang qianfan yanjiu’, p. 88; GZSL 1276: 10b–11a, 52.3.6 (23 April 1787); Waley-Cohen, Exile in Mid-Qing China, p. 167.

⁷¹ Xing'an huilan, 6: 20a–b, 30 October 1828; XZSL 143: 16b–17b, 8.9.22 (30 October 1828). See Qi Qingshun, ‘Qingdai Xinjiang qianfan yanjiu’, p. 88.

⁷² Although local law was still applied in Xinjiang for minor criminal cases, more serious crimes, or those involving Han and Manchu soldiers or civilians, were dealt with under Qing law.

⁷³ FHA Minzu lufu 655.15, Daoguang 21.3.26 (17 April 1841). The Turkic Muslim name Adir is a hypothetical reconstruction from the Chinese A-di-er. The other names in this account, Turkic and Manchu, are attested in non-Chinese sources.

⁷⁴ MacGowan, D. J., ‘On the Banishment of Criminals in China’, Journal of the North China Branch of the Royal Asiatic Society, 3, December 1859, p. 295Google Scholar; Peking Gazette, 19 June 1857.

⁷⁵ Kim, Hodong, Holy War in China: The Muslim rebellion and state in Chinese Central Asia, 1864–1877 (Stanford: Stanford University Press, 2004), pp. 129–31, 133–34Google Scholar.

⁷⁶ For use of Chinese labour under Ya'qūb bek, see Henry Bellew, ‘History of Kashgar’ in Forsyth, Report of a Mission to Yarkund, pp. 209–10.

⁷⁷ Not only does Islam regard manumission as a pious act, but enslavement as a punishment for crime or as a result of debt contravenes Islamic law. See Davis, Slavery and Human Progress, p. 39; and Clarence-Smith, ‘The British “Official Mind” and Nineteenth-century Islamic Debates over the Abolition of Slavery’, in Hamilton, Keith and Salmon, Patrick (eds), Slavery, Diplomacy and the Empire (Brighton: Sussex Academic Press, 2009), pp. 131–32Google Scholar. For indications that the slave trade and hard labour were curtailed under Ya'qūb bek, see Forsyth, Report of a Mission to Yarkund, p. 56; Shaw, Robert B., Visit to High Tartary, Yarkand and Kashgar (formerly Chinese Tartary), and Return (London: John Murray, 1871), p. 469Google Scholar.

⁷⁸ BLIO L/P&S/7/98, Macartney to the Government of India enclosing a proclamation issued by the District Magistrate of Yarkand (April 1897).

⁷⁹ Millward, Beyond the Pass, p. 305.

⁸⁰ The Shanghai riots of 1906, for example, were triggered as a result of a dispute over whether two women suspected of kidnapping and trafficking young girls should be held in the Mixed Court gaol or the Municipal gaol. See ‘Report on incident at the mixed court Friday morning 8 December’, Public Records Office (London) FO 228/2510 (December 1905).

⁸¹ Clarence-Smith, ‘The British “Official Mind”’, p. 127; and Keith Hamilton and Farida Shaikh, ‘Introduction’, in Hamilton and Salmon, Slavery, Diplomacy, p. 4.

⁸² Hopkirk, Peter, The Great Game: On secret service in High Asia (London: John Murray, 1990), p. 205Google Scholar.

⁸³ Significantly, Macartney did not take up the case of the slaves from Chitral. BLIO L/P&S/7/75 (1 February 1895).

⁸⁴ Xu Jianying suggests that the reason the Qing authorities were so willing to release the British slaves was because if there were no British residents in Xinjiang and no British trade, then there would be no need for a British consulate there. 许 建 英, Xu Jianying, Jindai Yingguo he Zhongguo Xinjiang (1840–1911) 近 代 英 国 和 中 国 新 疆 (1840–1911) (Harbin: Heilongjiang chubanshe, 2004), p. 247Google Scholar.

⁸⁵ BLIO L/P&S/7/77, Macartney to the Government of India (9 August 1894 and September 1894).

⁸⁶ See, for example, BLIO L/P&S/7/98, Macartney to the Government of India (23 December 1897).

⁸⁷ Grenard, Mission scientifique, Vol. 2, p. 166.

⁸⁸ Grenard, Mission scientifique, Vol. 2, pp. 166–67.

⁸⁹ BLIO L/P&S/7/98, Macartney to the Government of India (23 December 1897); Xu Jianying, Jindai Yingguo, p. 257.

⁹⁰ 吳 豐 培, Wu Fengpei (comp.), Tao qinsu gong zouyi yigao 陶 勤 肅 公 奏 議 遺 稿 (Beijing: Quanguo tushuguan wenxian suowei fuzhi zhongxin, 1987), 3: 22a–23aGoogle Scholar, 31 August 1897; BLIO L/P&S/7/75 Macartney to the Government of India (1 February 1894).

⁹¹ BLIO L/P&S/7/75, Macartney to the Government of India citing a letter from the District Magistrate of Yarkand to the Circuit Intendant at Kashgar (1 February 1894).

⁹² BLIO L/P&S/7/75, Macartney to the Government of India (1 February 1894).

⁹³ Interestingly, there is no mention of Xinjiang in the proposals for the abolition of slavery throughout China that culminated in the recommendations of the commissioner of Law Revision, Shen Jiaben, and their endorsement by the court in January 1910. See Meijer, ‘Slavery’, pp. 338–52.

⁹⁴ Roerich also notes that children were being transported to Turkestan to be sold and others became slaves through not being able to pay their debts. See Roerich, Trails to Innermost Asia, p. 64.