1. For ongoing discussions of the likely historicity of the Xia dynasty and the identification of its principal capital city, see Zhongshu, Xu 徐中舒, ed., Xia wenhua lunwen xuanji 夏文化論文選集 (Zhengzhou: Zhongzhou guji chubanshe, 1985)Google Scholar; Huber, Louisa G. Fitzgerald, “The Bo Capital and Questions Concerning Xia and Early Shang,” Early China 13 (1988), 46–77CrossRefGoogle Scholar; Thorp, Robert L., “Erlitou and the Search for the Xia,” Early China 16 (1991), 1–38CrossRefGoogle Scholar; and more recently a series of articles by Jinhuai, An 安金槐, Mingxiang, Pei 裴明相 and others in Zhongyuan wenwu 中原文杨(1993.3), 1–47Google Scholar.

2. See Henau chutu Shang Zhou qingtongqi 河南出土商周靑銅器(Beijing: Wenwu chubanshe, 1981)Google Scholar, nos. 28, 52 (Zhengzhou); nos. 50-51 (Kaifeng); no. 11 (Shangqiu). As indicated by ceramic inventories, the Erlitou sphere of influence reached as far north as Inner Mongolia (see below) and southwest to the Chengdu area of Sichuan, where at the Guanghan 廣漢 site of Sanxingdui 三星堆 we find attenuated versions of the kui 菜and dou 豆 based upon Erlitou counterparts (Kaogu xuebao 考古學報 1987.2, 240Google Scholar, fig. 13: 6, 9, 16, 17, 21; pl. 10: 5-7). In the Shang period the Guanghan region became a center for the production of bronze and jades, both of which continue to reflect some degree of influence from the North China Plain (Wenwu 文物 1987.10, 1–15Google Scholar; pls. 1-5; Wenwu 1989.5, 1–19Google Scholar; pls. 1-5.

3. Kaogu xuebao 1982.4, 454Google Scholar, fig. 22:1

4. Kaogu xuebao 1990.1, 59Google Scholar, fig. 12; see p. 67 for references to microlithic implements from other sites in the Anyang area, including Hougang and Tangyin, as well as Handan in southern Hebei. The pottery from the Longshan level at Dahancun combines utilitarian corded ware vessels typical of western Henan and adjacent southeast Shan-xi with ‘serving’ vessels of entirely different shapes having polished dark gray or black surfaces comparable in style to those of the late Shandong Longshan tradition (compare the Xi illustrated in Kaogu xuebao 1990.1, pl. 2: 8 with Miaodigou yu Sanliqiao 廟底溝與三里橋 [Beijing,: Kexue chubanshe, 1959]Google Scholar, pl. 81: 3 [Sanliqiao II] and Kaogu 考古1986.2, Ï03, fig. 8: 2 [Longwangyai龍王崖]; compare the drinking and pouring vessels on pl. 2: 3, 6 with Kaogu 1990.7, 593Google Scholar, fig. 4:12 [Linqu, 臨胸]: and Zouxian Yedian 鄒縣野店 [Beijing: Wenwu chubanshe, 1985]Google Scholar, pl. 80:1).

5. Kaogu 1992.4: 304–309Google Scholar; pls. 3-4. Available carbon-14 dates for Dadianzi range from 1735 B.C ± 135 (ZK 480) to 1695 B.C.± 135 (ZK 402) (Zhongguo kaogu xue zhong tan shisi niandai shujuji: 1965-1981 中國考古學中碳十四代數據集 [Beijing: Wenwu chubanshe, 1983], 25)Google Scholar.

6. See also, Kaogu 1992.4, pl. 3: 5, 6. The two vessels were previously published in Xin Zhongguo de kaogu faxian he yanjiu 新中國的考古發現扣硏究 (Beijing: Wenwu chubanshe, 1984), 341Google Scholar, fig. 85: 12, 13. For Erlitou II counterparts to the Dadianzi M 677 kui, see Kaogu 1992.4, 297Google Scholar, fig. 4: 3; Kaogu 1985.12, 1090Google Scholar, fig. 5: 15; pl. 3: 5; for counterparts to the jue, see, p. 1090, fig. 5:19. The only actual metalwork recovered at Dadian-zi is a copper finger-ring from M 677 (Kaogu 1992.4, 307Google Scholar, fig. 4:1).

The incised geometric patterns seen on the Dadianzi vessels exhibiting sheet-metal traits differ markedly from the curvilinear designs painted on the other types of pot-teiy. These geometric patterns are reminiscent of those incised on Andronovo pottery and may reflect the influence of northern nomadic groups (compare Kuzmina, E.E., Drevneischie skotovodi ot Urala do Tyan-Shanya [Frunze: Ilim, 1986]Google Scholar, fig. 19, following p. 80). Somewhat similar striated patterns occasionally appear at Erlitou on the pottery jue and like, but they are not seen on other vessel types (Kaogu 1983.3, 203Google Scholar, fig. 9: 1, 2, 8). See nn. 89 and 106 below.

7. See p. 63 and n. 105; Huber, Louisa G. Fitzgerald, “The Relationship of the Painted Pottery and Lung-shan Cultures,” in Keightley, David N., ed., The Origins of Chinese Civilization (Berkeley: University of California Press, 1983), 207–209Google Scholar. The most vocal proponent for the existence of a sheet-metal tradition has been Bagley, Robert W., Shang Ritual Bronzes in the Arthur M. Sackler Collection (Washington, D.C.: The Arthur M. Sadder Foundation, 1987), 15–16Google Scholar. For an opposing view, compare Barnard, Noel, “Wrought Metal Working Prior to Middle Shang (?) – A Problem in Archaeological and Art-Historical Research Approaches,” Early China 6 (1980–1981), 4–30CrossRefGoogle Scholar.

8. Kaogu 1992.4, pl. 4:1 (color). Debased versions or the same sort of painted decoration are seen on pottery recovered from earlier excavations at Dadianzi (Kaogu 1975.2, 101Google Scholar, fig. 3: 1, 2, 4, 5; pls. 6, 7 [color]). For painted ceramics from Fengxia, see Kaogu 1976.3, 206Google Scholar, fig. 12: 3; 208, fig. 17.

The route of transmission between the Erlitou culture and the eastern Inner Mongolian sites is unknown. It should be mentioned, however, that painted designs comparable to those at Dadianzi also appear at contemporaneous sites of the Yueshi 岳石 culture, which replaced the classical Longshan 龍山 phase in central Shandong. It is best represented by the Sishui site of Yinjiacheng (Sishui Yinjiacheng 洒水尹家城 [Beijing: Wenwu chubanshe, 1990], 207Google Scholar, fig. 139: 4, 10). The vessel types associated with the Yueshi culture are distinct from those of the Lower Xiajiadian culture, but they share the same li-vessel distinguished by a band of applique around the neck and another running up the body of the vessel from the point where the adjacent legs join (211, fig. 142: 3, 5, 6, 8; Kaogu 19763, 204, fig. 11: 3; pl. 10: 6). This type of li, unknown at Erlitou and uncharacteristic of the Shandong pottery assemblages generally, is more common in the Northern Zone. The question therefore remains open whether the painted ware of the Yueshi culture owes directly to Erlitou, or to transmissions from more northerly sites. The same question pertains in the case of metal objects, including fragmentary awls, knife blades, a single arrowhead and a bracelet, found in the Yueshi stratum at Yinjiacheng (Sishui Yinjiacheng, 203, fig. 137:1-8; pl. 81).

The Yueshi culture, first identified near Dongyueshi 東岳石 in Pingduxian on the Shandong peninsula (Kaogu 1962.10, 509–518)Google Scholar, is also represented on the northeastern tip of the peninsula at the Mouping 牟平 site of Zhaogezhuang照格庄, where sherds of painted pottery similar to that at Yinjiacheng have been found, along with a metal awl (Kaogu xuebao 1986.4, 463Google Scholar, fig. 17; 470, fig. 22: 18). Carbon-14 dates for the Yueshi finds at Zhaogezhuang range from 1890 B.C. ±135 (ZK 868) to 1745 B.C. ± 130 (ZK 870) (p. 474).

9. Kaogu 1984.1, pl. 4: 1 (color). A similarly organized design of interlocking C-curls is seen again on a turquoise-inlaid frontlet from Erlitou Period IV (Kaogu 1986.4, 321Google Scholar, fig. 6:1; pl. 7:1 [color]). See n. 14 below.

10. A fragment of lacquerware has been unearthed from M 2 in Sector 3 at Erlitou (Period III) which shows a pair of oval eyes surrounded by a series of C-curls (Kaogu 1983.3, 203Google Scholar, fig. 9: 9). Other items of lacquered wood, including coffins and vessels from 81YL tombs 3, 4, 5 assigned to Periods II-III, are noted in Kaogu 1984.1, 39–40Google Scholar.

11. Kaogu xuebao 1990.1, 64–65Google Scholar; pl. 6:1-3.

12. Kaogu 1975.5, 305Google Scholar, fig. 4: 3; pl. 9: 2; Kaogu 1976.4, 260Google Scholar, fig, 4; pl. 5:3; Kaogu 1978.4, 270Google Scholar, fig. 1: 2; pl. 12:1; Kaogu 1983.3, 203Google Scholar, fig. 9: 4; pl. 1:1 (color); Kaogu 1986.4, 320Google Scholar, figs. 4, 5; pl. 7: 2: 5 (color); pl. 8: 6, 7; Kaogu 1991.12, 1138Google Scholar, fig. 1; pl. 8; Kaogu 1992.4, 296Google Scholar, fig. 2: 2; pl. 2: 1.

13. Kaogu 1983.3, 204Google Scholar, fig. 10: 8, 9; Kaogu 1992.4, 296Google Scholar, fig. 1:3; Kaogu 1975.5, 305Google Scholar, fig. 4: 1, 2 (chisels); Kaogu 1976.4, 260, fig. 3Google Scholar; pl. 5: 4-6; Henan chutu Shang Zhou qingtongcji, no. 4 (axe and dagger-axes); Kaogu 1984.1, 38Google Scholar, fig. 5:1; pl. 4: 1 (color); Kaogu 1986.4, 321Google Scholar, fig. 6, top; pl. 7:1 (color); Kaogu 1992.5, 296Google Scholar, fig. 2:1; pl. 1 (color) (frontlets); Kaogu 1965.5, pl. 5: 20; Kaogu 1984.1, pl. 4: 2 (color); Kaogu 1986.4, 321Google Scholar, fig. 6, lower; pl. 8: 5; Kaogu 1992.4, 296Google Scholar, fig. 3:1; pl. 2: 2 (bells); Kaogu 1976.4, 261Google Scholar, fig. 5; pl. 5:1, 2 (disks).

14. Lin Yun 林漯 is the first scholar to propose that Chinese bronze implements exhibit influence from the Northern Zone as early as the period of the Erlitou culture. In his view, the Erlitou knife with perforated grip (Fig. 2) and the battle-axe from the same site show that “bronzes of the Northern Complex already existed in the period of the late Erh-li-f ou culture.” He comments that “in the past researchers within and without China frequently were constrained by the preconceived notion that bronzes of the Northern Complex were fairly late. The subjective hypothesis that the bronzes of the Northern Complex are an offshoot of the bronzes of the Chung-yüan Complex is untenable, as is the idea that they originated in the northwest” (”A Reexamination of the Relationship between the Bronzes of the Shang Culture and of the Northern Zone,” trans. Goodrich, David, in Chang, K.C., edv Studies of Shang Archaeology [New Haven: Yale University Press, 1986], 250Google Scholar). Lin Yun discusses late Shang bronze artifacts from the north unearthed from Tomb 5 at Xiaotun 小屯 on pp. 250-258; see also, Yinxu Fu Hao mu” 殷墟婦好墓 (Beijing: Wenwu chubanshe, 1980)Google Scholar, colorplate 12: 1-2; pl. 66:1-6; pl. 68:1, 3.

Lin Yun's contribution was preceded by that of Max Loehr, who had long advocated a ‘northern’ source for many of the Anyang weapons and tools. Well before the Zhengzhou and Erlitou sites were known, Loehr had also surmised the existence of a pre-Anyang phase based on his investigation of northern metalwork. In his words, “the term ‘Northern’ is vague in chronological regard, in that it has a connotation of earliness or, to put it more exactly, of a pre-Anyang stage. The weapons of the Northern group are not only foreign in the Shang milieu by virtue of their decor, they are also typologically early throughout. Therefore, we must consider them as possibly older than Anyang, the more so since some of them (shaft-hole axes; dagger-axes with sockets or shaft-rings) go out of fashion early. At the same time, this pre-Anyang metal stage — not hitherto actually discovered at a particular site — would go some length to explain Anyang's extraordinary technical niveau” (Chinese Bronze Age Weapons: The Werner Jannings Collection in the Chinese National Palace Museum, Peking [Ann Arbor: University of Michigan Press, 1956], 104Google Scholar).

A particular question pertains in the case of the small bells and the so-called horse frontlets represented at Erlitou and found together in the same tombs. Again, neither type of items has precedents in the immediate geographical area, and both would seem more congenial in the context of a pastoral society. The existence of the horse frontlet, if this designation is correct, implies indeed more than this, namely that the horse was either ridden or employed as a draft animal at Erlitou itself. The earliest of the excavated frontlets, from Period II, was found next to a bell in 81 YLM 4 (Kaogu 1984.1, 37Google Scholar, fig. 2). The two later frontlets belonging to Period IV were similarly discovered with a bell lying close by (M 11: Kaogu 1986.4, 319Google Scholar, fig. 3, right; M 57: Kaogu 1992.4, 296Google Scholar, fig. 2:1). The observation that the two objects have so far been consistently found together— and the frontlet never without an accompanying bell — may lend weight to the original identification of the former as a horse trapping, if one takes into account the likely function of a bell as the accoutrement of an animal, enabling its location to be discerned as it moves about. No remains of horses have come to light at any Erlitou site, but they are in evidence at a roughly contemporary Qijia settlement in Gansu (see n. 56 below). It may additionally be noted that small metal bells have been found in association with the sacrificed remains of oxen, horses and dogs in graves M 5 and M 16 at Huangjiazhai 黄家寨, Datong in Qinghai, assigned to the late phase of the Qijia culture or to the period immediately thereafter (Kaogu 1994.3, 195Google Scholar, fig. 4; 196, fig. 6).

15. Mallory, J. P., In Search of the Indo-Europeans: Language, Archaeology and Myth (London: Thames and Hudson, 1989), 62–63Google Scholar; 225-226. Chernykh, E.N. dates the Afana-sevo to the first half of the third millennium (Ancient Metallurgy in the USSR: The Early Metal Age, tr. Wright, Sarah [Cambridge: Cambridge University Press, 1992], 182–183)Google Scholar.

16. Mallory, , In Search of the Indo-Europeans, 62, 223, 226Google Scholar.

17. Mallory, , In Search of the Indo-Europeans, 227, 231Google Scholar; also see recent calibrated dates provided by David Anthony, referred to in the present text, n. 83.

The earlier dates now ascribed to the arrival of the Afanasevo and Andronovo pas-toralists in the east accord with Lin Yun's observation that the investigation of the early Bronze Age sites in Inner Mongolia is yielding increasingly older dates (“Bronzes of the Shang and Northern Zone,” 249-250).

18. Anthony, David W. and Brown, Dorcas R., “The Origins of Horseback Riding,” Antiquity 65 (03 1991), 22–38CrossRefGoogle Scholar; Anthony, David W., Telegin, Dimitri Y. and Brown, Dorcas, “The Origin of Horseback Riding,” Scientific American (12 1991), 94–100Google Scholar.

19. Telegin, Dmitriy Y., Dereivka: Y. Settlement and Cemetery of Copper Age Horse Keepers on the Middle Dnieper, tr., Pyatkovskiy, V.K., edv Mallory, J.P. (Oxford: B.A.R. 1986)Google Scholar. Telegin states that “animal husbandry was the main subsistence occupation of the SSC [Sredny Stög culture] population. It was an expressly horse-breeding culture in which the horse was the major form of livestock” (p. 72). He further notes that ”an essential feature of Dereivka's livestock was the predominance of the horse (74%), followed by cattle (19%) and, to a much lesser extent, ovicaprids, pigs and dogs” (p. 82).

20. Bibikova, V.I., “A Study of the Earliest Domestic Horses of Eastern Europe,” in Telegin, , Dereivka, 143Google Scholar.

21. Telegin claims that “it is impossible to pasture herds of these swift-footed animals without riding one of them,” and he adds that the “absolute necessity of having mounted herders for grazing horse droves is also emphasized by other scholars”, (Dereivka, 82). The concensus of opinion derived from the presentations at the conference organized by David Anthony on “Early Horsekeepers of the Eurasian Steppe” held at Petropovlovsk, Kazakhstan, June 18-24, 1995, strongly favors the concept of early horse domestication and the almost equally early advent of horse riding. The papers from the conference are unpublished, although abstracts have been distributed among the participants.

22. Telegin, , Dereivka, 82–83Google Scholar.

23. Chernykh, , Ancient Metallurgy, 42–43Google Scholar; also see p. 48.

24. Telegin is of the opinion that the domestic horse was a factor in steppe cultures stretching as far as eastern Siberia and possibly the Altai by the late fourth and third millennia:” … the domestic horse was present in the Neo-Eneolithic settlements of the Southern Ural region … [where] the amount of this animal [is estimated] at over 40%. … Considering that the domestic horse was known also in the Afanasyevo culture, … the eastern horse domestication area extends far beyond the Volga into the East Siberian steppe. It may be mentioned in this connection that some authors … include even the Turan-Altai area in the geographical range of horse domestication” (Dereivka, 86).

25. Mallory cites the work of Maryana Khlobystina and Elga Badetskaya regarding the striking similarities between the pointed bottom ceramic vessels characteristic of the Pit-Grave and Afanasevo cultures (In Search of the Indo-Europeans, 223, 225-226). Burial positions (hoeker or supine with the legs flexed) are the same in both cultures; and in both, the body is sprinkled with ochre (p. 223). Mallory also notes that Tamilla Potemkhina has uncovered comparable burials at points midway between the Volga and the Yenisei at Verkhnyaya Alabuga and Ubagan 1 in the Tobal river area (pp. 225226). Chernykh, who is likewise convinced of a connection between the Pit-Grave-Poltava communities and the Afanasevo, remarks on the presence of a single copper shaft-hole axe comparable to those of the Pit-Grave-Poltava culture from Potnikoyaya in the Altai region where Afanasevo finds have been made (Ancient Metallurgy, 183; the axe is illustrated on p. 87, fig. 28: 29). The Pit-Grave culture, datable between 36002200 B.C., is described by Mallory (pp. 210-215) and by Chernykh (pp. 83-91). Both authors draw attention to the large number of horse bones (80%) among the faunal remains uncovered at the Repin Khutor site on the Don (Mallory, 212; Chernykh, 85). Mallory further mentions wooden cheekpieces from a Pit-Grave kurgan at Vinogradovka, near Odessa, as possible evidence of horse-riding (p. 213).

26. Chernykh, , Ancient Metallurgy, 133Google Scholar.

27. Xiaxian Dongxiafeng 夏縣東下瑪 (Beijing: Wenwu chubanshe, 1988), 72, fig. 77Google Scholar: 20; 119, fig. 119:16; 120, fig. 120: 3; pl. 23: 2; p. 214.

28. Xiaxian Dongxiafeng, 75, fig. 78:10-11; 120, fig. 120: 5; pl. 23: 5.

29. Barnard, Noel and Tamotsu, Satō, Metallurgical Remains of Ancient China (Tokyo: Nichiōsha, 1975), 24–25Google Scholar.

30. Xiaxian Dongxiafeng, 163, fig. 149:14; pl. 69: 3, center.

31. Kaogu 1980.2, 100, fig. 7: 13Google Scholar; Kaogu 1984.12, 1068–1071Google Scholar; pl. 3. It is unclear whether the ceramic bells from the same site should be considered as prototypes of the metal version, or as copies (p. 1069, fig. 1: center and right).

32. Fengxifajue baogao 澄西泰掘報吿 (Beijing: Wenwu chubanshe, 1962), 43–69Google Scholar; pls. 19-36. For a detailed study of the Kexingzhuang II culture, see Xingpeng, Liang 梁星彭, “Shi lun Shaanxi Miaodigou erqi wenhua” 論陝西廟底溝二期文化, Kaogu xue-bao 1987.4, 407–411Google Scholar; and ”Shi lun Kexingzhuang erqi wenhua” 試論客省庄二文化, Kaogu xuebao 1994.4, 397–424Google Scholar. A complete list of the carbon-14 dates for the Kexing-zhuang II sites, which range from approximately 2400-1900 B.C., is found on p. 411 of Liang's 1994 article. See also Xizhang, Yang 楊錫萆, “Huanghe zhongyou de Longshan wenhua” 黄河中游的龍山文化,Xin Zhongguo de kaogu faxian he yanjiu, 79–81Google Scholar, translated by Thote, Alain in “Les traditions Longshan de la région du cours moyen du fleuve Jaune,” Arts Asiatiques 43 (1988), 69–72Google Scholar;

33. Fengxifajue baogao, pl. 25: 1-2; Kaogu xuebao 1990.1, pl. 2: 8.

34. Fengxifajue baogao, pl. 28.

35. Fengxi fajue baogao, pl. 29: 3; compare the ding from Meishan II in Kaogu xuebao 1982.4, pl. 2: 1.

36. Fengxifajue baogao, pl. 29: 1, 2; compare the Cuoli III lihe in Kaogu 1978.1: 11, fig. 12:11; fig.12:11; fig. 12: 6; Fengxi fajue baogao, pl. 31.

37. Kaogu 1975.5, pl. 2; Kaogu yu wenwu 1980.3, 11Google Scholar, fig. 20 Qiangzhai 姜寨 Period V).

38. Miaodigou yu Sanliqiao, 92-102; pls. 80-84.

39. Kaogu 1990.7, 584Google Scholar, fig. 10.

40. Kaogu 1989.4, 296Google Scholar, fig. 12: 9; Kaogu 1975.5, pl. 2: 4; Kaogu yu wenwu 1980, 3, 11Google Scholar, fig. 20:1-2. Other features of comparison between the Xinzhou finds and those from Ke-xingzhuang include crisscross strips of appliqué as surface decoration (Kaogu 1989.4, 296Google Scholar, fig. 12:2).

41. Wenwu 1989.3, 1–21Google Scholar; pls. 1-2. Baiyan Period I represents the late Yangshao phase at this site (p. 3, fig. 4); Period II, belonging to the Miaodigou I stage, includes a jia similar to examples from the early period at Taosi (p. 4, fig. 5; 7, fig. 7; pl. 1: 4); Period III is of the late Longshan phase (p. 8, fig. 8); Periods IV-VI are attributable to the late Erlitou through late Erligang stages (p. 13, fig. 12; 17, fig. 16).

At Taigu, as at Xinzhou, we note a second type of li ornamented with a strip of applique reinforcing the groin (Wenwu 1989.3: 8Google Scholar, fig. 8: 6; Kaogu 1989.4, 292Google Scholar, fig. 4:1-2).

The same li is also found during the Erlitou period in the context or the Lower Xiajia-dian culture in the Beipiao area of eastern Inner Mongolia, and again amid the remains of the Yueshi culture in Sishui, Shandong (see n. 8 above).

42. Kaogu 1977.3, pl. 1.

43. Kaogu xuebao 1988.3, 315Google Scholar, fig. 18; 320, fig. 23. The Zhukaigou site by no means represents the first time people from the Wei River area came to inhabit Inner Mongolia. It is long preceded, for instance, by the fourth millennium Miaodigou I 廟底溝一期 Yangshao 仰韶 settlement at Bainiyaozi 白泥窖子 at Qingshuihe located near the northeast bend of the Huanghe (Kaogu 1988.2, 102Google Scholar, fig. 7). Traces of Majiayao 馬家暂 ware similar to that from Gansu are also found there (112, fig. 5: 8).

44. Kaogu xuebao 1988.3, 329Google Scholar.

45. Kaogu xuebao 1988.3, 317Google Scholar; 321, fig. 24:4.

46. Kaogu xuebao 1988.3, 325Google Scholar, fig. 29; pl. 8: 3, 4, 7.

47. Kaogu 1977.3, pl. 3: 3. Jade blades discovered at the Shimao site suggest a relationship to the Erlitou culture (155, fig. 2; 156, figs. 1, 2, 5; pl. 4:1-6).

48. Kaogu xuebao 1988.3: 307Google Scholar, fig. 9:1; 315, fig. 18:13-14; 318, fig. 20:1-2; 319, fig. 21:1.

49. Kaogu 1989.2, 101Google Scholar; 102, fig. 7.

50. Noel Barnard provides a list of four copper ore locations in close proximity to Qijia settlement areas: Ledu in Qinghai; Gulang, near Wuvvei, in the Gansu Corridor; and Yongjing and Qingyuan, north of Lanzhou, (Metallurgical Remains, 24–25Google Scholar; fig. 11 on p. 25). It has not been determined whether tin was to be had in the same vicinity. According to Tian Guangjin 田廣金, there are known copper sources in the Huhe-haote region of Inner Mongolia, reaching as far south and east as Laohushan and as far west as the western reaches of the Huanghe (Katheryn Linduff, personal communication, July 28, 1993), It may have been technology, as much as trade, that the Qijia brought into this area.

51. Liang Xingpeng shares the view that the Kexingzhuang II culture derives from Miaodigou II (“Shi lun Shaanxi Miaodigou erqi wenhua,” 407-411; “Shi lun Kexing-zhuang erqi wenhua,” 412-417).

52. Kaogu xuebao 1974.2, 56Google Scholar; pl. 6: 3 (millet); Kaogu xuebao 1974.2, pl. 7; Kaogu xuebao 1975.2, 87, fig. 23Google Scholar; Kaogu xuebao 1978.4, 441Google Scholar; pl. 3: 8-9 (oracle bones).

53. The Qijia sites, numbering over 300, have been divided by Hu Qianying胡謙盈 into two types according to their respective locations and the differences in their inventories. One category (A) comprises the sites associated with Dahezhuang 大何庄 and Qinweijia 秦魏家 in Yongjingxian, in central Gansu; the other (B), sites located farther to the north and west, such as Huangniangniangtai 皇娘娘台, Wuwei, in the Gansu Corridor and Liu wan 柳灣 near Ledu in Qinghai. The second category (B), in which the pottery incorporates features of the foregoing Machang 馬廠 tradition, is surmised to be the later of the two (”Shilun Qijia wenhua de butong leixing ji qi yuan-Jiu” 試論齊家文化的不同類型及其流源, Kaogu yu wenwu 1980.3, 77–82Google Scholar; fig. 1 on p. 78 provides a distribution map; also see (Duanju, Xie 謝端据, “Shilun Qijia wenhua” 試論齊家文化, Kaogu yu wenwu 1981.3, 79–80Google Scholar). The Qinweijia and Dahezhuang sites are further subdivided chronologically into upper and lower strata (Xie Duanju, p. 80 and “Lun Dahezhuang yu Qinweijia Qijia wenhua de fenqi” 論大何庄與秦魏家文化的分期 Kaogu 1980.3, 248–254Google Scholar). The upper level at Qinweijia and the lower level at Dahe-zhuang appear to be closely contemporary (compare Kaogu 1980.3, 254Google Scholar, fig. 5 and 251, fig. 2). An analysis of the evolution of various pottery forms throughout the Qijia sites — including the progressive narrowing of the tall-necked, originally amphora-shaped, shouldered guan 罐 and the eventual loss of its handles by the time of Dahe-zhuang; the presence of the three-handled guan at Liuvvan, Huangniangniangtai and the lower level at Qinweijia, and its subsequent absence from both the upper level at Qinweijia and from Dahezhuang; and the introduction of the two-handled pear-shaped guan in the upper level at Qinweijia and its continuation at Dahezhuang — tends to support the chronological priority of the northern and western sites over those from the central Gansu region. These observations may, however, to some degree be conditioned by considerations of regional variation. The pottery from other Qijia sites in Qinghai and Ningxia is generally best accommodated toward the earlier end of this developmental spectrum (see, for example, the Qinghai sites in Xi'ning and those in Hualong, published in Kaogu 1986.4, 306, 309–315Google Scholar and Kaogu 1991.4, 318Google Scholar, fig. 7; and the Longde site of Yehezi 頁河子 in Ningxia; Kaogu 1990.4, 292, fig. 4Google Scholar; 293, fig. 5; pl. 1: 1, 2, 6). The foregoing may be compared to Zhang Zhongpei's 張忠培 detailed analysis of Qijia pottery in Zhongguo beifang kaogu wen ji 中國北方考古文集 (Beijing: Wenwu chubanshe, 1986), 105–131Google Scholar.

The origins of the Qijia ceramic tradition are poorly understood. Hu Qianying singles out the Changshan Lower 常山下層 ware as a possible source. The Chang-shan tradition, dated to 2930 B.C. ±180, is distributed along the upper Jinghe in eastern Gansu and southern Ningxia (Kaogu yu wenwu 1980.3, 77–82, 33Google Scholar). Certain Changshan vessels, such as the triple-handled drinking cup (pl. 12: 2) would appear to presage the mature Qijia types. Others, including the single-handled mugs and two-handled corded ware jars (pl. 4: 6; pl. 12: A, 6; pl. 13: 3-4) find parallels at the Anban 案板 Period III site near Fufeng in western Shaanxi (Kaogu yu wenwu 1988.5, 6, 180Google Scholar, fig. 13: 4, 13, 14), dated to 2670 B.c.±135 (ZK 1378) and 2575 8.C.±150 (ZK 1377) (Kaogu xuebao 1987.4, 406)Google Scholar. Similarities between the two traditions continue into Anban Period IV with straight-walled vessels of various types whose sides flare slightly toward a crimped edge at the bottom (Kaogu yu wenwu 1980.3, pl. 12: 5; pl. 13:1; Kaogu yu wenwu 1988.5, 181Google Scholar, fig. 14: 14; 187, fig. 19: 12). A high percentage of Anban Period III-IV vessels belongs within the Miaodigou II tradition. Several vessels said still to be of the early phase of the Qijia culture are found at the Shizhaocun 師趙村 site (Period VII) at Tianshui in eastern Gansu in the context of wares belonging to the Kexingzhuang II tradition (Kaogu 1990.7, 584Google Scholar, fig. 10; 585).

54. Kaogu xuebao 1975.2, pl. 11: 12 (from Yongjing Qinweijia, lower stratum). Even more striking is an analogous yan 魁 found at Wuwei Huangniangniangtai (Kaogu xuebao 1978.4, pl. 10: 2). Two li vessels closer to the Kexingzhuang II protoype in shape, and whose surfaces are covered with cord marks rather than combed impressions, were unearthed from ashpit H 4 at Lingtai Qiaocun 橋村 in southeast Gansu (Kaogu yu wenwu 1980.3, pl. 3:1-2).

55. Kaogu 1980.3, 248–254Google Scholar.

56. Kaogu xuebao 1974.2, 56Google Scholar; Kaogu xuebao 1975.2, 88Google Scholar; Kaogu yu wenwu 1981.3, 81Google Scholar. Horse remains from the Wuwei site of Huangniangniangtai are mentioned in Kaogu xuebao 1978.4, 448Google Scholar only in the English summary.

57. The practice of suttee (sati) is in evidence at the Yongjing site of Qinweijia (Kaogu xuebao 1975.2, 65Google Scholar, fig. 6; pl. 3: 1-3) and Huangniangniangtai, Wuwei (Kaogu xuebao 1978.4, 429Google Scholar, figs. 13-14; 431, fig. 17; pl. 2: 2-3).

Edwin Pulleyblank, whose research in linguistics has led him to infer that a branch of the proto-Tocharians ought to have been located as far east as western China by the second millennium B.C., calls particular attention to the practice of suttee within the Qijia culture (”The Indo-European Connection,” paper presented at the meeting of the American Historical Association, New York, 12 28, 1985, 8–10)Google ScholarPubMed.

Mikhail Gryaznov conjectures that the widespread practice of suttee in the Eurasian steppe was the result of a changed economy that combined agriculture with an increased exploitation of domestic animals for food and transport, leading to the abandonment of hunting as a means of subsistence. According to Gryaznov, the altered economy entailed a greater dominance of the male members of the society, whose services were required for the pasturing and protection of the herds. Within the new social structure, it became acceptable for men to capture their wives from other clans; a wife obtained in this manner lost the protection of her clan and so could be made to accompany her spouse in death (The Ancient Civilization of Southern Siberia, tr. Hogarth, James [Geneva: Nagel, 1969], 67, 95Google Scholar). Why members of the Qijia community might have accepted the practice remains a question.

Whatever the derivation of their burial rites, limited Osteometrie analysis has indicated that the Qijia themselves were a Mongoloid people (Yan, Yan 顏閬, “Gansu Qijia wenhua mu zang zhong tougu de chubu yanjiu” 甘蕭齊家文化墓葬中頭骨的初歩硏究, Kaogu xuebao 1955.9, 193–197Google Scholar; Duanju, Xie, “Shilun Qijia wenhua,” 82Google Scholar).

An additional aspect of the Qijia culture possibly indicative of contact with northern nomadic groups is a feature unique to the settlement at Dahezhuang, namely the five stone circles, each measuring approximately four meters in diameter, with a gap in each suggestive of an entrance. The stone circles were unearthed in both the lower stratum (F 3, 5) and the upper stratum (F 1, 6, 12), and they were found interspersed with square-shaped house foundations and burials of the same two periods (Kaogu 1960.3, 10Google Scholar; pl. 1:2; Kaogu xuebao 1974.2, 32Google Scholar, diagram, fig. 4; p. 38; pl. 2:2 and pl. 3:1). The original interpretation of these stone circles as perhaps related to religious practices, suggested mainly by the ritual sacrifice of a cow in the immediate area, has remained unchallenged for lack of other evidence that might clarify their actual purpose (Duanju, Xie, “Shilun Qijia wenhua,” 82Google Scholar; Kwang-chih, Chang, The Archaeology of Ancient China, 4th ed., rev. and enl. [New Haven and London: Yale University Press, 1986], 285)Google Scholar.

In other geographical areas of the world, such stone circles are often associated with nomadic cultures, and their significance in these contexts is clear. As Robert Neuman has noted in regard to the archaeology of the North American grasslands, the “most common evidences of nomadic culture habitations are characterized by ‘tipi ring’ sites containing several or as many as 170 circular arrangements of stones. The circles may range from two to ten meters in diameter. … Ethnohistoric data have documented that these stones functioned as weights to anchor down the wall edges of the … tent. … When the tent was moved, the stones remained generally in position thus outlining the position of the tent and the settlement pattern of the encampment” (“Structural Evidence of Pastoral Nomad Habitations in Northwest Xinjiang Province, PRC,” paper presented at the North China Archaeological Conference, Huhehaote, August 11-18, 1992, 2-3). The author mentions similar stone rings, some of recent date, found in the meadows northwest of the Altai and southwest of Lake Sayrum in Xinjiang (p. 4). Roger Cribb illustrates comparable remains from a campsite at Sari ay din Yayla in southern Turkey, (Nomads in Archaeology, [Cambridge: Cambridge University Press, 1991], 171Google Scholar, fig. 9: 4). In my own experience they are seen to mark abandoned present-day campsites across the grasslands of Mongolia. The stones composing the Dahezhuang circles differ by their greater regularity in size and placement, but the chance they could relate to nomadic tent structures may nonetheless deserve consideration.

58. For illustrations of metal implements from Qijia sites, see Kaogu xuebao 1974.2, pl. 18:10 (Dahezhuang); Kaogu xuebao 1978.4, 437Google Scholar, fig. 21 (Huangniangniangtai); Kaogu 1985.11, 978Google Scholar, fig. 2: 10, 11 (Xinglin); Zhang Zhongpei, Zhongguo beifang kaogu ji, pl. 14: 8-13 (Huangniangniangtai and Qinweijia). Zhang provides a table summarizing the results of metallurgical analysis of the Qijia artifacts (143). An Zhimin 安志敏 refers briefly to some of the Qijia metalwork in ”Zhongguo zaoqji tongqi de jige wenti” 國早期銅器的幾個問題, Kaogu xuebao 1981.3, 277–280Google Scholar. William Watson, who has dealt at length with the question of the early Chinese contact with the north, proposed the contrary theory that the socketed axe was indiginous to China (Cultural Frontiers in Ancient East Asia [Edinburgh: Edinburgh University Press, 1971, 54, 56, 57)Google Scholar.

It is probably incorrect to ascribe to the Qijia culture the term ‘Chalcolithic’ in its proper sense as a period when metalworking was limited to the use of copper alone and when deliberate alloys of copper to create bronze were unknown. By Qijia times, metallurgists throughout Eurasia were skilled in the production of bronze, and we may assume that those who influenced the early phase of metalworking in Gansu and Qinghai carried this technology with them. The fact that many of the Qijia metal implements were fashioned of copper, and only a few of bronze, may have had more to do with the ready availability of high-grade copper, whose properties are approximate to those of bronze in terms of hardness and durability, and to the possible scarcity of tin, rather than with any technological shortcoming. The designation ‘early metal age’ in regard to the Qijia would seem more appropriate. Similarly we should be wary of assigning objects fashioned by more ‘primitive’ sheet-metal techniques to a supposed stage preceding the use of casting. Again, both techniques were familiar to the Eurasian metalworkers, and both are in evidence at Qijia.

59. Chernykh, , Ancient Metallurg, 215–234Google Scholar.

60. Chernykh's conviction that the Seima-Turbino included among their numbers craftsmen skilled in metalworking is supported by ‘founders' burials’ at Rostovka and Sopka that contained stone molds used in casting socketed axes and other tools (Ancient Metallurgy, 215, 218, 222, fig. 76 and Fig. 7d in the present text). He has found no evidence that the Seima-Turbino were involved in mining or metallurgy (218). Casting molds for a variety of implements have been uncovered elsewhere at other Seima sites (see, for example, Chernykh, and Kuzminykh, , Drevnyaya metaliurgiya Severnoi Evrazii [Moscow: Nauka, 1989], 155, fig. 79: 3Google Scholar; 159, fig. 82; 160, fig. 83.)

It is noteworthy that horses, depicted in the round, occasionally appear as the pommels of Seima knives (Chernykh, , Ancient Metallurgy, 223Google Scholar, fig. 77: 1-2; see also n. 63 below).

61. Chernykh, , Ancient Metallurgy, 215–216Google Scholar.

62. Chernykh, , Ancient Metallurgy, 227Google Scholar.

63. Chernykh, , Ancient Metallurgy, 227Google Scholar. Chernykh's view of the direction in which the Seima-Turbino were travelling is based in part upon his conclusion that the tinbronze alloy that characterizes their metalwork can only have been obtained in the Rudny Altai area, which is rich in both copper and cassiterite ores. He reports that spectographic analysis of some 353 pieces of Seima-Turbino metalwork revealed almost 50% to be tin-bronze, and that this alloy is virtually the only one represented in the Siberian distribution zone (222-224). He points out that these ores are absent from western Siberia, and that the Urals lack cassiterite (224).

Chernykh speculates that some of their travels may have taken place in winter (p. 227). Nonetheless, the figure of a man on skiis holding the reins of a horse that pulls him along, seen atop the pommel of a Rostovka knife (p. 228; pl. 22, center, right), suggests a skill that surely was more a matter of sport than an effective means of transport.

64. Chernykh, , Ancient Metallurgy, 200, 224, 226–227Google Scholar.

65. Chernykh, , Ancient Metallurgy, 226Google Scholar.

66. Chernykh, , Ancient Metallurgy, 229Google Scholar.

67. Chernykh, , Ancient Metallurgy, 226Google Scholar.

67. Chernykh, , Ancient Metallurgy, 226Google Scholar. Nephrite disks from Seima-Turbino sites are illustrated in Chernykh and Kuzminykh, , Orevnyaya metaliurgiya, 245Google Scholar, fig. 109 and Bader, O.N., Drevneischie metallurgi Priuralya (Moscow: Nauka, 1964)Google Scholar, colorplate following p. 92. For nephrite rings from the Baikal area, see A.P. Okladnikov, ”Neolit i bronzovy vek Pribaikalya,” Materialy i issledovaniya po arkheologii SSSR 43 (1955), 179, fig. 75Google Scholar; 180, fig. 76; 184, fig. 79; 270, fig. 127; 271, fig. 128; 273, fig. 130: 6-7; colorplate 6 following p. 224.

69. Chernynkh, , Ancient Metallurgy, 226Google Scholar.

70. Chernykh, , Ancient Metallurgy, 218Google Scholar.

71. Chernykh, , Ancient Metallurgy, 220, 222Google Scholar; Chernykh and Kuzminykh divide the Seima-Turbino socketed axes with ring-shaped lugs into several categories and offer maps showing the geographical range of the various types (55, fig. 17: 4 [K-20]; 57, fig. 19 [K-22, 24, 26]; 58 fig. 20: 1, 3 [K-28]; 153, fig. 78 [K-46, 48, 50]; distribution maps appear on p. 49, fig. 11 and 59, fig, 21). It should be pointed out that the entire range of Seima weapons, including the socketed spearhead, the dagger and the shaft-hole axe, is conspicuously absent from the Qijia sites. (Illustrations of spearheads may be found in Chernykh, 220, fig. 74: 17-20; 221, fig. 75: 11-14; daggers and shaft-hole axes are shown in Chernykh and Kuzminykh, , Drevnyaya metaliurgiya, 112Google Scholar, fig. 64 and 127, fig. 70, respectively.)

72. The termination of the handle of the northern-style knife from Erlitou matches the Seima handles with their irregularly circular perforation within the bulge at the top (Figs. 2, 8a). Both are distinguished from the true ring-handeled terminations on knives and daggers of the later Karasuk type, which followed upon the earlier Seima form. The true ring-handled knife, long familiar from late Shang finds at Anyang (Yinxu Fu Hao mu, pl. 66: 2-6) is now known to have existed in the Northern Zone by Zhengzhou times, as evidenced by the long knife recovered from the Erligang level at the Inner Mongolian site of Zhukaigou (Fig. 4: 3).

73. The Huangniangniangtai piece is mentioned by Zhimin, An in “Zhongguo zaoqi tongqi de jige wenti,” 278Google Scholar.

74. In view of the corresponding decorative patterns on the Huangniangniangtai and Seima-Turbino knife handles, it may be worth pointing out the similarity between the patterns typically seen on Seima metalwork, especially on the socketed axes, and those that appear on the Qijia ceramics in the rare instances when they carry painted decoration. The Qijia decorative style, quite distinct from any other early Chinese painted pottery tradition, comprises a narrow range of sparingly used geometric motifs analogous to those on the Seima bronze implements. Both the Seima and the Qijia motifs follow one of two arrangements: they appear either as a row of pendant triangles encircling the objects, or else as sets of linked triangles (or rhombic shapes in the case of the Seima designs) placed tip-to-tip and aligned with the central axis of the object. The painted patterns on the Qijia jars from Liuwan, Ledu in Qinghai may be compared with those cast on the socketed axes from Seima and Rostovka (Qinghai caitao 靑海彩陶׳ [Beijing: Wenwu chubanshe, 1980]Google Scholar, colorplate 27, and Qinghai Liuwan 靑海柳灣 [Beijing: Wenwu chubanshe, 1984], vol. 2, colorplate 4Google Scholar; Chernykh, and Kuz-minykh, , Orevnyaya metaliurgiya, 55Google Scholar, fig. 17: 1, 4). In nearly all instances the geometric units are filled by fine parallel lines, either adjacent to one side of the unit, or forming a cluster of nested shapes within it. Less frequently the units are filled by cross-hatched lines (Qinghai caitao, pl. 159; Chernykh and Kuzminykh, 53, fig. 15: 6). The coincidence of shared patterns extends as well to the occasional use of designs in openwork in evidence on Seima knife-handles and on the handles of small Qijia jars (Chernykh and Kuzminykh, 111, fig. 63: 1, left; Qinghai caitao, colorplate 27).

75. Kaogu 1986.4, 309–315Google Scholar.

76. Kaogu 1986.4, 313Google Scholar, fig. 8: 3-6. Microlithic blades are a standard component of the Seima-Turbino inventory (see Bader, O. N., Drevneishie metallurgi Priuralya, 97–107, figs. 90-99Google Scholar; Chernykh, and Kuzminykh, , Orevnyaya metaliurgiya, 231–233, figs. 104-106Google Scholar).

77. Kaogu 1988.4, 313Google Scholar, fig. 8:1—2, 25.

78. For further examples of Okunevo bone-mounted awls, see Vadetskaya, E. B., Leontev, N. V. and Maksimenkov, G. A., Pamyatniki okunevskoi kultury (Leningrad: Nauka, 1980), 109Google Scholar, fig 18:1, 8.

79. Chernykh, and Kuzminykh, , Orevnyaya metaliurgiya, 106, fig. 61: 13Google Scholar; Chernykh, , Ancient Metallurgy, 184, fig. 65: 3Google Scholar; 221, fig. 75: 7. The same technique of haf ting metal blades at an angle to a bone mount is also known in the Baikal area during the Glaz-kovo period (Chernykh, 271; 272, fig. 93: 6; Okladnikov, , “Neolit i bronzovy vek Pri-baikalya,” 34Google Scholar, fig. 14: 1b, 3).

80. Two other bone-mounted awls from Rostovka appear in Matyushchenko, V. I. and Lozhnikova, G. B., “Raskopki mogilnika u derevni Rostovka bliz Omsk v 1966-1969,” Iz istorii Sibiri 2 (1969)Google Scholar, fig. 16:13, following p. 161.

81. Metal blades inserted vertically into bone or wooden handles continue to occur in Qinghai and Xinjiang subsequent to the Qijia period in contexts evincing strong connections to the northern Eurasian pastoralists. A short serrated triangular metal blade with a bone handle was recovered from M 5 at Huangjiazhai 黃家寨 near Da-tong in Qinghai, (Kaogu 1994.3, 201, fig. 9: 7)Google Scholar. The site is assigned on the basis of its pottery to the late Qijia or early Kayue 卡約 culture (p. 198, fig. 7; 200, fig. 8; pp. 204¬205). M 5 also contained a set of horses hooves, while ox heads and several sets of their hooves, along with a bone tube decorated with deer figures, were unearthed from a nearby grave (M 16) (p. 194, fig. 2; 195, fig. 4; 203, fig. 11:1; also see n. 14 above regarding metal bells from the same burials). A metal awl with a wooden handle was excavated from M 53 at the Yanbulake 焉不拉克 cemetery near Hami (Kaogu xuebao 1989.3, 346, fig. 24: 18Google Scholar; pl. 12: 7). Carbon-14 tests ascertained that the site dates from around the mid-Shang period (p. 354). The same cemetery revealed several skulls of ‘European’ type, as well as those of Mongoloids (p. 354). Carved wooden figures from the Yanbulake burials suggest connections with similar figures known from the Baikal region of Siberia (p. 349, fig. 27: 14-22; pl. 14; compare Okladnikov, , “Neolit i bronzovy vek Pribaikalya,” 286Google Scholar, fig. 139; 287, fig. 140; 294, figs. 145-146).

Other traits of the Qijia cultural complex likewise favor a connection with the eastern branch of the Seima-Turbino people, including specific features of the suttee burials. Gryaznov remarks upon the geographical distinctions among the south Siber-i an burials of this type, noting that in the case of those uncovered in the Minusinsk basin, the bodies are placed with the female facing the male, who in turn, lies with his back to her, while the Ob valley burials differ from this practice insofar as the male lies on his left side with the female on her right side facing him (Southern Siberia, 94). In the Qijia burials the male rests in an extended supine posture, while the female, whether positioned to his left (Qinweijia) or to his right (Huangniangniangtai), almost always faces her partner (Kaogu xuebao 1975.2, 65, fig. 6Google Scholar; pl. 3: 1-3; Kaogu xuebao 1978.4, 423Google Scholar, fig. 3; 429, figs. 13-14; 444-447; pl, 2). In a single Qijia burial at variance with this custom (Huangniangniangtai M 76), the female is placed with her back to the male (Kaogu xuebao 1978.4, 430Google Scholar, fig. 15; pl. 2: 1). The particular arrangement of the two bodies in the Qijia suttee graves corresponds more closely to those in the Ob valley and is thus more likely to indicate influence stemming from the territory associated with the Seima-Turbino than from other areas, such as the Yenisei.

82. According to Chernykh, “Very close links with well-dated Abashevo and Sin-tashta sites generally fix the period [following the Seima-Turbino entry into eastern Europe] in the sixteenth century B.C. It is quite reasonable to suppose that the consolidation of the mobile Seima-Turbino groups had taken place in the previous century” (Ancient Metallurgy, 229). Accepting a seventeenth century date for their original consolidation, Chernykh conjectures that the Seima-Turbino transcultural phenomenon endured for hardly more than two centuries: “The Eurasian component in the metal, which we see mainly in assemblages from the European zone, dates to the earliest phase of the Seima chronological horizon of the EAMP [Eurasian Metallurgical Province]. In the late phase it is virtually absent. It must be considered possible that these newcomers to eastern Europe could have completely disappeared by the fifteeenth century B.C., having been absorbed and assimilated by the mass of local peoples. Individual small groups apparently made their way through to the eastern Baltic region, but the same fate awaited them there. The assemblage from the Borodino hoard in Moldavia also points to the same time. … It has been reliably shown that the Borodino objects were contemporaneous with the period of Mycenaean influence in the Danube region and the Carpathians (sixteenth to fifteenth centuries B.C.)״ (p. 229). However, the earlier dating of the formation of the Seima-Turbino that I have proposed based on Anthony's redating of the Sintashta finds would imply that the existence of this group may not have been of such short duration as Chernykh thought (see below, n. 83).

83. David W. Anthony, “Bronze Age Burials in the Ural Steppes,” paper presented at the 92nd Annual Meeting of the American Anthropological Association, Washington, D.C., November 1993, 2, 12. The cluster of calibrated dates derived from tests performed on four samples of bone from two horse skulls from the Sintashta-Petrovka chariot burials at Krivoe Ozero were first calibrated to fall within the range of 22001800 B.C. Similar dates are reported to have been arrived at by independent tests carried out at Oxford (p. 6). Anthony has since slightly revised his dates to 2136-1904 B.C. (within a two sigma range); 2026 B.C. is the weighted average now given for all four dates (Personal communication, March 8, 1994; Anthony, , “Birth of the Chariot,” Archaeology 48.2 [March-April 1995, 38Google Scholar.

84. Mair, Victor H., “Progress Report for Project Entitled ‘A Study of the Genetic Composition of Ancient Desiccated Corpses from Xinjiang (Sinkiang), China’” Early China News 6 (Fall 1993), 4–6Google Scholar; also see Haddington, Evan, “The Mummies of Xinjiang,” Discover 15.4 (04 1994), 68–77Google Scholar and Mair, , “Mummies of the Tarim Basin,” Archaeology 48.2 (03-April 1995), 28–35Google Scholar.

A sense of the terrain involved in such a journey and the difficulties entailed may be gleaned from Thomas W. Atkinson's absorbing tales of his extensive travels by horseback in the mid-nineteenth century from the Altai mountains southeast to Barkol and beyond, and his return along the north side of the Tianshan, in Oriental and Western Siberia: A Narrative of Seven Years' Explorations and Adventures in Siberia, Mongolia, the Kirghis Steppes, Chinese Tartary, and Part of Central Asia (London: Hurst and Blackett, 1858), 441–573Google Scholar.).

85. An account of a trek by camel along the Edsingol to the Gansu Corridor in the 1930s is provided in DeFrancis, John, In the Footsteps of Genghis Khan (Honolulu: University of Hawaii Press, 1993), 181–227Google Scholar.

86. Given Chernykh's theory of the close relationship between the eastern Seima (whose physical type has not been ascertained) and the Okunevo, who are known to have been Mongoloid, it might prove difficult to recognize among the Qijia sketetal remains any belonging to northern population groups of Seima affiliation (see n. 57 above).

87. See p. 30 and nn. 13, 27 above; Chernykh, and Kuzminykh, , Orevnyaya metallurgiya, 129, fig. 71Google Scholar.

88. The stone macehead from Huangniangniangtai is shown in Kaogu xuebao 1978.4, 432Google Scholar, fig. 18: 5; pl. 3: 11. Bactrian maceheads of various materials are illustrated in Pottier, Hélène-Marie, Matériel funéraire de la Bactriane méridionale de l'Age du Bronze (Paris: Éditions recherche sur les civilizations, 1984), pl. 6: 44Google Scholar; pl. 7: 47-48.

It is Philip Kohl who first called attention to the possibility of a relationship between Bronze Age sites in northern Bactria and northwest China (”The Namazga Civilization: An Overview,” in Kohl, , ed., The Bronze Age Civilization of Central Asia [Armonk; M.E. Sharpe, 1981], xxi–xxiiiGoogle Scholar). Askarov's discovery of silk at Sapalli-tepe, however, has never been verified (as pointed out by Frederik Hiebert in a conversation, July 1, 1992).

89. Kaogu 1976.4, pl. 5: 1-2; the same excavation at Erlitou yielded three other slightly smaller inlaid disks from YLVIK 3 (p. 260, pl. 5: 2). For the Qijia mirror, see Wenwu kaogu gongzuo sanshi nian 1949-1979 文物考古工作三十年 (Beijing: Wenwu chubanshe, 1979)Google Scholar, pl. 14: 4; Kaogu 1980.4, 365, fig. 1Google Scholar; 366-368; Zhongguo meishu quanshu 中國美術全書, vol. 4 Qingtongqi 靑銅器 (Beijing: Wenwu chubanshe, 1985)Google Scholar, no. 1 (color illustration). An Zhimin mentions a second Qijia disk, measuring only 6 cm. in diameter, from the Gansu site of Qijiaping齊家呼in Kuangwu (“Zhongguo zaoqi tongqi de jige wenti,” 278). The striations around the star on the Gamatai disk, similar to those on the kui from Dadianzi, recall the kind of striated patterns ubiquitous to Andronovo pottery decor (see p. 20 and n. 6). This suggests that the star motif on the Qijia disk, which derives ultimately from Bactria, arrived in the Far East by way of an object made by the Andronovo and reflecting their style of decoration.

For a recent discussion of the disk from Gamatai, see O'Donoghue, Diane M., “Reflection and Reception: The Origins of the Mirror in the Bronze Age in China,” The Bulletin of the Museum of Far Eastern Antiquities 63 (1990), 19–20Google Scholar. The author makes no mention of the Erlitou mirror.

90. Yinxu Fu Hao mu, 104, fig. 65: 2; colorplate 12: 2. The diameter of the mirror is 11.8 cm. (p. 103). Lin Yun has pointed out the similarity between the Tomb 5 mirror and one found in Inner Mongolia (“Bronzes of Shang Culture and the Northern Zone,” 253; both mirrors are illustrated on p. 252, fig. 51: 2, 8). Lin Yun stresses that the ׳ Ν Northern Zone’ mirror from Tomb 5 has no prototype at Minusinsk, where the mirrors, all dating from the later Karasuk culture, are undecorated (pp. 251-253). Another small mirror with a decor analogous to that on the Tomb 5 example was found at a site in Pingliang in eastern Gansu and is assigned to late Shang (Wenwu 1991.5, 96, figs. 1-2).

91. See di, V.I. Sari ani, Drevnie zemledeltsy Afganistana (Moscow: Nauka, 1977), 87Google Scholar, fig. 45: 1 and fig. 46: 1, 5 for seal-amulets designed as crosses from Dashli 3. An example from the same site in the shape of a star is shown in fig. 46: 2. Dashli 3 has been dated between 2100-1800 B.c.; see Kohl, Philip, Central Asia: Paleolithic Beginnings to the Iron Age (Paris: Éditions recherche sur les civilizations, 1984), 230Google Scholar. Two recalibrated dates for Dashli 3 are offered by Hiebert, Frederik T. in Origins of the Bronze Age Oasis Civilization in Central Asia (Cambridge, MA: Peabody Museum of Archaeology and Ethnology, Harvard University, 1994), 83Google Scholar. Bronze mirrors of various types are represented at Dashli 3 and other Bactrian sites (for example, Sarianidi, , Afghanistana, 81, fig. 40Google Scholar; Mandelshtam, A.M., “Pamiatniki epokhi bronzi v iuzhnom Tadzhikistane,” Materiaii i issle-dovaniya po arkheologii SSSR 145 (Moscow: Nauka, 1968), 146, fig. 5: 2-4Google Scholar; 147, fig. 6; 148, fig. 7; 149, fig. 8:1.

92. For seal-amulets made of arsenical copper from Altyn-tepe, see Masson, V.M. and Kiiatkina, T.P., “Man at the Dawn of Civilization,” in Kohl, , ed., The Bronze Age Civilization of Central Asia, 111, fig. 2Google Scholar. Examples from the Gonur oasis are illustrated in Sarianidi, V.I., “Margiana in the Bronze Age,” in Kohl, , ed., The Bronze Age Civilization of Central Asia, 178, fig. 7:11, 26Google Scholar.

93. A diagram illustrating the evolution of pottery decoration from Geoksyur I through Altyn-tepe IV is available in Kircho, L.B., “The Problem of the Origin of the Early Bronze Age Culture in Southern Turkmenia,” in Kohl, , ed., The Bronze Age Civilization of Central Asia, 103, fig. 1Google Scholar. The patterns shown in the diagram date from the late fourth to the mid-third millennia (pp. 100-101). See also Kohl, , Palaeolithic Beginnings, 231–233Google Scholar; pls. 5b, 7c, 9a-b).

94. Masson, V.M. and Sarianidi, V.I., Central Asia: Turkmenia before the Achaemenids, tr., Tringham, Ruth (London: Thames and Hudson, 1972)Google Scholar, pl. 25, lower right.

95. Kohl, , Palaeolithic Beginnings, 126Google Scholar.

96. Tosi, Maurizio and Wardak, Rauf, “The Fullol Hoard: A New Find from Bronze Age Afghanistan,” East and West 22.1–2 (03-June 1972)Google Scholar, figs. 4 and 8 following p. 12. The authors tentatively assigned the hoard to the period 2600-1700 B.C. (p. 17).

97. Shanghai bowuguan cang qingtongqi 上海博物館藏靑銅器 (Shanghai: Shanghai renmin meishu chubanshe, 1964)Google Scholar, vols. 1, 2, no. 26. Possible stone prototypes for the Shanghai axe include a large jade yue axe from Erlitou 81YLM 6 belonging to Period IV. The blade measures 23 cm. across and has a large central perforation, but it lacks the extended tang of the bronze version (Kaogu 1984.1, 38Google Scholar, fig. 5: 2; pl. 3: 8).

98. A second large bronze yue formerly in the Eumorfopoulos collection may be as early as the Shanghai axe. The Eumorfopoulos version, which measures 36 cm. from the cutting edge to the top of the tang, is undecorated, and its central perforation is slightly oval (Umehara, Sueji 梅原末治, Ōbei shūho Shina kodō seika 歐米蒐儲支那古銅精華 [Kyoto: Yamanaka and Company, 1933]Google Scholar, part 2, vol. 2, pl. 93).

99. Bactrian sheet-metal vessels may be seen in Sarianidi, , Afghanistana, 80Google Scholar, fig. 39: 4; color plate 1: 8, opposite p. 80; and Ligabue, Giancarlo and Salvatori, Sandro, eds., Bactria: An Ancient Oasis Civilization from the Sands of Afghanistan (Venice: Erizzo, 1988)Google Scholar, pls. 74-75; 77-78; p. 160, fig. 3; 162, fig. 4; 163, fig. 5. Examples from the Norbert Schimmel collection in the Metropolitan Museum of Art, New York appear in Pittman, Holly, Art of the Bronze Age: Southeastern Iran, Western Central Asia and the Indus Valley (New York: The Metropolitan Museum of Art, 1984), nos. 30 and 31Google Scholar; see 61, 64-65 for Pittman's description and commentary. It is important to note that one of the drinking vessels from the Schimmel collection exhibits the use of rivets, which serve to attach separately-made ornamental bird figures to the rim (p. 31).

100. A selection of the Shahdad metal vessels was made available by their excavator Hakemi, A. in Catalogue de l'exposition: Lut Xabis (Shahdad) (Teheran: Premier symposium annuel de la recherche archéologique en Iran, 1972)Google Scholar, pls. 16-19; pl. 20 b; pl. 21 a. The vast majority of the Shahdad vessels await future publication. I am indebted to Holly Pittman for having pointed out to me the existence of the photographic archive of Shahdad finds presented to the Peabody Museum, Harvard University in 1985 by Hakemi and Maurizio Tosi; and I am especially grateful to Professor C.C. Lamberg-Karlovsky for his kindness in allowing me access to this material.

Hiebert argues cogently that the Shahdad material should be regarded as an offshoot of the BMAC: “This assemblage outside of Central Asia can be interpreted as a clear indication of the movement of Central Asian people, probably for resource acquisition through conquest or trade. Ironically, because of the late chronology used by the excavators of the Central Asian sites, the finds of Central Asian artifacts far outside of Central Asia have been perceived as antecedents of the BMAC. …” (Bronze Age Oasis Civilization, 163-164). The Bactrian-Margiana complex as it relates to Shahdad and other sites outside Central Asia is discussed at greater length in Hiebert, Frederik and Lamberg-Karlovsky, C.C., “Central Asia and the Indo-Iranian Borderlands,” Iran 30 (1992), 1–15CrossRefGoogle Scholar and Hiebert, , “Production Evidence for the Origins of the Oxus Civilization,” Antiquity 68/259 (06 1994), 376, 380, 384CrossRefGoogle Scholar. A convienient diagram illustrating the close similarity between artifacts found at Shahdad and at the BMAC sites is provided by Sarianidi, V., “Soviet Excavations in Bactria: The Bronze Age,” in Ligabue, and Salvatori, , eds., Bactria, 121, fig. 13Google Scholar.

The Shahdad finds also raise new questions concerning a puzzling and hitherto unique tumbler-shaped metal drinking vessel (height, 17 cm.) from the collection of Sir Herbert Ingram in the Ashmolean Museum, Oxford. The vessel's prior provenance is undocumented (Shelagh Vainker, personal correspondence, November 9, 1993). Loehr, Max, who published the vessel in Ritual Vessels of Bronze Age China (New York: Asia House Gallery, 1968), no. 1Google Scholar, remarked upon its apparent earliness and tentatively assigned it to “early Shang,” but at the same time expressed some doubt whether it was indeed Chinese (p. 18). No vessels of this type have been recovered from subsequent excavations at either the Erlitou or Zhengzhou sites. On the other hand, a smaller copper tumbler (height 10.5 cm) of precisely the same shape was unearthed at Shahdad, and a second tumbler made from alabaster was found at the roughly coeval site of Tepe Hissar III C in northern Iran (Hakemi, Catalogue de l'exposition: Lut Xabis [Shahdad], pl. 17a [cat no. 260]; Schmidt, Erich F.Excavations at Tepe Hissar Damghan [Philadelphia: University of Pennsylvania Press, 1937], 217, fig. 129)Google Scholar. Because the material culture of the Bactrian Margiana archaeological complex is represented at Hissar III C, as it is at Shahdad, the tumblers discovered at these two sites confirm this vessel form as a BMAC type (for reference to Hissar III C, see Hiebert and Lamberg-Karlovsky, “Central Asia and the Indo-Iranian Borderlands,” 11). But the question whether the Ashmolean tumbler is an actual BMAC vessel, or whether it is a Chinese rendition of the BMAC type remains unresolved. A close examination of the tumbler, however, reveals a single vertical seam up the wall of the body and clear evidence of crimping around the foot (noted during my visit to Oxford, June 30, 1995). The observation that the tumbler is a sheet-metal vessel and, moreover, that the specific features of its manufacture coincide with those exhibited by both sheet-metal and pottery vessels from the BMAC area, favors the conclusion that the vesssel is a BMAC production.

101. Hakemi reports the existence of copper mines in the mountains bordering the Lut Desert and sources of silver ore in the vicinity of Kerman and in the mountains east of the Lut. Copper, he feels, may have been an item of export (Catalogue de l'exposition: Lut Xabis [Shahdad], 11, 15).

102. Compare Fig. 12 with the slightly more advanced jue shown in Henan chutu Shang Zhou cfingtongqi, nos. 2, 5, 7, 9.

103. A brief discussion of the Longshan tradition of the elevated vessel may be found in Huber, Louisa G. Fitzgerald, “The Traditions of Chinese Neolithic Pottery,” The Bulletin of the Museum of Far Eastern Antiquities 53 (1983), 113–125Google Scholar, and “The Relationship of the Painted Pottery and Lung-shan Cultures,” 177-210.

104. Miaodigou yu Sanliqiao, pl. 88, fig. 4; Andersson, “Researches into the Prehistory of the Chinese,” pl. 35: 1. It is perhaps worth observing that one metal pouring vessel from Bactria in the Louvre shows a slight protuberance at one side that conceivably served in gripping the vessel when it was lifted. The Louvre vessel is illustrated in color in V. Sarianidi, Die Kunst des alten Afghanistan, tr. Sabine Grebe (Leipzig: E.A. Seeman Verlag, 1986), no. 69 on p. 193; also see Amiet, P., “Antiquités de Bactriane,” La Revue du Louvre et des Musées de la France 54 (1978), 109, fig. 16:1Google Scholar.

105. An example of a bronze gu from Zhengzhou is illustrated in Henan chutu Shang Zhou qingtongqi, no. 15. For Erlitou pottery gu, see Kaogu 1983.3, 205, fig. 11: 6Google Scholar; >Kaogu 1992.4, 297, fig. 4: 10, 11Kaogu+1992.4,+297,+fig.+4:+10,+11>Google Scholar. The first of these pottery gu, which appears from the drawing to echo the feature of a strengthened, crimped rim characteristic of sheet-metal workmanship, is closer in shape to the BMAC versions.

106. For examples of ceramic lihe from Erlitou, see Kaogu 1984.1, 39, fig. 6: 1Google Scholar; pl. 3: 4 (Period II); Kaogu 1983.3, 203, fig. 9: 7-8 (Period III)Google Scholar; Kaogu 1992.4, 302Google Scholar, fig. 9: 4 (Period ÏV). Apart from the imitation rivets at the top of the handles, there appear others (usually two in number) located at the front of the domed cover in positions suggesting that on the sheet-metal originals they served in the attachment of the upright spout, which would have been fashioned as a separate unit. Also see n. 7 above.

Precursors of the like, lacking any overt sheet-metal traits, have been found in a late Longshan context at the Luoyang site of Cuoli (Period III) in Henan and along with Kexingzhuang II ceramics at the Lintong site of Jiangzhai, east of Xi'an in Shaanxi (Kaogu 1978.1, 11, fig. 12: 6Google Scholar; Kaogu 1975.5, pl. 2: 3), The domed Cover is seen again on a fragmentary vessel, evidently lacking li-legs, from the Longshan stratum at Xiawang-gang, 下王岗 in Xichuan, 淅川, southwest Henan (Wenwu 1972.10, 13, fig. 14:11Google Scholar; p. 18, fig. 42; Xichuan Xiawanggang 淅川下王岗 [Beijing: Wenwu chubanshe, 1989], 260, fig. 253: 2Google Scholar; pl. 91:6).

107. Andersson, J.G., “Researches into the Prehistory of the Chinese,” The Bulletin of the Museum of Far Eastern Antiquities 15 (1943), pl. 37: 4Google Scholar; compare pl. 38.

108. Hakemi, , Catalogue de l'exposition: Lut Xabis (Shahdad), 19 bGoogle Scholar.

109. Hiebert, , Bronze Age Oasis Civilization, 12, 27, 69–70, 134–135Google Scholar. The Fedorovo sherds were found both at campsites along the outskirts of urban complexes and in rare instances within the towns themselves. Hiebert remarks upon the significance of the Andronovo to the Margiana cultures generally: “Interaction with distant nomadic peoples by Period 2 at Gonur is a key to understanding the interaction of the oasis of Margiana with a broader region. … The underlying factor in the development of the BMAC was the ability of the Bronze Age inhabitants of the oasis to bring in raw materials from distant areas. If these Andronovo nomads were involved in long-distance migrations, as is suggested by the nomadic ceramic parallels from distant areas, they may have played an important role in the development of the oasis economy” (p. 135).

E. E. Kuzmina provides illustrations of Fedorovo pottery and two maps showing the distribution of this culture across Eurasia, (Drevneischie skotovodi ot Urala do Tyan-Shanya, figs. 8, 18, following p. 48; figs. 19, 31, following p. 80)Google Scholar. See also nn. 6 and 89 in the present text.

Moreover, in a conversation that took place June 27, 1995, Kuzmina called my attention to a small Andronovo Alakul metal disk from Bylkyldak III bearing the design of a six-pointed star, a motif she considers to derive from the seal-amulets of the Bactrian Margiana complex; see Kuzmina, , “Kultumaya i etnicheskaya atributsiya pastu-sheskikh piemen Kazakhstana i srednei Azii epokhi bronzy,” Vestnik drevnei istorii 1988.2, 52, fig. 2:5 (left)Google Scholar. This observation lends support to the view that BMAC seal-amulets and their designs may have been carried far from their homeland by the Androvono. Disks showing a version of the cross motif occur among the remains from Bylkyldak I (fig. 2:10 [left]).

110. Hiebert, , Bronze Age Oasis Civilization, 77, 79–80Google Scholar.

111. The dissemination of small distinctive metal artifacts across this same vast geographical area in the succeeding Shang period is documented by the trumpet-shaped earrings of Andronovo type found in places as widely separate as Liujiahe 劉家河, near Beijing, and Dashti Kozy, located in Tajikistan on the Zershan graphical area in the succeeding Shang period is documented by the trumpet-shaped earrings of Andronovo type found in places as widely separate as Liujiahe 劉家河, near Beijing, and Dashti Kozy, located in Tajikistan on the Zerashan River fifty miles from Pendjikent. The Liujiahe site can be dated to the beginning of the Anyang period, and a similar date around 1300 B.C. is estimated for Dashti Kozy. The Dashti Kozy site is considered to have been inhabited by the Tazabagyab branch of the Andronovo (Wenwu 1977.11, 6, fig. 18; Isakov, A. I. and Potemkina, T. M., “Mogilnik piemen epokhi bronzey v Tadzhikistane,” Soveteskaya arkheologiya 1989.1, 145–167Google Scholar; figs 5: 6 and 8:1).

In the area intermediate between these two locations, the same trumpet-shaped earrings are encountered east of Dashti Kozy at other Andronovo sites in northern Kirgizia, such as Tash-tyube II near Tal as and Tegirmen-sai in the vicinity of Frunze (Kuzmina, Drevneischie skotovodi ot Urala do Tyan-shanya, figs. 33:1, left, and 34, upper right, following p. 80).

In China, an additional earring of this type has been recovered in the Tangshan region of Hebei, and at least two sites in the northeastern province of Liaoning have yielded further examples: Pin'ganbao, 平安堡 in Zhangwu, and Pingdingshan, 平頂山 in Fuxin, (Kaogu xuebao 1985.2, 140, fig. 4: 8Google Scholar; Kaogu xuebao 1992.4, 452, fig. 14: 19Google Scholar; Kaogu 1992.5, 403, fig. 8: 5Google Scholar; pl. 1: 6). Pingdingshan I, like Dadianzi two hundred kilometers to the west, is assigned to the Lower Xiajiadian culture. If Pingdingshan I is close in time to Dadianzi, then the earring unearthed there would constitute evidence of a link between northern pastoralists of Andronovo affiliation and the local populations of northeast China at the time of the Erlitou culture.

112. Kohl, , in Chernykh, , Ancient Metallurgy, xviGoogle Scholar.