do not know. The two locations were in the dry Yellow River drainage of the north and the moist, warm Yangzi drainage of the south (on Neolithic and early urban China, see Liu and Chen 2012; Underhill 2013). In the former, millets were the domesticates. “Millet” is a catchall term for any grain with small seeds—significantly smaller than those of wheat. There are a good half dozen millets in China, not very closely related to each other. Foxtail millet (Setaria italica) was probably domesticated first, being a better grain all round, but panic, or broomcorn, millet (Panicum miliaceum) was almost or quite as early. Both are well adapted to the dry, summer-hot, winter-cold north. Broomcorn millet spread rapidly west across the steppes, reaching Europe by 4000 BCE. In China, it was never adopted far from the dry northern interior. Foxtail did not spread west till much later; preferring more rain and warmth, it moved south instead, becoming important throughout China in later millennia. It eventually became a minor but significant grain in the uplands of Southeast Asia and locally in Central Asia and Europe.
In north China, agriculture began by 8000 BCE, possibly before 9000 (see, e.g., Crawford and Shen 1998; Higham and Lu 1998; Liu 2004; Liu and Chen 2012). Foxtail millet was domesticated by 8000–7500 BCE (K.-c. Chang 1999: 44–45; Liu 2004; X. Liu et al. 2009; T. Lu 2005; see Sagart et al. 2005, passim; Yan 2002; Zhao 2011). The early center of foxtail millet agriculture is a large area, from the Wei River Valley down the Yellow River and then south into the Huai River drainage.
Domesticated broomcorn millet may go back to 8500 BCE, but the finds are not securely attested (Zhao 2011). It was domesticated in northern China or Central Asia, somewhere between the Aral Sea and the Tian Shan. Genetic comparison of existing strains seems to pinpoint domestication there, and this seems logical given the early appearance of the grain in both China and Europe (Kimata 2012 and pers. comm.). Along the Yellow River, the Cishan culture may have been growing Panicum miliaceum as early as 8300 BP (around 6500 BCE; Bettinger et al. 2010: 703; Zhao 2011).
“Panic” is just the Latin root for “millet”—it has nothing to do with the Greek word for extreme fear. (The latter came from the god Pan; he keeps people away from his favorite spots by giving them an irrational terror when they go there.) Millet is a pretty obvious thing to domesticate; in pre-Columbian times, a foxtail millet was briefly domesticated in Mexico, and P. sonorum was independently domesticated on the Lower Colorado River. Interestingly, foxtail millet was replaced by maize in Mexico around 3000–2000 BCE, and then in China in the nineteenth and twentieth centuries CE. Also, several related (but different) millets were independently domesticated in Africa and India.
New data show that agriculture may even have been earlier than that. Starch grains on pots and grinding stones show extensive use, implying cultivation and some dependence, by 9500–9000 BCE at Nanzhuang tou, and 9500–7000 at Donghulin (Xiaoyan Yang et al. 2012). If this evidence means what it seems to mean, agriculture was at least as early in north-central China as in the Near East.
For this time period and later, Ian Morris, in his “big history,” Why the West Rules … For Now, provides a chart purporting to show that the Near East was always earlier than China is every advance except pottery and “rich grave goods” (Morris 2010: 130; he thinks the reason pottery came earlier in the East was that the Easterners boiled food more). Unfortunately, much of this is at best speculative (including the boiling). Leaving aside some errors, the whole chart depends on the luck of the excavation. Archaeology is far more developed in the Near East. Moreover, China’s archaeology is handicapped by the intensive occupation of the landscape. It is good for the residents that China’s first real city—Zhengzhou—persists today as a flourishing metropolis, far bigger than it was in 1500 BCE, but it certainly does not help archaeologists! The Near East was ahead in many things, but the domestication of plants may not have been one of them.
Both kinds of millets were widespread and basic to many local cultures by 6000 BCE. The Peiligang culture, in the upper Huai drainage, flourished in 7000–5000 BCE and, along with Cishan, “signals the emergence of food production and ritual complexity in the region” (Liu et al. 2010: 816–17). However, much of the food of the Peiligang people comprised acorns, as shown by abundant acorn starch grains in their well-made grinding stones (Liu 2012; Liu et al. 2010). Acorns are still somewhat widely used in China; in ancient Zhejiang they were made into an acorn jelly (Liu et al. 2010: 830), evidently similar to that which remains a common food in Korea today. In any case, in Peiligang the acorns and wild yams of an earlier age were suddenly and dramatically supplemented (but not replaced!) by great quantities of domesticated millet and rice (Liu 2012). The Peiligang and other early cultures had small settlements, 1–8 ha in size. Agriculture reached Inner Mongolia by around 6000 BCE (Xinglongwa culture; Shelach et al. 2011). Here and elsewhere, tree crops were so important in those days that Li Liu suspects deliberate tree management—resource husbandry—as in ancient California (Liu and Chen 2012: 265–67).
Around 5500 BCE, people, pigs, and dogs in central north China suddenly shifted toward eating a lot more millet. One way we know is that their bones all show markers of subsistence on plants that use the C4 pathway of carbon metabolism (Barton et al. 2008; Jing and Campbell 2009). (C4 is found largely among tropical grasses. Most other plants use the C3 pathway.) In this area, the only common C4 plants are millets, so this is evidence for reliance on agriculture. Wild plants and other cultigens in the area are C3. The only other important C4 plant in China is maize, which did not reach China until the sixteenth or seventeenth century CE. Jing and Campbell (2009: 101) report a very odd case of two skeletons showing a C3 diet among the many showing C4. Were these strangers? Hunter-gatherers from the uplands? Migrants from rice regions to the south?
At Dadiwan, we have the unique advantage of an almost continuous record of 80,000 years. Dadiwan is in the dry loess plateau lands (around 20” annual rainfall) of the Wei River drainage, northwest of Xi’an, but the climate was wetter during at least some of the Neolithic period. The site shows millet agriculture appearing slowly from 5500 BCE and intensifying between 5000 and 4000 into full Neolithic (Bettinger et al. 2010). Most of the loess plateaus of interior China were grassy or brushy, with sagebrush steppes and wild jujube scrub. These dominated on level lands. On loess soil, rainwater seeps in quickly and deeply, leaving the surface both dry and fire prone. In areas as dry as this, grass takes over. The steeper slopes were brushy, because water ran off too fast to allow much tree growth. (In addition to cited sources, I have my own observations of the loess plateaus to go by, as well as careful scanning of satellite photographs. For magnificent photographs of Chinese Neolithic sites and objects, see Yan 2002; Zhang Zhongpei 2002.)
However, the vast loess plateau is broken by many valleys and ravines and by higher hills and mountains. These were, and sometimes still are, densely forested. At Dadiwan (which has an archaeological record from 6000 to 1800 BCE) and nearby Xishanping, there is a good record of pollen and charcoal from 3200 to 2200. It reveals that the area had surprisingly diverse and rich forests, dominated by maple, elm, oak, and similar trees (Liu et al. 2013). Hazelnuts, chestnuts, wild cherries, and acorns from the oaks would have provided food. Most of these were probably on the hill and mountain ranges. Spruce and birch were common higher on the ranges, indicating cool moist conditions there. A wetter climate had also allowed warm-temperate plants like bamboo and sweetgum to flourish in the valleys, now totally farmed. Today, any area not too steep to be terraced is now used intensively for agriculture. This area is now cold and dry.
Rice (Oryza sativa) was domesticated somewhere in or near the Yangzi River drainage. Theories of Southeast Asian or Indian origin and of multiple sites of domestication have now been disproved; recent archaeology and genetic analysis (Molina et al. 2011) suggest that domestication was a single event that occurred in central China around 6000–7000 BCE.
Rice was cultivated and very possibly domesticated by around 8000 BCE (K.-c. Chang 1999: 46; Jiang and Liu 2006, earliest site, Shangshan in Zhejiang; Liu 2004; T. Lu 2005, 2011; MacNeish and Libby 1995; Yan 2002). Crawford (2006), Zhao (2011), and many others doubt domestication by this early date, finding certainty only by 6500 BCE, but Kuzmin (2008a) has definite evidence for it by 7000 BCE. It is now clear that China, specifically the Yangzi Valley and environs, was the place of origin of domesticated rice and of rice agriculture, though rice was quite early in the Yellow River drainage