The extensive set of direct dates, on the largest early assemblage of wheat and barley in China, provides important new evidence on the arrival of West Asian crops, and western stimulus, into China. Rowan Flad, Li Suicheng, Wu Xiaohung and Jimmy Zhao, have recently reported new archaeobotanical evidence and AMS dates from the Gansu corridor site of Donhuishan in short article in The Holocene, "Early wheat in China: Results from new studies at Donghuishan in the Hexi Corridor." Importantly, the evidence includes the first large assemblage of wheat rachis remains. These provide the first clear confirmation of what has long been taken for granted, that the early wheat in China is only hexaploid bread wheat, leaving the eastern margins of tetraploid naked wheats (durum) somewhere in central Asia and to the south in India. This article also provides an updated review of all the early wheat finds that have been published from China.
In my version of their map (left), I have colour-coded wheat reports by broad period. First of all, it can be seen that finds line up along the eastern line of the classical silk road, running through the Gansu corridor and along the Lower Yellow river basin. It should noted that while the earliest finds are all attributed to 2500-2000 BC, or even more than 2500 BC in the case of Xishanping, none of these earliest finds is directly AMS dated. The case on Donghuishan reported by Flad et al., provides a warning call against putting too much faith in single or few or associated dates, as the earlier evidence from Donghuishan has suggested the wheat could be closer to 2500 BC rather than the 1600-1500 BC age indicated by numerous dates, including 4 directly on barley (but not wheat). Nevertheless it still seems plausible that wheat and barley entered China by ca. 2500 BC, even if the wheat crop did not take off until closer to 2000 BC. This period of arrival ion China is paralleled by the adoption from the west also of sheep, cattle and probably copper metallurgy (with a possible parallel spread to Southeast Asia-- as argued recently by Whyte and Hamilton). The counter current was provided by Chinese millets, as both Panicum miliaceum and probably Setaria italica arrive in northwestern India around this period, and Panicum is also reported in Yemen (ca. 2200 BC) and in Sudan by ca. 1700 BC. On the dispersal of westwards and southwards through the Indus to Arabia and Nubia, see discussion the paper "Cattle, Crops and Commensals" that I recently published with Nicole Boivin in the French periodical Etudes Ocean Indien [pdf]. I have also argued that japonica rice followed this route west from the Yellow river and in India (see discussion in my article in the recent rice issue of Archaeological and Anthropoloigical Sciences)
Interestingly the barley from Donghuishan, like that from Xishanping, is notable since most sites in China that have yielded wheat have lacked barley. This indicates that the adoption of wheat went through a strong cultural filter in which is was only wheat rather than wheat and barley that was adopted in much of central China. This provides a curious contrast from other regions of Asia, whether west Asia, central Asia of South Asia where wheat and barley are almost always found together archaeologically. In India these two winter cereals are also often found with evidence for pulses crops like lentils, peas or chickpeas. None of these Southwest Asia crops appears to have made it into prehistoric China. Thus the diffusion of crops into (and out of) China was selective process of cultural choice.