The River, the Plain, and the State

by Ling Zhang

  Millet had dominated dry-land farming in north China for many centuries before wheat became a significant staple grain. Millet was the grain “constantly eaten [by people] in the northern lands (north China in general).”23 In the Tang Dynasty, millet served as a standard for evaluating the weight and price of other grains.24 A fall crop, millet was sown by the third lunar month in mid-spring, it sprouted in late spring and ripened in mid-autumn. In Hebei during the eleventh century, the harvesting of millet usually started around the middle of the eighth lunar month and finished by the ninth lunar month, before the frost descended in the tenth month. Given geographical and climatic differences, millet was harvested ten to fifteen days earlier in southern Hebei than in northern Hebei and the western mountainous areas. Millet yields did not improve much even into the mid-twentieth century, when the unit yield was around 60 kilograms per mu of land. During the Tang–Song period, the unit yield was between 40 and 50 kilograms. In places where labor and water supplies were limited or the weather was cooler, such as in northern Hebei, the production tended to be even smaller. Despite its limited yield, millet was nevertheless the most favored crop for dry-land farmers. Its stability, low water consumption, and high resistance to the dry climate made it a sturdy crop that did not demand much human care.

  Millet's predominant role in northern agriculture seemed to be challenged by winter wheat in the Tang period. Agricultural historians have suggested that winter wheat was widely planted; because food produced with wheat flour tasted better, it was more appealing to consumers than millet.25 More importantly, as a winter crop, winter wheat was sown in the eighth lunar month and harvested in the late fourth and fifth month of the next year. It utilized the winter months, when other grain crops could not grow, and thereby it added one more crop to the land's annual production. Hence, it dramatically extended the land's growing season and sped up the rotation of crops on a same swath of land: in the past, a unit of land could support one grain crop per year, but with winter wheat, it could produce three grain crops in two years, pushing the annual rotation index from 100 percent to 150 percent. Thanks to winter wheat, a unit of land could produce more food every year; such increase in land productivity was remarkable. Given these virtues, scholars have considered winter wheat a revolutionary crop, which gradually replaced millet as north China's leading crop.

  Combing through Tang–Song historical sources, however, we find that in comparison with millet, winter wheat had many disadvantages. Given those disadvantages, the lower socio-economic tiers of the population as well as regions like eleventh-century Hebei, which suffered various environmental and socio-economic difficulties, might not be able or willing to pursue large-scale wheat plantation. Without denying that the crop was a technological innovation that could improve land productivity and that was cultivated more in the Tang–Song period than before in some parts of China and by certain kinds of farmers, I argue that previous scholars have idealized the adoption of the crop and disregarded environmental and socio-economic specificities and their constraints. An advanced technology did not make a leap into mass production easily; its application in actual production was confined by various conditions. It was certainly subjected to the opinion of farmers – as individuals in the context of small farming communities – who judged whether or not they had the capital to pursue the new technology, whether or not their production would be more profitable with the technology, and what costs would incur. In a society like late tenth-century Hebei, where agriculture had just recovered from war damage and was growing at a steady, slow pace, and where environmental and socio-economic challenges led to an economy based on a principle of “subsistence first” rather than “profit maximization,” farmers would more likely choose traditional, low-risk crops like millet than high-risk, high-cost crops like winter wheat.26

  Let us look at winter wheat's disadvantages. Firstly, in comparison with millet, wheat demanded more nutrient and water resources and was more sensitive to climatic variations. In fields of similar soil conditions, millet required less fertilizer than wheat and made more efficient use of water.27 Wheat was far less drought resistant. It had to be planted in low-lying fields that could be irrigated.28 On the Hebei Plain where water was constantly in short supply and artificial irrigation was less developed (certainly in middle-period China), dryness threatened all stages of the growth of wheat plants.29 This meant that cultivating wheat was a challenging and risky business; it required more human care and capital input than millet did.

  In spring, when wheat sprouted and desperately needed water, the rainy season had not yet arrived and “dry hot winds” blew across the North China Plain. The rapid rise in temperature and dryness caused the fragile plants to wither. When droughts hit north China (including Hebei) hard in the 1030s, 1040s, 1060s, and 1070s, winter wheat must have been vulnerable. Also, in the 1040s, 1055, 1089, and 1090, warm winters led to a shortage of snow.30 Without snow to cover the ground and shield the underground seeds from the cold air and retaining moisture, wheat seeds would fail to sprout. After 1048, floods from the Yellow River and Hebei's local rivers took place nearly every year, sometimes even several times within a single year. While early floods happened in the fourth and fifth lunar months to damage the harvest of winter wheat, later floods often came in the seventh and eighth months to prevent the sowing of winter wheat. Clearly, it was not easy to carry out the cultivation of winter wheat in a sustainable manner in eleventh-century Hebei.

  Secondly, while consuming more labor and water resources than millet, winter wheat did not produce a more desirable yield. Based on numbers collected from various sources, winter wheat's unit yield appears to have been ten kilograms lower than millet (see Table 4). Hebei farmers had to reason carefully about whether or not it was worthwhile to pursue a smaller yield for winter wheat at higher costs. Given the escalating environmental pressure and socio-economic hardship, planting winter wheat did not appear to be a secure, profitable practice for ordinary people.

  Table 4. The Unit Yield of Winter Wheat and Millet (in kg)

  Yield in HebeiAfter D. H. PerkinsaAfter W. Wagnerb earlyAfter R. H. MyerscAfter J. L. Bucke

  Crop1000s1600–19001900s1917–19571932d1929–19331948–1952f1953g

  Winter 33.3 43–73 62 67 46

  Wheat

  Millet 40–50 52.4–66.7 72 96 68.7

  Notes:

  a Perkins (1969: 330); b Cited from Bray (1984: 448); c cited from Bray (1984: 476); d Cressey (1955: 100); e Buck (1937: 204); f Zhongguo nongye baike quanshu bianji weiyuanhui (1991: 611); g Zhongguo nongye baike quanshu bianji weiyuanhui (1991: 523). Bray (1984: 476) finds that Perkins underestimates the yields and believes in higher yields in Hebei and Shandong.

  The third disadvantage contradicts previous scholars’ belief that winter wheat brought about an improved rotation pattern; that is, three grain crops in two years was achieved in north China since the mid-Tang Dynasty.31 I do not challenge the technical validity of this rotation pattern. But I argue that, at least in Hebei, this pattern could work only in optimal conditions and was not widely practiced in reality. In this matter, timing was a major obstacle. Winter wheat was sown by the end of the eighth lunar month in Hebei, to be done earlier than in other regions, because temperatures dropped quickly in Hebei. The earlier it was sown in that month, the fewer seeds were required and the healthier sprouts would be, meaning lower costs for higher returns.32 However, in that month the millet harvest had not finished. As fields were still occupied by millet and had not yet been cleared of debris, they were not ready for the plantation of the winter crop. Frequent flooding from the Yellow River and other local rivers in the autumn certainly made the sowing of winter wheat more difficult. In the meantime, most farmers were busily engaged in harvesting and processing fall crops, and managing their payment of autumn taxes. We certainly should not forget that by that time a significant portion of Hebei's male population was getting ready for their winter military training, leaving the elderly, women, and ch
ildren to handle a great deal of the agricultural work. Thus, it is doubtful that there was sufficient labor for both the fall harvest and wheat plantation at the same time. Now, let us look at the summer time. The harvest of winter wheat was carried out in the fifth lunar month, when young shoots of millet had already broken through the earth in neighboring fields. The wheat fields would only become available for cultivation again in the late fifth month or the sixth month, either for planting a vegetable crop or fast-growing beans, or simply to lie fallow until the eighth month when they could be used for winter wheat again.

  Obviously, it is very unlikely that millet and winter wheat were rotated on a same field to achieve the intensified use of land in the three grain harvests over two years system that the previous scholarship has suggested. A more plausible farming practice would have been that in addition to the predominant millet plantation, which rotated at best one crop a year,33 a certain number of fields was used for winter wheat, with two minor vegetable crops immediately before and after the wheat crop. Such practices would bring the crop rotation in some fields up to three crops within two years, but with only one grain harvest, far less than three grain crops that some previous scholarship has idealistically suggested. In fact, even this modest practice – one wheat crop and two vegetable crops in two years – could not be achieved all the time. In 990, for instance, some 680 qing (68,000 mu) of summer crops (e.g., winter wheat or peas) in southern Hebei were damaged by drought; by the eleventh month of the year, merely 130 qing of the fields held fall crops.34 In this peculiar case of crop rotations, only 19 percent of the fields were utilized again for the next season. The remaining 81 percent lay fallow from the summer through the winter. Across Hebei, the ratio of fields that afforded three crops in two years in a sustainable manner could not be high.

  In the eleventh century, especially after 1048, there were many years when crops were harmed by disasters and the land failed to support additional crops. The changes in environmental conditions often led to irregular practices. For fragile crops like winter wheat, a delay in sowing of even ten or fifteen days might cause a substantial reduction in the eventual harvest. Historical records show that farmers often planted spring wheat instead of winter wheat.35 Like millet, spring wheat was sown in the spring and harvested in the fall. In contemporary China, this crop usually produces a lower yield than winter wheat, and its flour is of an inferior quality.36 It is only planted in northeastern provinces where temperature is too low for wheat seeds to survive the winter months. In eleventh-century Hebei, however, spring wheat was widely planted. In 1074, for instance, an official document mentioned that there were “households that ought to plant spring wheat (ying zhong chunmai hu 應種春麥戶)” in Jizhou of central Hebei.37 This suggests that spring wheat was not a minor crop, but more likely a major, regular staple crop in the economic life of some people. The significance of spring wheat also suggests the unpopularity of winter wheat in Hebei, due to the crop's fragility and unstable yield against Hebei's peculiar environmental conditions.

  We should note that, apart from the timing issues, there were many other factors determining whether or not a more intense crop rotation was possible. As agricultural historian Francesca Bray has pointed out, intensified land utilization can only be realized by “a consequent increase in the use of fertilizer; by more thorough tillage, improving soil structure and fertility; by careful regulation of the water supply; and by better care of plants, weeding frequently, hoeing and watering to ensure that each single stem bears fruit.”38 Not all farming households could afford such inputs of material and labor resources. In practice, “[t]he abbreviated growing season, together with other problems – lack of irrigation, lack of fertilizer, and alkaline soil – has severely limited the possibilities of land use over the winter months.”39 In eleventh-century Hebei, in particular after 1048, the population loss to the military, disasters, and various kinds of socio-economic turmoil sharply reduced labor supplies. The overall impoverishment of the Hebei society, among not only lower levels of the population but also among the landholding, wealthy upper classes, prevented farmers from purchasing seeds and fertilizer or from constructing irrigation infrastructure. A great number of people were at the edge of poverty and suffered frequent food shortages. Those people did not possess much capital. Not many of them could afford to experiment with intensifying land utilization and cultivating high-cost, high-risk crops like winter wheat. They were more likely to rely on traditional technology and produce stable, low-cost crops like millet, despite its low quality and inferior taste.

  Given the obstacles imposed by timing, environmental conditions, and resource requirements, a rotation pattern of three grain crops in two years could not be applied widely in real farming practices in eleventh-century Hebei. In fact, this rotation pattern was not achieved in Hebei even in the early twentieth century. Studies show that the rotation index in central Hebei in the 1930s reached only 126 percent, and the average level for all of Hebei was even lower, at a mere 122.56 percent. These numbers suggest a cultivation of five crops over four years.40 The kind of revolutionary growth in agriculture based on an intensified use of land and a consequential surge in land productivity and overall production – like what previous scholars have believed in winter wheat and what the early-ripening paddy rice brought to southeast China – did not happen in the disaster-ridden land of Hebei. A technological innovation, like adopting winter wheat, was unable to blossom into mass production, given the unfavorable environmental and socio-economic conditions.

  With all of this in mind, a more realistic assessment can be made about winter wheat. Focusing on the temporal dimension of agricultural practices, previous scholars have considered winter wheat a supplementary crop that filled in the winter gap between two major crops. The above pages have proven this temporal dimension invalid; winter wheat and millet could not rotate one after another on a same patch of field in Hebei. Now stressing a spatial dimension, my analysis suggests a different possibility: winter wheat was a competitor to the dominant grain, millet, because it competed for both land and other resources, like labor, water, and fertilizer. Rather than fitting into the existing rotation pattern that centered on millet, supplementing millet and accelerating the year-round rotation, winter wheat occupied certain fields that were originally for millet cultivation. Hence, the issue becomes that, as the acreage of winter wheat rose, the acreage available to millet decreased. As wheat fields produced wheat (at a smaller unit yield than millet) and two vegetable crops every year, the overall size of millet fields shrank. Moreover, some resources (such as water, fertilizer, and labor) originally dedicated to millet cultivation switched to serve wheat fields, so the remaining resources for millet fields had to decrease. As a result, in an area where the plantation of winter wheat took place, the overall production of millet had to drop. The popularity of winter wheat came at the cost of the millet production. In this situation, the quantity of total staple foods produced from the same land was not doubled or tripled, as happened with the intensified production of paddy rice in the lower Yangzi valley. At best, the total output of staple foods made available to human consumption by the millet-winter wheat combination remained at a similar level as in the past when millet was the dominant staple grain, if not lower, given the fact that winter wheat had a small unit yield.

  Therefore, the change to agriculture in north China did not happen in terms of the quantity of staple food – people did not get more food to eat. Rather, it happened to the food structure, the changing composition of various grains and their differentiating consumers. Even given a fairly consistent quantity of staple food, there was a proportional adjustment among various grains. A growing number of the economically well-to-do, in particular urban citizens, began to consume more food made of wheat flour, such as noodles, steamed buns, and dumplings.41 Wheat production and consumption might have increased, but the increase was not revolutionary. In less urban or rural areas, within the lower tiers of the society, and amo
ng those who were environmentally and socio-economically challenged, people could neither afford to cultivate winter wheat, nor to consume wheat flour on a daily basis. Millet still dominated their subsistence-oriented staple food production and consumption.42

  Let's now compare two groups of tax quotas from the late eleventh century to see how significant the two crops contributed to the state finance. In the Northern Song, the state levied two major taxes, respectively in summer and fall. The summer tax included grain (mainly winter wheat and mungbean in north China versus rice and winter wheat in south China), grass (like straw), silk, and cash in copper coin. The autumn tax included grain (mainly millet and soybean in north China versus rice in south China), grass, and cash. Table 5 shows empire-wide summer and autumn tax quotas in 1077. These numbers suggest that, throughout Song China, people were expected to pay double the amount of tax in the fall as in the summer. The grain they submitted to the government in the fall, a combination of millet-beans-rice, was 4.2 times the wheat-beans-rice combination that they submitted in the summer. Table 6 shows the tax quotas for Hebei in 1080, which indicates an even sharper contrast. People of Hebei were supposed to pay an autumn tax 5.6 times the amount of the summer tax. Their autumn tax of grain, namely millet and soybean, was 11 times their summer tax of grains, namely wheat and beans.