Part I
Journal articles and book chapters

Changes in Climate, Land, and Human Efforts: The Production of Wet-f iled Rice in Jiangnan During the Ming and Qing Dynasties1

Primary Remarks2

Farm output results from a combination of human exertions and the natural environment, the most relevant aspects of the latter being climate and land. Traditional Chinese agricultural theorists referred to climate as “Heaven” (Tian 天), “the seasons of Heaven” (Tianshi 天时), or “the Way of Heaven” (Tian Dao天道). Land was “Earth” (Di 地), “benef its drawn from the Earth” (Di li 地利), or “the way of the Earth” (Di Dao 地道). Human efforts were “man” (ren 人),“the coordination of the people” (renhe 人和),3 “human doings” (renshi 人事), or “the Way of men” (ren Dao 人道). Together they constituted the three basic elements of farm production.

With this doctrine of the “Three Powers” (san cai 三才) as its nucleus, traditional Chinese agricultural theorists produced penetrating explanations of the interactions between the three elements just mentioned and farm output.4 By the Ming and Qing Dynasties, the thinkers and agronomists of the lower Yangzi Valley ( “Jiangnan” ) had elucidated this further. It was the view of Zhang Lüxiang 张履祥 that “the essential concern of agriculture and sericulture is to make use of the Way of Heaven, to depend on human labor power, and to promote the benef its provided by the Earth. This is an absolute truth.”5 According to Lu Shiyi 陆世仪:

It is not only in military matters that one must take into account seasons and weather, the advantages afforded by terrain, and appropriate human action. This applies to everything, and with particular force to farmland. “Seasons and weather” refer to flood and drought, “advantages afforded by terrain” to whether the soil is fertile or barren, and “appropriate human action” to the opening up of land and its good management ... We have to place the greatest weight [of practical concern] on appropriatehuman action, with land in second place, and the weather last. But [climate and] the weather is the element of the greatest [ultimate] importance among the three, followed by land, and with human action last. Thus, if rainfall and sunshine are appropriate, poor-quality land can give as good a yield as top-quality land. If the land is richly fertile, a low-level farmer can get as good a yield as a top-level farmer. The reason for this is that there are big differences in labor inputs in the above cases (lao-yi dun shu gu ye 劳逸顿殊故也). Nonetheless, if one has the advantage of [good] weather and of [good] land, and adds to them (ji zhi 济之) with appropriate human action, one will even so always reap more than others. This is why we must honor appropriate human action. 6

This provides a relatively correct explanation of the roles of climate and weather, land, and human efforts, while further elucidating, at least to a f irst approximation, the existence of certain types of mutual dependence, control, and substitutivity among them.

These elements are always in a process of ongoing change. Thus if one of them changes, this induces changes in the others, leading to changes in agriculture. In this sense there is no “stagnation” in agriculture, and traditional Chinese farming provides no exception to this rule. Since China has long been an agrarian country, agriculture has occupied the most important position in the national economy; an inquiry into the changes in the three main elements that constitute agriculture, and the influences these changes have had, is of special signif icance for its socioeconomic history.

The present story focuses on Jiangnan during Ming-Qing times because of the comparative abundance of historical documentation. The period may be more precisely def ined as the late Ming, from the Jiajing reign (1522-66) until the end of the dynasty, and the early and middle Qing, from the Qing troops entered Shanhaiguan until the end of the Daoguang reign in 1850. In other words, from the early sixteenth century to the middle of the nineteenth. This starting point has been chosen partly because prior to this date documentation is comparatively sparse, and partly because, in the view of many scholars, the development of the Jiangnan economy went through a turning point broadly around the Jiajing reign, so that it was clearly different afterward from what it had been before.7 The reason for placing the terminus in the Daoguang reign (1821-50) is that this reign was followed by the most extensive socioeconomic destruction witnessed by the region for two millennia, in the form of the Taiping War.8 These points of chronological demarcation are clearly broadly identical with the period that most Mainland historians call “the sprouts of capitalism.” I shall leave this latter question to one side, but it may be aff irmed that the development of the Chinese economy during this period, especially in this region, has its own particularity, and constituted a distinctive phase, which makes this periodization a rational one.

The term “Jiangnan” used in this chapter is another name for the Yangzi Delta or the drainage basin of Lake Tai. It comprises the eight Ming-Qing prefectures of Suzhou (plus the Taicang department, which was split off from Suzhou), Songjiang, Changzhou, Zhenjiang, Yingtian (or Jiangning), Hangzhou, Jiaxing, and Huzhou. The principal reason for selecting these boundaries is that these prefectures are cut off from the areas outside them by the Yangzi River to their north, the Qiantang River to their south, the ocean to their east, and several lines of mountains to their west. Internally they form a dense network of waterways consisting of a lake (Lake Tai), two streams (the Jing 荆溪 and the Tiao 苕溪, the main sources of Lake Tai), “Three Rivers” (san jiang 三江),9 “Five Lakes”(wu hu 五湖),10 and numerous creeks, canals, lakes and marshes, and dammed reservoirs. In consequence, this region not only has a high degree of hydrological and ecological unity, but is also an economic area with unusually dense internal interrelationships. In Ming-Qing times people saw its characteristic economic and cultural traits as setting it apart from the other areas of Jiangsu and Zhejiang provinces, hence the term “Jiangnan,” and the justif ication for treating these eight prefectures as a distinct unity. 11

Changes in Climate, Land, and Human Action in Jiangnan During the Ming-Qing Period

Climate should not be confused with weather, and refers to the state of the long-term macroclimate of the earth. The most important components of climate that have a close connection with agriculture are the energy of solar radiation, duration of sunlight, temperature, and precipitation. When the present chapter (somewhat metaphorically) speaks of “earth,” this not only includes such external characteristics as geology, geomorphology, and hydrology, but also such internal properties as the nature of soils, their structure, composition, fertility, depth of arable layer, permeability, and acidity or alkalinity. Agronomists of China’s mainland, normally consider the key elements of climate and land as light, heat, oxygen and carbon dioxide, soil, water, and fertility. 12 By combining with each other in determinate patterns these elements collectively constitute the basic natural environment of agriculture. Our inquiry into the changes in the environment of wet-f ield rice production in Jiangnan during the Ming-Qing period will essentially concentrate on the changes in these elements. When we speak of changes in “human efforts” we shall likewise focus on such aspects as large-scale improvements in hydraulics and land management, and advances in cultivation technology and their diffusion.

Changes in Climate

Zhu Kezhen 竺可桢 did the initial work on changes in the climate in China (and particularly in Jiangnan) during the last f ive hundred years.13 This f ield of research has flourished during the last ten years but seems to have concentrated on fluctuations in temperature and on disasters due to floods and droughts.

In 1972 Zhu Kezhen showed that between 1400 and 1900 China had two periods of warm winters (1550-1600 and 1720-1830) and three periods of cold winters (1470-1520, 1620-1720, and 1840-90). The seventeenth century was the coldest of these latter, especially between 1650 and 1700. Zhu’s sources were mainly local gazetteers from eastern and central China, and thus his conclusions may be taken as representing the general trend of changes in temperature in Jiangnan.14 Research undertaken from a different angle during the 1980s conf irmed that the trends signaled by Zhu were in the main correct.15 Minor divergences aside, researchers agreed that the seventeenth century and the middle of the nineteenth century were cold periods in China (including Jiangnan), and that the later sixteenth century and the eighteenth century were warm periods. It also became apparent from some of their work that the mean annual temperature in Jiangnan during the eighteenth century was about 1℃warmer than it had been in the seventeenth. 16

On the basis of many years of meteorological records it is broadly true to say that for every additional 100 meters in height the mean annual temperature falls by 0.45℃ (about 0.42℃ in winter and 0.51℃ in summer), while precipitation increases and hours of sunlight and radiant energy decrease. Thus, if the mean annual temperature were to fall by 1℃, the climate of the present-day Jiangnan plain would be close to, or slightly worse than, that of the present-day hill country in southern Zhejiang. In other words, the mean annual temperature would fall from above 16℃ to below 15.5℃; the total annual warmth received in the range between 10℃ and 20℃ would decrease from 4,100℃ to below 4,000℃; the mean quantity of radiation would decline from above 110-14 to below 101-05 kilocalories per square centimeter; the annual hours of sunshine would diminish from in excess of 2,000 hours to under 1,900; while annual precipitation would increase from less than 1,200 mm to more than 1,500 mm.17 The climate in the eighteenth century was marginally cooler than it is today, from which it may be deduced that the climate of the Jiangnan plain in the seventeenth century was slightly more cold and overcast, and more humid, than that of the hilly districts in the mountains of southern Zhejiang is at the present time.

Zheng Zhaojing’s 郑肇经 research on floods and droughts in Jiangnan during the Ming-Qing period suggests that floods, and hence a rainy climate, were predominant in the four hundred years from the sixteenth to the nineteenth century, apart from 1523-68 and 1625-62, which were dominated by droughts, and hence a dry climate. During the last flood-dominated period, however, namely that from 1663 to 1900, the relative annual frequency of floods was only 60 percent, whereas in the two preceding rainy (flood-dominated) periods the corresponding rates were 85 percent and 90 percent. Likewise the percentage of “major floods” out of all floods in the f inal period was 29 percent, lower than the 36 percent and 32 percent levels for the two preceding periods. The frequency of floods during the Ming Dynasty (1368-1644) was one in 3.7 years, while during the Qing Dynasty (1644-1911) it was one in 4.0 years.18 The same trend of falling frequency from the Ming to the Qing is shown by Fang Zaihui’s f igures on floods in the three prefectures of Hangzhou, Jiaxing, and Huzhou, namely one flood every 1.3 years under the Ming, and one in every 1.7 years under the Qing.19 Exceptionally severe floods and droughts, defined as those affecting more than two-thirds of the entire area, had the following pattern: in the sixteenth century there were three such floods and two droughts; in the seventeenth century there were two such floods and two droughts, both types of disaster manifesting on these occasions the greatest severity for the entire period; in the eighteenth century there was one such flood; and in the nineteenth century there were two such floods and two droughts. 20

There was thus a degree of correlation21 in Jiangnan between the changes in rainfall implied by Ming-Qing flood and drought statistics and the changes in temperature discussed earlier. Floods to some extent varied inversely with the temperature. The main reason for this is that the climate of eastern China belongs to the regime of the East Asian monsoon: rain falls when northwardmoving water vapor from the tropical southwest Pacif ic meets cold air moving south from northern Asia. If the temperature falls, the warm air over southeastern China lacks the strength to continue on northward once it has reached the lower Yangzi Valley, and is obliged to remain here, where for a long period it forms an unstable saw tooth-shaped front with the cold northern air, causing prolonged rains and continuing overcast skies. Since the main part of Jiangnan, the Jiangnan Plain, is low-lying and easily inundated, a slight increase in precipitation can cause flooding. Floods have historically been the main natural disaster in the region, and prolonged rains are their chief cause. 22 Rains and overcast weather for an extended period have a serious impact on both the number of hours of sunshine and the intensity of the radiation. In Jiangnan they normally occur between the f ifth and eighth months of the traditional lunar calendar,23 when the hours of sunlight and intensity of radiation reach a potential maximum.24

Changes in the Land

There were both natural and anthropogenic changes in the Jiangnan terrain. Among the former were the silting up of the exits of Lake Tai to the sea, and the reduction of the zone affected by the tide. Among the latter were projects for water control and land management.

Most of those who have studied water control in Jiangnan under the Ming and Qing have not valued its achievements highly. The prevailing view is that even if the off icials of this time did to some degree exert themselves, what they were able to do was constrained by the lack of unif ied objectives and a long-term unif ied strategy. As a result, it is maintained, the entire water-control system fell into disorder, and the capacity to handle disasters declined. 25

There is a certain rationale for this view, but it would seem that it is hard to say that it is soundly based. There were limits to the technical capacity of the period to prevent major natural changes in the terrain, such as whether the exits to the sea of Lake Tai flowed freely or were silted up or changed their courses, or the zone under tidal influence altered. Deeper studies are needed to determine if the reasons for any disorder in the Jiangnan water-control system were ultimately natural or anthropogenic. Whether the major natural changes in the terrain in the last analysis harmed Jiangnan agriculture or benef ited it also needs reexamination. Thus, changes in the zone under the tidal influence (which were linked with changes in Lake Tai’s exits to the sea) may have been detrimental to the cultivation of wet-f ield rice in the coastal districts in northeastern Jiangnan, such as Taicang and the northern parts of Suzhou and Songjiang, but possibly benef icial to the coastal districts along the Qiantang River in the Southeast, that is, the eastern part of Hangzhou prefecture. Overall, these changes may have helped wet-f ield rice cultivation in the region more than they hurt it.26 What is more, the theory that the capacity of the Jiangnan water-control system to handle disasters declined does not f it with the lower frequency of floods in Jiangnan under the Qing as compared with the Ming. Even if this lower frequency was the consequence of the climatic factors recounted above, there is no doubt that the water-control system still played an important function. The historical record also shows that advances in the system also brought about a manifest reduction in the frequency of floods in certain areas that were particularly susceptible to them. A typical example is the county of Qingpu county 青浦县, one of the lowest-lying areas in Jiangnan. After the method of building polders by “circumvallation” (wei 围) and “striking head-on” (qiang 抢) advocated by Sun Jun 孙峻 had been put into effect, the county suffered no more floods for the next thirty years. 27 In summary, we have to a large degree to attribute the credit for improvements in water control in Jiangnan during the period under consideration to the water-control system.

Important improvements in the land took place in Ming-Qing times, although the scholarly community has not in the past assigned much importance to them. Recently, Hamashima Atsutoshi 滨岛敦俊 has expressed the view that while the opening up of lowland f ields in Jiangnan was already basically complete by the early Ming, large-scale improvement began in the middle of the f ifteenth century, reaching what was by and large completion by the middle of the seventeenth century. 28 Kitada Hideto’s investigation into the ecology of Jiangnan farmland has shown that even at the beginning of the nineteenth century, in certain areas, particularly along the eastern seaboard, there was still a fairly large quantity of land awaiting improvement. 29 In Morita Akira’s 森田明view, the key to improving farmland was drainage, and the method most widely relied on was the collective operation of square-pallet chain pumps (longguche 龙骨车 or dapeng-che 大棚车) in large numbers. 30 Although this method of operation had appeared long before, it became widespread only in Qing times. My view is therefore that the process of improving farmland only reached completion by the middle of the nineteenth century.

Prior to the middle of the f ifteenth century, the main objective in the development of land in Jiangnan was to increase the cultivated acreage, or what Hamashima calls “extensive development.” The principal method was the establishment of polders (weitian 圩田 or weitian 围田—there was no great difference in meaning between these two terms by Ming-Qing times), in other words the transformation of low-lying, wet, uncultivated land by surrounding it with a man-made dike. These polders were usually fairly large. Down to the Xuande reign of the Ming (1426-35), it was still the case in Suzhou prefecture that polders might have “as many as 6,000 to 7,000 mu, and, where these were fewer in number, still 3,000 to 4,000 mu.”31 Even in the Wanli reign (1573-1620) there were still places in Jiangnan where polders were above 1,000 mu.32 An example is Xiushui county 秀水县 in Jiaxing prefecture, which in the Wanli reign still had 17 polders over 5,000 mu, 34 of 3,000 to 5,000 mu, 104 of 1,000 to 3,000 mu, 47 of 500 to 1,000 mu, and 30 below 500 mu.33 The larger a polder the worse its shortcomings were. Because of the inequalities in elevation in its interior, the considerably enclosed low and swampy areas could not be effectively utilized, and Hamashima has pointed out that large polders contained a lot of uncultivated land and lakes and ponds.34 Further, while the large polders were effective at preventing the incursion of water from outside, because of the limited capacity of the machines used at that time for drainage, if there were prolonged rains or extended drought, the situation would be that described by Zhao Menglin 赵梦麟, who was the county magistrate of Wujiang 吴江 during the Wanli reign: “If a large polder meets with heavy rains, then the pumps and bailers will not be up to the task, and one will have to sit and watch it become submerged. If there is a drought then the water brought in by the pumps and bailers will not reach [everywhere it should], and this amounts to standing by and waiting for it to become dried out.”35 In the 1810s, Sun Jun, a local hydraulics expert in Qingpu county, also pointed out the shortcomings of large polders:

Some of the land in the field plots in a large polder may differ in height by from one to 1.4 or 1.5 feet (chi 尺), or even two to three feet. Although the dikes can protect the polder against the tides from outside, the water inside will drain downward, and the rice sprouts still be drowned. If one wants to remedy the situation by bailing, [this is not feasible because] there is a flooded expanse across [internal] interlinked dike paths covering sever-al hundred mu. 36

Since large polders thus had many undrainable flooded f ields in the rainy climate of Jiangnan, it was common—from Song to the early Ming—for much of the farmland to be under water.37 Prolonged submersion is not good for the soil of rice f ields, severely reducing its productive capacity. 38

Hamashima has called the improvement of farmland that took place from the mid-f ifteenth century “intensive development.” Its main objectives were to do away with the “internal frontier” by opening up the extensive wastelands inside the polders, and to “dry out the f ields,” by transforming arable areas that were low-lying and wet, and increasing the degree of maturation of their soils.39 The method most commonly used was the increased “subdivision of polders” (fen wei 分圩), or in other words splitting a large polder into a multiplicity of smaller ones. The criterion def ining a “small” polder decreased in consequence: from 500 mu in the Xuande reign to 100 mu in Wanli.40 These smaller polders were appropriate for the draining and irrigating capacities of the machinery of those times, making it possible both to open up the wasteland in their interiors and to dry them out by pumping out water that had ponded in their f ields. The control of the larger polders was also rationalized, by such means as subdividing them into smaller areas for management, and restructuring their water systems both internally and externally, with substantial progress resulting between late Ming and mid-Qing.41 This also improved the soil. By the middle of the nineteenth century, it had become possible to grow winter dry-f ield crops in what had previously been low-lying and wetland occupying a large proportion of the arable area of Jiangnan.42 It needs stressing that when this alternation of wet and dry conditions became feasible, it had a very positive effect on the improvement of the f ields.

Thus, even if hydrological changes and various kinds of problems with hydraulic installations had some damaging effects, overall the condition of the land in Jiangnan during the Ming-Qing period was continually improving.

Changes in the Patterns of Human Effort

Jiangnan farmers during the Ming-Qing mainly used the following methods to raise productivity per mu in wet-rice agriculture: (1) improvement of hydraulic installations and farmland; (2) f ine-tuning spatial cropping patterns, as by growing commercial crops like cotton and mulberry trees in areas unsuited to wet-f ield rice; (3) the heavy application of fertilizers; (4) improvement of the techniques of cultivation; and (5) extending the use of the best cropping systems. Since the f irst of these has been discussed above, and I have dealt with the second and third elsewhere,43 the following analysis will be conf ined to developments in the fourth and f ifth of these approaches.

Techniques of Cultivation

Tang Qiyu 唐启宇 has taken the position that from the mid-sixteenth to the mid-nineteenth century no new techniques, procedures, or types of seeds were introduced into wet-f ield rice farming (and more particularly not in Jiangnan).44 That no new techniques were discovered does not mean, however, that technology was stagnant. Mark Elvin has pointed out that in studying technological progress it is necessary to distinguish (1) invention (or the initial discovery of a technique), (2) innovation (or the f irst incorporation of this newly discovered technique into the structure of actual production), and (3) dissemination (the generalization of the use of the technique).45 The use of this approach is important for the accurate evaluation of changes in technology in Qing-Dynasty Jiangnan. In this context, it should not be overlooked that certain important techniques invented and/or innovated in Ming times were widely disseminated in Qing times. Nor should we, however, ignore the invention and innovation in Qing-Dynasty Jiangnan of some techniques in wet-f ield rice agriculture-and their diffusion.

There were two main subspecies of rice in Jiangnan, long grained orindica (xiandao 秈稻) and round-grained or japonica (jingdao/gengdao 粳稻). Indica has comparatively high requirements for warmth and does not endure low temperatures, whereas japonica has lower requirements for warmth and is relatively resistant to cold.46 The indica and japonica subspecies may each of them be further subdivided into “early” and “late” varieties, on the basis of their seasonal pattern of growth. The main environmental factor distinguishing them is the effect of the length of the day, which can be the result either of latitude or of the time of year-the latter being more signif icant in Jiangnan. The “early”rices are either sluggishly responsive to the length of the day or not responsive at all: given suitable warmth, their growing period is much the same whether the days be short or long. The late rices are sensitive to day length, and are typical short-day crops. To put it generally, indica and japonica are subspecies primarily developed to suit different regimes of temperature, and their “early” and “late”varieties are secondary developments to suit different patterns of day length. Furthermore, there is a transitional variety midway between the “early” and the “late,” namely “intermediate rice” (zhongdao 中稻). 47 The slowly maturing strains of intermediate rice have a relatively strong photosensitivity and in this are close to late rices; but the early-maturing strains of intermediate rice have weak photosensitivity or none, and are close to the early rices.48 One key factor in raising wet-f ield rice production has been matching the biological differences of these various types and varieties with their varying needs as regards temperature and the length of the day. Over a long period of natural and artif icial selection that has lasted at least since the Song Dynasty, the main strains of early rice in Jiangnan have come to be indicas, and those of intermediate and late rice to be japonicas.49 The understanding of both farmers and agronomists in Jiangnan regarding the particular characteristics of early, intermediate, and late rice was, however, far from clear and exact over a long period of time.

Suzhou farmers already knew about the difference between early and late rice by Northern Song times; and by the Southern Song Hangzhou farmers could distinguish between early, intermediate, and late japonica rice (as could those of nearby Mingzhou, the present-day Ningbo). By the late Yuan or early Ming at the latest the same was true for the Suzhou region: in the Suzhou prefectural gazetteer for the Hongwu reign (1368-98), juan 42, which is on local products, contains part of a work called the Wumen shilei 吴门事类 [Matters concerning the Suzhou area, arranged by categories], which for the f irst time in Jiangnan gives the criteria for distinguishing the three strains. The gazetteer for the Jiajing reign (1522-66) for Taicang department, which was under the jurisdiction of Suzhou, also cites this passage, suggesting the same criteria were in use there.50 What is surprising is that after this time and until the end of the Ming, even agricultural theorists in Jiangnan had no very clear understanding of early, intermediate, and late rice. Thus the mid-Ming Bian min tuzuan 便民图纂 [Collected illustrations on matters of convenience for the common people] as well as the late-Ming Shen-shi nongshu 沈氏农书 [Mr. Shen’s agricultural encyclopedia],51 two important agricultural encyclopedias, both style rice maturing and harvested in the ninth month of the old lunar calendar as “early rice,”52 whereas, according to the def inition provided by the Wumen shilei, this ought to be a typical late rice. By the mid-Qing, the conception of early rice had become clear and exact, and knowledge of the intermediate and late strains had also deepened. Not only were the criteria used by the Wumen shilei to differentiate the three strains widely disseminated in areas outside Suzhou,53 but Jiangnan farmers also had precise knowledge of the basic biological dissimilarities deriving from early rice being an indica and the other two being japonicas, as well as the principal other differences deriving from a long or short growing season, and not just from whether the seasons for transplanting and harvesting were early or late in the year. They also had a clearer perception of differences in yields and quality. 54

Tang Qiyu holds the view that the most important achievement in the technology of rice cultivation in Ming-Dynasty Jiangnan was the introduction of the application of supplementary fertilizer (zhuifei 追肥).55 The term “supplementary fertilizer” principally indicates fertilizer applied during the growth of the inflorescence. The fertilizer applied before transplanting, or “basic fertilizer” (jifei 基肥), was called “banking up the foundation” (diandi 垫底), hence the later top-dressing was seen as a supplement. In Mr. Shen’s agricultural encyclopedia, it says in the section “Yun tiandi fa” [How to manage the f ields] that once the period Autumn Begins has arrived (August 7-22), the effect of the basic fertilizer will have been exhausted, hence supplementary fertilizer is applied to provide continuity. For this reason, another name for supplementary fertilizer in Ming-Qing Jiangnan was the f igurative one of “carrying on the power” (jieli 接力).

The technique of adding supplementary fertilizer appeared early in China,56 and as early as the Song Dynasty the peasants of Yuqian county 於潜县 in the mountainous region of western Jiangnan were using it on their seedling plots.57 But even by the late Ming this technique had not been generally mastered for the period of production in the main f ields. The Zhi fu qi shu [Wonderful book for making you rich],58 which dates from the mid-Ming, although it mentions the application of supplementary fertilizer on the main rice f ields, does not indicate the quantity, though it is very precise about this as regards the “basic fertilizer.”Both Li Le’s 李乐 Wuqing zhi 乌青志 [Gazetteer of Wu-Qing town]59 from the Wanli reign and Mr. Shen’s agricultural encyclopedia from the Chongzhen reign (1628-1644), which develops its discussion of this point from the preceding treatise,60 talk about the heavy application of supplementary fertilizer and the quantities. The latter, however, emphasizes the great diff iculty of using the method:

The general explanation is that, while all the [other] tasks of the farming life are easy, in the single case of the application of supplementaryfertilizers for “carrying on the power” you must scrutinize the season and the weather, and examine the appearance [of the crops]. This has the most serious implications for the farmer. Families that lack resources will suffer from meager harvests due to insuff icient fertilizer; and families with an abundance of fertilizer will always be menaced by grain that only consists of husks because of an excess of [this] fertilizer. 61

Thus the application of supplementary fertilizer on main rice f ields was not generally mastered by Jiangnan peasants even by the late Ming, as may also be seen from the absence of any reference to it in Xu Guangqi’s 徐光启 Nongzheng quanshu 农政全书 [Complete treatise of farm administration] 62and the rarity of references in late Ming local gazetteers from the region.

The Qing period was very different. The application of supplementary fertilizer on main f ields in wet-f ield rice cultivation is repeatedly referred to in the agricultural encyclopedias, and is far from infrequent in the local gazetteers. Not only was the practice widespread by the early mid-Qing, but quantities had become by and large uniform: on the order of 40 jin per mu of beancake or its equivalent,63 indicating technological standardization. Tang Qiyu has pointed out that “mastery of the circumstances and timing for the application of supplementary fertilizer on wet-land f ields is exceedingly diff icult.”64 The diffusion of this technique represents major technical progress in the Qing period.

It should be stressed at this point, however, that supplementary fertilizer has relatively little effect in increasing the yield of early rice but is highly important for raising that of intermediate and late rice.65 Thus supplementary fertilizer was closely linked to the more widespread use of intermediate and late rice.

Cropping Systems

A cropping system is a technological complex with many facets. Here we shall only touch on multiple cropping rotations.

The alternation of rice and wheat that lies at the heart of the system of two crops per year had already appeared in Jiangnan in Tang times,66 after which it gradually spread to become the dominant system in evermore localities. It was only around the Ming-Qing transition, however, that (in Kitada’s words) it acquired a “commanding” position in the Jiangnan plain.67 It is Kitada’s view that the Ming-Qing system of two crops per year was different from its Song predecessor in two respects: (1) the earlier system was mostly practiced on the “high-lying f ields” of western Jiangnan, whereas the late one was practiced in the “low-lying f ields” of the plain; and (2) the earlier system took early rice as its principal crop, and wheat as its secondary one, whereas the new system took late rice as its main crop and wheat, rape, and beans as its subsidiary crops.68 What needs to be emphasized at this juncture is that, even if the two-crops-peryear system had already taken a dominant position in the Jiangnan plain by the Ming-Qing transition, the process of its generalization was still far from f inished. It may be seen from Kitada’s work that down to the turn of the eighteenth and nineteenth centuries the one-crop-per-year system was still being practiced in a considerable number of wet-f ields. It would seem that only by the middle of the nineteenth century was the diffusion process complete. 69

Anthropogenic changes in wet-f ield rice cultivation during the Ming-Qing period may thus be epitomized as the spread of a new system of two crops per year, closely linked with the diffusion of the cultivation of late-ripening intermediate rice and early-ripening late rice, and the extension of the use of supplementary fertilizer.

Last of all, we briefly consider the interrelationships between the changes in the climate, the land, and human activities. They had two manifest particularities: climatic changes played the leading role in such interrelationships; and changes in climate and human activities, especially the former, had a major influence on changes in the land.

Shifts in precipitation and temperature have a decisive effect on the aridity or moistness of the soils. Speaking with respect to the present-day situation in the Jiangnan plain, the aridity index, or in other words the ratio of evapotranspiration to precipitation over a given period of time, lies between 1.0 and 1.3 from April through June, and thus falls in the humid to semi-humid range, while from July through September it lies between 1.5 and 2.0, which indicates semi-aridity.70 Since most of the Jiangnan plain is low-lying, with abundant surface water, once the semi-arid conditions of July through September have passed away and the aridity index is down, there is an excess of water over the year as a whole. An excess of water over a long period can damage the structure of the soil, and lead to an increase in toxicity and stronger redox reactions, which reduces the fertility of the soil. If, on the contrary, the temperature is comparatively elevated and precipitation comparatively slight, the aridity index climbs, and the water stored in the soil diminishes, which improves the soil condition. In a manner of speaking one might say that climate was also to be thanked for achievements in the efforts to dry out the rice f ields in Jiangnan during the Qing Dynasty, and especially the eighteenth century.

Further, different meteorological environments can have different effects on activating the latent fertility of soils. This latter is dependent on weather conditions. When the temperature is high, microbial activity increases, the organic content of the soil is more swiftly decomposed, the level of activation of the latent fertility is high, and the supply of nutrients that can be effectively absorbed by crop growth-hence the eff iciency of fertilizer is also high. If the contrary holds, then the level of absorbable nutrients is low, and the eff iciency of fertilizer is reduced. If water stands for a long time in a paddy-f ield, then even if the latent fertility in the soil is high, the very low proportion of the total nutrients that are in an effectively usable state will result in a very low capacity to supply fertility. The same line of argument holds good as regards organic fertilizers that are applied to farmlands. Hence, given a certain level of latent fertility in soil, comparatively more organic fertilizer has to be applied to a rice f ield under conditions of low temperature and high water content in order to produce the same yield as in one with high temperature and low water. 71

Human efforts also played a considerable part in effecting changes in the land at a more modest level. This was done by improving water systems and soil quality, increasing the input of fertilizers, and rotating alternate wet-f ield and dry-f ield crops in the paddies. According to the Ming-Dynasty Jiangnan agronomist Ma Yilong 马一龙, “there are two [types of] methods of increasing soil fertility: those that take effect through human efforts include irrigation, hoeing out weeds, and making f ields and reed-beds;72 and those that take effect through the powers of material things include mud, ordure, lime, oil-cake (shen 糁), straw, and vegetable matter.”73 Han Mengzhou 韩梦周, who lived in the Qianlong reign (1736-95), wrote that “in Jiangsu and Zhejiang the land is rich, cultivating is done properly, and beancake and ordure are applied to increase the fertility ... The rice that they bring to fruition is all bursting with plumpness, and every mu has a harvest of from 800 to 900 jin.” 74 From these we may derive an impression of the part played by the farmers of Jiangnan during Ming-Qing times in improving the fecundity of their lands.

The Impact of These Changes in Weather, Land, and Human Actions on Wet-Field Rice Production in Jiangnan During the Ming-Qing Period

Jiangnan agronomists were conscious of the importance of the effects of variations in the climate, soil conditions, and human inputs on the growth of rice. Thus Ma Yilong wrote:

While it pertains to human beings nourishing [rice], bringing it to maturity pertains to Heaven. The inflorescence must be exposed to the sunlight before it will open. If rain continues for a long time it will close its apertures and not flower. If the winds are rough, the inflorescence will be damaged and not fruit. Both are threats to the as yet immature grain. When the ripened grain is about to be harvested, too dry a soil will cause the rice-kernels to be damaged by dryness; excessive immersion due to an abundance of water will produce black rot. These latter two are also afflictions that destroy the harvest. How can it be within the power of human beings to cause overcast or sunny weather, or arid or humid conditions?

He also said of paddy-f ields: “If they are constantly under cropping, their energy-vitality (qi 气) will invariably weaken. (By ‘weakening’ I mean that the power of the soil weakens ... The rice sprouts draw their resources from the soil in order to grow, and when the power of the soil is lacking, it becomes weak. [Original note.]).75 Bao Shichen said: “No matter whether there is flood or drought, when one is managing farmland an additional dressing of manure will increase the grain by two dou2 per mu, and the addition of one [man’s] working day (yi gong 一工) will likewise augment it by two dou.”76 There is an excessive simplicity in the analysis by these two authors, such as Bao’s view of the effect of incremental human labor inputs on productivity, and a certain lack of clarity with respect to certain important concepts, an example being the failure to distinguish between climate and weather. There may also perhaps be an excessive narrowness, as in their restricting the concept of “land” to one of fertility. While we may take exception to this, however, they did realize that the weather, the land, and human labor were all variable inputs, changes which had a major impact on wet-f ield rice production.

The nature of these impacts is a complex question, and only two aspects will be discussed in what follows: (1) The effect of changes in climate and land, with the technique of cultivation held constant; and (2) The effect of changes in cultivation technology, with climate and land held constant. On this foundation we shall then be able to examine what the consequences of these effects were on the productivity of rice fields per mu in Jiangnan during the Ming-Qing period.

The Impact of Changes in Weather and Land on Wet-Field Rice Production

Climate has decisive effects on the rearing of rice seedlings, transplantation, application of fertilizer, and irrigation. We give a résumé of these below, on the basis of the work of Gao Liangzhi and Li Lin. 77

CLIMATE AND SEEDLING GROWTH. The environmental conditions required during the period of seedling growth relate to temperature, water content of the soil, oxygen, sunlight, and nutrients.

The lowest temperature at which Oryza sativa japonica, the main subspecies of rice used in Ming-Qing Jiangnan, will sprout is about 1℃, and the lowest temperature at which seedlings will grow is 12℃, with the range most suitable to sturdy shoots being between 20℃ and 25℃. Speeds of germination, emergence, and growth are all positively correlated with temperature.

The speed of uptake of water, a suff icient quantity of which is essential for germination, rises with temperature from a point below 10℃, but this effect vanishes above 15℃.

Sprouting and growth demand adequate oxygen, which is necessary for the materials stored in the endosperm to be metabolized, the leaves and roots to grow well, and the sprout to be healthily nourished. Insuff icient oxygen causes weak seedlings that may rot if the weather is unfavorable. For this reason the soil of the seedling plot needs good ventilation for the maintenance of a suff icient supply of oxygen. This implies good irrigation and drainage.

Before the seedling reaches the three-leafed stage, organic nutrients are mainly supplied by the endosperm, but afterward it relies on photosynthesis by the leaves to make them. Hence adequate sunlight is indispensable. At the same time, after leaving the embryo stage it relies on its roots to absorb inorganic nutrients; hence the presence of effective nutrients in the soil is a key factor.

CLIMATE AND TRANSPLANTATION. Given a high temperature and enough sunshine, a single stalk of rice will rapidly sprout numerous tillers,78 which will result in numerous ears of grain. The high temperature also facilitates the activation of the fertility latent in the soil. If, under such conditions, the spatial density of the transplanted seedlings is reduced, a dense growth of stalks and ears may be obtained while economizing on seeds, seedling plots, fertilizers, and labor. 79

CLIMATE AND FERTILIZATION. The quantity of fertilizer that should be applied is commonly approximated by the formula

(F_a-S_f) / u_f

Where F_a is the total quantity of a given fertilizing nutrient needed for the anticipated yield, S_f is the quantity of fertilizing nutrient supplied by the soil, and u_f is the rate of utilization of the fertilizing nutrient supplied in both forms.

The relationship between the capacity of the soil to provide fertilizing nutrients and the weather has already been discussed. Here we will only touch on two further points:

(1) During the phase of vegetative growth, the capacity of the root system to absorb nutrients is greatly influenced in a positive sense by temperature and humidity. During the middle and f inal phases of growth the root system tends to decay, its capacity to produce oxygen little by little declines, and it often depends upon oxygen obtained from the aeration of the soil to maintain its regular vital functions. If there are heavy rains at this time, and the paddy-f ields are full of standing water for a long period, this can impede the root system’s absorptive function and ingestion of nutrients.

(2) Solar radiation is the most evident of the many factors affecting the rate of utilization of fertilizer. If radiation diminishes, so does photosynthesis in the rice plant. In cases where the level of the nitrogen compounds so necessary for a good yield is low, the production of constituent elements is not suff icient, the rate of seed formation is low, and the yield is small. Even if a large quantity of nitrogenous fertilizer is applied, it will be of no help, but merely a self-created waste of fertilizer.

CLIMATE AND IRRIGATION. Adequate water content in the soil is necessary throughout the process of the growth of wet-f ield rice, but irrigation is not required at every phase. On the contrary, in South China it is common, when the rice plants sprout tillers, to dry out the f ields (kaotian 烤田), so that for a time the soil is short of water. This has two effects: (1) It reduces the absorption of water and nutrients by the main stalk, and also photosynthesis, so that the soil only meets the needs of the main stalk and of the tillers that branched off early, so preventing an ineffective proliferation of further tillers. (2) It causes a large quantity of oxygen to enter the soil surface, so that there are fewer products of reduction processes, the activities of aerobic bacteria and the mineralization of organic matter are increased, and the number of black stalks is reduced, while the development of the root system is promoted, so that after the water has returned it will have a strong capacity for absorbing fertilizer and water, with substantially enhanced photosynthesis. In this way the plant’s resistance to lodging80 can be increased, and the material basis for the formation of large ears augmented. The period when single-crop rice in a wheat-stubble f ield (maicha麦茬) sprout tiller is often at the late-spring-early-summer season of the “plum rains” (or “rotting rains” ) when the weather is mostly overcast and rainy with little sunshine, which has an adverse effect on the drying out of the f ields and hence serious implications for rice production.

OTHER EFFECTS OF CLIMATE. Changes in climate had many other effects on rice production. For example, it was mentioned above that each further 100 meters above sea level reduce the mean annual temperature by about 0.45℃, while rainfall increases and hours of sunlight and intensity of solar radiation diminish. The investigation has shown that for each further 100 meters in altitude, the growing season for wet-f ield rice shortens by four to seven days, and the accumulated temperature at above 10℃ diminishes from 100℃ to 150℃, with consequences for the spatial distribution of the crop.81 On this basis, since the mean annual temperature in the seventeenth century was about 1℃ less than in the eighteenth century, the growing season for wet-f ield rice would have been one to two weeks shorter.82 The same effect would also have made itself felt to some degree on the secondary crops. Falling temperature means an increased frequency of damage from cold, to which wet-f ield rice is highly vulnerable in Jiangnan. Low temperatures at seedling time may cause the stalk to grow poorly, clearly with a reduction of the number of ears in consequence; and it may delay the heading stage, with the renewed risk of meeting with cold. In the growing stage there is a direct effect on yield. For example, in the summer of 1976, the temperature was abnormally chilly, and the wet-f ield rice yields in Jiangsu were 46.5 kilograms less per mu than they had been the year before. In 1980, when it was again cold, overcast, and rainy, the decline in the per-mu yields of single-crop late rice and second-crop late rice were 17 and 101.5 kilograms respectively. Although excessive heat can also have a bad effect on the growth of wet-f ield rice, this cannot easily happen in Jiangnan because of its proximity to the ocean, a body of water that exercises a moderating effect and keeps the air humidity high, so that temperatures rise slowly and extremes of heat are out of the usual run of things. 83

To sum up, when climatic factors reduce the water content in the soil, then environmental conditions are favorable for the growth of wet-f ield rice. A hypothetical analysis by Zhang Jiacheng 张家诚, based on macro theory, argues that a change of 1℃ in mean annual temperature would alter the accumulated annual temperature in South China by up to 365℃. This would amount to a difference in accumulated annual temperature of from 100℃ to 200℃ for each major crop, depending on its period of maturation. Hence for every change of 1℃ in temperature, the corresponding change in the “category of ripening”(in other words, from “early-ripening” to “middle-ripening” or vice versa, and so on) will by and large be two grades for each of the crops. According to the economics of agricultural production, a difference of one “category of ripening”implies a change of about 10 percent in output, or 20 percent for two degrees, as here.84 More research will be needed to establish whether this analytical f inding applied to Ming-Qing Jiangnan, but there are a few records that show that the categories of ripening of wet-f ield rice varied with changes in climate. 85Thus relatively good yields can be obtained under conditions of unchanging practice in cultivation, and with comparatively small inputs of labor, seeds, and fertilizer, provided that climate and soils are favorable.

The Impact of Changes in the Pattern of Human Efforts

In this section we explore the effects of changes in the style of farming, with a focus on varieties of seeds, the application of supplementary fertilizer, and the cultivation system.

As has been recounted, the much-enhanced understanding of the different types of wet-f ield rice in the Ming-Qing period was accompanied by wider use of slow-ripening intermediate rice and early-ripening late rice, these being better suited to the present-day natural conditions of Jiangnan, especially climate. The slowness of the return of warmth in the early spring and the wide range of temperatures encountered during this period created a damaging insecurity in the cultivation of early rice.86 Since early rice was grown early in the year, and the temperature at the time of growth was somewhat low, a comparatively heavier dressing of fertilizer was needed.87 The short growing period of early rice meant that yields were low and the quality of the grain was poor, while there was also a conflict in the timing of its cultivation with that of the previous year’s secondary crop that ripened in the summer. Although late rice gave high yields per mu and was of good quality, the long duration of the growing season required a relatively large input of fertilizer.88 There was also a clash with the timing of the cultivation of the secondary crop that would ripen in the summer of the following year. Hence it was that slow-ripening intermediate rice and early-ripening late rice were varieties better suited to Jiangnan conditions.

In Southern Song times, according to the Siming zhi 四明志 [Ningbo gazetteer] for the Baoqing reign (1225-27),89 in Mingzhou, which was to become modern Ningbo, and was next door to Jiangnan, most of the rice grown was intermediate rice that ripened in the solar period Limit of Heat (roughly equivalent to the middle ten-day period of the seventh month of the lunar calendar, approximately August 23 to September 7), with early rice that ripened in the solar period Autumn Begins (roughly equivalent to the f irst ten-day period of the seventh lunar month) in second place. Of the late rice that ripened in the eighth lunar month there was very little. The situation was much the same in Yuezhou, which was to become the present-day Shaoxing prefecture.90 There is something of a difference of opinion between Kawakatsu and You Xiuling, the former thinking that late rice was dominant in northern Zhejiang (southern Jiangnan) and the latter that the Lake Tai area (or Jiangnan plain) resembled southern Zhejiang (in which Ningbo and Shaoxing are located).91 According to the criteria in the Wumen shilei (cited above), “intermediate rice” in Song times belonged to the category of “early rice,” while Song “late rice” belonged to that of “intermediate rice.” On this account it would appear that during the Song Dynasty, early rice or early-ripening intermediate rice predominated, and that early rice continued to hold a considerable place in Jiangnan until the middle Ming. According to the list of 21 important varieties of non-glutinous Jiangnan rice given in Huang Shengceng’s sixteenth-century Varieties of rice, there are still 9 in the “early” group, and only 5 in the “intermediate” and 7 in the “late.”92 Although the number of varieties does not necessarily represent the area planted, the small number of intermediate and late varieties must have had an important impact on the degree to which they were generally cultivated. As proof of this we may instance the discussion of the growing period of wet-f ield rice in Suzhou and Songjiang by Yu Ruwei 俞汝为 at the end of the Ming: “By and large, the rice in the wet-f ields all relies on water for its nourishment ... Irrigation water has to be brought in for the 120 days in which the rice is born, grows, ripens, and bears ears.”93 As it happens, by present-day standards this duration of the growing season makes it early rice,94 which thus was still widespread in at least these two important prefectures down to the end of the dynasty. On the other hand, it appears from Mr. Shen’s agricultural encyclopedia and the Supplement to the Agricultural Encyclopedia that in the Huzhou/Jiaxing region the main rice was a late-ripening late variety that was harvested in the tenth lunar month.95 By the mid-Qing early rice had to all intents and purposes gone from the Jiangnan region, and late-ripening late rice seems mostly to have been grown where cotton and mulberries were comparatively common.96 If one has regard to the times of transplanting and harvesting, slow-ripening intermediate rice and early-ripening late rice were the main varieties in most places.97 The generalization of these two varieties was one of the basic conditions for the double-cropping system with over-winter secondary crops (chunhua 春花).

I have argued in the past that the quantity of fertilizer per mu of rice paddy increased in Ming-Qing Jiangnan, and that a large proportion of this was supplementary fertilizer.98 There are two peaks in the absorption of nitrogen by single-crop rice, at go and 60 days after transplanting;99 hence the great signif icance of supplementary fertilizer for yields (especially from intermediate and late rice). It is not, however, easy to determine its specif ic contribution in isolation from other factors, but as regards nitrogen content, 3 kilograms of soybean cake are broadly equivalent to 1 kilogram of ammonium sulfate.100 Further, according to the input-output analysis by Dwight Perkins, 1 kilogram of ammonium sulfate will increase grain yield by 6 kilograms in China.101 In other words, 1 kilogram of soybean cake produces an increase of 2 kilograms of grain. In the Qing period, about 40 jin of cake were applied to each mu as supplementary fertilizer, which could thus cause an increase of 80 jin per mu (on the order of 600 kilograms per hectare). These calculations indicate the importance of the effect but take no account of specif ic conditions of season or locality, and will need therefore to be followed up by determinations based on f ield investigations.

The two-crops-a-year rotation of wet-f ield rice with wheat, rape, or beans involved alternating wet-f ield with dry-f ield cultivation, which was benef icial to the soil. Zhang Lüxiang observed that growing wheat in a paddy-f ield would make the soil dry out and increase its porosity, which helped the next crop of rice.102 Fang Zaihui in modern times has pointed out that the advantage of rotating wet-f ield rice with barley or wheat, and in particular with rape or broad beans, is a rational exploitation of the nutritional elements in the soil: a balance is created between the supply and demand of nutrients, which makes the use of fertilizer more effective and hence reduces the quantity needed. It can also influence the humic content of the soil, its total nitrogen content, the unit weight of the arable layer, and the porosity. The result is more fertile soil with improved plowing characteristics, a possible reduction of toxins, and a lessening of the damage done by diseases, insects, and weeds.103 Alternative systems of cultivation, in contrast, have the defect of leaving the farmland in too humid a condition, which leads to physiochemical degradation. This is evidently the case for one-crop-a-year rice, apart from the period in the winter when the f ield is drained and the soil turned over and dried, and for double-cropped rice. Even the rotation of wet-f ield rice with green manures such as lucerne (alfalfa) and Astragalus sinicus (milk vetch) on a two-crops-a-year basis does not, however, permit turning and drying the soil in the winter, with the result that neither crop grows well and yields are low.104

Since even more wet-f ield rice land was transferred to the cultivation of cotton and mulberry trees in the Ming-Qing period, most of it being of a kind relatively unsuited for rice growing,105 the relative quality of the remaining rice land, and its per mu yield, ought on the basis of this alone to have risen. We now look to see if this was so in fact.

The Increased Yields of Wet-Field Rice per Unit of Area in Ming-Qing Jiangnan

What was the change in yields per mu? Some scholars think that it was constant, some that it fell, and some again that it went up. In the 1980s this divergence was very evident, with Yu Yefei 余也非 taking the traditional position that output per mu had remained unchanged through the Yuan, the Ming, and the Qing.106 Wu Hui 吴慧 likewise asserts that under the Qing from 1644 to 1840 yields per mu of wet-f ield rice in Jiangnan were not greatly different from what they had been in Ming times.107 Min Zongdian 闵宗殿, however, categorically maintains that there was a falling trend in yields per mu in the Lake Tai region in the Ming-Qing period, and even regards the Qing level as being no more than 83 percent of that in Ming times. In contrast to the foregoing three scholars, Wu Chengming 吴承明 is of the opinion that yields went up in the early and mid-Qing.108 Li Longqian 李龙潜 considers that in such places as Suzhou and Jiaxing in the early and mid-Qing Dynasty, wet-rice yields per mu were evidently higher than they had been at the end of the Ming. 109

Before we ask who is wrong and who is right as regards this question, several preliminary points need making. (1) There are not many works that focus specif ically on Jiangnan, and even fewer that concentrate on the different periods during the Ming and the Qing. (2) The majority of the works that discuss permu yields of wet-f ield rice in Jiangnan make their case by means of examples, relying on items taken from collections of documents and relating to this matter. (3) Some scholars regularly use per-mu yields from Song times and from modern times (particularly the 1930s and the 1950s) to serve as the basis for evaluating their estimates of Ming and Qing yields in Jiangnan. It is our view that while the above methods are in some respects reasonable, they are also in other respects questionable. First, they do not differentiate the Ming and Qing into distinctive periods for special study, thus having no way of discerning whether or not there were in fact changes. Second, there are large differences between the permu yields recorded in the historical material on Ming-Qing Jiangnan already in our possession, ranging from 1 shi 石 or less to 4 shi or more, and many of these f igures relate to such special circumstances as bumper crops or extremely good-quality land, which makes any judgment about the changes in the average level diff icult.110 Obtainable records of yields are also limited in number and insuff icient to support a quantitative analysis of the instances contained in them. Data from the Song and from the present time are valuable reference material when determining Ming-Qing yields, but require us to impose the following conditions: (1) that they be accurate; (2) that attention be paid to the effects of differences in climate and soil conditions in Song and modern times; and (3) that no assumption be made of “linear progress.” In other words, these materials serve the function of reference, not of proof. In what follows, a macroscopic analysis of the yield problem is based on the relationship between consumption and production, after which some local records are used to test the conclusions.

As a necessary preliminary, we begin with changes in cultivated acreage under wet-f ield rice, population levels, and rice consumption.

Although there are sizable discrepancies between the off icial statistics for cultivated land in Ming and Qing times and the reality, among the various important economic regions in China, the f igures for Jiangnan’s farmland are comparatively close to the truth. Among the statistics for the period prior to 1950, it is generally recognized that those for the 1580s are relatively reliable. Here we conflate the data from the f ive prefectures of Suzhou, Songjiang, Changshu, Zhenjiang, and Jiangning from juan 17 of the Wanli Da Ming huidian万历大明会典 [Collected statutes of the Great Ming] and the Wanli land f igures given in juan15 of the Kangxi-reign Zhejiang tongzhi 浙江通志 [Comprehensive gazetteer for Zhejiang province] for the three prefectures of Hangzhou, Jiaxing, and Huzhou to arrive at an approximate total of 45 million mu. This f igure is higher than the total obtained by adding together the later totals for individual prefectures and counties, which entitles us to say that concealment must have been low and this f igure close to the actual situation. 111

Because of the expansion of the cultivation of non-cereal crops for the market, the area under wet-f ield rice in Ming-Qing Jiangnan tended continuously to contract. Here, for the sake of simplicity, I shall approximate this for both periods as 90 percent.112 This is methodologically conservative in that it tends to cause an overestimation of productivity in Ming times, in the sense that it works against the thesis that I would like to establish here, namely that productivity was rising; thus, once this rise is even so shown to have occurred, it makes the conclusion all the more certainly correct.

Thus the Ming-Qing wet-rice cropping area in Jiangnan may be estimated as 4.05×107 mu.

The population of Jiangnan in Ming and Qing times is a thorny question. On the basis of my work on demographic change in Jiangnan, it is possible to make a rough investigation.113 By following the method adopted by Wang Yejian王业键,114 but using the population f igures for 1850 go established by Liang Fangzhong 梁方仲,115 and the f igures in the Board of Revenue’s acreage and tax registers for 1850, it can be determined that the total population in Jiangnan for this latter year was about 36 million. Although G. W. Skinner has pointed out that the off icial population f igures for many places in China in 1850 are too high,116 it is not possible to obtain better ones, and so they are used here.

There are no usable population f igures for the late Ming, so we are obliged to calculate the average annual growth rate from the reliable data for 1393 in the early Ming to 1850. For 1393, the Wanli Collected statutes of the Great Ming, Vol.19, and the Kangxi Comprehensive gazetteer for Zhejiang, Vol.15, give us a total population in Jiangnan at this time of 8.7 million people. This implies an average annual growth rate of 0.31 percent. The population in 1620, before the onset of the late Ming socioeconomic crisis, would therefore have been 17.6 million. Since the growth rate of the population in Jiangnan in Ming times may have been slightly higher than it was in Qing times, the true total may have been nearer to 20 million. 117

Consumption of wet-f ield rice mainly comprised three categories, namely food, seeds, and the raw material for the production of alcohol. Although the payment of tax on grain is not strictly speaking “consumption,” attention must also be paid to this aspect. To simplify the analysis, we shall set aside the use of grain for seeds, tax, and alcohol, although the amounts were of some signif icance, and the proportion used for alcohol in Qing times was far above that in Ming times. 118

A number of authors have discussed rice consumption in Jiangnan, of whom the deepest and most reliable is Shiba Yoshinobu.119 In his view the annual consumption per person in Ming-Qing Jiangnan was about 3.6 market shi (roughly equivalent to Ming or Qing shi). Bearing in mind the muchincreased production and interprovincial import of wheat in Qing times, and more numerous records of wheat consumption in this latter dynasty, here we provisionally take Bao Shichen’s estimate of 3 shi of rice per person per year, which leaves 0.6 shi over for wheat. For the Ming Dynasty, we take the average rice consumption per person per year as 3.3 shi, which leaves 0.3 shi for the wheat component.

On the basis of the foregoing f igures it can be calculated that the total consumption of rice in Jiangnan was 6.6×107 shi in 1620 (that is, 3.3×2×107), and 10.8×107 shi in 1850 (that is, 3.0×3.6×107). The interprovincial imports must, however, be taken into account. I have estimated that annual interprovincial imports of rice into Jiangnan in the middle of the nineteenth century were about 1.5×107 shi, on the basis of studies by Quan Hansheng 全汉升 and Wang Yejian, Wu Chengming, and Wang Yejian and Huang Guoshu 黄国枢.120 Thus 9.3×107 shi were grown in Jiangnan at this latter date.

These f igures show that in the early seventeenth century the productivity of wet-f ield rice land was on the order of 1.63 shi per mu (i.e., 6.6/4.05), and in the mid-nineteenth century of 2.3 shi per mu (i.e., 9.3/4.05). This was an increase of 0.67 shi per mu or of 41 percent (i.e., 0.67/1.63). Given the bias toward selecting larger estimates for the Ming where we had a choice, this represents a minimum increase.

We now look at the documentation from the two prefectures for which it is densest, namely Suzhou and Songjiang, to see if the above conclusion about the trend is supported.

Perkins has collected 17 f igures on per-mu productivity of wet-f ield rice in Suzhou and Songjiang covering the sixteenth to the eighteenth century: 11 of these relate to the sixteenth and seventeenth centuries and range from 2.0 to 6.0 shi of unhusked grain, which may be converted to half these values for husked rice. The mean is 3.0, or 1.5 shi of husked rice. Six items are from the eighteenth century: they range from 1.0 to 7.2, with the mean being 3.8 shi (i.e.,1.9 husked).121 The increase is 0.8 shi per mu or 26.6 percent. Wu Chengming has collected 26 f igures for the Jiangnan region covering the period 1621 to 1850, excluding double-cropped rice, some 8 of which are explicitly for Suzhou or Songjiang. From 1621 to 1644 productivity was from 1.0 to 3.0 shi per mu; from 1644 to 1720 it was from 1.0 to 2.0 shi per mu; and from 1796 to 1850 from 2.0 to 3.0 shi per mu, exclusively for Suzhou.122 There would seem to have been a rise overall. In addition to these data, Min Zongdian has 6 f igures for Suzhou, Songjiang, Jiaxing, and Huzhou from 1488 to 1644: they range from 1.5 shi per mu to 3.0 shi. For the Qing period down to 1850, he has 10 f igures for the same area: these range from 1.5 to 3.6.123 Min f inds the Ming average to be 2.3 shi and that of the Qing to be 2.0. If, however, we remove the f igures for the Guangxu reign (1875-1907) and f igures for Jiangxi that Min has mistakenly assigned to Jiangnan, then the mean for the Qing down to 1850 is 2.7 shi of husked rice per mu, up 0.4 shi from Ming times, some 17 percent. Wu Chengming’s f igures for parts of Jiangnan other than Suzhou and Songjiang also show a continuous rise in productivity.

There may of course have been variations in productivity in the f irst two centuries of Qing rule. Liu Yongcheng 刘永成 has analyzed the rent records for 1747-55 in the Shenyu-tang zubu 慎余堂租簿 [Rent books from the Shenyu hall] by a Shen family 沈氏 of Suzhou, and nine cases in the archives of the Board of Punishments relating to rents for farmland in Wujin, Yixing, Jiangyin, Fengxian, Taicang, Jingui, and Huating during the Kangxi, Qianlong, and Jiaqing reigns (1662-1820). He found that the per-mu output of wet-f ield rice was clearly higher in the f irst two of these reigns than in the third one. Thus in 1722 the f igure for Wujin was 5.2 shi of husked rice, between 1745 and 1755 the average for Suzhou was 4.12 shi, and from 1745 to 1784 six departments or counties averaged 4.16 shi; but between 1798 and 1801 Huating and another county produced only 1.86 shi.124 However, the data on per-mu output collected by Guo Songyi 郭松义 from f ifteen counties in Jiangnan show an average for the Qianlong reign of 2.14 shi, and for the Jiaqing and Daoguang reigns of 2.15 shi (with the Daoguang level at 2.45),125 which does not show any clear decline. Since the weather in the early part of the nineteenth century was somewhat worse than it had been in the eighteenth century, it is possible that some areas in Jiangnan, and especially those where yields had been particularly high, should manifest some reduction. Overall, though, there does not seem to have been a drop.126 Although these sparse data cannot provide a satisfactory answer to the problem, when taken together with the preceding quantitative analysis, they provide grounds for believing that there was in fact a signif icant increase in productivity per mu in Jiangnan wet-rice farming during the Ming-Qing period in Jiangnan.

If, per contra, it is argued that there was no increase, then the two following questions have to be answered: (1) What factors then prevented the changes described in the earlier part of this chapter from causing a rise in per-mu productivity? (2) How was the large consumption def icit that mid-Qing Jiangnan would still have faced, even after allowing for the interprovincial import of grain, resolved? We have not yet seen any credible answers to these questions.

As regards using either Song Dynasty or modern f igures for productivity in Jiangnan in order to deny any increase in Ming-Qing productivity, even harder questions have to be met. Both the natural and the socioeconomic conditions were different, or very different, in all three periods. This makes simple comparisons impermissible; and they explain nothing. Shiba has also pointed out that previous Song estimates are biased toward the high side.127 Chen Hengli and Wang Da, for their part, indicated as long ago as 1958 that the permu productivity of the main farm crops in Jiaxing and Huzhou in the 1930s was clearly less than it had been in the late Ming.128 We are therefore inclined to think that traditional wet-rice farming in Jiangnan (i.e., prior to about 1955) reached its peak in the mid-Qing. It is worth noting that this increase was not accompanied by a clearly marked increase in the input of labor,129 which means that it is hard to explain it in theoretical terms as “involution” (neijuan-hua 内卷化). Wang Yejian has argued for a close correlation between changes in the weather and changes in rice prices,130 which happens to f it with the foregoing conclusions. In the f irst half of the nineteenth century, the population of Jiangnan had already reached its pre-1950 maximum, and was mainly fed, in spite of considerable interprovincial imports of grain, from locally produced rice. Although there were other important factors involved, such as changes in the velocity of the circulation of money, this increase in unit-area productivity in wet-rice cultivation, which nonetheless did not fall into the trap of “involution,” was the main reason why Jiangnan rice prices did fall over a long period. It would, however, also seem that this increased productivity and the resulting economic benef its were in large measure due to changes in the natural environment.

I would like to conclude by emphasizing that the links between farming on the one hand and changes in the natural environment and in human activities on the other are at one and the same time important and complex. The above analysis is only a preliminary inquiry. Though I have expended considerable effort on it, and received the advice and help of a number of colleagues, it is only a beginning. Many important aspects have not been touched on; examples include the changes that occurred in the social system governing agriculture and in the organization of farming, and the impact of various sorts of changes that took place outside the farming sector. I will look both more widely and more deeply in future work.

(Translated by Mark Elvin)

---

1　For valuable criticisms of my original paper presented to the 1993 Conference in Hong Kong on the History of the Environment in China, I would like to offer my thanks to Professor You Xiuling of the Agricultural History Unit of Zhejiang University of Agriculture; Professor Li Genpan of the Institute of Economics of the Chinese Academy of Social Sciences; Professor Guo Songyi of the Institute of History of the Chinese Academy of Social Sciences; Professor Mark Elvin of the Research School of Pacif ic and Asian Studies of the Australian National University; and Professor Liu Ts’ ui-jung of the Institute of Economics of the Academia Sinica.

2　Please note that units for the volume of grain (shi 石) used during the Song, Ming, and Qing Dynasties varied in the ratios 0.6185, 1.000, and 0.9644, respectively, while the unit for the area of farmland varied in the ratios 0.8313, 1.000, and 0.9610. See Gang Deng, Development Versus Stagnation: Technological Continuity and Agricultural Progress in premodern China (Westport, Conn.: Greenwood, 1993), pp.xxv. Thus shi per mu varied in the ratios 0.744, 1.000, and 1.004, which is signif icant for comparisons between late medieval and late traditional times, but not as between Ming and Qing.

3　An alternative rendering is “appropriate human action,” to indicate the aspect of adaptation to natural circumstances. [Trans.]

4　Li Genpan Zhongguo gudai nongye, Tianjin: Tianjin jiaoyu CBS, 1991, ch.3, sec.5, provides a complete analysis: “‘Heaven’ and ‘Earth’ indicate the weather and the soils, the topography, and suchlike elements of the natural environment of farming, while ‘man’ indicates its core productive component. Farming represents the unif ication of natural reproduction and economic reproduction. Insofar as they reproduce naturally, farm products (or, in other words, the harvests) are inseparable from their circumambient natural environment. Insofar as they constitute economic reproduction, they are inseparable from those in charge of farming. Agriculture consists of the mutual interdependencies—the ecological system of mutual control, and the economic system—that link farm products, the natural environment, and human beings. This intrinsic character of farming is already touched on in the propositions of the Shenshi in the Lü-shi chunqiu (third century B.C.) relating to the interconnections between the farm crops and the ‘Three Powers. ’ ”

5　Zhang Lüxiang, Bu nongshu [Supplement to the agricultural encyclopedia (of Mr Shen)], “zonglun,”in Chen Hengli and Wang Da, eds., “Bu nongshu” jiaoshi, rev. ed., Beijing: Nongye CBS; 1983, p.152. An alternative rendering of the last clause (zui shi zhicheng wuwei) would be “The activities of farming and sericulture are the most practical.” [Trans.]

6　Lu Shiyi, Sibianlu jiyao (N. p.: Jiangsu shuju, 1877), j.11, “Lun qutian.”

7　Fu Yiling, Mingdai Jiangnan shimin jingji shitan, Shanghai: Shanghai renmin CBS, 1957, p.1.

8　The population of Jiangsu, Anhui, and Zhejiang may have been reduced to almost half its previous level. See Dwight Perkins, Agricultural Development in China, 1368-1968, Chicago: Aldine, 1969, pp.210-12.

9　A term whose identif ication is debated but which most ancient historians regard as indicating the main waterways once draining Lake Tai into the sea.

10　Again debatable as regards indentif ication, but all lakes on the Jiangnan plain.

11　See Li Bozhong, “Jian lun ‘Jiangnan diqu’ de jieding”, in Zhongguo shehui-jingji-shi yanjiu No.4, 1990.

12　Fang Zaihui et al., Zhejiang-de gengzuo zhidu, Hangzhou: Zhejiang kexue-jishu CBS, 1984, pp.2-3; Guo Wentao et al., zhongguo nongye keji fazhan shilüe, Beijing: Zhongguo kexue-jishu CBS, 1988, p.87.

13　Zhu Kezhen, “Lishi shidai shijie qihou de bodong”, Renmin ribao (Beijing), May 7, 1961.

14　Zhu Kezhen, “Zhongguo jin wuqian nian lai qihou bianqian de chubu yanjiu”, Kexue tongbao 1 (1972).

15　Ren Zhenqiu, “Zhongguo jin wuqian nian lai qihou de yichang qi ji qi tianwen chengyin tantao”, Nongye kaogu (1986); Wang Zichun and Gao Jian’ guo, “Zhongguo jin erqianwubai nian lai zhiwu chonghua lishi jilu zhi wuhouyanjiu”, Nongye kaogu 1 and 2 (1985); and Wang Yejian [Yeh-chien Wang], “Secular trends of rice prices in the Yangzi Delta, 1632-1935,” in T. Rawski and L. Li, eds., Chinese History in Economic Perspective, Berkeley: University of California Press, 1992, citing the work of Zhang Peiyuan [Piyuan] et al.

16　Ren Zhenqiu, “Abnormal periods in China’s climate,” f ig.1.

17　Fang Zaihui et al., Zhejiang farming, pp.48, 50, 53, 61, and table 2-2.

18　Zheng Zhaojing, ed., Taihu shuili jishu shi, Beijing: Nongye CBS, 1987, p.255.

19　Fang Zaihui et al., Zhejiang farming, p.90.

20　Zheng Zhaojing, Lake Tai water control, ch.10.

21　Limited by periods of cold winters (e.g, 1620-1720) partially overlapping with dry spells (e.g., 1625-62).

22　In the Lake Tai region, from A.D.300 to 1900, the frequency of floods was about twice that of droughts. Of the floods, some 70 percent of those for which the causes have been recorded were the result of prolonged rains. See Zheng Zhaojing, Lake Tai water control, p.240. Since there is an abundance of groundwater in Jiangnan and a fully developed irrigation system, droughts caused by an insuff iciency of rainfall have not on the whole been severe, with the exception of those in a few high-lying areas, and in a few years in which rain has been especially short.

23　Zheng Zhaojing, Lake Tai water control, pp.242-43.

24　Fang Zaihui et al., Zhejiang farming, p.61, compare the present-day Hang-Jia-Hu plain with the hill country in the mountainous region of southern Zhejiang, f inding that the plain enjoys more hours of sunlight, “richer illumination,” and a comparatively greater intensity of solar radiation. The main reason for the difference is that there is relatively less rain, and that the dry period in the second half of the year is relatively prolonged.

25　Zheng Zhaojing, Lake Tai water control, pp.94-95; Philip Huang [Huang zongzhih], Stanford: Stanford University Press, 1990, pp.34-35 and pp.38-40.

26　Kitada Hideto, “Chūgoku Kōnan sankakusu ni okeru kanchō chiiki no hensen”, Tōyō gakuhō 63 (1982).

27　Sun Jun, Zhuwei tushuo, 1869 ed., preface by Chen Qiyuan.

28　Hamashima Atsutoshi, “Tudi kaifa yu keshang huodong—Mingdai zhongqi Jiangnan dizhu-zhi touzi huodong”, in Zhongyang yanjiu-yuan di-erjie guoji Hanxue huiyi lunwen-ji, Taipei: Academia Sinica, 1989.

29　Kitada Hideto, Sō-Gen-Min-Shin-ki Chūgoku Kōnan sankakusu no nōgyō no shinka to nōson shukōgyō no hattatsu ni kansuru no kenhyū, Report on the results of research, Tokyo: mimeographed by Sutaa shōkai, 1988.

30　See Morita Akira’s speech in Watabe Tadayo and Sakurai Yūmio, eds., Chūgoku Kōnan no inasaku bunka-sono gakusaiteki kenkyū, Tokyo: Nihon Hōsō shuppan kyōkai, 1984, p.198.

31　Kuang Zhong, Ming Kuangtaishou zhi Su zhengii quanji (1764 ed. ), j.9, “Xiujun tianyu ji jianghu shuili zou.”

32　Hamashima Atsutoshi, “Land development and nonlocal merchants.”

33　Watabe Tadayo and Sakurai Yūmio, Rice-growing culture of Jiangnan, pp.194-95.

34　Hamashima Atsutoshi, “Land development and nonlocal merchants.”

35　Dong Fen, Biyuan ji, “Wujiang Ming fu Zhao-hou yizheng bian xu”, in Liu Chenggan, ed., Wuxing congshu.

36　Sun Jun, How to build polders.

37　Kitada Hideto, Study of agricultural progress, ch.2.

38　Gao Liangzhi and LiLin, Shuidao qixiang shengtai, Beijing: Nongye CBS, 1992, p.391.

39　Hamashima Atsutoshi, “Land development and nonlocal merchants.”

40　Hamashima Atsutoshi, “Guanyu Jiangnan ‘wei’ -de ruogan kaocha”, Lishi yanjiu 7 (1988).

41　Zheng Zhaojing, Lake Tai water control, pp.120-36.

42　Up to about 1850 there was still a small quantity of land in Jiangnan that was either too low-lying or subject to too much rainfall, or both, for the alternation of wet-f ield rice and dry-f ield “spring-flowering crops” (chunhua). The counties in the western part of Songjiang prefecture are a case in point. After the severe floods of 1823 the weather continued to be abnormal, and very few low-lying f ields were planted with “spring-flowering crops.” See Jiang Gao, Pumao nongzi, Shanghai tushuguan, 1963.

43　Li Bozhong, “Ming-Qing shiqi Jiangnan shuidao shengchan jiyue chengdu-de tigao”, Zhongguo nongshi 1 (1984); and Li Bozhong, “Ming-Qing Jiangnan nongye ziyuan de heli liyong”, Nongye kaogu 2 (1985).

44　Tang Qiyu, Zhongguo zuowu zaipei shigao, Beijing: Nongye CBS, 1986, p.39.

45　Communication to the author, May 1, 1994.

46　Gao Liangzhi and Li Lin, Ecology of wet-f ield rice, pp.11 and 201; You Xiuling, “Taihu diqu daozuo qiyuan ji qi chuanbo he fazhan wenti”, in Zhongguo nongye kexue yuan and Nanjing nongye daxue Zhongguo nongye yichan yanjiushi, eds., Taihu diqu nongshi lunwenji, Nanjing: Nanjing daxue, 1985.

47　The quotation marks around the terms “early,” “intermediate,” and “late” are dropped after this point, but the reader is asked to bear in mind that these are technical terms, somewhat variable in sense, and usually having a meaning that is not accurately conveyed by the simple dictionary def initions.

48　Nanjing nongxue-yuan and Jiangsu nongxue-yuan eds., Zuowu zaipei-xue, southern series, shang ce, Shanghai: Shanghai kexue-jishu CBS, 1979, pp.26-28.

49　You Xiuling, Rice cultivation in Lake Tai area.

50　Sutō Yoshiyuki, Sōdai keizai-shi kenkyū, Tokyo: Tōkyō daigaku shuppansha, 1962, pp.146-47; Amano Motonosuke, Chūgoku nōgyō-shi kenkyū, augmented ed. (Tokyo: Ochanomizu shobō, 1979), pp.212-13.

51　Shi shenghan and Kang Chengyi, eds., Bian min tuzuan, Beijing: Nongye CBS, 1982, and the Shenshi nongshu in Chen Hengli and Wand Da, Supplement to the agricultural encyclopedia.

52　Chen Hengli and Wand Da, Supplement to the agricultural encyclopedia, pp.21, 38.

53　Qingpu xianzhi, 1788 ed., j.11, “Chanwu” ; Bao Shichen, Qimin si-shu, in Bao Shichen Anwu Sizhong, Daoguang reign ed., j.1, Nong 1 shang, “Bian gu.”

54　It seems to have been already known in the Southern Song that earliest kind of rice was indicas and most late one was japonicas. See Kawakatsu Mamoru, Min-Shin Kōnan nōgyō keizai-shi kenkyū (Tokyo: Tōkyō daigaku shuppansha, 1992), pp.20-21. In mid-Ming, Huang Shengceng had a relatively clear appreciation of these sorts of differences, observing that “Indica ripens early; hence it is called ‘early rice. ’ Japonica matures late; hence it is termed ‘late rice. ’ ” See his Lisheng yujing daopin, in Wang Wenlu, ed., Bai ling xueshanji: Wanli reign-period. But he was unaware of the most important differences; otherwise he could not have said (as he did), “The smaller japonicas are called ‘indicas.’”By the mid-Qing few agronomists confused subspecies in this way. For example, in Bao Shichen’s Four Arts, we read that “Japonica (jingdao) is today the common source of food grain. The early-ripening type is called indica. Its harvest is f inished by the solar period “Autumn Begins” [August 7-22]. The most slowly-ripening type is called ‘late rice,’ being white, long-grained, and of a beautiful fragrance. It is especially good for people’s health, and the ears mature in the ninth month. The variety that ripens in the eighth month is given the special name of jing [usually read as geng by agriculturists] rice, and it is considered that its red color is of benef it to people’s well-being. Its yield is about the same. Only the early rice is somewhat more meagerly productive. Advantage is taken of its earliness to bridge the gap between harvests, and to avoid the droughts of the fall.” Likewise the Jiang Gao, Report on agriculure in the Huangpu and Mao [lake] area, says: “Early rice: ... Here they call it red ‘rice’(chi mi) and also xian rice [indica]. It is planted in the f ifth month and is ripe by the seventh. Even in years of plenty, however, its yields per mu do not exceed 1.4 to 1.5 shi. The labor and capital expended do not give comparable returns to those on late rice.” Or again: “The mid-autumn rice matures in the eighth month, and the late rice at about the time that the frosts are f irst falling. There is a popular saying, ‘If the green rice sprouts are not up when the dew grows cold [the solar period approximately equivalent to October 8-22], when the frosts fall [the period from October 23 to November 6] they will all lodge [i.e., collapse]. ’ ” The jing/geng rice referred to in the f irst of these works as ripening in the eighth month, and the “autumn rice” referred to in the second, are intermediate rice. It may be seen from these that by the f irst half of the nineteenth century Jiangnan people had full knowledge of the biological characteristics of the three types of rice.

55　Tang Qiyu, Chinese food crops, p.29.

56　Cao Longgong, “Woguo daozuo shifei fazhan shilue”, in Zhongguo nongshi No.1, 1989.

57　Lou Shu, Gengzhi tu, f ig.7, “Yu yin,” as discussed in Amano Motonosuke, Agricultural history, pp.226-28.

58　Zhifu qi shu (Kimura kendō ed. In Nihon Naikaku Bunko).

59　Parts of this now-lost work are cited in the Wucheng xianzhi (1879). See also the following note.

60　You Xiuling, “‘Shen-shi nongshu’ he ‘Wuqing zhi’”, Zhongguo keji shiliao No.4, 1989.

61　Mr. Shen’s agricultural encyclopedia, “Yun tiandi fa.”

62　Shi Shenghan, ed., Shanghai: Shanghai guji CBS, 1985.

63　Li Bozhong, “Increased intensif ication of rice production.”

64　Tang Qiyu, Chinese cultivation of crops, p.29.

65　Gao Liangzhi and Li Lin, Ecology of wet-f ield rice, pp.377-78; Nanjing nongxue yuan et al., Crop cultivation, p.77.

66　Li Bozhong, “Rice-wheat multicropping.”

67　While it must be pointed out that there were a considerable number of f ields in Jiangnan in Tang times that were under a system in which wet conditions alternated with dry and gave two crops a year (Li Bozhong, Tangdai Jiangnan nongye de fazhan, Beijing: Nongye CBS, 1990), pp.118-19, and that this area clearly expanded under the Song, the work of Ashitatsu Keiji and ōsawa Seishō has shown that the greater part of such land at this time was on the hills in the river valleys. Most of Jiangnan is of course a low-lying plain and at this period gave one crop of rice a year, or even one every two years, being under a “cultivation system for excessively wet land.” See Ashitatsu Keiji, “Sōdai Ryōsetsu ni okeru suitō no seisanryoku suijun”, Kumamoto daigaku bungakubu ronsō 17 (1985); Osawa Seisho,“‘So-Ko juku, Tenka soku’ -kyozō to jitsuzō no aida”, Atarishii rekishigaku no tame ni 179 (1985); and Kitada Hideto, Agriculural progress and village handicrafts in the Jiangnan delta. This corresponds with the model of historical agricultural progress in Jiangnan advocated by Shiba Yoshinobu, Sōdai Kōnan keizai-shi no kenkyū (Tokyo: Tōkyō daigaku Tōyō bunka kenkyūjo, 1988), pp.169-74, and so we may take it as correct.

68　Kitada Hideto, Agricultural progress and village handicrafts in the Jiangnan delta, chs.2 and 3. It is my opinion, however, that the “early” rice of which Kitada speaks here should be “intermediate” rice. This is because the main-f ield growing periods of early rice and of the secondary crop overlap, which makes a yearly cycle of rotation impossible. See Li Genpan, “Zhongguo gudai gengzuo zhidu-de ruogan wenti”, Gu-jin nongye 1 (1989). As indicated above, down to the end of the Ming, and even in such a major work dedicated to agricultural technology as Mr. Shen’s agricultural encyclopedia, early-ripening strains of late rice were still being taken to be “early rice.” Kitada may thus have used Ming technical terminology to describe the situation under the Song.

69　Li Bozhong, Agricultural Development in Jiangnan, 1620-1850, London: Macmillan, forthcoming, ch.2.2.

70　Fang Zaihui et al., Zhejiang farming, p.79.

71　Gao Liangzhi and Li Lin, Ecology of wet-f ield rice, pp.364, 375, 377, 378.

72　The last is a tentative reading of tudang. [Trans.]

73　Ma Yilong, Nongshuo, cited in Xu Guangqi, Complete treatise, j.2 “Nong ben,” pp.47.

74　Han Mengzhou, Litang wenji [Collection of Li Tang’s works], 1823 ed., waiji, “Quanyu yedian jitian jia fen wen.” A jin is a measure of weight, historically variable, and strictly speaking incommensurable with units of capacity. Gang Deng, Development Versus Stagnation, pp.XXV, and Chen Hengli and Wang Da, Supplement to the agricultural encyclopedia, indicate that in modern times one modern shi (of husked rice) weighs, on average, 140-50 modern jin.

75　Ma Yilong, Discussion of farming, p.51.

76　Bao Shichen, Four techniques for the governance of the common people, j.2, Nong er.

77　Gao Liangzhi and Li Lin, Ecology of wet-f ield rice, ch.8.

78　A tiller is the shoot of grass, usually lateral and at or near the base, and more or less erect.

79　The question of whether to transplant seedlings at a high density on rich soils was so long a topic of debate in China. With suitable seeds and techniques, dense transplanting will indeed give high yields per mu. The advantages and disadvantages, however, depend on the type of rice. During the Ming-Qing period, the principal wet-f ield rice was long-stalked: if this was densely transplanted, the growth of stalks and leaves would consume a relatively large proportion of fertilizer, and the lack of space between stalks, which impeded the circulation of air, would make the prevention of insect pests diff icult. Hence widely spaced transplanting was comparatively better, as those alive at this time were aware. According to Ma Yilong, “A low or high density is a concomitant of rich or poor soil [respectively] ... If the soil is rich, high density is particularly inappropriate” (see Mark Elvin & Liu Ts’ui-jung eds. Sediments of Time: Environment and Society in Chinese History, Cambridge University Press 1998, pp.50). Xu Guangqi leveled the following criticism against dense transplanting: “The way in which people today make use of seeds is to plant more than a dou per mu, to [trans]plant densely but apply little fertilizer, which makes it diff icult to weed, with the result that harvests are meager” (Complete treatise, j.25, Shuyi zhong). The Wuxing zhanggu ji (in the Wuxing congshu, 1914ed. ), j.13, Wuchan lei, pointed out that “In Huzhou they plow deeply but transplant seedlings at a low density. Their land is essentially very fertile, and many of those who do not transplant at a low density suffer from insect pests in the autumn.” An agricultural handbook from Huzhou in the Daoguang reign (1821-50) also said: “In general, the grain sprouts delight in the wind passing through them. Insects are comparatively rare in places where there is wind.” See Nongshi youwen, partially preserved in Nanxun zhenzhi (1863), j.21, Chuchong. Low-density transplanting in rich soils was common knowledge to many agronomists in Jiangnan in this period.

80　Falling over.

81　Gao Liangzhi and Li Lin, Ecology of wet-f ield rice, p.418.

82　This effect is not conf ined to rice. Yeh-chien Wang has pointed out, on the basis of research by A. B. Appleby and others, that in sixteenth-and seventeenth-century Europe whenever the mean annual temperature fell by 1℃, this could shorten the cropping season by three to four weeks, which was equivalent to increasing the altitude of cultivation by 500 English feet. See Yeh-chien Wang, “Rice prices.”

83　Gao Liangzhi and Li Lin, Ecology of wet-f ield rice, pp.425-26 and 447-48.

84　Zhang Jiacheng, Qihou yu renlei, Zhengzhou: He’ nan kexue-jishu CBS, 1988, pp.123-25.

85　For example, in the eastern part of Huzhou toward the end of the Ming, late-ripening late rice was the main crop. By the Kangxi reign (1662-1722), however, the villagers here were harvesting in the ninth lunar month, and “did not dare to go past the solar period ‘Falling Frosts’ [commonly in the latter half of the ninth lunar month, or from October 23 to November 6.]” A local saying described the ninth month as “the busy month” or the “month when gold is put in baskets and jewels obtained.” See the Wuqing wenxian [Records from Wu-Qing], cited in the Wuqing zhenzhi, 1746 ed., j.2; and Huzhou-fu zhi, 1874 ed., j.29, “Si-shi su, shang,” and j.32, “wuchan, shang.” Thus the category of ripening for late rice in the eastern part of Huzhou changed from a late-ripening to an early-ripening variety.

86　Fang Zaihui et al., Zhejiang farming, pp.71-72.

87　Gao Liangzhi and Li Lin, Ecology of wet-f ield rice, p.377.

88　Nanjing nongxue yuan et al., Crop cultivation, p.136.

89　Siming zhi [Ningbo gazetteer] (1854), j.4, “Wuchan.”

90　Kawakatsu Mamoru, Farming in Jiangnan in Ming and Qing, p.20.

91　You Xiuling, Rice cultivation in the Lake Tai area, pp.10-41.

92　J.21. See also Amano Motonosuke, Agricultural History, p.299.

93　Cited in Xu Guangqi, Complete treatise, j.16, “Zhejiang shuili.”

94　Amano Motonosuke, Agricultural History, p.292.

95　See Chen Hengli and Wang Da, Supplement to the agricultural encyclopedia, pp.38-39. According to the “Monthly tasks” section of Mr. Shen’s agricultural encyclopedia—a late Ming work—the late rice in the eastern part of Huzhou, the main crop in this area, was harvested in the tenth lunar month. The Supplement to the agricultural encyclopedia, which is slightly later, says of transplanting in Tongxiang county, Jiaxing prefecture, that “it mostly takes place after the summer solstice, being harvested at the end of the autumn,” which makes it unquestionably a late rice. The “Monthly tasks” section of Mr. Shen’s agricultural encyclopedia also describes rice ripening in the ninth lunar month as “early rice,”and the “Yun tiandi fa” section also says that “early white rice” (the same variety) is hard to grow, as “it is not easy to apply the fertilizer in a properly balanced way.” However, though “the kernels are coarse and hard, it is much used for food, [so]a large quantity should be planted.” The principal grain at this time was thus late-ripening late rice, which Shen calls “late rice” or “yellow rice” (huang dao).

96　Li Bozhong, “Ming-Qing Jiangnan zhongdao nonghu shengchan nengli chutan”, Zhongguo nongshi No.3, 1986.

97　A rough statistical analysis of local gazetteers shows that from mid-Ming to mid-Qing (and especially in the Qing) transplanting occurred mainly in the f ifth lunar month, with the maximum frequency between the solar periods of Grain In Ear and Summer Solstice (approximately June 6 to July 6), and harvesting mainly in the ninth lunar month. See Li Bozhong, “Productive capacity of rice farmers,”and Agriculture in Jiangnan 1622-1850, ch.2.2. Seen from a present-day point of view, early, intermediate, and late rice have growing periods of, respectively, 120, 150, and 180 days. See Amano Motonosuke, Agricultural history, p.291. The seedlings take about a month to grow, hence a posttransplanting growing season from, say, the f irst half of the f ifth lunar month to late in the ninth lunar month-implying a total growing season of about 30 + 15 + (3×30) + 15+=150+ days-would tend to indicate slow-ripening intermediate rice or early-ripening late rice.

98　Li Bozhong, “Raising the intensiveness of cultivation.” The increase was due in part to the generalization of the use of supplementary fertilizer, and in part to an increase in the amounts of basic fertilizer applied.

99　Nanjing nongxue yuan et al., Crop cultivation, p.74.

100　Beijing nonglin ju, Nongye changyong shuzi shouce, Beijing: Nongye CBS, 1980, pp.186-87, 193.

101　Dwight Perkins, Agricultural Development in China, 1368-1968, Chicago: Aldine Publishing, 1969, p.73.

102　Supplement to the agricultural encyclopedia, “Bu Nongshu hou,” p.106.

103　Fang Zaihui et al., Zhejiang farming, pp.161-67.

104　Ibid., pp.40, 309-10.

105　Li Bozhong, Rational utilization of agricultural resources.

106　Yu Yefei, “Zhongguo lidai liangshi pingjun mou-chanliang kaolüe”, Chongqing shifan xueyuan bao No.3, 1980.

107　Wu Hui, Zhongguo lidai liangshi muchan yanjiu, Beijing: Nongye CBS, 1985, pp.169, 175-77.

108　Wu Chengming and Xu Dixin, eds, Zhongguo zibenzhuyi fazhanshi diyijuan—Zhongguo zibenzhuyi de mengya, Beijing: Renmin CBS, 1985, p.191. The English edition of this book was published by Palgrave Macmillan in 2000, with the title of Chinese Capitalism, 1522-1840.

109　Li Longqian, Ming-Qing jingji shi, Guangzhou: Guangdong gaodeng jiaoyu CBS, 1988, p.407.

110　You Xiuling has pointed out that at the present time the highest yields are from three to f ive times the average, and that it takes more than twenty years to raise the level of the large number of average f ields to the level of the small number of outstanding ones. See You Xiuling, Rice cultivation in the Lake Tai area, pp.10-41. In Ming and Qing times, with a social organization and techniques of technology transfer inferior to those of the present day, it would have taken longer.

111　Although the f igures of the Wanli survey were not able to escape from those factors indicated by He Bingdi [Pingti Ho]as making for the off icial overestimation and underestimation of land throughout Chinese history, in comparison with other regions in China these were of slight effect in Jiangnan, at least since Southern Song times. See He Bingdi, “Nan Song zhi jin tudi shuzi de kaoshi he pingjia”, Zhongguo shehui kexue No.3, 1985. The Wanli survey began in Jiangnan and was most conscientiously carried out in that region. Furthermore, the opening up of new lowland farmland ( “extensive development” ) had basically drawn to its close in Jiangnan by this time, and later fluctuations would not have changed its magnitude to any great extent. That Wanli off icial f igures for farmland were basically used throughout the Qing Dynasty was mainly due to these reasons, and not necessarily to a purely conservative retention of the original numbers or negligence in investigating.

112　For a detailed discussion, see Li Bozhong, Agriculture in Jiangnan, 1620-1850, ch.2.2.

113　Li Bozhong,” Kongzhi zengzhang yi bao fuyu: Qingdai qian-zhongqi Jiangnan-de renkou xingwei”, Xin shixue (Taipei) 3.3 (1994).

114　Yeh-chien Wang [Wang Yejian], “The Impact of the Taiping Rebellion on Population in Southern Kiangsu,” Papers on China 19 (Harvard) (1965).

115　Liang Fangzhong, Zhongguo lidai hukou, tiandi, tianfu tongji Shanghai: Shanghai renmin CBS, 1980, pp.273-79.

116　C.W. Skinner, “Sichuan’s Population Data in the Nineteenth Century: Lessons from Disaggregated Data,” Late Imperial China 8.1 (1987).

117　See Li Bozhong, “Preserving prosperity.”

118　The alcohol industry made wine, vinegar, soy sauce, and brewer’s yeast (qu), which latter, apart from being used for the production of wine, was also employed in the making of vinegar, soy sauce, dyes for foodstuffs, and traditional Chinese medicines. See Hong Guangzhu, Zhongguo shipin ke-ji shigao (Beijing: Zhongguo shangye CBS, 1984), pp.76-88. Jiangnan people had long been drinkers of rice wine, especially “yellow wine” (huang jiu), so the quantity would have been substantial. Bao Shichen states that in Suzhou around the beginning of the nineteenth century, the quantity of rice used each year for wine amounted to more than several million shi. See Four Techniques for the Governance of the Common People, j.2, Nong er. He also observes that in Suzhou at this time, “Five or six jin of yellow wine was not a large quantity for someone to drink.” Zhang Lvxiang said of the early Qing “It is normal for a man to drink several sheng a day [1 sheng= 0. 01shi]. As for heavy drinkers, there is no counting their consumption.” See Chen Hengli and Wang Da, Supplement to the agricultural encyclopedia, pp.160. Even assuming no change in the level of per capita consumption of wine in Jiangnan between the Ming and the Qing, the quantity of rice used for making alcohol would have roughly doubled by mid-Qing in contrast with late Ming. See Li Bozhong, Qing economic history, ch.4, sec.2.1.

119　Shiba Yoshinobu, “Sōdai no shōhi, seisan suijun shitan”, Chūgoku bunka1.1 (1991).

120　Quan Hansheng and Wang Yejian, “Qing Yongzhengnianjian de mijia”, “Zhongyang” yanjiuyuan lishi yuyan yanjiusuo jikan 30 (1959); Wu Chengming, “Lun Qingdai qianqi woguo guonei shichang”, in Wu chengming, Zhongguo zibenzhuyi yu guonei shichang, Beijing: Zhongguo shehui kexue CBS, 1985, pp.257-58; and Wang Yejian and Huang Guoshu, “Shibashiji Zhongguo liangshi gongxu de kaocha”, in “Zhongyang” yanjiuyuan jindaishi yanjiusuo, Jindai Zhongguo nongcun jingjishi lunwenji, Taibei: “Zhongyang” yanjiuyuan jindaishi yanjiusuo, 1989. The estimate is in Li Bozhong, Qing economic history, ch.4, sec.2.

121　D. Perkins, Agricultural development, pp.318-19, table G4.

122　Wu Chengming and Xu Dixin, ed., Sprouts of Capitalism, pp.40-41 and 190-91.

123　Min Zongdian, “Song-Ming-Qing shiqi Taihu diqu shuidao muchanliang de tantao”, in Zhongguo nongshi, No.3, 1984.

124　Liu Yongcheng, “Cong zuce, Xingdang kan Qingdai Jiangsu diqu de liangshi muchanliang”, paper presented to the Zhongguo jingjishi xuehui second annual conference proceedings, Hu’ nan: Zhangjiajie, 1993.

125　Guo Songyi, “Liangshi shengchan de fazhan” in the Qingdai jingji shi [Qing economic history], Beijing: Zhongguo shehui kexue CBS, forthcoming.

126　There are reasons that probably explain this situation. The weather was not too bad in the twenty or thirty years before 1850 (see Yeh-chien Wang, “Secular Trends of Rice Prices in the Yangzi Delta, 1632-1935,” in T. Rawski and L. Li, eds., Chinese History in Economic Perspective, Berkeley: University of California Press, 1992) and would have had a limited adverse effect on wet-f ield rice production. There was, on the other hand, a great increase in the interprovincial import of fertilizer, in the form of beancakes, and the effect of this should not be underestimated. The sheer hard work of the Jiangnan people clearly also had an effect. In Huang Ang’s Xi-Jin zhi xiao lu [Records of small matters in Wuxi and Jingui] (eighteenth century), j.1, “Bei can shang,” he says: “ [The peasants’] diligence in working the farmland was not equaled in earlier times. When they met with droughts or floods they previously used to be in dread of the diff iculties and stopped midway [in their efforts to resist]. Today they exert themselves to the utmost to undertake rescue operations.” Although this refers to the region of Wuxi and Jingui during the Qianlong reign, there is no reason to think that this was not also the case with many parts of Jiangnan after this time.

127　Shiba Yoshinobu, Jiangnan economy under the Song, pp.7-65.

128　Chen Hengli and Wang Da, Supplement to the agricultural encycdopedia, ch.2 and pp.26-34, 146-47, 164, 181-84, 189-90, 250.

129　My 1984 analysis in Li Bozhong, “Intensif ication of cultivation,” did not take into account the labor involved in the application of fertilizer or in pumping water. Their inclusion should raise Qing labor inputs above the Ming level. In normal years, however, the require-ment of rice for water is constant, so that changes in double-cropping apart, the input of labor per mu could not have varied greatly. (See also, however, the comments on pp.455-56 above. ) Labor input for mid-Qing times may be aff irmed to have been greater than in Ming. On the other hand, when beancakes were used in the place of other fertilizers, the Supplement to the Agricultural Encyclopedia states that there were “economies both in labor and in the amount of fertilizer,” and the use of beancakes spread widely in Qing (especially mid-Qing) times. Thus the increase of labor inputs for fertilizing would not have increased commensurately with the increase in the use of fertilizer. Furthermore, according to the Report on Agriculture in the Huangpu and Mao [lake] region, the combined labor requirement for weeding and the application of (supplementary) fertilizer in the rice paddies in the western part of Songjiang during the Daoguang reign was only one working day (per crop) per mu. And the labor required for pumping in the rice paddies of Songjiang, according to calculations based on records in Shen Jingxian’s Maodong caotang biji, on which see Wei Jinyu, “Guanyu Zhongguo nongye zibenzhuyi mengya de jige wenti” in Nanjing daxue lishixi Ming Qing shi yanjiushi ed, Zhongguo zibenzhuyi mengya wenti lunwenji, Nanjing: Jiangsu renmin chubanshe, 1983, was less than 2.5 labor-days per mu (per crop). This was less than a quarter of the 10 labor-days needed for “plowing leveling, weeding, and harvesting.” There would thus appear to have been limits on the quantity of extra labor spent on the application of extra fertilizer and pumping even by the mid-Qing. Thus any increase is unlikely to have been great when compared with late Ming times.

130　Wang Yejian [Yeh-chien Wang], “Secular trends.”