NOTES
1.
Pan (1998: 10—11). Lest this seems absurdly high, Pan also provides a plausible reconstruction of daily labour effort during busy months that shows how these labourers could have consumed that much.
2.
See Marks (1991: 77—8) for a justification of this figure.
3.
Guo (1994: 46—7) estimates grain production in that year as 275,737,216,000 jin of unhusked rice and its equivalent, which (converting at an average of 55% to edible grain) becomes 151,655,468,800 jin, or just over 1,000,000,000 shi of edible rice equivalent. Subtracting perhaps 15% for next year's seed would give us 850,000,000 unhusked shi for people to eat. If the population was roughly 225,000,000 at the time (Guo uses the official figure of 184,000,000, but that is generally considered too low), we would have about 3.77 shi per person per year: about 70% above the 2.2 shi ‘average’ suggested by Marks and adopted here. This would convert to about 300 k of rice per person per year: enough to support 3,200 calories per day for everyone, or well over 4,000 calories per day per adult equivalent. It should be noted, however, that Guo's estimating techniques are quite rough, and no more precise calculation is possible with the materials known to be available.
4.
See Perkins (1969: 301).
5.
6.
7.
8.
9.
10.
11.
Cited in Aymard (1979: 6).
See, for example, Wong (1982, 1997), Will (1990: 298-9, 302-10), Tilly (1975), Grantham (1989), Shiue (1998). See, for example, Post (1977: 44-6) for famine conditions in various portions of Germany, and especially in the Habsburg lands; and Post (1977: 40, 108) for a comparison with the less extensive, but still regionally devastating famines of 1845-7.
Guthrie (1971: 71) cites a biological value (BV) of 86 for rice protein (egg=100, fish and beef=75, and wheat gluten=44), with a BV of 70 considered capable of supporting growth as long as quantities are adequate.
Guthrie (1971: 69) tells us that it takes 1,186 kcal of rice to provide 22 g of protein (versus fewer than 200 of most meats and fish). At this rate, it would take 3,500 calories of rice alone to meet the US RDA for an adult male from rice alone (Guthrie 1971: 68)—more than the average Chinese intake, but less than that of the farm labourers cited above, who would take in almost 85 g of this high-quality protein. It has been more conventional to argue that rice protein, like most other vegetable proteins, is an incomplete protein, which needs to be supplemented by others, which have large amounts of the amino acids it has relatively little of. This argument, however, is based on rat studies, which are not conclusive, and the biological value scale for protein is based, to the extent possible, on human studies. The point is still disputed, but the trend in opinion seems to favour a more generous valuation of at least some vegetable proteins. See especially D. J. Milt-Ward et al. (1992) who say older methods ‘overestimate the value of some animal proteins for human needs while underestimating the value of some vegetable proteins'.FAO (1974: 19), Guthrie (1971: 65, 68). See also Buck (1937/1964: 420) for a slightly lower figure.
Much of the reason for the switch was that soybean cake was far more concentrated—providing perhaps twenty times as much nutrient per pound applied—so that making the switch saved vast amounts of labour, see Pan (1994: 36-8, 110-13). That advantage would obviously be lost if each farm owner or labourer continued to apply just as much night soil while also adding the new work of applying soybean cake.
Buck (1937/1964: 259). Given Buck's estimate of the relative fertilizer output of humans and large livestock (1930/1971: 224), this would suggest that the ‘Yangzi Rice/Wheat' area should have had considerably less than half the number of large animals per person as the rest of the country. Buck's figures however, suggest that the number of large animals per farm, per person, and per crop acre in this region varied from a bit under 0.6 to a bit over 0.6 for China as a whole (1937/1964: tables: 246, 255).
Eitherway, the basic point being made here would hold.12 Early twentieth-century population in the Yangzi Delta was not much above mid-eighteenth-century figures, but those figures were already 75-100% above those for the late Ming, see Li (1994^: 30-3).
13' Note also Perkins' comment (1969: 301-3) that even some of Buck's grain availability data seem implausible unless China was converting much more grain than he finds plausible to either meat or alcohol. While Buck's figures may indeed be too high in many cases, the possibility of at least some additional indirect consumption cannot be ruled out, even for a period conventionally considered one of agrarian crisis.
14 Lis and Soly (1979: 113). For conversion of beef into grams of protein, see Watt and Merrill (1963: 12-13).
15 Cited in Aymard (1979: 6-7, 14).
16.
17.
18.
19.
I assume here that most of the starch in the diet described by Toutain (making up roughly 70% of its calories) was in the form of relatively nutritious whole wheat bread, and estimate protein content based on Watt and Merrill (1963: 18-19).
Oddy (1990: 274). This figure appears to be a conversion of the estimate on p. 269 of the amount of meat in the diet.
Buck (1930/1971: 371) combining his categories of legumes and vegetables, with almost 90% coming from the latter category, and cabbage alone making up one-third of the total.
Compare Lavely and Wong (1998), especially table II and figure III, and Lee and Campbell (1997: 79) with Wrigley and Schofield (1981: 230, 708-13) and see Razzell (1993: 757-63) for a suggestion that these figures are too high. (Razzell's suggested adjustment for infant mortality alone would bring life expectancy at birth of 37.0 down to somewhere between 31.6 and 34.0.) For continental examples, see Knodel (1988: 68-9) and Blayo (1975) (showing a much lower life expectancy in France).
20.
21.
22.
23.
24.
25.
26.
See Li Zhongqing (1994: 3).
Li Bozhong (1994^: 32-4); compare with McEvedy and Jones (1978: 28-9).
Shammas (1990: 136). Here she relies on late eighteenth-century figures, though there is little reason to assume that consumption of other foods rose for the poor until after 1840, see Mokyr (1988: 90-1).
See the complaint about peasants' ‘gaudy’ clothing at religious festivals by the official Chen Hongmou in Chen (1820/1962: 4a-6b).
His ‘gentry’ is sufficiently loosely defined that it would include many of the richest merchants, though only if they purchased honorary degrees—a more common phenomenon in the nineteenth than the eighteenth century.
See Pomeranz (2000a: 136-7); Lindert and Williamson (1982: 393, 396-7, 400-1).
Reed (2000: 206-7). He adds a casual estimate (the point is not important for his argument) that it would have cost about 10 taels to feed an adult for a year at this time and place, but I suspect the actual number would be significantly lower. See note 29 for a much lower estimate of the cost of eating at roughly the same time and not very far away.
27.
28.
Reed (2000: 45-51, 146-8). My figure for runners averages Reed's estimates of statutory runners and extrastatutory runners for 1848 and 1876, omitting the abnormally low one for 1855 (a year of civil war).
29.
30.
31.
In Pomeranz (2000a: 249-50) I discuss some (among many) possible reasons for Chinese families of this period not to automatically head for areas with higher per capita incomes. Sommer (2000) is a book that puts great emphasis on the increased number of people on the move in eighteenth-century China, and shows that the Qing were quite worried about what these movements suggested, but there is no clear evidence to show that their numbers grew faster than one would expect given the general population rise and on-going commercialization. Cited in Xu and Wu (2000: 344-5).
This assumes that we can use both the retail grain prices reported by another worker in the salt works, which refer to an unusually large local dou (a volume measure), as cited in Xu and Wu (2000: 345).
Much of the data was collected in the 1950s and 1960s projects aimed at demonstrating the importance of proletarian class struggle over the centuries in China, and thus may have selected examples showing the largest disparities. This would be particularly true for late Qing data, which sometimes was based on the recollections of elderly workers. Also, to the extent that there were uncounted in-kind supplements to cash wages, there is no reason to assume that they rose proportionately with cash wages, since no worker would need eight times (much less twenty times) as much grain as he could eat. On the other hand, in-kind wages may have increased in step with cash wages after all. The government paid certain workers with grain allowances that far exceeded their needs and that they then sold on the market, and the practice might have been copied elsewhere. The salt works material also tells us of certain kinds of in-kind payments and extra allowances (e.g. for tobacco) that were only granted to those higher up the ladder.
32.
33.
34.
35.
Again, assuming the use of local measures—see note 31.
Pomeranz (2003). See also Pomeranz (2000a: 319) and compare Horrell and Humphries (1995: 102—3).
The Qing were intensely concerned about people leaving the land, and complained about what was undoubtedly a rising number of people doing so—but China remained, in comparative perspective, a remarkably rural society, and the share of urban population was probably flat or falling slightly in the eighteenth century.
Particularly striking accounts may be found in the novels Jin ping mei and Xingshiyinyuan ghuan—striking in part because they deal with a medium-sized city and a small town, respectively, in north China rather than with any of the country's great metropolises. For some reflections on consumption in China by a leading historian of early modern European consumption, see Burke (1993: 148—61). I deal with this at much greater length in Pomeranz (2000a: ch.
3).36.
37.
38.
Cited in Dermigny (1964: III, 1253).
DaCruz in Boxer (1953: 106, see also 99).
By starting with the quantity of land reported on the tax rolls, we build in a big conservative bias, since underreporting was chronic throughout China. I have used the highest estimates I could plausibly defend of the amount of land that was under basic grain crops, and where estimating cash-crop production for an area was particularly tricky, I have simply omitted it from national totals, even though contemporaries may have remarked often that it produced the good in question. In the case of sugar, for instance, I have counted only output in Guangdong and Taiwan plus known imports, though we know that mainland Fujian was also a major producer, and production scattered through the rest of China was estimated by a contemporary to be about one-ninth of the total of Guangdong, Taiwan and that uncounted mainland Fujian output (cited in Daniels 1996: 97, 105). And within Guangdong itself I have used a figure for cash-cropping area more than 20% below that generated in Robert Marks' study of that province, and assigned only one-tenth of this cash-cropping area to sugarcane: a figure that Marks suggests is almost certainly too low For further discussion, see Pomeranz (2000a: ch. 3).
39.
40.
41.
42.
43.
44.
Buck (1937/1964: 442 on floor space, 443 table 5, on building materials).
See, for example, Kuhn (1990: 43—7, 111—12), Sommer (2000: 12—14, 99—101).
I analyse fashion, partly along these lines, in Pomeranz (2000a: 152—62).
This remains the pattern even today in much of rural China—much to the state's dismay—a high priority for many families as they have become more prosperous since 1978 has been to extend and strengthen their web of social ties through hosting more ritual events, and greatly expanding the guest lists at them, see Yan (1996: 225—6). See, for example Rowe (1992), Mann (1997), Sommer (2000).
A classic statement of the stagnation thesis is Fairbank, Reischauer, and Craig (1973: 435—44, 643—6). Eastman (1988: 80—1, 92), representing a later generation of scholars,
nonetheless sees the choice as one between two different explanations of why Chinese peasants (in Tawney's famous phrase) were in water up to their necks, and then, as I do, emphasizes the role of politics in explaining the increased frequency of ‘ripples' that could drown such a person.
45 See Lee and Wang (1999: 84-92, 97-9), Li (1994fc 41-2, 46-52).
6 Y. C. Wang (1989: 423-30, on rice), Lu (1992: 493, on raw cotton), Marks (1997: 172-6, on sugar), Li Bozhong (2000: 321-7, on timber), Adachi (1978, on soybeans).
47' Compare Braudel (1981: 127).
48' Skinner (1977: 234-5, 713, n. 30-2), Ch'uan and Kraus (1975: 77). If Ch'uan and Kraus are right, and Perkins is right about the 1930s, these shipments had declined by somewhere between 73% and 82%. A decline of this magnitude seems hard to believe, but it is also not necessary to my argument.
4,- Lu (1992: 482-3), Bray (1997: 217). Bray quotes a source saying raw cotton exports from the North had already become trivial in the eighteenth century, but this seems an exaggeration.
50 See Menzies (1996: 562) and Xu and Wu (2000: 240-6); on transport problems, see Osborne (1989: 52-6, 71, 81-5); on how these problems were complicated by increased exactions by officials and local strongmen, see Lu and Chen (1986: 31-2).
5 See Lee and Wang (1999: 116) for a useful map showing this pattern (but using provinces as the unit of analysis), and Pomeranz (2000^: 45) for a table producing some salient examples.
5 See Pomeranz (2000a: 123-4).
53' Pan (1994: 115-23, 344-83, see 382-3 for a summary statement).
54 See Pomeranz (2000^: 50-3) for a brief account; Pomeranz (2000c) makes the case in more detail.
5 See for example Pomeranz (2000a: 237-8), Pan (1994: 116-18), Xu and Wu (2000: 231-3, 257-9, 287), Vermeer (1991: 311, 325-8).
5 6 On temperatures, see Marks (1997: 217-19). On the unusually severe El Ninos and their effects, see Davis (2001: 64-79, 248-51, 271, 341-75).