NOTES
Tokugawa era, in which Japan was under the control of the Tokugawa shogunate in Edo (present Tokyo), begins in 1603 and ends in 1868.
According to Hayami (2001: 46), the rapid population increase in the seventeenth century was mainly due to large- scale cultivation of new land that took place across different regions of Tokugawa Japan.
Hence, the seventeenth century is called the ‘period of great land reclamation'.Examining data on the physical well-being and nutritional status of residents in large cities such as Tokyo and Kyoto, Hanley (1983) argued that the standard of living in mid-nineteenth-century urban Japan was not only higher than that in the einghteenth century, but relatively high in comparison to the industrializing West. However, her arguments do not seem to be applicable to living standards in rural villages, especially those in the northeast.
Sericulture became popular in the region at that time, and mulberry leaves were major cash crops. Existing historical records show that several neighbouring villages located in flat land had much higher proportions of fields used for growing mulberry trees (Narimatsu 1985: 53—4).
According to a survey by the domain government in 1828 on the use of agricultural land, around 30% of the dry field in Niita was cultivated as mulberry field whereas only 5—10% of the dry field in Shimomoriya was used to grow mulberry trees (Nihonmatsu-shi 1982: 581).
Saito (2002) reports that according to one source (Ogashima 1894) there were twenty-eight major famines recorded in the Tokugawa period, and according to another (R. Saito 1966) there were sixty-one estimated famines from 1600 to 1900.
Being only around 40 km to the west, Aizu is the neighbouring domain of Nihonmatsu. Unfortunately, rice price series of the Nihonmatsu market are not available, being lost some time in the history.
Thus, this study uses rice prices in the Aizu market. Given the geographical proximity of Nihonmatsu to Aizu, the use of Aizu price series is thought to be appropriate. Further, the Niwa family which was the domain lord of Nihonmatsu was in debt to Aizu merchants from the late sixteenth century and the debt was never paid off (Nagata, Kurosu, and Hayami 1998). For this reason, the Aizu rice prices probably had strong influences on the tax rates and finances of the Nihonmatsu domain government and may reflect changes in the tax rates imposed upon peasants in the domain. One koku is equivalent to approximately 5 bushels.According to Hamano (2000) who examined different measures of household economic status, household landholding is the most appropriate and most widely used indicator of household economic status in pre-industrial Japanese villages.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10 Among the major regions of Tokugawa Japan, the northeast was in general the one with the least economic development and lowest proto-industrialization. Compared to central and southwestern Japan, the overall level of proto-industrialization in the northeast is considered to have been much more limited.
11 The population registers in Shimomoriya and Niita, like all other communities in the Nihonmatsu domain, were ninbetsu-aratame-cho (NAC), rather than shumon-aratame-cho (SAC). Though similar in terms of information collected by these two types of registers, SAC was carried out, in its original purpose, to hunt hidden Christians whereas NAC was, as the name indicates, primarily for population registration and investigation (Narimatsu 1985: 11—14, 1992: 10-12).
12' The size of de jure population can also be computed for both villages because records were kept as far as one's permanent (legal) domicile was in the villages. However, for persons whose legal domicile was in the village but were not present (residing) there, information on individual circumstances, including demographic events occurring while away from the villages, are generally unavailable.
13 Eighteen out of the nineteen disappearances from Shimomoriya's NAC and seven out of the thirteen disappearances from Niita's registers occurred during 1851-70, the last years of the Tokugawa period.
14' In addition to chronological age (i.e. age according to the Gregorian calendar) and NAC age, there is also the traditional Japanese method of counting age. As in the rest of East Asia, it regards a child as age 1 at birth and adds an additional year on each New Year's Day thereafter. Consequently, if counted by the traditional Japanese method, most newborns, if they survived, appear in population registers at the age of 2 sai although in extreme cases they could be on the second day of life. If population registration was conducted on each New Year's Day (which was rarely the case), traditional Japanese age (in sai) minus one is equivalent to NAC age.
15' According to Guilkey and Murphy (1993), when records are repeated over five times, the problem of intercorrelation among observations becomes serious and affects the estimation results.
16 The formula was independently discovered by White (1980) and is also known in the econometrics literature as White's method. A summary discussion of the method is given in the STATA Reference Manual (Stata Corporation 1995: 457-65).
17. As mentioned earlier, nine registers for Shimomoriya and five registers for Niita are missing. Thus, for certain two- year periods (and the five-year period from 1863-7 for Shimomoriya and the three-year period from 1856-8 for Niita) it is impossible to deter mine in which one-year interval an event or right censoring occurred. Accordingly, we restricted our analysis to the years for which an immediately succeeding register is available.
18 Though some women gave birth before age 15 in the two villages, the number was very small. Out-of-wedlock childbearing was also minimal. Only 0.9% of recorded births (i.e.
twenty-three births) were out of wedlock. For details, see Tsuya and Kurosu (1998).”■ The one-year lag for husband is due to the assumption that given the gestation period of approximately nine months, a woman could be at risk of giving birth as long as her husband was present in the village at the previous enumeration.
0 Though the youngest recorded age at marriage was 6 for males and 3 for females, such cases were very rare and seem to require a separate explanation. We therefore exclude these outliers from the analysis.
21 Though substantial numbers (304 men and 728 women) migrated into the villages upon first marriage during the 154 years of observation, these individuals had to be excluded from the multivariate analysis.
22 When migration consisted of more than one individual (i.e. migrants did not move alone), we counted the event pertaining to each individual as one. For example, when a household of three members moved out of the village, we counted three events; and all the three were used in the analysis. This multiple counting of events in the case of migration of a whole (or part of) household would not seriously affect the results of our analysis because a large majority (90%) of out-migration was by a lone individual.
22 A large majority (82%) of out-migration of male non-natives was mostly due to end of service. As for outmigration of female non-natives, two most common reasons were end of service (54%) and marital disruption (38%).
24' For details on various rice price series in early modern Japan, see Iwahashi (1981), and for the relationship between local rice markets and the Osaka central market in TokugawaJapan, see Miyamoto (1988: 386—430).
25' Since the two villages under consideration almost exclusively consisted of peasants, we cannot use occupation of household head as an indicator of household socio-economic status as occupation of household head would be a constant, rather than a variable.
6 For the table showing the estimated coefficients of logged rice prices with different time lags on the probability of dying by sex and life stage as well as the tables on marital fertility, first marriage, and out-migration, contact N. Tsuya (tsuya@econ.keio.ac.jp).
7 Previous studies on pre-industrial Japanese villages also found that family control was not limited to the times of economic and environmental hardships. See Saito (1992), Skinner (1987), and Smith (1977: 59—85).
28' The mean number of children ever-observed was 2.8 for a woman whose first marriage was completed.
2,- We also conducted the analyses of first marriages by type of marriage for males and females in these two subperiods separately. The analyses of male first marriages by type of marriage showed that the temporal patterns of responses of all male marriages were virtually the same as those of responses of male virilocal marriages, as expected from the fact that a large majority of male first marriages were virilocal. As for female first marriages, we could not find any clear and discernible temporal patterns by type of marriage because the number of events for each of the three types of marriages became too small by slicing the data many times. Given these findings, we decided to discuss in the text only the results of the analyses on all first marriages by two sub-periods for males and females, respectively.
30 For the tables showing the results of the event history analyses of the four demographic responses to rice price fluctuations and household landholding by the two sub-periods, contact N. Tsuya (tsuya@econ.keio.ac.jp).