2016 Volume 124 Issue 3 Pages 185-197
The life history of a female individual skeleton (ST61) from the Edo period (AD 1603–1868) was investigated by using multi-tissue and multi-isotope analyses. Her gravestone and historical documents revealed that ST61 was a grandmother of a chief retainer of the Akashi clan who died in 1732 aged 77 years. Radiocarbon and sulfur stable isotope analyses indicated that the contribution of marine foods to the ST61 diet was relatively low (17.2% protein) despite the relatively higher nitrogen isotope ratio of the rib bone collagen. Carbon and nitrogen stable isotope analysis of the serial section of tooth dentin along the growth lines indicated that breast milk was not the major protein source of ST61 after roughly 1–1.5 years of age, although this weaning pattern was not evident from the oxygen stable isotope ratios of her tooth enamel serial sections. The carbon stable isotopes in tooth dentin collagen and tooth enamel apatite suggested that her diet from 0.5 to 5 years of age possibly contained a small proportion of C4 plants. Stable isotope ratios of the rib bone and the tooth dentin collagen differed, consistent with historical documents describing a residential change at the age of 27. The calibrated radiocarbon ages of the associated rice hull were at least 80–120 years older than the year of death of ST61. Sulfur stable isotope ratio of the rice hull suggested that fish fertilizers might have been used for paddy rice at that time. Multi-tissue and multi-isotope analyses can provide information of several kinds from different time windows even from an individual skeleton.
Detailed reconstruction of the life history of individuals is one of the important topics of bioarchaeology, and isotope analyses are valuable analytical tools in this sense. Isotopes of different elements provide different information about life history (e.g. chronological period, diet, and migration). The greater the number of isotope analyses combined, the more detailed the information that can be obtained. Different types of human skeletal tissues from archaeological sites record isotopic signals accumulated during different time windows in their life history. For example, most adult bones record average isotopic signals for 10–20 years before death (Stenhouse and Baxter, 1979; Hedges et al., 2007; but see Jørkov et al., 2009), and tooth retains the isotopic signals recorded during non-adulthood (reviewed by Humphrey, 2014; Tsutaya and Yoneda, 2015). Sometimes, isotope analysis of associated remains and/or evidence from historical documents further provide valuable information about life history (e.g. Van Strydonck et al., 2016). Several studies using multi-element isotope analyses on multi-tissues have been conducted to reveal life history at individual levels (e.g. Sealy et al., 1995; Cox and Sealy, 1997; Schroeder et al., 2009; Knudson et al., 2012; Lamb et al., 2014), although the number of these studies is not large compared with those carried out at the population level.
The objective of this study is to apply multi-tissue and multi-isotope analyses to an individual skeleton (ST61, a grandmother of a chief retainer of the Akashi clan in the Edo period, AD 1603–1868) to reconstruct her life history. Detailed analyses of physical characteristics of her skeletons have already been conducted (Nagaoka et al., 2013). Although the diets of people of different social status in different regions of Japan during the Edo period have been reconstructed at a population level in several isotopic studies (e.g. Yoneda et al., 2011; Tsutaya et al., 2016), the dietary change through life history at the individual level has never been studied for individuals from the Japanese archipelago. Among these works, isotopic studies on high-status people are limited, and this study provides valuable data of people from the Edo period. The availability of historical documents on this individual is another advantage of this study. Isotopic results will be discussed by considering the information from these sources, thereby providing more detailed evidence.
Isotope analysesCarbon and nitrogen stable isotope analysis has been applied to collagen extracted from ancient skeletons to reconstruct past human diet (Lee-Thorp, 2008). Carbon isotope ratios (δ13C values) of plants differ by the type of photosynthesis (i.e. C3, lower; C4, higher), and this difference is reflected in the isotope ratios in the consumers of the ecosystem (O’Leary, 1988; Schoeninger and DeNiro, 1984). Thus, the nitrogen isotope ratios (δ15N values) are higher in organisms at higher tropic levels because of bioenrichment (Minagawa and Wada, 1984; Schoeninger and DeNiro, 1984). Organisms from marine ecosystems usually possess higher δ15N values than terrestrial ones because of their longer food chains. The δ13C and δ15N values of bone collagen mainly reflect those of dietary protein (Ambrose and Norr, 1993).
Sulfur isotope analyses have been utilized to assess the marine or freshwater foods versus terrestrial ecosystem foods ratio (Privat et al., 2007; Fornander et al., 2008; Nehlich et al., 2011). Sulfur in bone collagen exists as essential amino acid residues of methionine (Eastoe, 1955), which originally derived from the diet. Sulfur isotope ratios (δ34S) of animals from marine ecosystems cluster between +15 to +19‰ (Nehlich et al., 2013; Nehlich, 2015) because of the uniformity of sea water δ34S (Rees et al., 1978; Chukhrov et al., 1980). On the other hand, those from terrestrial ecosystems could vary owing to the wide range of δ34S in freshwater and terrestrial organic matter (from −22 to +22‰; Peterson and Fry, 1987). The sulfur isotopic difference between preys and consumers is fairly small (i.e. +0.5 ± 2.4‰) in mammals as well as in other animals (reviewed by Nehlich, 2015). These signatures of δ34S provide a good indicator for the dietary sources of the consumers.
Marine organisms possess ‘older’ radiocarbon ages because of the marine reservoir effect, which can be used to estimate the proportion of marine foods in diet (Craig et al., 2013). Because ‘older’ atmospheric carbon dioxide dissolves into seawater and is preserved in deep water for a long time, marine products typically possess 400 year older radiocarbon ages (Stuiver et al., 1986; Stuiver and Braziunas, 1993). Organisms consuming marine products also have ‘older’ radiocarbon ages, and the difference is larger for organisms consuming larger proportions of marine products. Usually, the marine reservoir effect is corrected by considering the proxy of the dietary proportion of marine foods (e.g. δ15N values of bone collagen) in the target individuals to obtain the true chronological age (e.g. Tsutaya et al., 2014a). However, if the true chronological age is known, the dietary proportion of marine foods can be back-calculated by considering the marine reservoir effect (Craig et al., 2013; see also Materials and Methods section). Although there is a local variation in the marine reservoir effect owing to the mixing between deep and surface water, several correction factors have been studied worldwide including Japan (Stuiver and Braziunas, 1993).
Teeth retain the sequential record of isotopic signals because of their incremental growth, whereas bones provide mixed information from a relatively wider time window for isotopic signals (Humphrey, 2014; Tsutaya and Yoneda, 2015; but see Bell et al., 2001). Dentin and enamel tissues in tooth are incrementally formed in the shape of stacked cones from dentin horn to root tip to cervix, respectively, during prenatal and postnatal periods (Schour and Massler, 1940; Hillson, 1996; Smith, 2008; Nanci, 2013). The ages at the start and the end of formation differ among the different types of teeth (Hillson, 1996; Smith, 2008). The isotopic signals in non-adulthood recorded during the incremental growth period remain throughout their life once formed, because dentin and enamel tissues do not remodel after being formed as long as they are not affected by any pathology and physical stress (Nanci, 2013). The sequential dietary change corresponding to these periods can be obtained by sequentially analyzing the carbon, nitrogen, and oxygen isotopic change along the tooth growth (Humphrey, 2014; Tsutaya and Yoneda, 2015).
Carbon and nitrogen stable isotope analyses of tooth dentin have been typically used to reconstruct diet in non-adulthood, especially for breastfeeding and weaning practices (Humphrey, 2014; Tsutaya and Yoneda, 2015). Because breast milk is more enriched in 15N than maternal diet, δ15N values typically increase 2–3‰ in subadult tissues during exclusive breastfeeding periods and decrease to adult values during weaning (Fogel et al., 1989; Fuller et al., 2006). The δ13C values also increase in exclusively breastfed infants, although this increase is slight (i.e. approximately 1‰: Fuller et al., 2006), and is not necessarily evident in archaeological human populations. During and after the process of weaning, δ13C and δ15N values can be used to reconstruct weaning foods and post-weaning diets. By analyzing δ13C and δ15N values from extracted collagen samples along the growth of tooth dentin, the information of the isotopic change in dietary protein during non-adulthood is obtained (e.g. Fuller et al., 2003; Eerkens et al., 2011; Beaumont et al., 2013).
Oxygen (δ18O) and carbon stable isotope ratios of tooth enamel also provide information on breastfeeding, weaning, and diet during subadulthood (Humphrey, 2014; Tsutaya and Yoneda, 2015). The δ18O value of human enamel apatite reflects δ18O values of the water sources (Luz et al., 1984; Koch, 1998). Since the δ18O values of exclusively breastfed infants seem to increase 2–3‰ (Roberts et al., 1988), stable isotope analysis of oxygen is also used for the reconstruction of weaning ages, as in the case of nitrogen (Wright and Schwarcz, 1998; Britton et al., 2015).
The sourcing for carbon in collagen is different than in apatite. While the δ13C values of bone and dentin collagen mostly reflect those of a dietary protein (Ambrose and Norr, 1993), the δ13C values of enamel apatite reflect those of total dietary carbon including carbohydrates, lipids, and protein (Krueger and Sullivan, 1984; Jim et al., 2004; Kellner and Schoeninger, 2007). The breastfeeding period and carbohydrate supplementation in subadult diet have been studied by using oxygen and carbon isotope analyses of tooth enamel (Wright and Schwarcz, 1998; Wright, 2013; Garcia et al., 2015).
By combining the evidence from multi-isotopes from several tissues, breastfeeding and weaning practices, dietary change through life, migration, and dietary contributions of specific food categories/macronutrients can be investigated further (e.g., Sealy et al., 1995; Cox and Sealy, 1997; Kellner and Schoeninger, 2007; Schroeder et al., 2009; Knudson et al., 2012; Craig et al., 2013; Lamb et al., 2014).
ST61An individual skeleton, ST61, was excavated from the Unseiji temple (Akashi, Hyogo, Japan: Figure 1). The Unseiji temple was established at 1613 and excavated during the 2003–2004, yielding approximately 50 burial remains and 40 human skeletons from the earlier Edo to Meiji (1868–1912) periods (Akashi City Board of Education, 2008, 2009; Nagaoka et al., 2013). Several kinds of burial records indicated that the buried people included family members of warriors of high social status (such as karo) belonging to the Akashi clan (Akashi City Board of Education, 2008, 2009). The individual ST61 was buried in a rounded wooden coffin inside a rectangle wooden coffin containing a burial record, parts of a religious item (juzu), and six coins (kan-ei-tsuho) of the Edo period (Akashi City Board of Education, 2009). Large amounts of rice hull were also discovered within her inner coffin (Nagaoka et al., 2013).
Location of the Unseiji temple and the other places mentioned in this paper.
The burial record showed ST61’s identity. She was a grandmother of a chief retainer (karo) in the Oda family of the Akashi clan who died in 1732 aged 77 years (Akashi City Board of Education, 2009). The analyses of the historical documents further indicated that she was born at 1655 in Ohno, Echizen (the former name of the land around the Fukui prefecture, Japan), married a male from the Oda family before 1678 (≤23 years old), probably stayed in Echizen until moving to Akashi at 1682 (27 years old), and gave birth to her children at 23, 28, and 36 years old (Saito, 1954; Tsuda, 1994; Shigematsu, 1981). Unfortunately, information of ST61’s siblings, mother, and family environment during her non-adulthood (e.g. existence of wet nurse) is not known.
The physical characteristics of the ST61 skeleton and the δ13C and δ15N values of her rib bone collagen and rice hull have already been reported by Nagaoka et al. (2013) as summarized below. Almost all parts of the skeletons were recovered except for the mandible, the right scapula, the left patella, and several vertebrae including the atlas and the axis. The lack of duplicated bones indicated that her coffin contained only one individual. Morphological traits of the cranium and pelvis suggested that ST61 was a female, while those of the pelvis suggested an age higher or equal to 50 years, which are consistent with the burial record. Sixteen (14 maxillary and 2 mandibular) teeth were recovered, one ante-mortem tooth loss (left upper permanent first molar), and no caries were found. Osteoarthritis was detected in vertebrae and almost all limb bones, and some of these showed severe symptoms.
Previous isotope analyses indicated a relatively higher δ15N value in ST61’s rib bone collagen compared with other skeletons in the Edo period, and the δ15N value of the rice hull was apparently higher for terrestrial C3 plants (Nagaoka et al., 2013). These isotopic values are given in Table 1. Nagaoka et al. (2013) interpreted the elevated δ15N values of rice hull as a result of manuring with human feces, and the major dietary protein sources of ST61 were estimated to be freshwater fish, a combination of terrestrial and marine foods, or crops showing elevated δ15N values. However, these hypotheses have not been investigated further.
| ID | Element | δ13C | δ15N | δ34S | %C | %N | %S | C/N | C/S | N/S | Yield |
|---|---|---|---|---|---|---|---|---|---|---|---|
| ST61-1 | Rib R first | −19.7 | 13.9 | — | 42.4 | 15.2 | — | 3.3 | — | — | 7.4 |
| ST61-2 | Rib L lower | −19.8 | 13.9 | 9.4 | 43.8 | 15.3 | 0.2 | 3.3 | 629 | 188 | 5.2 |
| ST61-3 | Rice hull (AAA) | −25.7 | 9.5 | 14.6 | 32.1 | 1.0 | 0.5 | 37.5 | 185 | 5 | — |
Data of carbon, nitrogen, and yield are from Nagaoka et al. (2013).
Bone collagen extracted from a left lower rib (ST61-2) using a modified Longin method (Longin, 1971; Yoneda et al., 2004), and chemically washed rice hull (ST61-3) by the acid–alcari–acid method described in Nagaoka et al. (2013) were used for sulfur isotope analysis.
Approximately 25 mg of the extracted collagen and approximately 10 mg of the washed rice hull powder with additional V2O5 were loaded into tin capsules and measured for sulfur isotopes by an elemental analyzer-isotope ratio mass spectrometry (EA-IRMS; Thermo Flash 1110 elemental analyzer, Finnigan ConFlo III interface, and Thermo Delta Plus mass spectrometer) at SI Science Co., Ltd. Elemental concentrations and isotope ratios were calibrated against the international standards NBS123, NBS127, and IAEA S-1. The analytical standard deviation (SD) was estimated to be c. 0.3‰ on the basis of standards.
Radiocarbon dating of bone collagen and rice hullCellulose was purified from rice hulls to remove organic contaminants and lignin according to the procedures proposed by Borella et al. (1998). Rice hulls were treated with a mixed solution of approximately 0.3 M sodium chlorite and 0.14% acetic acid at 70°C for 1 h, and this procedure was repeated until the rice hulls became white in color. The treated sample was then washed with pure water. The washed sample was treated with a c. 0.4 M NaOH solution at 80°C for 1 h, and this procedure was repeated three times. The treated sample was washed with pure water followed by a treatment with weak HCl (pH 2.0) to neutralize the sample. The resultant cellulose was freeze-dried.
Approximately 2.5 mg of extracted gelatin and 2.3 mg of purified cellulose were burned into carbon dioxide, purified, and finally converted into graphite with an iron powder catalyst. Radiocarbon ages were measured by using accelerator mass spectrometry by Palaeo Labo Co., Ltd. and University Museum, the University of Tokyo for gelatin and purified cellulose, respectively.
Radiocarbon ages were calibrated with the software OxCal (Bronk Ramsey, 1995) and atmospheric and marine data sets (IntCal13 and Marine13, respectively: Reimer et al., 2013). A correction of −7 years for the local marine reservoir effect (Yoneda et al., 2000) was used for back-calculation of the dietary proportion of marine foods. Although Akashi faces Seto Inland Sea, local correction values for the sea have not been reported (Figure 1). The correction values for the Pacific coast of western Japan are affected by the warm Kuroshio Current from the south-west (Shishikura et al., 2007), and values for the nearby Kii Peninsula were used (Figure 1; Yoneda et al., 2000). The dietary proportion of marine foods (p) was calculated by using the following equations:
| (1) |
| (2) |
| (3) |
In these equations, 14CST61 is the uncalibrated 14C-age of ST61 bone collagen. 14CTerrestrial and 14CMarine are the expected radiocarbon ages of ST61 based on 0% and 100% marine proportions in her diet (117.9 and 536.3 years BP, respectively). Thus, the denominator and numerator of equation (3) represent the expected shift in 14C age of ST61 based on her having a 100% marine diet and the actual shift, respectively. IntCal and Marine are the functions back-calculating conventional radiocarbon ages from calibrated ones, according to the IntCal13 and Marine13 datasets (Reimer et al., 2013), respectively. ΔR is −7 years of the local marine reservoir effect (Yoneda et al., 2000). p is the fraction of difference in the radiocarbon age of ST61 versus those with 100% marine diet. Because the above equations were not developed in the Bayesian paradigm, the uncertainty of p was not calculated. We assumed that the rib bone collagen of ST61 recorded average radiocarbon signals during the 10 years before her death (i.e. 1722–1732). These procedures are schematically indicated in Figure 2. The computations described above were performed by using R software, version 3.3.0 (R Core Team, 2016).
Schematic illustration of the radiocarbon measurements results and calibration against the atmospheric IntCal13 calibration curve (Reimer et al., 2013).
The lower left permanent canine (C) and the upper right permanent first molar (M1) were subjected to stable isotope analysis. Longitudinal half-sections were made with a dental drill and a diamond disc, and lingual and buccal halves were used for C and M1, respectively. Dentin and enamel tissues were carefully separated for further analyses.
Sequential sections of the dentin samples were obtained by using the method proposed by Eerkens et al. (2011) and Beaumont et al. (2013), and collagen was extracted using the modified Longin method (Longin, 1971; Yoneda et al., 2004). First, the longitudinal half sections of the dentin were cleaned by sandblasting and washed with a 0.2 M NaOH solution at 4°C overnight. Then, the samples were washed with pure water, and decalcified with a 0.5 M HCl solution at 4°C for several days. The HCl solution was replaced daily, and the decalcification process proceeded until the samples became rather transparent and produced a pseudomorph of the dentin structure. The decalcified dentin samples retained their original shapes and were sequentially cut into approximately 1 mm thick transverse sections with a sterile scalpel and optical loupes. The sectioned subsamples were gelatinized with weak HCl (pH 4.5) at 80°C for over 24 h, filtered using a glassfiber filter (Wattmann GF/F), and finally freeze-dried.
During the sectioning procedure, a porous structure of inner dentin was found in the three nearest subsamples from the apical root of M1, while the similar appearance and texture of their mantle dentin was obtained from another subsample (Figure 3). These porous structures are tertiary dentin formed in response to strong stimuli. Although these subsamples were also processed and measured, their isotope ratios differed somewhat as compared to other subsamples (shown below).
Photograph of the apical roots of ST61’s maxillary right M1 before chemical treatments. The porous inner dentin was found in the apical part of the buccal root.
Approximately 0.35 mg of extracted gelatin subsamples was measured without duplicates by EA-IRMS (Thermo Flash 2000 elemental analyzer, Finnigan ConFlo IV interface, and Thermo Delta V mass spectrometer) at the University Museum, the University of Tokyo. Carbon and nitrogen stable isotopic data were calibrated by laboratory alanine standards. The measurement error was estimated as a rough indication based on the alanine standards, being c. 0.1‰ for both δ13C and δ15N values.
The age at the formation of each subsample of tooth dentin serial section was extrapolated from those at the start and the end of formation of C and M1. The ages at the end of formation of apical root of C and M1 were referred from Kaneda (1957), who analyzed 671 modern individuals of Japanese school-aged subadults by using X-ray. The ages of the end of formation for mandibular C and maxillary M1 of Japanese girls were found to be 12.1 and 10.0 years, respectively (Kaneda, 1957). Unfortunately, the age at the start of formation was not reported by Kaneda (1957), and no systematic investigations were done on Japanese teeth, to the best of our knowledge. Thus, the ages at the start of dentin formation in C and M1 were referred from Hillson (1996), who compiled several estimates of age ranges for the formation of tooth dentin in modern populations. Ages of 0.375 years and 0.0 years for the start of the formation were obtained for mandibular C and maxillary M1, respectively (Hillson, 1996; see also Schour and Massler, 1940). Assuming that the rate of dentin formation was the same throughout a single tooth, the ages at formation of each section were assigned according to its total section numbers. For example, the M1 dentin takes 10 years to fully form, and 16 sections were obtained herein, with each 1 mm section representing 1/16th of 10 years. The midpoint of age ranges was used for each section.
Carbon and oxygen stable isotope analyses of tooth enamelThe separated enamel tissues were subdivided as proposed by Wright (2013) and washed according to Koch et al. (1997) to remove exogenous non-apatic mineral phase. The enamel samples were divided into occlusal and cervical ones, and the latter was divided into three transverse sections with equal section heights. Sectioned enamel samples were cleaned with sandblasting, washed with pure water, dried, and crushed into a fine powder with a mortar and pestle. The powdered subsamples were soaked in a 2% sodium hypochlorite solution at room temperature overnight, washed with pure water, soaked in a 1.0 M acetic acid buffer solution at room temperature for 8 hours, washed with pure water, and finally freeze-dried.
Approximately 0.5–0.6 mg of the freeze-dried enamel subsamples were measured by using Kiel-IRMS (Thermo Kiel IV carbonate device and Thermo MAT253 mass spectrometer) at the Department of Geochemistry, National Museum of Nature and Science, Japan. Carbon and oxygen stable isotopic data were calibrated by NBS-18 and NBS-19 international standards. The measurement error, estimated based on the standards, was lower than 0.1‰ for both δ13C and δ18O values.
The age at the formation of each subsample of tooth enamel serial section was assigned following the same method as for dentin. The ages at the start and the end of formation were referred from Reid et al. (1998), who analyzed four individuals from a medieval site in France by using a histological method. The ages at the start and the end of formation estimated from the archaeological specimens were 0.55 and 5.96 years for C, and −0.05 and 2.78 years for M1, respectively (Reid et al., 1998).
The results of sulfur isotope analysis are shown in Table 1 and Figure 4. The rib bone collagen from ST61 indicated acceptable atomic concentration, C/S, and N/S ratios based on the quality criteria proposed by Nehlich and Richards (2009: 0.15–0.35%, 600 ± 300, and 200 ± 100, respectively). Because there are no criteria for evaluation of the diagenesis of sulfur in plant materials, it is possible that the δ34S value of rice hull was affected by the diagenesis.
Results of sulfur isotope analysis on rib bone collagen and associated rice hulls samples of ST61. The marine range of δ34S values are also indicated (Nehlich et al., 2013; Nehlich, 2015).
Although marine foods and consumers dependent exclusively on marine foods typically possess δ34S values mostly ranging from +15 to +19‰ (Nehlich et al., 2013; Nehlich, 2015), the δ34S value of the bone collagen from ST61 apparently differed from those of the marine range (Figure 4). The δ34S value of rice hull was closer to the marine range as compared to the bone collagen sample (Figure 4), although it was possible that the δ34S value of rice hull was affected by diagenesis.
Radiocarbon ages and marine reservoir effectConventional and calibrated radiocarbon ages of bone collagen and rice hull samples are summarized in Table 2 and schematically indicated in Figure 2. Calibrations were done with the IntCal13 dataset (Reimer et al., 2013) and the effect of marine carbon was not considered in this stage. The calibrated radiocarbon age of the bone collagen samples represents three possible ranges as a result of the fluctuation in the atmospheric radiocarbon concentrations during the industrial revolution and nuclear testing after the 1900s (Suess, 1955; Hua et al., 2013). The calibrated radiocarbon ages of rice hull were at least 80–120 years older than the actual death age of ST61 (1732) (Table 2).
| ID | Element | 14C age (BP) | 2SD ranges of calibrated 14C age (calAD) | Probability (%) | True age (AD) |
|---|---|---|---|---|---|
| ST61-2 | Rib L lower | 190 ± 15 | 1662–1684 | 20.8 | ≤1732 |
| 1735–1805 | 53.3 | ||||
| 1933–1954 | 21.3 | ||||
| ST61-4 | Rice hull (cellulose) | 297 ± 23 | 1499–1502 | 0.4 | — |
| 1513–1600 | 67.4 | ||||
| 1616–1653 | 27.5 |
The back-calculation of the proportion of marine foods in ST61 diet yielded 17.2%. The expected difference in the conventional radiocarbon ages of ST61 with a 100% marine diet owing to the marine reservoir effect was +418.4 years, with the actual difference being +72.1 years.
Carbon and nitrogen stable isotope ratios of tooth dentin collagenResults of carbon and nitrogen isotope analysis are shown in Table 3 and Figure 5. All the dentin collagen subsamples of ST61 indicated acceptable C/N ratios and yields based on the quality criterion proposed by DeNiro (1985: 2.9–3.6) and van Klinken (1999: over 1%), respectively.
| Tooth type | ID | δ13C | δ15N | %C | %N | C/N | Yield | Age | ||
|---|---|---|---|---|---|---|---|---|---|---|
| Lower | Upper | Midpoint | ||||||||
| C (Lower left) | ST61C-1 | −19.2 | 12.9 | 41.1 | 14.9 | 3.2 | 10.3 | 0.4 | 1.0 | 0.7 |
| ST61C-2 | −18.7 | 12.5 | 41.6 | 15.2 | 3.2 | 9.9 | 1.0 | 1.5 | 1.3 | |
| ST61C-3 | −18.9 | 12.1 | 41.0 | 15.1 | 3.2 | 10.1 | 1.5 | 2.1 | 1.8 | |
| ST61C-4 | −19.0 | 12.1 | 41.2 | 15.1 | 3.2 | 10.3 | 2.1 | 2.7 | 2.4 | |
| ST61C-5 | −19.1 | 12.4 | 41.1 | 15.1 | 3.2 | 10.0 | 2.7 | 3.3 | 3.0 | |
| ST61C-6 | −19.2 | 11.9 | 41.5 | 15.2 | 3.2 | 9.7 | 3.3 | 4.1 | 3.7 | |
| ST61C-7 | −19.3 | 11.7 | 42.2 | 15.5 | 3.2 | 8.1 | 4.1 | 4.9 | 4.5 | |
| ST61C-8 | −19.3 | 11.4 | 42.0 | 15.4 | 3.2 | 8.5 | 4.9 | 5.7 | 5.3 | |
| ST61C-9 | −19.3 | 11.5 | 41.9 | 15.2 | 3.2 | 9.2 | 5.7 | 6.4 | 6.0 | |
| ST61C-10 | −19.2 | 11.5 | 42.1 | 15.3 | 3.2 | 9.1 | 6.4 | 7.2 | 6.8 | |
| ST61C-11 | −19.1 | 11.4 | 41.8 | 15.1 | 3.2 | 9.7 | 7.2 | 8.0 | 7.6 | |
| ST61C-12 | −19.1 | 11.5 | 41.4 | 15.2 | 3.2 | 9.6 | 8.0 | 8.6 | 8.3 | |
| ST61C-13 | −19.1 | 11.7 | 41.6 | 15.1 | 3.2 | 9.9 | 8.6 | 9.2 | 8.9 | |
| ST61C-14 | −19.1 | 11.6 | 41.3 | 15.1 | 3.2 | 9.0 | 9.2 | 9.8 | 9.5 | |
| ST61C-15 | −19.1 | 11.7 | 40.6 | 14.9 | 3.2 | 7.5 | 9.8 | 10.3 | 10.0 | |
| ST61C-16 | −19.3 | 11.9 | 41.5 | 15.1 | 3.2 | 9.6 | 10.3 | 10.9 | 10.6 | |
| ST61C-17 | −19.5 | 12.1 | 40.5 | 14.8 | 3.2 | 8.7 | 10.9 | 11.5 | 11.2 | |
| ST61C-18 | −19.6 | 12.6 | 41.6 | 15.0 | 3.2 | 9.1 | 11.5 | 12.1 | 11.8 | |
| M1 (Upper right) | ST61M1-1 | −18.9 | 14.5 | 41.8 | 15.1 | 3.2 | 9.5 | 0.0 | 0.6 | 0.3 |
| ST61M1-2 | −19.1 | 12.6 | 41.8 | 15.3 | 3.2 | 9.4 | 0.6 | 1.1 | 0.8 | |
| ST61M1-3 | −19.0 | 12.2 | 41.6 | 15.3 | 3.2 | 9.5 | 1.1 | 1.7 | 1.4 | |
| ST61M1-4 | −18.7 | 12.2 | 41.7 | 15.2 | 3.2 | 9.8 | 1.7 | 2.2 | 1.9 | |
| ST61M1-5 | −18.9 | 12.4 | 41.9 | 15.3 | 3.2 | 7.8 | 2.2 | 2.8 | 2.5 | |
| ST61M1-6 | −18.8 | 12.1 | 41.6 | 15.2 | 3.2 | 9.1 | 2.8 | 3.3 | 3.1 | |
| ST61M1-7 | −19.0 | 12.0 | 42.0 | 15.3 | 3.2 | 9.1 | 3.3 | 3.9 | 3.6 | |
| ST61M1-8 | −19.1 | 11.7 | 42.4 | 15.5 | 3.2 | 9.0 | 3.9 | 4.4 | 4.2 | |
| ST61M1-9 | −19.1 | 11.4 | 41.8 | 15.1 | 3.2 | 9.4 | 4.4 | 5.0 | 4.7 | |
| ST61M1-10 | −19.0 | 11.6 | 41.6 | 15.2 | 3.2 | 9.3 | 5.0 | 5.6 | 5.3 | |
| ST61M1-11 | −19.2 | 11.7 | 41.8 | 15.3 | 3.2 | 9.8 | 5.6 | 6.1 | 5.8 | |
| ST61M1-12 | −19.3 | 11.7 | 41.9 | 15.2 | 3.2 | 9.7 | 6.1 | 6.7 | 6.4 | |
| ST61M1-13 | −19.2 | 11.6 | 41.6 | 15.1 | 3.2 | 9.3 | 6.7 | 7.2 | 6.9 | |
| ST61M1-14 * | −19.3 | 12.0 | 41.8 | 15.2 | 3.2 | 9.5 | 7.2 | 7.8 | 7.5 | |
| ST61M1-15 * | −19.4 | 12.3 | 41.9 | 15.1 | 3.2 | 9.2 | 7.8 | 8.3 | 8.1 | |
| ST61M1-16 * | −19.4 | 14.1 | 39.5 | 14.2 | 3.3 | 9.3 | 8.3 | 10.0 | 9.2 | |
The assigned ages for each section are also shown. Samples with an asterisk (*) were excluded because of the porous structure of inner dentin.
Results of carbon and nitrogen stable isotope analyses on collagen extracted from the tooth dentin serial sections of ST61. The isotope ratios of rib bone collagen are also indicated (Nagaoka et al., 2013). Gray diamonds in M1 represent results from the excluded subsamples owing to their porous structure.
The three subsamples from the apical root of M1 showing a porous inner structure indicated a 1–3‰ increase in δ15N and approximately a 0.3‰ decrease in δ13C values as compared to another subsamples from C in similar ages (Table 3 and Figure 5). The δ13C and δ15N values of these three subsamples were excluded from further analyses, since these subsamples might not reflect the dietary isotopic signals in their assigned formation ages.
Although the assigned ages are not so precise because of chronological overlaps in subsamples and varying nature of tooth growth (discussed later), isotopic changes along ST61’s life course were evident in the tooth dentin serial sections. The δ13C values of the dentin serial sections increased 0.4–0.5‰ during the age of 0.5–2 years, remained relatively higher during 2–10 years, and decreased 0.2–0.4‰ during 10–12 years (Table 3 and Figure 5A). The δ13C values and patterns of change were consistent between C and M1 (Figure 5A), but those of M1 were approximately 0.2‰ higher during 2–5 years of age (Figure 5A).
The δ15N changes in dentin serial sections showed evidence of breastfeeding, weaning, and further dietary changes after the weaning process. The δ15N values of the subsamples corresponding to 0–0.6 years of age indicated a 2–3‰ increase (Table 3 and Figure 5B) as a result of breastfeeding (Fogel et al., 1989; Fuller et al., 2006). These increased δ15N values abruptly dropped until 1–1.5 years of age, which suggests that breastmilk contribution in the dietary protein became less prominent by the age of 1–1.5 years. The δ15N values slightly (c. 0.3‰) increased again around 2–3 years, decreased c. 1‰ at 3–5 years, remained relatively lower during 5–10 years, and finally increased c. 1‰ during 10–12 years of age (Table 3 and Figure 5B). The δ15N values and patterns of change were consistent between C and M1 (Figure 5B).
The δ13C and δ15N values of tooth dentin collagen after the end of the weaning process (2–10 years of age) were 0.6‰ higher and 2.1‰ lower in average than those of the rib bone collagen, respectively. Mean δ13C and δ15N values of the sectioned subsamples during 2–10 years of age were −19.2 ± 0.1‰ and 11.7 ± 0.3‰ for C; and −19.1 ± 0.2‰ and 11.8 ± 0.3‰ for M1, respectively.
Carbon and oxygen stable isotope ratios of tooth enamel apatiteResults of carbon and oxygen isotope analysis are shown in Table 4 and Figure 6. Mean δ13C and δ18O values of the three subsamples were −12.8 ± 0.5‰ and −5.6 ± 0.2‰ for C; and −13.6 ± 0.3‰ and −6.5 ± 0.1‰ for M1, respectively.
| Tooth type | ID | δ13C | δ18O | Age | ||
|---|---|---|---|---|---|---|
| Lower | Upper | Midpoint | ||||
| C (Lower left) | ST61C-a | −13.2 | −6.7 | 0.6 | 2.4 | 1.5 |
| ST61C-b | −12.9 | −6.7 | 2.4 | 4.2 | 3.3 | |
| ST61C-c | −12.3 | −6.4 | 4.2 | 6.0 | 5.1 | |
| M1 (Upper right) | ST61M1-a | −13.9 | −6.6 | −0.1 | 0.9 | 0.4 |
| ST61M1-b | −13.5 | −6.4 | 0.9 | 1.8 | 1.4 | |
| ST61M1-c | −13.3 | −6.6 | 1.8 | 2.8 | 2.3 | |
The assigned ages for each section are also shown.
Results of carbon and oxygen stable isotope analyses on a serial section of the tooth enamel apatite of ST61.
The δ13C values of enamel continuously increased at c. 0.5–5 years of age (Table 4 and Figure 6A), which suggests a δ13C increase in the total dietary carbon during this period. The δ13C values and patterns of change were consistent between C and M1 (Figure 6A).
The δ18O values of enamel remained relatively constant during c. 0.5–5 years of age (Table 4 and Figure 6B). The δ18O difference among subsamples (≤0.3‰) was relatively low (Figure 6B).
Multiple lines of results were consolidated and discussed from the viewpoint of agricultural and burial practices in the Edo period, ST61’s diets in older ages and subadulthood, and ST61’s dietary change through her life.
Agricultural practicesThere are two possible causes for the relatively higher δ15N values of the associated rice hull (Table 1) measured by Nagaoka et al. (2013). First, denitrification increases the δ15N values of nitrate (NO3−) in the soil (Mariotti et al., 1988), and plants take up and assimilate these NO3− into their tissues (Evans, 2001). Paddy rice fields generate suitable conditions for bacterial denitrification, in which NO3− is reduced to dinitrogen monoxide (N2O) and the remaining NO3− enriches 15N (Xing et al., 2002). Second, the use of animal fertilizers increases the δ15N values of crops (reviewed by Szpak, 2014). Historical documents of the Edo period indicated that manure and dried marine fish (e.g. sardine: hoshika) were used as fertilizers for crops (Ando et al. 1982; Hanley, 1997; Ehara et al., 2009; Harada, 2009). The preferential loss of 14N in the form of volatile gaseous ammonia from manure resulted in enrichment of 15N, with the δ15N values of the manure and crops increasing as a result (Bogaard et al., 2007; Kim et al., 2008). Nitrogen from fish revealed higher δ15N values than plants, and thus crops that take up and assimilate nitrogen from fish fertilizer are expected to show increased δ15N values. More information on the soil conditions and nitrous levels is needed to fully understand the δ15N results of rice hull.
The results of δ34S analysis supported the possible use of fish fertilizer, although this is not conclusive. Most plants take up sulfur from soil as a sulfate (SO42−) and from the atmosphere as sulfur dioxide (SO2) (Nehlich, 2015), and researchers revealed consistent δ34S values for soil and plant tissues (Chukhrov et al., 1980; Peterson and Fry, 1987). Since the δ34S value of the rice hull is closer to the marine range (Figure 4), the paddy field of rice would be influenced by marine sulfur. It is possible that decomposed marine fish fertilizer generated SO42−. This SO42− is taken up and assimilated by rice and subsequently translated to rice hulls, which reflected marine δ34S signals. However, this inference is not conclusive because the background δ34S values of the local freshwater around Akashi at the Edo period are unknown. If the background δ34S values are higher, the higher δ34S values of the rice hull are not necessarily the result of the contribution of marine sulfur. Unfortunately, the δ34S values of modern terrestrial fauna cannot be used as a proxy of past ecosystems because of the pollution of sulfur from modern industries (Peterson and Fry, 1987). Alternatively, it is possible that the higher δ34S values of the rice hull is simply a result of the sea spray effect (e.g., Zazzo et al., 2011) or diagenesis. Sulfur stable isotope analyses of fauna and other terrestrial plant remains and evaluation of diagenesis in plant remains are thus needed to investigate further on this topic.
Burial practicesThe calibrated radiocarbon ages of the purified cellulose from rice hulls are at least 80–120 years older than the year of death of ST61 (Table 2), which strongly suggests that the associated rice hull was intact and preserved from the time of ST61’s burial. A historical document about agriculture, Tsukurimono shiyo, written before 1831 by Kinshichiro Kuge, a farmer around Tamba (the former name of the land around the Hyogo and the Kyoto prefectures) in the Edo period, advises that rice hulls should be collected and kept because they can be utilized for several purposes (Ando et al., 1982). Considering this description, it seems possible that rice hulls collected 80–120 years ago were carefully kept and used for burial of dead individuals.
Alternatively, if the marine fish fertilizer was actually used, it might affect the radiocarbon ages of the rice hull. Because most carbon was assimilated from atmospheric CO2 via photosynthesis in terrestrial crops, the dissolved organic carbon in paddy water and soil does not affect the carbon isotopes in plant tissues. However, it is possible that the fish fertilizer decomposed and produced a high concentration of CO2 containing ‘older’ radiocarbons from the marine ecosystem, and rice took up and assimilated this ‘older’ CO2. A previous study indicated that plants growing in sites where ancient volcanic emissions of 14C-free CO2 took place showed pseudo ‘older’ radiocarbon ages (Bruns et al., 1980). However, this is only an inference, and experiments in paddy field are needed to further investigate this possibility.
Diet in older agesMulti-isotope analyses (34S and 14C) of rib bone collagen indicated that the contribution of marine foods in ST61 diet was relatively small despite its higher δ15N value. The δ34S value of the bone collagen from ST61 (+9.4‰) was well below that of the marine range (+15 to +19‰, Nehlich et al., 2013; Nehlich, 2015) and similar to the terrestrial range (−5 to +15‰, Nehlich, 2015). Back-calculation of the marine reservoir effect in the radiocarbon age indicated that the dietary contribution of marine foods was 17.2%. The higher δ15N value of the bone collagen from ST61 would result from the consumption of freshwater resources and/or fertilized crops showing increased δ15N values. Isotope analysis of fauna (e.g. Yoneda et al., 2004; Tsutaya et al., 2014a) or compound-specific nitrogen isotope analysis of bone collagen (Naito et al., 2016) are needed to further investigate the major contributor to the ST61’s relatively higher δ15N value.
Although it is possible that ST61’s relatively higher δ15N value was produced by some pathology, this effect would be minimal. Symptoms of osteoarthritis were found in several types of bones from ST61 (Nagaoka et al., 2013). A previous isotopic study indicated that bones with osteoarthritis sometimes show increased δ15N values, although this magnitude was relatively small and mostly within the normal intraindividual variation of bone collagen isotope ratios (<1‰: Olsen et al., 2014). Furthermore, bone parts without evident pathology were used herein, which seems to result in non-affected isotope ratios (Olsen et al., 2014).
Breastfeeding, weaning, and diet in subadulthoodThe δ15N values of collagen extracted from dentin serial sections rapidly decreased until 1–1.5 years of age, which suggests that the weaning process proceeded rapidly in ST61 (Figure 5), although it is possible that the assigned ages are not so precise (discussed later). The dentin δ15N values gradually decreased further until 5 years old but slightly increased again at c. 2–3 years. It is possible that partial breastfeeding continued until five years and the breast milk contribution slightly increased again, considering the relatively higher δ15N values of breast milk as compared to the adult diet (Reynard and Tuross, 2015). Although the complete cessation of breastfeeding is not clear for ST61, it is evident that breast milk was not the major protein source of ST61 after c. 1–1.5 years of age.
The pattern of rapid weaning reconstructed in ST61 was not common in the Edo period. Several historical documents of this period, such as Shoni hitsuyo sodate-gusa published in 1703, recommended that infants should be gradually weaned and breastfed until approximately 3 years of age (Yamazumi and Nakae, 1976; Sone, 2011; see also Tsutaya et al., 2014b). A previous isotopic study of subadult bone collagen indicated a gradual weaning process, with 2.1–4.1 years being the weaning end age in the Hitotsubashi population of Tokyo in the earlier Edo period (Tsutaya et al., 2014b). A gradual weaning process was also reported in another isotopic study for the Fushimi population in Kyoto during the entire Edo period (Kusaka et al., 2011). It is possible that the rapid weaning process was characteristic of higher-status peoples in the Edo period, because skeletons from the Hitotsubashi and Fushimi sites represent ordinary people (Kusaka et al., 2011; Tsutaya et al., 2014b). Alternatively, it is also possible that the rapid weaning process of ST61 was an exceptional case (e.g. due to her mother’s or wet nurse’s death). Reconstructions of the weaning ages in other higher-status individuals in the Edo period and/or studies of historical documents are needed to further investigate these hypotheses.
The δ18O values of the enamel samples did not increase as a result of the breastfeeding (Table 4 and Figure 6B). This could be because the major contribution of breast milk in ST61’s diet was limited to a short time period (c. <0.5 years after birth), as indicated in Figure 5B. The increase in δ18O values as a result of breast milk consumption was masked by seasonal fluctuations of the δ18O values or averaged out by the enamel formed during other period (see also Wright, 2013).
A small dietary contribution of higher-δ13C foods, such as C4 plants, is suggested from the δ13C results of tooth dentin and enamel samples, although the bulk δ13C values are not within the C4 range. The δ13C values of enamel serial sections continuously increased (+1.4‰) during 0.5–5 years of age (Table 4 and Figure 6A), while those of dentin showed no continuous increase during 0.5–5 years (Table 3 and Figure 5A). These results indicate that foods with relatively higher δ13C values were introduced in ST61’s diet during at least 0.5–5 years of age, and this contribution was evident in the total dietary carbon (reflected in the enamel apatite, Krueger and Sullivan, 1984; Jim et al., 2004; Kellner and Schoeninger, 2007) but not necessarily in the protein carbon (reflected in dentin collagen, Ambrose and Norr, 1993). C4 plants show higher δ13C values and usually contain larger amounts of carbon in carbohydrates than in proteins. If the diet of ST61 during 0.5–5 years of age contained a small proportion of C4 plants, the patterns of δ13C change mentioned above would be evident.
There are two issues in the sectioning methods of tooth dentin and enamel used herein. First, because the tooth dentin and enamel grow in form of stacked cones, the horizontal sections cross the growth lines of dentin and enamel, especially in the root and cement–enamel junction, respectively (Schour and Massler, 1940; Hillson, 1996; Smith, 2008; Nanci, 2013). Thus, the sectioned subsamples include dietary information from several time periods and represent moving averages. In dentin, although the sections around the root portion include several growth lines, those around the dentin horn (corresponding to infancy and earlier childhood) show relatively precise dietary signals for the assigned ages because of the flat horizontal growth lines in this zone (Schour and Massler, 1940). Second, although we hypothesized that the rates of dentin and enamel growth are constant throughout their formation period, this is not the case. Histological studies indicate that the rates of dentin and enamel growth change during the formation period (Dean et al., 1993; Dean and Scandrett 1995; Reid et al., 1998). Because of this issue, the assigned ages to the sectioned subsamples sometimes differ to a certain degree from the actual formation periods, although the relative trends of isotopic change were still precise. See Beaumont et al. (2013) and Eerkens et al. (2011) for detailed discussions on this topic.
Dietary changeThe δ13C and δ15N values of the collagen synthesized during subadulthood (tooth dentin) and older ages (rib bone) apparently differed (Figure 5), thereby suggesting a dietary change between these time periods in ST61’s life history. Previous studies indicated that there is no significant isotopic difference among adult skeletons in different ages for an Edo period population (Kusaka et al., 2011; Tsutaya et al., 2014b, 2016), which suggests that δ13C and δ15N values do not apparently differ as a result exclusively of the different stages of life history after adulthood. Because ST61 moved from Echizen to Akashi at 27 years of age (Saito, 1954; Shigematsu, 1981; Tsuda, 1994), the differences in ecological isotopic baselines and/or dietary practices could account for this difference. Considering the above-mentioned discussion, the δ15N values of rib bone collagen could have increased relatively because of the consumption of freshwater resources and/or δ15N-increased crops. However, these inferences require further data on faunal isotopes.
Furthermore, ST61 seemed to experience another dietary change just after 10 years of age: the δ13C and δ15N values of the tooth dentin collagen were found to decrease and increase, respectively (Figure 5). However, the cause of this change is not clear owing to the lack of historical information on her life history during teenage. Alternatively, it is also possible that these changes were affected by secondary dentine which deposited during older ages. Carbon and nitrogen stable isotope analysis of serial sections of later-forming teeth (e.g. permanent third molars) is valuable to further investigate ST61’s dietary change after 10 years of age (Eerkens et al., 2016).
The three subsamples from the apical root of M1 showing different isotope ratios would indicate a mixture of dietary signals from childhood and older ages. Although tooth tissues do not typically remodel after their formation and retain intact isotopic signals, low amounts of secondary dentin gradually grow on the surface of the pulp cavity and tertiary dentin is formed in response to strong stimuli to the tooth (Nanci, 2013). We suspect that the porous inner dentin in the apical portion of buccal root of M1 was a tertiary dentin, which records isotopic signals in older ages. Nevertheless, Mjör et al. (2001) reported that the apical region of modern human teeth dentin tends to show an irregular and variable structure (e.g. areas of resorption and repaired resorption, varied amounts of irregular secondary dentine, and cementum-like tissue lining the apical root canal wall). The effect of tertiary dentin would be larger in more apical parts, and thus its isotope ratios were closer to those of rib bone collagen than recordings of dietary isotopic signals in older ages (Table 3 and Figure 5).
Carbon, nitrogen, and sulfur stable isotope analyses and radiocarbon measurements of bone collagen, serial section of tooth dentin collagen, tooth enamel apatite, and associated rice hull samples from ST61 revealed the following evidence of her life history.
Multi-tissue and multi-isotope analyses can provide several kinds of information from different time windows, even from an individual skeleton. Multi-tissue and multi-isotope analyses could provide more precise information about an individual’s life history from the past.
Materials used in this study were kindly provided by descendants of ST61. Kohei Yamazaki provided assistance with carbon and nitrogen stable isotope analysis of tooth dentin. Megumi Saito-Kato and Nozomi Suzuki helped with carbon and oxygen stable isotope analysis of tooth enamel. Wataru Morita provided comments on tooth morphology and histology. The attendees of the meeting of the Advanced Core Research Centre for the History of Human Ecology in the North provided helpful comments on this study. We would like to thank all these people. We are also grateful to editors and anonymous reviewers for providing comments that improved this article. This study was supported in part by Grants-in-Aid for Scientific Research (KAKENHI: 18500769, 20370095, 24-785, 26350377, 15J00464, and 15K07241) from the Japan Society for the Promotion of Science.
