Introduction

Archaeological excavations of ancient tombs unveil fascinating glimpses into the rich tapestry of past civilisations, revealing intricate cultural, social and economic patterns1. This study focuses on a burial site Ju’an (居安) Cemetery (Fig. 1), Shaanxi Province, China, which could date back to the early to mid-Spring and Autumn(春秋) periods (770 BC to 541 BC). This was a period which the vassal states continuously carried out diplomacy and waged wars for the expansion of their domains.

Fig. 1: The location of Ju’an cemetery and its relative position with the surrounding states of Qin(秦), Jin(晋), Zhou(周) and Chu(楚).
figure 1

Some important cities during the Spring and Autumn Period were also marked on the map.

Full size image

Historical literature reveals that during the early Spring and Autumn Period, the Qin (秦) state presided over a territory situated to the west of Guanzhong basin in the northwest of China. Over time, their influence gradually expanded eastwards, reaching the west bank of the Yellow River, while the east bank remained primarily under the control of the state of Jin (晋)2,3. For centuries, the two states engaged in numerous wars to determine the ownership of the Hexi(河西) region. In this historical and military context, the Qin state strategically constructed Wangguan(王官) city, a well-fortified military stronghold built along a ravine and easily defensible due to its elevated position. This historical relic has been excavated in the form of the Ju’an Cemetery and city site4,5.

The research focuses on the tombs of nobles with rich burial goods including bronze tripods (Fig. 2), particularly, with human sacrifices, who might have been servants4,6. By inferring from the specifications of the funerary objects and the size of the tombs, it is speculated that they were belongs to Qin State6.

Fig. 2
figure 2

Part of the burial objects unearthed in the Ju’an cemetery (The original pictures Photo by Jinlei Ma).

Full size image

Dietary habits reflect not only nutritional choices, but also socioeconomic status and cultural traditions7,8,9,10,11. There have always been differences in staple foods between southern and northern China. The staple food in the northern region was millet12,13,14,15, and wheat was gradually accepted and spread after it was introduced into China9,15,16,17,18,19,20. Moreover, in southern China, rice remains a staple food21,22,23. Apart from crop foods, the consumption of animal-based foods is often regarded as a reflection of one’s social class. Typically, individuals with an elevated social status enjoy a greater intake of meat, whereas the general populace predominantly adheres to a vegetarian diet 11,24.

Conducting carbon and nitrogen stable isotope analysis on bone collagen and teeth, is used to understand the contribution of C4 or C3-based foods in human diets from δ13C, as well as the categories of animal protein on individual’s dietary from δ15N25,26,27,28. Oxygen isotopes (δ18O) in tooth enamel further provide environmental context through their correlation with local water sources 27,29,30,31.

Furthermore, researchers have demonstrated that dentine isotopes provide dietary information during tooth formation32. Dentine isotope analysis investigates the dietary patterns of adults during their tooth development through high-resolution sequential sampling of dentin sections aligned with growth increments32,33,34, and multiple teeth offers a comprehensive view of the individual’s dietary from birth to adolescence33,35,36, whereas enamel δ¹³C and δ18O values capture plant consumption patterns and hydrological conditions specific to enamel mineralisation periods31,37,38,39,40. Additionally, starch grain analysis of dental calculus can provide direct evidence of food intake, revealing the types of plants consumed41,42,43.

Skeletal elements differ in their temporal representation of isotopic signals: rib bone collagen is typically considered to reflect a more recent period of life compared to long bone collagen, which may average dietary inputs over a longer timeframe due to slower remodelling rates. This multi-tissue approach is frequently employed to investigate dietary changes across life stages44,45,46,47.

This research allows for an exploration of their breastfeeding and weaning, child rearing and adolescent activities. This presents a rare opportunity to unpack the shift of personal dietary. Through an interdisciplinary approach encompassing archaeological findings, historical records and scientific analyses, this research aims to delve into the dietary secrets of the aristocracy in the Spring and Autumn period.

Methods

Samples

The Ju’an cemetery contains more than 4000 tombs, and two tombs of Qin nobility have been systematically excavated, a young male tomb(M8) and a young female accompanied by two sacrificial individuals (M158)4,5. The tomb occupant lies in an east-west orientation, interred within a nested arrangement of one coffin enclosed in an outer coffin. Positioned at the head of the deceased is a headbox, replete with a multitude of burial artifacts. It is noteworthy that Tomb M158 features a niche on both the north and south walls, each housing one sacrificial person in a crouched position.

The study included all of four people from two burials, including tomb owner and sacrificed individuals and animal bones buried with them, including cows and dogs. Samples of human bones included limb bones, ribs and multiple teeth. The enamel and dentin were separated for stable isotope analysis, and a part of the tooth with dental stones was used for starch granule analysis in addition (Table 1).

Table 1 Information of samples in this study
Full size table

The extraction of enamel

The analysis method of carbon and oxygen isotopes from tooth enamel was referenced from Koch et al.27. After the tooth surface was cleaned by ultrasonic waves, 10 mg of enamel powder was taken out with a dental drill. To minimise the influence of the enamel formation time on the data35,48, samples were taken from occlusal surface and the areas adjacent to the surface of the tooth. The sample was centrifuged after being immersed in 50% NaOCl to remove impurities, and then soaked in 1 mol/L CH3COOH. After completion, it was washed to neutrality and freeze-dried using a freeze dryer. The carbon and oxygen isotope data were obtained using the VPDB standard. And then the δ18O were converted to the V-SMOV standard and subsequently used to calculate the δ18O drinking water49,50. The conversion formula is as follows:

$${{\rm{\delta }}}^{18}{O}_{x/{VSMOW}}=1.45\times \frac{{{\rm{\delta }}}^{18}{O}_{x/{VPDB}}+29.99{\rm{\backslash }}{permil}}{0.97001}-35.23$$

The extraction of dentine

The methods for preparing bone collagen samples and the extraction and pretreatment came from Richards and Hedges48, and was modified with ultrafiltration before lyophilization51. Dental samples were prepared using Beaumont’s method 235, involving surface assessment, immersion in 0.5 M HCl at 4 °C, and subsequent deacidification. The tooth was segmented with a sterile scalpel and optical loupes. The tooth was first cut in half, and the portion with half of the crown and one intact root was taken off. Sectioning started from the crown part, meanwhile, the thickness of the dental slices taken from the tooth crown was refined to 0.5 mm, ensuring a closer correlation with the distinct growth rates of teeth at different times33. And the dentine slices (~1 mm) were obtained for uniform age estimation at the root of the tooth. After heating and immersion in 0.01 M HCl, samples were filtered, freeze-dried to obtain collagen.

The stable isotope values of C and N contents of the collagen samples were measured by the laboratory of the Department of Archaeology and Anthropology, University of Chinese Academy of Sciences. The samples were measured on an IsoPrime-100 IRMS. Two-point calibration was used to calibrate the raw isotope values. Stable isotope abundances of carbon were measured and calibrated to the VPDB and AIR standards, and then converted to δ values to reflect isotopic variation, expressed in per mille. The precision for C and N was determined to ±0.2‰ of calibration standards, check standards.

The extraction of starch granules

A whole piece of dental calculus was found on the teeth of the owner of M158. This part was extracted separately and subjected to starch grain analysis. Following EDTA treatment, a microscopic examination is performed to scrutinise the morphological features of starch grains, providing a direct means of acquiring pertinent information about plants present in food42,52.

Radiocarbon dating of collagen

Two skeletons of the tomb owners M8 and M158 were sent to the Institute of Archaeology, Chinese Academy of Social Sciences for AMS Radiocarbon.

Data processing and analysis

For the purpose of this research, We have utilised Python 3.10 in conjunction with SPSS to conduct rigorous mathematical analysis, specifically employing the statistical techniques of Analysis of Kruskal‒Wallis test and independent t-tests to scrutinise the data53,54. There is pertinent to note that a p-value threshold of 0.05 was set as the criterion for determining statistical significance in all analytical procedures55.

Results

Radiocarbon dating of collagen

The results show that the tomb M8 dates back to 2500 ± 25 cal a BP, while Tomb M158 dates to 2485 ± 20 cal a BP (Table 2). It showed that these people did indeed live in the Spring and Autumn Period, which is in accordance with the conclusion derived from archaeological relics.

Table 2 Radiocarbon dating of collagen
Full size table

δ 13C and δ 15N results

High-quality collagen samples (n = 102) from all bones and tooth, with suitable C/N ratios (2.9–3.6), were obtained for analysis, excluding degraded or low %C/%N samples28,56. Due to degradation, three samples did not meet the required sample size for testing, whereas the remaining samples produced valid results, as detailed in Supplementary Table.

The average δ13C in all Ju’an people was −9.2‰, ranging from −15.48 to −7.25 (SD = 1.27, n = 95). The convex hull diagram (Fig. 3) shows that the distribution range of isotope data varies among different individuals. This indicates possible differences in their daily dietary intake. The Kruskal-Wallis results indicate that while there is no substantial difference between M8 and M158s (sacrifices of M158) (p = 0.23 > 0.05), a significant disparity arises when comparing M158 to other people (p = 0.032 < 0.05).

Fig. 3
figure 3

Carbon and nitrogen stable isotope results of human bones(B) or dentine(D).

Full size image

Additionally, the findings from the single-factor T-test54 revealed substantial changes in the food components consumed by individuals during different life phases (Table 3), with the exception of the δ¹³C value of M8, no significant changes were observed. Notably, a high standard deviation among M158 indicates potential drastic changes in their living environment, which may have significantly impacted their diet.

Table 3 Single-factor T-test for individuals in different life phases
Full size table

The average δ15N in all Ju’an people was 8.5‰, ranging from 5.33 to 12.72 (SD = 1.76, n = 95). The Kruskal‒Wallis test analysis of nitrogen isotopes also revealed differences among individuals (p < 0.001). However, unlike carbon isotopes, M8 is significantly different from M158 and M158s. The single-factor T-test of sequence nitrogen isotopes at different age stages revealed significant differences in all individuals (Table 3).

Figure 4 shows the results of stable isotopes of carbon and nitrogen for individuals at different ages. The carbon isotope curve depicted in Fig. 4A reveals that, for the majority of the time, all individuals exhibit a dietary preference towards C4 foods. However, after lactation, the decrease in the δ13C value was observed across three individuals, indicating an increase in the consumption of C3 foods. During this process, M158 experienced the longest and steepest decrease, reaching its lowest δ13C value at approximately seven years of age. At this point, a significant portion of her diet consisted of C3 foods. Furthermore, the δ13C value for M158 reaches its peak at approximately 12 years of age, followed by a subsequent downwards trend.

Fig. 4: Development of stable isotopes in tooth sequence with age.
figure 4

With age as the x value, the y value is calculated as the δ13C (A) and δ15N (B) values of individuals in each year. The reference line is the data of bone collagen in each individual limb bone to represent the long-term isotopic values before death. The colour is consistent with the scatter plot of that person. Blue is M8, Red is M158, and green is M158s.

Full size image

The nitrogen isotope curve is shown in Fig. 4B. An obvious downwards trend in the data can be observed during the infant period, indicating a transition from breastfeeding to weaning. Subsequently, the δ15N level of M8 increased and remained relatively high, while the δ15N values of the other two female individuals fluctuated at relatively low levels. However, the δ15N values of M158 show a significant upwards trend after the age of 12. Combined with the bone δ15N results, these findings suggest that her high protein intake persisted into adulthood and until shortly before her death. Unlike those of the two aristocrats, the δ15N levels of the sacrificed female did not remain relatively stable after entering childhood but fluctuated continuously, suggesting that her diet may not have been in a steady state.

δ 13C and δ 18O results of enamel

Similar δ13C values imply that during M1 formation period, people’s plant-based food sources were comparable. The isotopes of oxygen present in the tooth enamel seems different (Fig. 5). The δ18O value of M8 is much higher than the others, suggesting that his environment or food consumption(especially liquids food) is different.

Fig. 5
figure 5

Carbon and oxygen isotopes of tooth enamel.

Full size image

The M2 results represent an individual’s childhood diet and water intake. Compared with maps of the environmental distribution of surface water in China57, the oxygen isotope results of M158 are significantly greater than the local environmental base values(about −10 to −11‰, VSMOW), indicating that she should have lived in a warmer and wetter region closer to the ocean than Ju’an. Meanwhile, the result of M8 is close to the local environment.

Starch granules in dental calculus

A total of four kinds of plant starch grains were identified from the molars of M158 (Fig. 6). The results suggested that her diet may have included Oryza (n = 18), Hordeum or Triticum (n = 10), Setaria italica or Panicum miliaceum (n = 15) and Coix (n = 14). The number of starches found was similar, suggesting that a variety of foods shared a certain proportion of her diet and the sources of her diet were mixed and varied.

Fig. 6
figure 6

Starch granules of dental calculus; 5-1 and 5-2 are Oryza clumps, 5-3 and 5-4 are Hordeum or Triticum, 5-4 and 5-6 are Setaria italica or Panicum miliaceum, and 5-7 and 5-8 are Coix.

Full size image

Discussion

The integrated isotopic analysis reveals significant social stratification in dietary patterns among individuals from the Ju’an site. The results showed that there were significant differences in eating habits among different individuals in Ju’an. In terms of food sources, M8 and M158s are mainly biased towards a diet dominated by C4 foods. In contrast, M158 had a greater proportion of C3 foods. In particular, during the juvenile period, M158 had a long-term high intake of C3 foods.

This finding needs to be contextualised within broader temporal and spatial frameworks. During the Neolithic to Bronze periods, the main crop in the surrounding area was the C4 plant millet58,59. Other stable isotope results based nearby archaeological sites in the Eastern Zhou, suggest that the main food sources of individuals are also C4 foods60,61. This is close to the case of the three individuals of the male nobleman and the sacrifices peoples.

Social hierarchy manifested most conspicuously in meat consumption patterns. In terms of meat intake, the trophic level of M8 is far greater than that of M158 and M158s, indicating a good supply of meat, which is a true reflection of his aristocratic status. Taking the results from the Liujiawa site, a nearby Eastern Zhou period site, as a reference, the average stable isotope value for commoners serves as the reference value, approximately 9.3‰, while for sacrifices it is 8.3‰61. The male nobility had a significantly higher meat intake than the commoners and sacrifices (probably slaves). During the Bronze Age, vegetarian food dominated the diets of most ordinary people and ‘meat eaters’ were used to refer to people of higher status62.

However, it should be noted that M158 may have been the one with the highest social status among these individuals. She had a wealth of funerary goods, and two servants were sacrificed as burial objects. Surprisingly, her diet contained the least amount of animal protein, not only lower than that of her minions but also below the average level of nearby ordinary residents61.

In comparison to other individuals, the dietary style of M158 distinctly features a greater consumption of C3 foods coupled with a lower intake of meat. This trait appears to be even more evident during her juvenile phase than during her youth phase. The variations in dietary composition could be attributed to factors such as social status, sex, or geographical differences. A more detailed exploration of her dietary deviations will be presented later.

Infant feeding patterns provide new insights into early socialisation processes. In the infant stage, people usually need breastfeeding to provide adequate nutrition63,64. This medical concept should have formed a social consensus in the Spring and Autumn Period. The Liji(禮記) is a book of norms for the life behaviours of the aristocracy at that time. The book stipulates that infants should have multiple female attendants, one of whom is a wet nurse65.

The results of stable isotope changes in dentine sequences revealed that different individuals experienced a stage of breastfeeding and then gradually added complementary foods for weaning. The weaning process should begin before seven months, but the end time varies. M158s were weaned at 16 months old (1.3 years old), M8s were weaned at 28 months old (2.4 years old), and M158s were weaned at 35 months old (2.9 years old). The total length of lactation given in ancient texts is ~3 years66, and the weaning time of the two noble individuals is consistent with this record.

However, the breastfeeding duration of the sacrificial victim was significantly shorter than that recorded in historical accounts, suggesting the possibility that both she and her mother did not possess free status. It is plausible that the mother may have been compelled to engage in labour during the period typically reserved for breastfeeding, thereby necessitating the early weaning of her child, or that the child was forcibly separated from her mother’s care.

The emergence of gender differentiation becomes apparent during childhood dietary transitions. After lactation, individuals enter childhood. Compared with the previous stage, everyone’s intake of C3 food has increased in a variable extent. This seems to be a ubiquitous phenomenon, stable isotope analyses of other Chinese archaeological sites have also revealed that in multiple populations where adults predominantly subsisted on C4 resources, children consumed a relatively higher proportion of C3-based foods in certain periods67,68,69. Children and adults have different status attributes in the family, so children do not receive the same food supply as adults. During this period, gender disparities in food supply became apparent. In Ju’an male aristocrats probably ate more meat than women, whether they were aristocrats or servants.

These gendered disparities align with differential educational paradigms. A pronounced patriarchal culture prevailed, positioning girls in a subordinate role relative to men and subjecting them to distinct educational paths and societal expectations65,70,71,72. As men embarked on the study of literature, martial arts and moral education—venturing beyond the confines of their families and into the larger society—women underwent training in household tasks such as weaving and cooking. Consequently, their domestic and social spheres became physically segmented, mirroring a similar divergence in their dietary patterns.

The living course perspective offers dynamic insights into dietary evolution. By utilising different physical materials, including bones and teeth, we were able to trace the dietary shifts of individuals throughout their lifespan, spanning over two decades from birth until two years prior to her death33,35. This comprehensive analysis encompasses their entire life journey—from being a newborn, transitioning through childhood and adolescence, entering marriage and ultimately facing the end of their life (Fig. 7).

Fig. 7: The stable isotope change trends of individuals at different stages.
figure 7

A, B represent individual M8, while C, D represent individual M158, and distinct colour bands signify diverse life stages of individuals.

Full size image

The first stage, from birth to about three years of age, spans from infancy to early childhood, when the child transitions from breast milk to solid food. During this period, the diet of M158 shifts from a predominantly C4-based diet during breastfeeding to the introduction of more C3-based complementary foods while gradually reducing the supply of high-protein foods.

The second stage ranges from the age of three to seven, during the childhood, Meat consumption among male aristocrats returned to high levels, while that of women remains steady in a low level. Nevertheless, the proportion of C3 foods in female overall diet continues to rise. The staple food of her tends to be rice or wheat, reaching a peak at the age of seven. But the source structure of the male individual plant food begins to remain in a relatively stable state until the end.

The third phase, spanning from age seven to twelve, marks the progression from childhood to adolescence. During this period, the primary food source of the M158 evolves from being predominantly C3-based to C4-based foods, which suggests a potential increase in the proportion of millet-based foods, just like that of the M8.

The age of seven has a special significance in ancient China. It is not only the age when boys and girls officially begin to separate and move towards socialisation learning and family learning71,72 but also the age when people begin to bear legal responsibilities according to the law73. This change in social identity is also reflected in their own dietary structure. A noteworthy observation is that gender-associated dietary differences emerge significantly earlier, specifically at three years of age, as opposed to the previously documented onset at seven years old.

The fourth stage, spanning from age twelve to fifteen, signifies the transition from adolescence to adulthood. During this period, for M158 was a diet shift, with rice or wheat becoming a more prominent part of the staple diet. Simultaneously, the proportion of meat consumption increased significantly. Considering the opposing trends in carbon and nitrogen stable isotope values, the increased meat intake during this stage likely originates from animals that predominantly consume C3 plants either. At this point, M8, an individual who had consistently maintained a high meat intake habit from childhood, unexpectedly exhibited a brief interval of decline.

Fifteen is also an age with special significance in traditional Chinese culture. It marks the beginning of adulthood. Adolescent males and females are initiating their response by accepting the responsibilities bestowed upon them by society70,71,72,73,74. China used a virtual age calculation method at that time, where people were considered one year old at birth and their age increased by one every New Year. Therefore, the fifteen years of age mentioned here actually corresponds to an age range from about twelve to fourteen75.

Under the legal stipulations of that period, adult males were mandated to enlist in the military upon reaching a height of seven feet (approximately 160 cm) or the age of twenty74,76,77. Nevertheless, in actuality, the exigent demand for soldiers during wartime compelled the conscription of boys aged fifteen or even younger, while the height prerequisite was lowered to six feet74.

In contrast to the ample food provisions of aristocratic daily life, the grain during military service were collected by the state and then distributed to the soldiers74,76. This change in circumstances might lead to his dietary shift, rendering it more similar to those of ordinary civilians upon his enlistment.

At that time, dietary structure of M158 has shown significant changes, including an increase in C3 foods and a higher intake of high protein foods. In accordance with the societal norms prevailing during that era, the age of fifteen was regarded as the threshold of adulthood for a girl, and there was a societal expectation for her to promptly enter into matrimony72,78. Unless she formally separated from her natal families through marriage and moved to reside with husband, aristocratic woman was rarely permitted to leave their homes or engage in any form of societal activities or employment. Hence, her significant change in diet may be attributed to marriage, reflecting the impact of marital status on lifestyle choices and dietary preferences.

In the Spring and Autumn period, expansionist aspirations led states to engage in wars and form alliances. During this era of alliances, aristocratic marriages emerged as a pivotal strategy, alongside regular alliance activities3,79.

Because of differences in geography and traditional customs, there are distinct disparities in dietary habits between different states. People in Qin consumed more millet and preferred animal meat, while people in eastern states primarily ate rice and their diet also included some meat or fish80. Tooth enamel oxygen isotope from M2 suggest that her childhood environment was probably warmer and wetter than Ju’an, while her diet, while isotope of dentine sequences indicated that her dietary intake was predominantly composed of C3 plants with low protein content, which aligns with the typical dietary patterns observed in eastern states. Based on this, we hypothesise that she may be a foreign bride from other regions.

From 16 to 20 years of age, after native settlement, her diet clearly showed a predominance of C4 foods and started on a diet richer in meat consumption to cater more closely to the local palate. The overall dietary structure reflects the integration of a foreign woman into local customs and habits while living with her husband.

Notably, even with her attempts to assimilate, her dietary preferences clearly exhibited traces of her natal region. The findings from the starch grain examination revealed that in addition to millet, she also consumed rice, wheat, and coicis. Her meat intake, while having escalated in comparison to her earlier habits, remained notably lower than the typical consumption of the local nobility and their attendees. Her aristocratic status and ample material resources provided her with the opportunity to maintain some autonomous dietary habits.

The fifth stage, extending from age fifteen through adulthood beyond twenty, marks the transition into the adult phase. The diet structure of this stage is basically a continuation of that of the previous stage. M8 returned to his high-millet, high-protein diet. However, the proportion of C3 foods of M158 slightly decreased, as millet may play a more central role in the staple diet. Additionally, there was a discernible decrease in high-protein food consumption, which is reflective of reduced meat intake.

This analysis culminates in three principal findings regarding Eastern Zhou dietary practices: that Eastern Zhou aristocratic diets were structured by age, gender and social roles, reflecting hierarchical norms of the Spring and Autumn period. From infancy to adolescence, dietary patterns shifted in tandem with life-stage transitions. Gendered divergences emerged early: males adhered to C4-based diets (millet) with higher animal protein intake, while females consumed more C3 plants (wheat/barley), likely influenced by regional exchange or gendered food practices. Sacrificial individuals exhibited distinct, marginalised diets, reinforcing their subordinate status.

These findings underscore that food consumption served as both a biological necessity and a sociocultural practice, encoding identity and power relations. By tracing dietary trajectories across lifetimes, this study highlights the utility of isotope analysis in reconstructing ancient social dynamics, offering a model for interrogating the interplay between subsistence strategies and cultural identity in early complex societies.