The study, published in Archaeological and Anthropological Sciences, applied a technique known as incremental dentin stable isotope analysis to individuals buried during the middle and late Dawenkou period, approximately 3500 to 2400 calibrated years BC. This is a critical moment in Chinese prehistory, when archaeologists have long observed dramatic differences in burial lavishness—some graves packed with more than twenty finely crafted jade ornaments and thin-walled black pottery cups, others little more than shallow pits with no furnishings at all. But burial goods can be misleading, and skeptics have long argued that mortuary display alone cannot prove that society was genuinely stratified. The new isotope evidence provides an independent, biological line of proof that social inequality was real, entrenched, and passed down through generations.

The method rests on a simple but powerful property of human teeth. Unlike bone, which continually remodels throughout life and averages dietary signals over decades, dentin forms in chronological layers during childhood and then stops changing. When a tooth’s dentin is demineralized in hydrochloric acid and sectioned into serial increments—the researchers cut crown dentin into 1-millimeter slices and root dentin into 2-millimeter slices—each slice corresponds to a specific window of the individual’s early life, sometimes less than a single year. Measuring the ratios of stable carbon isotopes (δ13C) and stable nitrogen isotopes (δ15N) in the collagen extracted from each slice produces a year-by-year dietary diary written into the tooth itself.

The chemistry works as follows. δ13C values distinguish between foods derived from C3 plants—trees, wild animals that browse on them, rice, and freshwater fish—and C4 plants such as foxtail and broomcorn millet, the staple crops of the Yellow River basin, along with pigs fed on millet. δ15N values track trophic level: breastfed infants sit roughly 2 to 3 per mil above their mothers because they are consuming a step up the food chain, so a declining nitrogen curve in dentin records the onset and completion of weaning. Sudden spikes of 1 to 3 per mil in δ15N, by contrast, are a well-established signature of physiological stress—malnutrition, illness, or metabolic disruption—because tissue catabolism enriches nitrogen isotopes even when diet does not change.

The analysis was performed on teeth, limb bones, and ribs from 22 individuals: 7 males, 7 females, and 8 of undetermined sex, drawn from elaborate, moderate, and simple burials in equal proportion where preservation allowed. First permanent molars were preferred because they form from the neonatal period to about age ten; canines and first incisors served as alternatives covering later childhood. Collagen quality was verified against standard criteria—collagen yields above 0.5 percent, carbon content above 13 percent, nitrogen above 4.8 percent, and atomic carbon-to-nitrogen ratios between 2.9 and 3.6—ensuring that the measured isotope values reflected ancient diets rather than burial contamination. Ages were assigned to each dentin increment using the London Atlas of Human Tooth Development and Eruption.

The results revealed extraordinary diversity in how Jiaojia children were raised. Breastfeeding was universal—the earliest dentin layers showed the elevated δ15N values characteristic of nursing in every individual—but weaning was anything but uniform. Where an inflection point in the nitrogen curve could be identified, weaning was completed between roughly 3.2 and 5.1 years of age, with a median of 3.8 years, slightly older than comparable estimates from the Yangshao and Baodun cultures. Ten of the 22 individuals permitted a weaning estimate; the remainder showed no clear transition, suggesting there was no single cultural rule about when children stopped nursing.

Sorting the isotope profiles by their shape, the team identified four distinct life-history groups. Seven individuals followed a “typical” weaning curve—a decline of about 2 to 3 per mil in δ15N accompanied by a simultaneous drop of 1.4 to 3.3 per mil in δ13C—indicating they were weaned onto C3-based foods, most likely meat porridges made from deer, fish, or shellfish, whose remains are abundant at the site. Six individuals showed instead a rise in δ13C during the weaning window, pointing to millet-based weaning foods, perhaps manured millet porridge, a practice consistent with the elevated nitrogen values observed in millet grains from contemporary sites in the region. Two individuals carried distinctly low childhood carbon values inconsistent with the local millet economy; given that one was buried with her head oriented to the south—nearly everyone at Jiaojia faced east—and that prior strontium isotope work identified some non-locals buried with unusual orientations, the researchers infer these two were immigrants who grew up in mixed millet-and-rice farming regions and adopted the local C4-dominated diet only after arriving as adults.

The most poignant group comprised seven individuals whose δ15N values fluctuated abnormally—sharp spikes or prolonged drops that signal physiological stress during childhood. Individual M98, for example, showed a rapid carbon and nitrogen shift between ages 2.7 and 3.5, likely a maladaptive weaning event, recovered, then deteriorated again at puberty, and died between 16 and 20 years old. Five of the seven stressed individuals died before reaching adulthood, while nearly everyone outside this group survived into adulthood—a stark demonstration that early-life stress casts a shadow across the entire lifespan, echoing modern epidemiological findings on the lasting consequences of childhood malnutrition.

The most consequential finding, however, came when the isotope profiles were compared against burial lavishness. Individuals interred in elaborate graves—with inner and outer coffins, jades, painted pottery, and tall-stemmed goblets—showed significantly lower δ13C and significantly higher δ15N values in their post-weaning dentin than those in moderate or simple burials: about 1.8 to 2.1 per mil lower in carbon, and about 0.7 per mil higher in nitrogen, differences that were statistically significant in Mann-Whitney U tests. This means that children from elite families consistently ate more high-protein C3 foods—deer, freshwater fish, and other premium resources—from the weaning period onward, while ordinary children remained on a millet-dominated diet. Previous bone-collagen studies had shown that Jiaojia elites ate better as adults; the new dentin data prove the disparity began in infancy and was maintained across the whole of childhood.

This continuity is the critical point. If diet differed only in adulthood, it might reflect individual achievement or occupational diet. But when the dietary signature of privilege appears in the teeth formed in the first years of life, it indicates that those children were born into wealthy households—families that controlled access to meat-rich foods and passed that advantage down. The researchers argue this constitutes direct evidence of intergenerational wealth transmission during the Dawenkou period, a phenomenon that economists and archaeologists identify as the essential precondition for enduring inequality. Elaborate burials of some elites clustered spatially in the cemetery further suggest these were family plots, consistent with the emergence of lineage-based households with heritable property such as land and livestock.

The mortuary record tells an equally complex story about how children were socially constructed. Of 215 burials excavated in the 2016–2017 seasons, 35 were subadults, and 29 of those included coffins or grave goods—investments that grew with age, implying a gradual socialization toward adult roles. Some children received extraordinary treatment: a 13-to-15-year-old, M124, was buried with up to 60 grave goods including jade rings, bracelets, and earrings. Yet others suffered the opposite fate. Two subadults were buried prone or bound beside elaborate adult graves, apparently as ritual sacrifices, and several infants were interred beneath house foundations or within rammed-earth walls, likely foundation offerings. The same society that lavished jades on its dead teenagers also placed some children in the ground beneath its houses—a stark illustration of how differently childhood roles could be constructed in the same community.

The Jiaojia findings mark the first time incremental dentin isotope analysis has been deployed to test whether dietary inequality between social classes existed in childhood in Neolithic China, and the answer is unambiguously yes. Beyond China, the study offers a template for archaeologists worldwide who seek to detect the intergenerational inheritance of status without relying solely on grave goods. It also carries a more human message: that the foods a family chooses for its infants—whether a porridge of millet or a mash of deer meat—can encode distinctions of class that persist for generations, and that the ancient children whose teeth were sliced for this study experienced childhood in ways as varied, and as consequential, as children do today. Future work combining isotopes with paleopathological stress markers and strontium-based mobility analysis promises to deepen the picture, but the outline is already clear: five thousand years ago, on the plains of the lower Yellow River, inequality was a family affair, learned at the breast and the bowl.