Reconstruction
In the early second millennium BCE, people buried at Xiaohe Cemetery occupied a river-fed oasis near Lop Nur, within one of the driest regions of the Tarim Basin. The landscape was not simply lifeless desert. Pollen and preserved plants indicate stands of reeds, cattails and Euphrates poplar beside surface water, surrounded by drought-tolerant tamarisk, ephedra and steppe vegetation. Seasonal runoff ultimately derived from mountain snow and ice made small-scale cultivation and livestock keeping possible.
When researchers analyzed dental calculus from seven individuals in Xiaohe’s earliest archaeological layers, dated approximately 2000–1700 BCE, every tested person yielded ruminant-milk proteins. Mass spectrometry identified beta-lactoglobulin, casein and alpha-lactalbumin, with diagnostic peptides attributable to cattle, sheep and goats. This is direct evidence that these individuals consumed dairy products. It does not, by itself, reveal how many calories milk supplied or prove that every community member was equally dependent upon it, but the consistency across all seven samples indicates that dairying was an established practice rather than an isolated experiment.
The proteins do not disclose whether the earliest consumers drank fresh milk, ate curds or used fermented products. Genetic analysis found no lactase-persistence alleles among the sampled Tarim individuals, making processing a plausible adaptation, although adult milk consumption without lactase persistence remains possible. Later Xiaohe burials provide firmer technological evidence: solid dairy masses dating mainly to the seventeenth through fifteenth centuries BCE contained ruminant proteins and microbial evidence consistent with kefir-like fermentation. Fermentation and straining could convert perishable milk into lower-lactose, more portable food, but the later cheese cannot automatically be projected onto every earlier meal.
Other remains place dairying within a mixed oasis economy. Cattle, sheep and goat bones or horns, wool textiles, hides, livestock hair and manure document close animal management. Wheat, broomcorn millet and occasional prepared cereal remains show that herding operated alongside crop use. Analyses of cattle dung indicate grazing or foddering on oasis plants, especially reeds and other grasses. Baskets may have served where pottery was uncommon, and a nearby Gumugou basket retained evidence of strained sour cattle milk.
Xiaohe matters because its people combined technologies with widely separated histories. Wheat came ultimately from western Asia, millet from northern China, and pastoral dairying had deep roots farther west and north, yet genomic evidence indicates that Xiaohe’s earliest population was not simply a colony of incoming steppe herders. Cultural knowledge moved even where large-scale migration did not. Modern kefir and strained fermented dairy illustrate the same broad preservation principle, but no unbroken lineage from Xiaohe to any present-day product has been demonstrated.
Historical context
Around 1900–1600 BCE, the Tarim Basin’s habitable zones were narrow riverine oases separated by immense deserts. Xiaohe belonged to an emerging Inner Asian exchange world in which Southwest Asian wheat and East Asian broomcorn millet crossed thousands of kilometres from their domestication regions. Far to the west, Mesopotamian cities used written law and organized irrigation; Egypt moved from the Middle Kingdom toward the Second Intermediate Period. In the Yellow River region, the Erlitou horizon preceded the historically documented Shang state. Xiaohe itself was a cemetery of roughly 2,500 square metres; about 170 graves had been excavated by the early twenty-first century. Its organic preservation is exceptional because aridity desiccated bodies, textiles, plants, manure and food residues that would normally disappear.
Evidence
Archaeological
StrongExcavated burials contain human remains, animal bones and horns, hides, wool textiles, manure, baskets, cereal remains and later solid dairy deposits in secure Xiaohe contexts.
Archaeobotanical
StrongIdentified bread wheat, broomcorn millet and lovegrass remains, together with pollen and other preserved plants, document cereal use and the oasis environment.
Zooarchaeological
ModerateBones, horns, hides, hair and manure support cattle and caprine management, although the complete herd structure and relative economic contribution of each species remain uncertain.
Residue chemical
StrongMass-spectrometric analysis directly detected ruminant milk proteins in dental calculus from all seven tested early Xiaohe individuals, including peptides assigned to cattle, sheep and goats.
Food identification
StrongThe dental-calculus evidence securely identifies consumption of ruminant dairy, but not the exact dairy foods or quantities consumed.
Dating
ModerateThe tested calculus samples are assigned to approximately 2000–1700 BCE using direct AMS dates for some individuals and archaeological context for others; published models place the earliest Xiaohe cemetery phase more narrowly around the nineteenth to eighteenth centuries BCE.
Preparation method
LimitedDental calculus does not distinguish fresh milk from fermented, strained or cooked dairy. Kefir-like fermentation is directly supported by later solid cheese deposits, not conclusively by the earliest calculus samples.
Geographic attribution
StrongThe cemetery is securely located near a former river channel in the Lop Nur region at approximately 40.34° N, 88.67° E.
Historical interpretation
ProbableThe combined evidence supports an established mixed agropastoral economy incorporating dairying, but claims of heavy caloric dependence cannot be quantified from protein detection alone.
Visual reconstruction
InterpretiveAny depiction of milking, fermentation, meals or settlement life is reconstructed from indirect economic evidence because the excavated context is primarily funerary and no preparation scene survives.
Sources
- 1.Fan Zhang, Chao Ning, Ashley Scott et al. (2021). The genomic origins of the Bronze Age Tarim Basin mummies. Nature 599, 256–261. doi:10.1038/s41586-021-04052-7Scientific literature
- 2.Jin-Feng Li, Idelisi Abuduresule, Francis M. Hueber et al. (2013). Buried in Sands: Environmental Analysis at the Archaeological Site of Xiaohe Cemetery, Xinjiang, China. PLOS ONE 8(7), e68957. doi:10.1371/journal.pone.0068957Scientific literature
- 3.Yimin Yang, Anna Shevchenko, Andrea Knaust, et al. (2014). Proteomics evidence for kefir dairy in Early Bronze Age China. Journal of Archaeological Science 45, 178–186. doi:10.1016/j.jas.2014.02.005Scientific literature
- 4.Yichen Liu, Bo Miao, Wenying Li et al. (2024). Bronze Age cheese reveals human-Lactobacillus interactions over evolutionary history. Cell 187(21), 5891–5900.e8. doi:10.1016/j.cell.2024.08.008Scientific literature
- 5.Ruiping Yang, Yimin Yang, Wenying Li, et al. (2014). Investigation of cereal remains at the Xiaohe Cemetery in Xinjiang, China. Journal of Archaeological Science 49, 42–47. doi:10.1016/j.jas.2014.04.020Scientific literature
- 6.Zhi-Wen Qiu, Yimin Yang, Xue Shang et al. (2014). Paleo-environment and paleo-diet inferred from Early Bronze Age cow dung at the Xiaohe Cemetery, Xinjiang, NW China. Quaternary International 349, 167–177. doi:10.1016/j.quaint.2014.03.029Scientific literature
- 7.Mingsi Xie, Anna Shevchenko, Binghua Wang, et al. (2016). Identification of a dairy product in the grass woven basket from Gumugou Cemetery (3800 BP, northwestern China). Quaternary International 426, 158–165. doi:10.1016/j.quaint.2016.04.015Scientific literature
Limitations
- moderatetitle and description
Milk proteins in all seven tested early individuals demonstrate established dairy consumption, but dental calculus cannot quantify dietary share. The terms “dairy-dependent” and “relied heavily” are stronger than the direct evidence permits.
- moderatedate
The dental-calculus cohort is reported at approximately 2000–1700 BCE, while modelled dates place Xiaohe’s earliest cemetery phase around 1884–1736 BCE. The supplied 1900–1600 BCE range is a broad site-level approximation and does not precisely delimit the individuals providing the dairy evidence.
