Dairying Enabled Early Bronze Age Yamnaya Steppe Expansions

The study investigates the drivers of Early Bronze Age Yamnaya expansions across Eurasia. Although large-scale migrations and genetic impact of steppe populations are well documented, the economic mechanisms enabling this mobility remain debated. Prevailing models posit a highly mobile pastoral economy based on horse traction, wagon transport, and heavy reliance on meat and milk. However, direct archaeological and biomolecular evidence for routine dairying and early horse domestication on the western steppe has been scarce or equivocal. The authors aim to test whether a transition to regular dairying coincided with, and potentially facilitated, the Early Bronze Age expansions, and whether horse milk consumption provides evidence for early horse domestication in the Pontic-Caspian region.

Prior research has shown a major influx of steppe ancestry into Europe during the Late Neolithic/Early Bronze Age and links between Yamnaya and Afanasievo populations across vast distances. Archaeological reconstructions emphasize horse traction and wagon transport as elements of early mobility, but the domestication status of early horses—especially Botai horses—has been reassessed, with Botai equids identified as Equus przewalskii rather than the lineage of modern domestic horses. On the eastern steppe, evidence suggests horses were neither ridden nor milked before about 1200 BC, and horse remains were uncommon in early pastoral assemblages. Stable isotope studies on the western steppe suggest but do not confirm dairying, and palaeoproteomics—the only method that can directly detect individual dairy consumption with taxonomic resolution—has been minimally applied. Previous proteomic/lipid studies provided limited or taxonomically ambiguous evidence (e.g., an Ovis/Bos peptide) and equivocal lipid indications of horse milking at Botai.

The authors conducted palaeoproteomic analysis of dental calculus from 56 individuals spanning the Eneolithic to Late Bronze Age (ca. 4600–1700 BC) across the Pontic–Caspian and adjacent steppe regions. Samples: Eneolithic (ca. 4600–3300 BC) included 19 individuals from Murzikha 2 (6), Khvalynsk 1 and 2 (9), Ekaterinovka Mys (1), Lebyazhinka 5 (1), and Khlopkov Bugor (2). Two additional individuals were sampled from Botai. Bronze Age samples (n=35) from 20 sites in the Volga-Ural steppes were divided into Early Bronze Age Yamnaya horizon (ca. 3300–2500 BC) and Middle–Late Bronze Age transition (ca. 2500–1700 BC; Sintashta-associated). Proteins were extracted from dental calculus, and peptide spectral matches (PSMs) were identified via mass spectrometry with authentication criteria based on oral microbiome proteins indicating preservation. Milk proteins targeted included β-lactoglobulin (BLG), α-S1-casein, α-S2-casein, and α-lactalbumin. Taxonomic assignments were made from peptide sequences at different ranks (infraorder Pecora to genus/species, including Ovis, Capra, Bos, and Equus). Preservation was assessed for each calculus (48/55 with strong oral protein signals). Accelerator mass spectrometry dates and context were used to situate individuals chronologically. Data were summarized by period and site, with histograms of taxonomic specificity of dairy PSMs per individual.

• Preservation and data yield: 55/56 calculus samples produced identifiable proteins; 48/55 (87%) had strong oral protein preservation signals.
• Eneolithic (ca. 4600–3300 BC): Of 19 individuals, 11 were well-preserved; 10/11 (91%) showed no evidence of dairy consumption. One Khvalynsk individual yielded two bovine α-S1-casein peptides without BLG, rendering the dairy identification equivocal due to poorer casein preservation and limited deamidation-based authentication.
• Botai: Two individuals showed adequate preservation but no milk proteins, not supporting widespread milk consumption at the site.
• Early Bronze Age (ca. 3300–2500 BC): Dairy peptides were detected in 15/16 (94%) individuals, typically including multiple PSMs to ruminant milk proteins (BLG and caseins). Taxa identified included Ovis, Capra, and Bos; many individuals showed mixed-species dairy signals. Importantly, Equus BLG peptides indicating horse milk were identified in 2 Early Bronze Age individuals from Krivyanskiy 9 (3305–2633 cal BC).
• Middle–Late Bronze Age transition (ca. 2500–1700 BC): 15/19 individuals were positive for ruminant milk proteins (BLG, α-S1/α-S2-caseins, α-lactalbumin) with taxonomic assignments to Pecora, Ovis, and Bos; no Equus-specific milk was detected in this phase.
• Temporal shift: Results demonstrate a marked transition from minimal or absent milk consumption in the Eneolithic to ubiquitous dairying in the Early Bronze Age on the Pontic–Caspian steppe.
• Broader implications: The presence of horse milk proteins in Early Bronze Age individuals implies horse domestication in the region by this time and aligns with genomic evidence distinguishing Botai horses from the DOM2 domestic lineage. The data support the Early Bronze Age adoption of secondary animal products as a key factor in steppe expansions.

The findings directly address the hypothesis that changes in pastoral economies underpinned Yamnaya expansions. Proteomic evidence shows a rapid onset of widespread dairying at the Eneolithic–Early Bronze Age transition, coincident with archaeological signs of increased mobility (kurgans in arid plateaus, wheeled vehicles) and known demographic dispersals. This shift would have provided reliable protein, nutrients, and fluids, supporting mobile lifeways in cold, arid environments. Species-specific identifications confirm milking of sheep, goats, and cattle, consistent with mixed-herd pastoralism adapted to steppe ecologies. Detection of horse milk proteins in Early Bronze Age individuals indicates that horses were domesticated and milked, bolstering the role of horses in facilitating mobility and possibly traction, even though riding or traction is not directly evidenced by proteomics. The apparent absence of routine dairying in Eneolithic steppe groups, despite neighbouring farming communities in Europe practicing dairying, suggests a cultural frontier. Integration with stable isotope and lipid data further supports a dietary transition from riverine/forest resources to pastoral animal products. These converging lines of evidence strengthen the case for a secondary products revolution in the steppe, contributing to the success and expansion of Yamnaya pastoralists.

This study provides direct biomolecular evidence that regular dairying began on the Pontic–Caspian steppe at the onset of the Early Bronze Age, aligning with the period of Yamnaya expansions. It documents milking of sheep, goats, and cattle, and identifies the earliest known horse milk proteins, implying horse domestication in the region by the Early Bronze Age and suggesting the Pontic–Caspian steppe as a likely epicentre for the DOM2 lineage’s domestication during the late fourth/third millennium BC. Together, dairying and horse exploitation likely contributed to the mobility and adaptability that enabled steppe-wide dispersals. Future research should expand proteomic sampling across broader geographies and time frames, integrate with genomics and archaeology to refine the timeline and processes of horse domestication and dairying adoption, and investigate variability in dairy processing and consumption relative to lactase persistence dynamics.

• Limited sample sizes at certain sites and periods; only two Botai individuals were analyzed, restricting inferences about that site.
• One Eneolithic dairy signal was equivocal, relying solely on α-S1-casein without BLG, and casein degradation complicates authentication via deamidation.
• Proteomics identifies consumption rather than production practices and cannot directly evidence horse riding or traction.
• Some taxonomic assignments are limited to higher ranks (e.g., Pecora), and Equus identification does not resolve species beyond genus, though archaeological context supports horse.
• Preservation biases may affect differential detection of milk proteins (e.g., BLG vs caseins), potentially underestimating dairying in earlier periods.