Human occupation of northern India spans the Toba super-eruption ~74,000 years ago

India is central to debates on when Homo sapiens arrived, what material culture signals mark their presence, how archaic populations were replaced, and how the ~74 ka Youngest Toba Tuff (YTT) eruption impacted hominins. The fossil record for this period in India is absent, but mitochondrial DNA suggests India was a stepping stone in the colonization of Australasia by modern humans. A key question is whether Homo sapiens reached India prior to YTT with non-microlithic African MSA technologies (Levallois and point technology), or around 50–60 ka with microlithic Howiesons Poort technology. Few Indian sites are securely dated to 80–50 ka, limiting tests of these hypotheses. This study reports detailed lithic assemblages and luminescence ages from the Dhaba locality on the Middle Son River, providing a sequence from ~80–40 ka that bridges earlier Middle Paleolithic/Late Acheulean sites (c. 140–>104 ka) and younger blade-based ‘Upper Palaeolithic’ assemblages, to contextualize technological change in India and assess implications for modern human dispersal.

The paper situates Dhaba within the South Asian Pleistocene record: Middle Son Valley has earlier Middle Paleolithic/Late Acheulean sites (Patpara, Nakjhar Khurd, Sihawal, Bamburi 1) dated c. 140–>104 ka, and blade-based Upper Palaeolithic technologies in Baghor formation deposits (~39 ka, though problematic). The Jwalapuram sequence in southern India shows a similar transition from Levallois technology to lamellar core reduction and microlithic industries, suggesting stepwise technological change pre- and post-YTT. Broader literature indicates: YTT dated by 40Ar/39Ar to ~74–75 ka; sparse skeletal record between Africa and South Asia shifts focus to lithics, modern DNA, and occasional symbolic/osseous artifacts. Genetic studies indicate a major out-of-Africa dispersal ~70–52 ka for all non-Africans, with possible minor earlier contributions. Fossil and archaeological evidence show earlier Homo sapiens in the Levant (200–185 ka), Arabia (~85 ka), China (<~80 ka), Southeast Asia (73–63 ka), and Australia (~65 ka), often with MSA/Middle Paleolithic technologies, supporting possible early dispersal routes across South Asia.

Excavation: Three step trenches (Dhaba 1, 2, 3) were opened on the north bank of the Son River, west of its confluence with the Rehi River. Trenches were positioned where artefacts were eroding in high density. Dhaba 1 comprised 4 lower steps and 12 upper steps (Dhaba 1a), covering 9–22 m above river level; Dhaba 2 had six steps (21–28 m); Dhaba 3 trench was 18 m long at 25–30 m above river level. Pits of 1 × 1 m were excavated in ~10 cm spits; all sediments were sieved (5 mm) and artefacts collected, weighed, bagged, and labeled. Sediments are colluvial/alluvial over Proterozoic bedrock; stratigraphy logs recorded pedogenic alteration, clays, silts, sands, gravels, carbonate nodules, rhizoliths. Artefact analysis: All artefacts were washed, catalogued in a relational database, assigned specimen numbers, weighed, measured with digital calipers, and photographed. Artefacts were classified technologically (cores, flakes, flaked pieces, retouched) and typologically where appropriate. Core measurement protocols followed Clarkson et al.; cores were 3D scanned (NextEngine). Selected artefacts were illustrated. Raw material types and dorsal scar patterning were quantified by stratum to track technological change. IRSL dating: Thirteen sediment samples were collected using opaque tubes hammered into cleaned section faces. In the lab, sand-sized K-feldspar (90–212 µm) grains were extracted via HCl/H2O2 treatments, sieving, heavy liquid separation (2.58 g/cm3), and 10% HF etching. IRSL measurements used a Risø TL-DA-20 reader with IR stimulation (875 nm) and beta irradiations from 90Sr/90Y; detection through BG-39 and 7-59 filters. Dose rates were determined from in situ gamma spectrometry (Exploranium GR-320), lab beta counting (Risø GM-25-5), cosmic-ray estimates, and internal dose from 40K and 87Rb contents. Water content corrections assumed 7 ± 2%. A multiple-elevated-temperature post-IR IRSL (MET-PIRIR) protocol was applied, measuring IRSL signals at 50–300 °C in 50 °C steps, with 320 °C preheats and 325 °C hot IR bleach between cycles. Dose response curves were fitted with single-saturating exponentials; characteristic saturation doses ranged from ~308–480 Gy depending on temperature. Performance tests included recycling ratios (~1.0 ± 0.1), recuperation (<5%), dose recovery, anomalous fading, and residual dose measurements (residuals increased with temperature, e.g., ~2 Gy at 50 °C to ~29 Gy at 300 °C; ~18 Gy at 250 °C, representing 5–10% of De). Chronological interpretations consider uncertainties and potential residual dose effects. Cryptotephra glass shard searches were conducted; six shards were identified in Dhaba 1 deposits consistent with YTT age, though possible contamination by human agency could not be excluded.

- Chronology: IRSL ages show Dhaba 1 lower unit 79.6 ± 3.2 and 78.0 ± 2.9 ka; Dhaba 1 upper unit 70.6 ± 3.9 and 65.2 ± 3.1 ka. Dhaba 2 deposited between 55.0 ± 2.7 and 37.1 ± 2.1 ka. Dhaba 3 between 55.1 ± 2.4 and 26.9 ± 3.8 ka. The sequence began accumulating just prior to the YTT eruption, with low probability of starting later (p < 0.08 or p < 0.15 assuming eruption true ages of 73.88 or 75.0 ka, respectively), and continued until near the Last Glacial Maximum. - Tephra: Six glass shards in Dhaba 1 occur in deposits bracketed by 78–71 ka, consistent with YTT presence regionally; contamination cannot be ruled out. - Lithic sequence: Three technological phases over ~55 kyr (c. 80–25 ka): • c. 80–65 ka (Dhaba 1): Dominant recurrent Levallois reduction (centripetal, bidirectional, unidirectional), producing Levallois flakes, points, blades; presence of scrapers, notches, redirecting flakes; raw materials mainly chert, mudstone, silicified limestone; red ochre present. • c. 55–47 ka (Dhaba 2 and 3 lower to middle strata): Continued dominance of Levallois technology with peak artefact deposition; Levallois absent above stratum E at Dhaba 2 dated 47.5 ± 2.0 ka. • Microlithic transition: Microlithic technology (microblade cores, microblades, backed microliths) appears ~48 ka (stratum D at Dhaba 2; stratum J at Dhaba 3), with limited overlap with Levallois (e.g., Dhaba 3 stratum J, 48.6 ± 2.7 ka; Dhaba 2 stratum E, 47.5 ± 2.0 ka). - Quantitative changes: Significant shifts in raw material proportions across strata (Pearson chi-square = 892.4; N strata = 9; N artefacts = 3512; p < 0.005) and in dorsal scar patterning of complete flakes (Pearson chi-square = 37.02; df = 9; N = 797; p < 0.005), reflecting transitions from centripetal Levallois to other core reduction strategies and microlithic production. - Broader affinities: The Dhaba lithic industry resembles African MSA, Arabian Middle Paleolithic, and the earliest Australian assemblages, suggesting attribution to Homo sapiens dispersing eastward out of Africa. Occupation persisted across the YTT event without major technological disruption until the later microlithic introduction.

The Dhaba sequence fills a critical chronological gap in the Middle Son Valley, documenting continuous human occupation and technological change from before the YTT eruption through to the microlithic transition. The persistence of MSA-like Levallois technology before and after YTT argues against a catastrophic and lasting demographic or cultural break due to the eruption in this region. The stepwise transition from Levallois reduction to single/multiplatform and lamellar core strategies, culminating in microlithic backed microblades, mirrors patterns at Jwalapuram and other Indian sites, indicating regional continuity and gradual innovation rather than abrupt replacement. Limited stratigraphic overlap of Levallois and microlithic elements indicates a transitional phase rather than an immediate replacement. In the wider dispersal context, Dhaba supports scenarios in which Homo sapiens were present in South Asia before ~65 ka, consistent with early modern human fossils and MSA/Middle Paleolithic technologies across Arabia, East and Southeast Asia, and early occupation in Australia by 65 ± 6 ka. Technological continuities along the proposed southern dispersal route, and genetic links between South Asian and Aboriginal Australian populations with evidence for Denisovan admixture, reinforce South Asia’s role as a key corridor. Dhaba thus provides an archaeological bridge linking similar technologies east and west of India and informs debates on the timing and nature of modern human dispersal into Asia and Oceania.

This study presents a well-dated, stratified archaeological sequence at Dhaba documenting long-term human occupation from ~80 to ~25 ka that spans the ~74 ka YTT eruption. The lithic record shows sustained MSA-like Levallois technology before and after YTT, with a later transition to microlithic technologies around 48 ka. These results indicate cultural continuity across YTT in the Middle Son Valley and support an early presence of Homo sapiens in South Asia prior to ~65 ka, contributing to evidence for a southern dispersal route into Australasia. Dhaba provides a key chronological and technological framework for the Indian Late Pleistocene. Future research should prioritize: expanding the number of securely dated stratified sites in India spanning 80–40 ka; integrating high-resolution tephrochronology to refine correlations with YTT; seeking direct human fossils or other diagnostic biomarkers; applying single-grain dating and complementary methods to test and refine chronologies; and conducting comparative quantitative analyses of lithic technologies across Africa–Arabia–South Asia–Australasia to further evaluate dispersal models and cultural transmission.

- Tephra attribution: Only six cryptotephra shards were recovered at Dhaba 1; while their age context is consistent with YTT, potential contamination from nearby thick YTT exposures visited by researchers cannot be excluded. - Chronological uncertainties: IRSL MET-PIRIR dating involves residual dose components (5–10% of De at 250 °C) and water-content assumptions; ages carry uncertainties that affect precise alignment with the YTT event and duration estimates of occupation pulses. - Proxy inference: Attribution to Homo sapiens is based on technological affinities; no human fossils were recovered. Cultural assignments are inferential rather than directly evidenced. - Site representativeness: Findings derive from three trenches at one locality; regional variability in technology and occupation intensity may not be fully captured. - Incomplete broader dating at comparative sites: Some Indian sequences used for comparison (e.g., Bhimbetka, Patne, certain Baghor formation ages) are not well-dated with modern methods, complicating regional syntheses.