From Dmanisi at 1.85 Ma to the last Bantu frontiers, sixteen well-supported dispersals form one deep-time chain
Asked:“Plot major human population migrations, detailing in each line complete known ancestry, assotiated stratigraphy and time frames”
Sixteen major hominin and human dispersals, consolidated from 224 source rows across three scanned result sets, spanning roughly 1,850 to 0.1 ka BP (ka = thousand years before present; Ma = million years ago). Each line carries its date range, migrating population, associated stratigraphy, and every ancestry component the sources actually state — nothing inferred.
Deep-time migration chart — log time axis, oldest at left
fossil / archaeological inference
ancient-DNA / genetic evidence
contested route or date
bar length = oldest → youngest bound in ka BP · hover a bar for ancestry & stratigraphy
Findings the chart cannot say alone
The Dmanisi group alone accounts for 34 of the 120 source-level rows — the single best-corroborated event, with skulls, Mode I tools and stratum B1 at 1.85–1.77 Ma.
nature.comEarly Homo sapiens reached the Levant by 194–177 ka at Misliya Cave with full Levallois technology — over 120 thousand years before the expansion that actually peopled the world.
science.orgOnly 7 of the 16 lines carry a genetically stated ancestry component; the other 9 rest on fossils and stratigraphy alone — a gap the sources are explicit about.
pmc.ncbi.nlm.nih.govThe first Remote Oceanians of the Lapita horizon (3.5–2.7 ka) carried nearly unadmixed East Asian ancestry; Papuan-related admixture arrived only after settlement.
pmc.ncbi.nlm.nih.govThe Bantu-speaking expansion is the longest-running line on the page: begun ~4 ka at the Nigeria–Cameroon border, its genetic signature was still moving within the last few centuries.
nature.com
Route matrix, oldest to youngest
| Migration | Route | Time frame (ka BP) | Population | Stratigraphy | Source |
|---|
Evidence cards
Caveats — genes, cultures and languages are not the same thing
- “Complete known ancestry” means every component the sources state, not a family tree. Where a line reads “not genetically resolved”, no ancient-DNA component was named in any supplied row; ancestry is never inferred from a stone-tool culture, a language family, or a taxon name.
- Early Pleistocene lines are fossil-only. Dmanisi, Java and the Acheulean dispersals are known from skeletons and stratigraphy; no genomic ancestry exists for them.
- Contested windows are drawn dashed. Sahul rows range 65–50 ka; Americas rows range 37–15 ka with ~25–15 ka the broadly supported window; MIS 5 Levant/Arabia pulses were largely dead ends.
- An over-broad aggregate was rejected. One shortlist row spanning 2.6 Ma–40 ka labelled “Out of Africa” conflates the erectus, early-sapiens and main-expansion events and was split accordingly.
- Steppe→South Asia and Indo-European language links are associations, not stated ancestry: the rows name steppe pastoralist source populations and Bronze-Age horizons, and the language connection remains inferential.
- Departure vs destination: Lapita voyagers left with East Asian ancestry; Papuan-related admixture in Oceania and steppe admixture in Beaker Iberia are destination-side mixtures, kept distinct in the cards.
Method: 16 consolidated migration lines drawn from three result sets — an 80-row cross-source shortlist, 120 source-level rows from Nature/Science/PMC scans, and a 24-row article synthesis (224 rows total, scanned 2025). Dates are oldest–youngest bounds in ka BP as stated by the rows; ranges, not point estimates. Duplicate Dmanisi and Austronesian rows were synthesized into single routes with subwaves; the 2.6 Ma–40 ka “Out of Africa” aggregate and single-source minor haplogroup movements were cut for space. Ancestry fields quote only components stated in the rows; blank means not genetically resolved.