Exhibit · the people already here
Two waves of men arrived. The people beneath both are in the hills, carrying the oldest lineages on the subcontinent.
Strip the politics away and the Y-chromosome record of South Asia is three layers. A population that was here from the Palaeolithic. A first arrival of highland herders from the Zagros, eight to six thousand years ago, who with that population built the Indus. A second arrival from the steppe, four to three thousand years ago, whose descendants sealed themselves at the top of a ranked order. The first layer is underneath both. It survives at highest concentration in the communities the order ranks lowest. This page is that record, lineage by lineage.
The three layers.
what the forager row shows
Every lineage, cross-referenced.
origin, peak, trough, meaning
sort by frequency, filter by lineage or region
| Lineage | Origin | Peaks in | Lowest in | What it marks |
|---|---|---|---|---|
| F*-M89 | Basal non-African; branched before G, H, I, J, K diverged | Hill foragers, 53%; Paniya, Paliyan, Irula | Brahmin-related — significantly absent | The deepest layer. Rare worldwide. The only lineage with clean population-specific clusters in Tamil Nadu — long isolation. |
| H-M69 / H1 | South Asian, ~48,000 years | Kannada-speaking tribes, 42.5%; Scheduled Castes; south and east | Northern Brahmins | The AASI signature. H1a1a-M82 is usually cited as the Romani founder lineage. No source is given for this on the page, and this museum has not located one. |
| C5-M356 | South Asian branch of C, very deep | Tribal and southern populations | Northern upper castes | Indigenous; low frequency, high antiquity. |
| R2-M124 | South Asian, 20,000+ years | Broadly across South Asia; the one lineage older in tribes than castes | — | Indigenous despite the "R". See below. |
| L1-M22 | Zagros, 20,600 years; into South Asia 8–6 kya | Balochistan and Sindh — highest anywhere; then Dravidian-speaking farmers | Northern Brahmins, hill foragers | The Indus layer. Its centre of gravity is the Indus, not the Ganges. |
| J2-M172 | West Asia, Neolithic | Indus region; moderate across India | Hill tribes | Second Indus-related marker; the Fertile Crescent farmer line proper. |
| R1a-Z93 | Steppe via Sintashta; expansion 4.5–4 kya | Gangetic and Bengali Brahmins — above 50% in some samples; Punjab, Kashmir | Hill tribes, southern non-Brahmin castes | The steppe layer. Tracks priestly rank and Indo-European speech. |
The page cross-references lineages against layers. Sorting is how a reader checks whether the three-layer reading survives the numbers.
Four questions the table raises.
answered
Why is L1 low in northern Brahmins?
Because the priestly population formed and sealed away from L1's centre. L1 arrived through Balochistan and Sindh and spread with agriculture; its density falls with distance from the Indus. The Gangetic plain, where Vedic culture crystallised after the Indus decline, was L1's periphery, not its core. When Brahmin-related jātis sealed with high R1a, whatever L1 was present was diluted by the founder effect and never replenished. Low L1 in northern Brahmins is not the absence of Indus ancestry — every South Asian carries that genome-wide. It is the absence of the Indus male line in a stratum that closed itself around a different one.
Why is R2-M124 "ambiguous"? It isn't — the name misleads.
R2 is a sister branch of R1, and R1 contains R1a (steppe) and R1b (western Europe). That shared letter makes R2 look steppe-related. It is not. R1 and R2 split more than 25,000 years ago — Upper Palaeolithic, long before any pastoralist existed. R2 then stayed in South Asia, where it is essentially confined today; R1 went north and west and only returned twenty millennia later as R1a.
R2 "travels with" R1a only in the sense that both appear across caste groups. The Tamil study found tribal R2 older than caste R2 — the one lineage where that held — which is exactly what an indigenous lineage looks like. R2 is the proof that a haplogroup's first letter tells you its phylogeny, not its history. It is as indigenous as H.
Does R1a really peak on the Ganges?
Yes, and specifically in Brahmin samples from Uttar Pradesh, Bihar and West Bengal, where frequencies above 50%, and in some studies above 60%, are reportedThree figures without their samplesThis figure, the Brahui percentages below and the R1/R2 split age are each given without a sample size or a named study; “in some studies above 60%” names no study at all. The underlying claims are not in doubt — the Brahui figures were checked here and hold, and the R1/R2 split is uncontroversially Upper Palaeolithic — but a page whose argument is that a haplogroup letter has misled people should not itself print frequencies without their denominators. — the highest concentrations of R1a-Z93 anywhere on earth, alongside Punjab and Kashmir. Not the Indus. The Gangetic plain. That is the region where the later Vedic texts were composed, where the Kuru-Pañcāla polities formed, and where the varṇa order was systematised. The lineage most associated with the steppe sits at highest density in the geography most associated with Brahmanical text.
What about Keeladi?
The Konthagai burial urns hold skeletal material. Whether any has been sequenced, and with what result, is not in the published record. The Tamil Nadu State Department has not announced a genome. The living population of the Vaigai valley carries the Tamil profile described on this page — indigenous lineages at roughly half, L1 and J2 at a quarter, R1a at an eighth — but no Sangam-era individual has been read. That gap is one of the most consequential on the subcontinent, and Keeladi's material is in exactly the tropical conditions that make it hard to fill.
A correction, and what it teaches.
the Brahui
This platform's page on the Brahui presents them as the western residue of a formerly wide Dravidian distribution, sharing 191 etymological roots with Tamil. The linguistic claim stands. A population reading of it does not.
The Brahui Y-chromosome profile is R1a and J at over 65%, with H under 1% and L under 8% — the Dravidian-associated lineages are nearly absent. Autosomally they are indistinguishable from their Balochi, Sindhi and Pathan neighbours, and high-resolution analysis fails to find any close affinity with other Dravidian speakers. A 2017 study proposes an elite-dominance model: the language was carried or imposed by a small group and survived while the population's genes converged on the region's.
The Brahui are the Tarim case and the Vedda case again, in the Dravidian family's own west. A Dravidian language on a Balochi gene pool. The 191 roots are real and the language is stranded — but what is stranded is a tongue, not a people. That distinction is the whole of historical population genetics, and this page carries the correction rather than the earlier claim.
What this settles.
what the evidence on this page supports
- The people with the most Indus in them are the ones the order ranks lowest. L1 and the AASI lineages concentrate in Balochistan, Sindh, Dravidian farmers, Scheduled Castes and hill tribes. The Vedic-associated lineage concentrates in Gangetic Brahmins. These are opposite ends of one cline.
- The claim that the Harappans were the Vedic people fails on the male line. The Harappan inheritance and the Vedic-associated inheritance have different centres of gravity a thousand kilometres apart.
- The forager communities are not a degraded branch of anyone. They are outside both arrivals, carrying lineages that predate both, at frequencies found nowhere else on the mainland.
- Their isolation is loss, not antiquity. The lineages are old; the isolation is what displacement into the hills did to them.
- And language is not ancestry. The Brahui, the Vedda and the Tarim mummies each prove it separately. Any argument that reads one from the other — in either direction — is doing what the evidence forbids.
The honest limits.
- Cited: per-group frequencies, significance tests, cluster structure and haplogroup ages (ArunKumar et al. 2012, PLOS One); L1-M22 origins and branches (Pathak et al. 2024, iScience); the Brahui Y and autosomal profile and the elite-dominance proposal (Pagani et al. 2017, and the 2025 Brahui–Oraon study in Human Population Genetics and Genomics, whose result is negative — it looked for Northern Dravidian ancestry in Brahui and found no signal of Oraon-related ancestry, placing the two on distinct branches); R1a distribution from the Underhill and Sengupta surveys; the R1/R2 split age from the Y phylogeny.
- Corrected here: the earlier presentation of the Brahui as a genetic residue of Dravidian distribution. They are a linguistic residue only.
- NOT claimed: that R1a's origin in South Asia is settled. The 2012 study argues against a single-wave introduction; Sengupta 2006 argues for deep antiquity; Underhill 2015 argues for steppe entry. The distribution — peaking in Gangetic Brahmins — is not disputed. The origin is.
- NOT claimed: that any Keeladi individual has been sequenced.
- Y-chromosome caution: every figure here is one ancestral line in a thousand. Genome-wide, every South Asian carries Indus-derived ancestry; the Y sharpens the sorting because the Y is what patrilineal endogamy sorts.
What is claimed, and how firmly
Several figures on this page name no study. The inspector shows which claims those are.