Human population bottlenecks

Among the most striking features of human genetics is how little diversity our species possesses. Despite numbering more than eight billion individuals and occupying virtually every terrestrial habitat on Earth, Homo sapiens carries less genetic variation than a single population of chimpanzees in a West African forest.⁸ This low diversity is the fingerprint of population bottlenecks — periods in which the number of breeding individuals contracted dramatically, erasing genetic variation that could never be recovered. Understanding when, where, and how severely these bottlenecks occurred is fundamental to reconstructing human evolutionary history, because bottleneck events shape not only the total amount of variation in a population but also the geographic distribution of that variation across the globe.^{5, 6}

The evidence from genomes

The first hints that the human lineage had passed through a severe population bottleneck came from mitochondrial DNA. In their landmark 1987 paper, Rebecca Cann, Mark Stoneking, and Allan Wilson showed that all living humans trace their maternal ancestry to a single woman who lived in Africa roughly 200,000 years ago — the so-called "Mitochondrial Eve." The limited divergence among modern human mitochondrial lineages implied a small ancestral population.¹ Subsequent analyses of Y-chromosome variation told a similar story for the paternal lineage, with a coalescence time of roughly 200,000 to 300,000 years.¹⁸

Nuclear genome data provided more precise estimates of population size over time. Using the pairwise sequentially Markovian coalescent (PSMC) method, Heng Li and Richard Durbin reconstructed changes in effective population size across the human lineage from single diploid genomes. Their analysis revealed a complex demographic history: a large ancestral population of approximately 10,000 to 20,000 effective individuals before 1 million years ago, a long period of fluctuating but generally declining population size through the Middle Pleistocene, a sharp reduction associated with the out-of-Africa migration, and a subsequent exponential expansion.¹³ Critically, the PSMC profiles of African genomes showed consistently higher effective population sizes than those of non-African genomes across the entire time span, confirming that the out-of-Africa migration involved a substantial founder effect.¹³

The concept of "effective population size" (N_e) is central to understanding bottleneck estimates. N_e is not the actual census population but a theoretical quantity that describes the population size that would produce the observed level of genetic drift in an idealized population. Estimates of N_e for the ancestral human lineage have converged on approximately 10,000 over most of the Pleistocene — a figure that has remained remarkably consistent across different methods and data sources.^{2, 13} This does not mean that only 10,000 humans were alive at any given time; the census population was almost certainly larger, perhaps by a factor of two to ten, but the effective breeding population was small enough to produce the low diversity observed in modern genomes.²

The out-of-Africa bottleneck

The best-documented population bottleneck in human history is the one associated with the expansion of Homo sapiens out of Africa roughly 50,000 to 70,000 years ago. The genetic signature of this event is unmistakable: non-African populations carry a strict subset of the genetic variation found in African populations, and heterozygosity (a measure of genetic diversity) declines in a nearly linear fashion with increasing geographic distance from sub-Saharan Africa.^{5, 14, 15}

This pattern, first demonstrated by Sohini Ramachandran and colleagues in 2005, is best explained by a serial founder effect: as small groups broke away from the parent population and migrated into new territory, each successive colonizing group carried only a fraction of the genetic diversity of the group it left behind.⁵ The initial departure from Africa was the most severe of these founder events, and each subsequent expansion — into the Middle East, across Asia, into Europe, and ultimately into the Americas — compounded the loss. The result is a smooth cline of decreasing diversity from Africa to the Americas, with African populations retaining the most variation and indigenous South American populations the least.^{5, 6}

Whole-genome sequencing data from the 1000 Genomes Project and the Human Genome Diversity Project have refined the picture. Anders Bergstrom and colleagues, analyzing 929 genomes from 54 diverse populations, confirmed that sub-Saharan African populations harbor approximately 20 to 25 percent more single-nucleotide polymorphisms per genome than non-African populations, a direct reflection of the out-of-Africa bottleneck.⁷ The data also revealed that the bottleneck was not instantaneous but extended over a period of perhaps 10,000 to 20,000 years, during which the migrating population remained small enough to lose substantial diversity through genetic drift.^{7, 13}

The Toba catastrophe hypothesis

One of the most dramatic explanations proposed for human population bottlenecks invokes a catastrophic volcanic eruption. Approximately 74,000 years ago, the Toba supervolcano on the island of Sumatra produced the largest eruption of the last two million years, ejecting an estimated 2,800 cubic kilometers of material and blanketing much of South and Southeast Asia in volcanic ash.⁴ In 1998, Stanley Ambrose proposed that this eruption triggered a "volcanic winter" lasting several years, followed by a global cooling episode of approximately 1,000 years, which nearly drove Homo sapiens to extinction. According to this hypothesis, the Toba eruption reduced the global human population to as few as 3,000 to 10,000 individuals, creating the genetic bottleneck responsible for the low diversity observed in modern humans.³

The Toba catastrophe hypothesis was compelling because it offered a specific, datable mechanism for the bottleneck and appeared to coincide roughly with the timing of the out-of-Africa expansion. However, subsequent research has significantly weakened its core claims. Climate modeling by Alan Robock and colleagues showed that the volcanic winter following the Toba eruption, while severe, would have lasted years rather than decades and would not have produced the millennium-long cooling episode originally proposed.¹¹ Archaeological evidence from India tells a more nuanced story: Middle Stone Age tool assemblages found both below and above the Toba ash layer at the site of Jwalapuram show continuity rather than disruption, suggesting that at least some human populations in South Asia survived the eruption without catastrophic population loss.¹⁰

Genetic evidence has also failed to support a Toba-specific bottleneck. Analyses of mitochondrial DNA diversity found no coalescent event uniquely datable to 74,000 years ago, and PSMC-based reconstructions of population size changes show a gradual decline through the Middle Pleistocene rather than a sudden crash at the time of the eruption.^{13, 17} Christine Lane and colleagues concluded in 2013 that there is "no evidence for a human population size bottleneck corresponding with the Toba eruption," and that the genetic bottleneck in modern non-African populations is more parsimoniously explained by the out-of-Africa founder effect rather than a global population crash.¹⁷ While the Toba eruption was undoubtedly a significant environmental event, the current consensus is that it did not bring humanity to the brink of extinction.^{10, 11, 17}

An ancient Middle Pleistocene bottleneck

A more recent and controversial proposal has pushed the timing of a severe human bottleneck much further back. In 2023, Wangjie Hu and colleagues published a study in Science using a novel coalescent method called FitCoal (Fast Infinitesimal Time Coalescent), which they applied to genomic data from 3,154 modern individuals across 10 African and 40 non-African populations. Their analysis inferred an extreme population bottleneck between approximately 930,000 and 813,000 years ago — during the Early to Middle Pleistocene transition — in which the effective population size crashed to approximately 1,280 breeding individuals and remained at that depressed level for roughly 117,000 years.¹⁶

If correct, this would represent one of the most severe and prolonged demographic crises experienced by any hominin lineage. Hu and colleagues proposed that the bottleneck coincided with a period of intense climatic instability marked by increased glacial severity, prolonged droughts in Africa, and the loss of approximately 65.85 percent of genetic diversity. They further suggested that the bottleneck may have been associated with a speciation event, potentially the divergence of the Homo sapiens lineage from other archaic Homo populations.¹⁶

The study has provoked vigorous debate. Critics have questioned whether the FitCoal method can reliably distinguish between a genuine population bottleneck and other demographic scenarios — such as population structure, admixture, or changes in generation time — that can produce superficially similar coalescent patterns. The inferred bottleneck is far more severe than anything suggested by earlier PSMC analyses of the same time period, raising questions about model sensitivity.¹³ Nevertheless, the proposal has drawn attention to the Early to Middle Pleistocene transition as a potentially transformative period in human evolution, one marked by significant environmental change that may have shaped the demographic trajectory of the lineage that would eventually produce Homo sapiens.¹⁶

Founder effects and regional diversity

Beyond the major bottlenecks that affected the entire human lineage or all non-African populations, smaller founder effects have shaped genetic diversity in specific regions. The colonization of Europe, East Asia, Oceania, and the Americas each involved additional reductions in diversity as small founding groups established new populations in previously unoccupied territory.^{5, 14}

The peopling of the Americas provides a particularly clear example. Genetic data indicate that the founding population that crossed Beringia was extremely small — perhaps only a few hundred to a few thousand effective individuals — and that this population experienced an extended period of isolation (the "Beringian standstill") before expanding southward into the Americas approximately 15,000 to 20,000 years ago. The result is that indigenous American populations carry the lowest levels of genetic diversity of any continental group, consistent with multiple sequential founder effects compounding the original out-of-Africa bottleneck.^{6, 7}

Oceania tells a different story. The ancestors of modern Melanesians and Aboriginal Australians left Africa as part of an early coastal migration, arriving in Sahul (the combined landmass of Australia and New Guinea) by at least 50,000 years ago. These populations experienced the out-of-Africa bottleneck but subsequently maintained relatively large effective population sizes in their new territory. They also acquired additional genetic diversity through admixture with Denisovans, partially compensating for the diversity lost during the founding event.^{7, 9}

Significance for modern humans

The population bottlenecks that punctuate human evolutionary history have had lasting consequences for modern biology and medicine. Low overall genetic diversity means that humans are more genetically similar to one another than would be expected for a species of our geographic range and census size. The classic observation by Richard Lewontin that approximately 85 percent of human genetic variation exists within populations rather than between them is, in part, a consequence of the severe bottlenecks that eliminated between-group variation during the out-of-Africa expansion.^{5, 14}

Bottlenecks also have medical implications. When a population passes through a severe reduction, deleterious alleles that would normally be kept at low frequency by purifying selection can increase in frequency through genetic drift — a phenomenon known as the "founder effect" for disease alleles. Several genetic diseases that are unusually common in specific populations — such as Tay-Sachs disease among Ashkenazi Jews, sickle cell disease in malaria-endemic regions, and cystic fibrosis among northern Europeans — reflect the amplification of rare alleles during population bottlenecks and founder events, though some of these also involve balancing selection.⁶

The study of human population bottlenecks thus connects deep evolutionary history to present-day patterns of genetic diversity, disease susceptibility, and population structure. Each bottleneck was a brush with reduced diversity from which the species recovered through subsequent expansion, but the scars of those contractions remain legible in every human genome. The remarkably low diversity of Homo sapiens — lower than that of chimpanzees, gorillas, or orangutans — is the cumulative result of a demographic history shaped by founder events, climatic crises, and the contingent pathways of migration and colonization that carried a small African primate to every corner of the globe.^{7, 8}