Population interconnectivity over the past 120,000 years explains distribution and diversity of Central African hunter-gatherers

The evolutionary history of African hunter-gatherers holds key insights into modern human diversity. Here, we combine ethnographic and genetic data on Central African hunter-gatherers (CAHG) to show that their current distribution and density are explained by ecology rather than by a displacement to marginal habitats due to recent farming expansions, as commonly assumed. We also estimate the range of hunter-gatherer presence across Central Africa over the past 120,000 years using paleoclimatic reconstructions, which were statistically validated by our newly compiled dataset of dated archaeological sites. Finally, we show that genomic estimates of divergence times between CAHG groups match our ecological estimates of periods favoring population splits, and that recoveries of connectivity would have facilitated subsequent gene flow. Our results reveal that CAHG stem from a deep history of partially connected populations. This form of sociality allowed the coexistence of relatively large effective population sizes and local differentiation, with important implications for the evolution of genetic and cultural diversity in Homo sapiens.

The evolutionary history of African hunter-gatherers may hold key insights into patterns and processes behind the evolution of modern human diversity. Recent genomic studies have revealed that these populations represent the oldest and most diverse human genetic lineages and have been genetically differentiated from one another since the origin of humans (1–3) (SI Appendix, Table S1). Therefore, a first question is whether their current ecological niches were also characteristic of early Homo sapiens populations. However, genetic data alone can neither determine the geographic distribution of hunter-gatherers in the past nor demonstrate a deep history of adaptation of hunter-gatherers to their current environments. In fact, various studies have proposed that farming expansions within the past 5,000 years (in particular by the ancestors of Bantu speakers) would have only recently displaced hunter-gatherers to marginalized regions less favorable to agriculture (such as rainforests and deserts) (4–7).For example, the central part of Africa, between latitudes 5°N and 5°S currently is inhabited by ∼20 scattered hunter-gatherer ethnic groups (8). These Central African hunter-gatherers (CAHG) form a genetic clade thought to have diverged from other African populations as far back as 120,000 to 200,000 years ago (2, 9). The lack of any major linguistic specificity between them is often implied to reflect extensive contacts with surrounding farmer populations (8, 10), and seen as evidence of recent displacement into marginal forest environments by expanding farming populations. However, anthropologists have remarked on the huge variability in lifestyle, habitat, techniques, and tools between CAHG (11), suggestive of long-term cultural diversification and adaptation to forest environments. Research on the drivers of demography and adaptation of CAHG populations remains extremely limited, which can be partially attributed to the lack of archaeological and osteological data resulting from a rapid disintegration of fossil remains in the rainforest’s acidic soils, in addition to social instability in the region (12). Therefore, we are still left with crucial questions regarding the time depth of occupation of Central Africa by hunter-gatherers, the breadth of the niche exploited by earlier populations in the region, and variations in levels of interconnectivity at different points in time.To address those questions, we first compiled ethnographic data on the distribution of 749 camps from 11 hunter-gatherer groups extending from West to East Central Africa. We used them as inputs for environmental niche models (ENMs) to determine the relative influence of several bioclimatic and ecological factors, as well as the presence of farming populations, on the distribution and abundance of CAHG (13, 14). Then, we used high-resolution paleoclimatic reconstructions and topographic maps to make continuous predictions about where CAHG could have lived over the past 120,000 years and the potential extension of their interaction networks. Next, we compiled all reliably dated archaeological assemblages ascribed to hunter-gatherer groups in the Congo Basin (n = 168) and confirmed the model’s ability to predict the location and date of the sites. We further contextualized genomic estimates of population divergences with changes in population densities and interpopulation connectivity predicted by our model. Last, we complemented these analyses with a detailed assessment of present and historical gene flow between nine CAHG populations (n = 265 individuals), which we used to assess recent interactions between previously diverged CAHG populations, after farming expansions. Our study therefore provides a causal link between past environmental changes and human population dynamics over evolutionary time, by predicting where and when populations across Central Africa could have exchanged genetic and/or cultural information throughout their evolutionary history.