Homo Sapiens: Anatomically Modern Humans (AMH)

Fig. 6. PCA of southern African–related ancestry.
(A) Map showing approximate sample locations. The lightest background color shows the desert and xeric shrublands biome. (B to D) PCA built with previously published southern African individuals (see references in table S2) that display <10% of missing data after masking the non–southern African ancestries (B), additionally including Angolan individuals (C), and other relevant modern (see references in table S2) or ancient groups (D) (16, 17). The axis labels include the eigenvalue (in parentheses) for each PC.

DISCUSSION
The Angolan Namib Desert provides an invaluable framework to examine the history and consequences of contact and admixture between different migratory waves into the wider region of southern Africa. Despite being culturally dominated by Southwest Bantu–speaking cattle herders, the area is remarkable for the presence of several impoverished groups with a peripatetic way of life, who have attracted a considerable degree of ethnographic interest (21, 27, 43, 46, 47).
Our results highlight the heterogeneity of the population landscape of the Angolan Namib by showing that, despite their high amounts of Bantu ancestry (~80%), all sampled peripatetic groups display elevated levels of an eastern African ancestry and a previously undocumented southern African–related component (KS-Namib) (Figs. 4 to 6). The co-occurrence of the two pre-Bantu ancestries among all peripatetic groups hints at a complex contact and admixture history. Because the eastern African component has also been detected in Khoe-speaking groups of the Kalahari Basin (Fig. 4), it is likely that it was introduced to southwestern Angola by the ancestors of the present-day Kwepe as part of the Kwadi branch of the Khoe-Kwadi pastoral dispersal. By contrast, KS-Namib has a more restricted distribution, being especially common in the Angolan Namib and appearing in residual amounts in other groups of southwestern Africa, including the Damara from Namibia who are historically linked to a peripatetic way of life (Fig. 5). This distribution suggests that KS-Namib was more likely associated with a resident foraging population of southwest Africa than with migrants from elsewhere.
While at present no hunter-gatherers resembling Kx’a- and Tuu-speaking groups exist in the Angolan Namib, an early account by 16th-century traveler Duarte Pacheco-Pereira states that the areas around the Kuroka River mouth were then inhabited by nomadic groups who lived on fishing and built houses from whale ribs that they covered with seaweed (48). This description is reminiscent of coastal foragers often referred to as “Strandlopers” who once lived near the coast in Namibia and South Africa but went extinct during the 19th century (49). Although historical records note them to be Khoekhoe-speaking, their culture differed from the herding groups further inland, and their origins may ultimately trace to resident hunter-gatherer groups, as suggested by the long history of maritime foraging in southern Africa (50). An ancient, prepastoral origin of southern African marine foragers was recently supported by genome-wide data from a 2241- to 1965-ka-old skeleton excavated at St. Helena Bay on the coast of South Africa (16, 51). However, the genetic profile of this individual is close to contemporary Tuu-speaking hunter-gatherers from inland areas of southern Africa and not to the KS-Namib component (Fig. 6). Hence, it is likely that only studies of ancient DNA from the southwestern coast of Angola will be able clarify the ancestral relationships between KS-Namib and extinct forager populations. While the archaeological record for the Angolan Namib is sparse, it is possible that prehistoric human remains associated with this ancestry may be recovered around shell midden deposits and vestiges of coastal settlements that have been reported in areas close to the Kuroka River mouth (52, 53).
An ancient occupation of the Namib coast is also supported by oral stories describing an encounter between the Kwadi-speaking Kwepe and resident peoples who had no fire and ate raw fish on the beach (27, 52, 54). Some anthropologists have equated these resident fishermen with the ancestors of the Kwisi and Twa, because of their present socioeconomic marginalization and historically documented association with hunting and gathering, which contrasts with the Kwepe’s higher reliance on small-scale pastoralism (27, 52, 54).
However, while our results show that present-day Twa and Kwisi can be genetically separated from the Kwepe (figs. S4 and S6), the three groups are virtually indistinguishable once the effects of genetic drift are attenuated (Fig. 3 and fig. S12). This pattern suggests that all extant peripatetic groups are equally related to different ancestries, thus challenging any attempts to establish continuity between specific modern populations and ancient foragers, beyond ethnographic considerations. The microcosmos of the Angolan Namib can therefore be considered a highly stratified polyethnic system (55) where groups with different genetic and ethnolinguistic backgrounds admixed but maintained sharp divisions based on their socioeconomic status.
This contact profile of southwestern Angola is remarkably similar to other areas of southern Africa where Khoe-Kwadi–speaking migrants encountered resident populations with different linguistic and genetic legacies. A defining characteristic of all these areas is an admixture history, which starts with the fusion between an eastern African ancestry and different resident southern African forager components, later followed by various degrees of admixture with Bantu speakers from the western and eastern streams of the Bantu migrations (19). Considering the available genetic, linguistic, and archaeological data (19, 40, 56), we hypothesize that proto–Khoe-Kwadi speakers split in the northwestern Kalahari, contrary to previous proposals that assume a southwestward movement from an intermediate homeland in northeastern Botswana (43, 57). Following their split, the Khoe-Kwadi diverged into different groups migrating into specific contact areas (figs. S1 and S22): Khoekhoe speakers moved south along the Atlantic coast, encountering !Ui-speaking groups; Kalahari Khoe speakers took an eastward route where they encountered Ju speakers in the northern Kalahari, and Taa and ǂHoan speakers in the central Kalahari; and Kwadi speakers migrated toward southwestern Angola, arriving in areas inhabited by a now extinct foraging group associated with the KS-Namib component. More recently, most Khoe-Kwadi–speaking peoples were further affected by East and West Bantu–speaking populations, adding to their diverse genetic makeups.
Together, our results show that contact areas associated with the confluence of different migratory waves can harbor the ancestry of vanished groups predating the arrival of food production in Africa. While the full diversity and geographical extension of these early foragers may ultimately be revealed by ancient DNA, detailed studies of highly admixed small-scale communities can still provide unique opportunities to probe the deep genetic structure of the continent.

Results
The effects of hominin group size and predator density on the competition for carrion between hominins and giant hyenas in two different ecological scenarios were evaluated using six simulation experiments (Table 1). Each experiment was replicated 70 times (see Methods and Supplementary Methods S1), and all the outputs of the 420 replicates (70 runs × six experiments) are shown in Supplementary Dataset 1. For the experiments, it was assumed that giant hyenas were strict solitary scavengers and hominins competed with them for the carcasses produced by large felids (Homotherium latidens, Megantereon whitei, and Panthera gombaszoegensis) in a trophic strategy of passive scavenging.


Carrion production in the Epivillafranchian

Estimated amounts of energy available in the carcasses abandoned by the three large predators were obtained based on the body weight of their preferred prey, the estimated hunting frequency, and the daily intake rate of the predators, as detailed in the “Methods” section. The values included in the experiments are listed in Table 2. These estimates are in general agreement with the values observed in recent ecosystems (Supplementary Methods S2) and confirm the assertion that the two sabre-toothed species generated far more scavengeable resources than recent and fossil pantherines.

The results for the two scenarios were similar. As expected, the final populations of both hyenas and hominins were larger in experiments with higher predator densities (Fig. 1). Indeed, the main difference between the results of both scenarios was that the available resources were insufficient to sustain a hominin population when the predator density was low and only two sabre-toothed cats were present (Fig. 1b). In the scenario of low resources, only giant hyenas survived, although at a low population density (Fig. 1b). In contrast, the presence of a third predator (P. gombaszoegensis) increased the available carrion enough to sustain the populations of both scavengers in most runs, even at a low predator density (Fig. 1a).