Why we are not all multiregionalists now Chris Stringer Department of Earth Sciences, The Natural History Museum, London, SW7 5BD, UK
Recent revelations that human genomes contain DNA introgressed through interbreeding with archaic populations outside of Africa have led to reassessments of models for the origins of our species. The fact that small portions of the DNA of recent Homo sapiens derive from ancient populations in more than one region of the world makes our origins ‘multiregional’, but does that mean that the multiregional model of modern human origins has been proved correct? The extent of archaic assimilation in living humans remains modest, and fossil evidence outside of Africa shows little sign of the long-term morphological continuity through to recent humans expected from the multiregional model. Thus, rather than multiregionalism, a recent African origin (RAO) model for modern humans is still supported by the data. Multiregional origins? Recent revelations that human genomes contain DNA introgressed through interbreeding with archaic populations outside of Africa have led to reassessments of models for the origins of our species. The fact that small portions of the DNA of recent Homo sapiens derive from ancient populations in more than one region of the world makes our origins ‘multiregional’, but does that mean that the multiregional model of modern human origins has been proved correct? Is it true, as one well-known blogger put it, that ‘we are all multiregionalists now’ ? Should we, according to another commentator, stop talking about evolutionary trees and replace them with ‘braids’ ? More remarkably, were RAO models, popularly known as Out of Africa models, founded on a hoax ? Models for modern human origins First, it is necessary to revisit what the RAO and multiregional models stipulate, although this is not straightforward, because scientists presenting these models have modified their views in the face of new data, and have shifted positions. Accepting new data and modifying views are commendable in scientists, of course, but if shifts of position are not made explicit, confusion will inevitably Corresponding author: Stringer, C. ([email protected]
). Keywords: Homo sapiens; human evolution; human origins; modern humans; Africa; multiregionalism; hybridisation; ancient DNA. 0169-5347/$ – see front matter ß 2014 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.tree.2014.03.001
Trends in Ecology & Evolution, May 2014, Vol. 29, No. 5
follow. In addition, if the shifts are significant, the models might converge on each other, causing further confusion . Therefore, I next present models as they were during the 1990s, as summarised by Aiello  and Stringer  (Figure 1): (i) the RAO model argues that modern humans first arose in Africa approximately 100 000 years ago and spread from there throughout the world. Indigenous premodern populations in other areas of the world were replaced by the migrating populations, with little, if any, hybridisation between the groups (Figure 1A); (ii) the RAO and hybridisation model is similar to the above, but allows for a greater extent of hybridisation between the migrating population and the indigenous premodern populations (Figure 1B); (iii) the assimilation model also accepts an African origin for modern humans. However, it differs from the previous models in denying replacement, or population migration, as a major process in the appearance of modern humans. Instead, this model emphasises the importance of gene flow, admixture, changing selection pressures, and resulting directional morphological change (Figure 1C); and (iv) multiregionalism differs from the previous three in denying a recent African origin for modern humans. It emphasises the role of both genetic continuity over time and gene flow between contemporaneous populations in arguing that modern humans arose not only in Africa, but also in Europe and Asia from their Middle Pleistocene forebears (Figure 1D). To give a further perspective on these models, I refer the reader to Figure 2, which compares the models in terms of the extent of African versus nonAfrican genetic contributions to present-day humans worldwide. On the left of Figure 2 is absolute RAO (sometimes dubbed by its opponents the ‘Eve Theory’, given that one version derived from mitochondrial DNA (mtDNA) reconstructions of an ancestral ‘African Eve’), with the total replacement of nonAfrican genes. At the other extreme of Figure 2 are models that give Africa no role in the evolution of modern humans (for example, ones that have suggested a West Asian or southeast Asian centre of origin; see ). In between the polar opposites is a ‘mostly out of Africa’ model in the form of Bra¨uer’s  RAO + hybridisation, and Smith  and Trinkaus’s  assimilation model, whereas towards the right is classic multiregionalism (i.e., the version generally espoused from 1984 to 2003, which denied Africa the major
Trends in Ecology & Evolution May 2014, Vol. 29, No. 5
TRENDS in Ecology & Evolution
Figure 2. A spectrum showing the percentage of recent African genetic contribution to living humans under various evolutionary scenarios from 0% recent African origin (RAO) under nonAfrican origin models to 100% under RAO with complete replacement (see ).
TRENDS in Ecology & Evolution
Figure 1. Evolutionary models of modern human origins. (A) Recent African origin; (B) recent African origin and hybridisation; (C) assimilation; and (D) multiregional evolution [4,5].
role in modern human origins [10,11]). Depending on the extent of nonAfrican archaic gene flow envisaged (less than, or greater than, 50%), the centrally positioned assimilation model could approach either RAO + hybridisation or classic multiregionalism. This is an important consideration in evaluating the veracity of the assimilation model, which has tended to argue that modern human features were assimilated into indigenous populations outside of Africa, rather than that early modern populations assimilated elements of the indigenous populations, as argued by RAO + hybridisation. An historical perspective on the models As recently as 1970, no palaeoanthropologist of whom I am aware held the view that Africa was the evolutionary home of modern humans (as distinct from more ancient species, such as Homo habilis or Homo erectus). The region was considered marginal or unimportant, with the pendulum of scientific opinion strongly swung toward nonAfrican, multiregional, or Neanderthal phase models (the latter postulated that human evolution everywhere had passed through a Neanderthal-like stage) . However, during
the course of the next two decades, workers such as Howells , Clark , Beaumont et al. , Hublin and Tillier , Stringer and Day , and Bra¨uer  began to focus on Africa as the possible homeland of modern humans. A convoluted African origin model was also advanced by Protsch , but given that this relied on dubious self-generated data to argue that modern humans had originated first, and had then given rise to archaic humans, it had little influence on the debate (contra Bednarik ). By the early 1990s, the pendulum was moving in favour of RAO because fossil evidence began to be increasingly reinforced by the clear signals of mtDNA and Y-DNA in recent human samples. By the late 1990s, the pendulum had swung even further toward a pure RAO, with growing fossil, archaeological, and genetic data supporting this model, including the distinctiveness of the first Neanderthal mtDNA sequences recovered from 1997 onwards . However, large amounts of autosomal DNA have now been recovered from Eurasian Neanderthals and from fragmentary fossils found in Denisova Cave, Siberia, revealing another ancient human population. Moreover, traces of this nonAfrican DNA, attributed to ancient interbreeding events, have been found in recent human genomes, indicating that they are not purely of recent African origin. These revelations have halted and reversed the pendulum swing, away from an absolute RAO. I would say that we are now looking at a version of RAO that most closely resembles Bra¨uer’s early formulation (out of Africa + hybridisation), or perhaps a version of the assimilation model of Smith and Trinkaus, with a strong African predominance. If the evidence for archaic assimilation in living humans remains modest and is restricted to Africa and to the dispersal phase of modern humans from Africa, constituting less than 10% of our genome, I think ‘mostly out of Africa’ is the appropriate designation and, for me, that is still RAO. Human species in the fossil record However, given that interbreeding did occur between modern and archaic humans, out of Africa and perhaps within Africa too [19,20], does this mean that we should abandon the different species names and lump the fossils of the past million years or more as H. sapiens, as multiregionalists have suggested? If hybridisation events between the various lineages prove to have been widespread and significant in both time and space, we might have to do that, but I do not think that point has been reached yet. Personally, I have never equated the use of a separate species designation for Neanderthals (morphological species concept) with complete reproductive isolation from H. sapiens (biological species concept) and, in my view, there are still good scientific reasons to give populations that had long and (largely) separate evolutionary histories different names, 249
Opinion species or otherwise. We can measure the amount of morphological variation in closely related primate species today, and compare this with the differences between, say, Neanderthals and recent humans (see, for example, [21,22]). Such comparisons demonstrate that the latter two are distinct enough to be classed as different species, regardless of whether they meet the biological species criterion of no interbreeding (a standard that numerous recognised primate species today do not achieve). If we nevertheless proceed to merge even Neanderthals and modern humans as a species, we end up with a H. sapiens that is simultaneously characterised by a high and rounded skull, and a longer and lower skull; by no continuous brow ridge, and a strong continuous brow ridge; by a well-developed chin even in infants, and minimal chin development; by no suprainiac fossa in adults, and a suprainiac fossa throughout ontogeny; by semicircular canals of modern shape, and semicircular canals of Neanderthal shape; by a narrow pelvis with a short, thick superior pubic ramus, and a wide pelvis with a long, thin superior pubic ramus, and so on. The disparate nature of H. sapiens would become even more extreme if we start to add in the features of species such as Homo heidelbergensis, antecessor, and erectus. Similar arguments have recently been made in connection with the proposal to widen greatly the diagnosis of H. erectus, based on new fossil discoveries and analyses from the site of Dmanisi (compare  and ). What the multiregional model really meant However, if our modern genes do come from more than one region, why has multiregionalism not been proved correct, as its supporters are now asserting? At this point, it is worth reminding ourselves what classic multiregionalism actually proposed. Here is a quotation from a paper written in 1994 by Milford Wolpoff and four other prominent advocates of the model at that time  (note that one of these authors was Smith, who had by then also proposed an early version of the assimilation model ): The evolutionary patterns of three different regions show that the earliest ‘modern’ humans are not Africans and do not have the complex of features that characterize the Africans of that time or any other... There is no evidence of specific admixture with Africans at any time, let alone replacement by them... There is indisputable evidence for the continuity of distinct unique combinations of skeletal features in different regions, connecting the earliest human populations with recent and living peoples. To reinforce that last point, we can add this caption from Thorne and Wolpoff  for archaic and modern human skulls from China spanning approximately 500 000 years: Series of Chinese skulls shows continuity in form without evidence of a replacement by African characteristics. From left to right, the male skulls are from the Zhoukoudian Lower Cave (Middle Pleistocene period), Dali site (early Upper Pleistocene period) and Zhoukoudian Upper Cave (late Upper Pleistocene). 250
Trends in Ecology & Evolution May 2014, Vol. 29, No. 5
Thus, multiregionalism gave Africa no special place in our evolution and claimed specific anatomical connections down to details of the skull and teeth between H. erectus fossils that were more than 500 000-years old and humans living in the same areas today. Despite recent revelations that support its inference of interregional gene flow (although so far at a relatively low level; see below), the multiregional model has otherwise been rather comprehensively falsified, as has the absolute recent African origin (‘Eve theory’) model (given that we know that modern humans are not derived purely from African ancestors). Trees and braids Finlayson  has argued that the extent of reticulation now apparent from genomic data is so great that we should abandon specific distinctions, and envisage evolutionary processes as braided rivulets rather than as branches. However, the view that we should switch to braids and abandon phylogenetic trees is also mistaken, because we need both concepts to understand recent human evolution. As Hawks  recognised: I admit that the braided stream is not a perfect analogy. Diverging rivulets within a valley almost always come together again, forming a complicated network as they form sandbars and islets. None of them flow into a cul-de-sac. Some human populations of the past did become extinct, they did not inexorably flow back into the mainstream of our evolutionary history. Some of them may have flowed back into the mainstream only through very small channels of genetic exchange. When we go far enough back, some populations really did branch off into their own direction.
Comparing 1000 recent human genomes reveals approximately 100 changes in amino acid coding shared by living H. sapiens but absent from the Neanderthal and Denisovan genomes , including ones that are known to influence the development and maintenance of brain cells. Over 3000 further genetic mutations that might have affected the behavioural or morphological evolution of modern humans are also uniquely fixed in our species. Such distinctions are likely to have accumulated during the several hundred millennia that the lineages of H. sapiens and those of Neanderthals and Denisovans were biogeographically separated (tree concept), and evidently did not diffuse into the genomes of Neanderthals and Denisovans known so far, despite any braids of gene flow that might have existed. This brings us on to a fundamental issue: does the relatively low prevalence of Neanderthal and Denisovan genes in H. sapiens reflect the rarity of ancient hybridisation opportunities, or their lack of viability ? Whereas multiregional and assimilation models often imply unconstrained interbreeding between ancient human groups whenever they had the opportunity, the reality is that many factors (demographic, cultural, and biological) could have militated against reproductive success, and we are now learning more about these constraints, including the fact that exchanges between genetically distant populations can have costs as well as
Opinion benefits [27,28]. New analyses suggest that, when modern humans and Neanderthals met and mixed, they were at the edge of biological incompatibility, such that there was reduced male fertility and rapid natural selection to remove the Neanderthal-derived variants that caused this sterility . Further evidence of the swift elimination of much of the introgressed DNA comes from the partial genome of the Tianyuan skeleton from China, dated to approximately 40 000 years ago, because this shows a no greater component of Neanderthal-derived DNA than recent Asian samples . RAO is still the most appropriate model The big picture is that we are predominantly of recent African origin, and RAO is not just about the sources of our shared modern morphology and most of our genes; it is also about the genesis of our shared patterns of behaviour. Inferred behavioural gaps between Neanderthals and modern humans have certainly narrowed from recent research, but in my view they have not disappeared. I think that the pre-eminence of Africa in the story of modern human origins was primarily a question of its larger geographic and human population size, which gave greater opportunities for morphological and behavioural variations, and for innovations to develop and be conserved, rather than the result of a special evolutionary pathway. By contrast, genomic data suggest that the lineages of the Neanderthals and Denisovans had much greater demographic attrition , perhaps related to the challenges posed by the unstable climates of Eurasia, and this might well have inhibited their cultural as well as physical evolution . ‘Modernity’ was not a package that had a single African origin in one time, place, and population, but was a composite whose elements appeared, and sometimes disappeared, at different times and places and then coalesced to assume the form we see in extant humans . However, during the past 400 000 years, most of that assembly took place in Africa, which is why a recent African origin still represents the predominant (but not exclusive) mode of evolution for H. sapiens. Rather than saying ‘we are all multiregionalists trying to explain the out-of-Africa pattern’ , it would be more appropriate to say ‘we are all out-of-Africanists who accept some multiregional contributions’. Acknowledgements I would like to thank David Reich, Laura Buck, and two reviewers for their helpful comments on this manuscript, and the Human Origins Research Fund and the Calleva Foundation for support of my research.
References 1 Hawks, J. (2010) Neandertals Live! (http://johnhawks.net/weblog/ reviews/neandertals/neandertal_dna/neandertals-live-genomesequencing-2010.html) [Accessed 10 March 2014] 2 Finlayson, C. (2013) Viewpoint: Human Evolution, from Tree to Braid. (http://www.bbc.co.uk/news/science-environment-25559172#) [Accessed 10 March 2014] 3 Bednarik, R. (2013) African Eve: hoax or hypothesis? Adv. Anthropol. 3, 216–228
Trends in Ecology & Evolution May 2014, Vol. 29, No. 5
4 Stringer, C. (2002) Modern human origins – progress and prospects. Philos. Trans. R. Soc. Lond. B: Biol. Sci. 357, 563–579 5 Aiello, L.C. (1993) The fossil evidence for modern human origins in Africa: a revised view. Am. Anthropol. 95, 73–96 6 Stringer, C. (2012) The Origin of Our Species, Penguin (published in the USA as Lone Survivors: How We Came to be the Only Humans on Earth by Times Books) 7 Bra¨uer, G. (1992) Africa’s place in the evolution of Homo sapiens. In Continuity or Replacement? Controversies in Homo sapiens Evolution (Bra¨uer, G. and Smith, F., eds), pp. 83–98, Balkema 8 Smith, F. (1992) The role of continuity in modern human origins. In Continuity or Replacement? Controversies in Homo sapiens Evolution (Bra¨uer, G. and Smith, F., eds), pp. 145–156, Balkema 9 Trinkaus, E. (2005) Early modern humans. Annu. Rev. Anthropol. 34, 207–230 10 Wolpoff, M.H. et al. (1994) Multiregional evolution: a world-wide source for modern human populations. In Origins of Anatomically Modern Humans (Nitecki, M.H. and Nitecki, D.V., eds), pp. 175–199, Plenum 11 Thorne, A.G. and Wolpoff, M.H. (2003) The multiregional evolution of humans (revised paper). Sci. Am. 13, 46–53 12 Howells, W.W. (1973) Evolution of the Genus Homo, Addison-Wesley Longman 13 Clark, D. (1975) Africa in prehistory: peripheral or paramount? Man 10, 175–198 14 Beaumont, P.B. et al. (1978) Modern man in sub-Saharan Africa prior to 49,000 years B.P.: a review and evaluation with particular reference to Border Cave. S. Afr. J. Sci. 74, 409–419 15 Hublin, J-J. and Tillier, A.M. (1981) The Mousterian juvenile mandible from Irhoud (Morocco): a phylogenetic interpretation. In Aspects of Human Evolution (Stringer, C.B., ed.), pp. 167–185, Taylor & Francis 16 Day, M. and Stringer, C.B. (October, 1982) A reconsideration of the Omo Kibish remains and the erectus–sapiens transition. In Pre´tirage First Inte´rnational Congress of Human Palaeontology. pp. 814–846, Nice, CNRS 17 Bra¨uer, G. (1984) The ‘Afro-European sapiens hypothesis’ and hominid evolution in east Asia during the Middle and Upper Pleistocene. Cour. Forsch. Senck. 69, 145–165 18 Protsch, R. (1975) The absolute dating of Upper Pleistocene subSaharan fossil hominids and their place in human evolution. J. Hum. Evol. 4, 297–322 19 Hammer, M. et al. (2011) Genetic evidence for archaic admixture in Africa. Proc. Natl. Acad. Sci. U.S.A. 108, 15123–15128 20 Harvati, K. et al. (2011) The Later Stone Age calvaria from Iwo Eleru, Nigeria: morphology and chronology. PLoS ONE 6, e24024 21 Harvati, K. et al. (2004) Neanderthal taxonomy reconsidered: implications of 3D primate models of intra- and interspecific differences. Proc. Natl. Acad. Sci. U.S.A. 101, 1147–1152 22 Lordkipanidze, D. et al. (2013) A complete skull from Dmanisi, Georgia, and the evolutionary biology of early Homo. Science 342, 326–331 23 Hublin, J-J. (2014) Paleoanthropology: Homo erectus and the limits of a paleontological species. Curr. Biol. 24, R82–R84 24 Hawks, J. (2013) The ‘Baided Sream’ at Year-end. (http://johnhawks.net/ weblog/topics/news/finlayson-braided-stream-2013.html) [Accessed 10 March 2014] 25 Pru¨fer, K. et al. (2014) The complete genome sequence of a Neanderthal from the Altai Mountains. Nature 505, 43–49 26 Currat, M. and Excoffier, L. (2011) Strong reproductive isolation between humans and Neanderthals inferred from observed patterns of introgression. Proc. Natl. Acad. Sci. U.S.A. 108, 15129–15132 27 Sankararaman, S. et al. (2014) The landscape of Neandertal ancestry in present-day humans. Nature http://dx.doi.org/10.1038/nature12961 28 Vernot, B. and Akey, J.M. (2014) Resurrecting surviving Neandertal lineages from modern human genomes. Science http://dx.doi.org/ 10.1126/science.1245938 29 Fu, Q. et al. (2013) DNA analysis of an early modern human from Tianyuan Cave, China. Proc. Natl. Acad. Sci. U.S.A. 10, 2223–2227