Modern Human Origins: The “Out of Africa” Debate

John H Relethford. Encyclopedia of Life Sciences. Volume 24, Wiley, 2007.

Anthropologists continue to debate on the evolutionary origin of modern humans. Although most researchers agree that modern human anatomy originated in Africa between 150 000 and 200 000 years ago, there is still debate over whether this was a speciation event that replaced archaic humans outside of Africa, or whether there was a genetic mixture of archaic and modern humans.


Although the broad picture of human evolution over the past 2 million years is generally agreed upon, the fine details are still being debated. In broad view, we see evidence of early humans (most often classified as Homo erectus or Homo ergaster) arising in Africa roughly 2 million years ago, with a smaller brain and larger face and teeth than living humans, skeletal proportions similar to living humans, and a stone tool technology. These early humans were the first hominids to leave Africa, dispersing into parts of Eastern Europe and Southeast Asia roughly 1.7 million years ago. Over time, the human brain size increased, and fossil humans dating back several hundred thousand years ago across the Old World are often referred to as ‘archaic humans’. These archaics had an average brain size similar to that of living humans, but with a larger face and a lower, differently shaped skull. There is considerable controversy over the classification of these archaics; some refer to them as early examples of Homo sapiens, whereas others see evidence of two (or more) separate species, such as Homo heidelbergensis and Homo neanderthalensis (the latter referring to the Neanderthals of Europe and the Middle East). Fossil evidence shows that by 130 000 years ago some human populations had evolved into what is often referred to as ‘anatomically modern humans’ with a high and well-rounded skull, reduction in the size of the face, and the development of a prominent chin, among other characteristics.

Viewed from a broad perspective, we could describe human evolution in terms of a transition from early to archaic to modern, with an increase in brain size (archaics) followed by a reduction in the size of the face and a change in the shape of the skull (moderns). Although accurate, this broad view does not provide us with much detail about the actual mechanisms of evolutionary change, the evolutionary relationship between populations across time and space, or the underlying question of classification. For example, was there more than one species in existence at any point during human evolution and, if so, which one are we descended from? Such questions of evolutionary change include the debate over modern human origins, which specifically deals with the questions of where, when, and how anatomically modern humans (Homo sapiens sapiens) first appeared.

Models of Modern Human Origins

Some brief historical review is needed in order to understand the current state of the debate over modern human origins.

History of the Debate

During the first part of the twentieth century, much of the research on human evolution over the past several hundred thousand years focused on Europe, specifically the evolutionary relationship between the archaic Neanderthals and early modern humans, such as were found at the CroMagnon site in France and elsewhere. A number of researchers viewed the Neanderthals as a different species from modern humans, which became extinct following the introduction of modern humans from outside Europe. At first, fossil evidence suggested a 5000-year gap between the Neanderthals, which were felt to have become extinct 35 000 years ago, and early modern humans, who appeared to have arrived in Europe some 30 000 years ago. This gap was interpreted by some as a replacement event under the assumption that the physical changes between Neanderthals and modern humans could not have taken place in such a short period of time. Others disagreed, and argued for a ‘Neanderthal phase’ in human evolution, while still others viewed Neanderthals as a variant of an evolving human lineage, perhaps physically different because of isolation and subsequent adaptation to a glacial environment (see Trinkaus and Shipman (1993) for a history of interpretations regarding Neanderthal evolution).

Over time, the fossil record grew and revealed much more complexity, including the existence of Neanderthals in the Middle East and some temporal overlap of Neanderthals and moderns in Europe. At present, we have evidence that this period of overlap lasted several thousand years, with the youngest known Neanderthal fossils dating to 28 000 years. Perhaps even more significantly, the expanding fossil record began to focus more on events outside Europe, transforming questions of modern human evolution to a global level.

By the early 1980s, this global perspective had led to the development of the multiregional evolution model of human evolution, proposed as a way of explaining why some evolutionary changes occurred worldwide (such as the increase in brain size) whereas, at the same time, some physical traits retained unique regional characteristics (Wolpoff et al., 1984). An example of this type of regional continuity is the higher frequency of shovel-shaped incisor teeth in East Asian populations today as well as in the past. The multiregional evolution model proposed that all human populations had been connected through gene flow, thus allowing common changes such as an increase in brain size to be shared across several continents, while genetic drift and selection acted to maintain some regional continuity traits. As initially suggested, the multiregional evolution model also proposed that the transition from archaic to modern humans did not occur in a single place or time, but instead resulted from the coalescence of changes taking place in different parts of the world at different times. Genetic changes occurring in one part of the species’ range were then spread throughout the rest of the species over time via gene flow.

By the end of the 1980s, this multiregional perspective was challenged increasingly by advocates of the ‘out of Africa’ model, which proposed that the transition from archaics to moderns occurred first in Africa between about 150 000 and 200 000 years ago (Stringer and Andrews, 1988). In its extreme form, the ‘out of Africa’ model hypothesized that the transition was a speciation event, and that modern humans then dispersed into Europe and Asia, replacing preexisting archaic human populations in these regions.

Current Debate

Debate between the multiregional and the ‘out of Africa’ models is often confusing, largely because there are variants of each model, with overlap in some interpretations. This confusion stems partly from the fact that the debate over modern human origins actually breaks down into two related debates: how the transition took place and where the transition took place (Relethford, 2001). The first question deals with the mode of evolution proposed, where the transition from archaics to moderns can be viewed as the formation of a new species (cladogenesis) or as evolution within a species (anagenesis). The second question looks at the geographic location of the transition, usually argued to either be solely in Africa or across the entire Old World. Different models of modern human origins can then be expressed in terms of different combinations of the answers to these questions. As first proposed, the multiregional evolution model argued for change within a species (anagenesis), where there was no single time or place for the initial transition from archaic to modern morphology. Rather, some changes occurred in Africa and some elsewhere, and these changes were eventually shared across the entire species through gene flow. In order to differentiate between this view and other multiregional models (broadly defined as models where there is genetic contribution from more than one geographic region), I have referred elsewhere to this model as the ‘regional coalescence’ variant of multiregional evolution. At the other extreme lies a version of the ‘out of Africa’ model known as the ‘African replacement model’. Here, the proposed mode of evolutionary change is speciation – modern humans are considered a separate species – and the place of origin is Africa.

There are variants of the multiregional model, which I refer to as ‘primary African origin’ models, which combine elements of both the regional coalescence model and the African replacement model (Relethford, 2001). Here, the transition from archaics to moderns is proposed to have started in Africa, but subsequent changes in human evolution occurred within a single evolutionary lineage, and there was no speciation. As the genetic and anatomical changes leading to modern humans took place in Africa, they were then shared with populations in the rest of the Old World through gene flow. In other words, the origin of modern humans was ‘out of Africa’ but within a multiregional framework. The best-known example of this type of model is the ‘assimilation model’ proposed by anthropologist Fred Smith (Smith, 2002).

The three major models of modern human origins each make specific predictions about the mode and location for the transition from archaic to modern humans:

  1. African replacement model: Modern humans arose in Africa as a new species between 150 000 and 200 000 years ago and then dispersed across the Old World, replacing archaic populations outside Africa that belonged to one or more different species.
  2. Primary African origin model (multiregional): The transition from archaic to modern humans first began in Africa between 150 000 and 200 000 years ago, and these changes then spread to populations in other geographic regions through gene flow. This gene flow could have occurred through the movement of individuals from one population to another, the movement of entire populations with subsequent genetic mixing, or both.
  3. Regional coalescence model (multiregional): The transition from archaic to modern humans did not take place in any single place or time, but rather occurred over time through the coalescence of changes across the Old World; modern traits appeared as a consequence of this mixing over time.

Another, perhaps more intuitive, way of looking at the debate is to imagine being able to search back through everyone’s family tree, generation by generation, noting where ancestors in any given generation lived. Where were our ancestors living 150 000-200 000 years ago? If the African replacement model is true, then all our ancestors lived in Africa at that time. If one of the multiregional models is true, then some of our ancestors were living in Africa at that time, but others lived outside Africa. As I will argue below, it seems likely that the correct answer is that most (but not all) of our ancestors came from Africa within the past 150,000 years or so.

Fossil Evidence

Thus far, the discussion has been rather abstract, dealing with models and hypotheses. How are these models supported by the fossil evidence of the transition from archaics to moderns?

The First Appearance of Modern Humans

One place to start is the distribution of archaic and modern humans in time and space, in particular looking to see where the first definitive ‘modern’ humans appear. Although there is continuing debate over defining characteristics of ‘archaic’ and ‘modern’ the general consensus is that modern anatomy appeared first in Africa. Early modern humans have been dated to about 130 000 years ago in Africa, with earlier forms dating back 160 000 years ago showing a transition from archaic to modern appearance (White et al., 2003). Modern humans appear later in other parts of the Old World: roughly 90 000 years ago in the Middle East, perhaps as much as 60 000 years ago in Australia, and between 40 000 and 30 000 years ago in different parts of Europe.

The distribution of the earliest modern humans argues strongly for an ‘out of Africa’ perspective, where modern anatomy first appears in Africa and later elsewhere in the world. The earlier appearance of moderns in Africa argues against the regional coalescence model, which proposes less distinct boundaries between archaic and modern morphology. Although the fossil record is sparse in many times and places, and as such cannot definitely rule out the regional coalescence model, the data at present suggest that the general hypothesis that moderns arose ‘out of Africa’ is more likely to be correct. However, this does not solve the debate entirely, as both the African replacement model and the primary African origin model propose an African origin. The later appearance of modern anatomy outside Africa can be explained by both models. If replacement took place, then it would take time for populations expanding out of Africa to replace archaic populations elsewhere. The same thing applies to the primary African origin models that combine an African origin with admixture outside Africa; it would take time for this genetic mixing to occur, resulting in a later appearance of modern morphology outside Africa.

Replacement or Admixture?

The main issue at present is whether fossil evidence shows evidence outside Africa for replacement or for genetic admixture, and, if the latter, then the amount of admixture that took place. Resolution of this issue is likely to be more difficult if (as some suspect) the overall level of admixture was relatively low, but nonzero. The major difference between the African replacement model and a primary African origin model is that the former proposes essentially zero admixture, whereas the latter proposes some admixture.

Is there fossil evidence for admixture? One specimen that has received considerable attention recently is the skeleton of a 4 year-old child from the Abrigo do Lagar Velho in Portugal, dating to about 24 500 years ago, which shows a mixture of Neanderthal and anatomically modern traits (Zilhão and Trinkaus, 2002). Although some have argued that this represents evidence of admixture, others have concluded that the skeleton is essentially a modern human.

Patterns of regional continuity have long been argued by supporters of multiregional models as evidence rejecting wholesale population replacement (Wolpoff et al., 1984; Smith, 2002). The persistence of regionally specific traits is often taken as evidence for some genetic mixing. Proponents of multiregional evolution have documented evidence for regional continuity in Europe, East Asia, and Australasia, although supporters of a replacement model have suggested that apparent continuity is instead the result of shared traits retained from an original common ancestor.

Another piece of evidence from the fossil record is the analysis of temporal changes in Europe, where the fossil record is more abundant than elsewhere. Some anatomical traits are found in moderate to high frequencies in European Neanderthals and are still present, although reduced in frequency, in the earliest known post-Neanderthal moderns in Europe. Some of these traits are still present in living Europeans, although at very low frequency (Wolpoff, 1999). The persistence of such traits suggests genetic continuity over time, and thus some mixture of archaics (Neanderthals) and moderns in Europe. Further, the decreasing frequency of these traits is consistent with a pattern of admixture that acts to dilute the ancestral contribution of Neanderthals over time. The ‘swamping’ of Neanderthal genes might reflect the end result of the mixing of a relatively small Neanderthal population with a much larger population of moderns that had expanded out of Africa (Relethford, 2001). If true, then most of our ancestry does come from ‘out of Africa’ but with some small amount of genetic contribution from archaic populations outside Africa.

Genetic Evidence

Since the mid-1980 s, the debate over modern human origins has increasingly incorporated inferences from the analysis of genetic variation in living human beings. Anthropologists and geneticists examine patterns of genetic diversity in different human populations today in an attempt to reconstruct the past: in other words, what could have happened in the past to give rise to what we see today in terms of genetic variation? There are a number of approaches that have been used to shed light on the question of modern human origins, but here I focus only on two major ones: analysis of gene trees and of genetic diversity.

Gene Trees and Common Ancestors

Given contemporary methods of molecular research and our increasing knowledge of the human genome, it is possible to construct genealogies based on genetic data, providing a means to show how closely related different people are to each other, which in turn can be used to make inferences about our ancestral history. In particular, various methods have been developed to construct genealogies of genes (gene trees) and to identify the most recent common ancestor shared by humans across the world and the date at which this ancestor lived. Initial studies focused on mitochondrial deoxyribonucleic acid (DNA), the small amount of DNA that exists in the mitochondria of the cells and is inherited solely through the female line (males do not contribute mitochondrial DNA to their offspring), as compared with nuclear DNA, which is inherited from both parents. Studies of mitochondrial DNA supported a common mitochondrial ancestor of humanity that lived in Africa roughly 200 000 years ago (Cann et al., 1987). As such, this date was considered too late in time to be compatible with a multiregional model that proposed common ancestry back to the origin of the genus Homo 2 million years ago. Although similar results were also found when examining Y-chromosome DNA (inherited only through the male line), some studies of nuclear DNA suggested older dates, and a common ancestor that likely lived outside Africa. Part of the problem in interpreting gene trees is confusing the history of a particular gene or DNA sequence with the evolutionary history of populations. Because of differences in mutation rates, genetic drift, and other factors, different gene trees may show different events in a population’s past. Although much more work is needed in this area, the results to date suggest that most gene trees reflect recent African ancestry, but not all of them, a view consistent with the fossil evidence for a ‘mostly out of Africa’ ancestry of living humans. Geneticist Alan Templeton has developed a method of combining information from different gene trees and concludes that the underlying pattern is due to several expansions out of Africa during the past 2 million years, each resulting in some admixture with previous non-African populations (Templeton, 2002). Again, this is compatible with the model that I proposed above, since recurrent expansion with admixture would mean that most of our ancestry (but not all) was out of Africa.

Genetic Diversity

Another source of historical inference from genetic data comes from the analysis of levels of genetic diversity in different parts of the world. DNA evidence (both mitochondrial and nuclear) shows that sub-Saharan African populations today have more genetic diversity than populations in other geographic regions. There are two possible reasons for this regional difference in genetic diversity: (1) diversity increases over time and (2) increases with population size. If modern humans arose as a new species in Africa and later dispersed into other regions, then African populations would be older and hence have greater levels of genetic diversity, which is what we find today. Another (and, in my view, more likely) possibility is that the population size of sub-Saharan Africa was larger than in other geographic regions throughout most human evolution, perhaps relating to differences in environment and carrying capacity. If true, then on this model, too, sub-Saharan Africa would have the greatest level of genetic diversity today (Relethford, 2001). Note, however, that the hypothesis of regional differences in population size does not favour replacement or admixture; it is compatible with either model. In a multiregional model, greater average population size in Africa could mean that most of our ancestry comes from Africa.

Ancient DNA

The above discussion deals with the examination of genetic variation as it exists in our species today in an attempt to figure out what happened in the past. Advances in molecular technology have also provided some direct information on genetic variation in the past; since 1997, mitochondrial DNA sequences have been extracted from several Neanderthal fossils. These studies show that Neanderthal mitochondrial DNA is quite different from that of living humans, and many scientists have concluded that this provides definitive proof for Neanderthal replacement (e.g. Krings et al., 1997). Of course, if true this does not tell us anything about the issue of replacement versus admixture elsewhere in the Old World. Moreover, the mitochondrial DNA evidence is not accepted by all; although Neanderthal DNA is different, it is less clear whether this difference necessarily means they were a separate species, incapable of genetically mixing with modern humans, or whether they were a different subspecies. Another issue to contend with is the fact that there are no equivalent DNA sequences in living people. Does the apparent extinction of Neanderthal mitochondrial DNA imply the extinction of the Neanderthals as a population, or could this pattern be due to admixture and random chance (genetic drift)? Although the arguments supporting these views continue, we need additional samples of Neanderthal DNA and early modern DNA to make more progress on this issue. One additional complication is the potential for contamination of ancient DNA with that of living human DNA introduced by people who have handled the fossil specimens.

Consensus and Future Directions

As is typical in much scientific research, studies of modern human origins have provided the answers to some questions, while raising new questions at the same time. Although the long-term debate over the number of species and evolutionary relationships within the genus Homo continues, progress has been made on several issues. The predominantly European focus of much of the early twentieth century has given way to the understanding that questions of modern human origins need to be addressed from a global perspective, and that Africa, not Europe, was the main stage for much of human evolution.

Fossil and genetic evidence has shown the strong likelihood that anatomically modern humans arose in Africa roughly 150 000 years ago, and then moved throughout the rest of the Old World over time. Our current knowledge of the fossil record demonstrates an earlier appearance of modern anatomical form in Africa than elsewhere, and it seems less likely that the multiregional model of regional coalescence is correct. At this level, proponents of an ‘out of Africa’ model appear to be vindicated. However, it is less clear whether the earlier African origin of modern human anatomy is best described by a speciation and replacement model, or by a model of genetic replacement through admixture within a single evolving lineage. In other words, the debate is beginning to focus less on the issue of an African origin and more on the relationship of early modern humans in Africa to preexisting archaic populations outside Africa. Were these archaic populations replaced, or were their genes assimilated into an expanding population of ‘modern’ humans?

Future research on modern human origins needs to be focused on addressing evidence for or against archaic admixture. Although current evidence from the fossil record favours some regional continuity, and thus admixture, additional specimens across time or space are needed to make this conclusive. In addition, hypotheses and analyses that can distinguish between admixture and continuity need to be formed versus the presence of shared traits resulting from earlier shared ancestry. Likewise, although the finding of ancient non-African roots from gene tree analysis tends to support some admixture, additional data and analyses are needed to rule out potential confounding factors, such as natural selection. Further, both fossil and genetic analyses need to move beyond simpler demographic models and deal with potential complexities due to varying population size, across both space and time.

We also need to grapple with potential problems that might arise if the true rate of archaic admixture was very low, but not zero (as implied by complete replacement). If, for example, the Neanderthals of Europe contributed only a very small amount (say less than 5%) of our species’ current genetic variation, how could we distinguish between this small amount of mixture and the zero amount of mixture expected under complete replacement? From a mathematical perspective, the difference between zero and a few per cent might seem trivial, but this difference implies a rather different underlying evolutionary mechanism: a form of genetic replacement within a species rather than a population replacement by another species. In addition to providing answers about our own ancestry, the study of modern human origins also has the potential to provide general insight into evolutionary processes.