What do we know about the ancestry of Homo erectus?

This morning a reader alerted me that a quote from one of my scientific papers was making the rounds of a creation science interest group on Facebook. Here's the quote:

“No gradual series of changes in earlier australopithecine populations clearly leads to the new species [Homo sapiens], and no australopithecine species is obviously transitional.”

I know I have a lot of readers who have run across selective quotation before. It's a common and obviously misleading tactic: take a long passage and quote a single sentence out of context. In this case the sentence on its own seems to suggest that our species, Homo sapiens, did not evolve in a gradual series of changes from earlier populations of Australopithecus.

You might guess from the word “australopithecine” that this is a fairly old quote, and indeed it comes from one of my first academic articles, published in 2000. The fossil evidence for human evolution has grown a lot since then. So what was the broader context of this quote, and what do I think today?

Three skulls from the Turkana Basin all dating to between 2.1 million and 1.6 million years ago, representing three species of Homo.

Genetic reshuffling

The quote comes from an article titled, “Population bottlenecks and Pleistocene human evolution”, which I wrote with several collaborators including my research mentor, Milford Wolpoff. In the article, we considered an emerging understanding from human genetics at that time: today's humans throughout the world came from a population that for most of the Pleistocene was quite limited in genetic variation.

The data on human genetic variation began to build during the 1970s and 1980s. At first the evidence of low variation came from protein polymorphisms and later from DNA evidence. The geneticists Masatoshi Nei and Dan Graur in a 1984 paper suggested that the low genetic variation must reflect repeated population bottlenecks, not only in humans but in many other species also. The geneticists Rebecca Cann, Mark Stoneking, and Allan Wilson in 1987 found that human mitochondrial DNA had exceptionally low variation and a common ancestor within the last 200,000 years, prompting the idea that every bit of the genetic ancestry of today's people came from African ancestral groups in that short time. But by the late 1990s, Henry Harpending, Alan Rogers, and other human geneticists had shown that the mitochondrial story was paralleled by nuclear DNA on a much longer timescale. Their work suggested that factors extending across the last million years or more must have been at play.

This is where I entered the story. In that 2000 paper, we wanted to see whether the fossil and archaeological records might align with this genetic evidence for longer-term constraints on variation. In that era, paleoanthropologists had come to see early Homo erectus as a big departure from every earlier hominin. The species had larger bodies, taller stature, larger brains, it was the first human relative to strike out from its African homeland, it seemed to have changes in diet, home range, and technology. These concepts followed after the discovery of the Turkana Boy skeleton, which had been described in detail within a 1993 volume edited by Alan Walker and Richard Leakey. Earlier scientists thought about an orderly sequence from Australopithecus to Homo habilis to Homo erectus, then ultimately onward to us. But by the 1990s it seemed that H. erectus was as old as the oldest Homo habilis fossils, or even older. And while the fossils of H. habilis weren't great, that species itself seemed pretty much like Australopithecus in its small body size, short stature, fairly big molars, and ratio of arm to leg length.

Skeletons of “Lucy” (Australopitheucs afarensis) and “Turkana Boy” (Homo erectus)

So, we thought, what if the first appearance of H. erectus was a rapid event? What would that look like? (In the following passage, we used the name “early Homo sapiens” instead of Homo erectus because at the time we argued that the lineage did not have later speciations that split it into different species. It's a long story.)

“The anatomy of the earliest H. sapiens sample indicates significant modifications of the ancestral genome and is not simply an extension of evolutionary trends in an earlier australopithecine lineage throughout the Pliocene. In fact, its combination of features never appears earlier; some of its characteristics are unique, such as the very large body sizes and long legs described below, while others can be found in isolation in various different Pliocene and penecontemporary hominid species.”
“If we assume these earlier australopithecines are a group of very closely related species, for instance, nearer to each other than Pan and Homo, we can expect that they differ much more in allele frequencies than in the presence or absence of specific genes for these features. Therefore, a reshuffling of existing alleles could result in the frequencies of features we observe in early H. sapiens. Thus, our second question is about this reshuffling, whether early H. sapiens is a consequence of rapid speciation with significant founder effect or the result of a long, gradual process of anagenic change. The first explanation, cladogenesis, is suggested by the fact that no gradual series of changes in earlier australopithecine populations clearly leads to the new species, and no australopithecine species is obviously transitional. This may seem to be an unexpected statement, because for 3 decades habiline species have been interpreted as being just such transitional taxa, linking Australopithecus through the habilines to later Homo species. But with a few exceptions, the known habiline specimens are now recognized to be less than 2 Myr old (Feibel, Brown, and McDougall 1989) and therefore are too recent to be transitional forms leading to H. sapiens.”

So there's the full context for that quote from the beginning of the post. Our idea was to explain how a rapid change might have brought about the combination of features found in Homo erectus for the first time. Maybe an initially small population of an Australopithecus-like species had started rapidly adapting to new circumstances. At first it wouldn't be much different from other kinds of Australopithecus, and most of the first set of changes would result from random genetic drift in a small population. But if some of the changes happened to click with the environment, natural selection would kick in.

If you looked at the resulting population after the fact, you would not find very many features that aligned closely with the ancestral form of Australopithecus. The ancestral features would be more of a mosaic, some resembling one species (maybe Homo habilis), some resembling another (maybe Homo rudolfensis or Australopithecus africanus), but most seeming pretty new. The main genetic pattern you would notice was a lot of drift across the genome reflecting the population bottleneck.

We called it genetic reshuffling, and it drew on some old ideas in biology including work by Ernst Mayr and Sewell Wright.

“Our interpretation is that the changes are sudden and interrelated and reflect a bottleneck that was created because of the isolation of a small group from a parent australopithecine species. In this small population, a combination of drift and selection resulted in a radical transformation of allele frequencies, fundamentally shifting the adaptive complex (Wright 1942); in other words, a genetic revolution (Mayr 1954; Templeton 1980).”

Little did I know at the time that I would help found a new evolution institute at the University of Wisconsin where Wright had spent the last part of his research career, or borrow some of the books that had been part of his library!

But did this reshuffling idea have legs as long as Homo erectus? Well, it turned out that Homo erectus didn't have legs as long as we thought.

Three skulls of Homo erectus from Dmanisi, Republic of Georgia. From left: D2700, D2282, and D4500. Image: John Hawks

New fossils

A lot has changed in the fossil record during the last 24 years. For one thing, we know more about the diversity of early hominins. Around 2 million years ago when H. erectus was making its first appearance in South Africa, the same region was home to Australopithecus sediba. That species is like Homo habilis in a lot of ways, from body size, arm length relative to leg length, and tooth sizes, with a more humanlike hand that H. habilis, a pelvis resembling H. erectus in some ways, but brain size and other features like Australopithecus africanus. In other words, it also seems like a mosaic that combines features found in other species, including later ones.

Australopithecus sediba is not alone in representing greater diversity of earlier human relatives. There's also Kenyanthropus platyops in the Turkana Basin and Australopithecus garhi and Australopithecus deyiremeda in the Afar region. A lot less is known about those species' anatomy. But one notable observation is that the tree of hominins at every point since 3.5 million years ago always had varied branches that existed at the same time.

Meanwhile, the fossil picture of Homo erectus has also changed. More fossils now document the first 500,000 years of its evolution, from its first appearances at Drimolen, South Africa and Garba, Ethiopia just around 2 million years ago, up to the Turkana Boy skeleton around 500,000 years ago. Most informative is the fossil sample from Dmanisi, Republic of Georgia, which includes five skulls and at least two partial postcranial skeletons. Those fossils do not share the very tall stature that was estimated for the Turkana Boy, and they have brain sizes in the same range as Homo habilis, not like the much larger Turkana Boy. With these discoveries, the early evolution of H. erectus no longer looks like a sudden pulse of evolution of body, brain, and behavior. If you put these early H. erectus fossils next to H. habilis, H. rudolfensis, and Au. sediba, the patchwork of their features seems like a group of Dungeons and Dragons players rolled dice to pick the strengths of their characters. These species were all close relatives, and they overlapped a lot in their biology.

Finding this kind of evidence of early Homo erectus has caused many paleoanthropologists to revise their views of the Turkana Boy skeleton itself. The individual was a child of around 8 years when he died. He would have grown larger as an adult, but recent estimates suggest a stature in between the tall 1990s-era estimate and the short Dmanisi statures. The 1990s-era reconstruction of his pelvis had a very humanlike shape, but other fossil pelves attributed to H. erectus are broader and not so much like recent humans, meaning Turkana Boy's probably was also.

Of course, the most important new information during the last decade has been the ancient DNA record. There is no ancient DNA from Homo erectus yet, nor for any other hominin fossil older than the early Neanderthals from Sima de los Huesos, Spain. The biochemical constraints on DNA preservation make it unlikely that direct DNA evidence from the earliest Homo species and Australopithecus will ever be found. But we have learned enough from later human groups, including the Neanderthals and Denisovans, to see how important mixture was to their evolution. This picture is confirmed by DNA studies of nonhuman primates, most of whom today have deep population lineages that were connected from time to time by gene flow.

A transition to Homo

Ideas about the first appearance of Homo have changed a lot over time. From the 1960s to the 1990s, most researchers thought of an Australopithecus species evolving into Homo habilis, and thence onward to Homo erectus. The classic “March of Progress” painting by Rudolph Zallinger illustrates exactly this picture. Then in the 1990s, many researchers came to accept that H. habilis and the similar H. rudolfensis were much more like Australopithecus in their body size, shape, and behavior than Homo erectus. The idea of a simultaneous shift in body size, brain size, diet, and technology took root. The appearance of Homo erectus seemed to be one of the major transitions of our evolutionary history.

Returning to the subject of the post, how do we look at this transition today? The idea of a big shift from an unknown Australopithecus-like species to Homo erectus has given way to an understanding that none of these hominins around 2 million years ago were very different from each other. Australopithecus and their close relatives were much more humanlike, and Homo erectus much more Australopithecus-like, than it once seemed.

If this were a Pleistocene version of Ancestry.com, we still cannot say exactly which form of Australopithecus or its relatives gave rise to Homo erectus. But that's no longer a central goal for understanding the evolution of our genus. Both Australopithecus sediba and Homo habilis show something about what the transition from early hominin anatomical patterns to Homo erectus may have looked like. These species were very close to each other, certainly diverging within a few hundred thousand years from a common ancestor. That makes them as close as modern people, Denisovans, and Neanderthals were to each other. What we know from DNA about the interbreeding of the later human relatives must have been true of early Pleistocene ones also. When they met, they would have exchanged some traits and behaviors along with genes. Just like modern-day genealogy discoveries, the ancestor of Homo erectus might be “all of the above”.

For a long time I've thought about revisiting my old 2000 article, examining how the evidence from both genetics and fossils has changed. The idea of repeated bottlenecks has faded to some extent, as human geneticists came to accept that many factors affect genetic diversity on the timescales involved. In the paper we noted that the genetic variation today is unlikely to preserve much signal from as early as two million years ago. With whole genomes, it's even more clear that such ancient population dynamics are largely masked by more recent events. And the fossils have changed a lot, as I note in this post.

But as I read through the paper now, I think we were on the right track with the core idea: reshuffling of standing genetic variation among closely related populations that were subject to strong genetic drift. Whole genomes have brought about the understanding that most human traits are highly polygenic, and natural selection on traits acts upon standing genetic variation much more rapidly than new mutations can arise. As local populations of Australopithecus species came to differ from each other in allele frequencies, they would explore a substantial space of phenotypic variation. Add in the fact that they sometimes met and exchanged genes, and you have a scenario where the anatomical features of species might come to look like reshuffled versions of each other. This would put random chance near the top of the reasons for the evolution of Homo erectus as we know it.

I'm just sure that nobody will pick up that last sentence and take it out of context!


Notes: The late 1990s were a time when paleoanthropologists ran free with classification. Two issues from the time are important for interpreting this post. Some researchers in that era argued that Homo habilis and Homo rudolfensis should instead be classified within Australopithecus, because they were so different from Homo erectus. At the same time some researchers, led by Milford Wolpoff, argued that the human lineage from early Homo erectus onward was a single evolutionary species, meaning that Homo sapiens included most Pleistocene fossil humans. Both these classification proposals lost steam during the early 2000s. The discovery of the Dmanisi fossils played a role in reframing ideas about classification in early Homo: they showed much greater similarity to Homo habilis than other known H. erectus samples, and much greater difference from recent humans.

The topic of ancient population sizes and genetic variation was also the core of my dissertation research. It's remarkable that the conclusions from four genes could arrive at basically similar results as whole genomes, for the basic estimation of effective population size. It's also remarkable how little we still know about the genetic structure of the adaptations that may have been important to fossil hominins.

I told my graduate students recently that this paper for me was a good model for how interdisciplinary thinking matters, even if the details turn out to be wrong. It's almost inevitable that the details in any one field will change as new data are collected, and that's especially true of fossil hominins. What matters across fields is not those details but how the underlying processes are reflected in different real-world phenomena. Both DNA and fossils had not yet reached the point in 2000 that they could clearly illuminate those processes for early Homo. We are getting closer now!

References

Cann, R. L., Stoneking, M., & Wilson, A. C. (1987). Mitochondrial DNA and human evolution. Nature325(6099), 31–36. https://doi.org/10.1038/325031a0

Hawks, John, Hunley, Keit, Lee, Sang-Hee, & Wolpoff, Milford. (2000). Population Bottlenecks and Pleistocene Human Evolution. Molecular Biology and Evolution17(1), 2–22. https://doi.org/10.1093/oxfordjournals.molbev.a026233

Nei, Masatoshi & Graur, Daniel. (1984). Extent of protein polymorphism and the neutral mutation theory. Evolutionary Biology17, 73–118.

Walker, A., & Leakey, R. E. (1993). The Nariokotome Homo Erectus Skeleton. Harvard University Press.