Header image

john hawks weblog

paleoanthropology, genetics and evolution

Photo Credit: 3D printing a Homo naledi cranium. John Hawks CC-BY-NC-ND

A paleoanthropological Thanksgiving

This is Thanksgiving weekend in the U.S., and friends on my social feeds have been forwarding a proverb: “When you have more than you need, build a longer table not a higher fence.”

I’m thankful for the opportunity to collaborate with such an outstanding group of scientists on a rewarding project. I’m also thankful for my many colleagues around the world who are so deeply engaged with the study of human origins. As paleoanthropologists, we are part of a select group, probably fewer than a thousand people worldwide, who care so much about the merest hints of evidence about our past.

It’s a strange family sometimes.

Paleoanthropology freedom from want
I think Karabo put something in the punchbowl...

The other traditional Thanksgiving activity, after giving thanks, is arguing with aging relatives about politics and religion. Sure, this has its unpleasant moments. Let’s face it, the paleoanthropology family has a few cranky uncles who fart at the dinner table just to get a rise out of people.

But setting aside emeritus attitudes, most arguments come from shared love for the science and a deep concern about its uncertain future. Population growth and development are encroaching upon many of our most important field sites. Funding for our science is declining globally, particularly when measured against the growing costs of fieldwork and analysis. Students are entering the field as professionals without ever seeing a cast of some of the most important discoveries in the history of our field.

Still there is much to give hope. The pace of paleoanthropological discovery has accelerated during the past fifteen years. A greater proportion of students than ever are being trained in field research methods, with field schools at some of the most iconic sites. New technologies have unleashed the data hidden within old fossils. These discoveries have transformed our view of ancient populations.

The roots of paleoanthropology are in the pre-evolutionary days of the nineteenth century. Today we work to understand one of the richest integrated datasets in twenty-first century biology. Even though there may be fewer than a thousand of us around the world, paleoanthropology is vastly larger than any single research team. The extended paleoanthropological family is growing, not by training new paleoanthropologists, but by greater integration of human geneticists and other biologists who can give new perspectives on the hominin fossil record and our evolutionary history.

It is an extraordinary time in our science, and for that I am thankful.

Tim White, from an essay in 2000:

In the armchair, do not assume that fieldworkers “stumble” over fossils or are rendered stupid by getting dirty and absorbing lots of ultraviolet light. Sometimes they can even correctly interpret what they find.


White TD. 2000. A view on the science: Physical anthropology at the millennium. American Journal of Physical Anthropology 113:287–292. doi:10.1002/1096-8644(200011)113:3<287::AID-AJPA1>3.0.CO;2-8

What is the 'braided stream' analogy for human evolution?

Two years ago, I pointed to a year-end retrospective that Clive Finlayson had done for BBC News: “Viewpoint: Human evolution, from tree to braid”. He reflected on the meaning of 2013’s significant paleoanthropological discoveries, from the new Dmanisi Homo erectus skull to the Sima de los Huesos and Denisova ancient DNA discoveries.

I was thinking about this old post again today because the topic of the “braided stream” analogy keeps coming up, and Finlayson’s piece was one of the clearest expressions of the idea.

The overall theme of the piece was a key realization of the last few years of paleoanthropology: the importance of gene flow between ancient populations that were morphologically much more diverse than today’s people. Finlayson emphasized the implications: that the more we discover, the more mixing seems to have been an important part of our evolution:

It seems that almost every other discovery in palaeoanthropology is reported as a surprise. I wonder when the penny will drop: when we have five pieces of a 5,000-piece jigsaw puzzle, every new bit that we add is likely to change the picture.
Did we really think that having just a minuscule residue of our long and diverse past was enough for us to tell humanity's story?
If the fossils of 1.8 or so million years ago and those of the more recent Neanderthal-modern human era were all part of a single, morphologically diverse, species with a wide geographical range, what is there to suggest that it would have been any different in the intervening periods?

The whole article provides a great account of how mixing matters to the way we understand human evolution.

When I wrote about this in 2006 with Greg Cochran, we noted that the development of genetic markers to examine variation in other species had begun to turn up widespread evidence of mixture between divergent populations. At that time, some of the most important action was in studying the process of domestication. As humans domesticated plant and animal species, they brought lineages into contact that had been partially or completely isolated in their natural geographic ranges. That mixture had huge effects on the resulting domesticated species, and subsequent introgression from the spreading domesticated populations likewise had huge effects on the genetics of remaining wild populations. It seemed clear that this would also be true of populations of ancient humans, who occupied large geographic ranges and were more closely related than subspecies of many primate species. But the evidence was not yet there.

Of course, that evidence began to emerge in 2010, showing that most living humans carry at least a small fraction of Neandertal genes. Later that year, the Denisovan genome confirmed that some living human populations have genetic ancestry in multiple archaic human populations not yet discovered from archaeological or fossil evidence. But this year, the reality of population mixing really blossomed, as ancient genomes themselves were incorporated into a yet deeper network of separating and remixing.

A braided stream
A braided stream. Photo Credit: Sam Beebe, Ecotrust via Compfight cc

The “braided stream” analogy captures different information about human origins than the usual branching tree. The branches of a tree do not reconnect with each other above the point where they initially separate. A tree will never admit to exchanging sap between its branches, and there are no little xylem hyphae between branches to carry sap anyway. Our evolution was truly a network in which multiple populations existed and contributed to our process of adaptation.

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. It’s just not clear yet which populations those were. Maybe an evolutionary swamp would be a better analogy, full of algae-covered bayous.

One of my long-time correspondents notes that there’s a name for this concept: reticulation.

Reticulated giraffe necks
Photo Credit: Just chaos via Compfight cc

The word comes from the Latin, for “little net”, and the concept of a network is increasingly central to the way we must consider genetic relationships between living and past people. It will be different networks for different genetic loci, involving different ancient populations stretching far back into the Pleistocene.

I like the braided stream, and it’s clear that its time has come. Ancient DNA has begun to show the process of genetic exchange was not a minor player in our evolution. All human populations today evidence some mixture of ancient populations that existed well before the “origin of modern humans”. Genetic exchanges between different populations were dominant in the formation of some human adaptations.

Meanwhile, some human populations can be understood only as the mixed descendants of other, yet more ancient ones, which archaeologists have not yet found. “Ghost population” is the term that geneticists have coined to talk about these mysterious populations known only from the slight mixture of their genes into more recent groups. Two such ghost populations seem to have contributed to the ancestry of living African populations. Another pre-Columbian ghost population contributed to South American indigenous groups. One Siberian ghost population seems to have contributed both to the ancestors of American Indian and to European peoples, as well as to the ancient Mal’ta individual from the Lake Baikal area.

It’s mixing all the way back.

What will be exciting moving forward is integrating this complex genetic picture with archaeology. In his year-end piece from two years ago Finlayson cited some examples of how our view of Neandertal behavior had become more resolved:

A recent paper, for example, supports the view that Neanderthals at La Chapelle-aux-Saints in France intentionally buried their dead which contrasts with reports of cannibalistic behaviour not far away at El Sidron in northern Spain.
Here we have two very different behavioural patterns within Neanderthals. Similarly, modern humans in south-western Europe painted in cave walls for a limited period but many contemporaries did not. Some Neanderthals did it in a completely different way it seems, by selecting raptor feathers of particular colours. Rather than focus on differences between modern humans and Neanderthals, what the examples show is the range of possibilities open to humans (Neanderthals included) in different circumstances.

The time of Neandertals is truly redawning. They are the first ancient population that we have begun to use our full range of genetic tools to understand. We have so many points of comparison between them, stretched across Central and Western Eurasia, and historical human populations. They faced a history of climate change and population contact unmatched in duration among our recent population history. They innovated with their behaviors sometimes in parallel with African populations of the Middle Stone Age, sometimes quite differently. Matching their ecological and archaeological circumstances with genetics will tell us how much of the human evolutionary record we are still missing for other groups.

Neandertals are now the first case of examining human evolution in the deep time dimension. It is the Neandertals who have given us both morphological and genetic evidence across a series of consecutive time slices. In many ways, they are the model organism by which we will come to understand the origins of people throughout the world.


Hawks, J., & Cochran, G. (2006). Dynamics of adaptive introgression from archaic to modern humans. PaleoAnthropology, 2006, 101-115. (PDF free online)

Should journals do post-pub review?

Nikolai Slavov recently published an opinion piece in eLife arguing the advantages of post-publication review of scientific papers: “Point of view: Making the most of peer review”

As with referee reports, I often read blogs and comments and find many—but not all—of them helpful. However, just as the scientific community is failing to take full advantage of the time and expertise that goes into referee reports, I feel that we are also failing to take advantage of the possibilities offered by post-publication comments on papers; indeed, many legitimate comments remain ignored by authors, journals and universities (PubPeer, 2014). By neglecting these comments we are missing an opportunity to reduce the time and resources that are spent trying to repeat and build on experiments that are not reproducible (Freedman et al., 2015).

Slavov’s piece suggests that journals should find ways to facilitate and track the post-publication commentary that follows their papers.

Of course, every post-publication comment on a paper does not have to lead to the publication of a formal challenge (such as a Technical Comment in Science or a Brief Communication Arising in Nature) or a retraction. However, I feel that many of the legitimate concerns about papers that are being raised in blogs and other platforms are being ignored by journals, so there is a clear need to make sure that comments that satisfy some basic criteria (see below) are acted upon by journals. The peer review process, the reproducibility of published results, and the scientific community as a whole stand to benefit tremendously from a more inclusive consideration of post-publication comments.

Most journals maintain reputation by perpetuating a fiction that they have thoroughly vetted papers during the editorial process. Authors who survive the editorial process rightly expect some protection from the journal from some kinds of critical attention. Of course, if the journal ignores well-justified critical commentary, the journal’s reputation must suffer.

As it is, unless a paper is very badly flawed, the journal and authors usually have shared interests with respect to commentary. Critical commentary on the papers must to some extent be adversarial to the journal as well as to the authors. Meanwhile, additional laudatory commentary must be independent (and therefore not promoted by the journal), or else it seems self-serving to both the journal and the authors.

In reality, no paper that answers the concerns of every possible reviewer. Commentary has a real value, because it can bring the attention of the community and the authors to unexamined aspects of the work. Many journals used to create a space for such commentary accompanying a paper, by inviting independent scholars to comment on a peer-reviewed draft of the article. Some journals still do this, including Current Anthropology.

Should journals take on even more responsibility than this? So far, journals have done very poorly managing commenting on published articles. Comments on journal sites often mix of one or two critical comments, sometimes very formally written, with many low-information comments and trolls.

Research by my UW-Madison colleagues Dominique Brossard and Dietram Scheufele has shown that readers’ perception of an online article is affected by the comments that follow it (“Online communication biases upon the public perception of science”). If a journal is going to publish comments after an online article appears, those articles will affect readers’ perception of the paper. Authors should not be protected from critical commentary, but authors deserve strong moderation to enforce some rigor and to filter out obvious trolls.

In other words, effective post-publication commentary is not going to happen alongside published papers unless journals invest a great deal of time and assume responsibility for post-publication editorial work. That’s not very different from the current system of technical comments, although actually investing effort into managing commentary would be a substantially faster system than the current one.

Speed in the service of better science is no vice.

I recently published a technical comment in Science, in response to a paper on a fossil hominin discovery. I submitted the comment on the same day the paper was published. The comment did not appear in Science until more than three months later. That delay is unnecessary in the internet era—it is an artifact of the space and scheduling limitations of paper, so that no article will end up with more than one or two published technical comments, and no paper issue of the journal will have more than one or two total. In practical terms, prolonged delays in publishing critical comments shield the journal and authors from answering criticism while they are promoting their work to the media.

Is there a solution? Not yet.

The PubPeer model has potential, because readers can return to a single place to find comments on published articles, and commenters can develop a reputation on the site. Commentary on the site can be rapid, which makes real exchanges possible. But PubPeer itself lacks the kind of moderation and links to most journals that would be necessary to actually draw readers into the commentary. So far, the site lacks the critical mass to raise standards of scientific discussion on most topics.

There is no effective answer yet, but many are approaching the topic of post-pub review with the same questions.

From Ethan Siegel: I’m weirdly fascinated by the ghostly image of a human face that emerges after averaging dozens of inanimate objects: “Averaging Inanimate Objects Can Produce Human Faces”.

The trick is that objects around us, from parking meters to tractors, have been designed with face-like symmetry.

It does make you wonder, though, whether there is reality beneath all this work from evolutionary psychology about “average faces”.

Google today (November 24) is running a Google Doodle commemorating the 41st anniversary of the discovery of the famous “Lucy” skeleton of Australopithecus afarensis.

Lucy Google Doodle

I really like this moving graphic for two reasons: First, the figures are female. Second, it’s only three figures. The usual five-figure progression creates a misleading impression that every species falls upon a single line of progress, with no side branches. Three figures shows clearly that one can be an intermediate, but to me leaves less of an impression of inevitable progress.

Maybe that’s just me.

A point of interest is that this Google Doodle is being shown throughout nearly the entire world. The map of countries where Google is showing the doodle is available from Google’s website.

Google doodle map

It’s a pretty cool moment for evolutionary science in society. Still, the lack of Lucy graphics in Iran, Turkey and Sudan I can understand, but Svalbard?

Young adults in the U.S. increasingly accept evolution by natural processes

After many years of stasis, the acceptance of evolution has taken a noticeable uptick for American adults. Most of this increase comes from the change in young adults 18 to 29 years of age.

Since its 2014 survey of the U.S. public, the Pew Research Center has issued many press releases and interactive features about its findings concerning public attitudes toward science topics. I’m actually quite amazed at the sheer number of press releases they roll out, targeted individually at every religion and issue.

The 2014 Pew survey results are reported on their website: “Americans, Politics and Science Issues”. With respect to the evolution questions, the survey results show a strong association between evolution acceptance and lower age, and between evolution acceptance and lower religious attendance. Other demographic factors had associations of medium strength with evolution acceptance, including political party affiliation, gender, and education.

The evolution results are not breaking news, but I was prompted to write by an article in Slate by Rachel Gross, which especially emphasizes the recent growth in evolution acceptance among young adults: “Evolution Is Finally Winning Out Over Creationism”.

This increase among young adults first was noticed in the results of a 2013 Pew Research Center survey, and continued in a 2014 survey. In the more recent survey, a slight majority (51%) of adults ages 18-29 agreed with the statement that humans and other living things have evolved over time by natural processes such as natural selection. A much greater majority, 73%, agreed that humans have evolved over time, and have not been in their present form since the beginning of time.

Any survey is tricky to interpret, especially because the way a question is worded can make a difference to the way people respond. The Pew survey asks its respondents about many different scientific topics, nearly all of them related to public policy, such as genetically modified organisms, climate change, and the space program. As far as evolution is concerned, the survey has included the same questions for many, many years, so at least we can see whether there are trends over time. The most obvious trend is the large increase in the acceptance of evolution among the youngest age cohort.

I quite like this figure available from the Pew website. It shows the result of one of the evolution survey questions across the age groups, and in comparison with the membership of the American Association for the Advancement of Science.

Pew evolution survey results
Pew survey results in groups with different ages. Bars indicate percent who agree that humans have evolved over time, without regard to whether evolution was by natural processes or divine guidance.

Happily, 98% of AAAS members agree that humans have evolved over time. Only 65% of the general public agree with this statement—with the gap shown by the bar at the top of this graph. But 73% of young adults agree, compared to only 54% of adults over 65 years old in the survey.

Science knowledge and educational attainment both correlate with evolution acceptance in the survey, although neither correlates as strongly as religious attendance. Still I can’t help but think that much of the increase in acceptance of evolution in particular must go along with an increase in science knowledge by younger adults:

Pew evolution education results
Percent who agree that humans have evolved over time, either by natural processes or with divine guidance, in groups with different levels of overall science knowledge.

In this, I have a similar intuition as the geneticist Sean B. Carroll, who wrote about the results of the previous 2013 Pew survey: “Is America Evolving on Evolution?”:

[A]n additional possibility that we in the science education profession would like to entertain is that today’s young-adult generation has received the most and best information about evolution. The last few decades have been a golden era for evolutionary science, from discoveries in the fossil record to the mining of DNA to reveal how evolution works at the molecular level. And over the past decade the concerted efforts of various academic and scientific organizations have led to greater emphasis in textbooks and curricula on the central place of evolution in understanding life.

Human evolution has become more vivid and interesting within the last 10 years. The increasingly widespread genetic sampling of the world’s peoples has demonstrated the recent evolution of our species, including the effects of natural selection during the past 10,000 years since humans invented agriculture and civilization. Ancient DNA evidence has expanded our knowledge of our genetic relationships to even more ancient populations, including the Neandertals. Those classic representatives of “extinct” humans are not fully extinct but instead contributed to the ancestry of living people. We can find their genes within living people, so that every month seems to bring a new discovery of how evolution within modern humans has changed our biology.

Meanwhile, paleoanthropologists have made stunning fossil discoveries.

In the last ten years, we’ve had the first publications of two skeletons of Australopithecus sediba, of one skeleton of Ardipithecus ramidus, new material of Australopithecus afarensis including a partial skeleton, new postcranial remains and a new skull of Homo erectus from Dmanisi, new remains of Homo rudolfensis and Homo habilis, a raft of new hominin material from Sima de los Huesos, and of course this year, Homo naledi. And I’m leaving many out! Just this week, a new skull of Homo erectus.

The avalanche of new fossil discoveries means that since the turn of the century, if we look beyond modern human and Neandertal remains, the fossil record of hominin evolution has grown by nearly a third.

In the face of such a relentless pace of discovery, advocates of creationism are failing to persuade young people that scientists have no evidence of our evolution. The evidence is within all of our cells, in billions of base pairs of information.

Human evolution is no longer a theoretical exercise in many classrooms. Students are browsing actual genome data to see the parts that have evolved. Teachers are printing their own casts from fossil data. These technologies make it possible for us to take the rich legacy of our evolution and make it visible for students. The strength of the evidence has made human evolution more and more important to biology education.

That’s good news for biology. After many decades in which laboratory biologists might work without much use of applied evolution, applying genomics has made the principles of evolution have more and more important to biological discovery. Evolutionary ideas have greater practical importance in medicine, agriculture, and pharmaceutical development than ever before.

It’s also good news for society. The twenty-first century is the age of biology, a time during which humans will apply technology to change nature and ourselves. Understanding how human populations can evolve will be important for many of the questions we face in the future.

Nautilus is running a nice interview with Lee Berger about Homo naledi and other things: “The Man Who Used Facebook to Find an Extinct Human Species”.

From the interview:

We had senior anthropologists saying there was nothing left to find, we should stop teaching graduate students, you’re not going to find fossils in the soil anymore, they’re a declining resource. The richest fossil hominin site on the continent of Africa was less than a mile from the most explored locality on the continent. What does that tell you? That there must be extraordinary findings still.

Many interesting points in the interview about how paleoanthropology has been changing into a team-oriented science.

Did humans approach the southern tip of South America more than 18,000 years ago?

Tom Dillehay has for many years investigated the archaeological remains at Monte Verde, Chile. These provide some of the earliest evidence of human habitation in the Americas, with well-documented cultural material from 14,500 years ago (calibrated). This early occurrence was for the site designated Monte Verde II, and has abundant cultural evidence including at least two structures, human footprints, and many artifacts. It is widely regarded as the earliest most solid evidence of human habitation in South America.

Monte Verde is a bit south of the midpoint of Chile, around 1500 km from the southernmost tip of the South American mainland. While archaeologists still debate the earliest evidence of human activity in North America, it is clear that people must have gotten to North America early enough for them to end up near the southern end of South America more than 14,000 years ago.

In PLoS ONE this week, Dillehay and colleagues report on the results of further archaeological survey work at Monte Verde, this time at the locality designated Monte Verde I. Earlier investigation of Monte Verde I (MV-I) had uncovered what seemed to be very early artifacts, but not with contexts that Dillehay and colleagues had judged to be adequate to substantiate their surprisingly early dates. But during the new excavation work, they uncovered artifacts together with localized burned features in association with radiocarbon ages that, when calibrated, imply that some of the features are at least 18,500 years old.

This paragraph from the paper tells the overall story of their excavation goals:

Based on our previous findings at MV-I, which revealed possible cultural evidence laterally dispersed in deeper, sandy levels of the sandur plain, our recent work centered on spatially intermittent excavations and core drillings across a 500 m area between the MV-I and Chinchihuapi sites in search of additional scattered remains down to and below these levels (Fig 2). The result was the discovery of twelve small, discrete burned features directly associated with fragments of burned and unburned faunal remains, spherical and manuport stones, and human-knapped flakes dated by 14C and OSL means between at least 18,500 and 14,500 cal BP (Fig 3, S4 and S5 Figs). The features and associated lithics and bones are spatially limited within discrete lenses, averaging ~33 by 42 cm in spatial extent and ~1.0 to 2.8 cm in thickness. Only one excavated unit, measuring 5 by 5 m in size, contained more than one feature, indicating their widespread and intermittent dispersion across the study area. Although both horizontally and vertically discontinuous, these remains appear to represent ephemeral seasonal activities laterally spread across uneroded, slightly elevated surfaces (~0.5–0.8 m high) between small, narrow and shallow channels of a braided drainage system buried in the SCH-Fm (Fig 2 and S1 Fig).

The depiction here is an ancient streamside landscape upon which humans periodically camped with fires. The artifacts come from a fairly systematic survey with test pits and cores drilled at regular intervals, but the overall excavation area is quite limited. The archaeological features are single small burned areas with artifacts, and are apparently in good stratigraphic context.

Figure 3 from Dillehay et al 2015
Figure 3 from Dillehay et al. 2015, showing some of the archaeological features in situ. Original caption: Excavated archaeological features. a: in situ burned feature (No. 3) with charcoal, burned bone fragment (top arrow), and ash associated with an embedded, knapped flake of unsourced limestone (bottom arrow) in Unit 56, site MV-I. The thin, light yellowish orange (10YR 8/3) tephra lens has been removed in and around portions of the feature to expose the underlying gray sand. Inset shows a close-up of the flake after the overlying lens was brushed from its upper surface; b: in situ scorched bone fragment (left arrow) and a percussion flake of unsourced limestone (see S7f Fig) directly associated with a burned feature (No. 6, dotted line defines dark ashy areas) in Unit 56, site MV-I.

Obviously more intensive archaeological investigation of the site may yield more information. But one of the most interesting things about the Monte Verde sites is that the overall situation is not very exceptional. The peat-rich sediments do preserve organic material and evidence for burned features very well, but otherwise the Monte Verde I occurrences are just streamside terraces with small-scale fires and some artifacts. There should be hundreds of stream banks in South America with the same archaeological potential.

There is a school of thought that population growth should have been rapid once humans entered the virgin landscape of the Americas. With sufficiently rapid growth, the earliest signs of habitation should have been rapidly followed by continent-wide evidence of human populations. By this logic, if people first arrived more than 18,000 years ago, then we should find abundant evidence of them long before 16,000 years ago throughout the Americas.

But there may be a problem with the assumption of rapid growth. It assumes that people could rapidly change their strategies to spread into the very different ecologies of inland North and South America.

Maybe it wasn’t so easy for them to develop the technical and organizational innovations necessary to move into those continental ecosystems.

Sure, humans may have spread rapidly within one relatively uniform environment. Moving down the west coast of the Americas, using coastal resources, may have been such a relatively uniform environment. That’s the proposal that Jon Erlandson and colleagues call the “kelp highway” hypothesis. The idea is that kelp forest ecosystems in the coastal waters of northeastern Asia and North and South America are highly productive habitats. Once people had the ability to move within the coastal waters, they would have had a very easy time hopping among those coastal kelp forests.

A coastal maritime lifestyle of some kind must have existed in Southeast Asia well before 25,000 years ago, and enabled humans to occupy Okinawa by 18,000 years ago.

We know from the experiences of the Polynesians within the last 2000 years that island-hopping does not lead inexorably to success on every island. Some populations find favorable spots and can grow rapidly, if they come equipped with the right technology—and for Polynesians that technology includes domesticated plants and animals. But other occupations are purely ephemeral. For a maritime gathering and fishing population, without domesticated plants and animals, I’m not sure we should consider it surprising if they did not spread inland for a few thousand years after they had first established some coastal presence in South America.

But even more interesting is the possibility that they did spread inland. As Dillehay and colleagues briefly suggest in their introduction, the earliest sites in North America might even represent a back-migration of people from South America.

I have probably been too influenced by the assumption of logistic growth for early inhabitants of the Americas, an assumption that greatly downplays the challenges of developing effective cultural means of supporting population expansion into new ecologies. The spread of modern humans from Africa seems to have taken tens of thousands of years. As a time-averaged rate, modern humans moving across southern Asia, or into Europe, or across Australia, all seem to have been much slower than many of us have been assuming for the Americas.

Sure, many of the Old World cases involve possible encounters with other human populations, but we now know that mixture was an option. Modern humans encountering well-adapted Neandertals, mixing with them and witnessing Neandertal cultures, should have expanded into Europe vastly more rapidly than ancient mariners who found themselves on the coast of South America.

Imagine being born into a culture with a rich legacy of thousands of years of cultural knowledge of coastal resources, and instead of following the ways of your ancestors, deciding to strike into Patagonia, or into the Amazon rainforest, or across the Atacama Desert.

The archaeologists who consider the initial habitation of the Americas have long thought about these logistical issues, and there are no easy answers. But South America may well have been home to a Last Glacial Maximum human population, one that took 4000 years to spread across both North and South America.

The earliest cranial remains we have from both North and South America are surprisingly variable in comparison to later peoples of the Americas. Those skulls suggest the possibility that they represent populations that had already experienced a lot of diversifying evolution by genetic drift. An earlier initial spread of humans across South America might explain that appearance.

There is nothing in the genetics today to exclude this early habitation scenario, particularly considering that revisions to a lower autosomal mutation rate have not been fully factored into most of the work on early Americans. But there is an additional interesting possibility: Maybe the first inhabitants of the Americas were mostly genetically supplanted by later peoples. Could there have been an earlier habitation, now only present in living Native American peoples as traces of a “ghost population” that we haven’t yet identified?

If so, that scenario might explain the evidence for “deep divergence in Native American populations” that Rasmussen et al. (2014) found to predate the 12,600-year-old Clovis-associated Anzick-1 burial.

These are bound to be interesting times for people interested in the habitation of the Americas. What I think is exciting is that the ancient genomes provide so much evidence and context for relationships, but actually do not answer many of the key questions. We need these different sources of data to start working together with each other.

And we need more exploration. If MV-I is really that old, there should be hundreds of other sites out there to be found.

Oh, and one more thought on this paper—isn’t it cool that they published it open access? More and more interesting results are coming out in these journals where the images and text can be freely used and read by anyone in the world.


Dillehay TD, Ocampo C, Saavedra J, Sawakuchi AO, Vega RM, Pino M, et al. (2015) New Archaeological Evidence for an Early Human Presence at Monte Verde, Chile. PLoS ONE 10(11): e0141923. doi:10.1371/journal.pone.0141923

Erlandson, J. M., Graham, M. H., Bourque, B. J., Corbett, D., Estes, J. A., & Steneck, R. S. (2007). The kelp highway hypothesis: marine ecology, the coastal migration theory, and the peopling of the Americas. The Journal of Island and Coastal Archaeology, 2(2), 161-174. doi:10.1080/15564890701628612

Rasmussen, M., Anzick, S. L., Waters, M. R., Skoglund, P., DeGiorgio, M., Stafford Jr, T. W., ... & Willerslev, E. (2014). The genome of a Late Pleistocene human from a Clovis burial site in western Montana. Nature, 506(7487), 225-229. doi:10.1038/nature13025

From Roland Kays in The Conversation: “Yes, eastern coyotes are hybrids, but the ‘coywolf’ is not a thing”.

Coyotes in the Northeast are mostly (60%-84%) coyote, with lesser amounts of wolf (8%-25%) and dog (8%-11%). Start moving south or east and this mixture slowly changes. Virginia animals average more dog than wolf (85%:2%:13% coyote:wolf:dog) while coyotes from the Deep South had just a dash of wolf and dog genes mixed in (91%:4%:5% coyote:wolf:dog). Tests show that there are no animals that are just coyote and wolf (that is, a coywolf), and some eastern coyotes that have almost no wolf at all.
In other words, there is no single new genetic entity that should be considered a unique species. Instead, we are finding a large intermixing population of coyotes across the continent, with a smattering of noncoyote DNA mixed in to varying degrees along the eastern edge. The coywolf is not a thing.

Sounds a little like humans around 45,000 years ago…

Notable paper: Roffet-Salque, Mélanie et al. 2015. Widespread exploitation of the honeybee by early Neolithic farmers. Nature 527: 226–230. doi:10.1038/nature15757

Synopsis: Roffet-Salque and colleagues sampled organic residues from Neolithic-era pottery sherds across much of Europe, Turkey and the Levant. They found evidence of beeswax in a small fraction of the sherds spanning much of this area, including from early Linearbandkeramik (LBK) Neolithic sites in north-central Europe, from Çatalhöyük in Anatolia as early as 9000 years ago, and for the first time from North Africa at Gueldaman, Algeria.

Interesting because: The widespread exploitation of beeswax across Europe as soon as pottery was made by Neolithic farmers shows that the systematic exploitation of honey and wax was part of the strategies of early farmers. The use of wax is itself interesting, as it is not principally dietary but is to make other things: candles, cosmetics, unguents and medicines.

Looking further back: Honey foraging is a major activity by hunting and gathering people around the world today, and constitutes the second-largest caloric contribution of men to their societies (behind hunting). Honey is a part of chimpanzee diets today, and foraging for honey has given rise to some of the most complex toolkits used by chimpanzees in their native habitats. Honeyguides are birds that exploit beehives that are destroyed by other honey foraging animals, and some species seek out people to guide them to the bees. The evolution of this cooperation probably indicates a long association of these birds and earlier hominins. Personally, I also wonder how far back beeswax was used by humans for cosmetic or other purposes, particularly after the recent revelation that milk was used by MSA people as a pigment additive.

Another Denisovan from Denisova Cave

Denisova Cave is one of the most fascinating places in the story of human origins. The cave is in the northern wall of the Anuy River valley, within the Altai Krai region of Russia but very near the border with the Altai Republic.

The cave’s East Gallery, where excavators still work through part of the Altai summer, feels like a walk-in refrigerator. Archaeologists work inside wearing rugged jackets as they work slowly down through sediments at five degrees Celsius. In the winter, the temperature inside the cave is below freezing. The exceptional cold has made this cave into a time capsule for ancient genetic sequences.

The fossil hominins from the cave are mere scraps, some of them possibly leftovers from hyena meals. But before now three of the bones and teeth have produced genetic sequences, including two from a previously-unknown population that we now call the “Denisovans”.

Today there’s something new. Another hominin tooth, Denisova 8, has now yielded a partial low-coverage genome, and we can welcome it as a new member of the Denisovan population. Its DNA sequence shows that this tooth is older than the other Denisovan specimens, maybe 60,000 years older. And its DNA increases the diversity of the known sample of Denisovan specimens.

What’s important about this? And how does it fit into the already-complicated Denisova family?

Denisova tooth photo sprite

The Denisova cast of characters

The story of the hominins from this cave has been fast-moving for the last five years. If you haven’t been following closely, it probably seems like a telenovella. Many characters jump out for attention, each with a special, distinctive story.

The first and most famous of the specimens is the distal phalanx from a fifth finger, Denisova 3. When ancient mitochondrial DNA from Denisova 3 was first reported in 2010 by Johannes Krause and colleagues, they found it was very different from any known human or Neandertal—so different that they called this hominin the “X-Woman”.

Denisova pinky bone
A modern human juvenile distal fifth phalanx (left) compared to the Denisova fragment (right).

That name didn’t last long. Later in 2010, David Reich and colleagues followed up on the mtDNA by sequencing a low-coverage nuclear genome from the specimen. They showed that even though the mtDNA of Denisova 3 last shared an ancestor with us more than a million years ago, the nuclear genome is much less divergent. The “Denisovan” really did belong to a previously unknown population, very different from living humans. But it shared an ancestry with Neandertals within the last half million years.

At the same time, Reich and colleagues sampled the mtDNA from a second hominin specimen. This one, Denisova 4, is a third molar, or “wisdom tooth” representing a different individual from the finger bone. The mtDNA of this tooth is not identical to the finger, but it does belong to the same highly divergent lineage.

Denisova Cave
Denisova Cave

In 2012, Mattias Meyer and colleagues did more work on the Denisova 3 sample, ultimately yielding a very high coverage genome. From this genome it was possible to show that Denisova 3 came from a very endogamous population, one that had been small for a very long time.

Then, in 2014, Kay Prüfer and colleagues sequenced the DNA from another Denisova Cave specimen. This one, a toe bone, came from a similar archaeological level as the original Denisova 3 sequence, although it lay a bit lower in the sequence and therefore is probably a bit older. This toe bone produced another high-coverage genome, the second from the site. And unlike the other two Denisova specimens, this toe belonged to a Neandertal.

That’s quite a trick, telling Neandertals from previously unknown populations based on fingers and toes. Anthropologists would never attempt such a thing from the bones themselves. Or, maybe more accurately, some anthropologists would make extravagant claims from the finger and toe bones, and they would be fools.

But a whole genome gives millions of bits of evidence about relationships. The two Denisova mtDNA sequences stand apart from any known Neandertal or modern human. The nuclear genome of Denisova 3 differs from any known Neandertal or modern human at millions of base pairs. Still, that Denisova 3 genome has large stretches of DNA shared with the Neandertal toe bone, indicating that the Neandertal population interbred with the ancestors of the Denisovan individuals at some point.

And the Neandertal toe bone itself was highly inbred—as homozygous across its whole genome as people whose parents are half-siblings.

It’s a Paleolithic telenovella.

Denisova 8

Now, Susanna Sawyer and colleagues report this week in Proceedings of the National Academy of Sciences on the genetic data from a third Denisovan individual, Denisova 8. And they add nuclear genetic sequence, at low coverage, from the Denisova 4 tooth.

Denisova 8 is another tooth, again, likely a third molar. The two teeth are morphologically different from each other, with little indication that they reflect a single gene pool, except that the teeth are both quite large relative to most later Pleistocene humans.

One of the main genetic results is that the two molars (Denisova 4 and 8) clearly group with the original pinky (Denisova 3) genome in their nuclear and mitochondrial DNA. However, Denisova 8 is an outlier to the other two, more different from them in mtDNA sequence than any Neandertals are from each other, and somewhat more different in the nuclear genome than would be typical for Neandertals.

Part of that difference is due to the age of Denisova 8. Its mtDNA branch is shorter than the other two specimens. It is missing evolutionary history that is present in Denisova 3 and Denisova 4, enough to suggest that Denisova 8 lived some 60,000 years earlier in time than those other two. The nuclear genome is consistent with such an age estimate, although the low-coverage sequence is not really sufficient to give an any accuracy on its own. So one reason why Denisova 8 increases the diversity of the Denisova sample is that it lacks tens of thousands of years of genetic drift that was shared in the ancestry of the later specimens.

Still, when we look at the variation in the two low-coverage genomes from Denisova 4 and 8 in comparison with the high-coverage Denisova 3 genome, all the Denisovans together seem to have been just a bit less inbred than Neandertals. Sawyer and colleagues provide a measure of genetic difference between genomes that is scaled to the difference between humans and chimpanzees. To accomplish this measurement, they take the preserved sequence that overlaps between two low-coverage genomes (or high-coverage genomes), examine every site within the overlapping sequence where either of the two hominin sequences differs from the chimpanzee genome, and examine the fraction of times that the two hominin sequences are discrepant from each other. This gives a measure of difference between the two genomes relative to their difference from chimpanzees.

For the Denisovan genomes, this difference averages 2.9 percent. For Neandertal genomes compared to each other, the difference averages 2.5 percent. So Denisovans are a bit more diverse than Neandertals by this measure.

For human populations, the differences range between 4.2 and 9.5 percent. The least diverse pair of individuals are Karitiana people from South America. This highly endogamous human group is still half again as diverse as Denisovans.

Both the Denisovans and Neandertals were very endogamous relative to today’s people, even just comparing to Europeans today. That endogamy must reflect something about the ancient population structure of these archaic people. One consequence of the endogamy — as two preprints by Graham Coop’s lab and Rasmus Nielsen’s lab show — is that Neandertals may have carried a load of slightly deleterious genetic variants. When the Neandertals and Denisovans mated with the more diverse population that dispersed from Africa in the Late Pleistocene, some of these deleterious genetic variants may have been slowly weeded out of the modern human population by purifying selection.

Denisova 8 and Denisova 4 compared to Neandertal and human
Figure 1 from Sawyer and colleagues 2015. Original caption: "Fig. 1. Occlusal surfaces of the Denisova 4 and Denisova 8 molars and third molars of a Neandertal and a present-day European."

How big were the Denisovans’ teeth?

We don’t know very much about what Denisovans may have looked like, with only a fingertip and two wisdom teeth to go on. So it’s tempting to take these slim data and make something more out of them than we probably should.

With two teeth, we at least can measure their sizes. They’re big. Bigger than most living humans, bigger than any Neandertals, as big as some third molars of Australopithecus.

We can’t make too much out of the large size of these third molars, because similarly large teeth do occasionally occur among Upper Paleolithic people. The supplementary material to Sawyer and colleagues’ paper gives a brief review, noting that the Oase 2 skull has such a large third molar, as does a tooth from a partial juvenile dentition of a probable Neandertal from Obi-Rakhmat, Uzbekistan.

Two Late Pleistocene specimens are comparably large in size, the M3s of the early Upper Paleolithic modern human Oase 2 and the M2/3 of Obi-Rakhmat 1 (14, 15). Oase 2 does not show large extra cusps, but instead strong crenulation (16). Obi-Rakhmat shows a large extra cusp, but mesially, not distally (Main text, Figure 1), and a large number of accessory cusps possibly due to gemination (17).

The Oase 2 specimen is the earliest modern human known from Europe. This is not the same individual as the Oase 1 mandible, but it may be relevant that the Oase 1 specimen had more Neandertal ancestry than any known living person, and that Neandertal ancestry was quite recent, possibly within four generations.

Is it odd that we have large third molars in some individuals in Eastern Europe and Central Asia, including the Denisovans? I think to answer this we will need a larger sample of fossil humans from other places. There is at least one Chinese third molar specimen from Xujiayao that approaches the size of these large teeth. Perhaps the Chinese specimen is a Denisovan, or maybe this is one morphological extreme that is distributed among Late Pleistocene populations.

The thing is, if there was one piece of morphological evidence I would throw away and pay no attention to above all others, it’s the morphology of the third molar. It is just enormously variable among living humans and living primates. I wouldn’t trust it to tell us about relationships of hominin groups.

OK, so maybe I would trust the third molar above the pinky finger.

UPDATE (2015-11-18): I’ve known about these results for a long time, as the authors kindly shared them with me. But I forgot that the results were also previously a part of the documentary “Sex in the Stone Age”, until I found a letter from a reader in my archives:

I just watched the National Geographic documentary "Sex in the Stone Age" and was surprised by the reference to the discovery of a 2nd Denisovan tooth, one whose mitochondrial DNA was distinct enough from that of the mtDNA in the finger and original tooth to indicate that the Denisovan population had as much genetic diversity as H. Sapiens currently has today. This is interesting, since if I recall correctly, Neanderthals had low levels of genetic diversity, with evidence of replacement of their western European population by an Eastern population. This perhaps indicates that the Denisoans had a larger population than that of the Neanderthals. I don't recall reading about this find on your website or anywhere else.

At the time I gave some context to that report from the documentary but had little more to say.

What’s interesting about this week’s publication is that the nuclear genome and mtDNA have contrasting patterns. The genetic variation within Denisovans is best assessed across the nuclear genome, and they are low in a similar way as Neandertals.

Still, the Neandertal sample covers a very large geographic area, from Spain to the Altai. Our intuition might lead us to expect that extensive range to encompass more variation than three Denisovan individuals from a single site. But the Denisovan genomes cover a longer time than the low-coverage Neandertal genomes, and across such a time they may actually sample very different populations. If we had a Neandertal low-coverage genome as early as Denisova 8, their variation might look different.

To me, the core observation is that both these populations were highly endogamous compared to any living human groups. I’ll have a bit more to say about the mtDNA discrepancy shortly…


Sawyer S, Renaud G, Viola B, Hublin J-J, Gansauge M-T, Shunkov MV, Derevianko AP, Prüfer K, Kelso J, Pääbo S. 2015. Nuclear and mitochondrial DNA sequences from two Denisovan individuals. Proceedings of the National Academy of Sciences, USA (early edition) doi:10.1073/pnas.1519905112

More on Denisova

“The Altai Neandertal”. A high-coverage genome from a new Denisova Cave specimen provides new details about ancient population mixture and modern human origins.

“The Denisova-Sima de los Huesos connection”. The most ancient human mtDNA yet recovered leads to a result that surprises some, not me.

“Hunting the Denisovan belt”. Some suggest that Denisovans mixed with humans in Wallacea. I’m skeptical.

Just some notes from a designer at Facebook worth sharing:

A successful team moves fast, and they do often leave a trail of failed experiments in their wake. But here’s the deal: quick experiments, even failed ones, are the best way to learn what works and what doesn’t. So when the first two things don’t work, try two more. And when that doesn’t work, try two more. And if those fail too, perhaps it’s time to approach something from another angle, but give it another two, and then two more.
Get comfortable with the fact that it might take ten tries to get a positive result from your experiment. It might take twenty or more.
Over the course of days and months and years, you’ll find that you’re no longer solely working on assumptions, but that both your knowledge base and the strength of your hypotheses are growing.

From Jasmine Friedl, “What I’ve Learned Designing Small Things at Facebook”. Seems relevant to science.

Some nice coverage of Svante Pääbo in the Washington Post on the occasion of him winning one of the 3-million dollar “Breakthrough Awards” last week: “3 science ideas you’ll be hearing more about in 2016”:

The traditional way to learn about extinct species is by digging up old bones or, in the case of early humans, examining stone tools they might have left behind. But what if you could learn about them by studying their DNA? Svante Pääbo of the Max Planck Institute for Evolutionary Anthropology won a Breakthrough Prize in Life Sciences for pioneering the sequencing of ancient DNA and ancient genomes, with a focus on understanding the origins of modern humans, their relationships to extinct relatives such as Neanderthals, and the evolution of human populations and traits.

How 3D printing fossils will change the way we look at human evolution

In September our team announced the discovery of the new species, Homo naledi. The species comes from the Dinaledi Chamber, deep within the Rising Star cave system in South Africa, where our team has uncovered more than 1500 fossil hominin specimens.

The team took the unprecedented step of releasing three-dimensional scan data for the key fossils at the same time we published their initial description. These models are available to the public from the MorphoSource site, an archive of 3D data sponsored by Duke University. Over the past two months, teachers, researchers and others around the world have downloaded the models more than 5000 times. Many of them are using 3D printers to create their own physical models of the fossils.

This move has gotten a lot of attention around the world. Human origins is a field in which the people who study fossils often seem very secretive, and many fossils are not available to other researchers. Now for the first time, people are downloading and handling fossils at the moment that they are published. People around the world are wondering, how will open access to data change the study of human origins?

Lucas Brouwers from the Dutch publication NRC Handelsblad interviewed me last month about why the Rising Star team took this open access approach. The interview was part of his article, “Voorouder komt zo uit de printer”. Brouwers kindly has allowed me to share all the questions that he wrote to me, along with the answers I wrote back. I’ve put a bit more into some of these answers before posting, because the question-and-answer was a great occasion to write about access to fossil data and how we have approached open access on the Rising Star project.

Q: An obvious question perhaps, but could you explain why the naledi-team decided to publish the 3D-files? What was your role in this decision? Was everyone on the team convinced this was a good idea from the start?

A: For many years I’ve been a vocal advocate for more open data access in paleoanthropology, so obviously I have my own perspective as a scientist and I don’t speak for everyone all the time.

Our team includes more than 60 scientists, and scientific decisions involve discussion with of a lot of people. The Dinaledi fossils are in the care of the Evolutionary Studies Institute (ESI) at the University of Witwatersrand, and most of us are associates of the University. Everything we do is shaped by the mission of the University and by South African law.

Steven Churchill was the key player in putting together a collaboration between Duke University, which hosts the MorphoSource archive, and Wits. Doug Boyer at Duke has also been tremendously helpful in facilitating our team uploading the data. The data themselves have been generated by more than a dozen of the team members, and they’ve uploaded these to MorphoSource as an archive of the work.

The move to more open access really started with Lee Berger. Lee had found the Malapa site in 2008, and assembled a team that described Australopithecus sediba from those fossil remains. He established a policy of open access to the fossils by researchers, which became a huge part of the Malapa project’s success.

So when we brought this large team of scientists to work on the anatomy of the Dinaledi fossils, Lee and the rest of us felt strongly that data generated during the project would be available to the community. We saw it as one of the most innovative things we could contribute, and it also solves two practical problems.

One of the practical challenges is the condition of the fossils. The Dinaledi fossils are very fragile and subjecting them to traditional molding would damage them. So we knew from the beginning that high-resolution 3D printing was going to be necessary to make copies of the fossils.

The other practical challenge is the collaboration among a huge team.

Providing research access to digital data is not free, and it’s not easy—somebody has to prepare those data in a form that can be useful to other scientists. You don’t want to have to do that in a different, non-standard way for every new investigator. And if the digital data are being curated by a dozen different individuals, that creates a series of bottlenecks for other people on the team who need the data. A project you could do in an afternoon suddenly takes weeks while you wait for the email chain.

The way to share data effectively across a large team is to establish an archive where the data are interoperable, in a standard format with metadata included. Once that culture of data access is in place, it becomes very easy to bring new people into the collaboration.

Q: According to your impression, how popular is 3D-printing right now amongst academics and educators? Have you had many requests for help with printing?

A: We are seeing an explosion of 3D printing. I’ve been so amazed to see thousands of people downloading these files and starting to produce prints. I showed up in Milwaukee the other day to give a lecture, carrying my bag full of fossil prints. When I got there, they had already printed up a whole set of their own! Homo naledi is everywhere—it’s like the tribbles!

We can’t do support, there are such a huge range of printers and equipment. We have heard from some people who were unable to print our surface models because the data were too high-resolution for their software or printer. One person even volunteered to make the data lower-resolution for us! Fortunately, if anyone has trouble with these, free software like MeshLab can load these files and reduce their quality to a smaller file size.

It’s been really cool to see people painting the models and customizing them, making them look like real fossils.

Q: Do you think releasing 3D-data will become common practice in anthropology? In the past, some paleoanthropologists have been very protective of ‘their’ fossils, only releasing photos and descriptions in tidbits and morsels. Do you feel this attitude is changing?

A: There is this myth that if you provide access to digital data, you are somehow losing the value of the fossils, that you should somehow make scientists pay to access the digital data. We have to learn to live in the 21st century.

The experience in South Africa at the ESI is that providing access to new fossils like Australopithecus sediba and now Homo naledi has taken the institution from a handful of scientific visitors per year up to hundreds. Access to the fossils increases scientific visits, increases scientific output, and creates long-term investment in the institutions that curate them.

I came from genetics, have done most of my research using open data archives. There is no significant research in human genetics today that does not use open access data in some way. These archives enable us to train our students on the newest research methods, and allow scientists to develop pilot data to explore hypotheses, that vastly strengthen their ability to get funding for those new ideas.

People would be surprised that many hominin fossils from other sites have been scanned lots of times, but scientists have signed agreements that prevent them from providing access to their own data. Sometimes a researcher will share data “under the table”, in violation of those agreements. That’s very negative because it means that people give data to their friends, or to people with something to trade, instead of making the data accessible to everyone. It becomes an “old boys’ network” of data.

This is one reason why we need to emphasize “access” and not “sharing”—many researchers will say, “well, I share my data”, but they’re sharing very selectively, it’s like a black market. And then new researchers come and scan the same fossils again and again. It’s such a waste of time and resources, and it’s so unnecessary.

We want to move the expectations of the field toward broader collaboration. Great ideas do not only come from one place, they take many people looking at the data.

When I think of the students who are entering paleoanthropology, they are going to be my scientific peers five, ten years from now. Why in the world would I not want them to have the best, most current data from our fieldwork? I’m struggling to understand the mindset of scientists who think that students shouldn’t see the fossils.

Screenshot from AfricanFossils.org
Screenshot of a page from the AfricanFossils.org site, showing a jawbone of the species Homo rudolfensis for viewing or download.

Q: For some museums, especially in Africa, selling high quality fossil casts to academics is one major source of income. Do think 3D-printing, in the future, will undercut this source of revenue?

A: Obviously this is a big concern for our team, because ESI has its own casting program and supplies casts internationally to many institutions.

Let’s be clear about this. Right now two of the countries that are making the most data available to the public online—South Africa and Kenya—are also the ones that are selling the most casts from their traditional casting programs. I want to draw attention not only to our own data archive, but also the AfricanFossils.org initiative from Kenya, led by Louise Leakey, and the archive of CT data from the Ditsong Museum of Natural History in South Africa. AfricanFossils.org is directed toward educational use casts, while both the Homo naledi surface models and the Ditsong CT data are research data, but both kinds of data distribution pose the same issue for traditional cast production.

What these examples show is that data access is not a zero-sum calculation. When the data are available online, many more people will have access to copies of the fossils. That drives much more interest in building collections of high-quality fossil reproductions in a broader range of institutions, including schools that never had them before.

That doesn’t mean we shouldn’t be concerned about the casting programs. We want to find ways to leverage interest into longer-term relationships for teachers and the public, so they can support the work that brings casts into South African schools, support the production of high-quality digital scans, and support new exploration.

Now I’m going to editorialize, speaking for myself only.

There are institutions that never sell casts, even to academics for research and teaching. Some countries don’t even allow casts to be exported. These institutions do not allow researchers to share digital data. “Prestige” journals permit researchers to publish without putting their data in an archive. So no one can have copies of the fossils or data from the fossils except for the original investigators.

Right now, American and European companies are selling artistic sculptures of the fossils that are curated at these institutions. Universities and schools buy those sculptural recreations, and most of them don’t even realize that these are sculptures and not based on real data. That investment doesn’t go back to the source institutions.

Try to envision NASA saying that no one can have data from the Hubble telescope, because they might make some money selling photographic prints. But then NASA doesn’t even print them! So private companies hire artists to paint space scenes, and universities buy these paintings so that students to see what distant galaxies look like.

You just can’t imagine this happening in a real science, right? But that’s what happens every day in paleoanthropology.

If “prestige” journals were to start requiring scientists to submit their digital data to an open archive, as a condition of publication, open access to data would happen overnight. How do I know this? Look at human genetics. The major countries that restrict access to digital data from fossils, right now, today, I can freely download human genetic data from all of those countries. Geneticists don’t withhold their data for years until they have published many papers—in fact large projects provide open access data before they publish their first paper.

There is no government in the world that would say that fossils are more precious than people’s blood. But the genes are in open access archives and the fossils are highly restricted.

So what’s the difference?

In human genetics everyone recognizes that open data access is necessary to have a real science. The human genetics community has developed for more than 20 years on the basis of public data archives and open data access, and journals require the deposit of data in open access archives when they publish the papers.

In paleoanthropology, we don’t publish data in open access archives, and we don’t require people to provide data when they publish.

Closed access science may not be flawed, but how can we be sure? How long could a NASA contractor have hidden the flawed Hubble mirror if no one could inspect the data but their own technicians?

I want to look at all the positive benefits of open data that go beyond the science. Imagine if you tell any company—Apple, Microsoft, Google—that teachers all over the world are begging to put the company’s most innovative product in the hands of every schoolchild. That’s what South Africa is doing right now with these fossils.

Think about the long-term effects if every kid on Earth could go to school and hold the most precious objects from Ethiopia, from Kenya, Tanzania and South Africa.

Those kids are going to feel that heritage in a way that has never been possible before, and some of them are going to remember those objects, dream about visiting the place where all humankind originated. We’re talking about shaping the future, and the cost of this is just putting data online to let teachers create for themselves.

Foot printed on template
A foot of Homo naledi, ready for assembly. Photo credit: John Hawks CC-BY-NC-ND

Q: Some anthropologists I spoke to were concerned with the quality of the scans published on MorphoSource (some questioned whether they’re research-grade).

A: We don’t have the time or personnel to provide any lower quality than we are using for the research. One of the great things about a proper archive like MorphoSource is that the metadata on the scanner type and the person who did the scan can be there archived with the data. Anyone can check that out.

The surface models have higher measurement fidelity than traditional casts because silicone molds used in casting can deform, and the epoxy used to make research-grade casts will shrink. So for research purposes that involve accurate shape and measurement, the surface models are much better than traditional methods. But they do not capture surface detail nearly as well as silicone molds, so the models cannot be used for research that examines microscopic details of the bone surface.

Of course the quality of prints depend on the printer. But we use the scan models for research, and prints are mostly useful for teaching and demonstration.

Some people may be confusing surface scans with microCT, which has much higher resolution and shows internal structure, and is superior for many small objects such as teeth and hand bones. We are undertaking microCT scanning of some of the Dinaledi sample, and this will take many months of work for our team.

Q: Why have only 86 scans been released so far?

A: We worked hard before publication to release the essential scans that document the anatomy of Homo naledi, especially the holotype and paratype specimens described in our eLife paper. Those are the key specimens that we’ve described, and for most people they are the first priority for examination and printing.

More than a dozen team members have helped upload models, and they’ve also added several good examples of variation—the femora for instance provide a great record of how the Dinaledi sample varies.

As we publish descriptions of other fossils from the collection, including some of the more fragmentary specimens, we’ll be adding more. With many elements, especially small ones like teeth, ankle and hand bones, it may just make sense to wait until we have microCT data.

Q: One more question, if you permit: could you elaborate a bit more on the decision to publish in eLife? Surely could’ve been a cover article for Nature or Science? Did you want to set an example by publishing in an OA-journal?

A: We have two main considerations—we have a responsibility to our colleagues to publish the science in the best way available to us. And we have a responsibility to the people of South Africa, who have funded and supported the work, and who are guardians of the heritage that we have uncovered.

Among the senior members of the team, we have published more than 20 papers in Science and Nature, so we had been through that process many times, we knew what that would be like.

When you look at those journals just this year in paleoanthropology, you see they are publishing articles that describe only one or two specimens. We really tried but we just could not work out a way to publish Homo naledi in one of those journals with the detail to support the science and enable colleagues to understand what we have done.

On the positive side, eLife was really a pleasure to work with. They gave us all the space we needed, high-resolution photos of all the key specimens integrated with the article, and the review process was thorough and professional. The result is two papers that—–if you print them out—–total 72 pages, with 34 color figures and additional supplementary data online.

From the science side, our colleagues worldwide have given us a tremendously positive response.

From the heritage side, we have been totally blown away by the excitement people have for the open access. Some of our colleagues used to say that people aren’t really interested in reading scientific articles. We’ve shown that’s totally wrong. More than 230,000 people worldwide have viewed the scientific paper describing Homo naledi, and 81,000 have viewed the geology paper describing its context. I’ve never encountered anything like this in our field. When you provide papers with the right level of detail and rich illustrations, people will make the effort to engage with the science.

The Vice Chancellor of Wits University, Adam Habib, spoke at the announcement of Homo naledi last month, and he really said something that I think resonates widely:

We often talk about science as having no boundaries, but in our world scientific knowledge has become commodified, and too often, what should be the bequest of the world, the bequest of a common humanity, is locked up under paywalls that postgraduate students and researchers cannot get access to. So what we did when we made this discovery, was we put cameras in the cave, and we streamed it live from day one.
"We partnered with eLIFE, an open access journal, to make sure that the discovery was available to all of humanity. And what we did in that practice, is create the first elements of a common global academy….We are not simply going to be beneficiaries of open access, but we are going to be contributors to open access, to the knowledge of a common humanity."

I’ve been sharing that quote a lot, because it expresses something really central. South Africa is stepping forward to lead in the area of open access, and this is not just one team of scientists, we have tremendous support from every level of the scientific enterprise and government. Our team has gotten a lot of attention for this because people aren’t used to seeing significant new discoveries published in such an open access way. But this is the future.

Paleoanthropology is about uncovering the history that all humans share. People are curious about it, they want to encounter that history, they want to explore their origins. As scientists we have the tremendous privilege of discovery, and it gives us a responsibility to enable people to understand our common past.