Saturday, January 28, 2012

The Sub-Saharan African Dental Complex

Map of Nigeria, showing the location of the Iwo Eleru rock shelter and the Iwo Eleru skulls. (Harvati et al., 2011)

Sub-Saharan Africans have an unusual complex of dental features:

[…] compared to other world populations, Africans south of the Sahara Desert are distinct dentally — especially in their expression of nine high- and two low-frequency morphological features. This suite of traits was termed the “Sub-Saharan African Dental Complex” (SSADC); it includes the world’s highest occurrences of Bushman canine, two-rooted UP1, UM1 Carabelli’s trait, three-rooted UM2, LM2 Y-groove, LM1 cusp 7, LP1 Tom’s root, two-rooted LM2, and UM3 presence, and among the lowest occurrences of UI1 double shoveling and UM1 enamel extension. (Irish, 2011)

The two low-frequency traits appear to be “derived.” They seem to have developed in sub-Saharan Africa after modern humans began to spread to other continents. The other traits, however, are ancestral:

[…] the same nine high-frequency traits are also ubiquitous in the dentitions of extinct hominids and many extinct and extant non-human primates

[…] The presence and, indeed, prevalence (see next section), of high-frequency Sub-Saharan dental traits in fossil and recent hominoids—some of which are probably direct ancestors of modern humans, suggests they have been around for a long time. (Irish, 1998, pp. 87-88)

In addition to these traits, Irish (1998) mentions a low-frequency trait that seems likewise ancestral and specific to sub-Saharan Africans:

A final ancestral feature found with some regularity in Sub-Saharan Africans, relative to other modern groups, is polydontia. Numerous cases of extra incisors, third premolars, and fourth molars have been noted […] In one study (Watters, 1962) the incidence reached 2.5-3% in several hundred west Africans; many of the extra teeth were fully formed and erupted. “Typical” mammals exhibit three incisors and four premolars (Jordan et al., 1992). Polydontia is also found in living non-human primates […] (Irish, 1998, p. 88)

Why are these ancestral traits much more common in sub-Saharan Africans than in other humans? There are several possible reasons. One is that non-Africans began as a small founder group and thus lost much of the dental variability that still characterizes Africans. Another reason might be that natural selection favored new forms of dentition outside Africa, perhaps as a response to new food sources or new ways of preparing food.

But there’s a third possible reason: archaic admixture. Just as modern humans mixed to some extent with Neanderthals in Europe and Denisovans in Asia, perhaps there was also mixture with archaic hominins in Africa, and perhaps this admixture introduced archaic dental features into present-day Africans.

But how could present-day Africans have archaic admixture? If modern humans originated in Africa, wouldn’t they have encountered archaic humans only in Europe and Asia?

Well, at first, modern humans did not occupy all of Africa. They were initially a small population somewhere in East Africa. Then, around 80,000 years ago, this population began to expand northward and eventually into Eurasia (Watson et al., 1997). Meanwhile, the same expansion was taking modern humans westward and southward into other parts of Africa.

Just whom exactly did these modern humans encounter during their expansion within Africa? Initially, they probably met hominins who looked the same but still lacked some of the mental rewiring that gave modern humans a competitive edge. These “almost-moderns” account for about 13% of the current sub-Saharan gene pool and may have been related to the Skhul-Qafzeh hominins who occupied the Middle East 120,000 to 80,000 years ago (Watson et al., 1997).

As modern humans spread further west and south within Africa, they encountered much more archaic hominins, and perhaps even lingering Homo erectus groups. About 2% of the modern African genome comes from an archaic population that split from ancestral modern humans some 700,000 years ago. This admixture is dated to about 35,000 years ago and may have occurred in Central Africa, since the level of admixture is highest in pygmy groups from that region (Hammer et al., 2011).

A more tangible sign of admixture is visible in a skull retrieved from the Iwo Eleru rock shelter, in southwestern Nigeria, and dated to approximately 16,300 BP:

Our analysis indicates that Iwo Eleru possesses neurocranial morphology intermediate in shape between archaic hominins (Neanderthals and Homo erectus) and modern humans. This morphology is outside the range of modern human variability in the PCA and CVA analyses, and is most similar to that shown by LPA individuals from Africa and the early anatomically modern specimens from Skhul and Qafzeh.

[… ] the transition to anatomical modernity in Africa was more complicated than previously thought, with late survival of “archaic” features and possibly deep population substructure in Africa during this time.
(Harvati et al., 2011)

Then there is the Broken Hill skull, found near Kabwe, Zambia and dated to 110,000 BP (Bada et al., 1974). It looks for all the world like a Homo erectus. Textbooks generally try to raise it to Homo sapiens status or argue for an earlier dating. Recently, a late dating has been confirmed by Stringer (2011).

Interestingly, when Irish (2011) compared dentitions from west, central, east, and south Africa, ranging in age from the late Pleistocene to the mid-1950s, the early Holocene Kenyans and Tanzanians were the sample that had the fewest ancestral traits of the Sub-Saharan African Dental Complex (SSADC). In other words, the SSADC seems to have been least present in the “homeland” of modern humans (East Africa) and more present farther west and south.

Given the high level of archaic admixture in sub-Saharan Africans, we may have to revise downwards the estimate of 1 to 4% Neanderthal admixture in Eurasians. Yes, Eurasians are closer than sub-Saharan Africans to the Neanderthal genome. But is this discrepancy solely due to Neanderthal admixture in Eurasians? Could it also be due to Sub-Saharan Africans becoming further removed from the Neanderthal genome through admixture with other archaic groups?

The past may be a stranger country than previously thought. When farming villages began to form in the Middle East, there may still have been archaic hominins roaming over parts of western and southern Africa.


Bada, J.L., R.A. Schroeder, R. Protsch, & R. Berger. (1974). Concordance of Collagen-Based Radiocarbon and Aspartic-Acid Racemization Ages, Proceedings of the National Academy of Sciences (USA), 71, 914-917.

Hammer, M.F., A.E. Woerner, F.L. Mendez, J.C. Watkins, and J.D. Wall. (2011). Genetic evidence for archaic admixture in Africa, Proceedings of the National Academy of Science (USA), 108, 15123-15128,

Irish, J.D. (2011). Afridonty: the “Sub-Saharan African Dental Complex” revisited, American Journal of Physical Anthropology, 144(supp. 52), 174

Irish, J.D. (1998). Ancestral dental traits in recent Sub-Saharan Africans and the origins of modern humans, Journal of Human Evolution, 34, 81-98.

Harvati, K., C. Stringer, R. Grün, M. Aubert, P. Allsworth-Jones, C.A. Folorunso. (2011). The Later Stone Age Calvaria from Iwo Eleru, Nigeria: Morphology and Chronology. PLoS ONE 6(9): e24024. doi:10.1371/journal.pone.0024024

Stringer, C. (2011). The chronological and evolutionary position of the Broken Hill cranium. American Journal of Physical Anthropology, 144(supp. 52), 287

Watson, E., P. Forster, M. Richards, and H-J. Bandelt. (1997). Mitochondrial footprints of human expansions in Africa, American Journal of Human Genetics, 61, 691-704.


Matt said...

Could it also be due to Sub-Saharan Africans becoming further removed from the Neanderthal genome through admixture with other archaic groups?

We, of course, would need some reason to think the early Homo Sapiens who expanded into Eurasia were Neanderthal shifted relative to other populations in Africa (the archaic groups) across their genome.

Would we really expect an East African African hominin to be more Neanderthal shifted than a South or West African one? That would require the East African hominin to either

- have been part of a breeding cline with Neanderthal populations, and for this to have been more the case than with the South and West African hominins. (note this gets us back to Neanderthal mixture again, just through intermediate species in North Africa and the Middle East)
- share a more recent common ancestor with Neanderthal than with the South and West African hominins

I find these totally unlikely propositions myself, but the fossil record is rather fragmentary.

Matt said...

To expand (probably unnecessarily) if it isn't already clear:

Lets say you have population A.

A splits so it leaves descendent B where it is, descendent C to the north and descendents D and E to the west and south.

All these populations drift from one another and there is an even rate of drift in each group. Theres no reason to assume any of these populations would be closer from drift than the others.

Let's say B mixes with D and E - would it become less close to C than B is? No, because B is mixing with groups which is no more or less dissimilar to C than B itself is.

Now, if B and C are either part of a breeding cline to the exclusion of D and E, or B and C share a more recent common ancestor, then for B to mixed with D and E will take it further from C.

Spandrell said...

13% of negroid genes are non-sapiens?
Is this certain?

Aaronovitch said...

I found this on the PBS website:

RACE - The Power of an Illusion

There's more genetic diversity within a group of chimps on a single hillside in Gomba than in the entire human species.

How is this possible? Is this just a noble lie? Or is this some sort of semantic shell game that is technically "true" while conveying incorrect information to people who aren't in on the joke?

Jason Malloy said...

Well since the extra teeth are a derived feature, it seems more likely that we've always had them, and the human diaspora first lost them in Eurasia through natural selection. And I find Rushton's case here compelling:

"Fig. 4 illustrates that with increasing brain size there is decreasing prognathism and a flatter face (Cranial and Mandibular Traits 9, 11, 19). Muscles are no longer available to hold up a heavy forward jutting jaw. Since smaller temporalis muscles cannot close as large a jaw, jaw size was reduced. East Asians average a flatter face than do Whites, who average a flatter face than do Blacks. Consequently, there is less room for teeth, resulting in smaller teeth, shorter roots, and fewer teeth (Mandibular Traits 16–18, 23)."

Anonymous said...

Prognathism may be due to gravity.

Anonymous said...

Since smaller temporalis muscles cannot close as large a jaw, jaw size was reduced. East Asians average a flatter face than do Whites, who average a flatter face than do Blacks.

East Asians supposedly have a more prognathic (forward jutting from the face) and absolutely anterior-posterior larger jaw than Whites.

According to the physical anthropology.

See the index of prognathism on (page 112-113 in the document). East Asians are a little more prognatheous than Whites, but not much, and have the higher basion-prosthion lengths than Whites (i.e. larger jaws).

East Asians relative to Whites (and I think Blacks as well) also have a higher face, and consequently have the brain less above the face and more behind the face, like our recent ancestors (although I think recent Europeans have this quality compared to paleolithic Europeans as well, despite having a much more lightly built skull).

East Asians do have a flatter midface of course.

M said...

@ Aaronovitch,

I complained to PBS a few years ago about "Race the Power of an Illusion". I received an email from the producer Larry Adelman attaching a document explaining their position. He also queried why it would matter if race didn't exist.

Sean said...

It seems that that Europeans may have significantly less archaic ancestry than other humans.

Peter Frost said...


Yes, it is reasonable to expect that modern humans were genetically closer to Neanderthals than to the unidentified archaic population in Africa.

There are two reasons:

1. The estimated time of divergence between modern humans and Neanderthals is about 500,000 years ago. According to Hammer et al., the estimated time of divergence between modern humans and the unidentified "Afro-archaics" is about 700,000 years ago.

2. Neanderthals were present in the Levant between 80,000 and 50,000 years ago. Gene flow from those Neanderthals should have more strongly impacted modern humans in East Africa than Afro-archaics in western and southern Africa.


The total figure may be higher than 13%. That's the estimate for admixture with a pre-sapiens population that underwent an expansion c. 111,000 years ago. They were probably related to the Skhul-Qafzeh hominins who occupied the Middle East 120,000 to 80,000 years ago.


It's true. It just doesn't mean what most people think it means.

Keep in mind that most genetic variability is of little or no adaptive value. If a population stays put in one place, like those chimps, it will accumulate more and more "junk" variability. If, however, it buds off and adapts to a new environment, it will lose much of that variability, while nonetheless undergoing considerable physical change.

In short, there is only a weak correlation between total genetic variability and actual anatomical, physiological, and behavioral variability, i.e., the kind that's actually expressed and does stuff.


I don't exclude a natural selection explanation. On the other hand, the incidence of 2.5 - 3% looks to me like a relatively recent admixture effect.


I remember reading someone who attributed prognathism to diet. Gravity is a new one for me.


Sending letters do have an impact. The problem is that most people don't react. If it's on TV or in the newspaper, they think it must be right.


Archaic admixture seems to be highest in sub-Saharan Africa and Melanesia.

Matt said...

"1. The estimated time of divergence between modern humans and Neanderthals is about 500,000 years ago. According to Hammer et al., the estimated time of divergence between modern humans and the unidentified "Afro-archaics" is about 700,000 years ago."

Thanks Peter.

I think other, deeper divergence times for Neanderthal are still in currency, and appear to me to be more common based on the recent genetic data, the Neanderthal genome (rather than the cranial data):

"Assuming an average DNA divergence of 6.5 million years between the human and chimpanzee genomes, this results in a point estimate for the average divergence of Neandertal and modern human autosomal DNA sequences of 825,000 years. We caution that this is only a rough estimate because of the uncertainty about the time of divergence of humans and chimpanzees."

The comparison here is a reference genome UCSC hg18 (which I believe is composed of anonymous donors of unknown ethnic origin - but I think may be CEU - White Utah).

older and mtDna only

"gene divergence time of 631-789 KY".


"The estimated average time of divergence between Denisovan and Neanderthal sequences is 640,000 years ago, and that between both of these and the sequences of modern Africans is 804,000 years ago. "

Of course, in the latter case divergence times are based on modern Africans, so perhaps could could be an overestimate for the reasons of archaic mixture into Africans...?

Sean said...

10,000 YE says that introgresssing alleles were strongly selected when agriculture presented humans with new problems.

The West of Africa was where the archiacs were. West Africa is also where the Bantu expansion started.

Insightful said...

If a population stays put in one place, like those chimps, it will accumulate more and more "junk" variability. If, however, it buds off and adapts to a new environment, it will lose much of that variability...

2 Questions for Peter:

Do Europeans have more genetic variability than Euro-Canadians since the latter budded off to become a founder population?

One more question. Which group has more genetic variation, Europeans or Melanesians? Thanks..

Anonymous said...

Spandrell said...
13% of negroid genes are non-sapiens?Is this certain?

plus 2%

Peter Frost said...


Yes, I was aware of those older time estimates for Neanderthal-modern human divergence. But the majority opinion still seems to favor 500,000 BP (see the Wikipedia entry on Neanderthals).

But you may be right. Green et al.'s estimate is suspect (it might have to be revised upwards because of contamination).

This being said, how would you reply to my second argument, i.e., that modern humans were geographically closer to the Levant Neanderthals and hence more exposed to gene flow from that source?


Well, OK. I'm not hostile to the "cherry-picking" hypothesis. But it just seems to be one long drumroll ...


Euro-Canadians are a product of multiple migration events. This is true even if we look only at English Canadians. Some came directly from England at different periods and others by way of the U.S.

French Canadians are definitely descended from a small founder group, only 10,000 or so founding immigrants (and some of them, being earlier, ended up contributing much more than others).

If I remember correctly, there is no difference between total genetic variability of Europeans and total genetic variability of Melanesians. I'm not sure I would expect much difference. Yes, there is Denisovan admixture, but the time depth is comparable, a bit more for the PNG component and a lot less for the Austronesian component.


Good point. Archaic admixture in sub-Saharan Africa encompasses a lot of admixture from "almost-moderns" and much less (but significant) admixture from a very archaic population.

Anonymous said...

you people are a bunch of sperglords

Sister Y said...

So as I understand it, the theory in the Watson 1997 paper, supported as far as I can tell by Labuda 2000, is that two groups of anatomically modern humans separate in Africa, then stay separate for 20,000 to 100,000 years. Then one of the groups becomes the behaviorally modern humans and expands all over Africa in the Upper Paleolithic, and mixes with the previously separate (non-behaviorally-modern?) population (resulting in Watson's 13% admixture figure) and (less) some super-archaics (the 2%), and also expand all over the world and mix with archaic populations where they find them, like Neanderthals and Denisovans.

Is the 13% figure controversial? You describe them as "almost human." What do we know about the two separate groups? I notice this kind of admixture is not even mentioned on the Wiki for human-archaic admixture, and it's certainly the first time I've heard of it, fifteen years after the Watson study. Has anyone supported or refuted it? I probably don't have the right search terms, because I can't find anything more recent than the Labuda 2000 paper.

Stephen said...

How much paleantology is done in places like Arabia, Persia and Iraq if Africa had so many Archaics perhaps the middle east is a more likely homeland of Sapiens that perhaps could be missed by a lack of digging. At the crossroads of the continents that region could also have more mutations available for selection.

Peter Frost said...

Sister Y,

The picture now seems a bit more complicated than Labuda et al proposed. The evolutionary sequence can be summarized as follows:

- 500,000 years ago, all of Africa was inhabited by an archaic population similar to the Neanderthals in Europe and to the Denisovans in Asia

- perhaps around 200,000 years ago, these African hominins began to differentiate into two populations: (a) an evolutionarily conservative lineage in western and southern Africa; and (b) a more evolving lineage in eastern Africa.

- around 110,000 years ago, the eastern African hominins expanded north into the Levant (Skhul-Qafzeh). They were close to being modern human. Their material culture was similar to that of the Neanderthals but their anatomy was almost modern, albeit with some archaic features

- beginning around 80,000 years ago, a sub-population among the eastern African hominins underwent a series of expansions. Then, around 60,000 years ago, there was a final "big bang" that eclipsed all of the others and gave rise to true modern humans.

The 13% figure is not controversial. What is controversial is the status of the Skhul-Qafzeh hominins who seem to be associated with that fraction of the African gene pool. They were long thought to be true modern humans. In recent years, they have been downgraded to "almost-modern" status.

There are fashions in anthropology. People are really only just now coming around to the idea that archaic admixture occurred in Africa as well.


Much less than in Israel, unfortunately. Ecologically, the Arabian Peninsula is part of Africa, so I would expect to see more evidence there of East African hominins.

Anonymous said...

Off topic perhaps, but this harkens back to some of your earlier topics.

One effect of different hair colors and eye colors among Caucasians is that it is harder for Caucasian women to cheat, say, than it is for Asian women.

I am not saying that these differences evolved for that purpose, but they would have had an effect once started.

Sister Y said...

It at least reduces the size of the pool of people available to cheat with, to the extent that folks understood the heritability of such traits.