November 01, 2011

Homo floresiensis: dramatically dwarfed Homo erectus descendant

Journal of Human Evolution doi:10.1016/j.jhevol.2011.08.008

Craniofacial morphology of Homo floresiensis: Description, taxonomic affinities, and evolutionary implication

Yousuke Kaifu et al.

This paper describes in detail the external morphology of LB1/1, the nearly complete and only known cranium of Homo floresiensis. Comparisons were made with a large sample of early groups of the genus Homo to assess primitive, derived, and unique craniofacial traits of LB1 and discuss its evolution. Principal cranial shape differences between H. floresiensis and Homo sapiens are also explored metrically.

The LB1 specimen exhibits a marked reductive trend in its facial skeleton, which is comparable to the H. sapiens condition and is probably associated with reduced masticatory stresses. However, LB1 is craniometrically different from H. sapiens showing an extremely small overall cranial size, and the combination of a primitive low and anteriorly narrow vault shape, a relatively prognathic face, a rounded oval foramen that is greatly separated anteriorly from the carotid canal/jugular foramen, and a unique, tall orbital shape. Whereas the neurocranium of LB1 is as small as that of some Homo habilis specimens, it exhibits laterally expanded parietals, a weak suprameatal crest, a moderately flexed occipital, a marked facial reduction, and many other derived features that characterize post-habilis Homo. Other craniofacial characteristics of LB1 include, for example, a relatively narrow frontal squama with flattened right and left sides, a marked frontal keel, posteriorly divergent temporal lines, a posteriorly flexed anteromedial corner of the mandibular fossa, a bulbous lateral end of the supraorbital torus, and a forward protruding maxillary body with a distinct infraorbital sulcus. LB1 is most similar to early Javanese Homo erectus from Sangiran and Trinil in these and other aspects. We conclude that the craniofacial morphology of LB1 is consistent with the hypothesis that H. floresiensis evolved from early Javanese H. erectus with dramatic island dwarfism. However, further field discoveries of early hominin skeletal remains from Flores and detailed analyses of the finds are needed to understand the evolutionary history of this endemic hominin species.

Link

10 comments:

Andrew Oh-Willeke said...

One implication of H. Florensis being an offshoot of Homo Erectus, is that it suggests Denvisovia was probably a H. Erectus site (the latest by far) and that the H. Erectus gene pool must have been pretty fixed all the way from S. Siberia to first contact with modern humans, probably in Southeast Asia somewhere when it wound up in the Papuan and Australian and Filipino Negrito genomes. Otherwise the Denisovian genomes wouldn't match the traces in these populations.

Mike Keesey said...

How does it suggest that Denisovans and Homo erectus are the same? Denisovan nuclear DNA diverged from Neandertal nuclear DNA after the initial sapiens-neanderthalensis split around 400 kya, and erectus (even in the strict sense) diverged before that.

The Denisovan mitochondrial genome diverged earlier, so that could have something to do with erectus, but it seems like that's still an untestable hypothesis without definitive erectus DNA.

eurologist said...

The Devisova location makes heidelbergensis-like people a much better candidate - who apparently spread east into much of northern and mid-latitude Asia around 300,000 - 200,000 ya. For them, a split off Neanderthal around 400,000 ya is also logical, if you think of an eastern population perhaps in West Asia that moved further East (as Neanderthal would do much later, but in a much colder climate).

Asian erectus seemed to have preferred tropical and perhaps sub-tropical regions, was much less advanced, and the least likely mating partner.

If the admixture with Denisovans took place before the Toba eruption (seems likely based on the geographic split and the data we have so far), then the location could have been almost anywhere within an early migration, and the most admixed ended up in protected (and later separated by water) areas at the fringe. Newer migrations diluted Denisovan admixture elsewhere.

terryt said...

"We conclude that the craniofacial morphology of LB1 is consistent with the hypothesis that H. floresiensis evolved from early Javanese H. erectus with dramatic island dwarfism".

Isn't that what most sensible people have thought since the Hobbits were first discovered? I agree there have been any number of diversions from that scenario proposed, but none have held up to any sort of scrutiny.

"How does it suggest that Denisovans and Homo erectus are the same?"

I agree. How do you come to that conclusion?

"the H. Erectus gene pool must have been pretty fixed all the way from S. Siberia to first contact with modern humans, probably in Southeast Asia"

You're making a few unjustified assumptions here.

Andrew Oh-Willeke said...

I am assuming:

(1) H. floresiensis is basically genetically similar to to local H. erectus from which H. florensiensis is derived.

(2) I am assuming that if H. florensiensis was derived from H. erectus ca. 100 kya, that H. erectus was the only archaic hominin in the SE Asia region with which early modern humans in the region admixed.

(3) I am assuming that the fact that archaic admixture in Melanesians and Australian aborigines looks like Denisovian admixture, and that they were actually admixing with SE Asian homo erectus of ca. 100,000-45,000 years ago, implies that Denisovians and H. erectus were genetically similar, and hence that Denisovians were H. erectus.

Also, if Denisovians were admixed Neanderthal-H. Erectus or some such, we'd expect that those with Denisovian admixture would have elevated Neanderthal admixture, which isn't the case.

Now, H. Erectus DNA ca. 100,000 yeras ago may differ quite a bit from H. Erectus DNA ca. 1,900,000 years ago, but given the apparent continuity (also suggested by the fact that Oldowan industry was found in the region from H. Erectus until the last archaic industry traces prior to modern humans), this doesn't mean that we can't call both subtypes of a H. Erectus continuity.

The "inferiority" of H. Erectus relative to Nenaderthal could explain lower admixture percentages in mainland Asians relative to H. Erectus, with small founding populations accounting for high admixture in Melanesians and Aboriginal Australians.

terryt said...

"I am assuming"

I agree with number 1. As far as number 2 goes I'm sure the first part is correct, 'I am assuming that if H. florensiensis was derived from H. erectus ca. 100 kya, that H. erectus was the only archaic hominin in the SE Asia region'. But 'with which early modern humans in the region admixed' is only partially correct. They could have mixed with at least one other 'archaic hominin' before thet reached SE Asia. So that leaves number 3 only partially correct also. I agree with, 'Melanesians and Australian aborigines looks like ... they were actually admixing with SE Asian homo erectus of ca. 100,000-45,000 years ago'. But those 'SE Asian homo erectus' were not necessarily the source of the Denisovan admixture.

"The Devisova location makes heidelbergensis-like people a much better candidate - who apparently spread east into much of northern and mid-latitude Asia around 300,000 - 200,000 ya".

Totally agree.

"For them, a split off Neanderthal around 400,000 ya is also logical"

Don't Denisovans split off before the human/Neanderthal split? That would make Denisovans more ancient than the African Line.

"if you think of an eastern population perhaps in West Asia that moved further East (as Neanderthal would do much later, but in a much colder climate)".

Yes. But members of that population didn't move south as far as Melanesia/Australia till humans carrying Africa haplogroups had passed through.

"the most admixed ended up in protected (and later separated by water) areas at the fringe. Newer migrations diluted Denisovan admixture elsewhere".

That's the way I see it.

eurologist said...

Don't Denisovans split off before the human/Neanderthal split?

In terms of contact, I think it may very well be that Denisovans split off Neanderthals roughly at the same time as modern humans did (the fact that we share ~50% of Denisovan-specific SNPs where we differ among ourselves indicates as much to me). However, that's not the same thing as human-Denisovan relation, since European erectus and heidelbergensis started to diverge from Africa at least a million years ago, and roughly does not mean the same as identical. As such, Neanderthals and Denisovans carry different, distinct sets of that one million year history, in them.

terryt said...

"In terms of contact, I think it may very well be that Denisovans split off Neanderthals roughly at the same time as modern humans did"

I was thinking specifically of the mtDNA. Wiki, because it's easiest:

http://en.wikipedia.org/wiki/Denisova_hominin

Quote:

"The mtDNA from the finger bone differs from that of modern humans by 385 bases (nucleotides) in the mtDNA strand out of approximately 16,500, whereas the difference between modern humans and Neanderthals is around 202 bases. In contrast, the difference between chimpanzees and modern humans is approximately 1,462 mtDNA base pairs.[2] This suggested a divergence time around million year ago".

That suggests the similarities between Modern, Neanderthal and Denisovan is a result of later interbreeding. More from the link:

"they concluded that in spite of the apparent divergence of their mitochondrial sequence, the Denisova population along with Neanderthal shared a common branch from the lineage leading to modern African humans. 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. They suggest that the divergence of the Denisova mtDNA results either from the persistence of a lineage purged from the other branches of humanity through genetic drift or else an introgression from an older hominin lineage".

My vote goes to introgression.

Some papers, in case you haven't seen them:

http://genetics.med.harvard.edu/reich/Reich_Lab/Welcome_files/2010_Nature_Denisova_Genome.pdf

http://www.eva.mpg.de/genetics/pdf/Krause_Complete_Nature_doi.pdf

eurologist said...

Of course I have seen those papers (whether I can remember details is another story...). ;)

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.

Given the definition of "divergence time," isn't that rather close to what I said, anyway?

On the mtDNA side, the difference between Denisovans' split-off and Neanderthals' is about a factor of two. That is a similar difference as between average world-wide human differences and the extreme. Given the few data points, I still think this can be explained without introgression. I would have pegged introgression at East Asian erectus times - i.e., ~2 million years.

terryt said...

"I would have pegged introgression at East Asian erectus times - i.e., ~2 million years".

Possibly. But I think it's more likely to have been an ongoing process, whenever the opportunity arose.

"The estimated average time of divergence between Denisovan and Neanderthal sequences is 640,000 years ago"

Doesn't that place their divergence to be late enough to have both sprung from heidelbergensis?