New blogs

Leherensuge was replaced in October 2010 by two new blogs: For what they were... we are and For what we are... they will be. Check them out.

Wednesday, April 1, 2009

Y-DNA timeline (4) - comparison with climatic data


In the second post of this mini-series, I suggested to do some fine tuning experiments in regard with my Y-DNA timeline analysis, by comparing the general graph (which shows pattern and proportions)
with actual climatic data.

I have done some of it, mirroring the glaciar climatic data in order to make it fit with the left to right orientation of my graph, and stretching it at different escales until it fit as well as I could figure with the apparent demic expansions' pattern.

The result is as follows:



I am pretty much satisfied with this result but I welcome alternative proposals if you think you can find something better. Just in case I tried to find other possible fittings between the two graphs and the best I could come up with was this:


Unlike in the previous case this alternative does not satisfy me, nor I think it will anybody else. Even if the oldest dates could fit better with some usual assumptions (OOA at 80-70 kya, main Eurasian expansions at 60-40 ka), it makes the "recent past" final line to be way too distant from the present, well in the Magdalenian period and in the middle of a cold phase. There are many other inconsistencies that suggest that this "short chronology" version is wrong.

But feel free to experiment yourself: maybe you can come up with some improved chronology that I have missed.

So I am sticking to the first graph, which I am pretty much satisfied with. What revised conclusions can we come up with?

1. The Y-DNA root and the OOA (CF split) epysode are close in time. It suggests that the whole process of Y-DNA diversification at high levels happened in an expansive phase that culminated in the OOA. It fits well with a climatic high and also with archaeological data, recently ratified with the datation of North African Aterian to c. 95,000 BP. It suggests that the OOA route was not via southern Arabia but via the Fertile Crescent instead, where the oldest Eurasian AMH fossils are found.

2. The DE split (OOA chapter 2) is the only anomalous reference point, as it is in a climatic low. Nevertheless this may well be due to the normal margin of error in the estimate and could have happened in the previous or subsequent climatic highs perfectly.

3. The F-K (or F-IJK-K-NOP) supernode (first expansion) is in the 80-70,000 BP window. It may have got the Toba epysode in between (sharp valley between two climatic peaks) or right after it (much larger and also deep climatic low). Whatever the case, it fits well with the evidence of South Asian cultural continuity at Jawalpura under and over the Toba ash layer.

Of some interest is that the first and second nodes in this supernode show divisions with West Eurasian exclusive subclades, as are G adn IJ, as well as South Asian only (H). It suggests that "the F clan" was estabilished between West and South Asia at the time.

A parallel expansion is that of macrohaplogroup D, which shows an East Asian distribution nowadays (from Andaman to Siberia). It is possible that D lineages represent, in the Y-DNA, the oldest Eastern Eurasian signature.

In Africa, haplogroup B seems also to have expanded in this window.

4. The second expansion corresponds to a better accepted period: that between 60 and 50,000 BP for the Eurasian expansion.

The expansion of macrohaplogroup C may have happened then (though there is great uncertainty, because this clade is not sufficiently well studied). A possible center for this haplogroup could have been SE Asia, though South Asia is also a possibility.

Other splits happening in this phase are (seemingly) those of H, IJ, NO and P (Q and R). It clearly seems an already regionalized expansion in South Asia, West Asia, East Asia and Central/South Asia respectively. Some time before, Melanesian clade M expanded too, what surely means that people were already thriving in New Guinea at that time.

In Africa, haplogroup E also began its early expansion at this time.

5. Another moment of particular interest is c. 41,000 BP, where the graph places the split between I2a and I2b, as well as of J1 and J2. This is so strictly coincident with Aurignacian expansion that I can't but take it as proof that the graph is well calibrated (as this was not taken in account initially at all). Further division of these clades happened closer to the Gravettian timeline, what makes me wonder if some J2 subclades, rare in West Asia but common in Europe and South Asia, are not pre-Neolithic signature as well.

Other clades expanding in this period (roughly 45-30,000 BP) are C2 (Melanesia), G (West Eurasia), N (East and North Eurasia), O (East and SE Asia), Q (South, Central and North Asia), R (South, Central and West Eurasia) and S (Australia).

In Africa, it is also the time of expansion of E1b1 and E1b1b (but not yet of E1b1a), as well as that of E2 (but not yet of E1a).

6. The last mark in cyan color in the graph is for the formation of the West European R1b1b2a1, dated to some 19,000 years ago. It really makes it look like a Solutrean or Magdalenian founder effect. A matter that surprises me but that I am going to research as much as I can, looking for possible precursors of either of these cultures in West Asia, where R1b coalesced almost without doubt since some 30-28,000 years ago. This date for R1b as a whole is nevertheless most intriguing, as it is oddly coincident with Gravettian expansion. I can only scratch my head at the moment, as the structure of R1b is only relatively well understood with many blanks in our knowledge of non-European (and even European) subclades.

The formation of the (rather rare) R1b1b1a subhaplogroups would still fall within a late Magdalenian (or early post-Magdalenian at most) time frame.

Other expansions at this "late" moment are those of E1a and E1b1a (Africa), C1 (Japan), L (South and Central Asia), T (West Eurasia), and probably R1a1 towards the easternmost edge of Europe.

7. The "recent past" final line of (relative) stability is dated to some 9,000 years ago, in the Epipaleolithic or early Neolithic (depending on the region). This may seem a little early to some but it's what the excercise of fine tuning has produced on its own. It does anyhow fit with a climatic high, what makes sense.

That's all, folks.

_______________________________________________________
Previous post of the Y-DNA timeline series: 1, 2, 3.
Related R1b structure post.
MtDNA miniseries: 1, 2, 3.

.

23 comments:

Ebizur said...

Maju said,

"haplogroup D...from the Andamans to Siberia"

Haplogroup D in Siberia? Maybe so, since Turkic-speaking people living in and around the Altai Mountains (the so-called "Altaians") may be considered to live in southernmost Siberia, and Kharkov et al. (2007) have reported finding 6/96 = 6.25% D-M174(xD1-M15) in their pool of samples of "Southern Altaians." However, I do not recall having encountered haplogroup D Y-DNA in any sample of a population in Siberia outside of the Altai region.

South Central Haplo said...

Maju,

My thinking is similar to your thinking. I believe in god with left orientation. I am from India. While going through Basque words two words strike me.

Brother : Anaia
Mother : Ama

I am familar with Indian,Arabic and Europian words.

Only other language I see Anaia for brother is South Indian Language Telugu(Dravidian). There may be others. Is there any phonetic study done on Basque words with Middle Eastern, Indian, Indo Europian etc?.

Maju said...

@Ebizur: you may be right and I overextended the haplogroup to the North somewhat.

Just visited Mathilda's blog, where she refers to an intresting recent study on Y-DNA D. The authors conclude that it spread from South to North (the main node could be in Thailand) long before the other East Asian clades did. They estimate 60,000 years ago for that process.

They do mention a 0.33% of D in North Asians but guess this is kind of residual.

Ebizur said...

Maju said,

"They do mention a 0.33% of D in North Asians but guess this is kind of residual."

I guess that this should be due to the 5%+ D(xD1) that has been found in Altaians; the supplementary data table of Hammer et al. (2005), most of which is based on the data of Karafet et al. (2002), also reports 5/98 = 5.1% D-M174(xD1-M15, D2-P37.1, D3a-P47) in an "Altai" sample. The inclusion of Altaians in the category of "North Asians" is questionable, but they are very likely responsible for the haplogroup D that is present in that study's "North Asian" pool.

South Central Haplo said,

"I am from India. While going through Basque words two words strike me.

Brother : Anaia
Mother : Ama"

Japanese
(Elder) brother : Ani
Mother : Omo [archaic]

Chinese
Mother : Ama ~ Mama ~ Ma
[literary]Mu ~ Muqin
(Some dialects use Niang ~ Aniang instead, which sounds sort of like Turkish)

Korean
Mother : eomi ~ eomeom ~ eomma ~ eomeoni ~ eomeo-nim
(In Middle Korean also eoma-nim)

Maju said...

@South Central Haplo:

Well, the word ama/amma/mamma is actually Proto-World.

...

No. Just kidding: happy fools day! Was yesterday? Who cares? I just noticed as someone announced the "discovery" of three cases of Neanderthal introgression in Scandinavia. That was fun as hell! :D

Seriously: ama/amma/mamma/mum is actually the first sound that babies are able to pronounce. They probably do not mean anything at all but all parents like to think the little cute thing is talking of them. There is somewhat more variation in their second word (ata/tata/dada/baba), which is made to mean father normally.

So ama and similar words is just a most extended word for mother but lacks any sort of real genealogy: it's re-invented every single day.

Anai is more intriguing maybe. Notice that the actual Basque word is "anai", not "anaia", which is just the basic nominative/direct object decination. It does not mean brother in general either but just brother of a man/boy. The brother of a woman is "neba". Similarly there are different words for sister of a man (arreba) and of a woman (ahizpa).

I don't know of any systematic study of a potential relation of Basque and Dravidian but I have made very basic comparisons with many languages and found no or very feeble connections.

In general the best candidate for a distant living relative of Basque is the Eastern Caucasian family, whose largest representative is Chechen, and that could well be a remnant of the Eastern Gravettian cultural area that penetrated in West Asia by the Caucasus and Zagros (Zarzian culture) in the Epipaleolithic. Anyhow the correlation is very hypothetical and conflicts with other, even more hypothetical, superfamiies like Sino-Dene-Caucasian or even Caucasian itself.

Of the dead languages the closest relative is of course ancient Aquitanian, that just seems to be Basque by another name. Then the best candidate is Iberian, which shows some clearly connected words (especially in personal names) and a general pattern of similar sound rules. Nevertheless Iberian has never been deciphered (we know how it sounded approximately, because it was occasionally written in Greek alphabet, but not what it means) and Basque has not helped in this at all. Once I transliterated an Iberian text from Murcia (originally in Greek alphabet) and presented to a native Basque speaking friend: he read it and said jokingly "it's not the dialect of Ondarru [his native town] but could from Lekitto [Lekeitio, the neighbouring one]".

At least one linguist has proposed an alternative reading of Tartessian (or Sudlusitanian or Western Iberian), generally considered a total isolate, and proposed it was close to Basque.

Other theories include that of Vasconic languages once extending through most of Europe, mostly based on toponimy, and stuff like that.

I have tried in several occasions to use the numbers' names to estabilish very feeble hypothesis on possible connections of isolated languages. My best hunch was to group Sumerian, Hurro-Urartean and East Caucasian, with a weaker link to Basque. But this is very tentative, as with dead languages we know only so much (extremely little in most cases) and with rare living languages happens much of the same too (I doubt any linguist speaks even poorly both Basque and Lezgian for instance, as to see if there is any real pattern of connection with more knowledge than a dictionary can give).

As said, I could not find any apparent connection between Dravidian and Basque by comparing lists like that of numbers. The word "anai(a)" is an interesting coincidence but if there is nothing more, I can only think of a term maybe spread in the religious context of Dolmenic Megalithism that (quite mysteriously) spread into several parts of Asia after it was finished in Europe ("brother" is a term often used in religious communities). Of course, it can be just a coincidence.

You can do-it-yourself with the Swadesh lists at Wikictionary. My first quick look at the Dravidian list does not appear much "Basquic" but guess you can find the occasional coincidence (this is normal: a mere product of randomness).

Maju said...

In a previous conversation, Ebizur, you said that:

it seems that about two out of every three Polynesian males should belong to the C-RPS4Y/M130 clade, and about one out of every four Polynesian males should belong to the O3-M122 clade. These proportions are roughly similar to those for the C and O3 clades in Mongolians, but there is little overlap between the Mongolians and the Polynesians in regard to the subclades of C and O3 to which they belong.

I woke up today thinking of this and the fact that in my graph the expansion (branching) of C and NO are rougly contemporary (unlike that of D, which seems older). Even the expansion (branching) of the derived C and O clades (excepted C1) appear roughly contemporary (c. 40-30 kya). Does this make any sense to you? In other cases (Han or SE Asians for instance) this apportion of C and NO/O does not seem to exist instead. Can we get any conclusions, even if tentative, about these patterns? Was there a 3-1 apportion in the original C-NO expanding population (or one of the original ones)?

I don't have any clear opinion yet just that, suddenly, this thought got me intrigued.

terryt said...

That's a topic that interests me too. As far as I know no-one has come up with a decent haplogroup tree of C, just all branching at the same time.

terryt said...

"Can we get any conclusions, even if tentative, about these patterns?"

It depends on whether we consider that C Y-chromosome reached Australia with the first settlers (at least 50,000 years ago) or arrived and spread rapidly after being introduced at the time of the Austronesian and Na-Dene expansions. There's no evidence for any substantial genetic input into Australia at the more recent time though. Perhaps some other time?

"Even the expansion (branching) of the derived C and O clades (excepted C1) appear roughly contemporary (c. 40-30 kya)".

This time scale for C's expansion would be too recent if the variant that gave rise to C4 arrived with the first people into Oz.

Maju said...

The estimates for C are quite uncertain because the lineage is not sufficiently well studied.

Still my estimate for the branching of C is of c. 60kya. 30-40kya would be for its main branches to split again (sign of expansion). This would certainly make sense for C3, IMO, but it's kind of odd for C4. Still it can be an internal process of drift/fixation within Australia, and the lineage have been there in small numbers much earlier (c. 60 kya).

terryt said...

Thanks for that explanation.

I'd agree that C3's expansion took place long after the basal expansion of C. But that still puts its origin as being fairly old, long before its expansion.

I'm inclined to accept that Y-hap C arrived in Australia with the first humans there. Therefore I think you're correct with the explanation of C4's expansion when you say, "it can be an internal process of drift/fixation within Australia". But the lineage may well have been there in large numbers much earlier (c. 60 kya). It's just that there was some sort of population bottleneck after that initial colonisation. It seems accepted that the Australian Aboriginal languages diversified much more recently than the population's arrival.

Maju said...

I really can't adress the issue of C properly (i.e. with some security) for the reasons mentioned before but I do notice that M, for example, appears much older even with the same kind of constraints. This may be some sort of distorting accident that I can't account for properly but might also be something real.

What I can say is that with the timeline posted here, C3 diversified well in time to participate in American colonization (and more if, as many suspect, only participated in a late wave). C3 is my best reference within C and its expansion timeline fits well with what we know about it and in general with the Eurasian northlands, where C3 may well be a relatively late developement. Q and N that are probably older in the area (at least in meaningful numbers), also appear older in the timeline and with similar timelines with each other (timelines that appear coincident on first sight with northern mtDNA expansion patterns too).

A reasonable possibility is that C subclades participated as minority clades in the expansive drives of other lineages like H, M, NO and S, becoming more dominant in some populations by local founder effects/drift.

Notice anyhow that the timeline for the other Australian major lineage, S, is not really too different from that of C4. S appears to have a expansion timeline c. 40,000 BP while C4 is more in the 30,000 BP zone but both fit well within a UP periodization, when there could well have been all kind of drift phenomenons, like the ones we observe in Europe and other places. MtDNA instead does not seem to have been altered so much by drift instead (what is logical) and appears to be very old in all regions, including Australia.

terryt said...

"both fit well within a UP periodization, when there could well have been all kind of drift phenomenons, like the ones we observe in Europe and other places".

Yes. But that drift could only have happened on populations already in Australia. If they had previously 'participated as minority clades in the expansive drives of other lineages like H, M, NO and S' we would expect to find remnants of the Australian haplogroups outside that continent. The haplogroups' ancestors presumably arrived originally with the first colonisers, perhaps 60,000 years ago, and therefore their separation from related haplogroups dates to that time.

"C3 may well be a relatively late developement".

But you would have a difficult task arguing that its separation from C's root was so much more recent than that of C4, unless further research eventually reveals that C4 branched first. Again, where it originated is a mystery but surely we should find remnant populations if it had originated anywhere other than within its current geographic range.

"and more if, as many suspect, only participated in a late wave".

I think the evidence for that late wave is fairly convincing. However that doesn't eliminate the possibility it survived in some small region for a long time before that wave.

"Q and N that are probably older in the area (at least in meaningful numbers)".

That piece in brackets sums up the situation, I believe.

Maju said...

Yes. But that drift could only have happened on populations already in Australia.

Yah, sure. In Sahul/Wallacea in any case.

If they had previously 'participated as minority clades in the expansive drives of other lineages like H, M, NO and S' we would expect to find remnants of the Australian haplogroups outside that continent.

The word "previously" is your personal contribution and that was not what I meant. I meant "in real time", "as they marched".

The haplogroups' ancestors presumably arrived originally with the first colonisers, perhaps 60,000 years ago, and therefore their separation from related haplogroups dates to that time.

It is possible though there is no solid evidence of AMH colonization of Australia before c. 40,000 BP.

But you would have a difficult task arguing that its separation from C's root was so much more recent than that of C4...

It's not what I said nor meant: C3, in fact its expansion, is a late developement in NE Asia, not in relation with C.

Maju said...

Anyway, we don't really know if Homo erectus actually used any manufactured craft to achieve these two crossings.

Swimming? It is very different to achieve such feat individually (as a sportman would maybe) than to cross several kilometers with your whole clan or family (or several of them maybe), what is needed to achieve a stable colonization. No parent would risk swimming 5 kilometers with two or more children, probably very young, on their backs, obviously. And even you could find one such dangerous parent, you would not find two.

Either they were brought by some sort of tsunami (but we find no animals that made it) or it was intentional colonization with some sort of boat or raft.

I doubt very much that 'the derived Oceanian lineages' themselves originated anywhere other than near where they survive today. It would have to have been their ancestors who moved through either Central Asia or India.

Take C4 and C2: ancestor is C, where do we find C lineages? In the Indo-Pacific arch. what does it likely mean? That C coalesced soemwhere in that IP arch, not in Central Asia.

Same with every other single clade.

Y-haps C3 and D and mt-haps A and Ket N. Mt-hap Y is also difficult to explain away as having arrived via India. In fact it is most unlikely any of these linages could have moved through India "without leaving some trace".

C has at least two lineages in India (C5 and C*, which may well be C6 for what I know). D is more and more obviously a very old remnant or a very early colonization of East Asia (from south to north) and was drifted out elsewhere, where it was never surely in sufficiently high numbers to persist.

A and Euro-Ket N* are too derived to be meaningful: they must have experienced extremely long coalescences before arriving to destiny. In fact mtDNA D is more meaningful ("only" 3 SNPs at root, still younger than many southern lineages) but I'm working atm in a revised mtDNA tree and tentative chronology and and don't want to advance on issues I'm still uncertain of.

Only if we 'believe' that all ancient haplogroups derive from just a single small region.

The explosion of M and N are simultaneous or in rapid sequence, their distribution patterns are extremely similar. A single area of origin is most likely, even if it doesn't need to be "small" (South Asia is actualy *huge* and more if we consider SE Asia as fully involved).

The human species has probably always been fairly widely distributed, if only through Africa, the Middle East, the Iranian Plateau and Southern Asia.

I have no evidence for the Iranian plateau nor in fact anywhere between Palestine and India. If you know of something, please let me know.

As for Africa, it's fairly irrelevant for the autonomous Asian developements but still it's likely that Central and West Africa were colonized after humans were already in Asia.

The belief in a single small region of origin is certainly a product of "the typical Judeo-Crhistian trying to 'prove' Noah's deluge or whatever".

Actually it's more what we find in the data, which insists in forming a pattern along the Indo-Pacific areas. Though, I insist, Tropical Asia is anything but small.

Ebizur said...

Maju said,
"and more if, as many suspect, only participated in a late wave."

terryt said,
"I think the evidence for that late wave is fairly convincing. However that doesn't eliminate the possibility it survived in some small region for a long time before that wave."

Maju said,
"Q and N that are probably older in the area (at least in meaningful numbers)."

terryt said,
"That piece in brackets sums up the situation, I believe."

Maju said,
"It's not what I said nor meant: C3, in fact its expansion, is a late developement in NE Asia, not in relation with C."

I'm sorry to have to say this, but I'm afraid you two are babbling nonsense.

All estimates of the ages of Y-DNA lineages in North Asia have firmly placed haplogroup Q (or haplogroup P, which subsumes Q) and haplogroup C (which subsumes, and in this region should be equivalent to, C3) together as the
oldest Y-DNA haplogroups in this region. It is most likely that haplogroup C3 has accompanied haplogroup Q during the original colonization of North Asia (and eventually the Americas).

Haplogroup N, on the other hand, is a much more recent development, and has nothing to do with the original Siberians or the original Americans, which are derived from the former.

Maju said...

I'm sorry to have to say this, but I'm afraid you two are babbling nonsense.

I would not be surprised, sincerely.

All estimates of the ages of Y-DNA lineages in North Asia have firmly placed haplogroup Q (or haplogroup P, which subsumes Q) and haplogroup C (which subsumes, and in this region should be equivalent to, C3) together as the
oldest Y-DNA haplogroups in this region. It is most likely that haplogroup C3 has accompanied haplogroup Q during the original colonization of North Asia (and eventually the Americas).

Haplogroup N, on the other hand, is a much more recent development, and has nothing to do with the original Siberians or the original Americans, which are derived from the former.


That is what I do not see in the SNP-based estimates that constitute the core of this post (and previous ones). Both Q and N appear of similar age, while C3 appears younger. This may be a false impression due to lack of data on haplogroup C in general, which may in fact be older than it appears here, but makes some sense anyhow, as Q and N are better estabilished among Siberians than C3. C3 is also suspected nowadays to be more recent (i.e. belonging to a second wave, Clovis) in America.

I'd like of course to keep a more prudent stand but it's difficult when you're being challenged in every single point by Mr. Terry, aka "no boats till I say so", aka "Papuans come from Siberia", what forces me to adress every single point in greater detail than I'd like to. But well...

But I also keep a most prudent stand (i.e. open disbelief) regarding standard TRMCA guesses: most of them make no sense at all. For example I have been working as of late in the mtDNA tree and chronlogy and, following SNP counts, I find that U cannot be older than R0 and that H (and most H subclades) must be in fact older than most U subclades. And I also find that that makes more sense than the usual TRMCA-based assumptions, as H (unlike most U clades) is clearly Eurocentric in its spread and U is rarer in Central and Western Europe (where human colonization is older) than in Eastern Europe (only properly colonized in Gravettian times). H spread would still be roughly contemporary with U6 which may represent the North African "Aurignacian" founder effect (Dabban industries - do not know much about them yet but seem to be the "other Aurignacian", both of which would be derived from Levantine industries).

But, well, I'm having a hard time with mtDNA timeline because, even if H is as old as Aurignacian, the main spread of M, N and R wouldn't be much older than 60,000 BP. This seems to force me to redefine the timeline for Y-DNA, because it makes no sense at all that male lineages would spread before female ones (they should be contemporary or younger than female ones).

So maybe the second graph, that I first disdained, should be taken more seriously after all. I am still not fully convinced and will take my time before posting on all that.

terryt said...

I'll leave you two to argue it out. I'm off for a few days.

terryt said...

"what forces me to adress every single point in greater detail than I'd like to. But well..."

I hope I have stimulated you to further thought though. Certainly you've now come up with a pretty good mtDNA timeline.

"It is possible though there is no solid evidence of AMH colonization of Australia before c. 40,000 BP".

I am strongly in favour of the argument that an interpretation of all the available evidence supports a date of at least 46,000 years ago for human arrival in Australia. Megafauna extinction alone is enough to support that date. After all it's not the only part of the world where species extinction is closely associated with human arrival, and no other explanation for the Australian extinctions actually holds water once looked at it closely. The 46,000 year date is also certainly believable for Y-hap C's arrival in Australia (C4), and presumably in New Guinea (C6).

So C4 and C6 (or their respective ancestors) must have separated from the core C Y-hap by at least 46,000 years ago. If all six C Y-haps separated around the same time then C1, C2, C3, and C5 must also have become separated by 46,000 years ago, which fits your original diagram.

"The expansion of macrohaplogroup C may have happened then ... A possible center for this haplogroup could have been SE Asia, though South Asia is also a possibility".

I'm pretty sure it's SE Asia. If we look at the distribution of C*(xC1,2,3,4,5,6) we find it distributed around what, for much of ancient prehistory, would have been a virtually landlocked south China Sea, with an apparent outlier in India.

View of the South China Sea bathymetry:

http://walrus.wr.usgs.gov/infobank/gazette/jpg/regions/scs.jpg

The route by which C reached SE Asia is as yet unknown.

Maju said...

I hope I have stimulated you to further thought though. Certainly you've now come up with a pretty good mtDNA timeline.That forces the costal migration model all the way from Djibouti.

I am strongly in favour of the argument that an interpretation of all the available evidence supports a date of at least 46,000 years ago for human arrival in Australia.You said 60,000 before.

Megafauna extinction alone is enough to support that date.Or could have been caused by Toba or whatever.

Fossils, including tools, are all in the 40 ky zone, AFAIK.

After all it's not the only part of the world where species extinction is closely associated with human arrival...Hmmm... like North American megafauna extinction... 10,000 after humans arrived and coincident withe the end of the Ice Age? Your evidence isn only a maybe, a hypothesis is not proof.

So C4 and C6 (or their respective ancestors) must have separated from the core C Y-hap by at least 46,000 years ago. If all six C Y-haps separated around the same time then C1, C2, C3, and C5 must also have become separated by 46,000 years ago, which fits your original diagram. Which one? This one makes them diverge c. 60 kya, 20ky before actuale evidence of anything.

But the second one may in fact be more realistic in some senses. Because, you know, women have spoken and they say that a short chronology seems more realistic from their viewpoint. What doesn't make any sense would be a bunch of males colonizing Eurasia... and women only arriving late when the process is already finsihed, somehow totally displacing the women that were before them. If that makes little sense with males, it makes much less sense with females, really.

I'm pretty sure it's SE Asia. If we look at the distribution of C*(xC1,2,3,4,5,6) we find it distributed around what, for much of ancient prehistory, would have been a virtually landlocked south China Sea, with an apparent outlier in India.Hmmm, makes some sense but you still seem to have high diversity of C in India, not just an "outlier". Plus everything points to a first expansion centered in South-SE Asia together not separatedly. As people came from East Africa, the beginning of that expansion surely happened in South Asia, with Eastern Eurasia being an offshot of it.

When people arrived to South Asia, they found opportunity to thrive and expanded: as simple as that. That expansion soon crossed the Ganges delta and extended via Eastern Eurasia, to the north an the south.

As the East was getting crowded, people began looking West and we do see a shift c. 52-48kya in the direction of migration, coincident with the R secondary explosion.

That's what mtDNA says very clearly. I'd rather discuss mtDNA these days: seems much more clear than Y-DNA. Obviously some geneticists had already done that before but we are blinded by our Patriarchalist background and try to decipher the Y-DNA. Fine. But it's not as neatly written as mtDNA.

terryt said...

"I'd rather discuss mtDNA these days: seems much more clear than Y-DNA".

It seems to me that the mtDNA is more complicated than the Y-chromosome though. Probably because, as you say, it's more ancient. It was in fact my study of both lines some years ago while looking at Polynesian origins that led me to believe that some ancient lines had moved across Central Asia.

"Fossils, including tools, are all in the 40 ky zone, AFAIK".

I'm pretty sure some C14 dates are also 40-50 k. I'll try to track them down. What time do you have mtDNA S in Australia separating? I can't remember just now and I'm in a bit of a hurry.

"What doesn't make any sense would be a bunch of males colonizing Eurasia... and women only arriving late when the process is already finsihed, somehow totally displacing the women that were before them".

It's not actually impossible if men were leading the expansion, and so breeding with women already there. We know from recent history that something similar has happened in regions heavily colonised by Europeans. Mind you, I'm not claiming that it happened in ancient times.

Maju said...

It seems to me that the mtDNA is more complicated than the Y-chromosome though. Probably because, as you say, it's more ancient.It's more diverse but in some aspects more simple than Y-DNA, because all subclades span regionally only from some node down. There's nothing like R1b, R1a, Q or E1b1b in mtDNA, spanning through several continents and/or huge continental regions. From the Southern Asian expansion process on, all expansions are localized in smaller contiguous areas - what makes total sense.

Once pioneer colonist women arrived to their destinations there was not much migration among them, certainly not long distance one (with very specific exceptions).

The basic mtDNA structure means:

1. A single original expansion zone (probably South Asia because of geography of dispersal and top level diversity).

2. An M-dominated expansive phase, with some N subclades alongside. This pulse moved mostly in Eastern direction (assuming India as origin).

3. An R-dominated expansive phase (with some M and N associated subclades). This pulse still affected the East (B, F and related clades) but actually marks a shift to a westward line of migration, that lingers for some longer time.

4. Regional expansion (events like the colonization of Europe, North Africa or North Asia notably).

These phases should be parallel to those of Y-DNA, so I think that phase 2 corresponds with F and D diversification, phase 3 with K, etc. I am unsure where to place C in this context: my estimates seems to make it expanding more in phase 3 or even 4 (but the data available is limited).

I'm pretty sure some C14 dates are also 40-50 k.40-42 kya in fact.

What time do you have mtDNA S in Australia separating?40 kya by the graphic in this thread. If all is adjusted to fit with mtDNA patterns (what is only logical to do), then it'd be more in the 30 kya area. This expansion anyhow would only reflect a regional phenomenon, because S hangs directly from F and therefore could have arrived to Asutalia a any intermediate time between when F and S nodes are estimated.

It's not actually impossible if men were leading the expansion, and so breeding with women already there.Re-read please. If they did that mtDNA would be locally older, just what I said. I argue that Y-DNA should not be older than (most) mtDNA anywhere (though the opposite is possible).

terryt said...

"This expansion anyhow would only reflect a regional phenomenon, because S hangs directly from F and therefore could have arrived to Asutalia a any intermediate time between when F and S nodes are estimated".

You're confusing mtDNA S with Y-chromosome S here. Mt-hap S is older than mt-hap F. The only thing they have in common is that they're both derived from N. Mt-hap S presumably arrived in Australia long before Y-hap S. The 40-30 k arrival would neatly fit other evidence for a second arrival that I covered in the essay "Into Australia".

Maju said...

Yah, true.

MtDNA S is (AFAIK) just one mutation away from N, so it's like 54kya. But again this is subject to adjustment as our knowledge improves. It's not rocket science, you know.